Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Integrated Dual Filter framework for forest management planning : a case study in the southwest Yukon Waeber, Patrick O. 2012

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
24-ubc_2012_spring_waeber_patrick.pdf [ 3.82MB ]
Metadata
JSON: 24-1.0072761.json
JSON-LD: 24-1.0072761-ld.json
RDF/XML (Pretty): 24-1.0072761-rdf.xml
RDF/JSON: 24-1.0072761-rdf.json
Turtle: 24-1.0072761-turtle.txt
N-Triples: 24-1.0072761-rdf-ntriples.txt
Original Record: 24-1.0072761-source.json
Full Text
24-1.0072761-fulltext.txt
Citation
24-1.0072761.ris

Full Text

 INTEGRATED DUAL FILTER FRAMEWORK FOR FOREST MANAGEMENT PLANNING: A CASE STUDY IN THE SOUTHWEST YUKON   by  Patrick O. Waeber  M.Sc., University of Zurich, 2002  A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF  DOCTOR OF PHILOSOPHY  in  THE FACULTY OF GRADUATE STUDIES (Forestry)  THE UNIVERSITY OF BRITISH COLUMBIA (Vancouver)  April 2012  © Patrick O. Waeber, 2012 ii  Abstract  Forest management planning is increasingly difficult due to the growing numbers of values and interests associated with forests, with climatic and socio-economic change adding further complexity. Forest planning hence can be described as a wicked problem. It cannot be tamed by linear deterministic approaches, but rather requires a more holistic approach. This study illustrated the development and testing of a planning and decision-support framework, the Integrated Dual Filter (IDF), for forest management. The IDF consists of two Filters, an Environmental and a Social Filter, and three States: Environmental representing the no- management landscape that can be used in planning as a ‗natural baseline‘ for the Desired State (an engineered landscape); the Management State represents a managed landscape to explore tactical aspects of planning and assess the feasibility of the Desired State. The IDF was tested using empirical data from the Champagne and Aishihik Traditional Territory (CATT), southwest Yukon. The CATT has recently experienced a spruce bark beetle outbreak affecting over 85% of its forests. The governance response was the development of the Strategic Forest Management Plan (SFMP). The SFMP was the starting point for the IDF-Social Filter to extract important values during working group meetings in the Yukon, deploying a ratings table to develop, judge and rank a hierarchy of criteria and sub-features for the Analytic Hierarchy Process. This resulted in five alternative forest management strategies for the long-term: manage forests for the timber industry; for multiple values and use; for fire risk reduction; for wildlife; and for the carbon industry. The IDF-Environmental Filter revealed that white spruce will still dominate the landscape after 200 years; that fire will remain an important disturbance to maintain a heterogeneous landscape; and that climate change will likely have no direct effect on tree species iii  in the overall landscape but will be important at the site level. The conceptual simplicity of the IDF makes it a valuable decision-making support system to identify system properties, constraints and concerns (by using the Filters) in order to simulate and project the planning landscape (using the States) into the future. It allows an extension of the existing planning horizon from 20 years for the CATT SFMP to >100 years; this will better inform management, especially under consideration of climate change.  iv  Preface  A version of Chapter 2 was submitted for publication in June 2011 titled ‗Filtering socially balanced forest management strategies for the Champagne and Aishihik Traditional Territory, southwest Yukon‘. The authors are P. O. Waeber, C. R. Nitschke, A. Le Ferrec, H. Harshaw and J. L. Innes. I developed the research design and framework of the manuscript, reviewed all the literature on the subject, conducted the working group meetings and did the analysis. I wrote most of the manuscript including discussion and conclusion. Craig Nitschke helped in the design and framework of this research, supported the analyses, and helped in the writing of the manuscript. Ambre Le Ferrec helped with the data management. Howie Harshaw helped in the manuscript final formulations. John Innes established the contacts with the Yukon working group and helped in the final formulations of the manuscript.  A version of Chapter 3 was submitted for publication in July 2011 titled ‗The sensitivity of tree species at the site and landscape level to disturbance and climate change in southwest Yukon, Canada‘. The authors are P. O. Waeber, C. R. Nitschke, R. Chavardès, S. Herrmann and J. L. Innes. I developed the research questions, designed the framework of the manuscript, reviewed all the literature on the subject and conducted all the modelling and testing. I wrote most of the manuscript including discussion and conclusion. Craig Nitschke helped with the modeling parameterization and calibration, and in the study design and refining of the manuscript writing. Raphaël Chavardès helped with the GIS, the database management, and the dendrochronology. Svenja Hermann did the kriging interpolations, and helped with the weather data organization. John Innes helped in the refinement of the manuscript writing. v  A version of Chapter 4 was submitted for publication in August 2011 titled ‗Integrated Dual Filter approach, a decision support tool for forest management‘. The authors are P. O. Waeber, C. R. Nitschke, R. Chavardès, H. Harshaw and J. L. Innes. I developed research questions, designed the framework of the manuscript, reviewed all the literature on the subject and conducted all the modeling and testing. I wrote all of the manuscript. Craig Nitschke helped with the modeling parameterization and calibration, and the study design. Raphaël Chavardès helped with the GIS mapping and the beetle and leaf miner calibrations. Howie Harshaw helped with the manuscript framework, and the manuscript final formulations. John Innes helped with the final drafting of the manuscript.  The research for Chapter 2 was approved by the UBC Behavioural Research Ethics Board. The Certificate Number of the Ethics Certificate is H08-02397 (Integrated Dual Filter Approach for Sustainable Forest Management in the Champagne and Aishihik Traditional Territory, Southwest Yukon).  vi  Table of contents Abstract .................................................................................................................................................... ii Preface ..................................................................................................................................................... iv Table of contents .................................................................................................................................... vi List of tables............................................................................................................................................ ix List of figures .......................................................................................................................................... xi Acknowledgments ................................................................................................................................. xii 1. Thesis introduction ......................................................................................................................... 1 1.1 Research context: Yukon and CATT strategic forest management planning ................... 2 1.2 Research objectives ................................................................................................................. 5 1.3 Research questions .................................................................................................................. 7 1.4 Research assumptions ............................................................................................................. 7 1.5 Coarse- and fine-filter concepts for forest management: Addressing the importance of spatial and temporal scales................................................................................................................. 9 1.6 Addressing increasing complexity in forest management ................................................. 13 1.7 Integrated Dual Filter: Filters, states, and an adaptive process ....................................... 18 1.8 Research methodologies ....................................................................................................... 21 1.8.1 Tools ................................................................................................................................ 22 1.8.2 Data types and sources ................................................................................................... 25 1.9 Research design ..................................................................................................................... 26 1.10 Outlook ................................................................................................................................... 29 2. Filtering socially balanced forest management strategies for the Champagne and Aishihik Traditional Territory, southwest Yukon ............................................................................................ 31 2.1 Introduction ........................................................................................................................... 31 2.2 Methodology .......................................................................................................................... 35 2.2.1 Yukon working group and ratings table .......................................................................... 36 2.3 Results .................................................................................................................................... 40 2.3.1 The AHP hierarchy structure .......................................................................................... 40 2.3.2 Characterizing the alternative forest management strategies ........................................ 40 2.3.3 The AHP judgments and priority relations ..................................................................... 42 2.4 Discussion ............................................................................................................................... 46 2.4.1 Strategic forest management planning............................................................................ 46 2.4.2 Alternative strategies and climate change ...................................................................... 47 2.4.3 Holistic approach for balancing ecology and economy.................................................. 52 vii  2.5 Conclusions ............................................................................................................................ 56 2.6 Outlook ................................................................................................................................... 58 3. The sensitivity of tree species at the site and landscape level to disturbance and climate change in southwest Yukon, Canada................................................................................................... 60 3.1 Introduction ........................................................................................................................... 60 3.1.1 Study region: Champagne and Aishihik Traditional Territory, southwest Yukon .......... 64 3.2 Methodology .......................................................................................................................... 68 3.2.1 Tree/site-level model: TACA ........................................................................................... 68 3.2.2 Spatially explicit landscape model: LANDIS-II .............................................................. 71 3.2.3 Historic weather data...................................................................................................... 72 3.2.4 Climate change projections ............................................................................................ 74 3.2.5 Soil data and calibration ................................................................................................. 75 3.2.6 Parameterizing the landscape model LANDIS-II ........................................................... 77 3.2.7 Factorial design and response variables ........................................................................ 81 3.3 Results .................................................................................................................................... 82 3.3.1 Climate and site effects on establishment ....................................................................... 82 3.3.2 Landscape-level effects of climate and fire ..................................................................... 88 3.4 Discussion ............................................................................................................................... 95 3.4.1 Site-level responses ......................................................................................................... 95 3.4.2 Landscape-scale responses: Fire, succession and climate change ................................. 98 3.5 Conclusions .......................................................................................................................... 103 4. Integrated Dual Filter approach for forest management planning – a synthesis .................. 106 4.1 Introduction ......................................................................................................................... 106 4.2 Description of the IDF decision support framework ....................................................... 111 4.2.1 The Dual Filters ............................................................................................................ 111 4.2.2 Forest State Space and its States .................................................................................. 112 4.2.3 The IDF iterative process ............................................................................................. 114 4.3 Application of the Integrated Dual Filter ......................................................................... 116 4.3.1 Social Filter .................................................................................................................. 116 4.3.2 Environmental Filter ..................................................................................................... 119 4.3.3 Engineering the Desired State ...................................................................................... 121 4.3.4 Management State and the adaptive cycling ................................................................. 123 4.3.5 Assessment of management actions at the landscape level ........................................... 124 viii  4.4 Results .................................................................................................................................. 127 4.4.1 Fire risk reduction – developing the Desired State....................................................... 127 4.4.2 Harvesting – the IDF adaptive iterative process .......................................................... 128 4.4.3 Environmental State and change over time .................................................................. 129 4.5 Discussion ............................................................................................................................. 136 4.6 Conclusions .......................................................................................................................... 145 5. General conclusions .................................................................................................................... 147 5.1 Integrated Dual Filter ......................................................................................................... 147 5.2 Limitations of the IDF-study .............................................................................................. 151 5.3 Key conclusions for the IDF ............................................................................................... 158 5.4 Future opportunities ........................................................................................................... 163 5.4.1 From further verification .............................................................................................. 163 5.4.2 … to validation .............................................................................................................. 164 5.4.3 …and accepting complexity .......................................................................................... 166 5.5 Contributions to management and forest management planning .................................. 166 Bibliography ........................................................................................................................................ 169 Appendices ........................................................................................................................................... 201 Appendix A: Summary of tools and data used in the Environmental Filter study ................... 201 Appendix B: 90 ecological research plots...................................................................................... 202 Appendix C: AHP calculations ...................................................................................................... 214 Appendix D: Soil clustering ........................................................................................................... 230 Appendix E: LANDIS-II modules and parameters ..................................................................... 231 Appendix F: Point fire ignitions .................................................................................................... 247 Appendix G: Seral stage distribution ............................................................................................ 248 Appendix H: Projected climate change CATT ............................................................................. 249 Appendix I: Harvesting distribution ............................................................................................. 250 Appendix J: Indicators listed in the SFMP................................................................................... 251 Appendix K: Calculation of indices used in this thesis ................................................................ 256  ix  List of tables  Table 2.1: Priorities of the five FMS for each AHP hierarchy element ............................................... 45 Table 3.1: Selected species parameters used in TACA .......................................................................... 70 Table 3.2: Observed F-values for the factorial ANOVA with corresponding P-values ...................... 83 Table 3.3: Mean and Standard Deviation of establishment scores for the three ecoregions .............. 85 Table 3.4: Observed F-values for the factorial ANOVA for the landscape ......................................... 89 Table 4.1: Criteria and indicators used in this study ........................................................................... 126 Table 4.2: Indices calculated for the Environmental State ................................................................. 135 Table B1: 90 ecological plots .................................................................................................................. 201 Table B2: Tree samples .......................................................................................................................... 202 Table B3: Identified plot-vegetation ...................................................................................................... 208 Table B4: Soil data .................................................................................................................................. 210 Table C1:  SFMP and corresponding AHP hierarchy elements ......................................................... 215 Table C2: Ratings table for SFM-experts ............................................................................................. 217 Table C3: From ratings to judgments ................................................................................................... 219 Table C4: Pair-wise comparison judgments ......................................................................................... 220 Table C5: Pair-wise comparison matrices ............................................................................................ 223 Table C6: Priorities and consistency checking ..................................................................................... 226 Table E1: LANDIS species parameters ................................................................................................ 230 Table E2: LANDIS ecoregions ............................................................................................................... 230 Table E3: LANDIS initial communities ................................................................................................ 231 Table E4: LANDIS Biomass per ecoregion .......................................................................................... 239 Table E5: LANDIS establishment probabilities (historic) .................................................................. 239 Table E6: LANDIS establishment probabilities (2020s)...................................................................... 240 Table E7: LANDIS establishment probabilities (2050s)...................................................................... 240 x  Table E8: LANDIS establishment probabilities (2080s)...................................................................... 241 Table E9: LANDIS harvesting scenario M3 ......................................................................................... 242 Table E10: LANDIS fuel types .............................................................................................................. 243 Table E11: LANDIS fire seasons ........................................................................................................... 244 Table E12: LANDIS fire region parameters ........................................................................................ 244 Table E13: LANDIS ageclass output ..................................................................................................... 245 Table E14: LANDIS forest types ........................................................................................................... 245 Table G1: Seral stage distribution ......................................................................................................... 247 Table J1: SFMP excerpt with goals, objectives and indicators .......................................................... 250 Table K1: Leafminer risk matrix .......................................................................................................... 257 Table K2: White spruce height and diameters ..................................................................................... 258  xi  List of figures  Figure 1.1: Integrated Dual Filter concept listing purposes and main tools used in this study ......... 28 Figure 2.1: AHP hierarchy ....................................................................................................................... 39 Figure 3.1: Temperatures and precipitations for Aishihik, Bison and Haines Junction weather stations ......................................................................................................................................... 66 Figure 3.2: Study landscape in the CATT, southwest Yukon ............................................................... 67 Figure 3.3: Ecoregions (A, B, H) and edaphic site types ....................................................................... 76 Figure 3.4: Current dominant forest types ............................................................................................. 79 Figure 3.5: Distribution of tree species ................................................................................................... 93 Figure 3.6: Distribution of dominant forest types over 200 years ........................................................ 94 Figure 3.7: Change in site moisture for Bison ecoregion ....................................................................... 94 Figure 4.1: Yukon fires from 1948-2004 ............................................................................................... 107 Figure 4.2: Integrated Dual Filter framework ..................................................................................... 115 Figure 4.3: Study landscape in the CATT showing areas of fuel treatments .................................... 118 Figure 4.4: Fuel treatment scenarios: Cumulative area burned over 200 years ............................... 128 Figure 4.5: Integrated Dual Filter adaptive cycling ............................................................................. 129 Figure 4.6: Normalized indices representing change compared to a baseline ................................... 134 Figure 5.1: Integrated Dual Filter framework ..................................................................................... 149 Figure A1: Summary of tools and data ................................................................................................. 200 Figure D1: Soil cluster analysis.............................................................................................................. 229 Figure F1: Digital elevation model CATT ............................................................................................ 246 Figure H1: Projected climate change CATT ........................................................................................ 248 Figure I1: Harvesting distribution for LANDIS modelling ................................................................ 249 xii  Acknowledgments  There is a long list of people I would like to share my gratitude with – without them this study would have never been realized. I owe a big thank you to my research supervisor John Innes and to my three research committee members Brad Hawkes, Nicholas Coops, and Rod Davis who endured me and my ideas through all these years. John Innes gave me the freedom to express myself in this research adventure from the very beginning. My committee gave me encouragement, challenges and guidance which enabled me to develop a fair understanding of the subject. Thank you very much. A special thank you goes to Craig Nitschke with whom I discussed my research back and forth exploring many dimensions. Craig taught me to enjoy the beauty of modeling, but even more so the beauty of sustainable forest management, and is an amazing buddy and field companion while crawling up and down the hilly slopes of the Yukon. The tree ring connection: Special thanks to Lori Daniels for introducing me to the tree rings and their secrets. A special big thank you I owe to Raphaël Chavardès who introduced me to GIS, tree rings, and Access database – but more so, to his wonderful petit monsieur Adam! The Yukon Team: The help and support of this team was crucial to the entire project. Finella Pescott was an inspiration in the lab and especially in the field. Marti Samis added with his photographing a special perspective to this research. ‗Merci mille fois‘ to the French connection: Jean-Sebastien Jacquet, whose enthusiasm and muscles helped unravelling many secrets below the Yukon FML-layers; Corentin Clément and Ambre Le Ferrec for their wealth of questions and discussions around landscape modeling and fire weather index calculations. ‗Herzlichen Dank‘ xiii  to the Germans: Svenja and Sonja Herrmann for the endless TACA-hours and weather making (I learned quite some ‗Schwaebisch‘ and I am a big fan of their ‗Kaese-Spaetzle‘!), and Hannes Fugmann for organizing the weather excel files with Matlab; a big ‗Dankeschoen‘ also to Julia Dordel for her help in the field, and for the many hikes together with her two dogs Luna and Tierra. Another big thank you is for Alyson McHugh for her field vigor and her untamed love for nature and wildlife. Jan Willem Klaassen is acknowledged for his endurance in soil and flower verification; Wendy Hirsh for her help with the dbf and excel data shuffling; Anne-Hélène Mathieu and Samuel Robinson for the Python code (breaker) relieving me from hundreds of hours of GIS roboting; Shyam Paudel for helping in the field. Thank you guys, I could not have done it without you! Around the corner: A warm thank you to Howie Harshaw, whose office is just around the corner and always with an open door – Howie was always very flexible and willing to listening to my extended explanations of forest states and filters, of AHP and decision making. Thank you for your understanding and patience. The local knowledge and network: The following people are acknowledged: John Trotta (Haines Junction Forestry office) for his availability when I had many questions regarding forest management in the area. Jean Paul Pinard for his invaluable insights into Haines Junction wind analyses and consultancy. Don Green (Yukon Government) for his support with weather data and consultancy. Aynslie Ogden (YTG Forestry Branch) for helping organizing the Yukon working group and providing me with Forestry Inventory Data and many other documents. The seven members of the Yukon working group are acknowledged for their availability and patience and enthusiasm in participating in this study. Special thanks go to Roger Brown (CAFN) who introduced me to the CATT and the CAFN forestry team. Together with him and Graham Boyd we visited all the field sites to learn about First Nations lands (the R-Blocks). A special thanks xiv  also to the CAFN who invited me and the Yukon team to celebrate the General Assembly in the beautiful Aishihik Valley – a fantastic experience! My office mates: A wonderful thank you is reserved for Guangyu Wang, Sang Seoup Lim and Lianzhen Xu with whom I had many discussions beyond the scope of sustainable forest management, and with whom I had excellent coffee-philosophy-breaks over the time of hundreds of espressos. The UBCs: I offer my enduring gratitude to the faculty and staff who were always very patient and supportive dealing with my many questions and requests. A thank you to the SFM lab members, who have inspired me to continue my work at UBC: Bogdan Strimbu, Garth Greskiw, Joleen Timko, Denise Allen, Judi Krzyzanowski, Angeline Gough, Reem Hajjar, Ajith Chandran, Monika Singh, Pano Skrivano, Suzi Malan, David Perez, Juan Chena, Tomoko Yoshida, Rebacca Klady, Harris Gilani, Futao Guo, and José Arias Bustamante. The soil connection: Thank you to Valery LeMay, Suzan Simard, Gary Bradfield with whom I was able to bounce ideas about soil and vegetation clustering, allowing me to build the edaphic inputs for the modeling. The LANDIS-team: Thank you to Rob Scheller, Brian Sturtevant, and Brian Miranda who evoked in me the pleasure of using LANDIS, a corner stone tool in my research, during the LANDIS 2010 workshop. The Yukoners: I would like to give a heartfelt thank you to Brent Liddle and Wenda Lithgoe for being excellent hosts and friends; and for teaching me the love and respect for the Yukon (―it is so full of biodiversity here‖); Pat Riley and Annette Sinclair for their encouragements during xv  coffee breaks, and especially Annette‘s unforgettable pies – a gem out in the wilderness. And Ginger, a Yukon spirit on four paws – I will always remember her! The Yukon, bigger than life, is acknowledged for its sincere and astonishing but unfinished beauty. Doing research in such a place on such a breathtaking landscape is a luxury I will never forget. This research would not have been possible without the funding support from the NSERC, the Yukon Territorial Government, and the UBC University Fellowship and Tuition Fee Awards. Lots of love to my family Carla, Waldi, Anna, Walti and Nadine for supporting me in all my ups and downs of that journey called Ph.D. and for sending endless packages of cheese, chocolate and ‗Nusstorte‘ over the big pond to Canada. Most of all, special thanks is reserved for Christine, my life partner for being as she is and accepting me as I am. The Ph.D. chapter is coming to an end. A new one is in its beginnings.  1  1. Thesis introduction  Forests are an invaluable ecological, economic, cultural and aesthetic resource on which the Earth‘s ‗homeostasis‘ relies. To manage forests properly, a profound knowledge of their status should be the basis of any planning. Managers should be able to recognize changes in the condition of forests and understand the implications of such changes in order to manage them adequately. There needs to be an understanding of changes and how to adapt them, including those caused by biotic and abiotic disturbances. This will involve conventional approaches, as well as allowing new management techniques operating at different spatial and temporal scales. A further and increasingly widely accepted external force impacting on forest systems is climate change, which is adding further uncertainty to the complexity of planning that already has to account for a growing number of interests and values in forest management. This thesis addresses the process of forest management planning, thereby touching on an array of challenges and topics related to sustainable forest management.  This introductory chapter presents the rationale for the research by setting the stage for the work, formulating the objectives and assumptions, and detailing the research questions examined in the thesis. The study design is presented and the techniques and tools deployed in this research are briefly introduced here.   2  1.1 Research context: Yukon and CATT strategic forest management planning  Boreal forests are expected to be amongst the ecosystems most affected by climate change (Chapin III et al. 2004, Solomon et al. 2007). The Champagne and Aishihik Traditional Territory (CATT), situated in southwest Yukon, Canada, has recently experienced an unprecedented spruce bark beetle outbreak (since the early 1990s), affecting about 380,000 ha of forests (Berg and Henry 2003, Berg et al. 2006, Yukon FHR 2009). The severity and extent has been causally linked to recent changes in climate within the CATT (ACIA 2004). In response to this epidemic, the Yukon Territorial Government and Champagne and Aishihik First Nations Government, together with the Alsek Renewable Resource Council (a government-independent body representing local communities‘ interests), engaged in a collaborative process to develop a Strategic Forest Management Plan (SFMP 2004) and Integrated Landscape Plan (ILP 2007) to ―… provide direction for sustainable forest management in the CATT. It is intended to provide a clear framework and practical guideline for forest managers and planners. … Generally they [= the CATT stakeholders] all express one common interest from a broad range of views: the desire for a ―balanced‖ approach to development. To most people in the region, this means an approach that balances maintaining a world-class wilderness environment with viable resource development opportunities …‖ (SFMP 2004:1).  3  The Strategic Forest Management Plan (SFMP) / Integrated Landscape Plan (ILP) is a guiding plan for the CATT. It provides guidance and a framework for practitioners in the CATT, and allows site-specific, short-term actions to be developed (e.g., 5–10 years), based on a hierarchical planning approach (SFMP-ILP-Harvesting Plan-Site Plan). For example, the Yukon government has allocated one million m 3  of the beetle-killed wood to be salvaged over the next 10 years (by 2016). To date this tenure has not been taken up and a local wood products mill (Dimok Timber) is currently harvesting an initial two to three thousand m 3  of spruce beetle killed wood. This management action primarily constitutes mitigation of climate change impacts, with the SFMP considered a reactive-indirect form of adaptation (Ogden and Innes 2008). The plan does not directly address adaptation but through its foundation in Goals/Objectives/Indicators it is believed that the plan will allow for adaptation to climate change. Although the majority of the SFMP is focused on short-term salvage harvesting there has been an increase in fuel-abatement treatments around the communities of Haines Junction and Canyon, which have been undertaken to reduce fire risk and improve community safety. The fuel management actions constitute proactive and direct adaptation to both current and future fire risk.  The Alsek Renewable Resource Council (ARRC) was created in 1995 after the Champagne and Aishihik First Nations (CAFN) Final Agreement was signed. The Final Agreement created the ARRC as the primary instrument for local renewable resources management in the region. The ARRC constitutes a voice for local community members in the management of the CATT‘s forests fish and wildlife populations (<http://www.yfwmb.ca/rrc/alsek>). The ARRC council consists of six members, half being from First Nations (CAFN representatives) and the other half being Yukon Territorial Government (YTG) representatives; all members are also locals. The 4  ARRC represents the community in resource planning processes that engage with the two governments, the YTG and the CAFN. In the Letter of Understanding signed in 1998 between four parties, the INAC (Indian Northern Affairs Canada), the CAFN, the YTG and the ARRC commits the parties to work together in the development of forest management plans and strategies. So far, the ARRC has helped the CAFN and YTG in the development of the SFMP (2004) and the follow up document Integrated Landscape Plan (ILP 2007). The ARRC is a guarantee that local, and especially FN concerns are integrated in planning documents such as the SFMP and ILP, and that their implementation is in line with local community desires.  The CATT context provides an excellent research opportunity for several reasons:  The two governments (i.e., Champagne and Aishihik First Nations Government, Yukon Territorial Government) are undertaking collaborative forest management, with the SFMP being one of the first major outcomes of this collaboration.  Climate change and the condition of the forest resource has resulted in a beetle epidemic affecting over 80% of the region‘s forests.  The current spruce beetle outbreak has affected the societal perception of forests in this region, including elevated fire risk and the future of the forest economy.  Although the SFMP is a government-guided forest management plan, it constitutes a bottom-up community-based compendium of regional values and concerns.   5  1.2 Research objectives  Forest management and planning is becoming increasingly difficult under conditions of uncertainty. The purpose of this study is to develop and test a new decision support framework, the Integrated Dual Filter (IDF; for a concept description see section 1.7) approach within the Champagne and Aishihik Traditional Territory context, and within its Strategic Forest Management Plan (SFMP)-framework for determining and implementing long-term sustainable forest management.  The research approach follows a global trend for an integrated qualitative and quantitative approach combining what Douthwaite et al. (2001) refer to as ―hard‖ and ―soft‖ science. For this, an array of (already developed and validated) tools are parameterized and calibrated to the natural and socio-cultural situation of the CATT. This integrated natural resource management research aims to test the proposed decision support system and its concept (see section 1.7) by combining environmental and social empirical data from the CATT, and by following the research questions stated in section 1.3. This combined should make the testing of the new forest management decision support and planning framework more academically robust and regionally meaningful. According to Lovell et al. (2002), resource management studies with a strategic scope (e.g., addressing larger spatial and temporal scales) that are able to increase the knowledge and understanding of a system may ease the development and formulation of policies. Lessard (1998) recommends the inclusion of two critical components into an adaptive framework: (i) To explore the Desired Future Conditions (DFC) of an ecosystem by reconciling values beyond the 6  biological scope, i.e., including socio-economic values; and (ii) the building of the context or program of management implementations to achieve identified and formulated DFCs. As a result, the following planning components are addressed and explored to varying extents in this research to develop and test the feasibility and applicability of the Integrated Dual Filter (or IDF) approach in the Champagne and Aishihik Traditional Territory (CATT):  Spatial scales: stand and landscape  Temporal scale: long-term, e.g., 100+ years  Ecosystem: boreal forest  Social system: communities of the CATT  Disturbances: fire, bark beetle, harvesting  External system drivers: climate change  Planning framework: SFMP/ILP CATT  Management context: experts and practitioners from the Yukon  The overall objectives of this research are to develop and test the new planning framework (Integrated Dual Filter framework, IDF), which allows to (i) integrate social and environmental values, (ii) addresses different temporal and spatial scales, (iii) incorporates climate change, natural disturbances and their consequences into forest management planning, and is (iv) able to be flexible under changing strategic directions and tactical implementations as needed.   7  1.3 Research questions  To test the concept described in section 1.7, I address the following working questions:  Social Filter; developing socially embedded alternative strategic directions on forest management: What is the desired long-term forest strategy for the CATT?  Environmental Filter; assessing possible systemic reactions to climate change: What trajectory will forest succession take over the next 200 years?  Desired State; does a threshold (e.g., minimum number of hectares that need to be treated to reduce area burned) at the landscape scale exist for fuel treatment (e.g., stand conversion from coniferous to aspen)?  Management State; illustrating the ‗adaptive cycling‘ of the IDF. Is intensified fuel treatment (e.g., during a shorter time period) more effective in reducing area burned than less intensified treatments?   1.4 Research assumptions  A research problem is built on research principles that in turn are based on a range of assumptions. They therefore provide the means for placing and framing the research within particular bounds. This thesis was developed under, and adheres to, the following assumptions: 8  Understanding eases decision making. We know that we have incomplete knowledge of a system, and that we have to accept that we do not understand it in depth. By increasing our knowledge and understanding we are able to make better and hence more robust decisions. Choosing constructivism rather than reductionism increases understanding. Real life cannot be described by one single discipline, or by reducing it to one single strategy (Born and Sonzogni 1995). Reality is formed by a multitude of constructions, each based on its own values and knowledge system (Guba 1990). This is particularly the case when uncertainty is high and a multiple scenario approach is more suitable for describing such complexity and increasing its understanding. Increased participation in planning helps plans to be more resilient to mistakes. Socially embedded desired forest management strategies and tactics help to satisfy communities dependent on forests. It also helps in situations where a previous decision made by forest managers was wrong (e.g., Lal et al. 2002). Change is real and continuous. We know that climate change is happening. We also know that economic change is happening. We are aware that ‗change‘ is taking place continuously but with changing (in the near future most likely at increased) pace (Solomon et al. 2007), and we therefore need to take change into account in forest management planning. Ignoring ‗change‘ is like deliberately investing in mismanagement. Our brain capacity can be enhanced. We know that the human brain is able to grasp and process only a limited amount of data at a time. Therefore, we should accept the help of machines able to process data several magnitudes higher/faster than ourselves. However, decision making is beyond a machines‘ capability, and a human brain in contrast is creative and flexible. 9  Consequently, brain and machine together can be more successful in extending a planning time frame from e.g., 20 to over 100 years. Simplicity helps address complexity. It is easy to ―goropise‖, i.e., making things more complicated as they seem (Leibnitz, in Feuer 1957). Or, as John Gall (<http://www.quoteland.com/qldb/author/59>) states: ―A complex system that works is invariably found to have evolved from a simple system that worked.‖   1.5 Coarse- and fine-filter concepts for forest management: Addressing the importance of spatial and temporal scales  The development of the coarse- and fine-filter concepts for managing ecosystems is a consequence of the hierarchical structure of ecosystems (Hunter 1991). Ecological processes are organized in hierarchies of processes that function at different levels (Urban et al. 1987, Wu and Loucks 1995, Holling 2001). Sustainable forest management operates within different levels of this organization, with different management goals and objectives typically defined at each level (Lindenmayer and Franklin 2002). The terms coarse- and fine-filter are used to implement scale; they are used in the context of management actions (or tactics) and systems and also in the context of measuring and monitoring the success of sustainable management. The choice of indicators for evaluating management actions reflects coarse- to fine-scale metrics.  10  Coarse-filter metrics focus on landscape elements such as ecosystem representation and seral stage distribution and typically rely on focal species to measure the impact of management. The coarse-filter approach focuses on managing ecosystem diversity and structure, with the objective of protecting representative areas of each ecosystem and seral stage in order to provide landscape-level heterogeneity (Kremsater et al. 2003). By preserving representative examples of ecosystems in a variety of seral stages, we assume that we will maintain the habitat structural elements, their spatial arrangements, and the inherent processes that determine biodiversity. The maintenance of patches of disturbance refugia, wintering habitat for species, late successional forests and riparian ecosystems are considered important elements of landscape-level coarse- filter management that should be encompassed by this filter (Lindenmayer and Franklin 2002). This approach therefore seeks to provide function and structure at the landscape-level by providing a broad range of habitats for a broad range of species (Noss 1987, Hunter 1991).  The fine-filter approach is a stand-specific approach and uses stand-level structural elements as surrogates for biodiversity (Kremsater et al. 2003). This approach aims to maintain key habitat elements such as large living trees, dead and dying trees, shrubs, hardwood tree species, coarse woody debris, and representative examples of edaphic communities. Fine-filter management maintains structural elements at the stand-level that will continually provide functionality for specific species or groups of species in defined spatial locations that in turn by scaling up will not impinge on the ability of the landscape to maintain its current state of biodiversity.  11  Identifying the historical spatial and temporal variation of ecological processes can contribute to our understanding of the processes that define ecosystems (e.g., Foster et al. 1998, Barber et al. 2000). Natural or historic variability is a key attribute of ecosystems, as well as a practical foundation for landscape scale (coarse-filter) management (Landres et al. 1999). Natural and historic variability are concepts that attempt to increase our understanding of ecological organization and functionality of an area, including landscape-scale effects of disturbances. This understanding may enable managers to mimic past and current variability and emulate these patterns in order to make existing and future conditions more heterogeneous and possibly maintain ecosystem health and vitality (Lertzman et al. 1997, Landres et al. 1999). The concepts can also be useful for setting management goals and objectives and give rise to the use of the natural variability approach (also called the natural range of variability) in forest management. This approach assumes that, for example, biodiversity at different levels of ecological organization will be preserved if the ecological structures and functions of forests are preserved (Hunter et al. 1988, Hunter 1999, Landres et al. 1999). Heterogeneity at many scales is generally accepted as an important concept for biodiversity (Pickett et al. 1997), since heterogeneity assures variation in resource availability (Lindenmayer and Franklin 2002). Historic conditions thus act as a coarse-filter from which management objectives are based on sustaining biodiversity at the landscape-level (Hunter 1990). The strength of the natural variability approach is that by approximating historic conditions it provides an understanding that can help us predict and reduce impacts to present-day ecosystems (Kaufmann et al. 1994). Two weaknesses of this approach are: 1) that it focuses at the higher levels of the ecosystem hierarchy where causality is much more complex than at lower levels; as a result, our understanding declines as the number of determinants increases, shifting our knowledge of the 12  system away from causal relationships to random, unpredictable outcomes; and, 2) it is based on historic and/or current assumptions that future conditions will equal current or past conditions (Bunge 1959, Kimmins 2004).  The management of forests based on historic range of variation is widely considered and used (Perera and Buse 2004). For example, in the province of British Columbia, Canada, a biodiversity guidebook for managing forest ecosystems based on natural disturbance patterns was developed so that landscape-level processes could be mimicked through the emulation of natural disturbance (Andison and Marshall 1999). In Ontario, Canada, the policy of natural disturbance emulation has led to the development of the Forest Management Guide for Natural Disturbance Pattern Emulation (McNicol and Baker 2004). In many Scandinavian countries, the use of natural forests and landscapes as templates for conservation or restoration of forest biodiversity in managed forests has been applied widely (Angelstam 1998, Fries et al. 1998). The Scandinavian approach is to link natural process at both the stand- and landscape-level where natural disturbance management attempts to maintain stand-level structures and landscape-level patterns that reflect the stand types and successional stages that develop over time synchronously with disturbance agents. In Sweden, the ASIO model has been developed to achieve desired stand-level and landscape-level structural complexity (Angelstam 1998). The ASIO model spatially allocates management zones based on the frequency of natural disturbances on the landscape and then prescribes different coarse-filter measures that are applied at the stand-level.   13  1.6 Addressing increasing complexity in forest management  A paradigm shift from purely utilitarian management towards multi-purpose based management is underway (cf. Bengston 1994, Mendoza and Vanclay 2008). Ecosystem management, or integrated natural resource management, is promoting and favoring ―management strategies that achieve some future desired state over strategies that produce some desired mix of resource outputs over time‖ (Weintraub and Bare 1996). A steadily growing dimension in natural resource management, including forestry, is the inclusion and recognition of social and cultural values. This dimension is becoming increasingly important in forest management planning (Bengston 1994). Sustainable forest management (SFM) calls for the balancing of diverse ecological, economic, and social values, usually represented in the form of multiple criteria and indicators (e.g., CIFOR 1999, CCFM 2003,) that often express conflicting management objectives. Accordingly, new forest management approaches and practices have been designed to reduce or possibly resolve potential conflicts of interests (e.g., Kimmins 1999 and 2002, Pommerening and Murphy 2004). Also, an increasing number of natural resource management approaches try to incorporate and address elements and principles that entail multiple disciplines spanning various spatial and temporal scales, and involving a multitude of stakeholders representing a growing array of values and interests in the planning and implementation process (Born and Sonzogni 1995, Holling et al. 1998, Lal et al. 2001, Mitchell and Beese 2002).  In the following paragraphs, I introduce the fields of forest management planning and decision making. Forest management must deal with growing complexity, with an increasing 14  number of expectations of the forest expressed by a growing number of values, an increasing number of stakeholders involved in the use of forest goods and services, and increasing risk and uncertainty due to changing frequencies or magnitudes of external forces such as climatic, economic or technological changes. This is requiring increasingly sophisticated ways to choose alternative strategies. Decision making involves choices between distinct alternatives (Kangas et al. 2008). Forest management planning is central to decision making (Kangas et al. 2008), and is based on three elements: alternatives, information, and preferences (Bradshaw and Boose 1990). According to Belton and Stewart (2002), decision-aid or support processes consist of three phases: problem structuring, model development, and using the model to inform decision makers and challenge alternative thinking. This approach emphasizes the importance of the ‗help in thinking‘ rather than the provision of ready-made solutions. As forest decision making includes a growing number of stakeholders and is being confronted with increasing external ‗shocks‘ (e.g., economic, climate change), decision making is increasingly becoming multi-dimensional and has to be practised under increased uncertainty. In other words, there may be uncertainty in all the parameters of a decision problem, and the future outcomes and consequences of some or all of these parameters and the decisions taken by the decision makers may be uncertain (Kangas et al. 2008). Consequently, decision problems are increasingly related to risk and uncertainty. A decision problem can be represented by three components (Kangas 1994): decision alternatives, states of nature, and consequences of actions. The consequences of actions are determined by the type of action, which is pre-determined by the decision alternatives and by the external forces or factors (characterizing a state of nature) that are beyond the decision makers‘ control (Kangas et al. 2008). Under risk, the probabilities of a state of nature as well as the distributions of 15  probabilities of the consequences of the actions are known; if they are unknown, then uncertainty prevails (Ananda and Herath 2003).  Scenario-based approaches have been used to address and deal with complexity and uncertainty and are particularly useful when used in the context of non-linear and paradigmatic changes. Scenario planning is a technique that has been used for several purposes (Lindgren and Bandhold 2003): for planning reasons, with the aim of developing practical results; as a guide for the development and filtering of ideas for projects or management; and for the evaluation of existing or new concepts and strategies. The greater the planning time frame, i.e., the further into the future we look, the greater the number of possibilities or possible futures, with some of the alternatives being more probable than others. The most desirable future may not even overlap with possible futures (Lindgren and Bandhold 2003). According to the same authors, the near future bears more certainty and less risk, and forecasts are powerful in these realms since they are based on certain relationship and deal with probable futures. Hence, decisions can be taken under greater certainty and reduced risk. However, with expanding planning time frames, structural stability is replaced by complexity and certainty by growing uncertainty. The reduction to a meaningful number of usually 5–10 plausible or possible alternative futures containing the most relevant dimensions of uncertainty is required for scenario planning. Another way of structuring a complex problem such as forest management planning is to use a hierarchical approach. Usually, hierarchical structures have been organized into three levels: the strategic level, the tactical level, and an operational level or unit (Sessions and Bettinger 2001, Jeakins et al. 2006). At the first level, basic ideas of what is wanted are identified and formulated for the longer run (e.g., usually more than 20 years). The tactical level is used to define and describe the 16  proximate mechanisms on how these goals are achieved (e.g., usually 5–10 years), and the operational level details the local actions on the ground (e.g., less than 1–5 years).  Forest management planning has a long history, focusing initially on the management of even-aged forest stands (Kilkki 1987, in Kangas et al. 2008). Tabular and mathematical functions were deployed in forestry to depict the relationships between for example stand volume per unit area, relative age and site quality (Buongiorno and Gilles 2003). A recurring theme in operations research, and which found its way into forestry in the 1960s, is to seek the optimal solution, i.e., finding the best solution amongst a possibly endless number of solutions (Buongiorno and Gilles 2003). Linear Programming was introduced to address large area problems using complex models to assess different forest developments under different scenarios, such as FORPLAN (Johnson 1986) and WOODSTOCK (Walters et al. 1999), developed in the USA and Canada, respectively, or AVVIRK-2000 in Norway (Eid and Hobbelstad 2000). Linear Programming is based on decision variables. There are three categories of methods to define decision variables in LP; Model I, II, and III (Bettinger et al. 2009): In a Model I LP problem, decision variables are used to keep track of the history of a stand during the entire planning time-frame. A Model II LP problem is utilized to track higher time resolutions, e.g., changes during a single harvesting intervention only. Both models are widely used in forest management planning problems. Model III stratifies stands of the same age class and keeps track of them throughout the entire planning period. Models based on LP have been criticized because they fail to adequately address non- linearity and uncertainty (Pukkala 2002), or because they have the inability to incorporate spatial problems (Baskent and Keles 2005). Simulation approaches have been developed that allow for stand development predictions by sequentially projecting forest conditions. However, they are 17  less suitable for dealing with multi-objective decision problems, where potential conflicting objectives are competing and where oftentimes optimal solutions are desirable (Baskent 2001, Baskent and Keles 2005). When the type of value assigned to a so called decision variable requires refinement, or when non-linear relationships have to be accounted for, or both, heuristic approaches are more suitable to address complex and large-scale decision and planning problems. According to Bettinger et al. (2009), a heuristic method is based on a combination of logic and rule of thumb approach to reach an efficient solution to a complex planning problem. Heuristics do not guarantee the optimal solution, but in combination with LP can address complexity and search for near-optimal solutions. Heuristic optimization techniques such as Monte Carlo integer programming (Nelson and Brodie 1990), Simulated Annealing (Lockwood and Moore 1993), Tabu Search programming (Murray and Church 1995), or combinations of these techniques have been developed (e.g., Nelson et al. 1991, Baskent and Keles 2005). These enable several decision problems to be addressed at a time. Multi-Criteria Decision Making (MCDM) presents a framework for analysing complex multi-objective decision problems (such as forest management issues). Such MCDMs usually define objectives, identify criteria to measure and assess the stated objectives, help the development of decision alternatives and assign weights to the decision criteria regarding their relative importance, rank and ability to choose decision alternatives (Ananda and Herath 2009). SMART (Simple Multi-Attribute Rating Technique) is a decision support method developed in the early 1970‘s (von Winterfeldt and Edwards 1986) which allows users to directly assign weights to decision criteria; a similar technique is the AHP (Analytic Hierarchy Process) developed by Saaty (1977, 2001) which is one of the most applied decision support techniques in natural resources management (Schmoldt et al. 2001). Mendoza and Martins (2006) in their review list more than 55 different applications 18  of MCDMs. Many of today‘s complex forest management planning challenges can be addressed and alleviated using Multi Criteria Decision Support (MCDS) or Aid (MCDA) methods, each with its own set of characteristics and techniques (e.g., for a classification of MCDAs see Diaz- Balteiro and Romero 2008) for a specific decision situation and planning context (Kangas and Kangas 2005).   1.7 Integrated Dual Filter: Filters, states, and an adaptive process  In this section I will briefly introduce the three concepts that are fundamental to the IDF framework: The Dual Filters, the Forest State Space, and the IDF ‗adaptive cycling‘.  Dual Filters: One of the main reasons for the growing complexity of forest management is the number of values that must be managed for. I will introduce two filters that allow summarizing and grouping many of these values: The Environmental Filter, representing all environmental values, and the Social Filter, representing all socio-cultural/economic values. The term ―filter‖ is based on the classic management approach with the ―coarse‖ and the ―fine‖ filters referring to spatial scales like landscape and stand scale (discussed in section 1.5). Both filters can encompass spatial and temporal scales. The Environmental and Social Filters implicitly incorporate different scales, i.e., coarse and finer scales. For example, the Environmental Filter can address single trees, a forested stand, or an entire landscape; the Social Filter can address for example temporal scales such as short-term tactics or long-term strategies, or spatial scales such 19  as single household income (finer scale) versus regional income (coarser scale). As a result, the Environmental and Social Filters reduce the planning complexity to only two dimensions; this allows a forest manager to focus on one or the other dimension at a time and eases the identification of important values and drivers, e.g., concerns and desires (Social Filter), or vulnerabilities (Environmental Filter).  Forest State Space: In the Integrated Dual Filter (IDF) planning approach the Forest State Space F, a new concept based on the definition of an ecosystem as an‘ n-dimensional space‘, describes specific forest states, depending on the perspective one is looking at a forest. The IDF approach uses three different forest states to describe the forest state space F: E = Environmental Forest State, representing the biophysical forest system; D = Desired Forest State, reflecting a socially desired forest state; M = Management Forest State, representing a managed forest system. Each state of the forest state space F can be represented as a landscape map at a given time t where all its components (or data layers) are merged. For example, the E (environmental) State is like a biophysical landscape (i.e., during the planning process no management is simulated on the landscape); the D (desired) State is like an ‗engineered‘ or ‗designed‘ landscape (i.e., it is designed to represent the social requirements/desires by answering the question of ‗what do we want‘?); the M State focus on the proximate level of planning and is like a ‗managed‘ landscape where the emphasize of planning is on management tactics and actions (by answering the question of ‗how do we do it?‘). Each of the three states enables planning to focus on one state at a time, increasing the knowledge and understanding of the planning landscape.  20  Adaptive management is a systematic approach and iterative process that allows management to proceed in a complex, uncertain biological and socioeconomic environment across different temporal and spatial scales. However, it requires continuous learning, an iterative evaluation of goals and approaches, and redirection based on an increasing information base and changing public expectations (Baskerville 1985). Or, as Murray and Marmorek (2003) put it, adaptive management is a systematic approach that involves the exploration of alternative scenarios to meet management objectives through anticipation of potential realities. Adaptive management hence requires explicit hypotheses about system structure, composition and function, a clear statement of management goals, defined management actions to be implemented and recognition of anticipated ecosystem responses (Holling 1978, Walters and Holling 1990, Holling 2001).  The Integrated Dual Filter approach adheres to the principles of adaptive management, with the IDF-iterative process consisting of two running modes: a ‗real world‘ mode and a ‗virtual‘ mode. The former consists of on the ground tactics and operations and monitoring at the stand/landscape scales. The virtual (or actual planning) mode allows for scenario-based studies (cf. Lindgren and Bandhold 2003) of the relations between climate change, operations/disturbances and their long-term effects. The virtual mode enables a manager to learn and acquire understanding of long-term consequences of management decisions, e.g., according to Shifley et al. (2000:8) ―even the decision to do nothing has consequences on the forest system‖. This scenario-based approach helps to assess which of the scenarios are favoring certain management approaches over others.  21  IDF ‘adaptive cycling’: The IDF adaptive iterative process is the study of how to bring the Environmental State closer to the Desired State through management (i.e., through the formulation of an M State). For this, each cycle starts from the E State by implementing some management tactics and comparing the M State with the D State. Each further management modification requires additional comparison between D and (a new) M. This ‗adaptive cycling‘ ends when the M ~ D (i.e., within the F state space, the E State is been transposed onto the D State).   1.8 Research methodologies  According to Mingers and Brocklesby (1997), a methodology is a framework of guidelines or activities to assist people or agents in research or other activities such as management planning. Methodologies are generally developed within a paradigm (i.e., a set of philosophical assumptions characterizing possible research and actions) and can embody a set of techniques as for example in operational research; these techniques refer to specific activities, such as the development of simulation models or the application of a statistical analysis, e.g., ANOVA (ANalysis Of VAriance). A tool refers to the performance of a particular technique (Mingers and Brocklesby (1997); examples include the use of a Linear Programming optimizer (e.g., Buongiorno and Gilles 2003) or the applications of a mechanistic model such as TACA (Tree And Climate Assessment, Nitschke and Innes 2008). In this thesis, the development of a decision support system or framework (the IDF) is based on a set of techniques such as ecological 22  modelling, statistical analyses, and systems identification. These in turn are based on a set of already established and widely applied qualitative and quantitative tools.  1.8.1 Tools The current spruce beetle outbreak in the southwest Yukon has impacted the societal perception of forests, including elevated fire risk and the future of a forest-based economy. As a governance response to these dramatic regional changes, the Strategic Forest Management Plan (SFMP, 2004) has been developed as a guidance and framework for practitioners in the CATT. Ogden and Innes (2008) classified the SFMP as a ―reactive-indirect‖ plan, which does not explicitly identify the impacts of climate change on ecosystems and communities, nor does it prescribe management actions that will reduce such impacts. Species‘ vulnerability and their responses to climate change constitute major challenges to forest managers when planning for the implementation of strategies and tactics for adaptive forest ecosystems management under climate change. Understanding and possibly anticipating potential responses in terms of species composition and abundance under possible growing sub-optimal conditions will allow forest management to make informed and pro-active decisions. These arguments are guiding the choice for the tools to be deployed in the IDF framework. The proposed framework of the Integrated Dual Filter approach (IDF) is composed of four tools, (i) an Ecological Modeling Tool Kit using a set of ecological simulation models to perform forest succession simulations by varying input parameters such as harvesting operations or natural disturbances such as beetle or fire under climate change, (ii) a widely used decision-making tool, i.e., the Analytic Hierarchy Process (AHP) of Saaty (1977, 2001), which is used to balance certain values to reach sustainable forest 23  management in the CATT, (iii) a database, and (iv) a monitoring tool (a set of Criteria and Indicators) to measure management effects and impacts.  To test the IDF, a set of qualitative and quantitative tools has been chosen, parameterized and calibrated to the CATT boreal context. In the following, the tools used to create the forest states are briefly characterized (for more details see Chapters 2-4, and for general and IDF case- specific limitations see Chapter 5.2): The D State reflects goals and objectives for sustainable forest management. More specifically, a combination of tools such as AHP (Analytical Hierarchy Process, Saaty 2001), a Yukon working group and a ratings table have been used to characterize the alternative D States. The regional socio-economic and ecological values were derived from the SFMP (the Strategic Forest Management Plan of the CATT) and reorganized into an AHP structure, with the bottom level of the AHP hierarchy representing different potential desired D States. The Yukon working group, consisting of forest experts and practitioners from the Yukon (i) helped to define/characterize a set of alternative forest management strategies (= potential desired D States at the bottom level of the AHP), and (ii) advised on the development of a hierarchy of constituent elements (sub-criteria) describing and characterizing the two potentially conflicting objectives, Functioning Forest Ecosystems (representing the ecological value), and Community Economic Sustainability and Benefits (representing the socio-economic value), both decision criteria representing two out of four goals from the SFMP (SFMP 2004). An expert group in sustainable forest management from UBC then weighted and rated individual components of the potential D States (the alternative forest management strategies). The AHP technique then calculated the 24  highest relative priority forest management strategy for the IDF approach to represent the D State in the IDF planning approach.  The E State represents the environmental conditions at a given time t. To describe the current environmental conditions of the CATT I used empirical data (e.g., raw data derived from field work, GIS thematic layers), using locally tailored monitoring protocols (i.e., Yukon Forestry Monitoring Program 2008). These data were then used for parameterizing and calibrating the LANDIS-II (LANdscape, DIsturbances, Succession) model, which is a spatial explicit landscape eco-model that enables the simulation of forest succession under different natural (e.g., fire) and anthropogenic disturbances (e.g., harvesting operations for the M States) on large spatial and temporal scales (Mladenoff 2004, Scheller et al. 2007). TACA (Tree And Climate Assessment) is a mechanistic aspatial model (Nitschke and Innes 2008) that analyzes the response of trees in their regeneration niche to climate-driven phenological and biophysical variables. The Canadian Forest Fire Danger Rating System (CFFDRS) (vanWagner 1987, Stocks et al. 1989) is a modular fire danger rating system with two primary subsystems – the Canadian Forest Fire Weather Index (FWI) System and the Canadian Forest Fire Behavior Prediction (FBP) System. The FWI System is calculated from weather observations. The FBP System uses FWI indices to predict fire behavior. Both TACA and CFFDRS outputs serve as libraries for LANDIS, by creating species- specific establishment coefficients, and expected fire spread and intensities. The combination of LANDIS, with TACA and CFFDRS was chosen to simultaneously model vegetation dynamics under different natural disturbances and climate change. The different models are introduced in Chapter 3. An overview of all the tools and data used in this component of the study are summarized in a flow chart in Appendix A. 25   1.8.2 Data types and sources The IDF approach has been developed and tested using qualitative and quantitative empirical data and information from the CATT region:  GIS layers on forest data inventory (Government of the Yukon‘s Forest Inventory Database, Government of the Yukon 2004);  Six research blocks containing 90 ecological plots were established in the CATT to sample information on soil, vegetation, tree age, stand structure. Details are presented in Chapter 3 and Appendix B.  Climate data from two weather stations established in 2008 in the Aishihik valley (southwest Yukon) as well as from existing regional stations. These data are used for parameterization and calibration of the model toolkit;  Climate data for Beavercreek, Haines Junction, and Whitehorse Airport weather stations (from Environment Canada);  Fire occurrence data from 1943–2004 (Yukon fire management branch);  Social judgment and ranking inputs from Yukon working group members and UBC expert group members, which feeds into the AHP.  SFMP/ILP documents were used to gather the values (criteria and sub-criteria) for subsequently analyzing and balancing of the potentially contradicting/competing values listed in this document (which is realized by the AHP technique).   26  1.9  Research design  According to Kangas et al. (2008), decision making can be descriptive or prescriptive. The former analyzes how people or agents make decisions without aid. The latter analyzes the proximate levels, i.e., how a decision should be made. In the approach described in this thesis, methods are used to aid people or agents in making a decision. My research is grounded in this decision making approach. In order to develop and test the IDF-decision support framework, this research follows elements of the structured decision-making approach described by Ohlson et al. (2005): (i) Problem identification; (ii) setting of management objectives; (iii) assessment of vulnerabilities; (iv) development of management alternatives; (v) evaluation of alternatives; (vi) decisions; and (vii) implementation. Figure 1.1 depicts a summary of the study design highlighting the prominent features of Chapters 2–4. Each chapter is briefly summarized below, emphasizing which part of the new IDF approach is introduced and analyzed where. The square brackets indicate which of Ohlson et al.‘s (2005) elements are addressed in the chapter.  Chapter 2: The purpose of this chapter is to present the Social Filter, and to inform the potential D States (i.e., strategic direction/emphasis, goals and objectives) of the Integrated Dual Filter approach. For this, the potential conflict between two central values of the Strategic Forest Management Plan for the CATT (SFMP 2004), the ‗ecological‘ and ‗economic‘ values, is addressed [Ohlson et al. (2005): (i) problem identification]. Then, a group of forest experts and practitioners from the Yukon developed and characterized alternative forest management strategies [Ohlson et al. (2005): (ii) management objectives, and (iv) alternatives]; the Analytic 27  Hierarchy Process was used to determine which of these alternatives best balanced the conflicting values ecology and economy [Ohlson et al. (2005): (vi) decision (of the socially desired forest management strategy to pursue)].  Chapter 3: This chapter explores the environmental conditions of the study area to represent the Environmental Filter of the IDF approach, and to inform a baseline landscape (the Environmental State) for further IDF planning. For this, a modelling approach was chosen that first assessed the ecological niche of the Yukon tree species, and then in a second step brought these mechanisms to the landscape scale to simulate forest succession with and without natural disturbance agents such as fire, and with and without climate change [Ohlson et al. (2005): (iii) assess vulnerabilities].  Chapter 4: This chapter synthesizes the research. The outcomes of the earlier chapters are used to show the IDF planning approach in its entirety: a detailed description of the IDF concepts (filters, states), plus the testing of its applicability, are presented in a case study [Ohlson et al. (2005): (i-vii)]. D State: I choose one of the alternative forest management strategies characterized in Chapter 2 to move from the strategic level to the tactical level. For illustrative purposes, I followed the principle of simplicity to avoid possible confounding factors and chose the strategy ‗manage forests to reduce fire risk‘ for this chapter. M State: different management tactics are applied [Ohlson et al. (2005): (vii) implementation], and their impacts (or efficacy) at the landscape level are discussed [Ohlson et al. (2005): (v) evaluate alternatives (here referring to tactics, not to the alternative strategies developed in Chapter 2)]. 28   Chapter 5: General Conclusion. This chapter summarizes the benefits and limitations of the proposed decision-making support approach for forest management planning, and highlights possible future research directions for the IDF approach.  Figure 1.1: Integrated Dual Filter concept listing purposes and main tools used in this study. Square brackets refer to the respective chapters in this thesis  In summary, the following highlights the material which was used from the Champagne and Aishihk Traditional Territory, the case study for the testing of the IDF planning framework:  Development of decision alternatives: The SFMP document has been used to identify and rank values to characterize alternative forest management strategies [Social Filter, Chapter 2]. 29   Development of a planning baseline: Ecological and environmental data and information has been used to assess the natural system responses under climate change [Environmental Filter, Chapter 3].  Designing of a desired landscape: A landscape has been designed and engineered to address one of the major concerns (fire risk) in the area [IDF Synthesis, Chapter 4].  Formulation of management implementations: Management actions have been implemented onto a natural landscape to assess the feasibility of a desired landscape [IDF Synthesis, Chapter 4].   1.10 Outlook  The Integrated Dual Filter approach consists of two major ‗filters‘ or dimensions of values: the Environmental and Social Filters, which together cover an ‗n-dimensionality‘ of ecological, economic and social forest values. These are all important for forest management planning, and add to its increasing complexity. The following chapter (Chapter 2) introduces the Social Filter of the IDF approach, using as an example the participatory forest management in the Champagne and Aishihik Traditional Territory, southwest Yukon, Canada. For this, a group of forest experts and practitioners from the Yukon developed and structured a set of alternative forest management strategies to evaluate which best represented sustainable forest management in the Champagne and Aishihik Traditional Territory (CATT). The characterization of these strategies is based on the Strategic Forest Management Plan for the CATT (SFMP 2004), with each strategy being described by its long-term emphasis (using the question ―what is our forest for?‖), on the rating of economic, social and environmental values (the so-called objectives-based 30  values), and on the short-term tactics or forest management actions at the landscape and stand scales.  In the IDF-planning approach, the Social Filter represents the framework for informing the D State (the desired future forest) for the next planning steps – what is the main socially desired strategic direction for sustainable forest management? This direction or emphasis characterizes the D State in the IDF planning approach. Each of the five alternative forest management strategies potentially represents a D State (e.g., managing for fire risk reduction, or managing for wildlife, or managing for timber production). In Chapter 2, after the characterization of the five alternatives, I used the Analytic Hierarchy Process (AHP, Saaty 1977, 2001) to assess which of the five alternatives would best balance the two potentially contradicting or competing criteria of ‗economy‘ and ‗ecology‘ to reach the goal ‗sustainable forest management in the CATT‘. 31  2. Filtering socially balanced forest management strategies for the Champagne and Aishihik Traditional Territory, southwest Yukon  2.1 Introduction  Sustainable forest management (SFM) calls for balancing diverse ecological, social, and economic values (e.g., Varma et al. 2000), usually represented in the form of criteria and indicators (Montreal Process 1995) that express sometimes conflicting management objectives. SFM includes the protection of a wide array of forest components such as measurable commodities, opportunities for outdoor recreation, aesthetic values, and ecosystem processes (Silsbee and Peterson 1993, Kant and Lee 2004). Forest managers, for example, are confronted with the challenge of resolving complex issues spanning spatial scales that range from stands of a few hectares to landscapes that are tens of thousands of hectares in size. Similarly, temporal scales can range from hours (e.g., forest fire control) to decades or centuries for strategic forest management planning. Forest managers therefore have to address objectives concerning different scales within a hierarchical decision-making process laden with seemingly competing management objectives (e.g., maintenance of productive capacity vs. conservation of biodiversity). Decision-making and priority setting requires an evaluation of the advantages and disadvantages of alternatives for management. However, many evaluations suffer from one or more shortcomings that hinder longer-term decisions (i.e., 100 or more years). Many evaluations focus on producing detailed information about a small set of impacts: an evaluation might, for example, examine the detailed effects of fire on forest stand recovery. Often the evaluation is 32  limited to a single spatial or temporal scale. ‗Real-world‘ decisions, however, involve trade-offs among multiple goals and objectives and different spatial or temporal scales of concern, or they involve several interest groups and stakeholders (Klenk and Hickey 2011). An adequate and especially balanced representation of key trade-offs is therefore critical for good decisions.  As complexity increases it becomes increasingly difficult for decision-makers to identify management alternatives or directions that can maximize, or best balance, all the decision criteria (e.g., Mendoza and Martins 2006, Ananda and Herath 2009). Few decisions in forest resources management are made unilaterally by a single stakeholder group (Schmoldt et al. 2001). Consultation and collaboration are becoming more common; an implicit belief in sustainable forest management is that an increase in public involvement in decision-making processes will lead to greater acceptance of the decisions made in such processes (Wondolleck and Yaffee 2000, Harshaw 2010). Collaborative decision-making is therefore becoming increasingly common. However, there are shortcomings associated with it: consensus among stakeholders is often not achieved and/or processes are time consuming and inefficient, leading to conflicts; or the public is not given reasonable choices amongst alternative management actions (Gregory 2002). Additionally, there is often a lack of credibility or perceived equity and fairness in the decision-making process (Gregory 2002). Forest managers are therefore typically confronted with outcomes from public processes that are too broad or that have potential conflicts with the sustainable forest management approach (Martin et al. 2000).  33  In the Champagne and Aishihik Traditional Territory (CATT) in southwest Yukon, Canada, an unprecedented spruce bark beetle outbreak has recently occurred (1993–2006), affecting about 380,000 ha of forest (Berg and Henry 2003, Berg et al. 2006, Yukon FHR 2009). The severity and extent of this disturbance has been linked to recent changes in climate within the CATT (ACIA 2004, Berg et al. 2006). The forest governance response to this epidemic was the development of a hierarchical approach to strategic forest management planning: The Yukon Territorial Government and Champagne and Aishihik First Nations Government, together with the Alsek Renewable Resource Council, engaged in a collaborative process to develop a strategic forest management plan (SFMP) (SFMP 2004). The SFMP outlines strategic directions for sustainable forest management and planning and provides a framework and practical guideline for forest management (SFMP 2004). For most people in the region, forest management is an approach that balances the maintenance of a wilderness environment with viable resource development opportunities (SFMP 2004). The SFMP addresses four fundamental objectives to achieving sustainable forest management in the CATT (SFMP 2004): (1) functioning forest ecosystems, (2) community sustainability and benefits, (3) cooperative forest management, and (4) building local human capacity. The management priorities of the SFMP are to reduce fire hazard, promote forest renewal, increase or maintain economic benefits, and preserve wildlife habitat.  The SFMP development team identified a broad range of values within the region; competing views about how the forest should be managed and used will need to be continually balanced were also identified (SFMP 2004). The SFMP explicitly incorporates a commitment to developing an adaptive management framework that includes monitoring the effects of forest 34  management activities and modifying practices as a necessary step to ensuring that the goals and objectives are being met. The stated time frame of the SFMP is 20 years. The plan organizes the region into 18 planning areas based on watersheds. A second planning document, the Integrated Landscape Plan (ILP 2007), was developed to guide landscape-level management. The ILP identified three management treatments that would form the basis of zoning for each of the planning regions identified in the SFMP: an intensive management zone, a provisional management zone, and a conservation zone (no forest harvesting). A third planning document that guides stand-level management actions was also developed to guide local harvesting tactics.  The SFMP was developed in response to the recent bark beetle outbreak which in turn has been related to recent climate change (see Berg et al. 2006). Interestingly, the SFMP does not directly address climate change. Ogden and Innes (2008) classified the SFMP as a ―reactive- indirect‖ plan, i.e., focusing on mitigation rather than anticipation: it does not explicitly identify the impacts of climate change or the vulnerabilities of ecosystems and communities, nor does it prescribe management actions that will reduce vulnerabilities. Although the SFMP prescribes ―best management‖ practices, these may not reduce the impacts of climate change (Ogden and Innes 2008). Ogden and Innes (2008) surmise that ―reactive-indirect‖ management may arise when managers are unaware of the vulnerabilities and/or changes that may occur due to climate change and/or how to manage for climate change. The development of a reactive and indirect plan in response to a climate change impact identifies that lack of knowledge on climate change impacts exists that will require increased research into the interaction between impacts of , vulnerabilities to, and management of climate change. Ogden and Innes (2009) identified that in the CATT, increased research and monitoring is needed to better understand how climate change 35  may affect the region‘s forests. The authors also stated that strategies and practices need to be explored that will better enable local managers and communities to adapt to future climate change. Alternative management strategies that are both socially and environmentally acceptable are therefore needed for understanding the long-term impacts of climate change and strategic planning decisions on achieving sustainable forest management in the CATT (Ogden and Innes 2009).  The objective of this study is to use the existing Strategic Forest Management Plan of the CATT as a foundation from which to develop socially balanced alternative forest management strategies for transitioning the forest management in the CATT from short-term and tactics-based management, to long-term and sustainable forest management. The study also explores how alternative strategies incorporate climate change impacts and actions that may reduce the risks associated with climate change. This may encourage the transition of the SFMP from being mainly a ―reactive-indirect‖ plan to a ―proactive-direct‖ approach for facilitating the adaptation of the forests and communities of the CATT to climate change.   2.2 Methodology Multiple approaches were used to develop and prioritise socially-balanced alternative forest management strategies (FMSs). A content analysis of the Strategic Forest Management Plan (SFMP 2004) and Integrated Landscape Plan (ILP 2007) was conducted with working group members through meetings and interviews; the working group also helped to develop a ratings 36  table for an SFM-expert group. These approaches were used to collect and collate data that was then used in an Analytic Hierarchy Process (AHP, Saaty 2001). According to Schmoldt et al. (2001) AHP is an ideal decision support tool which allows using subjective and judgmental information in a transparent and straightforward manner.  The Analytic Hierarchy Process (AHP, Saaty 1977, 2001, 2008) is a method used for analyzing a problem in a systematic manner. It consists of three basic parts: (1) Decomposition of the problem (i.e., structuring it into a hierarchy consisting of a goal and its criteria, and continuing to decompose the latter into subordinate features); (2) Evaluation of the problem: judging and pair-wise element comparison at each hierarchy level (i.e., people judge each element of the AHP structure by assigning ranks to the elements — the criteria and sub-criteria); and, (3) Synthesis (i.e., propagation and calculation of relative weights and local priorities to global and overall priorities using the subjective judgments). Subordinate features can include objectives, scenarios, events, actions, or outcomes that best describe the criteria listed at level two of the hierarchy. The final result of the AHP is a numeric value that indicates the relative priority of each of the bottom-level alternatives. The decision maker then selects the alternative with the highest score (i.e., highest overall relative priority) as the ‗best‘ forest management strategy (FMS) or alternative for achieving sustainable forest management in the CATT.  2.2.1 Yukon working group and ratings table In this study, the working group discussions were led by a facilitator; all group decisions were based on consensus reached through open discussions. The working group consisted of 37  seven local experts and practitioners from the Yukon, who were well-acquainted with the SFMP and forest management planning. All participants of the Yukon working group (three of whom represented FN interests) were guaranteed confidentiality; thus, no further details on the type of group or identity of the participants are disclosed in this chapter. The working group identified important values listed in the SFMP and reorganized them in order to develop a static hierarchy of constituent elements (criteria and sub-features) for the AHP that best described the two potentially competing criteria ‗functioning forest ecosystems‘, and ‗community economic sustainability and benefits‘ (labelled L2A and L2B in Figure 2.1, respectively). All the AHP hierarchy elements originated from the SFMP (SFMP 2004), including the overarching goal ‗sustainable forest management in the CATT‘. The terminology used in the SFMP was adopted in order to allow clear identification of the values chosen from that document (Table C1 in Appendix C).  During the working group meetings, a ratings table (Table C2 in Appendix C) was designed by the participants to help complete the AHP (i.e., judgment of the hierarchy elements). The table consisted of three parts: (1) Objective-based values that were grouped into cultural, environmental and economic categories (e.g., tourism, hunting, and timber production); (2) landscape-level planning approaches (e.g., zoning); and (3) stand-level management actions (e.g., thinning). The latter two represent management tactics and actions for the short-term, e.g., 5-10 years. The three groups of elements characterize the alternative and competing forest management strategies that represent the bottom level of the AHP hierarchy (see Figure 2.1) based on their respective importance. Landscape and stand scale tactics are based on forest 38  management actions and operations that may be undertaken to achieve the respective forest management strategy (FMS).  The developed ratings table was then completed by an independent SFM-expert group (12 SFM researchers from the University of British Columbia (UBC), Vancouver who have published widely on SFM-related issues in Canada and the Yukon) who assigned ratings of ‗low‘ for ‗least important‘ to ‗high‘ for ‗very important‘ to the elements of the questionnaire (Table C2 in Appendix C). Each member of the SFM-expert group assigned the ratings individually; the ratings table were completed anonymously. The resulting hierarchy box ratings (for details see Step 2 in C2 in Appendix C) were then used to generate pair-wise judgments for the AHP hierarchy (Table C4 in Appendix C). The AHP structure, with its judgments and pair-wise comparisons for obtaining the weights of the criteria on all the levels including the alternatives at the bottom level, has been synthesized following Kuusipalo and Kangas (1994) (for details see Step 3 in Appendix C, Tables C5 and C6). 39    Figure 2.1: Hierarchy of the Analytic Hierarchy Process (AHP). Each box describes a criterion or sub-feature of the AHP structure, and has a unique label referring to its position in the hierarchy. FMS1-5 constitute the bottom level of this structure representing the five alternative forest management strategies. 40    2.3 Results  2.3.1 The AHP hierarchy structure The hierarchy elements are those that best describe the two L2-criteria ‗functioning forest ecosystems‘ (L2A), representing the environmental values, and ‗community sustainability and benefits‘ (L2B), representing the socio-cultural and -economic values of the CATT region. The resulting hierarchy consisted of four levels, with three elements broadly characterizing each of the L2 criteria. Level 4 of the hierarchy provides a refinement of the L3 elements, representing ecological and economic process and structures. The criteria and sub-criteria of Figure 2.1 can be grouped into socio-cultural (e.g., L3BB and sub-features), environmental (L3AA/L3AB, and sub-features), and economic values (L3BA and sub-features). The bottom level of the hierarchy represents the competing alternative forest management strategies (FMS1–5). The SFM expert group did not add any meaning to these elements; they simply rated the elements listed in the ratings table (Table C2 in Appendix C) that the Yukon working group developed during earlier meetings.  2.3.2 Characterizing the alternative forest management strategies The working group developed five forest management strategies as alternatives for achieving SFM within the CATT. The five strategies were: 41  1) Manage to support and enhance a sustainable forest industry: the main goal of FMS1 (Figure 2.1) is to create a sustainable forest industry in the region. Four out of seven economic elements have been rated by the SFM-expert group with medium-high to high importance. The promotion of a timber and biomass industry was advocated by the working group, while the SFM-expert group placed higher importance on maintaining forest productivity and enabling and encouraging forest-based activities that stimulate employment opportunities (Table C2 in Appendix C). 2) Managing for multiple values and use (the holistic strategy): the main goal of FMS2 (Figure 2.1) is to balance a broad variety of different socio-cultural, environmental and economic values. The SFM-expert group assigned medium to high ratings in all three categories; however, the environmental category received the highest importance (Table C2 in Appendix C). 3) Managing for fire risk reduction: the goal of FMS3 is to reduce fire risk in the three communities and at the landscape scale of the CATT (for delineation of boundaries of community see Figure 4.2, for boundaries of interface and landscape zones see ILP 2007). Fuel treatments were assigned the highest ratings in terms of management actions (Table C2 in Appendix C). 4) Managing for wildlife: the goal of FMS4 is to manage the forest landscape with a strategic emphasis on maintaining and enhancing habitat for locally important wildlife that are utilized primarily for hunting and as a food source. The highest importance was given to the environmental values. The cultural values were rated between medium to high, and the economic values between low to moderately high. The most important management tactics identified related to the protection of core areas containing high quality habitat and the provision of appropriate wildlife movement corridors between habitat areas at the landscape-level. The 42  control of invasive species at the stand-scale was also identified as an important management tactic. 5) Managing for a carbon economy: the goal of FMS5 was to strategically manage the forest to maximise carbon sequestration and storage. The SFM-expert group rated ‗maintain forest productivity‘ (economic) and ‗maintain both forest ecosystem health and resilience‘ (environment) as the most important objective-based values.  Over all the five FMS, environmental objective-based values were assigned relatively high scores (e.g., on average 7–9), and the cultural values received medium ratings, whereas the economic values show the greatest variability in terms of ratings, with differences occurring in response to the objectives of each scenario.  2.3.3 The AHP judgments and priority relations The pair-wise Saaty judgments for the criteria and sub-features of L2–L4 were found to be relatively similar (e.g., between 5.6 and 8.2, see Table C4 in Appendix C), indicating that the elements within these levels are of similar degrees of importance. The pair-wise comparison between L5 (the alternative FMS) and their covering L4-elements revealed a broader range of judgments (e.g., between 5.0 and 9.0). There are 18 (i.e., 1+2+4+11) pair-wise comparison matrices of the type presented in Table C5 in Appendix C.  43  At the L3-level, four priority relations were identified (see equation (2) in Step 3 in Appendix C). Forest Management Strategy 4 (‗wildlife strategy‘) scored the highest for the sub-criteria ‗maintain/enhance natural processes‘ (Table 2.1: Natural Processes, 0.2191), ‗maintain/enhance ecosystem diversity‘ (Table 2.1: Diversity, 0.2308) and ‗enable and encourage forest-based activities to stimulate employment opportunities‘ (Table 2.1: Employment, 0.2138); FMS1 (‗timber strategy‘) scored the highest global priority of 0.2356 (Table 2.1: Industry). The ‗multiple values strategy‘ (FMS2) had the second highest scores in all four L3-sub-criteria (Table 2.1).  At the L2-level, two priority relationships (see equation (3) in Step 3 in Appendix C) were identified. The normalized priorities (L2 of Table 2.1) show that with a score of 0.2249, FMS4 scored the highest for the criterion ‗functioning forest ecosystems‘ closely followed by FMS2 with a score of 0.2153, and the other three alternatives ranging within 82–84% respective to FMS4. For the second criterion ‗community sustainability and benefits‘, FMS2 has the highest scores with the timber strategy FMS2 and the carbon strategy FMS5 scoring 98 and 96%, respectively, and the fire risk reduction strategy FMS3 scoring 86% respective to FMS2.  The normalized overall priorities calculated for each FMS (L1 of Table 2.1) identified FMS2 ‗manage for multiple values‘ as the ‗best‘ strategy for balancing the two criteria of ‗functioning ecosystems‘ and ‗sustainable economy‘ in order to reach the goal of ‗sustainable forest management in the CATT‘; FMS4 (‗wildlife strategy‘) was found to be 97% as effective as 44  FMS2, while FMS1 ‗timber strategy‘ and FMS5 ‗carbon strategy‘ both scored 92%; FMS3 ‗fire risk reduction‘ had the lowest score (86%). 45  Table 2.1: Priorities of the five FMS for each AHP hierarchy element (bold), with respective weights listed above each hierarchy element. L4 priorities are local priorities, L3 and L2 are global priorities, L1 are overall priorities. "ideal." = idealized vector with the highest priority being 100%. L1  SFM ideal.       FMS1 0.1972 0.92       FMS2 0.2142 1.00       FMS3 0.1843 0.86       FMS4 0.2080 0.97       FMS5 0.1962 0.92       L2 Weight 0.5082672  0.4917328   Environment ideal.  Economy ideal.   FMS1 0.1869 0.83  0.2079 0.98   FMS2 0.2153 0.96  0.2131 1.00   FMS3 0.1845 0.82  0.1841 0.86   FMS4 0.2249 1.00  0.1906 0.89   FMS5 0.1883 0.84  0.2043 0.96   L3 Weight 0.5084264  0.4915736  0.4840006  0.5159994   Natural Process ideal. Diversity ideal. Industry ideal. Employment ideal. FMS1 0.2005 0.91 0.1730 0.75 0.2356 1.00 0.1819 0.85 FMS2 0.2170 0.99 0.2136 0.93 0.2147 0.91 0.2116 0.99 FMS3 0.1811 0.83 0.1880 0.81 0.1720 0.73 0.1954 0.91 FMS4 0.2191 1.00 0.2308 1.00 0.1658 0.70 0.2138 1.00 FMS5 0.1823 0.83 0.1946 0.84 0.2119 0.90 0.1973 0.92 L4 Weight 0.337426  0.3238194  0.5457708  0.3309392   Productivity ideal. Habitat ideal. Timber ideal. Tourism ideal. FMS1 0.2096 0.94 0.1719 0.74 0.2397 1.00 0.1930 0.91 FMS2 0.2223 1.00 0.2122 0.91 0.2218 0.93 0.2130 1.00 FMS3 0.1807 0.81 0.1865 0.80 0.1692 0.71 0.1930 0.91 FMS4 0.2032 0.91 0.2332 1.00 0.1579 0.66 0.2080 0.98 FMS5 0.1842 0.83 0.1962 0.84 0.2115 0.88 0.1930 0.91 Weight 0.3657249  0.3536153  0.4542292  0.3314877   Succession ideal. Water ideal. Biomass ideal. Hunting ideal. FMS1 0.1971 0.90 0.1805 0.82 0.2308 1.00 0.1743 0.80 FMS2 0.2190 1.00 0.2000 0.91 0.2062 0.89 0.2110 0.97 FMS3 0.1825 0.83 0.1902 0.87 0.1754 0.76 0.1988 0.92 FMS4 0.2190 1.00 0.2195 1.00 0.1754 0.76 0.2171 1.00 FMS5 0.1825 0.83 0.2098 0.96 0.2123 0.92 0.1988 0.92 Weight 0.2968491  0.3225654  0.3375732   Disturbance ideal. Connectivity ideal.  NTFP ideal. FMS1 0.1942 0.82 0.1658 0.69  0.1784 0.82 FMS2 0.2086 0.88 0.2299 0.96  0.2108 0.98 FMS3 0.1799 0.76 0.1872 0.78  0.1946 0.90 FMS4 0.2374 1.00 0.2406 1.00  0.2162 1.00 FMS5 0.1799 0.76 0.1765 0.73  0.2000 0.93 46  2.4 Discussion  2.4.1 Strategic forest management planning According to Martell et al. (1998), strategic forest management planning can take place on small parcels of land or vast management units spanning several hundred thousand hectares. Often, a planning time frame can run for over 100 years (Naesset 1997) and can span planning horizons of centuries (Martell et al. 1998). The planning process is typically structured into hierarchical levels (Davies and Martell 1993, Jeakins et al. 2006) to better address the different scales of the system under planning: (1) A strategic level outlining the key values at stake and allocating the forest resources for the entire planning period; (2) a tactical level that schedules actions and operations for the short-term (e.g., 10 years); (3) a third level that characterizes the operational details at smaller spatial and temporal scales. The Strategic Forest Management Plan (SFMP) represents a Level 1 plan as it identifies key issues related to the CATT forestry (e.g., reduction of fire hazard, community and economic benefits, or preservation of wildlife habitat) and formulates goals and objectives. However, the SFMP has a relatively short planning time frame of 20 years (with a reassessment of implemented and achieved management actions after the first 10 years (SFMP 2004)), and it lacks a stated long-term strategy and direction that clearly points out how sustainable forest management will be achieved in the CATT. The landscape- level planning document, the Integrated Landscape Plan, identifies actions for the CATT for the coming 5-10 years and thus constitutes a Level-2 plan. For example, the Yukon Territorial Government has allocated one million m 3  of the beetle-killed wood to be harvested over a period of 10 years (by 2016). This management focus primarily constitutes mitigation of climate change impacts since it identifies actions that will be taken in response to change. According to Ogden and Innes (2008) it is indirectly doing a reasonable job of minimizing the risks associated with 47  climate change because of its adherence to the principles and practice of sustainable forest management and use of an adaptive management framework. Although the SFMP does not directly address adaptation, it is believed that the plan will allow for adaptation to climate change (Ogden and Innes 2008, 2009). While the majority of the CATT forestry planning is focused on short-term and unsustainable salvage harvesting of beetle-affected stands (cf. Lindenmayer et al. 2008), there has been an increase in fuel-abatement treatments around the communities of Haines Junction and Canyon, which have been undertaken to reduce fire risk and improve community safety. These management actions constitute proactive and direct adaptation to both current and future fire risk (Hajjar et al. 2009). The SFMP in its current form and with its hierarchical planning approach allows for short-term management actions and tactics; however, it is not a strategic plan that allows for proactive adaptation to climate change. Nonetheless, as demonstrated here, this strategic plan provides a good foundation for developing alternative forest management strategies and directions for the purpose of forest management planning in order to achieve sustainable forest management in the CATT under environmental uncertainty.  2.4.2 Alternative strategies and climate change In order to balance potentially competing ecological and economic values to achieve sustainable forest management in the CATT, the Yukon working group and SFM-expert group have characterized five alternative forest management strategies that all scored between 0.1843 and 0.2142 (L1, Table 2.4) where the strategy with the lowest priority (FMS3, ‗manage to reduce fire risk‘) constitutes 86% of similarity to the ‗best‘ scoring strategy (FMS2, ‗multiple value strategy‘. The Yukon working group developed and structured values and objectives that the SFM-expert group then ranked and judged. 48   The main goal of FMS1 is to create a sustainable forest industry in the region, with a focus on timber and biomass production. Other areas that were ranked as important were the maintenance of forest health and resilience (Table A2). Although currently there is only a small- scale forestry industry established in the southwest Yukon, this strategy reflects one of the main key issues identified by the current SFMP (2004), to increase community and economic benefits from forestry. The participants of the working group developed tactics and actions that could be used to achieve each identified FMS. These tactics focused primarily on promoting forest productivity (timber and biomass production) and stimulating forest renewal, which is of particular interest in the context of recent climate change and forest disturbance (e.g., the recent beetle outbreak or the progression of leaf miners (Yukon FHR 2009)). Currently, forest productivity in the CATT region is low, with 80% of the 272,000 hectares of forested land classified as ‗poor‘; heights of 10–14 meters are expected after 100 years (SFMP 2004). The SFM-expert group encouraged the establishment of both single and mixed species plantations following harvesting. To maintain landscape heterogeneity, a variety of harvesting and retention practices were proposed; in particular, variable retention harvesting. To date, the only harvesting in the region is a tenure issued by the Yukon Government for the salvage harvesting of beetle- killed white spruce. This has not been realized and initial harvesting of two to three thousand m 3  of spruce beetle killed wood is being undertaken by a local wood products mill (Dimok Timber). The real obstacles for the establishment of a timber industry is market access given the expensive transportation costs associated with the CATT‘s geographic position and low forest productivity. Whether the establishment of a timber industry is a realistic and sustainable option remains an 49  open question, as it will be heavily influenced by future market access issues (be that timber or carbon) and potential changes in forest renewal and productivity due to climate change.  ‗Managing for fire risk reduction‘ (FMS3) incorporates one of the SFMP‘s key objectives. Fire is perceived as one of the biggest threats in the region given the recent beetle outbreak and consequent increase in dead fuel loads within the forests (Garbutt et al. 2006). FMS3 has the goal of reducing fire risk at the landscape scale with a focus on risk reduction within the community zones (e.g., interface and community zones defined in the Integrated Landscape Plan for the CATT (ILP 2007)). Special emphasis was placed on respecting traditional hunting and trapping lines, and also on maintaining the current level of biodiversity within management areas. This strategy entails tactics such as fuel treatments around the communities, and/or the promotion of different seral stages throughout the landscape. The tactics defined for this FMS are strongly related to the Fire-smart management paradigm that emphasizes the use of forest management practices to reduce the risk of fire by altering fuel composition and structure in order to decrease fire behavior potential and increase the ability to suppress fires successfully (Hirsch et al. 2001, 2004). Predicted climate change is expected to lead to increases in forest diseases, drought events, and forest fires (Flannigan et al. 1998, Fleming and Candau 1998). Fire occurrence and area burned is expected to increase by two to three times respectively in the Yukon under future climate change (McCoy and Burn 2005). Due to the expected increase in fire behavior and occurrence, some local practitioners within the CATT identified that the current use of fuel abatement through thinning may be insufficient to reduce fire risk in the long-term and that the only effective management action should involve the conversion of stands around the communities from white spruce to deciduous species to better provide for community safety. 50  Hirsch et al. (2001) recommended the use of fuel reduction burning to reduce potential fire behavior; however, this tactic was not defined by the working group.  The goal of FMS4 ‗maintain, protect and enhance fish and wildlife habitat‘ is to conserve the habitat of wildlife in the region (SFMP 2004). The focus of this strategy is on wildlife that are important to the communities and to plants that have local and cultural significance (e.g., Labrador tea (Ledum groenlandicum) and Soopolallie (Shepherdia canadensis)). Forest tactics and operations were devised to focus on actions that promote a diversity of seral stages at the landscape-level and particular tree and shrub species at the stand-level. The expert group acknowledged the importance of maintaining healthy forest ecosystems in order to provide for a high diversity of wildlife. This is important as some wildlife are attractive for tourism (e.g., grizzly bears (Ursus arctos)), while others represent important cultural values; particularly (meat-) hunting (e.g., moose (Alces alces) and wood bison (Bison bison athabascae)) and trapping (e.g., pine marten (Martes americana)). Identified management actions were therefore developed that focused on maintaining landscape connectivity to provide wildlife with natural and seasonal migration routes and/or maintain range patterns, reducing fragmentation in the landscape, and protecting important habitat areas (riparian areas, winter ranges and calving areas). At the stand-level, actions focused on the control of invasive species and the retention of course woody debris and dead and dying trees, shrubs, and hardwoods to increase stand-level heterogeneity. Both stand and landscape-level heterogeneity are regarded as important attributes for conserving biodiversity (Lindenmayer and Franklin 2002). Heterogeneity may also be important for reducing the impacts of climate change on forest ecosystems (Noss 2001). 51   Maintaining forest productivity (economic) and ecosystem health and resilience (environment) were identified by the expert group as the most important objective-based values (Table 2.1) for ‗managing for carbon economy‘ (FMS5). This strategy may become more important with the development of a global carbon economy. For this to be realised in the CATT however, an increase in forest productivity and/or a reduction in future disturbance events (fire and bark beetle in particular) will be required to provide a sustainable and positive carbon balance.  The main goal of ‗managing for multiple values and use‘ (FMS2) is to provide a variety of different forest values and therefore represents a holistic strategy. This strategy reflects the current SFMP and is the closest in terms of scope, by having its focus on two of its stated key issues (fire risk reduction and promoting wildlife); this strategy also represents the other four alternatives developed by the working group (fire, timber, wildlife, carbon). The participants of this research opted for the broadest variety of tactics to apply at both the landscape and the stand scale, which allows managing the future forest towards a heterogeneous state (Lindenmayer and Franklin 2002). This strategy would certainly meet most demands of the local communities and also entail the highest number of proactive direct climate change activities. However, it remains unknown whether all the values can be addressed under the context of SFM and climate change.  52  2.4.3 Holistic approach for balancing ecology and economy Although the SFMP (2004) has a clear direction towards a timber strategy to promote and establish a timber industry in the CATT, the AHP performance for FMS1 is modest as it scored the highest amongst the five alternatives in three hierarchy elements (Table 2.4, L3BA ‗promote a forest timber industry in the region‘, and L4BA1/2 ‗strengthen local timber harvesting and processing capacity‘/‗promote harvest for biomass‘). Overall, the timber strategy and the carbon strategy scored relatively low with respect to the overall goal of ‗sustainable forest management in the CATT‘. However, these strategies are still 92% similar to the best alternative strategy, FMS2. The recent beetle outbreak has caused an unprecedented amount of dead fuel in the forests which is perceived as a great concern because of increased fire risk (Garbutt et al. 2006). Given the relatively dry conditions of the region due to its location in the rain shadow of the Mt. Elias range, the occurrence of severe fire weather is frequent, and under climate change, extreme fire weather may occur more frequently (see McCoy and Burn 2005). It is therefore interesting that the AHP analysis revealed relatively low scores for FMS3: priorities for all the hierarchy elements are never above third place; in 12 out of 18 elements it scored fourth place (Table 2.4). This could be explained by the structure of the hierarchy and the elements listed in the ratings table, with the Yukon working group placing little emphasis on fire management; also, many elements promote heterogeneity at the landscape or stand scale and therefore implicitly address fire risk reduction tactics.  The SFMP (2004) states the importance of a world class wilderness for the regions‘ economy and culture, which is also reflected here in the AHP results: under the ‗ecology‘ criterion (L2A) seven out of eight hierarchy elements (Figure 2.1., Table 2.1) have the highest scores for the 53  ‗wildlife strategy‘, underlining the importance of wildlife for the entire environmental sector of the region. Amongst the ‗economy‘ criteria (L2B) and sub-features, the ‗wildlife strategy‘ scored the highest priorities under the socio-cultural features defined as ‗enable and encourage forest based activities‘ (L3BB) and its sub-features ‗trapping and hunting‘ (L4BB2) and ‗non-timber forest products‘ (L4BB3). This not only reflects the importance of the biodiversity component for the regional economy, but it also shows how strongly embedded biodiversity elements are in the regional culture.  FMS5: The ‗carbon strategy‘ is becoming a strategic management alternative when the role that forests may play in mitigating climate change is considered (e.g., Greig and Bull 2009). This strategy has performed the same as the ‗timber strategy‘; however, it never had the highest scores in any of the 18 hierarchy elements. This again, can be explained by the AHP structure of Figure 2.1., which is missing any direct element representing carbon economy.  The ‗multiple value strategy‘ FMS2 had the highest priorities in 5, and second highest in 13 out of 18 value dimensions (Table 2.1, Figure 2.1) and therefore best balanced the two criteria ‗ecology‘ and ‗economy‘. This is a clear indication that economic, environmental and cultural values seem to be best reflected when balanced against each other within a holistic strategy.  The verbal rankings for each objective value ranged from low to high and differed according to the five forest management strategies (FMS); however, the ratio scales for the comparison 54  matrices were all 1 or 2:1 (with a maximum possible ratio of 9:1 for the AHP synthesis process). This relatively low ratio could be due to the fact that the verbal ranking of six options (Table 2.1) was not enough to capture sensitivities of the criteria and its sub-features (Figure 2.1). Sometimes, as Kuusipalo and Kangas (1994) note, a ratio of 2:1 may represent a ―big difference‖ for a pair-wise comparison, but a ratio of 9:1 on the other hand can sometimes be insufficient to explain smaller differences. A more sophisticated verbal ranking is advised in order to capture individual rankings in a more sensitive way, e.g., graphic ranking as proposed by Kuusipalo and Kangas (1994).  The Strategic Forest Management Plan document is the result of a community-based process that took nearly ten years to complete (SFMP 2004). The SFMP can therefore be regarded as set of values representing local stakeholders‘ views and perspectives to inform adaptive and sustainable forest management in the region. My research extended this participatory approach: The combined approach of working group discussions, ratings table and AHP allowed the participants to identify, develop and characterize a set of alternative forest management strategies in a reasonable time (within the course of 3–4 days of meetings). The strategies represent a wide range of management emphases based on existing SFMP objectives. The outcome of the AHP analysis (Table 2.1) is dependent upon the hierarchy structure and composition (which criteria and sub-features, e.g., the elements in Figure 2.1). The rankings and judgments depend on the type of group(s) that participate(s) in such an approach. The AHP is sensitive to these judgments (Schmoldt et al. 2001). The pair-wise judgments would certainly look different and impact on the final outcome. Originally, the research design was set up to have the Yukon working group provide the ratings and judgment of the hierarchy elements (Figure 55  2.1). However, due to specific circumstances (i.e., the small number of people involved would make it easy to identify the members of this working group) there was a general unease among working group members to provide these judgments. One working group member commented that ―we do not feel comfortable in providing ratings for the Carbon strategy since we simply do not know enough about it to provide ratings and judgments‖. Thus, an outside working group was used for the ratings of the values.  In social research, especially in the context of land use and management, it is important to respect the participants‘ opinions (but see CCFM 2003). It is also important to have social sustainability within a planning process in order to allow the forests of the future to maintain their social functions, and to allow the communities depending on these forests to have fair and transparent management processes for their forests (Pukkala 2002). This combined participatory approach linking lay input (from the community level) condensed in the SFMP document, with input from practitioners and experts (Yukon working group allowing for open discussions and free sharing of ideas), and the opinions of expert researchers (UBC) enabled the characterization and filtering of alternative forest management strategies from an already existing plan (i.e., the SFMP), and the formulation of clear and desired directions for SFM in the region. The ‗multiple use‘ strategy has received the highest ratings and scored the best priorities in the AHP analysis; I can therefore suggest that this strategy is socially balanced and may ‗best‘ represent the achievement of SFM in the CATT. It remains unclear, however, whether this strategy amongst the five characterized alternatives would best facilitate adaptation to climate change and achieves sustainable forest management in the Champagne and Aishihik Traditional Territory. In order to test this, a scenario-based approach using quantitative modelling tools that allows for the 56  integration of management scenarios with environmental change is required. This is the focus of further research. In lieu of this scenario-based analysis, this study provides socially-based and accepted forest management directions to be explored with further research. This study has also quantified important objectives and their relationships to alternative strategies of SFM. Testing of these socially-based strategies will provide a greater degree of realism that will enable the transition of forest management in the CATT from being reactive to proactive in the context of climate change.   2.5 Conclusions  I believe that the combination of planning documents (SFMP/ILP) with AHP and working group inputs can function as a reasonable addition to the planning approach proposed by the Yukon governments (SFMP-ILP-Harvesting Plan-Site Plan). A working group allows for open discussions and sharing of ideas, and with the help of a facilitator it can be possible to find a shared decision. Further, the application of the AHP tool allows the comparison of different values (with different units), resulting in the prioritizing of the values listed in the major SFMP document. It enables the generation of socially-based scenarios (the alternatives). However, due to the limited number of natural resources managers and experts in the Yukon, their anonymity in this study could not be assured, and members were not comfortable to participate in an evaluative context (i.e., the ratings); hence, another group external to the Yukon was used. It is important that local knowledge is included in forest management planning. However, as with all 57  sources of information, it can add a bias by including (or excluding thereof) values that are deemed important. An external group such as the UBC experts can be regarded as knowledgeable on sustainable forest management issues in general, and bring some impartiality (i.e., fewer personal or political viewpoints) to the ratings of the value structure (developed by the local working group). However, an external group has the limitation that it will not be familiar with the local situation and may be unaware of local constraints, such as regulations and bylaws. The two groups should therefore be seen as being complementary. A weakness of this study as reported here is that there was insufficient capacity to re-engage with the working group to discuss the expert ratings and triangulate the results.  The alternative forest management strategies can be compared against each other to test potential management questions such as: does a social embedded scenario with a higher priority (see AHP outcome) really achieve sustainable forest management? How does a social scenario perform under changing climatic conditions? How sustainable are these scenarios under climate change in the long-run? The answer to these questions will only be found through simulation of these strategies using modelling and/or application of the strategies followed by long-term monitoring.  The SFMP in its current form, even with the ILP, cannot alone guide long-term forest management in the region. The documents only provide for short-term forest management planning that prescribes indirect and reactive management tactics. However, by using the SFMP and its follow-up plans in combination with techniques utilized in this chapter, the SFMP can 58  become a true SFM plan that can drive direct and proactive tactics that foster adaptation to future climate change.   2.6 Outlook  Disturbances as ecological processes can form patterns, and patterns can ease or constrain processes (Wu and Loucks 1995, Kuuluvainen 2002). Disturbance events can vary in size, frequency, and intensity. Large natural disturbances are ecologically important because they can structure an ecosystem as well as create further ecological legacies (Turner 1989). Ecological processes are strongly influenced by climatic and topographic factors (delMoral et al. 1995). They may also be determined by eventualities such as the timing of the availability of propagules, the spatial arrangement of surviving entities, or barriers to the spread of disturbances (Turner and Dale 1998). Hence, it is difficult to predict the effects of large disturbances on subsequent ecological processes, and even more so to include the probability of future disturbances. Ideally, empirical data and experiments would help the understanding of these complex interactions, but the broad temporal and spatial extent that must be addressed limits such approaches, making spatial modelling essential. However, the modelling of such interactions often requires spatially explicit data sets, which are rarely available (Mladenoff 2004). Also, using modeling to develop an understanding of the long-term dynamics of ecological systems across different spatial scales allows us to assess the variability in simulation outputs. 59   In order to ensure a holistic planning approach for the Champagne and Aishihik Traditional Territory, we need to explore both the social and environmental dimensions of this complex planning space. In Chapter 2, I introduced the Social Filter characterizing the potential desired D States. In the next chapter (Chapter 3), I will introduce the importance of the Environmental Filter to accomplish the ―Dual‖ part of the Integrated Dual Filter approach. For this, I use a combination of two models. The first is an ecological niche model (TACA, Nitschke and Innes 2008) that I use to explore how Yukon tree species may perform under different climate change scenarios. These mechanistic relationships are made spatially explicit through the use of a second model; this is an ecological landscape model (LANDIS-II, Scheller et al. 2007) that enables the simultaneous consideration of the most threatening natural disturbance for the region (e.g., fire) and forest succession. The resulting successional trajectories over a period of 200 years constitute the E (environmental)-states in the IDF planning approach. It is also the first time that these two models are been combined to directly include climate change at the site level. 60  3. The sensitivity of tree species at the site and landscape level to disturbance and climate change in southwest Yukon, Canada  3.1 Introduction  In the Yukon Territory, the mean annual temperature has risen between 1 and 3°C between 1955 and 2005, with some regions warming more than others (Réale et al. 2003, Field et al. 2007). The past 50 years of climate change has resulted in an increase in winter and growing season temperatures along with a reduction in soil moisture availability that has reduced tree growth at the northernmost extent of white spruce (Picea glauca) (Barber et al. 2000) and facilitated an unprecedented 380,000 ha spruce bark beetle (Dendroctonus rufipennis Kirby) outbreak (Berg et al. 2006, Yukon FHR 2009). Within the Champagne Aishihik Traditional Territory (CATT) of the southwest Yukon the spruce bark beetle outbreak affected about 85% of the forests between 1993 and 2006 (ACIA 2004, Strategic Forest Management Plan 2004, Yukon FHR 2009). This has resulted in a dramatic increase in dead fuel that has potentially increased the risk of forest fires occurring within the area (Garbutt et al. 2006); fire is recognized as a major threat to the communities of the CATT (SFMP 2004, Garbutt et al. 2006, or see Chapter 2). Concurrent with the bark beetle outbreak, the Yukon has also been experiencing increases in wildfire severity and frequency (Hogg and Wein 2005). Fire occurrence and area burned is expected to increase by two to three times respectively in the Yukon under future climate change (McCoy and Burn 2005).  61  Ecosystems are dynamic systems where the only constant is change (Agee 2003). The type, severity, extent, and frequency of disturbances regulate the structural mosaic that develops in a given forested landscape (Pickett and White 1985, Spies and Turner 1999). In forest ecosystems, disturbances are dynamic and complex hierarchical processes where different kinds of autogenic, biogenic and allogenic agents affect forest structure at a given location (Kimmins 2004). Climate change is a key stressor that directly and indirectly affects forest ecosystems through its interaction with disturbance agents (Nitschke and Innes 2006, 2008). It is predicted to have significant influences on forests at multiple scales. Climate directly affects the functions of individual organisms (physiology, i.e. growth, and behavior), modifies populations (size, structure, spatial and temporal distribution), and can affect ecosystem structure and function (decomposition rates, nutrient cycling, water flows, species composition and interactions) and the distribution of ecosystems within landscapes (Gitay et al. 2002). For example, Hamann and Wang (2006) predicted an upwards shift in elevation and northern shift in latitude of both tree species and forest ecosystems of British Columbia, Canada.  Climate change is expected to increase the severity and frequency of forest fires in many regions (Flannigan and Van Wagner 1991, Stocks et al. 1998). An increase in fire in the boreal forest has been attributed to recent climate change (Flannigan et al. 1998, 2005). The annual area burned in boreal forest regions of Alaska and western Canada has exhibited a pronounced upward trend between 1959 and 1999 (Xiao and Zhuang 2007). McCoy and Burn (2005) predict a 20% increase in burned area in the central Yukon under predicted climate change by the 2080s. Weber and Flannigan (1997) highlighted that a change in fire regime due to climate change in the boreal may be more influential than the direct effects of climate change in shaping the future 62  distributions of species. Bradshaw et al. (2000) state that frequent burning gives fire-adapted pine species a competitive advantage over many fire-evading species. As a result, fire-adapted species such as lodgepole pine (Pinus contorta) are predicted to expand in areas where future climate causes increased fire frequencies at the expense of other species such as subalpine fir (Bartlein et al.1997).  Climate change can act on the physiological and phenological mechanisms that tree species utilize; for example, the timing of flowering and bud burst are driven largely by temperature and moisture cues (Chuine and Beaubien 2001, Morin et al. 2007). Climate change can interfere with these mechanisms and thus prevent a species from reproducing, regenerating and/or growing. Species are also highly susceptible to frost damage during their regeneration phase (Goulet 1995). For example, Nitschke and Innes (2008) modeled an increased occurrence of frost damage in coniferous species in British Columbia due to early bud burst events under predicted climate change. Changes in phenology due to climate and climate change can impact the distributions of species (Chuine and Beaubien 2001, Morin et al. 2007). Drought is another important mechanism impacting a species‘ establishment ability and distribution (Sykes and Prentice 1995, Chhin and Wang 2008, Chhin et al. 2008). During the establishment phase, drier conditions can prevent regeneration by inducing the mortality of seedlings (Spittlehouse and Childs 1990, Hogg and Wein 2005). Nitschke and Innes (2008) found that drought was a major mechanism driving species response under climate change scenarios in southern British Columbia. Minor climate and disturbance events that may not affect adult plants can have a significant impact on seedlings (Ibanez et al. 2007). Climate change may impact the timing and abundance of regeneration which could potentially lead to shifts at stand and landscape-level (Walck et al. 2011). 63   The predictions for the Yukon region are that climate change will result in increases in droughts, fires and outbreaks of insect and diseases and that these impacts will pose significant challenges for forest managers (Ogden and Innes 2007ab, 2008). Understanding which vulnerabilities exist, and where, is thus a logical next step in reducing the uncertainty associated with the challenges managers will face under future climate change (Nitschke and Innes 2008). Understanding the response and sensitivity of species at multiple scales can thus be used to provide guidance on how to manage for the risks associated with climatic change (Turner et al. 2003).  In this paper I model the response of key species within the CATT region at both the site- and landscape-level under historic climate and predicted climate change. The objectives of this paper are to:  Link the tree/site-level model TACA (Nitschke and Innes 2008) with the landscape-level model LANDIS-II (Scheller et al. 2007).  Investigate the impact that climate change may have on species regeneration potential across an edaphic gradient using TACA.  Explore the landscape-level impact of climate change and fire on species abundance and distribution at the landscape-level using the LANDIS-II model with establishment driven by the site-level model TACA.  64  3.1.1 Study region: Champagne and Aishihik Traditional Territory, southwest Yukon The study area is located in the CATT, southwest of the Yukon Territory. The study landscape is 53,621 hectares in area, adjacent to Kluane National Park (Figure 3.2). The boundary of the research area/landscape is based on several decision criteria: (i) avoidance of First Nation owned land (known as R-blocks) and Kluane National Park for the establishment of 90 research plots used to parameterize the TACA and LANDIS-II models; (ii) inclusion of harvested, beetled, burned forest and ‗undisturbed‘ forest (i.e., no recorded disturbance <60 years) areas.  The CATT lies in the Boreal Cordillera ecozone including the St. Elias Mountains, the Ruby ranges, and Yukon-Stikine Highlands ecoregions (Smith et al. 2004). The St. Elias Mountains are the dominating feature in the region; they intercept the moisture from the Pacific Ocean, making the study area one of the driest in the Yukon. Based on 1961–1990 climate normals from the Haines Junction airport weather station, the mean annual precipitation amounts to 305 mm; half of the annual precipitation occurs as snow between October and April. The mean annual temperature is -2.9°C. Extreme temperatures range from -53.9 to +32.8°C and occur mainly in the lower valley floors. Frost can be expected at any time of the year. There are between 16 and 86 frost-free days annually (see Figure 3.1).  The soils of the region have been determined largely by glaciation (Krebs et al. 2001, Smith et al. 2004). In the past 3,000 years, advances of the Lowell Glacier have blocked the Alsek River at least five times resulting in the formation of the Neoglacial Lake Alsek. This has caused the repeated flooding of the Haines Junction area, with the most recent occurrence being from 65  1848 to 1891 (Clague and Ramptom 1982). Glacial movements (advances/retreats) have covered most hillslopes with glacial till, and left stagnant ground ice features (Smith et al. 2004). According to Heginbottom et al. (1995), the southwest Yukon is situated in zones classified as sporadic (10–50%) and extensive discontinuous (50–90%) permafrost. Permafrost layers 7–22 meters thickness have been reported (Smith et al. 2004). The maximum permafrost thickness recorded at Haines Junction is 12 m, with a mean near-surface ground temperature of -0.4°C (Burgess et al. 1982); however, most ground is only frozen seasonally (Smith and Burgess 2002). Forest ecosystems in southwest Yukon (CATT) are dominated by white spruce (Picea glauca [Moench] Voss) representing over 85% of the forests, and trembling aspen (Populus tremuloides Michx.) accounting for about 15% of the forests; to lesser extent black spruce (Picea mariana [Mill.] B.S.P.) and balsam poplar (Populus balsamifera L.) are found in these forests. Many forest stands in the region are characterized by an open canopy of mature P. glauca. Willow (Salix spp.) with a moss and lichen community and low shrub groundcover is typical of mature P. glauca forests (Douglas 1974).  66    Figure 3.1: Minimum and maximum temperatures (Tmin, Tmax) [ C] (top), and precipitation (ppt) [mm] (bottom) for the three weather stations, Aishihik, Bison and Haines Junction.  67  Figure 3.2: Study landscape (dashed line) in the Champagne and Aishihik Traditional Territory, southwest Yukon. Haines Junction (60° 46‘ N; 137° 35‘ W). 68    3.2 Methodology  3.2.1 Tree/site-level model: TACA The ecological model, TACA (Tree And Climate Assessment) (Nitschke and Innes 2008), was modified and parameterized for use in the boreal ecosystems of the southwest Yukon. TACA is a mechanistic species distribution model that analyses the response of trees in their regeneration niche to climate-driven phenological and biophysical variables. It assesses the probability that species can regenerate, grow and survive under a range of climatic and edaphic conditions. The modeling of species presence reflects the fundamental regeneration niche of a species, since presence is directly related to establishment (McKenzie et al. 2003). The TACA model was found to perform well in southern British Columbia demonstrating that the model‘s logic is correct (Nitschke and Innes 2008); however, species responses are sensitive to their defined parameter values. A conservative approach was used both by Nitschke and Innes (2008) and in this study for assigning species parameters. The minimum or maximum value (depending on parameter) reported in the literature was assigned as a parameter value. In addition, a multiple scenario approach was also used to define the boundaries of species response under both current and future conditions. The multiple scenario approach is a form of sensitivity analysis that allowed for the incorporation of variability and uncertainty that exist in climate and soil parameters (Nitschke and Innes 2008). The original TACA model developed by Nitschke and Innes (2008) was modified here to incorporate a frost-free period mechanism. Hamann and Wang (2006) found that the annual number of frost days had a significant interaction with 69  observed species. The phenological component of TACA was also improved to increase the interaction between chilling, heat sum accumulation, frost, and budburst, based on Bailey and Harrington (2006). The new phenology component integrates the obtainment of a species chilling requirement with the accumulation of heat sum that then interacts with frost events that delay bud burst and/or cause frost damage after bud burst occurs. In addition, the soil component of TACA was expanded to allow for three different soil types (texture, coarse fragments, and rooting zone depth) to be run simultaneously, enabling the representation of the multiple edaphic conditions across the resource gradient used in this study. The final adjustment to the TACA model was with the termination of the growing season. The growing season was set to end for all species by September 15; however, the growing season could end prior to September 15 if temperatures fall below a species-specific basal temperature (see Nitschke and Innes 2008). The static end date was based on observations of the growing season ending from late August to early September in the region. The species-specific parameters used for TACA are summarized in Table 3.1. 70  Table 3.1: Selected species parameters used in TACA. Tbase is the species-specific threshold temperature; BB is the required heat sum for starting bud burst process; CR is the chilling requirement; Tmin is the lowest temperature a species can survive; Drought represent the percentage of water deficit a species can survive in a season; Frost is a species specific forest modifier; GDD represent the boundaries of growing degree days a species requires to grow (min), and is limiting its growth by heat stress (max); Wet Soils represents the species‘ tolerance to flooding conditions (indicated in the Results and Discussion section as site S4). Species Tbase [°C] BB [GDD] CR [days] Tmin [°C] Drought Frost GDDmin GDDmax Wet Soils Lodgepole Pine 2.9 116 63 -85 0.42 0.9 185.6 3374 0.5 Subalpine Fir  2.6 119 60 -67 0.2 0.9 197.6 5444 0.75 Black Spruce  3 123 56 -69 0.3 0.9 144 3060 1 Engelmann Spruce  3.1 145 49 -45 0.2 0.9 74.4 1344 0.5 Paper Birch 3.7 231 77 -80 0.3 0.9 236.8 4122 0.3 Trembling Aspen 3.5 189 70 -80 0.4 0.9 226.8 4414 0.3 Mountain Hemlock 3.6 214 91 -50 0.25 0.5 273.6 6103 0.3 White Spruce 2.7 147 42 -70 0.34 0.9 129.6 3459 0.5 Balsam Poplar 2.1 93 49 -80 0.13 0.9 126 7852 0.55 Tamarack 2 111 42 -76 0.2 0.9 150.8 3331 0.75 71  3.2.2 Spatially explicit landscape model: LANDIS-II LANDIS-II is a spatially explicit forest landscape model that simulates ecological processes and their interactions across a large heterogeneous area over long time periods, with user-defined spatial and temporal resolutions. LANDIS has been widely used across many different ecosystems with several sensitivity analyses concluding that the model can accurately simulate landscape processes and dynamics provided that the model parameters are logical and reflective of the studied landscape and species (see Xu et al. 2005, Sturtevant et al. 2009, Liu et al. 2010). In LANDIS-II, users define individual site types which represent homogeneous environmental conditions (i.e., soil, aspect, elevation, etc.); site types are aggregated into ecoregions that represent broader climatic zones (Mladenoff et al. 1996, He et al. 1999, Mladenoff 2004, Scheller et al. 2007). Ecological processes simulated in the model include forest succession and the following natural disturbances: fire, wind, and biological disturbance agents (e.g., insects), as well as the anthropogenic disturbance of harvesting. The model simulates regeneration based on establishment coefficients in conjunction with seed dispersal, vegetative reproduction (sprouting), serotiny and light availability at the site/stand level (He et al. 1999, Mladenoff 2004, Scheller et al. 2007). The model operates on variable time steps with the user defining the temporal scale that each process is operated at (Scheller et al. 2007).  Succession is the main ecological process that is simulated in every time step in LANDIS-II. There are three steps in this process: (1) cohort ageing and mortality, (2) computing shade for forest sites, and (3) cohort reproduction by seeding and resprouting. Seed dispersal is a spatially explicit process. Whether a species can successfully become established at a site depends on the seeding distances of a species, and on the seed availability on site. If a species is already on site 72  then establishment depends on its shade tolerance and an establishment coefficient (Mladenoff et al. 1996, Mladenoff and He 1999). The species establishment coefficient at a site is not modeled mechanistically by LANDIS; instead, the coefficients are provided as parameters from a library or other model (He et al. 1999), in this case from TACA.  In LANDIS-II, a dynamic fire extension component allows fire to interact dynamically with vegetation (e.g., fuel types), climate (e.g., fire weather) and topographic features such as aspect and slope (Sturtevant et al. 2009). The fire growth equations of this module are based on the Canadian Fire Behavior Prediction System FBP (Forestry Canada Fire Danger Group 1992). The fire extension component is tightly linked to a fuel component representing a variety of fuel types that are influenced by disturbances, vegetation type, and the presence of dead cohorts (Sturtevant et al. 2009).  Climate change is incorporated in two ways: (i) indirectly, through the outputs of the climate- dependent establishment coefficients modeled using TACA; (ii) directly, by inputting unique fire weather for the different time periods 2020s, 2050s, and 2080s. To allow for the simultaneous use of several establishment parameters, the biomass succession extension (Scheller and Mladenoff 2004) was applied versus the age-only cohort succession extension, as well as the dynamic fuel and fire extensions versus the ―Base Fire‖ extension.  3.2.3 Historic weather data The weather data input was derived from three weather stations: Haines Junction airport (595.3 m a.s.l., 60º 42.2‘ N, 137º 34.8‘ W), ‘Aishihik‗ (967 m a.s.l., 61º 11.998‘ N, 136º 59.503‘ 73  W) and ‘Bison‗ (781m a.s.l., 61º 0.443‘ N, 137º 2.504‘ W). The latter two stations were established in 2008 to facilitate this research. Climate data from Haines Junction (1946–2009) were used to create extrapolated climate histories for the Aishihik Valley for both temperature and precipitation. Daily minimum and maximum temperatures collected from 2008 to 2010 at the Aishihik and Bison weather stations were compared to Haines Junction data for the same time period to develop empirical relationships between the stations. Linear regression analyses were conducted and significant relationships (P ≤ 0.05) with high r2 (> 0.95) were found for both minimum and maximum temperatures (e.g., Haines Junction–Aishihik Min T regression function: y=0.9x-1.8883; Haines Junction–Bison Min T regression function: y=0.9178x-2.2049).  To develop longer-term climate records for daily precipitation in the Aishihik Valley, a kriging interpolation technique (see Weber 1997) was performed using three reference stations: Beaver Creek, Haines Junction, and Whitehorse Airport, which were the only weather stations within the region with a sufficient data set (1969–2006) (data source: <http://climate.weatheroffice.gc.ca>). The empirical temperature and precipitation relationships were used to create a historical climate record for the Aishihik Valley spanning the period 1946– 2009. For each of the three weather stations in the study area a rank and percentile test was applied and used to select the five years representing 90%, 75%, 50%, 25% and 10% for mean annual temperature and annual precipitation, respectively. This resulted in 10 climate scenarios representing the historical range of observed climate variation for the region. The 30 climate scenarios (10 per station) were entered into the TACA model and used in the development of the fire weather inputs for LANDIS-II following Van Wagner and Pickett‘s fire weather equations (1985). Noon (Pacific Standard Time) weather observations were extracted for the Fire Weather 74  Indices (FWI). A wind correction factor was applied for the HJ-weather station to correct for open conditions (wind velocity * 0.60), and to correct for wind measurements taken below a height of 10 meters (Aishihik and Bison weather stations, wind velocity * 1.54) (Lawson and Armitage 2008).  3.2.4 Climate change projections An ensemble of six climate change scenarios were used in the modeling, representing high (A1B) and low (B1) climate change. Projections from the Model for Interdisciplinary Research on Climate‘s MIROC 3.2 (hires), the Hadley Centre‘s HadGEM3 and HADCM1, the Goddard Institute for Space Studies GISS-EH, Max Plank Institute for Climate Research‘s MPI- ECHAM5, and the Meteorological Research Institute‘s MRI CGCM 232A were used to generate the climate scenarios. The data for climate projections were retrieved from the Pacific Climate Impacts Consortium website at <http://pacificclimate.org/tools/select>. The following gives a summary of the climate models and SRES-AR4 emissions used: High: 1. SRES-AR4-UKMO HADGEM3 A1B Run 1 2. SRES-AR4-MIROC 3.2 (hires) A1B Run 1 3. SRES-AR4-GISS-EH A1B Run 3 Low: 1. SRES-AR4-UKMO_HADCM1 B1 Run 1 2. SRES-AR4-MPI_ECHAM5 B1 Run 3 3. SRES-AR4-MRI_CGCM 232A B1 Run 1 75  Due to the study area‘s unique geographic position (St. Elias ranges with a complex topography) and the limited coverage of weather stations (16 for the entire Yukon), the GCM‘s model performance might be impacted, increasing model uncertainty (Bonsal et al. 2001, 2003). The same authors state that temperature is predicted well by most GCMs; however, precipitation in particular is poorly predicted by GCMs in general, and especially so beyond 60ºN.  3.2.5 Soil data and calibration Soil and site characteristics (slope, aspect, elevation, texture, rooting depth, and coarse fragment content) from 90 ecological plots (Appendix B) were analyzed using a top-down cluster analysis. Soil and site variables were standardized using a dissimilarity matrix prior to the cluster analysis, which was conducted using SAS 9.2. The analysis created five soil clusters based on Manhattan distance representing five edaphic soil groups ranging from xeric, sub-xeric, submesic, mesic to subhygric sites (Appendix D).  The study area has been divided into three climatic ecoregions based on topographic factors (e.g., aspect and elevation): Haines Junction, Aishihik and Bison (HJ, A, B, respectively, in Figure 3.3). In order to calibrate the soil parameters of TACA, soil data from the 90 plots were used to create three soil profiles for each edaphic soil group and two regions (HJ and Aishihik Valley). This resulted in 15 soil profiles for the study region. The calibrated soils for the HJ region (5 soil profiles) ranged from sand to organic textures, with rooting zones ranging from 30–60 cm and coarse fragment contents ranging from 5–20%. For the Aishihik Valley (Aishihik 76  and Bison Ecoregions) the 10 soil profiles had textures that ranged from sand to organic textures, with rooting depths of 15–40 cm and coarse fragment contents ranging from 15-35%.   Figure 3.3: Ecoregions (A, B, HJ) and edaphic site types. A (Aishihk), B (Bison), HJ (Haines Junction) are separated to allow better distinction of delineations. This landscape corresponds to the dashed-lined landscape (the study landscape) in Figure 3.2 77  3.2.6 Parameterizing the landscape model LANDIS-II 3.2.6.1 Ecoregions A factorial analysis using SAS 9.2 was performed to relate soil types to topography and the vegetation variables that were recorded at the landscape-level. The analysis identified 8 factors that explained 84% of the variability. The factor ―Topography‖, which was comprised (factorial loads in brackets) of Elevation (0.85198), Slope (0.76062), and Aspect (0.61639), was selected and used to extrapolate the soil types to the landscape-level. To assign a soil type to each cell in the landscape, a Multiple Discriminant Analysis (Manly 1986) was performed to develop a model of topographic centroids; the resulting model had a predictive power of 53%. The model was tested on a random subset of 40% of the 90 samples, resulting in a model fit of 52%. A Fisher‘s discriminant analysis was then performed to extrapolate from the 90 plots to the entire landscape (53,621 cells). This soil map was combined with the climate region map to generate 15 unique ecotypes representing a combination of soil type, topography and climate (Figure 3.3).  3.2.6.2 Initial tree communities Data on forest composition and structure were collected from 90 plots between June and September 2008 (from six research blocks, see Figure 3.2, or Appendix B). Data collected included: species composition, shrub, tree, herb and bryophyte abundance, tree density, age, tree height, diameter at breast height (DBH), coarse and fine woody debris, crown cover and live crown percent. These data were used to parameterize and calibrate the initial communities and biomass parameters for use in LANDIS-II (Appendix E). For extrapolation to the landscape scale, GIS thematic layers from the Government of the Yukon‘s Forest Inventory Database 78  (Government of the Yukon 2004) on dominant tree species and average stand height were used. A coarse filter classifier for landscape disturbances such as bark beetle, fire or harvesting was applied to adjust for maximum age cohort for a given site. For this study, additional non-CATT tree species were included. Cells on the edge of the landscape were populated with tree species that might not be in the current landscape due to migration lag, as shown for Pinus contorta (Johnstone and Chapin 2003), or due to habitat unsuitability (Figure 3.4). In the west, these consisted of the remaining four Yukon tree species: lodgepole pine (Pinus contorta Dougl. ex. Loud. var. latifolia Engelm.), tamarack (Larix laricina [Du Roi] K. Koch), subalpine fir (Abies lasiocarpa [Hook.] Nutt) and the deciduous paper birch (Betula papyrifera Marsh.). P. contorta and B. papyrifera are represented by a few individuals adjacent to the study region (e.g., within the Kluane National Park (Douglas 1974)). Two additional species from northern British Columbia were placed on the southernmost edge of the study landscape: Mountain hemlock (Tsuga mertensiana (Bongard) Carrière) and Engelmann spruce (Picea engelmannii Parry ex. Engelm.). Both species occur at higher latitudes (i.e., Alaska) and show a high level of frost tolerance (Klinka et al. 1998). Figure 3.4 represents the current forest type distribution in the landscape.  79    Figure 3.4: Current dominant forest types in the study area. 80  3.2.6.3 Fuel type assignments and fire calibration Fuel types were assigned according to the Canadian Fire Behaviour System (Forestry Canada Fire Danger Group 1992) to account for the 10 tree species placed in the CATT landscape. The following fuel types were defined: C2 for mature conifer stands; C3 for conifers in the pole or sapling stage (Foote 1983); C4 representing a regenerating conifer stand; D1 for deciduous sapling and mature stands; O1a representing the seedling/herb stage of a deciduous stand; M2 for mixed stands; D1 has been assigned for the non-forested land types classified as alpine; and non- productive (Government of the Yukon‘s Forest Inventory Database, Government of the Yukon 2004) representing shrubs like alder (Alnus spp.) or American dwarf birch (Betula glandulosa Michx.).  To best reflect the initiation of the region‘s fire season, the fire season was assumed to begin in the spring when all snow had melted and end in the autumn with the presence of a continuous snow pack. The spring fire season start Julian Dates were 138, 132, 128, 127 for historic, 2020s, 2050s, and 2080s climatic scenarios, respectively. The autumn end Julian Dates were 304, 306, 309, 313, respectively. Summer fire season was defined as lasting from June 1 to August 31. The seasonal proportion of fire ignitions was based on an analysis of more than 100 point fire ignitions (Appendix F) for the region from 1943–2004 from the YTG (2004): 19% spring, 73% summer, 8% autumn.  Due to a paucity of fire data (only five fires have been recorded in the area since 1948), I calibrated the fire return interval to reflect a mean fire return interval of 150–200 years, with an 81  average fire size of 200–450 hectares according to Francis‘ (1996) fire study of the Shakwak Trench, and on Hawkes‘ (1983) study of Kluane National Park and Reserve.  3.2.7 Factorial design and response variables At the site-level, a balanced two-way Analysis of Variance (ANOVA) was performed to test the impact of site type (e.g., dry, mesic, moist) on the establishment probability of the 10 tree species on the CATT landscape, under historic and future low and high climate change conditions. Each scenario was replicated three times.  At the landscape-scale, a 2*2 factorial design (two ecological treatments * two climate change scenarios (historic climate, future high climate change)) using a balanced factorial ANOVA was used to assess the effect of treatment (e.g., no disturbance (succession only) vs. disturbance (fire)) and climate (historic versus climate change) or the interaction of treatment and climate on the succession outcome over a period of 200 years. For this, response variables such as forest type (P. glauca forest, P. tremuloides, boreal mixedwood, deciduous forest), species abundance (P. glauca, P. mariana, P. tremuloides, P. balsamifera, P. contorta), and seral stage distribution (early, mid, mid-late, late seral) were assessed at ten years time steps. Results are shown only for the years 0, 20, 50, 80, 150 and 200. LANDIS-II is a stochastic model and produces outcome that is very variable; hence, each landscape scenario was replicated 10 times in order to reduce model uncertainty and reflect natural range of variability (Mladenoff et al. 1996, Liu et al. 2010).   82  3.3 Results 3.3.1 Climate and site effects on establishment Factorial two-way ANOVAs for each of the three ecoregions (Aishihik, Bison, Haines Junction) revealed significant effects of site (F (0.05, 4, 105) = 2.45), climate (F (0.05, 6, 105) = 2.17) and the interaction of site and climate (F (0.05, 24, 105) = 1.61). Site was represented by five site types ranging from xeric to subhygric, while climate was represented by historic, low and high climate change for the periods 2020s/50s/80s. The results for the three ecoregions are presented in Table 3.2. Site type showed a significant effect for all the tree species in the three ecoregions. Climate in the Haines Junction ecoregion showed a significant effect on establishment probabilities for six tree species; only P. tremuloides, B. papyrifera, P. mariana and A. lasiocarpa seemed unaffected by climate in this ecoregion. In the Aishihik and Bison ecoregions, climate had a significant effect on all tree species except P. glauca for the Aishihik, and P. mariana for the Bison ecoregion, respectively. Four significant interactions between site type and climate were found for the Aishihik (A. lasiocarpa, P. balsamifera, B. papyrifera, T. mertensiana), three for the Bison ecoregion (A. lasiocarpa, P. balsamifera, T. mertensiana), and two for the Haines Junction ecoregion (T. mertensiana and Picea engelmannii) (Table 3.2).  83  Table 3.2: Observed F-values for the factorial ANOVA with corresponding P-values; n.s. = not significant. Response variables: Pl=lodgepole pine, Bl=subalpine fir, Sb=black spruce, Se=Engelmann spruce, Ep=paper birch, At=trembling aspen, Hm=Mountain hemlock, Sw=white spruce, Ab=balsam poplar, Lt=tamarack.      Aishihik  Bison  Haines Junction  class F observed P F observed P F observed P Pl treatment 197.05 <.0001 211.71 <.0001 44.03 <.0001  climate 3.74 0.0028 13.78 <.0001 7.35 <.0001  treatment*climate 1.78 0.0324 1.47 n.s. 1.06 n.s. Bl treatment 177.48 <.0001 105.51 <.0001 204.62 <.0001  climate 285.42 <.0001 344.19 <.0001 0.28 n.s.  treatment*climate 9.01 <.0001 14.26 <.0001 1.02 n.s. Sb treatment 182.32 <.0001 307.74 <.0001 143.13 <.0001  climate 1.12 n.s. 3.13 0.0089 1.92 n.s.  treatment*climate 0.51 n.s. 0.68 n.s. 0.33 n.s. Se treatment 190.37 <.0001 159.39 <.0001 145.09 <.0001  climate 24.05 <.0001 25.27 <.0001 6.42 <.0001  treatment*climate 0.83 n.s. 1.17 n.s. 2 0.0134 Ep treatment 180.56 <.0001 218.35 <.0001 162.4 <.0001  climate 76.95 <.0001 282.44 <.0001 2.17 n.s.  treatment*climate 1.64 n.s. 7.23 <.0001 0.54 n.s. At treatment 176.26 <.0001 176.79 <.0001 47.94 <.0001  climate 8.29 <.0001 39.56 <.0001 3.46 0.0047  treatment*climate 1.35 n.s. 0.67 n.s. 1.19 n.s. Hm treatment 41.36 <.0001 17.88 <.0001 38.45 <.0001  climate 177.06 <.0001 70.56 <.0001 71.48 <.0001  treatment*climate 9.13 <.0001 11.66 <.0001 5.25 <.0001 Sw treatment 211.66 <.0001 274.47 <.0001 111.76 <.0001  climate 5.14 0.0002 1.82 n.s. 2.35 0.04  treatment*climate 0.92 n.s. 1.15 n.s. 0.89 n.s. Ab treatment 245.83 <.0001 272.47 <.0001 200.4 <.0001  climate 5.44 0.0001 26.17 <.0001 1.06 n.s.  treatment*climate 1.43 n.s. 1.85 0.0249 0.44 n.s. Lt treatment 270.35 <.0001 218.12 <.0001 233.06 <.0001  climate 3.44 0.0049 3.95 0.0018 2.62 0.0237  treatment*climate 0.93 n.s. 0.22 n.s. 0.48 n.s. 84  For the Aishihik and Bison ecoregions, Bonferroni and Tukey Dunn post-hoc tests revealed significant differences at an alpha ≤ 0.05 for a range of pairwise comparisons of site types. The driest sites (xeric and subxeric) had the greatest effect on the establishment probabilities for all ten tree species (Table 3.3). In the Aishihik ecoregion, a significant effect of climate on P. contorta was detected for the high 2050s and 2080s climate, and for high 2080s on L. laricina. For P. tremuloides, the multiple comparison tests revealed a significant effect between the current and high 2050s climate and the low and high 2050s. Low and high 2020s, 2050s, and 2080s climate had a significant effect on P. glauca and P. balsamifera, and the low and high 2020s and 2050s climate for P. engelmannii. Low and high 2080s climate effects were also detected for A. lasiocarpa. In the Bison ecoregion, climate affected more species than in the Haines Junction area. The low and high 2020s climate showed an effect on all species except P. contorta, T. mertensiana, and P. mariana and P. glauca. The low and high 2050s and 2080s climate affected P. balsamifera, P. tremuloides, T. mertensiana, B. papyrifera, P. engelmannii, A. lasiocarpa and P. contorta. L. laricina were affected by the low and high 2050s climate. For the Haines Junction ecoregion, the multiple comparison tests revealed significant effects of the low and high 2020s climate on T. mertensiana, high 2020s climate on P. engelmannii, low 2020s climate on P. contorta, low 2050s climate on P. glauca, P. tremuloides and P. contorta, and a high 2080s climate effect on L. laricina and T. mertensiana. A significant impact of increases in drought was found on dry sites for all tree species, except for P. balsamifera where the species exhibit higher impacts on moist sites (Table 3.3). 85  Table 3.3: Mean and standard deviation of establishment scores for the three ecoregions (A=Aishihik, B=Bison, HJ=Haines Junction) according to climate (current and projected climate) and edaphic site. Values represent transformed establishment probabilities derived from TACA ((180/PI())*ASIN(SQRT(TACA score)). μ=mean, σ=SD.    Xeric Subxeric Submesic Species Stats Historic 2020s 2050s 2080s Historic 2020s 2050s 2080s Historic 2020s 2050s 2080s Pl μ 35.9 32.5 30.9 31.0 42.0 41.3 39.4 38.0 42.3 43.6 42.7 41.5  σ 2.3 2.7 2.0 2.1 0.6 2.7 2.9 2.9 0.0 0.0 1.3 1.1 Bl μ 3.2 0.9 4.4 10.1 6.3 4.4 15.1 16.9 10.2 9.7 23.3 24.5  σ 2.7 1.6 2.2 2.0 2.7 2.0 3.7 3.2 0.8 2.1 3.3 2.9 Sb μ 24.3 23.0 21.7 21.6 36.2 33.7 31.7 30.3 43.8 43.0 41.2 38.5  σ 2.9 2.2 2.0 2.2 6.6 5.4 5.3 4.5 1.7 3.3 3.9 3.7 Se μ 9.0 3.3 9.9 10.6 13.1 9.4 15.6 17.2 25.9 14.3 24.0 25.1  σ 0.0 0.7 0.9 1.4 5.1 0.6 2.1 4.2 3.8 5.4 3.4 2.6 Ep μ 5.5 10.6 10.8 12.1 9.3 18.7 18.5 19.3 13.9 24.6 24.7 25.6  σ 1.3 1.1 1.1 1.6 2.1 3.3 2.7 3.7 2.5 2.0 3.3 3.3 At μ 32.9 30.0 27.3 28.4 39.8 39.0 34.8 36.6 40.7 42.0 38.8 41.0  σ 3.9 2.1 1.5 3.1 1.7 2.9 3.4 3.6 0.0 0.0 1.2 1.0 Hm μ 0.0 0.0 1.1 2.2 0.0 2.3 2.4 3.6 0.0 3.5 4.6 4.6  σ 0.0 0.0 0.0 0.0 0.0 2.0 0.8 0.8 0.0 0.0 1.3 0.0 Sw μ 34.1 27.7 26.7 24.9 44.2 39.4 37.5 36.7 48.3 46.3 45.1 44.3  σ 3.6 3.1 3.0 2.7 3.9 5.1 4.1 4.0 0.0 1.6 2.5 3.7 Ab μ 9.5 8.4 8.9 9.3 13.0 14.7 14.2 16.6 23.5 23.8 22.7 24.6  σ 0.0 0.0 0.4 0.0 3.7 4.6 3.9 3.8 6.2 5.4 5.0 5.5 Lt μ 16.6 11.3 12.9 13.0 26.7 20.6 22.3 22.5 39.8 33.0 32.8 31.9  σ 4.2 2.6 2.3 2.5 7.1 3.9 4.2 4.4 5.3 6.4 4.6 4.4 86    Mesic Subhygric Species Stats Historic 2020s 2050s 2080s Historic 2020s 2050s 2080s Pl μ 42.3 43.6 43.6 43.2 42.3 43.6 43.6 43.5  σ 0.0 0.0 0.0 0.4 0.0 0.0 0.0 0.0 Bl μ 10.6 16.1 29.7 30.1 10.6 19.9 34.8 37.6  σ 0.0 2.5 2.2 2.7 0.0 0.0 0.0 0.0 Sb μ 44.8 46.5 46.1 44.0 44.8 46.5 47.0 46.6  σ 0.0 0.0 0.8 0.8 0.0 0.0 0.0 0.0 Se μ 29.5 22.3 30.5 29.1 29.5 24.2 33.6 32.7  σ 0.0 1.5 0.9 1.5 0.0 0.0 0.0 0.0 Ep μ 15.4 27.5 28.9 30.1 15.4 27.5 30.1 32.3  σ 0.0 0.0 1.0 0.8 0.0 0.0 0.0 0.0 At μ 40.7 42.0 40.1 42.9 40.7 42.0 40.1 43.2  σ 0.0 0.0 0.0 0.4 0.0 0.0 0.0 0.0 Hm μ 0.0 4.2 5.4 4.7 0.0 4.6 6.5 4.9  σ 0.0 0.6 0.0 0.2 0.0 0.0 0.0 0.0 Sw μ 48.3 47.6 48.0 48.9 48.3 47.6 48.6 50.5  σ 0.0 0.0 0.8 1.4 0.0 0.0 0.0 0.0 Ab μ 32.1 34.5 31.8 33.6 35.1 41.4 42.8 47.0  σ 3.2 5.1 5.3 4.2 0.0 0.0 0.0 0.0 Lt μ 45.2 42.2 41.5 39.2 31.5 30.7 32.1 32.8  σ 0.0 1.8 2.6 2.8 0.0 0.0 0.0 0.0  87  Modeling showed that the four CATT-species (P. mariana and P. glauca, P. tremuloides and P. balsamifera) are present on almost all sites under current conditions, with the exception of P. balsamifera on the driest site in the HJ-region. Under future climatic conditions the CATT species generally performed better, i.e., they show either no response or an increase in establishment probabilities, on all the moist sites (Table 3.3). On xeric sites, all the species show a decline in establishment under future climate conditions. Under current conditions, establishment probabilities (EP) were relatively low and ranged from 0.05 for P. balsamifera to 0.35 for P. contorta; under climate change the range of EP contracted (e.g., from 0.05 for P. balsamifera to 0.3 for P. contorta and P. tremuloides). On the subxeric site, species EP ranged from 0.05 for P. balsamifera to 0.55 for P. glauca under current conditions; while under future conditions EP were similar to xeric sites and less favorable ranging from 0.4 for P. tremuloides, P. contorta and P. glauca to only 0.01 for P. balsamifera. On subxeric sites, all species show the same trend as for the xeric site suggesting a potential decline in conditions suitable for establishment. On sub-mesic sites, P. balsamifera and P. tremuloides performed better under climate change. Also, the EP range was wider compared to the drier sites ranging from 0.14 for P. balsamifera to 0.58 for P. glauca under current climate and from 0.08 for P. balsamifera to 0.51 for P. mariana in the 2080s climate. On mesic sites, all the species exhibited a similar response under current and future conditions, although some (e.g., P. contorta) had improved EPs under climate change. In comparison to the sub-mesic sites, the range of EP contracts to a greater degree from 0.19 for P. balsamifera and almost 0.59 for P. glauca under current conditions to 0.24 and 0.59 under the 2080s climate. On the moist sites, all the tree species show increased establishment probabilities under future conditions. P. balsamifera, in contrast to the drier sites, increases its establishment probability to 0.55 under climate change; P. tremuloides 88  increases to 0.41 (in the Haines Junction ecoregion), and P. glauca reaches the highest modeled establishment probability of any species and climate period, with a value of 0.63.  3.3.2 Landscape-level effects of climate and fire The balanced two-way ANOVA analysis of landscape-level changes in species distributions and landscape structure resulted in fire having a significant effect on landscape composition and structure compared to climate change. The results are summarized in Table 3.4. All the response variables were significantly affected by treatment (fire/no fire) at some point in time during the 200 years of the simulation (Table 3.4). Climate change had a significant impact on forest type (increase in aspen dominated forest in year 150 (Table 3.4), and on two seral stages (e.g., early and mid seral stages, Table 3.4). The interaction between fire and climate had a significant effect on P. glauca and on the early seral stage by year 80 in the simulation (Table 3.4). 89  Table 3.4: Observed F-values for the factorial ANOVA, with Fcritical (0.05, 3, 39) = 2.84. var=response variable; Class=Factor (T=treatment, C=climate, TxC=interaction); ****=p-value <0.0001; "-" indicates no significance.    year 20 year50 year80 Response variable Class F-value p-value Class F-value p-value Class F-value p-value Sw T 29.68 **** - - - T,TxC 91.85,4.64 ****, 0.0379 Sm T 5.46 0.0251 T 21.33 **** T 21.54 **** At T 5.59 0.0235 T 7.54 0.0094 T 38.36 **** Ab - - - - - - T 8.49 **** Pl - - - - - - - - - white spruce T 32.39 **** T, C 99.15,4.24 ****, 0.0468 T 106.64 **** aspen T 40.18 **** T 99.79 **** T,C 142.95, 5.86 ****, 0.0206 boreal mixed wood T 17.15 0.0002 T 41.08 **** T 9.16 0.0045 deciduous C 4.47 0.0414 C 16.29 0.0003 C 6.3 0.0167 early seral T 38.17 **** T 109.76 **** T,C,TxC 131.49,7.95,7.94 ****, 0.0078, 0.0078 mid seral T 25.36 **** T,C 42.88,4.28 ****, 0.0459 T 91.56 **** mid late T 25.93 **** T 28.81 **** T 119.78 **** late seral T 25.52 **** T 88.81 **** T 149.4 **** 90   year150 year200 Response variable Class F-value p-value Class F-value p-value Sw - - - T 35.71 **** Sm T 42.01 **** T 39.09 **** At T 101.7 **** T 171.43 **** Ab T 35.59 **** T 54.87 **** Pl T 4.67 0.0375 T 6.83 0.013 white spruce T 218.63 **** T,C 595.69,5.49 ****,0.0247 aspen T,C,TxC 208.33,8.75,5.04 ****,0.0054,0.0309 T,C 217.41,7.48 ****,0.0096 boreal mixed wood T 16.56 0.0002 T 281.81 **** deciduous T 7.85 0.0081 T 9.87 0.0034 early seral T 76.04 **** T 81.81 **** mid seral T 78.89 **** T 131.87 **** mid late T 50.06 **** T 47.98 **** late seral T 311.33 **** T 404.07 ****   91  P. glauca is currently the dominant tree species in the CATT landscape and throughout the 200 years of simulation remained so (i.e., it occupied >30,000 hectares out of a total 37,004 hectares of forested land; see also Figures 3.4 and 3.5). Climate change in general did not affect P. glauca except in the interaction between climate change and fire treatment for year 80 (Table 3.4). Under fire-only treatments, P. glauca occupied less than 90% of the landscape during the 200 years. P. tremuloides, in contrast, benefited from fire occurrence, occupying significantly more area than under the null-fire (i.e., succession only) treatment; this becomes apparent from year 80 onwards (Figure 3.5). P. tremuloides was not significantly affected by climate change (Table 3.4). Picea mariana responded similarly to P. glauca, i.e., it occupied less area when the landscape was subjected to a fire regime. The fire-adapted species, P. contorta, benefited from fire, especially from year 80 onwards; however, climate change alone did not affect its abundance (Table 3.4). Populus balsamifera responded similarly to P. tremuloides in that it benefited from fire but not climate change. These effects of fire and climate are also reflected in the distribution of forest types (P. glauca vs. P. tremuloides vs. boreal mixedwood vs. deciduous forests). The P. glauca-type dominated the landscape during the entire simulation period (i.e., from >27,000 ha in year 0 (see Figure 3.4) to >21,000 ha in year 200, Figure 3.6). Under the null fire treatment, P. glauca forest occupied less than 70% (i.e., <26,000 ha) of the landscape until year 150, and after 200 years it covered >80% of the total area. Under a fire treatment, P. glauca forest occupied less than 70% of the area throughout the 200 years of simulation (Figure 3.6). P. tremuloides forests, in contrast, almost doubled the area they occupied under a fire treatment (i.e., from ca. 3,000 ha in year 0 to ca. 5,000 ha in year 200, Figure 3.6), and showed a significant positive response under climate change from the years 80 and 150 (Table 3.4), and to the interaction between fire and climate for the year 150 onwards. The area occupied by boreal 92  mixedwood forest decreased by almost 100% under the null fire treatment. Under the fire treatment, however, the area it occupied remained the same (Figure 3.6). Climate change had no effect on the boreal mixedwood type. For the deciduous forest type, climate had more effect during the first 80 years; whereas fire had more effect on the forest distribution from year 150 onwards (Table 3.4).  The seral stage distribution provides a coarse metric of the forest stand age structure of a landscape. Following the null fire treatment, at year 20, the mid-seral stage (forest stands between 20 and 80 years) and the late seral stage (>140 year old stands) occupied around 40% and 50% of the landscape, respectively. Late seral stage forest increasingly dominated the landscape over time, as would be expected, and after 200 years occupied nearly 100%. Under a fire treatment, late seral forests never reached 80% cover. For early seral stage forests (i.e., stands less than 20 years old), both fire and climate*fire-interaction had a positive effect by year 80 (Table 3.4). Fire also had a positive effect on mid- and mid-late seral stages (see Appendix G for averages and standard deviations). 93   Figure 3.5: Distribution of five tree species [hectares] over 200 years and four treatments: Forest succession under current climate (dark line), projected climate (dotted dark line), current climate and fire disturbance (grey line), projected climate and fire disturbance (dotted grey line). 94   Figure 3.6: Distribution of dominant forest types [hectares] according to the four scenarios (from left to right: control (i.e., succession under current climate), climate change, fire, fire*climate change) and time (first bar = year 0, second bar = year 20, year 50, y80, y150, y200).    Figure 3.7: Change in site moisture for Bison ecoregion. Y-axis represents a percentage of the current.. 95  3.4 Discussion  3.4.1 Site-level responses Changes in climate can affect the relationships among the controlling variables and processes that define and drive an ecosystem (Nitschke and Innes 2008). Changes in climate can ameliorate or exacerbate a species‘ response to disturbance; this is especially so during a species‘ regeneration phase (Bradshaw et al. 2000, Spittlehouse and Stewart 2003). Species have a minimum temperature threshold below which frost can kill an individual (Burton and Cumming 1995). In this study frost was found to limit species establishment. For example, under historic conditions, Tsuga mertensiana could not establish at any of the study sites due to cold temperatures. However, under future high climate it could potentially establish in the region. The ability for this species to become an abundant species on the landscape however is unlikely given that it had very low establishment scores even under high climate change scenarios due to the regular occurrence of spring and fall frost events and periodic occurrence of killing frosts. Climate change is expected to shift the timing of bud break to earlier dates in the season, thereby increasing a species‘ exposure to spring frosts (Aitken et al. 2008). In this study, changes in mean temperatures resulted in some species performing more poorly than expected as increases in temperature pushed bud burst earlier into spring and/or extended the growing season later into the fall, leading to increases in frost damage. Picea glauca in the 2020s and Populus tremuloides in the 2050s exhibited this response.  Although precipitation in the CATT is expected to increase by 5–25%, temperature is expected to increase by 1–4°C between the 2020s and 2080s (Appendix H). The increase in 96  temperature caused an increase in the water-holding capacity of the atmosphere which in this case led to the mean annual precipitation being exceeded by the mean annual potential evaporation. In effect, the increase of evapotranspirational demand of the atmosphere by 9–33% equated to a 5–8% decrease in effective precipitation (Goldammer and Price 1998), which in turn affected soil moisture availability (Figure 3.7) and potentially increased moisture stress in some species. Drought is another limiting factor that can affect the distribution of a species (Hogg and Wein 2005). Figure 3.7 shows that all sites other than the subhygric sites may exhibit an increase in soil moisture deficits under climate change (especially the driest sites in the Aishihik and Bison climate regions). Nitschke and Innes (2008), for example, found that a modeled reduction in soil moisture content could force species to contract from lower elevations at the landscape level and from drier to wetter sites at the site-level. In this study, the landscape-level effect was not detected but a site-level effect was evident for many of the species that were modeled. All species exhibited declines in establishment ability on the xeric sites. P. glauca was found to be negatively affected on the driest site types, which is supported by Hogg and Wein (2005) who found that this species is particularly sensitive to soil moisture deficits during the regeneration phase. Delays in regeneration in the region are known to occur following fire if climate is too warm and dry, resulting in a regeneration lag ranging from years to decades (Hogg and Wein 2005, Brad Hawkes pers. comm.). The regeneration modeling results in this study showed that P. glauca performed best on mesic to moist sites; in Alaska this species dominates well-drained floodplains and upland soils (Wirth et al. 2008) which supports the species response modeled in this study. In the western boreal of Canada, P. mariana is commonly found on organic imperfectly drained (hydric) wetland depressions (Cumming et al. 1996, Hall et al. 1997). In this study, P. mariana was modeled to establish across sites that ranged from having slightly dry to 97  wet soils, a finding supported by Klinka et al. (2000) who identified that P. mariana occur on site types in British Columbia ranging from slightly dry to very wet soils.  P. balsamifera typically occupies sites in riparian zones but can be found on soils that range from being slightly dry to very moist (Klinka et al. 2000). Yarie (2008) identified that P. balsamifera in Alaska had higher growth rates on floodplains than on adjacent upland sites. Both, Shaw et al. (2001) and Yarie (2008) have found that P. balsamifera productivity is affected by soil moisture while even mild soil moisture deficits will prevent germination (Zasada and Densmore 1980, Krasny et al. 1988). Nitschke (2009) suggested that if upland sites become drier due to climate change P. balsamifera will have to compete increasingly with other species better adapted to such conditions, such as P. tremuloides. Our results suggest that P. balsamifera can be present on drier sites, which was validated by the field sampling, but with increasing moisture deficits expected due to climate change it may be unable to establish on dry to mesic sites, although its regeneration ability on wetter sites may increase. In comparison, P. tremuloides may exhibit increases in its establishment ability on submesic, mesic and subhygric sites. The species exhibited a 13.7% and 6.2% decline in establishment on xeric and subxeric sites, respectively, under the 2080s climate change scenarios. P. tremuloides regeneration is known to be moisture- limited in the southwest Yukon, which can result in slow and patchy regeneration over a period of 40 years (Hogg and Wein 2005). The outcomes of this study suggest that moisture limitations on regeneration may increase on some but not all sites within the study region.  98  Betula papyrifera is found on slightly dry to very moist sites (Klinka et al. 2000). B. papyrifera regeneration is typically higher on moist sites compared to dry sites (Cater and Chapin 2000). In this study B. papyrifera was only able to establish under current conditions on eight out of 15 modeled sites, with only two of them currently representing drier sites. Under future high climate change conditions, however, it was modeled to occupy all 15 sites due to a warmer future. P. contorta is found across a broad range of edaphic regimes, from very dry (xeric sites) to wet (hygric) sites (Klinka et al. 2000). P. contorta had higher establishment probabilities, except under moist conditions compared to the four CATT species (P. glauca, P. mariana, P. tremuloides, P. balsamifera). P. contorta has a higher tolerance to drought (Lotan et al. 1985, Klinka et al. 2000), which allowed it to have higher establishment scores on the drier sites compared to the four CATT species. Abies lasiocarpa can occur across a broad range of edaphic conditions, from dry to wet sites (Klinka et al. 2000), however it was not modeled to occur in the Aishihik Valley except under climate change. A. lasiocarpa was modeled to be present at only five sites (Table 3.3), with low establishment scores on the drier sites, which is supported by Bigler et al.‘s (2007) findings. Under the high climate change scenario, A. lasiocarpa was modeled to occupy all the (15) CATT site types due to warmer temperatures and higher precipitation in the region. Interestingly, A. lasiocarpa showed a preference for wetter sites, which is supported by Peterson et al. (2002) who suggest that moisture is a limiting factor for A. lasiocarpa seedling establishment.  3.4.2 Landscape-scale responses: Fire, succession and climate change Wildfire is the dominant disturbance in many Canadian forests (Flannigan et al. 2005); this is especially the case for the boreal forest (Johnson 1992, Xiao and Zhuang 2007). Fire-return 99  intervals (i.e., the average of return times between fires at a point (Reed 2006)) of 50–150 years are common throughout the Canadian boreal. This means that most forests in a given area are at some stage of recovery from a fire (i.e., in a successional stage) (Johnstone et al. 2004). Generally, tree recruitment is associated with disturbance events that remove established individuals and open gaps in the forest canopy, which allow new individuals to become established.  Picea glauca regeneration over areas affected by large-scale disturbances (e.g., bark beetles or fire) is strongly affected by distance from undisturbed edges or surviving trees within the disturbance patch, with little regeneration more than about 100 meters from a source tree (Greene and Johnson 2000). In contrast to fire-adapted conifers, P. glauca does not have serotinous cones, but instead relies on periodic masting to ensure adequate seedling regeneration after a disturbance (Greene et al. 1999, Danby and Hik 2007). The degree to which mast years correspond to periods of disturbance can have a strong effect on spruce regeneration (Peters et al. 2005). LANDIS-II currently does not take masting into account. Another factor affecting successful regeneration is the seedbed condition. A disturbance may change substrate quality (Bhatti et al. 2002), which, coupled with competition with shrubby understory vegetation, may limit establishment in cases where the disturbance leaves large amounts of residual vegetation or litter on the soil surface. The occurrence of severe fires is more likely during years of reduced precipitation and higher temperatures (e.g., increased drought) (Hogg and Wein 2005), and hence can slow down the regeneration of P. glauca (Hogg and Wein 2005). All these factors can result in high variability in the recovery success of P. glauca after a fire disturbance resulting in heterogeneous, patchy landscapes consisting of uneven-aged stands (Hogg and Wein 2005). The 100  seral stage distribution suggests that fire promotes heterogeneity, as proposed by several authors (e.g., Franklin et al. 2002, Noss et al. 2006).  Succession in Picea glauca stands can be ‗spruce to spruce‘, where an understory of the same species succeeds the dying overstory. The other successional pathway can be ‗aspen, or aspen/willow to spruce‘ (Garbutt et al. 2006). The latter pathway is common after stand- removing disturbances but only if there are available aspen and willow seed sources. Once the spruce overstory has died (e.g., due to beetle attack), spruce regeneration is likely to compete for dominance with common deciduous shrubs such as Salix and Betula spp. (Garbutt 2004). Fire is an important disturbance that allows B. papyrifera, P. balsamifera and P. tremuloides to regenerate successfully provided that soil moisture is not limiting (Johnstone and Chapin 2006a- b). This was confirmed by our modeling, where P. balsamifera and P. tremuloides occupied a greater area when fire was present in the landscape.  Calef et al. (2005) predicted that with 5 ºC of warming, the area of deciduous forests could expand by 354 to 500% within Interior Alaska, replacing P. mariana and P. glauca forests. They also modelled that deciduous forest could increase by up to 233% if fire frequency increases, leading to the replacement of P. glauca forests. Interestingly, in my analysis, deciduous- dominated forests increased by only 12.5% under climate change/no fire, 214% under no climate change/with fire and 306% under climate change/with fire. The latter two estimates are in line with the findings of Calef et al. (2005) but the former demonstrate the importance that transient dynamics may play in vegetation responses to climate change. P. glauca forests were found to 101  decline by 0.3% under climate change/no fire, by 15.6% under no climate change/with fire and by 18.2% under climate change/with fire. Calef et al. (2005) predicted that a 2.8 °C increase in temperature and 6 % increase in precipitation would result in a 78% decline in P. glauca forests in Interior Alaska. However, my study suggests that changes in P. glauca forests are likely to be less extreme. Interestingly, Calef et al. (2005) predicted that P. glauca forests would decline by 10 % due to an increase in fire with no climate change and in this study we modeled a 15.6% decrease under no climate change/with fire. P. mariana was modelled to decline by 3.5% under climate change/no fire, by 25.6% under no climate change/with fire and by 29.7% under climate change/with fire, which is in contrast to the predictions of Calef et al. (2005) for Interior Alaska. Possible explanations include an insufficient change in the fire regime, and that the drier site conditions favour P. tremuloides over P. mariana and P. glauca, with P. glauca still outcompeting P. mariana (the results of Calef et al. (2005) are based on a statistical approach and competition was not considered).  Fire is also important in maintaining boreal mixed forests. In the absence of fire, the spruce component typically outcompetes the shade-intolerant and shorter-lived deciduous species (Burns and Honkala 1990). In this study, after 100 years of succession the mixedwood evolved into a pure spruce forest due to the lack of hardwood regeneration (Figure 3.6). However, our results also show that under a null-fire treatment, P. glauca is likely to dominate the entire landscape. This is rather unlikely to happen given that P. glauca requires an exposed mineral layer for abundant regeneration (which typically occurs after a fire disturbance) (Astrup et al. 2008). LANDIS-II currently does not take changes in soil substrate due to disturbance or lack thereof into account. 102   Fire is an important ecological process for the regeneration of P. contorta. Astrup et al. (2008) state that fire events can maintain or increase the dominance of P. contorta in a landscape as they create open canopy conditions (i.e., creating gaps for light) and expose mineral substrates, both being requirements for pine regeneration. Like P. mariana, its serotinous cones require heat to open and release seed. In the Yukon, recruitment rates are highest in the first five years after a fire event (Johnstone et al. 2004). Fire also benefits P. contorta in that it exposes mineral soil that would otherwise be a limiting factor for its germination. With Pinus generally being a weak competitor, fire removes competition by grass and deciduous tree species (Bradshaw et al. 2000), and opens up gaps in the canopy, increasing light conditions (pine being a less shade-tolerant regenerator (Klinka et al. 2000)). Lotan et al. (1985) suggest that with the absence of fire, P. contorta would likely be outcompeted by more shade-tolerant species such as A. lasiocarpa. In this study, climate had no significant direct effect on P. contorta (Table 3.4); however, our results indicate that under increased fire occurrence due to climate change, P. contorta could occupy almost six times the area compared to conditions under a null fire treatment after 200 years, or three times the area under historic climate conditions. Interestingly, it would take pine almost 150 years to increase its area of occupancy significantly (Figure 3.5).  Hamann and Wang (2006) modeled P. contorta distribution using a biogeoclimatic envelope technique. They found that by the 2080s pine will generally expand to higher elevations and latitudes. This study supports the findings of Hamann and Wang but also highlights that range shifts in the area will be a transient process governed by disturbance (e.g., fire). Our results 103  indicate that species benefiting from fire, such as P. tremuloides, P. balsamifera and P. contorta, will be less affected directly by climate change. With the increasing area burned under climate change, these three species might even increase their frequency on the landscape. This is consistent with Aitken et al. (2008) who state that species showing pioneer attributes might occupy new areas more quickly under favorable climatic conditions. However, given the relatively short time period (200 years) explored in this study, the full impact of climate change and successional dynamics may not have been captured. Still, the study highlights that it will take longer than 200 years for climate change and fire to overcome the present dominance of Picea glauca on the CATT landscape. The increase in fire activity on the landscape would increase the heterogeneity of both species composition and seral stages, which could benefit biodiversity and also reduce the susceptibility to bark beetle (e.g., see Howe and Baker 2003). The recent outbreak of the spruce bark beetle in the CATT occurred when the landscape was dominated by older, mid- to late seral, P. glauca forests (SFMP 2004).   3.5 Conclusions  To better understand forest dynamics and impacts of climate change, transient dynamics and interactions between species and disturbance need to be considered at multiple scales. The southwest Yukon is a cold and dry region which is expected to be warmer and wetter in the future. These changes will reduce growing season frosts and provide a warmer and longer growing season which will benefit species establishment and subsequent growth on mesic to moist sites however not on xeric sites where drought is expected to limit species establishment. 104  The direct effects of climate change on species at the site-level were not detected at the landscape-level; instead, climate change in interaction with fire drove species responses. A 20% increase in area burned was modeled to occur under climate change. Interestingly this study found that white spruce is likely to remain the dominant tree species in the Champagne and Aishihik Traditional Territory over the next 200 years even with marked shifts in climate. This is in contrast to the findings of Calef et al. (2005) who modeled a large decline in white spruce forests in neighboring Alaska. Modeled changes to the region‘s fire regime suggest that disturbance will increase the sensitivity of the region‘s forests to climate change and that the dominance of white spruce in the region will be reduced in favor of tree species with pioneer characteristics such as Populus tremuloides, P. balsamifera and Pinus contorta leading to an increase in their abundance at the landscape-level. Interestingly such an increase in fire may reduce the landscape‘s susceptibility to future spruce bark beetle infestations.  Local forest managers and practitioners in the region have concerns about potential climate change due to increases in fire occurrence, a repeat of the recent spruce bark beetle epidemic and failures in forest establishment. This study has identified that the interaction between species, climate and fire needs to be consider when assessing the impacts of climate change on forest landscapes which in turn may impact forest management. Fires will likely increase in the region which will drive forest succession in favor of deciduous species which will subsequently have impacts on forest biodiversity, carbon, timber and cultural values within the region (see Chapter 2 for locally important values). The increase in fire will likely increase community risk to wildfire but also may reduce the extent of future spruce bark beetle outbreaks within the region. These interactions at the landscape may require management actions that 105  reduce fire risk around communities (see Chapter 4, Desired State) but will also require consideration of species at the site-level as species will exhibit divergent responses to disturbance and management at this scale. The outcomes of this study should help the Yukon (YTG) and Champagne and Aishihik First Nation (CAFN) governments develop more informed adaptive management strategies within their current SFMP framework (SFMP 2004) allowing a transition from the current reactive/mitigation approach that exists to more pro-active/adaptive management of their forests under future climate change.  106  4. Integrated Dual Filter approach for forest management planning – a synthesis  4.1 Introduction In Chapter 2, I presented the Social Filter of the Integrated Dual Filter. The Yukon working group developed five alternative management strategies for the CATT. The strategy ‗manage for multiple values and use‘ scored the highest in the AHP-synthesis. The strategy ‗manage to reduce fire risk‘ scored the lowest. Possible reasons for this are addressed in Chapter 2. However, the working group emphasized that fire is a major concern in the region, as already prominently expressed in the Strategic Forest Management Plan for the CATT (SFMP 2004). The holistic management strategy will also address fire as an important system component to be taken into account by management, at both the landscape scale and at the community level.  In Chapter 3, I presented and used the Environmental Filter of the IDF to explore system constraints, and to describe possible forest succession trajectories under fire disturbance and under current and projected climate conditions. The results show fire as an important natural disturbance in the region, even if fire occurrence is rarer compared to other regions of the Yukon, as for example shown by McCoy and Burn (2005). Figure 4.1 provides a visual impression of this. The St. Elias Mountains are the dominant topographic feature in the region, intercepting moisture from the Pacific Ocean and being responsible for the reduced number of lightning events in the study region (Northern Climate ExChange 2006).  107   Figure 4.1: Yukon fires from 1948–2004. Study landscape CATT (Champagne and Aishihik Traditional Territory) is shown in black contours in the southwest of the Yukon. Fire data (in grey) based on Forestry Inventory Database of the Government of the Yukon Territory (2004).  An increasing number of natural resource management approaches try to incorporate and address elements and principles that entail multiple disciplines spanning various spatial and temporal scales. They involve a multitude of stakeholders representing a growing array of values and interests from the very beginning, i.e., in the planning phase and throughout the management implementation process (Born and Sonzogni 1995, Holling et al. 1998, Messier and Kneeshaw 1999, Lal et al. 2001, Mitchell and Beese 2002). Lal et al. (2001) see a major challenge in natural 108  resource management research in the testing and applicability of proposed integrated natural resource management frameworks. Forest resource management and planning is also becoming more complex and challenging due to climate, economic and technological changes, which are difficult to predict due to their non-linear nature (Carpenter and Gunderson 2001, Walker et al. 2002). Increasing complexity poses a growing challenge for decision-makers seeking to identify management alternatives or directions that can maximize, or best balance, all the decision criteria due to increased risk and uncertainty (e.g., Mendoza and Martins 2006, Ananda and Herath 2009). In recent years, useful tools and approaches such as game theory, cost-benefit analysis and resilience management have helped to address spatially and temporally complex and integrated, unpredictable system problems (e.g., Grimble and Wellard 1997, Walker et al. 2002, Baskent and Keles 2005). Dealing with complexity means making distinctions and decisions, i.e., favoring some options (or values thereof) over others.  Components of structured decision-making include the following steps: recognizing and defining a problem; formulating management objectives; assessing vulnerabilities and system properties; identifying management options; evaluating these against each other; making a decision; implementing management actions; and monitoring the impacts of the same (Ohlson et al. 2005). Forest management decision support systems implicitly or explicitly incorporate these steps. According to Rauscher (1999), a generic decision support system (DSS) entails a number of subsystems: (i) People (i.e., the managers/decision makers and stakeholders), (ii) spatial-/non- spatial data management, and (iii) knowledge-management (i.e., knowledge base, simulation models, data visualization management). Ultimately, people, representing the first DSS- subsystem, are the most important part of a DSS and will make the final decisions; however, 109  these decision processes require the support of tools where human judgment alone is constrained by the limitations of data processing. The second DSS subsystem organizes data about the natural resource to be managed (e.g., a forest or a landscape), allows the generation of alternative management scenarios and forecasts possible consequences of management activities. The third subsystem deals with knowledge in its many forms, and organizes the quantitative and qualitative data to support the decision-making process. This framework can consist of simulation models, visualization tools, and expert and knowledge-based systems (Schmoldt and Rauscher 1996).  Decision support systems share two features: simplification of reality (i.e., the model representing an abstracted system) and translation from reality to such a model (Bunnell and Boyland 2003). The same authors identified four purposes of DSS, namely (i) to aid in research in describing ―new but true relations‖, i.e., the more precisely (or less generally) a relationship is described the more confidence there is that it will inform decisions; (ii) to guide and support management  through robust relationships and to anticipate possible trajectories; (iii) to convey knowledge from experts to practitioners and stakeholders about the long-term dynamics of the system under scrutiny; and (iv) to facilitate publicly the evaluation of management tradeoffs in order to receive a ―social license‖ (i.e., socially accepted management decisions) (Bunnell and Boyland 2003). In order to increase credibility for public evaluations, the models and tools employed in DSS should be scientifically rigorous (Kangas et al. 2000). Seely et al. (2004) suggest that the models should be based on both empirical and mechanistic relations so that they can better address the possibly broad needs of decision support. Decision support systems developed in North America include the McGregor Model Forest ECHO-planning system 110  (McGregor Model Forest Association 2001), which is based on the interplay of three Models to address each specific/different spatial and temporal scales. Model A deals with the highest planning level overseeing strategic planning scales (long-term and landscape scales), whereas Model C deals with the highest planning resolution (short term operations at the site level). Models A and B use non-integer goal programming while Model C uses a stochastic meta- heuristic simulated annealing technique for optimization of forest plans. Another DSS is the NED-2, used by the USDA (e.g., Rauscher et al. 2000) to develop goals and measure current and future conditions for SFM in support of resource planning. This goal-oriented framework combines a set of tools such as wildlife and vegetation growth models and GIS with knowledge databases. Seeley et al. (2004) have developed a DSS for British Columbia, which integrates stand-level simulations (FORECAST, Kimmins et al. 1999) with forest-level simulations and harvesting schedules (FPS-ATLAS, Nelson 2003), habitat modeling (SIMFOR, Daust and Sutherland 1997) and a visualization system (CALP, Cavens 2002).  In this chapter I present the development of a new decision-support framework, the Integrated Dual Filter (IDF) approach, combining the Social and the Environmental Filters with the three States (the respective concepts are introduced in Section 1.7). The approach also enables the assessment of current and future conditions; during this process I bring these conditions into juxtaposition to enable the planning process to address questions such as ‗what is the (current) situation?‘ (by using the Environmental State), ‗what do we want (in the future)?‘ (by using the Desired State) and ‗how to do (or achieve) it?‘ (by using the Management State). To do this, I use a set of established tools and techniques such as decision-making and re- combine them with scenario-based approaches based on simulation modeling. The tools used in 111  this IDF should enable management questions at different scales to be addressed. When describing IDF‘s structure and components I will make reference to decision support systems (DSS). The DSS was designed to aid resource managers and practitioners in strategic management planning. A case study is presented in which the IDF framework was applied to illustrate how the different components of the IDF interact in a planning process for the Champagne and Aishihik Traditional Territory (CATT), southwest Yukon. The use and application of the two filters (Social and Environmental) is briefly summarized as they are presented in more detail in Chapters 2 and 3, respectively. The focus of this chapter is on the demonstration and development of the three states, the Environmental, Desired, and Management States, as well as on the adaptive iterative process that brings the E towards the D State by using the M State. To assess management consequences at the landscape I have developed a set of indicators that are based on discussions with the Yukon working group and that lean towards the indicators of the SFMP (2004) and ILP (2007).   4.2 Description of the IDF decision support framework  4.2.1 The Dual Filters One of the main reasons for the growing complexity in forest management is the large number of economic, ecological, and social values that are relevant to forest management. The simplification of these values provided by models (or incomplete representation of reality) is a prerequisite to considering and addressing the relations within a complex system (Bunnel and 112  Boyland 2003). The Integrated Dual Filter (see Figure 4.2) uses two filters that allow the summarizing and grouping of many of these values: The Environmental Filter represents all ecological values and the Social Filter is the umbrella for all socio-cultural/-economic values. The term ―filter‖ is based on the classic management approach with ―coarse‖ and the ―fine‖ filters referring to spatial scales (e.g., Hunter 1990, 1991, 1999, Kremsater et al. 2003). The terms Environmental Filter and Social Filter are based on the resilience management vocabulary describing socio-ecological systems (Walker et al. 2002). The use of the term environment also emphasizes the explicit incorporation and addressing of so-called ‗system externalities‘ or external forces such as climate change (see Figure 4. 2), since planning systems, like natural resource systems, are not closed systems (Oreskes et al. 1994). The Environmental and Social Filters both implicitly incorporate different scales (i.e., coarse and finer scales). For example, the Environmental Filter can address single trees (fine scale), a forested stand, or an entire landscape (coarse scale), while the Social Filter can address temporal scales such as short-term tactics or long-term strategies. Both filters can address the short- and long-term fine and coarser spatial scales of a resource system. As a result, the Environmental and Social Filters can reduce the planning complexity to two dimensions. This allows the forest manager to focus on one or the other filter at a time and eases the identification of important values (e.g., system drivers), the formulation of system processes and mechanistic relationships, and the quantification of vulnerabilities and system resilience.  4.2.2 Forest State Space and its States An ecosystem could be described as an n-dimensional space (O‘Neill et al. 1986, Kay 1991). Similarly, in the IDF planning framework, the forest state space F can describe specific forest 113  states, depending on the perspective from which one is looking at a forest. The IDF approach uses three different forest states to describe the forest state space F: E = Environmental forest state, representing the biophysical forest system (e.g., with no forest management); D = Desired forest state, reflecting a socially desired forest state; and M = Management forest state, representing a managed forest system. Each state of the forest state space F can be represented as a landscape map such as a GIS thematic layer. If more simplicity is required for a first planning phase, each state can be represented as a landscape map at a given time t where all its components (or data layers) are combined. For example, the E (environmental) State is like a landscape describing the biophysical conditions (any existing infrastructure such as roads or power lines on the landscape is ignored). The D (desired) State is like an ‗engineered‘ landscape or a desired condition to be achieved. It is designed to represent the social requirements/desire (e.g., the landscape is structured/modified in GIS to serve a certain function such as reducing the area burned). The M state represents a ‗managed‘ landscape or a management condition where the focus of the resource planning is on management tactics and operations. Note that the purpose of the IDF approach is not to understand a state as the end point of a trajectory. For example, the Desired State should not be associated with Rauscher‘s notion of management, which he defines as ―making decisions about and controlling systems to achieve desired ends‖ (Rauscher 1999: 179). Rather, the IDF enables a description of the evolution of the state components through time. This allows a comparison of all three states at any given time t. Each of the three states enables the planning to focus on one state at a time to increase the understanding and/or knowledge of the planning system. Each state has its own type of ‗planning question‘: The E State deals with the question ‗where do we start from?‘; the D State addresses the question ‗what do we want?‘, and the M State uses the question ‗how do we do it?‘. The 114  disentangling of the forest state space F into three states enables the scrutiny of a complex problem by keeping the planning focus for example during a model simulation run on environmental system properties (E State) only, or to emphasize the management (tactical) aspects of the planning landscape (M State).  4.2.3 The IDF iterative process The Integrated Dual Filter approach adheres to the principles of adaptive management (in its broadest sense following the four steps of planning, acting, monitoring, evaluating) with the IDF- iterative process consisting of two running modes: a ‗real world‘ mode and a ‗virtual‘ mode. The former consists of on the ground tactics and operations and monitoring at the stand/landscape scales. The virtual mode allows for scenario-based studies of the relationships between climate change, (virtual) operations/disturbances and their long-term (virtual) effects. Hence, the virtual mode enables a manager to learn and acquire understanding of the long-term consequences of management decisions. This scenario-based approach helps assess which of the scenarios favors certain management approaches over others, as all decision-making processes involve some sort of forecasting techniques (Baskerville 1985), and allows the identification of system vulnerabilities (Berry et al. 2006).  The IDF iterative process is the study (or planning process) of how to bring the Environmental State closer to the Desired State through management (i.e., through the formulation of an M State). For this, each IDF (planning) cycle starts from the E State (the current biophysical information) by implementing a certain set of management tactics onto the 115  landscape; by doing so, the E State (i.e., a natural landscape) becomes the M(i) State (i.e., a managed landscape). In continuing in the IDF iterative process, the M(i) State is then compared with the D State (an engineered landscape). Each further management modification requires additional comparison between D and (a new) M State (M(i+1)). This ‗adaptive cycling‘ ends when similar values are reached, e.g., the M(n) State approximates the D State (within the F state space, the E State is been transposed onto the D State).  Figure 4.2: Integrated Dual Filter framework, showing the two Filters (Environmental and Social) and the three States (E = Environmental, M = Management, D = Desired). Each Filter contains a Toolbox that allows the development of a Library. Arrows indicate direction and flow of information. All components in fine white lines indicate that current tools presented here can be replaced with tools deployed in other decision support systems and planning environments. The Resource Planning and Decision Environment is not a closed system, i.e., it is open to external forces (e.g., climate change).  116  4.3 Application of the Integrated Dual Filter  The Integrated Dual Filter approach (IDF) comprises four parts (Figure 4.2): (i) an ecological modeling tool kit (the Toolbox in the Environmental Filter) that uses a set of ecological simulation models to perform forest succession simulations by varying input parameters such as harvesting operations or natural disturbances such as fire under climate change; (ii) a decision- making tool (i.e., the Analytic Hierarchy Process (AHP) of Saaty (1977, 2001), which is used to balance certain values (assigned by users such as the Yukon working group) to reach sustainable forest management (see Toolbox in the Social Filter, Figure 4.2); (iii) a database library in each of the two Filters; and (iv) a monitoring tool (a set of Criteria and Indicators) to measure management effects and impacts on the landscape. These tools are used in the IDF process to develop the forest states (Figure 4.2) and are also employed during the iterative adaptive cycle to bring the E State closer to a defined D State by using the Management (M) State. In the following section the specific use (e.g., data type, parameterization, and timing) of the respective tools within the IDF framework is presented using a case study based on the Champagne and Aishihik Traditional Territory (CATT) of southwest Yukon Territory, Canada.  4.3.1 Social Filter The Strategic Forest Management Plan (SFMP) of the Champagne and Aishihik Traditional Territory (CATT), a framework for sustainable forest management, was developed during a 10- year community-based process. In its current form, the SFMP utilizes a hierarchical planning approach (as defined by Jeakins et al. 2006) to initiate the salvage harvesting of beetle-killed 117  white spruce (Picea glauca) and fuel-abatement treatments to reduce current and future fire risk around local communities.  Chapter 2 analyzed and balanced potentially contradicting values listed in the SFMP by applying an Analytic Hierarchy Process (AHP) and discussed desired outcomes and local thresholds with a group of stakeholders (referred to here as the Yukon working group) to develop a set of socially embedded forest management strategies (manage forests for timber industry; manage forests for multiple values and use; for fire risk reduction; for wildlife; for carbon industry). These alternative forest management strategies are listed in the Social Library of the IDF (see Figure 4.2). Using a ratings table (see Table 2.1 in Chapter 2), practitioners and experts characterized five alternative forest management strategies stemming from the SFMP. The AHP overall priority for the alternative strategy ‗manage for multiple values‘ (the holistic strategy) had the highest score and therefore ‗best‘ balanced the potential conflicting SFMP-values in achieving sustainable forest management in the CATT (see Table 2.4 in Chapter 2). To illustrate the application of the IDF approach, I chose one of the five formulated alternatives (listed in the Social Library, Figure 4.2), ‗managing for fire risk reduction‘, which incorporates one of the SFMP‘s key objectives. The goal of this alternative strategy is to reduce fire risk at the landscape scale; the objective is to focus management activities on stands surrounding the communities (i.e., outside community boundaries; see the interface and community zones highlighted in Figure 4.3, which follow the boundaries delineated in the Integrated Landscape Plan for the CATT (ILP 2007)). The Yukon working group, which consisted of ca. 50% of FN- representatives, placed special emphasis on respecting and maintaining cultural values such as opportunities for hunting and trapping, and also highlighted the importance of maintaining the 118  current level of biodiversity within management areas. For example, moose is a very important local cultural value because of its significance for hunting. Identified tactics for this alternative strategy entail fuel treatments around the communities. These tactics are related to the Fire-smart management paradigm that emphasizes the use of forest management practices to reduce the risk of fire by altering fuel composition and structure in order to decrease fire behavior potential and increase the ability to successfully suppress fires (Hirsch et al. 2001, 2004).  Figure 4.3: Study landscape in the CATT showing areas of fuel treatments. Dark grey = fuel treatment; grey = communities; light grey = forested cover type; white = non-forested cover type. Row 1: left to right = 10-30% scenario; row 2: left to right = 40%, 50%, 100% fuel treatment. Percentage of treatment refers to forested area in the management zone; Aishihik Valley (northeast part of the landscape) is conservation zone (i.e., no management) according to Integrated Landscape Plan (ILP 2007). 119  4.3.2 Environmental Filter The Environmental Filter represents the environmental dimension to resource management. To describe the current environmental conditions of the CATT, I used empirical ecological data derived using locally tailored monitoring protocols (Yukon Forestry Monitoring Program 2008). These data were then used for parameterizing and calibrating the LANDIS-II model (for specific details refer to Chapter 3), which is a landscape eco-model that enables the simulation of forest succession under different natural and anthropogenic disturbances on large spatial and temporal scales (Mladenoff 2004, Scheller et al. 2007). Another used model in the IDF was TACA, a mechanistic aspatial, scale-less model (Nitschke and Innes 2008) that analyzes the response of trees in their regeneration niche to climate-driven phenological and biophysical variables. Another model, CFFDRS (Canadian Forest Fire Danger Rating System) (vanWagner 1987, Stocks et al. 1989) is a modular fire danger rating system with two primary subsystems – the Canadian Forest Fire Weather Index (FWI) System and the Canadian Forest Fire Behavior Prediction (FBP) System. The FWI system is calculated from daily weather observations (temperature, relative humidity, rain, wind) and uses FWI indices such as Fine Fuel Moisture Code, Duff Moisture Code, Drought Code, amongst others to predict fire behavior. Both TACA and CFFDRS outputs serve as libraries for LANDIS, by creating species-specific establishment coefficients, and expected fire spread and intensities (for details see Chapter 3). Both TACA and CFFDRS outputs serve as input libraries for LANDIS (Figure 4.2), with the former creating species-specific establishment coefficients for current and projected climate, and the latter fire weather indices.  120  LANDIS-II is a spatially explicit forest landscape model that simulates ecological processes and their interactions across large heterogeneous areas over long time periods, with user-defined spatial and temporal resolutions. Users define individual site types that represent assumed homogeneous environmental conditions (i.e., soil, aspect, elevation, etc.). Site types are then aggregated into ecoregions that represent broader climatic zones (Mladenoff et al. 1996, He et al. 1999, Mladenoff 2004, Scheller et al. 2007). Ecological processes simulated in the model include forest succession and natural disturbances such as fire, wind, and biological disturbance agents, as well as the anthropogenic disturbance of harvesting. The model simulates regeneration based on establishment coefficients in conjunction with seed dispersal, vegetative reproduction (sprouting), serotiny and light availability at the site/stand level (He et al. 1999, Mladenoff 2004, Scheller et al. 2007). The model operates on variable time steps with the user defining the temporal scale that each process operates at (Scheller et al. 2007).  The dataset for the CATT case study landscape encompasses an area of 53,621 hectares. In Chapter 3, I concluded that climate change would have important impacts at the site level in terms of altering edaphic conditions that could lead to shifts towards drought-tolerant species on drier sites and the maintenance of the current species on moist sites. At the landscape-level, fire would play an important role in shaping a heterogeneous landscape. For example, the models suggested that mixedwood forests would require the occurrence of fire to maintain at least 20% occupancy of the landscape over a period of 200 years. Also, aspen forests could increase their extent under the current fire regime, but could decline in the absence of fire. Fire will be an important disturbance, maintaining the forest in a heterogeneous state by promoting forest renewal and reducing the area covered by late seral stage stands of white spruce by about 20%. 121  However, 200 years of succession and fire disturbance would be insufficient to overcome the dominance of white spruce in the study landscape.  In the Integrated Dual Filter approach, the Environmental E State represents the CATT landscape under no management conditions (with infrastructure such as houses, roads, power lines being ignored). In LANDIS, only forest succession and fire/fuel are active modules in order to compare the E State (or its driving variable ‗area burned‘) with a Desired D State. To develop a D State, I first needed to find a threshold where fuel treatment begins to have an impact on the landscape.  4.3.3 Engineering the Desired State Climate change is expected to lead to increases in forest diseases, drought events, and forest fires (Flannigan et al. 1998, Fleming and Candau 1998). Fire occurrence and area burned are expected to increase by two to three times, respectively, in the Yukon under future climate change (McCoy and Burn 2005). In the study area, white spruce is the dominant tree species and will likely continue to dominate over the next two centuries (see Chapter 3). The area burned will likely increase under projected climates (Chapter 3); fire will remain a major potential threat in the region. Due to the expected increase in fire occurrence, some local forest practitioners within the CATT have identified that the current use of fuel abatement through thinning may not reduce fire risk in the long-term. They have argued that the only effective management action should involve the conversion of stands around the communities from white spruce to deciduous species, as this will better provide for community safety. 122   In the Social Filter part of the IDF ‗managing for fire risk reduction‘ is one of the alternative forest management strategies for the coming 150 years developed by the Yukon working group. Although it was the lowest ranked strategy (Chapter 2), I have used this strategy in the following to showcase the IDF process (i.e., engineering/designing a Desired State; testing whether such a designed landscape could be also achieved through management actions (M State). According to Parisien et al. (2006), a fuel treatment entails modifying the flammable portion of a forest stand to cause a change in the physical fire behavior; ideally this modification will achieve a decrease in fire intensity or severity, or its rate of spread Parisien et al. (2006). From a fuel treatment perspective, the aggregation of less flammable fuel types at the landscape scale will likely lead to a reduction of fire spread potential (Loehle 2004, Parisien et al. 2006). For the placement of fuel treatments, i.e., the conversion of coniferous to aspen stands (Figure 4.3), the following criteria were followed: (1) Highly combustible fuels targeted: mainly mature, beetle-killed white spruce (Hirsch et al. 2004, Parisien et al. 2006); (2) landscape features used as fuel-breaks: pure deciduous or mixed wood stands, wetland, alpine, rivers and lakes, roads (Mermoz et al. 2005, Parisien et al. 2006); (3) area of most fire ignitions (lighting and person-caused) since 1943 (data from Government of the Yukon 2004); (4) prevailing wind directions (i.e., W-E); 123  (5) fuel management restricted to the green management zone (ILP 2007) and areas outside the community/interface zone targeted; touching on the wildland-urban interface zone only (CAFN Fuel Treatment Project 2005, ILP 2007).  It is important to note that in the IDF planning approach the engineered (fuel treated) landscapes as presented in Figure 4.3 are not modelled yet, they have been designed in GIS. In a next step, these landscapes were input to LANDIS using the fire module to assess whether such a treatment would show an effect (i.e., is there a threshold?) in terms of area burned. A balanced factorial ANOVA was used to assess the effect of treatment (e.g., no fuel treatment (0% stand conversion), i.e., the Environmental State, versus 10, 20, 30, 40, 50, 100% stand conversion) and climate (historic versus climate change), or the interaction of treatment and climate, on the area burned over a period of 200 years. Each landscape scenario was replicated 10 times in order to reflect natural variability (Mladenoff et al. 1996, Liu et al. 2010). In the Desired State the stand conversions (e.g., 10%, 20%, …,100%) were assumed to be distributed across the landscape and started to function as fuel treatments at time zero of the LANDIS model-simulation.  4.3.4 Management State and the adaptive cycling The Management State describes the management tactics/silvicultural techniques applied to engage in the adaptive iterative process of the Integrated Dual Filter approach. This means that the tactics that were deployed were developed to bring the Environmental State (i.e., the area burned during the entire planning period) towards a Desired State (i.e., significantly reduced burns at the landscape scale compared to the Environmental State). The M State starts from the E 124  State‘s initial conditions (i.e., the same initial tree cohort distribution) and applies harvesting tactics that bring the driving variable (i.e., cumulative area burned) closer to the Desired State level. For this, I developed harvesting interventions that mimic the fuel treatment of the engineered D State, with the difference being that in the M State I tried to reach the same fuel treatment structure through harvesting (different harvesting intensities to reach the 20% landscape stand conversion from white spruce to aspen stands [identified in 4.4.1]). The M1 scenario distributes the harvesting of 5,920 hectares over a period of 150 years (i.e., the least intense harvesting approach), the M2 scenario distributes harvesting over a period of 100 years and the M3 over a period of 50 years. (For a description of the harvesting distribution over time for M1 to M3, respectively, see Appendix I). Harvest block sizes ranged within the recommendations given by the ILP (2007) (harvest block sizes of 1 to 200 ha) and DIAND (1998) (maximum of 2,000 ha block sizes for the Yukon). The maximum block size was 280 hectares, which is in the range of the mean fire sizes for the area of 200 to 450 hectares, according to Francis‘ (1996) fire study of the Shakwak Trench, and on Hawkes‘ (1983) study of Kluane National Park and Reserve.  4.3.5 Assessment of management actions at the landscape level Criteria and Indicators (C&I) constitute a critical tool for the planning and implementation of sustainable forest management. They enable the effectiveness of forest management to be evaluated through an assessment of progress towards established goals and objectives. C&Is can be used to assess, monitor and report on sustainable forest management, enabling planning directions and actions to be assessed. C&I can also be used to guide forest management planning (Karjala and Dewhurst 2003). According to Natural Resources Canada (2011), an indicator 125  constitutes a ―measurable (quantitative) or descriptive (qualitative) variable that can be used to observe trends as a criterion changes over time‖. As ecosystems and socio-cultural systems vary between planning landscapes, C&I need to be defined and implemented at different scales if sustainability is to be assured (Karjala and Dewhurst 2003). Table 4.1 gives an overview of the few selected C&Is used in this study to demonstrate the assessment of landscape level forest management interventions (e.g., the M1, M2, M3). These indicators are based on discussions with the Yukon working group (see Chapter 2), and also stem from the Strategic Forest Management Plan. The SFMP (2004) represents a locally drafted planning framework with C&I proposed by the CATT communities that also include FN perspectives. Appendix J represents an excerpt of indicators identified during the participatory assessment for the SFMP development by the YTG, CAFN and ARRC; Appendix K details the development of the indicators listed in Table 4.1. All the indices calculated for the M and D States were normalized to reflect change compared to a baseline (the E State), with values ranging from -1 (for maximum decrease) to +1 (for maximum increase). The normalization removed any units and enabled a better visual assessment of change and comparison within and between the different indicators. 126  Table 4.1: Criteria and indicators used in this case study. Square brackets indicate the origin of the indicator (e.g., from Yukon working group discussions, or from SFMP document, etc.). Criteria Indicator Description Significance fire area burned cumulative area burned a major concern in the CATT [SFMP 2004, Yukon working group] forest health leafminer risk classes of leafminer attack on aspen and poplar currently the biggest outbreak in the CATT [Yukon working group, Yukon FHR 2009]  spruce bark beetle risk classes of beetle attack on white spruce recent spruce bark beetle outbreak was the biggest in Canada [SFMP 2004, Yukon working group, Yukon FHR 2009] forest types white spruce   seral stage distribution (young, mid, mid- late, late) dominant coniferous [SFMP 2004, Yukon working group]  aspen  dominant broadleaf [SFMP 2004, Yukon working group  mixedwood  reflects forest renewal [Yukon working group]  spruce-pine  pine can become interesting under climate change; currently not in the region [Yukon working group]  boreal black and white spruce the two dominant coniferous in the CATT [SFMP 2004]   127  4.4 Results  4.4.1 Fire risk reduction – developing the Desired State Figure 4.4 shows the cumulative area burned in hectares during 200 years for each scenario (with the Environmental State or no management condition being represented by the 0% treatment, 10–100 = percentage of landscape that has been fuel-treated). The ANOVA showed a treatment effect (F=4.35, p<0.001) for log10-transformed fire data regarding cumulative area burned. There were no significant interactions between climate and treatment. Post-hoc Bonferroni tests confirm a significant difference between the E State (no fuel treatment) and the fuel-treated scenarios involving 20–100% stand conversion, and between the 10% treatment and the other treatments (20/30/40/50/100% fuel-treated landscapes). These results suggest that an effect (i.e., reduced area burned) occurs at a threshold at 20% of the treated landscape (Figure 4.3). Given the stated objective of the minimum fuel treatment around the communities, this scenario served as the Desired State in the further planning process of the Integrated Dual Filter approach.   128   Figure 4.4: Fuel treatment scenarios: Cumulative area burned over 200 years [hectares]. Comparison of fuel treated area (x-axis = percentage of treated area; see Figure 4.3 for respective maps). 0%-treatment corresponds to the Environmental State; 10–100% constitute the potential Desired State.   4.4.2 Harvesting – the IDF adaptive iterative process The result of the IDF adaptive process is shown in Figure 4.5. The cumulative area burned during a period of 200 years under the no-management condition (i.e., the Environmental State) can be reduced and brought close to the Desired State through the Management State. M1 is the least intense harvesting intervention (i.e., fuel treatments spread over a 150-year period) while M3 is the highest intensity management intervention (i.e., stand conversion accomplished within 50 years). All three management scenarios targeted the same area, but M3 harvested 16,000 m 3 , M2 20,000 m 3 , and M1 removed 28,000 m 3 due to the ageing white spruce stands across the landscape.   0 5000 10000 15000 20000 25000 30000 0 10 20 30 40 50 100 climate change current climate 129   Figure 4.5: Integrated Dual Filter ‗adaptive cycling‘. M1—M3 are representing the Management State that bring the Environmental (E) State closer to the Desired (D) State in terms of cumulative area burned [hectares] over 200 years.   4.4.3 Environmental State and change over time Figure 4.6 shows the response of a range of indicators to management actions aimed at reducing fire risk (i.e., cumulative area burned, under the Desired and Management States). The values in Figure 4.6 reflect the degree of change of a specific indicator (i.e., normalized values ranging between -1 (=maximum relative decrease) to +1 (=maximum relative increase) compared to the baseline with absolute values measured at the E State (in Table 4.2)).  Overall, white spruce forests (Figure 4.6a) decline under the Desired and the Management States (M1-M3). Aspen forests (Figure 4.6b) generally increase on the landscape due to the stand conversion from white spruce to aspen stands. In the M1 scenario, stand conversions occur 0 5000 10000 15000 20000 25000 30000 E M1 M2 M3 D climate change current climate 130  within the first 150 years, in M2 within 100 years, and in M3 within 50 years. This is also reflected in the increase of the M1 values compared to the E State (i.e., index values >0); almost the entire 150 years show a positive change compared to the baseline (the E State, see Table 4.2) in all the seral stages. M1 seems also to be the scenario that will promote the most mixedwood on the landscape (Figure 4.6c).  When compared to the Environmental State, pure coniferous stands decreased in the study area in all the scenarios and through the 150 years of simulation. The reason for this likely lies in the spatial distribution of the species. I populated the eastern margins of this landscape with pine for the climate change study (described in Chapter 3), which was then harvested and replaced by aspen. However, the TACA model suggested that pine could reach and become established in the Aishihk valley. This area was not targeted for management interventions as it is in the conservation zone of the CATT (according to ILP 2007 plan) (see Figure 4.3).  Overall, there were smaller changes in black and white boreal spruce (BWBS) stands, which may be due to the spatial distribution of these stands (they are restricted to the Aishihik Valley) and hence were not targeted by management. Interestingly, there seemed to be renewal of BWBS forests at all times, and this induced frequent fire disturbances (Figure 4.6e). The fuel treatment to reduce fire risk had a long-lasting change at the landscape level as it reduced the white spruce dominance in the Environmental State (Table 4.2) and favored a more mixed forest at the landscape scale and increased diversity (for more details see Chapter 3). 131   Spruce bark beetle risk decreased under all scenarios due to the large amount of aspen replacing white spruce in the landscape (5,910 hectares of total conversion). Beetle risk increased again after 150 years (but only for the low risk class). The reverse was true for leaf miner risk, as the risk increased throughout the study period and landscape due to the increased presence of aspen in the landscape. M3, like the D State, resulted in a drastic increase in aspen during a relatively short period (M3 within 50 years, D from the very beginning of the simulation runs) and hence changed the landscape to more favorable conditions for forest defoliators such as the serpentine leaf miner.  a) White spruce forests:      132  b) Aspen forests:    c) Mixed forests:      133  d) Spruce-pine forests:    e) Black and white boreal spruce forests:    134  f) Spruce bark beetle risk:     g) Leaf miner risk:  Figure 4.6: Normalized indices (a-g) representing change compared to a baseline over a period of 150 years (i.e., the Environmental State, see Table 4.2 for absolute values): Values > 0 show an increase, values < 0 show a decrease in the respective index compared to the baseline; a value of 0 indicates no change.  135  Table 4.2: Indices calculated for the Environmental State; all are given as a proportion of the forested landscape. Proportional distribution of forest types (e.g., all types listed sum to 100%); Sw=white spruce, At=aspen, Mix=mixed forests, S-P=Spruce-Pine forests, BWBS=Black and White Boreal Spruce forests; 1-4=early, mid, mid-late, late seral stage; L=Leaf miner risk, B=Spruce bark beetle risk, l=low, m=medium, h=high.  E0 E20 E50 E80 E150 Sw1 25 20 1 1 1 Sw2 5 1 18 16 1 Sw3 36 33 14 4 14 Sw4 8 14 32 39 43 At1 6 6 5 6 6 At2 1 0 4 4 3 At3 1 2 1 0 2 At4 0 0 0 1 0 Mix1 7 12 1 2 3 Mix2 0 0 12 13 4 Mix3 9 10 6 1 13 Mix4 0 0 4 10 5 S-P1 0.0 0.1 0.1 0.1 0.3 S-P2 0.0 0.0 0.1 0.2 0.1 S-P3 0.0 0.0 0.0 0.0 0.2 S-P4 0.0 0.0 0.1 0.3 0.2 BWBS1 0.0 0.0 0.0 0.0 0.0 BWBS2 0.1 0.0 0.1 0.1 0.0 BWBS3 0.0 0.1 0.0 0.0 0.2 BWBS4 0.1 0.1 0.1 0.1 0.2 Lh 16.6 20.3 16.3 17.2 19.1 Lm 7.2 23.5 15.6 17.5 17.2 Ll 3.5 4.5 3.2 0.6 3.7 Bh 10.8 11.8 18.8 32.4 31.8 Bm 21.3 22.1 20.4 21.2 26.1 Bl 21.1 35.7 34.6 23.5 18.1 136  4.5 Discussion  A hierarchical approach has been recommended to the solution of integrated planning problems (Jeakins et al. 2006) since it encourages managers to organize data and knowledge into appropriate levels of planning (Barber et al. 1996) or, as Connelly (1996) defines it: ―[hierarchical organization] is the organizing of information for making decisions at different levels, when the quality of the decisions at one level is dependent upon decision or information at other levels…levels may be defined temporally or spatially, where the scope of the higher level fully encompasses the scope of the lower level‖. In this case study, the Social Filter with its tool box (Figure 4.2) informs decision makers (in the center of Figure 4.2) of major concerns and desires. The Social Filter helped the organization and structuring of data (e.g., values stemming from the SFMP (SFMP 2004) into alternative forest management strategies (listed in the Social Library, Figure 4.2), and formulating long-term emphases, objectives and goals (see Chapter 2). Although the ‗holistic‘ management strategy received the highest scores (detailed in Table 2.1, Chapter 2) and would be the desirable one to follow in a real-world scenario, I have chosen ‗manage for fire risk reduction‘ to illustrate the completion of the Integrated Dual Filter approach to avoid confounding factors that might distract from the IDF. For this, I have tested whether a threshold in the landscape exists that would show an effect of fuel treatment. As Parisien et al. (2006) noted, to show an effect at the landscape level, an enormous amount of coniferous forest needs to be converted: 20% of the manageable forested landscape (the so-called green management zone according to the delineations shown in the Integrated Landscape Plan for the CATT, 2007).  137  I have not considered economic costs in this study as the main purpose was to illustrate the planning process and step-wise completion of the IDF. The Environmental State demonstrated that the landscape is and could be dominated by white spruce even under projected climate conditions. Fire plays an important role in forest renewal and is responsible for increased heterogeneity in the region (Chapter 3). To reduce fire risk, the stand conversion also introduced increased heterogeneity at the landscape scale throughout the entire planning period. Fire risk was significantly reduced, as less area was burned, and beetle risk was reduced due to the removal of white spruce. The recent bark beetle outbreak, which lasted for almost 15 years (from 1992 to 2006), affected over 85% of the CATT white spruce forests (Berg et al. 2006, SFMP 2004).  There seems to be a recurring pattern: every 70–80 years the risk peaks (Figure 4.6), indicating that resource availability may be more important than climate change in the future if climatic conditions stay favorable for the beetles‘ increased survival and reproduction rate, as has been suggested by several authors (e.g., Berg 2003, Berg et al. 2006, Garbutt et al. 2006). The fire risk reduction management action removed a lot of white spruce and hence reduced the fire risk significantly. However, the spruce was replaced by aspen, which favors a defoliator, the aspen serpentine leaf miner (Phyllocnistis populiella). Currently, an ongoing outbreak that extends beyond the range of historic variability in the southwest Yukon is affecting every deciduous tree in the area (Yukon FHR 2009). Usually, the defoliation does not kill the trees, but if the mining affects both the upper and lower epidermis layers, the tree will shed its foliage three to four weeks earlier than usual (Wagner et al. 2008)). At high latitudes, this constitutes a 138  significant shortening of the growing season. Younger cohorts are at greatest risk, and 20–25% of these trees may die (Yukon FHR 2009).  An aspect not addressed in this study is related to the consequences of management (i.e., the M States) on wildlife. Adding almost 6,000 hectares of aspen in lieu of coniferous species onto a landscape could potentially affect the habitat quality of different key wildlife species. For example, moose (Alces alces) has been identified as one of the most important species in the area for hunting (see Chapter 2). Moose are large, wide-ranging herbivore (Ahlén 1975) that require a mixture of habitat attributes to maintain a viable population and are commonly found in riparian areas and deciduous stands (Courtois et al. 2002, Dussault et al. 2006). Deciduous tree species and shrubs are used for food year round; mature coniferous stands are used to reduce the risk of predation (reduced visibility) and are seasonally used as cover to shelter them from snow in winter. Dussault et al. (2006) highlight the importance of the spatial distribution of these attributes, and provide a habitat-suitability model to assess trade-offs between open areas, food, and shelter opportunities. This model could be linked with LANDIS using its species and age distribution output. One of the main moose predators in the area and a species of regional importance in terms of tourism attraction and hunting (see Chapter 2) is represented by the grizzly bear (Ursus arctos). The grizzly bear is a wide-ranging omnivore that uses a variety of seral stages and habitat elements (Carroll et al. 1999, McCormick 1999). Assessing grizzly habitat based on food distribution would require using LANDIS output in combination with shrub and other vegetation mapping to be linked in order to allow assessment of consequences of management actions as shown for M1–M3. This could reveal important information and increase the understanding of management actions on wildlife. For example, the thinning as fuel 139  treatment within the community boundaries of Haines Junction (see CAFN Fuel Treatment Project 2005, or ILP 2007) has opened up stands which allowed shrub species to increase their abundance; as a consequence, the increased availability of berries attracted unusually more grizzly bears into the communities posing increased risk to the local human population (Brad Hawkes, pers. comment).  The three management scenarios used here represent different harvesting intensities; the M1 scenario distributes harvesting interventions over 150 years, and is compared to the D State; then M2 that runs harvesting interventions over 100 years is compared to the D State; and finally, M3 runs over 50 years only, and is compared to the D State. The M3 scenario brings the E State the closest to the Desired State, i.e., there is the least cumulative area burned over the 200 years (Figure 4.4). However, M1 distributed harvesting over the longest period, allowing the growth of younger cohorts and resulting in higher harvesting outputs compared to M3 or M2. Planning is a continuous process, and values and social desires can change during the course of a planning horizon. Social values are assumed to have an influence on planning decisions, as changes in values can lead to changes in decisions (Dietz et al. 2005). Hence, during a planning process, the Desired State can be re-formulated, and M States then need to be re-adjusted accordingly. For example, the advantages of M1 (stand conversions spread over 150 years) or even M2 (stand conversion within 100 years) over M3 is that during the longer management timelines (100–150 years), such changes could be integrated.  140  People and institutions are an integral part of the planning process (Schmoldt et al. 2001). Planning and decision-making environments (e.g., Figure 4.2) are integrated socio-ecological systems. Hence, they need to be treated as single rather than separated systems (Chapin III et al. 2004). The IDF, with its social and environmental filters, enables the simultaneous integration of a society within its ecosystem. The IDF enables the identification of system knowledge through the communication and exchange between the two filters (indicated by the top arrow in Figure 4.2). For example, the Environmental Filter conveys information about disturbance as a coherent part of the CATT system that is required to maintain a heterogeneous landscape and the renewal of the forests (Chapter 3). The Social Filter identifies concerns (e.g., fire risk) and formulates alternative strategies, with management emphases, goals and objectives to be evaluated and tested with empirical data (see Figure 4.2, or Chapter 2).  As with the Social Filter process of the IDF, in socially-based resource management planning it is important to respect the participants‘ opinions (CCFM 2003). It is also important to address social sustainability within a planning process in order to allow the forests of the future to maintain and increase their social functions, and to allow the communities depending on these forests to have fair and transparent management processes for their forests (Pukkala 2002). This is also important if decision makers are to acquire the ―social license‖ described by Bunnell and Boyle (2003). For this case study, I combined a participatory approach linking non-expert input (from the CATT community level, which also included First Nation perspectives and interests) condensed in the SFMP document, with input from practitioners and experts (Yukon working group), and expert researcher opinions (UBC) (see Chapter 2). This enabled the development (characterization and filtering) of alternative forest management strategies from an already 141  existing plan (i.e., the SFMP), and the formulation of clear and desired directions for SFM in the region to feed the Social Library (Figure 4.2). The Social Filter was a starting point for the IDF (e.g., giving the strategic direction and coarser management tactics). In this thesis, I was able to show how the IDF process with its two filters and three states works, but I was not able to answer the question whether the engineered D State really was able to meet/achieve the social requirements because the formal process was completed. This would allow the stakeholders and practitioners to provide feedback and voice their opinions on this outcome in order to enable re- adjustments of the Desired State, with consequent re-adjustment of the Management State (e.g., building new M States to bring the E closer to the new D State); the Environmental State would still be the same in this case.  The long-term productivity and functionality of a landscape requires a management approach that is also adapted to climate change if uncertainty is to be reduced and resilience maintained (Holling 2001, Gunderson et al. 2002). Adaptive management, like the shifting of the E State to the D State using the M State, is a systematic approach and reiterative process (e.g., building M1 and comparing with D, then building M2 and comparing with D, then building M3, etc.) that allows management to proceed in a complex, uncertain biological and socioeconomic environment across different temporal and spatial scales. However, it requires continuous learning, a reiterative evaluation of goals and approaches, and redirection based on an increasing information base and changing public expectations (Baskerville 1985). The same author states ―learning can only proceed by the identification of mistakes‖ (Baskerville 1985:172). Along the same lines, a success story seldom allows us to learn since we do not want to change it. Only when we make mistakes are we keen to change system parameters or rethink the role of values 142  that allow us to reach back to the good trajectory (Carpenter and Gunderson 2001). The paradigm of adaptive management is certainly useful (and assumed to be widely integrated in management planning) if we allow the planning and decision-making system the time to accept mistakes. However, in the fast pace and short-term mandates of today‘s decision-making environment, we rarely have time for such reflection. Although learning constitutes more than just a paradigm in resource planning, there seem to be very little room for learning or for exploratory or non-conventional approaches.  The primary goal of this study was to demonstrate a new decision support framework, the Integrated Dual Filter approach. A relatively simple scenario-set (e.g., manage for fire risk reduction) was selected to remove confounding factors and to emphasize and focus on the development and demonstration of the individual IDF process steps such as the Desired State or the Management State, or the process of bringing the Environmental State closer to the Desired State. I believe that the conceptual structure of the new decision support system with the Environmental and Social Filters, and with the three States (Environmental, Management and Desired) presented in this study is sound and sufficiently robust to withstand further testing. However, est modus in rebus (there is a meaning in all things); the frameworks‘ performance is only as good as its components (e.g., used tools), and a model is only as good as its input data and the number of scenarios that can be examined (Bunnell and Boyle 2003). For example, I have not demonstrated here that bottom-up and top-down communications (e.g., from stand to landscape and vice versa) can reveal important understanding for managers. For this, further testing is required using finer-scaled tools that can be linked with the current ones. Basically, all parts of the components presented in Figure 4.2 that are depicted with fine white lines can be 143  exchanged. Depending on the system/region under scrutiny this means that we could also take the IDF core components/concepts (e.g., filters and states) and export them to another context beyond the southwest Yukon. For example, instead of using the Yukon working group for the expert and practitioner input, or instead of using the Strategic Forest Management Plan and its associated documents, other existing and locally tailored and corroborated databases could be deployed and incorporated. There are many parts of the world for which such information is available, especially in countries such as Canada that have detailed planned regulations and where forest management certification is widely practiced. For example, it would be interesting to test the IDF framework using the strategic forest management plan of the Teslin and Tlingit Traditional Territory that has been developed by the Yukon Government and the Tlingit First Nations in 2007 (TRRC 2007). This region lies in the south-central part of Yukon Territory, and its planning currently does not take climate change (e.g., a so-called external force in the IDF) into account. This may be especially interesting to test given that the recent spruce bark beetle (Dendroctonus rufipennis) outbreak stopped just short of the Teslin area, and there is a general concern that the spruce forests of the south-central Yukon may be next affected. Potentially aggravating the beetle risk in the Teslin is the considerable amount of lodgepole pine that is highly vulnerable to the mountain pine beetle (Dendroctonus ponderosae). So far, the pine beetle is only present in British Columbia (BC), south of the Yukon Territory, but it could reach Yukon along the Rocky Mountain Trench (Yukon FHR 2010). In BC, model forests have been established to acquire long-term research for SFM. For example, the John Prince Research Forest in north-east BC would be an ideal context to validate the IDF for several reasons. The forest is co-managed by the University of Northern British Columbia and the Tl'azt'en First Nation, representing a unique opportunity to test the IDF with all the tools used in this research where 144  the different management scenarios could be assessed using the locally developed traditional knowledge framework (see Karjala et al. 2004). Since it is located in the Sub-Boreal Spruce (SBS) biogeoclimatic zone the parameterization of TACA and LANDIS-II would be similar to the CATT.  LANDIS-II is the main tool used in the IDF and presented in this study. LANDIS represents a powerful tool that enables the simulation of forest succession with consideration of different natural and anthropogenic disturbances, and their interaction under climate change. This was useful especially in an area like the CATT to increase the understanding of such interactions, and to address the paucity of Yukon specific species and system knowledge. However, LANDIS is based on a number of parameters; for example it requires at least eight vital attributes for each species used in the simulations, and more than 15 parameters have to be adjusted for the disturbances of interest. A further complication is that LANDIS is based on rasterized maps that require sophisticated preparation techniques since specific cell-based initial conditions (e.g., see the number of Initial [tree] Communities in Appendix E) are needed for input (He and Mladenoff 1999). The IDF framework, as with all DSS, is a simplification of reality. To add more complexity, we need more tools, but in adding them we also risk adding more errors (Bunnell and Boyle 2003). This needs to be taken into account in any future use of the system.      145  4.6 Conclusions  Resource management problems are complex and pose a challenge to forest managers and decision makers. An integrated approach is advisable if values are to be balanced. For solving complex problems, oftentimes relatively simple concepts can help understand system processes and properties so that socially acceptable, informed decisions can be made. Such decisions are often better than those made based on greater uncertainties and higher risks.  This study constitutes a first round of a planning process in which alternative management strategies are developed through a Social Filter and possible tactics are addressed. The Environmental Filter provided a first impression of system trajectories under current and projected climate conditions. The IDF adaptive process with the E, M and D States enabled a potential decision maker to be informed (i.e., placed in the center of the Figure 4.2) and to evaluate the array of planning scenarios shown here for the alternative strategy of fire risk reduction. A next step would be to go back to the working groups and define thresholds and even finer resolution goals, objectives and tactics in order to re-define/adjust the Desired State and then modify the M State for the ‗adaptive cycling‘. I believe that due to its conceptual simplicity, this IDF decision support system will help to identify system properties, constraints and concerns in order to feed an adaptive management cycle with the three states E, M, D. The IDF constitutes a platform where different knowledge databases and different techniques and tools can be applied and tested in combination for an exploratory forest management planning. Here, I combined AHP with LANDIS and GIS. For further validation I recommend using a different 146  planning context where the IDF can function as a vessel for different/complementing tools to explore more fine-scaled and higher resolution analyzes of a planning problem. 147  5. General conclusions  5.1 Integrated Dual Filter  The primary aim of this research was to develop a new decision-making support tool, the Integrated Dual Filter (IDF) approach, and to illustrate its conceptuality and applicability for resource management planning through the example of the Champagne and Aishihik Traditional Territory (CATT) context. The CATT constituted an ideal research opportunity, as a recent bark beetle outbreak affected over 85% of the spruce forest of the region between 1993 and 2006 (Berg et al. 2006, Garbutt et al. 2006). This insect outbreak has been linked causally to climate change (ACIA 2004). This dramatic environmental change has caused a change in public awareness and attitude towards its forests. Great concerns in the communities have been raised, mainly revolving around their economic future (SFMP 2004), and also about increased fire risk due to the increased fuel loads associated with the dead coniferous trees (SFMP 2004, Garbutt et al. 2006). The governance response has been the development of a management and planning framework, the Strategic Forest Management Plan (SFMP). This strategic plan provides the background to determining and implementing long-term (e.g., 20 years) sustainable forest management in the region that adheres to the principles of adaptive management (SFMP 2004). My research used the CATT context as a starting point for developing a new decision support tool, and provided an opportunity to test it in a case study approach (chapter 4).  Management and planning within socio-ecological systems (Chapin III et al. 2004) such as the CATT can be regarded as a ‗wicked‘ problem (cf. Rittel and Webber 1973). It is a complex 148  and complicated problem (cf. Pietronero 2008) for forest managers and decision makers since they have to deal with the complexity of forest ecosystems, consider public input, be sensitive to the needs of the nascent forest industry, and in the meantime address risk and uncertainty (Church et al. 2000). The tricky part with wicked problems is that they do not have a stopping rule, i.e., when is a solution reached? Such problems cannot be addressed by so-called ‗tame solutions‘-based methods such as deterministic models as wicked problems do not have an end state (Rauscher 1999). Complex problems hence require approaches that combine quantitative and qualitative tools in an integrated framework (Rauscher 1999, Lal et al. 2001). Decision support systems (DSS) constitute an integrated approach to address complex management planning problems.  The Integrated Dual Filter approach presented in this study entails the generic elements or subsystems of a Decision Support System as listed by Rauscher (1999): (i) People, (ii) spatial- /non-spatial data management, and (iii) knowledge-management. People build the most important part in the IDF, as in any DSS. In this study, the stakeholders were represented directly in two ways. The first was a working group consisting of Yukon forest practitioners and experts that were able to voice their opinions and rank forest values in the Social Filter portion of this research. However, the majority of the CATT stakeholders‘ opinions were represented indirectly in this research. These are condensed in the Strategic Forest Management Plan (SFMP) document (SFMP 2004). The SFMP is a compendium of forest values and voices from the region which was developed in a community-engaging process taking almost a decade. Real life cannot be described by one single discipline, or be reduced to one single strategy (Born and Sonzogni 1995) as presented in the SFMP (SFMP 2004). Hence, this research took the SFMP one step 149  further and developed together with the Yukon working group the alternative management strategies stemming from the SFMP document‘s goals/objectives/indicators (see Chapter 2, or as a digest in Figure 5.1).  The data utilized in this research consisted mainly of empirical ecological data, feeding into the Environmental Filter, and social data informing the Social Filter of the IDF (e.g., Table 2.1). Due to the paucity of ecological data in the region, considerable time was used in the establishment of 90 ecological plots to build and corroborate an eco-database from which the models deployed in this study were parameterized and calibrated, and resource maps developed (see Chapter 3). For example, information on soil enabled the development of an edaphic site map; tree ring data from over 250 white spruce trees were the basis for the initial tree species cohorts to populate the study landscape; two weather stations established for this research provided two years of weather data for a valley where no data were available (the Aishihik Valley); understory/ground vegetation information and forest stand structure were combined with forest inventory data (Government of the Yukon 2004) to build the base input for the simulation modeling of this study (Chapter 3). Through modeling, the Environmental, Desired, and Management States (see Chapter 4 for description) were developed, representing planning projections to be compared in the IDF adaptive process (described in Chapter 4) to inform decision-making (e.g., see centre of Figure 4.2).  The third subsystem deals with knowledge in its many forms, and organizes the quantitative and qualitative data to support the decision-making process (see arrows in Figure 4.2). As 150  described by Rauscher (1999), this framework can consist of simulation models (e.g., LANDIS- II, Scheller et al. 2007), visualization tools (e.g., GIS), and expert and knowledge-based systems (Schmoldt and Rauscher 1996). For example, the Yukon working group consisted of members that had been engaged in the development of the SFMP (2004) document and know the forest planning and management environment of the CATT very well and could help with insights and details to inform the IDF.    Figure 5.1: Integrated Dual Filter framework, showing the two Filters (Environmental and Social) and the three States (Environmental, Management, Desired). For each filter the main findings of the thesis are listed.   151  5.2 Limitations of the IDF-study  An understanding of system vulnerabilities especially under climate change is important knowledge that must be used to manage possible risks associated with climate change (e.g., see bark beetle outbreak, Berg et al. 2006 or Garbutt et al. 2006). For forest management it is therefore advisable to consider climate change during the planning process. In the IDF, the simulations deployed were also run under high climate change projections to show possible forest successional trajectories (Chapter 3). However, care is advised when using projected data, especially for an area like the CATT. The unique topography of the study area (St. Elias mountain ranges with a big elevation range and diverse relief), and the low number of weather stations with continuous data records add to the difficulty of interpolation of climate data. For example, the historic weather data for the climate normals were based on a few weather stations all more than 100 km away from the study landscape (see Chapter 3). Also, according to Bonsal et al. (2001, 2003), the Global Circulation Models‘(see Chapter 3 for details) performance might be impacted and hence contribute to increased model uncertainty; especially the prediction of precipitation is poor beyond 60ºN — the study area is at around 61ºN.  Climate change has been linked through the TACA (Tree And Climate Assessment) model (Nitschke and Innes 2008) to feed into LANDIS-II — this constituted the first linking of the two models (Chapter 3). LANDIS-II was the main tool deployed for building the three states (E, D, M) used in this research (Chapters 3 and 4). LANDIS-II is a forest landscape dynamics model that has been designed to address questions regarding bigger spatial extents, e.g., landscapes ranging from ~ 10 4  to 10 8  ha in size, and to address temporal extents of between 50–2,000 years 152  (Scheller et al. 2010). The model has flexible resolution, the spatial ranging from 10x10 m 2  up to 500x500m 2 , and temporal resolutions of 1–40 years (Scheller et al. 2010). The model allows the simultaneous employment of an unlimited number of tree species (the challenge here is to find the respective species parameters). This allowed adding six more species currently not present in the study area (see Chapter 3). This tool was very useful in that it allowed questions related to the interaction of natural and anthropogenic disturbances and resulting patterns at the landscape scale to be addressed (Mladenoff 2004, Scheller et al. 2007). LANDIS therefore constitutes an ideal tool to engage in a planning process where simulation over a longer time period (e.g., >100 years as in this planning study) and bigger landscapes (e.g., >50,000 ha like the study landscape) is required. Assuming that LANDIS reflects biophysical realities (e.g., LANDIS has been applied and tested widely, also in the eastern boreal of the Canada (e.g., Sturtevant et al. 2009) and shown to do a sufficient job at larger scales) it ideally could guide and constrain social desires as identified in the Social Filter portion of the IDF. This, however, requires further testing. In addition, the fire module in LANDIS does not account for fire spotting (Sturtevant et al. 2009), which in an area like the CATT with strong prevailing winds (Jean Paul Pinard, pers. comment) could impact on the outcome of the fire risk analyses conducted in Chapter 4 (reducing fire risk at the landscape) by likely shifting the threshold from 20% to higher values for the landscape scale. Other tools that can incorporate this aspect of fire behaviour should be deployed to test and compare the efficacy of these fuel treatment approaches.  For the addressing of finer scale questions, translations and communication between stand and landscape planning levels, a limitation in this IDF study, the inclusion of other tools (e.g., ecosystem models) is encouraged so that decision-making can be informed by finer resolution 153  answers to planning questions regarding stand or lower spatial scales, and shorter time frames. For harvesting (simulating the anthropogenic disturbance), the LANDIS harvesting module follows management libraries based on stand attributes defined by the user, which it then selects randomly (Gustafson et al. 2000). For more precise and flexible harvesting scheduling at the landscape and stand scale, other tools would probably allow greater flexibility and precision in harvesting planning and intervention. Common spatial forest planning problems such as the shape and distribution of patches or intervention units, adjacency and green-up restrictions and constraints, or other landscape-level attributes such as connectivity and fragmentation and distribution of patch sizes and road considerations (Baskent and Keles 2005) could be better addressed, and could possibly be linked to LANDIS. For example, TEAMS (Terrestrial Ecosystem Analysis and Modeling Systems), a goal-oriented decision support system designed to achieve optimal management for the short- and long-term, and encompassing strategic and tactical level planning (Covington et al. 1988), or SELES (Spatially Explicit Landscape Event Simulator), a model that has been deployed in several land-use planning projects to support forest landscape-level decisions (e.g., Fall et al. 2004) could be used. The SELES harvest submodule captures the same management regimes and data requirements as the aspatial Forest Service Simulator FSSIM deployed by the British Columbia Ministry of Forests, Lands and Natural Resource Operations for timber supply analyses (BCMOF 2002). Generally, to overcome the lack of tactical level tools in the IDF, other tools could be considered that have this planning level as a focus, such as the Stanley model in the Woodstock/Stanley DSS), which is composed of three components: Woodstock, spatial Woodstock, and Stanley. The aspatial Woodstock (Huettmann et al. 2005) allows the formulation of a planning problem through linear programming, usually presented as a Model I or II problem (cf. Bettinger et al. 2009). It is 154  organized into 12 modules, with the landscape module representing an important module of the planning system. In comparison, the IDF uses the LANDIS-II model to address landscape questions while the stand scale is more difficult to address with the harvesting module. In the Woodstock landscape module, current and future development/allocation types are described (e.g., forest strata defined by specific landscape attributes), and where development types are directly linked with growth and yield information. The IDF in contrast uses two different States, namely the E and D States, to address current (environmental) and future (desired) conditions; in both States the same model is deployed (e.g., LANDIS-II). Forest strata are defined in species and ecosystem attribute files, growth is guided by the succession module, and yield is documented in the output modules (see Appendix E). In the Woodstock control module parameters are specified such as the number of runs for a Monte Carlo simulation. The transition module allows the analysis of the planning system responses to management actions. In the IDF the transition is represented by the M State, where management actions are deployed on the E State to bring the system parameters closer to the D State. Similar to LANDIS, the input to Woodstock can be as GIS files. Spatial Woodstock is the spatial analytical tool and data manager that is linked to GIS, and which allows the visualization of management actions (Walters and Cogswell 2002). It also addresses spatial issues as discussed in Baskent and Keles (2005). Stanley addresses the tactical forest planning level (in contrast, the LANDIS harvesting module focuses on the landscape level, although with adjustments of the Management Areas and the harvesting criteria list, the harvester can be better guided at higher resolution). Stanley takes the Woodstock output and applies it to specific sites by respecting spatial considerations as adjacency and green-up areas. Hence, this model complements the Woodstock model by trying to minimize the differences between itself and the Woodstock output (Huettmann et al. 2005). 155   The ‗Analytic Hierarchy Process‘ (Saaty 1977) was the main resource/tool for the systematic approach adopted in the Social Filter portion of this research. This multi-criteria decision support tool is today widely applied in operations research and environmental management (Schmoldt et al. 2001, Mendoza and Martins 2006). The AHP is specifically useful in the realms of complex planning problems such as forest management in the CATT as it simultaneously considers a great number of economic, ecological, environmental, and socio-cultural values. The AHP searches for a reasonable compromise between potentially competing values and conditions in multiple-use planning such as in the present study. This tool, in combination with a working group and a ratings table permitted the hierarchical structuring of the values listed in the Strategic Forest Management Plan (the values in the SFMP are nor ranked nor prioritized) (Figure 2.1 showing the AHP structure of criteria and sub-features). This allowed, in the second step, the assignment of ratings and judging of these values and the development of alternative forest management strategies (Chapter 2).  The AHP is also open to an endless number of stakeholder groups being engaged in the ranking and judgment processes (Schmoldt et al. 2001), which on the one hand increases the computing time but on the other hand enables the planning research to become more socially representative. In the present study only one group helped in the ranking and judgment of the values listed in the AHP hierarchy (Figure 2.1). The ratings table (see Table C2) developed by the local working group (i.e., structuring the values deemed to be important) helped an external expert group to rank the Saaty elements (Figure 2.1). The inclusion of an external group has the advantage that criteria and subcriteria of an AHP can be ranked without regional bias simply 156  focusing on the pair-wise importance of the respective relationships (i.e., assuming that fewer personal or political biases are linked to the elements during the rating process). Hence, I believe that the use of an external group is a logical and valid approach. Another concern was raised and addressed during the working group meetings regarding the pair-wise comparisons, which could be seen as a reductionist view and unrepresentative of First Nation views. According to a Yukon working group member, the Champagne and Aishihik FN perceives socio-economic and anthropocentric aspects concentrically embedded in the natural environment sphere. The ‗boxes‘ (e.g., Figure 2.1) in the AHP are clearly not structured as concentric ‗spheres‘. However, all the criteria and subcriteria identified for the AHP hierarchy stemmed from the SFMP document (SFMP 2004) (see Appendix C1). This document was developed over a ten-year period and involved a participatory approach that included numerous community meetings, where FN members were present and brought forward their views and concerns. Hence, the SFMP can be seen as a framework reflecting local and traditional knowledge values and concerns. The AHP is based on these values. Also, the AHP brings the economic and ecological boxes (Figure 2.1) together, e.g., synthesizing them when building the eigenvalue vectors.  Two of the most important criticisms regarding the AHP approach are related to uncertainty checks and rank reversal: (i) AHP itself does not allow for dealing with uncertainty inherent in the data; i.e., the only way to address uncertainty is through the consistency analysis of the pair- wise matrices (Alho et al. 1996); (ii) rank reversal can occur when introducing a new alternative, such as an additional alternative strategy as presented in Figure 2.1. This means that the inclusion of a new alternative may change the existing rank order (Kangas et al. 2008): A > B > C. With a new D, the situation may change so that B > A. Different propositions exist to alleviate 157  this problem. For example, regression techniques can be used instead of eigenvalue techniques to reduce the number of comparisons and estimate preferences (Alho et al. 1996). In this study I did not encounter this problem as the example (Figure 2.1) was a static hierarchy, with a fixed number of alternatives (e.g., FMS1–5). However, if I had had the opportunity to discuss the results with the participants of the Yukon working group, new alternatives could have replaced the existing ones, which might then have led to the problem described. Hence, a more flexible methodology than AHP could be envisioned in this context. ANP, the Analytic Network Process (Saaty and Vargas 2006) is an extension of the AHP, which enables the inclusion of feedback and dependences between the criteria. According to Saaty (2001) the ANP is able to grasp reality better than AHP given its emphasis on interdependencies of the criteria. Another option would be to use the AHP basic model in combination with a SWOT (Strengths, Weaknesses, Opportunities, and Threats) analysis (e.g., Hill and Westbrook 1997). This would be less elaborative than an ANP, and would enable the evaluation of decision alternatives in respect to each SWOT factor using the AHP. SWOT would basically provide the decision framework, with AHP guiding the analysis to prioritize the alternative strategies (Kangas et al. 2008).  The lack of feedback from and with the working group (due to circumstances outside my control) represents a major limitation of this study; during the IDF process I was not able to present the modeling output (TACA, LANDIS). It would be crucial to a planning process to have continuous feedback from the planning team involved (e.g., the Yukon working group) in order to modify/refine the Social Filter, and to develop thresholds for the D and the M States. The main tool of the IDF, LANDIS simulation model, is tightly linked with GIS. Using GIS maps to illustrate visually how the landscape changes, and how the different management approaches 158  translate through time and space could be incorporated into the IDF into a more explicit spatial DSS that allows for more planning interactivity like Tang and Bishop proposed in their 2002 paper. Sheppard and Meitner (2005) for example integrated in their multi-criteria analysis modelling-based expert evaluations of an array of alternative forest strategies with a 3D landscape visualization tool (CALP, Cavens 2002) to inform and evoke feedback from stakeholders engaged in a planning and decision making process; the same tool is also deployed in Seeley et al.‘s DSS for British Columbia (Seeley et al. 2004).  This study aimed to illustrate and discuss the new concepts of Filters and States by deploying a number of already established tools simultaneously in combination; this is the IDF. However, I did not explicitly explore risk and uncertainty neither in the Social or Environmental Filter. LANDIS-II for example is based on scenario approaches which according to Lindgren and Bandhold (2005) are used as an effective strategic planning tool for medium and long-term timeframes under uncertain conditions. Accordingly, Gadow (2001: 2-3) states: ―Information on the uncertainty involved is essential in environmental decision making. One refers to decision making under risk, if the decision alternatives with their possible consequences, and the probabilities of these consequences occurring, are known. If the probabilities or even the consequences of the alternative actions are not known, the decision is made under uncertainty.‖ In Section 5.4, I discuss some suggestions on how to increase verification and validation of the IDF. By doing so,  the dual problem of risk and uncertainty could be better addressed.   5.3 Key conclusions for the IDF 159   The core concepts of the IDF, the two filters and the three states, stem from other fields and have been combined here into a new decision support and planning framework. For example, the two filters are based on the conservation and management of biodiversity concepts of coarse and fine filters that refer to scale issues in natural resources management (e.g., Hunter 1990, 1991, 1999, Kremsater et al. 2003). The states are based on the hierarchical organization of ecosystems (e.g., Kay 1991) and on the current planning and decision making terminology such as the term ―desired future condition‖ (e.g., Lessard 1998). The IDF process consists of an iterative ‗adaptive cycle‘ that allows the comparison of E and D States, and re-adjustment of the M State to allow for a continuous new comparison of the E and D States. The adaptive approach is not new in decision making and decision support systems. However, the conceptual simplicity of the IDF represents a forest management planning process follows a heuristic approach (cf. Bettinger et al. 2009) that can be used to explore system opportunities and challenges. The inherent scenario-based approach in particular enables complexity and uncertainty to be addressed (Lindgren and Bandhold 2003). The most prominent characteristics of the IDF are described below.  The Social Filter of the IDF highlights regionally important socio-economic and environmental values. The combination of tools (AHP, ratings table, working group, expert group) assisted in the development of five alternative forest management strategies (stemming from a publicly available planning document as the SFMP). The ‗holistic‘ strategy scored the highest in the Analytic Hierarchy Process (Table 2.4), and is therefore assumed to best reflect the SFMP in its entirety. The Social Filter also emphasizes regional desires and concerns, which are 160  reflected in the strategic directions (the alternative forest management strategies), with their objectives and goals (Chapter 2). Ecosystem sustainability for Bormann et al. (1994, in Gadow et al. 2000) represents ―the degree of overlap between what people collectively want – reflecting social values and economic concerns …‖ (Gadow et al. 2000:14). In the IDF case this has been identified in the Social Filter, and transposed to the landscape (the Desired State). ―…and what is ecologically possible in the long term‖ (Gadow et al. 2000:14), has been identified in the Environmental Filter, and simulated using the Environmental State.  The Environmental Filter of the IDF identifies the biophysical characteristics of the planning system. This filter entails the environmental data acquisition, as well as simulation modeling (with the E State). The E State (representing the no-management landscape) allows the simulation of a landscape with the presence of forest succession and other ecological processes such as fire or insects, in the absence of any anthropogenic processes. This allows the question ‗what does the system do‘ to be addressed (Rittel and Webber 1973), which in turn gives the planning a ‗natural‘ baseline against which to compare the impacts of management actions on a landscape (M State). This is especially helpful when projecting a landscape into the future under different scenarios (e.g., historic versus climate change). For example, fire, although less frequent than elsewhere in the Yukon (McCoy and Burns 2005) is an important natural disturbance in the CATT that encourages forest renewal and keeps the landscape in a heterogeneous state. Climate change is affecting species at the site level, but not directly at the landscape level (Chapter 3).  161  The Desired State is based on the stakeholder inputs, and represents a landscape which should achieve socially desired outcomes. The managers engineer a landscape and test whether the implemented structure meets the stated goals (i.e., reduce amount of area burned). The focus of this state is on the ‗what to achieve‘, or ‗what should the system do‘ (Rittel and Webber 1973). Therefore, the D State allows the testing of how to transpose a strategic direction (which entails the social desire, e.g., manage forests for fire risk reduction) onto a landscape.  The Management State sets the planning focus on the testing of management tactics to reach similar outcomes as for the D State: how and where on the landscape to act? The ‗how‘ is informed by the Social Filter indicating possible management tactics for a specific forest management strategy. The ‗where‘ is guided by the D State. M and D States are compared in an iterative way (adaptive process), where M is been modified until the D State is reached. The M State gives answers to the feasibility of what has been engineered and designed in the D State.  Extended planning time frames allow for a more informed management. To gain understanding of the planning system (e.g., the CATT forest landscape), the study of successional trajectories is advised. The extension of timeframes is fundamental, especially when dealing with some ‗system components‘ that have longer lifespans (e.g., trees). For example, white spruce still dominated the landscape after 200 years, and pine started to spread only after 80 years (Chapter 3). When considering system disturbances that constitute an integral part of an ecosystem (Rykiel 1985), such as beetle or fire, longer time horizons are necessary to understand the possible response patterns and trajectories of the landscape (Turner 2005). The current beetle epidemic in the CATT lasted for some 15 years (Yukon FHR 2009). In order to enable 162  management not only to react to an outbreak (i.e., salvage logging of beetle induced mortality in white spruce, SFMP (2004)) but to take proactive measures to possibly reduce future beetle risk, it is crucial to study beetle outbreak cycles. The risk assessment (Chapter 4) revealed that within 80 years of simulation there might again be enough suitable white spruce on the landscape for a spruce bark beetle outbreak.  The Strategic Forest Management Plan (SFMP 2004), as shown in this study, is not an end point for forest management planning, but rather should be seen as a starting point to engage in exploratory forest management planning through techniques such as the IDF proposed here.  The conceptual simplicity of the IDF makes it a valuable decision support framework to identify system properties, constraints and concerns (e.g., the Filters) in order to simulate and project the planning landscape for the long-run (>100 years): the E State to allow for a no- management baseline; the D State to allow for the assessment (landscape designing) of desired forest strategies; the M State to allow for the exploration of management tactics. The IDF could act as a vessel for decision making tools such as MCDAs, MCDSs, MODSs, and MODAs (Ananda and Herath 2000), in that current tools (e.g., AHP, harvesting module LANDIS) are been replaced by other tools that have already been verified and validated elsewhere. Most decision aid techniques and tools address certain characteristics of a decision problem but do not alleviate all bottlenecks (Kangas and Kangas 2005). DSSs do not compensate for but rather complement each other (Kangas et al. 2008); so does the IDF.   163  5.4 Future opportunities  5.4.1 From further verification This research is wide in breadth, touching on many forestry topics and is based on an extensive sampling of empirical data to parameterize and calibrate the tools deployed in the IDF. This research also constitutes the first round of a planning process (cf. Mintzberg et al. 1976) in which alternative management strategies are developed through a Social Filter (Figure 5.1) and possible tactics addressed (see Table C2). The Environmental Filter provided a first impression of system trajectories under current and projected climate conditions (Figure 5.1). The IDF adaptive process, with the E, M and D States, informed potential decision-makers about an array of planning scenarios, illustrated in this study with the example of fire risk reduction (e.g., cumulative area burned). However, a real world planning environment is more complex and will require further study. As Craig (1996) has stated, sustainable forest management is only possible if there are sustainable social and economic costs involved. In a next further step these should be addressed. This could be envisioned by consulting the stakeholders involved in this research, and by presenting the outcome of this study. This would allow the Yukon working group (the stakeholders) to define thresholds for the different tactics and strategies, something that was not possible during the working group meetings as they occurred at the outset of this study. Also, finer resolution goals and tactics could be developed in order to re-define/adjust the Desired State and then modify the M State for the ‗adaptive cycling‘. As the D and M comparison represents a continuous stepwise adjustment, the exchange with the stakeholders should be iterative and continuous so that they are exposed to the modeling and planning output. This would enable re-adjustments and modifications based on social input. In order to increase social 164  acceptance of the planning process and outcomes further, stakeholder groups (i.e., the ones involved in the SFMP community process) should be shown the outcomes of the IDF and their feedback sought. This would not only enable re-adjustments, but would also enable a social license to be acquired (i.e., social sustainability) at the regional level (Pukkala 2002, Bunnell and Boyland 2003). To further verify and test the IDF, all the alternative forest management strategies (Figure 5.1) could be tested in the IDF for sustainability to allow a cross comparison between each scenario and also to increase the information for the decision maker to potentially allow for better informed decisions. This could stimulate an interesting opportunity for the forest governance stewards of the CATT (i.e., the Champagne and Aishihik First Nations and the Yukon Territorial governments) to help reassess their SFMP after the first 20 years of implementation (SFMP 2004). This would also constitute a further verification of the IDF in a real world environment (i.e., outside the research realms of this study).  5.4.2 … to validation A decision support system‘s performance or decision-making support is only as good as its subsystems and components: the more reliable model processes are (e.g., the less logic errors they contain) the more reliable their output is. With the same logic, a DSS such as the IDF is only as good as its input data: for example, the more accurate (finer-scaled) data and knowledge the two Filters can extract from the management planning and decision-making environment (Figure 4.2), the more reliable the model results will be (Bunnell and Boyle 2003). To use Sterman‘s words (2002:501) ―Systems thinking requires understanding that all models are wrong and humility about the limitations of our knowledge‖. Uncertainty in planning can thus be addressed by increasing knowledge, which can be done by collecting more empirical data (i.e., 165  improving geographic representation, Longley et al. 2005), or by assessing a higher number of forecasting scenarios (Schwartz 1988, Lindgren and Bandhold 2003), or, less common though, by backcasting (i.e., using a starting point in the future and trying to reach the current conditions; Dreborg 1996). Also, the deployment of other models is encouraged to verify the outcome of for example the results generated by LANDIS.  Further data layers and ecological processes should be investigated and tested in an area like the CATT where permafrost conditions are likely to change under future climate, possibly impacting soil temperature, water cycling, nutrient cycling, and decomposition processes. This would allow for a more comprehensive and realistic testing of the CATT situation. It would, however, require the addition of other models with different resolution in scope than the ones used in this study. Basically, all parts of the components presented in Figure 4.2, such as the libraries and toolboxes, can be exchanged with other tools and filled with other types of data and knowledge. Depending on the system/region under scrutiny this means that the IDF framework could be exported to other contexts beyond the boreal ecosystem of the southwest Yukon. For example, instead of using the Yukon working group for the expert and practitioner input, or instead of using the Strategic Forest Management Plan and its associated documents (SFMP 2004), other existing and locally tailored and corroborated databases could be deployed and incorporated to validate the IDF.  166  5.4.3 …and accepting complexity For forest management planning, engaging in the IDF framework involves making abstractions from complex and complicated realities. Accepting that forest resource management planning is a wicked problem, it can be tamed by deploying and integrating a combination of tools that have already been used by people dealing with wicked problems; these tools are also known as knowledge, organization, judicious simplification, and inspired leadership (Rauscher 1999).   5.5 Contributions to management and forest management planning In this final section of the thesis, I present what I believe to be the main contributions and innovations to forest management and planning:  For this study, 90 ecological monitoring plots were established to inform modelling parameterization and calibration. Combined with the establishment of two weather stations in the Aishihik Valley these could provide data and contribute to long-term studies of climate change, forest disturbances and system responses in the region. This information (e.g., knowledge, understanding) is crucial for forest management in a cold- dry boreal forest area such as the CATT.  Forest disturbances are of great concern in the region. The current spruce bark beetle outbreak, coupled with the increased amount of dead white spruce fuel has increased awareness of forest values and change in the region. The Strategic Forest Management Plan (and follow-up plans) represent a first attempt towards an adaptive management 167  approach for sustainably managing the forests of the future. This study is the first to use the SFMP‘s Criteria and Indicators to derive a set of alternative forest management strategies (building the Social Filter of the IDF), and to inform scenario-based planning and simulation approaches in the CATT.  It is the first time where TACA and LANDIS-II have been combined into a single modelling approach. The mechanistic ecophysiology model TACA was used to translate climate change data into tree establishment probabilities and used these as input for the landscape ecological model LANDIS-II. Within the IDF framework managers can assess species responses under climate change at the site level. The LANDIS-II model represents the core tool for the Environmental Filter of the IDF decision support system; it helps forest managers to consider simultaneously forest disturbances (e.g., fire) and climate change, thereby enabling the assessment of the consequences on forest succession at a landscape scale.  The IDF allows forest management planners to identify and assess tangible and intangible forest values by using the combination of a widely applied decision support tool (AHP) with a simulation tool (LANDIS). This facilitates the formulation of values, issues and concerns to inform a Social Library (the Social Filter, see Figure 4.2) and to develop a set of alternative strategies for further evaluation. This will inform the Desired State of the IDF which then can be compared in the further planning process with the Environmental State (informed by the Environmental Library, Figure 4.2). A third state, the Management State transitions the E State towards a defined D State. During this process insights regarding tactical thresholds and adjustments can be gained. 168   The Environmental State allows developing a baseline natural landscape for forest management planning to assess current and possible future conditions of the natural system. The Desired State allows the graphical visualization of a future landscape. The Management State allows assessing the management implications onto a natural landscape and its analyses can give insights on whether a desired landscape can be achieved (reality check) through management, and also allows to assess possible consequences of the planned management implications on an array of locally important indicators.  This multi-modelling approach (or decision support system) allows the modelling of climate change and potential forest responses to be used as a dialogue tool to inform future forest management planning in the CATT area and offers support for adaptive management.  169  Bibliography  ACIA (Arctic Climate Impact Assessment). 2004. Impacts of a Warming Arctic: Arctic Climate Impact Assessment. Cambridge University Press, Cambridge. Agee, J.K. 2003. Historical range of variability in eastern Cascades forests, Washington, USA. Landscape Ecology 18: 725–740. Ahlén, I. 1975. Winter habitats of moose and deer in relation to land use in Scandinavia. Swedish Wildlife 9: 45-192. Aitken, S.N., Yeaman, S., Holliday, J.A., Wang, T., and Curtis-McLane, S. 2008. Adaptation, migration or extirpation: climate change outcomes for tree populations. Evolutionary Applications 1: 95–111. Alho, J., Kangas, J., and Kolehmainen, O. 1996. Uncertainty in the expert predictions of the ecological consequences of forest plans. Applied Statistics 45: 1–14. Allen, J.L., Wesser, S., Markon, C.J., and Winterberger, K.C. 2006. Stand and landscape level effects of a major outbreak of spruce beetles on forest vegetation in the Copper River Basin, Alaska. Forest Ecology and Management 227: 257–266. Ananda, J., and Herath, G. 2003. The use of Analytic Hierachy Process to incorporate stakeholder preferences into regional forest planning. Forest Policy and Economics 5: 13– 26. Ananda, J., and Herath, G. 2009. A critical review of multi-criteria decision making methods with special reference to forest management and planning. Ecological Economics 68: 2535– 2548. Andison, D.W., and Marshall, P.L. 1999. Simulating the impact of landscape level biodiversity guidelines: a case study. The Forestry Chronicle 75: 655–665. 170  Angelstam, P. 1998. Maintaining and restoring biodiversity in European boreal forests by developing natural disturbance regimes. Journal of Vegetation Science 9: 593–602. Astrup, R., Coates, K.D., and Hall, E. 2008. Recruitment limitation in forests: Lessons from an unprecedented mountain pine beetle epidemic. Forest Ecology and Management. 256: 1743–1750. Bailey, J.D., and Harrington, C.A. 2006. Temperature regulation of bud-burst phenology within and among years in a young Douglas-fir (Pseudotsuga menziesii) plantation in western Washington, USA. Tree Physiology 26: 421–430. Barber K., Butler, R., Caird, D., and Kirby, M. 1996. Hierarchical approach for national forest planning and implementation. In: Proceedings: Hierarchical approaches to forest management in public and private organizations. 25–29 May 1995, Toronto, Ontario. Canadian Forest Service, Petawawa National Forestry Institute, Petawawa, Ontario. Information Report PI-X-124. Pp 36–44. Barber, V.A., Juday, G.P., and Finney, B.P. 2000. Reduced growth of Alaskan white spruce in the twentieth century from temperature-induced drought stress. Nature 405: 668–673. Bartlein, P.J., Whitlock, C., and Shafer, S.L. 1997. Future climate in the Yellowstone National Park region and its potential impact on vegetation. Conservation Biology 11: 782–792. Baskent, E.Z. 2001. Combinatiorial optimization in firest ecosystem management modelling. Turkish Journal of Acgriculture and Forestry 25: 187–194. Baskent, E.Z., and Keles, S. 2005. Spatial forest planning: A review. Ecological Modelling 188: 145–173. Baskerville, G. 1985. Adaptive management, wood availability, and habitat availability. The Forestry Chronicle 61: 171–175. 171  BCMOF. 2002. Morice timber supply area analysis report. British Columbia Ministry of Forests, Victoria, BC. Belton, V., and Stewart, T.J. 2002. Multiple Criteria Decision Analysis: An Integrated Approach. Kluwer, Dordrecht. Bengston, D.N. 1994. Changing forest values and ecosystem management. Society & Natural Resources 7: 515–533. Berg, E.E., and Henry, J.D. 2003. The History of Spruce Bark Beetle Outbreak in the Kluane Region as Determined from the Dendrochronology of Selected Forest Stands. Parks Canada Report. Berg, E.E., Henry, J.D., Fastie, C.L., De Volder, A.D., and Matsuoka, S.M. 2006. Spruce beetle outbreaks on the Kenai Peninsula, Alaska, and Kluane National Park and Reserve, Yukon Territory: Relationship to summer temperatures and regional differences in disturbance regimes. Forest Ecology and Management 227: 219–232. Berry, P., Rounsevell, M., Harrison, P., and Audsley, E. 2006. Assessing the vulnerability of agricultural land use and species to climate change and the role of policy in facilitating adaptation. Environmental Science & Policy 9: 189–204. Bettinger, P., Boston, K., Siry, J.P., and Grebner, D.L. 2009. Forest Management and Planning. Academic Press, Burlington, MA, USA. Bhatti, J.S., Apps, M.J., and Jiang, H. 2002. Influence of nutrients, disturbances and site conditions on carbon stocks along a boreal forest transect in central Canada. Plant and Soil 242: 1–14. Bigler, C., Gavin, D.G., Gunning, C., and Veblen, T.T. 2007. Drought induces lagged tree mortality in a subalpine forest in the Rocky Mountains. Oikos 116: 1983–1994. 172  Bonsal, B., Zhang, X., Vincent, L., and Hogg, W. 2001. Characteristics of daily and extreme temperatures over Canada. Journal of Climate 14: 1959–1976. Bonsal, B.R., Prowse, T.D., and Pietroniro, A. 2003. An assessment of global climate model‐simulated climate for the western cordillera of Canada (1961–90). Hydrological Process. 17: 3703–3716. Bormann, B.T., Brooks, M.H., Ford, E.D., Kiester, A.R., Oliver, C.D., and Weigand, J.F. 1994. Volume V: A Framework for Sustainable Ecosystem Management. Gen. Tech. Rep. PNW- 331. U.S. Department of Agriculture, Forest Service, Pacific Northwest Research Station. Portland, Oregon. Born, S.M., and Sonzogni, W.C. 1995. Integrated environmental management: strengthening the conceptualization. Environmental Management 19: 167–181. Bradshaw, J.M., and Boose, J.H. 1990. Decision analysis techniques for knowledge acquisition: combining information and preferences using Aquinas and Axotl. International Journal of Man-Machine Studies 32: 121–186. Bradshaw, H., Holmqvist, B.H., Cowling, S.A., and Sykes, M.T. 2000. The effect of climate change on the distribution and management of Picea abies in southern Scandinavia. Canadian Journal of Forest Research 30: 1992–1998. Bunge, M. 1959. Causality: The Place of the Causal Principle in Modern Science. Harvard University Press, Cambridge, Massachusetts. Bunnell, F.L., and Boyland, M. 2003. Decision-support systems: it‘s the question not the model. Journal for Nature Conservation 10: 269–279. Buongiorno, J., and Gilles, J.K. 2003. Decision Methods for Forest Resource Management. Academic Press, Burlington, Massachusetts. 173  Burgess, M.M., Judge, A.S., and Taylor, A.E. 1982. Yukon ground temperature data collection - 1966 to August 1981; Earth Physics Branch Open File 82-1; Energy, Mines and Resources Canada, Ottawa, Ontario. Burns, R.M., and Honkala, B.H. 1990. Silvics of North America. USDA Forest Service Agriculture Handbook 654, Washington, D.C., USA. Burton, P., and Cumming, S. 1995. Potential effects of climatic change on some western Canadian forests, based on phenological enhancements to a patch model of forest succession. Water, Air, & Soil Pollution 82: 401–414. CAFN Fuel Treatment Project 2005. CAFN Strategic Forest Management Plan Implementation. Fire Risk Abatement Technical Working Group, (FRATWG) Members, January 31, 2005. Calef, M.P., David McGuire, A., Epstein, H.E., Scott Rupp, T., and Shugart, H.H. 2005. Analysis of vegetation distribution in Interior Alaska and sensitivity to climate change using a logistic regression approach. Journal of Biogeography 32: 863–878. Carpenter, S.R., and Gunderson, L.H. 2001. Coping with collapse: ecological and social dynamics in ecosystem management. Bioscience 51: 451–457. Carroll, C., Paquet, P.C., and Noss, R.F. 1999. Modeling carnivore habitat in the Rocky Mountain region: A literature review and suggested strategy. World Wildlife Fund-Canada, Toronto. Cater, T.C., and Chapin III, F.S. 2000. Differential effects of competition or microenvironment on boreal tree seedling establishment after fire. Ecology 81: 1086–1099. Cavens, D. 2002. A Semi-Immersive Visualisation System for Model-Based Participatory Forest Design and Decision Support. M.Sc. Thesis, University of British Columbia, Vancouver. 174  CCFM (Canadian Council of Forest Ministers). 2003. Defining Sustainable Forest Management in Canada: Criteria and indicators 2003. Natural Resources Canada, Canadian Forest Service, Ottawa. Chapin III, F.S., Shaver, G.R., Giblin, A.E., Nadelhoffer, K.J., and Laundre, J.A. 1995. Responses of arctic tundra to experimental and observed changes in climate. Ecology 76: 694–711. Chapin III, F., Peterson, G., Berkes, F., Callaghan, T., Angelstam, P., Apps, M., Beier, C., Bergeron, Y., Crépin, A.S., and Danell, K. 2004. Resilience and vulnerability of northern regions to social and environmental change. Ambio 33: 344–349. Chhin, S., and Wang, G.G. 2008. Climatic response of Picea glauca seedlings in a forest-prairie ecotone of western Canada. Annals of Forest Science 65: 207. Chhin, S., Hogg, E., Lieffers, V.J., and Huang, S. 2008. Potential effects of climate change on the growth of lodgepole pine across diameter size classes and ecological regions. Forest Ecology and Management 256: 1692–1703. Chuine, I., and Beaubien, E.G. 2001. Phenology is a major determinant of tree species range. Ecology Letters 4: 500–510. Church, R.L., Murray, A.T., Figueroa, M.A., and Barber, K.H. 2000. Support system development for forest ecosystem management. European Journal of Operational Research 121: 247–258. CIFOR (Center for International Forestry Research). 1999. The CIFOR Criteria and Indicators Generic Template. CIFOR, Jakarta. Clague, J.J., and Rampton, V. 1982. Neoglacial Lake Alsek. Canadian Journal of Earth Sciences 19: 94–117. 175  Connelly, W. 1996. A definition for hierarchical analysis for forest planning. In: Proceedings of a workshop on hierarchical approaches to forest management in public and private organizations. 1996. Martell, D.L.; Davis, L.S.; Weintraub, A., May 25-29, 1992, Toronto, Canada. Natural Resources Canada, Canadian Forest Service, Petawawa National Forestry Institute, Chalk River, ON. Information Report PI-X-124. Courtois, R., Dussault, C., and Potvin, F. 2002. Habitat selection by moose (Alces alces) in clear- cut landscapes. Alces 38: 177–192. Covington, W.W., Wood, D.B., Young, D.L., Dykstra, D.P., and Garrett, L.D. 1988. TEAMS: A decision support system for multiresource management. Journal of Forestry 86: 25–33. Craig, L.E. 1996. Letter from the United States Senate, Committee on Energy and Natural Resources, Washington, DC to The Honorable Dan Glickman, Secretary of Agriculture. 20 June 1996, 5 pp. Cumming, S., Burton, P., and Klinkenberg, B. 1996. Boreal mixedwood forests may have no "representative" areas: some implications for reserve design. Ecography 19: 162–180. Danby, R.K., and Hik, D.S. 2007. Variability, contingency and rapid change in recent subarctic alpine tree line dynamics. Journal of Ecology 95: 352–363. Daust, D.K., and Sutherland, G.D. 1997. SIMFOR: software for simulating forest management and assessing effects on biodiversity. In: Thompson, I.D. (e.d.), The Status of Forestry/Wildlife Decision Support Systems in Canada: Proceedings of a Symposium, Toronto, Canada, 1994. Natural Resources Canada, Canadian Forestry Service, Great Lakes Forestry Centre, Sault St. Marie, Ontario, pp. 15–29. 176  Davis, R.G., and Martell, D.L. 1993. A decision support system that links short-term silvicultural operating plans with long-term forest-level strategic plans. Canadian Journal of Forest Research 23: 1078–1078. delMoral, R., Titus, J.H., and Cook, A.M. 1995. Early primary succession on Mount St. Helens, Washington, USA. Journal of Vegetation Science 6: 107–120. DIAND (Department of Indian Affairs and Northern Development). 1998. Timber Harvest Planning and Operating Ground Rules. DIAND, Whitehorse, Yukon. Diaz-Balteiro, L., and Romero, C. 2008. Making forestry decisions with multiple criteria: A review and an assessment. Forest Ecology and Management 255: 3222–3241. Dietz, T., Fitzgerald, A., and Shwom, R. 2005. Environmental values. Annual Review of Environment and Resources 30: 335–372. Doak, P. 2004. The impact of tree and stand characteristics on spruce beetle (Coleoptera: Scolytidae) induced mortality of white spruce in the Copper River Basin, Alaska. Canadian Journal of Forest Research 34: 810–816. Douglas, G.W. 1974. Montane zone vegetation of the Alsek River region, southwestern Yukon. Canadian Journal of Botany 52: 2505–2532. Douthwaite, B., de Haan, N.C., Manyong, V., and Keatinge, D. 2001. Blending ―hard‖ and ―soft‖ science: the ―follow-the-technology‖ approach to catalyzing and evaluating technology change. Conservation Ecology 5: [online] <http://www.consecol.org/vol5/iss2/art13> Dreborg, K.H. 1996. Essence of backcasting. Futures 28: 813–828. 177  Driscoll, W., Wiles, G., D‘Arrigo, R., and Wilmking, M. 2005. Divergent tree growth response to recent climatic warming, Lake Clark National Park and Preserve, Alaska. Geophysical Research Letters 32: L20703. Dussault, C., Courtois, R., and Ouellet, J.-P. 2006. A habitat suitability index model to assess moose habitat selection at multiple spatial scales. Canadian Journal of Forest Research 36: 1097–1107. Eid, T. and Hobbelstad, K. 2000. AVVIRK-2000 – a large-scale forestry scenario model for long-term investment, income and harvest analyses. Scandinavian Journal of Forest Research 15: 472–482. Fall, A., Morgan, D. and Edie, A. 2004. Morice landscape model. Report to the Morice Land and Resource Management Planning Process, British Columbia Ministry of Forests, Victoria, BC. Fettig, C.J., Klepzig, K.D., Billings, R.F., Munson, A.S., Nebeker, T.E., Negrón, J.F., and Nowak, J.T. 2007. The effectiveness of vegetation management practices for prevention and control of bark beetle infestations in coniferous forests of the western and southern United States. Forest Ecology and Management 238: 24–53. Feuer, L.S. 1957. The principle of simplicity. Philosophy of Science 24: 109–122. Field, C., and Mortsch, L. 2007. North America. In Climate Change 2007: Impacts, Adaptation and Vulnerability. Contribution of Working Group II to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change. Flannigan, M., and Van Wagner, C. 1991. Climate change and wildfire in Canada. Canadian Journal of Forest Research 21: 66–72. 178  Flannigan, M.D., Bergeron, Y., and Wotton, B.M. 1998. Future wildfire in circumboreal forests in relation to global warming. Journal of Vegetation Science 9: 469–476. Flannigan, M., Logan, K., Amiro, B., Skinner, W., and Stocks, B. 2005. Future area burned in Canada. Climatic Change 72: 1–16. Fleming, R.A., and Candau, J.-N. 1998. Influences of climatic change on some ecological processes of an insect outbreak system in Canada's boreal forests and the implications for biodiversity. Environmental Monitoring and Assessment 49: 235–249. Foote, M.J. 1983. Classification, description, and dynamics of plant communities after fire in the taiga of interior Alaska. Research Paper PNW-307. Portland, OR: U.S. Department of Agriculture. Forest Service, Pacific Northwest Forest and Range Experiment Station. Forestry Canada Fire Danger Group. 1992. Development and structure of the Canadian Forest Fire Behavior Prediction System. Forestry Canada, Science and Sustainable Development Directorate, Information Report ST-X-3. Foster, D.R., Knight, D.H., and Franklin, J.F. 1998. Landscape patterns and legacies resulting from large, infrequent forest disturbances. Ecosystems 1: 497–510. Francis, S., 1996. Fire history of the Shakwak Trench. M.Sc. Thesis, University of British Columbia, Vancouver. Franklin, A.B., Noon, B.R., and George, T.L. 2002. What is habitat fragmentation? Studies in Avian Biology 25: 20–29. Fries, C., Carlsson, M., Dahlin, B., Lämås, T. and Sallnäs, O. 1998. A review of conceptual landscape planning models for multiobjective forestry in Sweden. Canadian Journal of Forest Research 28: 159–167. Garbutt, R. 2004. Forest Health Assessment Yukon, 2000–2004 Establishment Report. 179  Garbutt, R.W., Hawkes, B.C., and Allen, E.A. 2006. Spruce beetle and the forests of the southwest Yukon. Natural Resources Canada. Canadian Forest Service, Pacific Forestry Centre, Rep. Information Report BC-X-406. Grimble, R., and Wellard, K. 1997. Stakeholder methodologies in natural resource management: a review of principles, contexts, experiences and opportunities. Agricultural Systems 55: 173–193. Gitay, H., Suarez, A., and Wilson, R. 2002. Climate Change and Biodiversity. Intergovernmental Panel on Climate Change, Rep. Technical Paper V. Goldammer, J.G., and Price, C. 1998. Potential impacts of climatic change on fire regimes in the tropics based on MAGICC and a GISS GCM-derived lightening model. Climatic Change 39: 273–296. Government of the Yukon. 2004. Government of the Yukon‘s Forest Inventory Database. Greene, D., Zasada, J., Sirois, L., Kneeshaw, D., Morin, H., Charron, I., and Simard, M. 1999. A review of the regeneration dynamics of North American boreal forest tree species. Canadian Journal of Forest Research 29: 824–839. Greene, D.F., and Johnson, E.A. 2000. Tree recruitment from burn edges. Canadian Journal of Forest Research 30: 1264–1274. Gregory, R.S. 2002. Incorporating value trade-offs into community-based environmental risk decisions. Environmental Values 11: 461–488. Greig, M., and Bull, G. 2009. Carbon management in British Columbia‘s forests: opportunities and challenges. FORREX Series. Guba, E.G. 1990. The alternative paradigm dialog. In The Paradigm Dialog. Sage, London. pp 17–30. 180  Gunderson, L.H., Holling, C.S., Pritchard L., Jr., and Peterson, G.D. 2002. Resilience of large- scale resource systems. In SCOPE 60: Resilience and Behaviour of Large-Scale Systems. Island Press, Washington, U.S.A. pp. 3–20. Hajjar, R., Gough, A., Mathey, A.H., Nitschke, C., Paudel, S.K., Skrivanos, P., Waeber, P.O., and Innes, J. 2009. Criteria and indicators for sustainable forest management in the face of decentralization: are they still relevant in their current form? Proceedings of the World Forestry Congress, Argentina. Hall, F.G., Knapp, D.E., and Huemmrich, K.F. 1997. Physically based classification and satellite mapping of biophysical characteristics in the southern boreal forest. Journal of Geophysical Research 102: 29567–29580. Hamann, A., and Wang, T. 2006. Potential effects of climate change on ecosystem and tree species distribution in British Columbia. Ecology 87: 2773–2786. Harshaw, H. 2010. Public participation in British Columbia forest management. The Forestry Chronicle 86: 697–708. Hawkes, B.C., 1983. Fire history and management study of Kluane National Park. Canadian Forest Service, Pacific Forest Research Centre, Victoria, British Columbia. He, H.S., Mladenoff, D.J., and Boeder, J. 1999. An object-oriented forest landscape model and its representation of tree species. Ecological Modelling 119: 1–19. Heginbottom, J.A., Dubreuil, M.-A., and Harker, P.A. 1995. Canada Permafrost, Plate 2.1, (MCR 4177). National Atlas of Canada (5th edition), Canada Permafrost, Plate 2.1, (MCR 4177), 1:7,500,000 scale. Hill, T., and Westbrook, R. 1997. SWOT analysis: It‘s time for a product recall. Long Range Planning 30: 46–52. 181  Hirsch, K., Kafka, V., Tymstra, C., McAlpine, R., Hawkes, B., Stagehuis, H., Quintilio, S., Gauthier, S., and Peck, K. 2001. Fire-smart forest management: a pragmatic approach to sustainable forest management in fire-dominated ecosystems. The Forestry Chronicle 77: 357–363. Hirsch, K.G., Podur, J.J., Janser, R.F., McAlpine, R.S., and Martell, D.L. 2004. Productivity of Ontario initial-attack fire crews: results of an expert-judgement elicitation study. Canadian Journal of Forest Research 34: 705–715. Hogg, E.H., and Wein, R.W. 2005. Impacts of drought on forest growth and regeneration following fire in southwestern Yukon, Canada. Canadian Journal of Forest Research 35: 2141–2150. Holling, C.S. 1978. Adaptive Environmental Assessment and Adaptive Management. John Wiley and Sons, Oxford, U.K. Holling, C.S., Berkes, F., and Folke, C. 1998. Science, sustainability and resource management. In Linking Social and Ecological Systems: Management Practices and Social Mechanisms for Building Resilience. Edited by F. Berkes and C. Folke. Cambridge University Press, Cambridge. pp 342–362. Holling, C.S. 2001. Understanding the complexity of economic, ecological, and social systems. Ecosystems 4: 390–405. Howe, E., and Baker, W.L. 2003. Landscape heterogeneity and disturbance interactions in a subalpine watershed in Northern Colorado, USA. Annals of the Association of American Geographers 93: 797–813. Huettmann, F., Franklin, S.E., and Stenhouse, G.B. 2005. Predictive spatial modelling of landscape change in the Foothills Model Forest. The Forestry Chronicle 81: 525–537. 182  Hunter, M.L. 1999. Maintaining Biodiversity in Forest Ecosystems. Cambridge University Press, Cambridge. Hunter, M.L., Jr., Jacobson, G.L. and Webb, T. 1988. Paleoecology and coarse filter approach in maintaining biological diversity. Conservation Biology 2: 375–385. Hunter, M.L., Jr. 1990. Wildlife, Forests, and Forestry: Principles of Managing Forests for Biological Diversity. Prentice Hall, New Jersey. Hunter, M.L., Jr. 1991. Coping with ignorance: the coarse-filter strategy for maintaining biological diversity. In Balancing on the Brink of Extinction. Edited by K.A. Kohm. Island Press, Washington DC. pp. 266–281. Ib  ez, I., Clark, J.S., LaDeau, S., and Lambers, J.H.R. 2007. Exploiting temporal variability to understand tree recruitment response to climate change. Ecological Monographs 77: 163– 177. ILP (Integrated Landscape Plan). 2006. Technical Assessment of Resouces, Management Priorities, and Guidelines for Timber Harvest Project Planning – Integrated Landscape Plan for the Champagne and Aishihik Traditional Territory. Yukon Government, Ministry of Energy, Mining and Resources. Jeakins, P., Sheppard, S.R.J., Bunnell, F.L., and Wells, R. 2006. A framework for sustainable forest management. Arrow Innovative Forest Practices Agreement (IFPA) Series: Extension Note 1: 37–49. Johnson, E.A. 1992. Fire and Vegetation Dynamics: Studies from the North American Boreal Forest. Cambridge University Press, Cambridge. Johnson, K.N. 1986. FORPLAN Version 1: an overview. USDA Forest Service, Land Management Planning Section, Washington, DC. 183  Johnstone, J., and Chapin, F.S. 2003. Non-equilibrium succession dynamics indicate continued northern migration of lodgepole pine. Global Change Biology 9: 1401–1409. Johnstone, J.F., Chapin, I., Foote, J., Kemmett, S., Price, K., and Viereck, L. 2004. Decadal observations of tree regeneration following fire in boreal forests. Canadian Journal of Forest Research 34: 267–273. Johnstone, J.F., and Chapin, F.S. 2006. Effects of soil burn severity on post-fire tree recruitment in boreal forest. Ecosystems 9: 14–31. Johnstone, J., and Chapin, F. 2006. Fire interval effects on successional trajectory in boreal forests of northwest Canada. Ecosystems 9: 268–277. Kangas, A., Kangas, J., and Kurttila, M. 2008. Decision Support for Forest Management. Springer. Kangas, J. 1994. An approach to public participation in strategic forest management planning. Forest Ecology and Management 70: 75–88. Kangas, J., Store, R., Leskinen, P., and Mehtätalo, L. 2000. Improving the quality of landscape ecological forest planning by utilising advanced decision-support tools. Forest Ecology and Management 132: 157–171. Kangas, J., and Kangas, A. 2005. Multiple criteria decision support in forest management – Fundamentals of the approach, methods applied, and experiences gained. Forest Ecology and Management 207: 133–143. Kant, S., and Lee, S. 2004. A social choice approach to sustainable forest management: an analysis of multiple forest values in Northwestern Ontario. Forest Policy and Economics 6: 215–227. 184  Karjala, M.K., and Dewhurst, S.M. 2003. Including aboriginal issues in forest planning: a case study in central interior British Columbia, Canada. Landscape and Urban Planning 64: 1– 17. Karjala, M.K., Sherry, E.E., and Dewhurst, S.M. 2004. Criteria and Indicators for sustainable forest planning: a framework for recoding Aboriginal resource and social values. Forest Policy and Economics 6: 95–110. Kaufmann. M. R., Graham, R. T., Boyce, D. A., Moirm W. H., Perry, L., Reynolds, R. T., Bassett. R. L., Mehlhop, l., Edminster, C. B., Block, W. M., and Corn, P. S. 1994. An ecological basis for ecosystem management. U.S. Forest Service, Rocky Mountain Research Station, Rep. General Technical Report RM-GTR-246. Kay, J.J. 1991. A nonequilibrium thermodynamic framework for discussing ecosystem integrity. Environmental Management 15: 483–495. Kimmins, J. 1999. Biodiversity, Beauty and the ―Beast‖: Are beautiful forests sustainable, are sustainable forests beautiful, and is ―small‖ always ecologically desirable? The Forestry Chronicle 75: 955–960. Kimmins, J.P., Mailly, D., and Seely, B. 1999. Modelling forest ecosystem net primary production: the hybrid simulation approach used in FORECAST. Ecological Modelling 122: 195–224. Kimmins, J. 2002. Future shock in forestry. The Forestry Chronicle 78: 263–271. Kimmins, J.P. 2004. Forest Ecology: A Foundation for Sustainable Forest Management and Environmental Ethics in Forestry. Prentice Hall, New Jersey. 185  Klenk, N.L., and Hickey, G.M. 2011. A virtual and anonymous, deliberative and analytic participation process for planning and evaluation: The Concept Mapping Policy Delphi. International Journal of Forecasting 27: 152–165. Klinka, K., Worral, J., Skoda, L., Varga, P., and Krajina, V.J. 1998. The distribution and synopsis of ecological and silvical characteristics of tree species of British Columbia‘s forests. Scientia Silvica Extension Series Number 10. University of British Columbia. Klinka, K., Worrall, J., Skoda, L., Varga, P, and Krajina, V.J. 2000. The Distribution and Synopsis of Ecological and Silvical Characteristics of Tree Species of British Columbia‘s Forests. Canadian Cartographics Ltd., Coquitlam, B.C. Krasny, M.E., Vogt, K.A., and Zasada, J.C. 1988. Establishment of four Salicaceae species on river bars in interior Alaska. Ecography 11: 210–219. Krebs, C.J., Boutin, S. and Boonstra, R. (eds.). 2001. Ecosystem Dynamics of the Boreal Forest: The Kluane Project. Oxford University Press, New York. Kremsater L., Bunnell, F., Huggard, D., Dunsworth, G. 2003. Indicators to assess biological diversity: Weyerhaeuser‘s coastal British Columbia forest project. The Forestry Chronicle 79: 590–601. Kuuluvainen, T., 2002. Disturbance dynamics in boreal forests: defining the ecological basis of restoration and management of biodiversity. Silva Fennica 36: 5–12. Kuusipalo, J., and Kangas, J. 1994. Managing biodiversity in a forestry environment. Conservation Biology 8: 450–460. Lal, P., Lim-Applegate, H., and Scoccimarro, M. 2001. The adaptive decision-making process as a tool for integrated natural resource management: focus, attitudes, and approach. Conservation Ecology 5: 11 [online] <http://www.consecol.org/vol5/iss2/art11/>. 186  Landres, P.B., Morgan, P. and Swanson, F.J. 1999. Overview of the use of natural variability concepts in managing ecological systems. Ecological Applications 9: 1179–1188. Lawson, B.D., and Armitage, O. 2008. Weather guide for the Canadian Forest Fire Danger Rating System. Canadian Forest Service, Northern Forestry Centre, Canada. Lertzman, K., Spies, T. and Swanson, F. 1997. From ecosystem dynamics to ecosystem management. In The Rain Forests of Home: Profile of a North American Bioregion. Edited by P.K. Schoonmaker, B. Von Hagen, and E.C. Wolf. Island Press, Washington, D.C., USA. pp. 261–382. Lessard, G. 1998. An adaptive approach to planning and decision-making. Landscape and Urban Planning 40: 81–87. Lindenmayer, D., and Franklin, J.F. 2002. Conserving Forest Biodiversity: A Comprehensive Multiscaled Approach. Island Press, Washington. Lindenmayer, D., Hobbs, R.J., Montague-Drake, R., Alexandra, J., Bennett, A., Burgman, M., Cale, P., Calhoun, A., Cramer, V., and Cullen, P. 2008. A checklist for ecological management of landscapes for conservation. Ecology Letters 11: 78–91. Lindgren, M., and Bandhold, H. 2003. Scenario Planning: The Link between Future and Strategy. Curran Publishing Services, Norwich. Liu, Z., He, H.S., Chang, Y., and Hu, Y. 2010. Analyzing the effectiveness of alternative fuel reductions of a forested landscape in Northeastern China. Forest Ecology and Management 259: 1255–1261. Lockwood, C., and Moore, T. 1993. Harvest scheduling with spatial constraints: a simulated annealing approach. Canadian Journal of Forest Research 23: 468–478. 187  Loehle, C. 2004. Applying landscape principles to fire hazard reduction. Forest Ecology and Management 198: 261–267. Longley, P.A., Goodchild, M.F., Maguire, D.J., and Rhind, D.W. 2005. Geographic Information Systems and Science. John Wiley & Sons Ltd., Chichester, U.K. Lotan, J., Brown, J. and Neuenschwander, L. 1985. Role of fire in lodgepole pine forests. In Lodgepole Pine the Species and its Management Symposium Proceedings. Edited by D. Baumgartner et al.Washington State University, Pullman. pp. 133-152 Lovell, C., Mandondo, A., and Moriarty, P. 2002. The question of scale in integrated natural resource management. Conservation Ecology 5: [online] <http://www.consecol.org/vol/iss2/art25> Manly, B.F.J. 1986. Multivariate Statistical Methods: A Primer. Chapman & Hall, London. Martell, D.L., Gunn, E.A., and Weintraub, A. 1998. Forest management challenges for operational researchers. European Journal of Operational Research 104: 1–17. Martin, W.E., Wise Bender, H., and Shields, D.J. 2000. Stakeholder objectives for public lands: Rankings of forest management alternatives. Journal of Environmental Management 58: 21–32. Matsuoka, S.M., Handel, C.M., and Ruthrauff, D.R. 2001. Densities of breeding birds and changes in vegetation in an Alaskan boreal forest following a massive disturbance by spruce beetles. Canadian Journal of Zoology 79: 1678–1690. McCormick, J.E. 1999. A food-based habitat-selection model for grizzly bears in Kluane National Park, Yukon. M.Sc. Thesis, University of British Columbia, Vancouver. McCoy, V., and Burn, C. 2005. Potential alteration by climate change of the forest-fire regime in the boreal forest of central Yukon Territory. Arctic 58: 276–285. 188  McGregor Model Forest Association. 2001. ECHO Planning System Overview Manual http://:www.mcgregor.bc.ca accessed 10 October 2011. McKenzie, D., Peterson, D.W., and Peterson, D.L. 2003. Modelling conifer species distributions in mountain forests of Washington State, USA. The Forestry Chronicle 79: 253–258. McNicol, J.G., and Baker, J.A. 2004. Emulating natural forest disturbances: From policy to practical guidance in Ontario. In Emulating Natural Forest Landscape Disturbances: Concepts and Applications. Edited by A.H. Perera, L.J. Buse, and M.G. Weber. Columbia University Press, New York. pp. 251–262. Mendoza, G., and Martins, H. 2006. Multi-criteria decision analysis in natural resource management: A critical review of methods and new modelling paradigms. Forest Ecology and Management 230: 1–22. Mendoza, G., and Vanclay, J. 2008. Trends in forestry modelling. Agriculture, Veterinary Scinece, Nutrition and Natural Resources 3: 010. Mermoz, M., Kitzberger, T., and Veblen, T.T. 2005. Landscape influences on occurrence and spread of wildfires in Patagonian forests and shrublands. Ecology 86: 2705–2715. Messier, C., and Kneeshaw, D. 1999. Thinking and acting differently for sustainable management of the boreal forest. The Forestry Chronicle 75: 929–938. Mingers, J., and Brocklesby, J. 1997. Multimethodology: Towards a framework for mixing methodologies. Omega 25: 489–509. Mintzberg, H., Raisinghani, D., and Theoret, A. 1976. The structure of ―unstructured‖ decision processes. Administrative Science Quarterly 21: 246–275. Mitchell, S., and Beese, W. 2002. The retention system: reconciling variable retention with the principles of silvicultural systems. The Forestry Chronicle 78: 397–403. 189  Mladenoff, D.J. 2004. LANDIS and forest landscape models. Ecological Modelling 180: 7–19. Mladenoff, D.J., Host, G.E., Boeder, J., Crow, T.R., 1996. LANDIS: a spatial model of forest landscape disturbance, succession, and management. In GIS and Environmental Modeling: Progress and Research Issues. Edited by M.F. Goodchild, L.T. Steyaert, and B.O.Parks. GIS World Books, Fort Collins, CO, USA. pp. 75–180. Mladenoff, D.J., and He, H.S. 1999. Design, behavior and application of LANDIS, an object- oriented model of forest landscape disturbance and succession. In Spatial Modeling of Forest Landscape Change: Approaches and Applications. Edited by D.J. Mladenoff and W.L. Baker. Cambridge University Press, Cambridge. pp 125–162. Montreal Process. 1995. Criteria and Indicators for the Conservation and Sustainable Management of Natural Forests. Natural Resources Canada, Ottawa. Mott, D.G. 1963. The analysis of the survival of small larvae in the unsprayed area. Memoires of the Entomological Society Canada 95: 42–52. Murray, A.T., and Church, R.L. 1995. Heuristic solution approaches to operational forest planning problems. OR Spektrum 17: 193–203. Murray, C., and Marmorek, D. 2003. Adaptive management: A science-based approach to managing ecosystems in the face of uncertainty. Prepared for presentation at the Fifth International Conference on Science and Management of Protected Areas: Making Ecosystem Based Management Work, Victoria, British Columbia, 11–16 May 2003. Naesset, E. 1997. Geographical information systems in long-term forest management and planning with special reference to preservation of biological diversity: a review. Forest Ecology and Management 93: 121–136. 190  Natural Resources Canada. 2011. Forest ecosystems of Canada. Measuring sustainability: Criteria and indicators. <http://ecosys.cfl.scf.rncan.gc.ca/enjeux-issues/mesurabilite- sustainability-eng.asp> accessed 10 May 2011. Nelson, J., and Brodie, J.D. 1990. Comparison of a random search algorithm and mixed integer programming for solving area-based forest plans. Canadian Journal of Forest Research 20: 934–942. Nelson, J., Brodie, J.D., and Sessions, J. 1991. Integrating short-term, area-based logging plans with long-term harvest schedules. Forest Science 37: 101–122. Nelson, J. 2003. Forest Planning Studio (FPS)-ATLAS Program: Reference Manual Version 6. Faculty of Forestry, University of British Columbia, Vancouver. Available at <http://www.forestry.ubc.ca/atlas-simfor/extension/docs.html#FPS_2003> Nitschke, C.R., and Innes, J.L. 2006. Interactions between fire, climate change and forest biodiversity. Perspectives in Agriculture, Veterinary Science, Nutrition and Natural Resources 1: 1–9. Nitschke, C.R., and Innes, J.L. 2008. A tree and climate assessment tool for modelling ecosystem response to climate change. Ecological Modelling 210: 263–277. Nitschke, C.R. 2009. Vulnerability and Adaptive Capacity of Yukon Tree Species to Climate Change. Technical Report RC08-1698. Northern Climate ExChange. 2006. Forest Management in a changing climate: Building the environmental information base for the southwest Yukon. Backgrounder: Climate Change and Major Forest Disturbance in the southwest Yukon. Available at <http://yukon.taiga.net/swyukon/extranet/disturbance_backgrounder2.pdf> 191  Noss, R.F. 1987. From plant communities to landscape in conservation inventories: a look at the Nature Conservancy (USA). Conservation Biology 41: 11–37. Noss, R.F. 2001. Beyond Kyoto: forest management in a time of rapid climate change. Conservation Biology 15: 578–590. Noss, R.F., Franklin, J.F., Baker, W.L., Schoennagel, T., and Moyle, P.B. 2006. Managing fire- prone forests in the western United States. Frontiers in Ecology and the Environment 4: 481–487. Ogden, A.E., and Innes, J.L. 2007a. Incorporating climate change adaptation considerations into forest management planning in the boreal forest. International Forestry Review 9: 713–733. Ogden, A.E., and Innes, J.L. 2007b. Perspectives of forest practitioners on climate change adaptation in the Yukon and Northwest Territories of Canada. The Forestry Chronicle 83: 557–569. Ogden, A.E., and Innes, J.L. 2008. Climate change adaptation and regional forest planning in southern Yukon, Canada. Mitigation and Adaptation Strategies for Global Change 13: 833– 861. Ogden, A.E., and Innes, J.L. 2009. Application of structured decision making to an assessment of climate change vulnerabilities and adaptation options for sustainable forest management. Ecology and Society 14: 11 [online] <http://www.ecologyandsociety.org/vol14/iss1/art11/> Ohlson, D.W., McKinnon, G.A., and Hirsch, K.G. 2005. A structured decision-making approach to climate change adaptation in the forest sector. The Forestry Chronicle 81: 97–103. O‘Neill, R.V., deAngelis, D.L., Waide, J.B., and Allen, T.F.H. 1986. A Hierarchical Concept of Ecosystems. Princeton University Press, Princeton, New Jersey. 192  Oreskes, N., Shrader-Frechette, K., and Belitz, K.. 1994. Verification, validation, and confirmation of numerical models in earth sciences. Science 263: 641–646. Parisien, M.A., Junor, D.R., and Kafka, V.G. 2006. Using landscape-based decision rules to prioritize locations of fuel treatments in the Boreal mixedwood of western Canada. In: Fuels Management–How to Measure Success: Conference Proceedings. 28–30 March 2006; Portland, Oregon. Proceedings RMRS-P-41. Fort Collins, Colorado: U.S. Department of Agriculture, Forest Service, Rocky Mountain Research Station. Perera, A.H., and Buse, L.J. 2004. Emulating natural disturbances in forest management: An overview. In Emulating Natural Forest Landscape Disturbances: Concepts and Applications. Edited by A.H. Perera, L.J. Buse and M. Weber. Columbia University Press, New York. pp. 3–7. Peters, V.S., MacDonald, S.E., and Dale, M.R.T. 2005. The Interaction between masting and fire is key to white spruce regeneration. Ecology 86: 1744–1750. Peterson, D.W., Peterson, D.L., and Ettl, G.J. 2002. Growth responses of subalpine fir to climatic variability in the Pacific Northwest. Canadian Journal of Forest Research 32: 1503–1517. Pickett, S.T.A., and White, P.S. 1985. The Ecology of Natural Disturbance and Patch Dynamics. Academic Press, San Diego. Pickett, S.T.A., Ostfeld, R.S., Shachak, M. and Likens, G.E. (eds.). 1997. The Ecological Basis of Conservation: Heterogeneity, Ecosystems, and Biodiversity. Chapman & Hall, New York. Pietronero, L. 2008. Complexity ideas from condensed matter and statistical physics. Europhysics News 39: 26–29. 193  Pommerening, A., and Murphy, S. 2004. A review of the history, definitions and methods of continuous cover forestry with special attention to afforestation and restocking. Forestry 77: 27–44. Pukkala, T. (ed.). 2002. Multi-objective forest planning. Managing forest ecosystems. Kluwer, Dordrecht. Rauscher, H.M. 1999. Ecosystem management decision support for federal forests in the United States: A review. Forest Ecology and Management 114: 173–197. Rauscher, H.M., Lloyd, T., Loftis, D.L., and Twery, M.J. 2000. A practical decision-analysis process for forest ecosystem management. Computers and Electronics in Agriculture 27: 195–226. Réale, D., McAdam, A.G., Boutin, S., and Berteaux, D. 2003. Genetic and plastic responses of a northern mammal to climate change. Proceedings of the Royal Society of London. Series B: Biological Sciences 270: 591–596. Reed, W.J. 2006. A note on fire frequency concepts and definitions. Canadian Journal of Forest Research 36: 1884–1888. Rittel, H.W.J., and Webber, M.M. 1973. Dilemmas in a general theory of planning. Policy Sciences 4: 155–169. Rykiel, E. 1985. Towards a definition of ecological disturbance. Australian Journal of Ecology 10: 361–365. Saaty, T.L. 1977. A scaling method for priorities in hierarchical structures. Journal of Mathematical Psychology 15: 234–281. Saaty, T.L. 2001. Decision Making for Leaders, the AHP for Decisions in a Complex World. RWS Publications, Pittsburgh. 194  Saaty, T.L., and Vargaas, L.G. 2006. Decision Making with the Analytic Network Process: Economic, Political, Social and Technological Applications with Benefits, Opportunities, Costs and Risks. Springer, New York. Saaty, T.L. 2008. Decision making with the analytic hierarchy process. International Journal of Services Sciences 1: 83–98. Safranyik, L., Simmons, C., and Barclay, H.J. 1990. A conceptual model of spruce beetle population dynamics. Forestry Canada, Pacific Forestry Centre, Victoria, BC. Information Report BC-X-316. Scheller, R.M., and Mladenoff, D.J. 2004. A forest growth and biomass module for a landscape simulation model, LANDIS: design, validation, and application. Ecological Modelling 180: 211–229. Scheller, R.M., Domingo, J.B., Sturtevant, B.R., Williams, J.S., Rudy, A., Gustafson, E.J., and Mladenoff, D.J. 2007. Design, development, and application of LANDIS-II, a spatial landscape simulation model with flexible temporal and spatial resolution. Ecological Modelling 201: 409–419. Scheller, R.M., E.J. Gustafson, B.R. Sturtevant, B.C. Ward, and D.J. Mladenoff. 2010. Increasing the research and management value of ecological models using modern software engineering techniques. Frontiers in Ecology and the Environment: 8: 253–260. Schmoldt, D.L., and Rauscher, H.M. 1996. Building Knowledge-Based Systems for Natural Resource Management. Chapman & Hall, New York. Schmoldt, D.L., Kangas, J., and Mendoza, G. 2001. Basic principles of decision making in natural resources and the environment. In The Analytic Hierarchy Process in Natural 195  Resources and Environmental Decision Making. Edited by D.L. Schmoldt, J. Kangas, G.A. Mendoza and M. Pesonen. Kluwer Academic Publishers. The Netherlands. pp 1–13. Schwartz, B. 1988. Forecasting and scenarios. In Handbook of Systems Analysis – Craft Issues and Procedural Choices. Edited by J.E. Bingham and G.W.P. Davies. John Wiley and Sons, Chichester. pp. 327–367. Seely, B., Nelson, J., Wells, R., Peter, B., Meitner, M., Anderson, A., Harshaw, H., Sheppard, S., Bunnell, F., and Kimmins, H. 2004. The application of a hierarchical, decision-support system to evaluate multi-objective forest management strategies: a case study in northeastern British Columbia, Canada. Forest Ecology and Management 199: 283–305. Sessions, J., and Bettinger, P. 2001. Hierarchical planning: pathway to the future? In Proceedings of the First International Precision Forestry Symposium. Edited by Briggs, D. University of Washington Institue for Forest Resources, Seattle, Washington, USA. pp 185–190. SFMP (Strategic Forest Management Plan). 2004. Strategic Forest Management Plan for the Champagne and Aishihik Traditional Territory. Champagne and Aishihik First Nations government and Government of the Yukon. Available at <http://www.emr.gov.yk.ca/forestry/fmp_champagne_aishihik_traditional_territory.html> Sheppard, R.J., and Meitner, M. 2005. Using multi-criteria analysis and visualisation for sustainable forest management planning with stakeholder groups. Forest Ecology and Management 207: 171–187. Shifley, S.R., Thompson, F.R., Larsen, D.R., and Dijak, W.D. 2000. Modeling forest landscape change in the Missouri Ozarks under alternative management practices. Computers and Electronics in Agriculture 27: 7–24. 196  Silsbee, D., and Peterson, D.L. 1993. Planning for implementation of long-term resource monitoring programs. Environmental Monitoring and Assessment 26: 177–185. Smith, C.A.S., Meikle, J.C. and Roots, C.F. (eds.). 2004. Ecoregions of the Yukon Territory. Biophysical properties of Yukon landscapes. Agriculture and Agri-Food Canada, PARC Technical Bulleting No. 04-01, Summerland, British Columbia. Smith, S.L., and Burgess, M.M. 2002. A digital database of permafrost thickness in Canada. Geological Survey of Canada, Open File 4173. 1 CD-ROM. Solomon, S., Qin, D., Manning, M., Alley, R.B., Berntsen, T., Bindoff, N.L., Chen, Z., Chidthaisong, A., Gregory, J.M., Hegerl, G.C., Heimann, M., Hewitson, B., Hoskins, B.J., Joos, F. et al. 2007. Technical Summary. In Climate change 2007: The Physical Science Basis. Cambridge University Press, Cambridge. Spies, T.A. and Turner, M.G. 1999. Dynamic forest mosaics. In Maintaining Biodiversity in Forest Ecosystems. Edited by M.L. Hunter. Cambridge University Press, Cambridge. pp. 95–160. Spittlehouse, D.L., and Childs, S.W. 1990. Evaluating the seedling moisture environment after site preparation. In Sustained Productivity of Forest Soils. Proceedings of the 7th North American Forest Soils Conference ed. University of British Columbia, Vancouver, Canada pp. 80–94. Spittlehouse, D.L., and Stewart, R.B. 2003. Adaptation to climate change in forest management. BC Journal of Ecosystems and Management 4: 1–11. Sterman, J.D. 2002. All models are wrong: reflections on becoming a systems scientist. System Dynamics Review 18: 501–531. 197  Stocks, B., Lawson, B., Alexander, M., Van Wagner, C., McAlpine, R., Lynham, T., and Dube, D. 1989. Canadian forest fire danger rating system: an overview. The Forestry Chronicle 65: 450–457. Sturtevant, B.R., Scheller, R.M., Miranda, B.R., Shinneman, D., and Syphard, A. 2009. Simulating dynamic and mixed-severity fire regimes: A process-based fire extension for LANDIS-II. Ecological Modelling. 220: 3380–3393. Sykes, M.T., and Prentice, I.C. 1996. Climate change, tree species distributions and forest dynamics: A case study in the mixed conifer/northern hardwoods zone of northern Europe. Climatic Change 34: 161–177. Tang, H., and Bishop, I.D. 2002. Integration methodologies for interactive forest modelling and visualization systems. The Cartographic Journal 39: 27–35. TRRC (Teslin Renewable Resource Council). 2007. Strategic forest management plan for the Teslin Tlingit traditional territory: strategic direction for sustainable forest resource development. Government of Yukon, Whitehorse. Turner, M.G. 1989. Landscape Ecology: The effect of pattern on processes. Annual Review of Ecology and Systematics 20: 171–197. Turner, M.G. 2005. Landscape ecology in North America: past, present, and future. Ecology 86: 1967–1974. Turner, M.G., and Dale, V.H. 1998. Comparing large, infrequent disturbances: What have we learned? Ecosystems 1: 493–496. Turner, M.G., Romme, W.H., and Tinker, D.B. 2003. Surprises and lessons from the 1988 Yellowstone fires. Frontiers in Ecology and the Environment 1: 351-358. 198  Urban, D.L., O'Neill, R.V., and Shugart, H.H., Jr. 1987. Landscape ecology. BioScience 37: 119–127. Van Wagner, C. 1987. Development and structure of the Canadian forest fire weather index system. Technical Report 35, Canadian Forest Service, Ottawa, Ontario. Varma, V.K., Ferguson, I., and Wild, I. 2000. Decision support system for the sustainable forest management. Forest Ecology and Management 128: 49–55. Von Gadow, K., Pukkala, T., and Tomé, M. 2000. Sustainable Forest Management. Springer, The Netherlands. von Winterfeldt, D., and Edwards, W. 1986. Decision Analysis and Behavioral Research. Cambridge University Press, Cambridge. Wagner, D., DeFoliart, L., Doak, P., and Schneiderheinze, J. 2008. Impact of epidermal leaf mining by the aspen leaf miner (Phyllocnistis populiella) on the growth, physiology, and leaf longevity of quaking aspen. Oecologia 157: 259–267. Walck, J.L., Hidayati, S.N., Dixon, K.W., Thompson, K., Poschlod, P. In press. Climate change and plant regeneration from seed. Accepted article to Global Change Biology. Walker, B., Carpenter, S., Anderies, J., Abel, N., Cumming, G., Janssen, M., Lebel, L., Norberg, J., Peterson, G.D., and Pritchard, R. 2002. Resilience management in social-ecological systems: a working hypothesis for a participatory approach. Conservation Ecology 6: 14 [online] <http://www.consecol.org/vol6/iss1/art14> Walters, C.J., and Holling, C.S. 1990. Large-scale management experiments and learning by doing. Ecology 71: 2060–2068. 199  Walters, K.R., and Cogswell, A. 2002. Spatial forest planning: Where did all the wood go? Forest Technology Group. Available at <http://www.remsoft.com/docs/library/spatial.forest.planning.pdf > Walters, K.R., Feunekes, H., Cogswell, A., and Cox, E. 1999. A forest planning system for solving spatial harvest scheduling problems. Available at <http://www.remsoft.com>. Weber, M. 1997. Aspekte zur Extrapolation von Tagesmittelwerten von Temperatur und taeglichen Niderschlagssummen an hochgelegenen Gebirgsstationen aus Klimadaten des oertlichen Klimamessnetzes in Nepal. Project Report 12/1996, Establishment of a Measuring Service for Snow and Glacier Hydrology. Kommission fuer Glaziologie der Bayerischen Akademie der Wissenschaften, Muenchen. Weber, M., and Flannigan, M. 1997. Canadian boreal forest ecosystem structure and function in a changing climate: impact on fire regimes. Environmental Reviews 5: 145–166. Weintraub, A., and Bare, B.B. 1996. New issues in forest land management from an operations research perspective. Interfaces 26: 9–25. Wirth, C., Lichstein, J., Dushoff, J., Chen, A., and Chapin, F. 2008. White spruce meets black spruce: dispersal, post-fire establishment, and growth in a warming climate. Ecological Monographs 78: 489–505. Wondolleck, J.M., and Yaffee, S.L. 2000. Making Collaboration Work: Lessons from Innovation in Natural Resource Management. Island Press, Washington DC. Wu, J. and Loucks, O.L. 1995. From balance of nature to hierarchical patch dynamics: A paradigm shift in ecology. The Quarterly Review of Biology 70: 439–466. Xiao, J., and Zhuang, Q. 2007. Drought effects on large fire activity in Canadian and Alaskan forests. Environmental Research Letters 2: 044003. 200  Xu, C.G., He, H.S., Hu, Y.M., Chang, Y., Li, X.Z., and Bu, R.C. 2005. Latin hypercube sampling and geostatistical modeling of spatial uncertainty in a spatially explicit forest landscape model simulation. Ecological Modelling 185: 255–269. Yarie, J. 2008. Effects of moisture limitation on tree growth in upland and floodplain forest ecosystems in interior Alaska. Forest Ecology and Management 256: 1055–1063. Yarie, J., and Van Cleve, K. 1983. Biomass and productivity of white spruce stands in interior Alaska. Canadian Journal of Forest Research 13: 767–772. Yukon FHR. 2009. Forest Health Report 2009. Government of Yukon Energy, Mines and Resources, Forest Management Branch. Available at <http://www.emr.gov.yk.ca/forestry/forest_health_reports.html> Yukon Forestry Monitoring Program. 2008. Field Manual and Monitoring Protocols. Available at: <http://www.emr.gov.yk.ca/forestry/pdf/monitoring_manual_jan2009> Zasada, J.C., and Densmore, R. 1980. Alaskan willow and balsam poplar seed viability after 3 years‘ storage. Tree Planters‘ Notes 31: 9–10. 201  Appendices Appendix A: Summary of tools and data used in the Environmental Filter study  Figure A1: Summary of tools and data used in Chapter 3 (Environmental Filter study). Rectangular boxes represent tools, rhomboids represent data input or output. Arrows indicate direction of process. Details are explained in Chapter 3. 202  Appendix B: 90 ecological research plots For the modelling input (e.g., LANDIS-II), 90 ecological plots grouped in six research blocks were established (see Table B1). Each research block represented a disturbance type, which could be either ‗undisturbed forest‘ (e.g., no records available before 1948), fire, beetle (beetled during the 1994-2007 spruce bark beetle outbreak), or anthropogenic disturbance (e.g., salvage harvesting of beetle-killed wood). Within a research block, a point of commencement was used to build a transect (on a random bearing) with plots established every 100 meters. On each (square) plot, tree density, altitude, slope, latitude, longitude, crown coverage, DBH, height and age of sample trees were collected following the field protocols of the Yukon Forestry Monitoring Program (2008) (the YFMP used round plots). A sample tree was selected from each identified stand structural class where available (dominant, co-dominant, intermediate, and suppressed) with the classes being relative to the respective forest stand; sample trees were used for tree ring analysis (e.g., counting rings to assess tree age) (see Table B2). Within each 400 m 2 - plot, three 5m x 5m nested subplots were randomly chosen to assess shrub, sapling and tree density. In the centre of each 25 m 2 -plot a 1 m 2 -subplot was established to assess ground vegetation species and abundance (see Table B3). To inform the TACA modelling, a soil pit was dug in the centre of a randomly selected 25 m 2 -plot (Table B4).  Table B1: 90 ecological plots Block 400 m2-plots Dominant trees and shrubs Elevation range [m] Major disturbances 1 13 white spruce, aspen, willow 933-967 undisturbed  forest 2 20 white spruce, willow, alder 887-1158 undisturbed forest and beetle 3 10 white spruce, aspen  754-786 undisturbed forest 4 25 white spruce (in beetled), aspen and willow (in burned patches) 651-693 fire and beetle 5 13 white spruce  778-791 harvesting 6 9 white spruce  773-781 harvesting   203  Table B2: Tree samples from the 90 400-m 2  plots. sn=Sample number (Block_Plot_SampleTree); h=height; c=dominance class assessed from height (do=dominant, co=co-dominant, in=intermediate, su=suppressed); d=diameter at breast height; cr=live crown; c.y.=core year; f.y.=first year ring; a=age (calculated from c.y.-f.y.). sn h [m] c d [cm] cr [%] c.y. f.y. a 1_01_1 11.9 do 27.1 80 2008 1917 91 1_01_2 11.8 co 16.8 90 2009 1946 63 1_01_3 7.3 in 9.1 70 2008 1943 65 1_02_1 14.3 do 25.6 95 2009 1931 78 1_02_2 11.8 co 20.4 90 2008 1939 69 1_02_3 8 in 10.2 50 2009 1939 70 1_04_1 13.1 do 27 80 2008 1938 70 1_04_2 9.1 co 15.5 90 2008 1960 48 1_04_3 5.5 in 12.3 75 2008 1943 65 1_05_1 10.8 do 23.6 90 2009 1933 76 1_05_2 8.3 co 12.2 70 2009 1934 75 1_05_3 5.5 in 9.9 70 2009 1956 53 1_07_1 4.8 in 7 95 2008 1985 23 1_08_1 12.3 do 21.1 90 2008 1929 79 1_08_2 9.7 co 16 80 2009 1915 94 1_08_3 7.3 in 11.1 90 2009 1937 72 1_09_1 13.3 do 27.3 95 2008 1950 58 1_09_2 10 co 22.3 95 2009 1941 68 1_09_3 6.5 in 8.2 90 2008 1961 47 1_10_1 10.8 do 27.6 100 2009 1961 48 1_10_2 4.8 in 7.8 95 2008 1961 47 1_11_1 11.4 do 26.7 95 2008 1947 61 1_11_2 8.1 co 12.2 95 2008 1950 58 1_12_1 12.9 do 19.9 99 2008 1958 50 1_12_3 10.3 co 19.5 80 2008 1943 65 1_12_4 9.5 co 14.5 90 2008 1946 62 1_12_5 6.5 in 10.7 70 2009 1941 68 1_13_1 14.7 co 18.2 95 2008 1966 42 1_13_2 6.3 in 10.2 95 2008 1978 30 2_01_1 20.2 do 30 0 2008 1764 244 2_01_3 9 in 10.2 40 2008 1812 196 2_02_1 9.7 co 15.9 50 2008 1897 111 2_02_2 8.4 in 12 50 2008 1903 105 2_02_3 4.8 su 7 40 2008 1916 92 2_03_1 9.3 co 12.5 60 2009 1889 120 2_03_2 7.1 in 8.9 50 2009 1893 116 2_03_3 6.4 in 8.6 50 2009 1889 120 2_04_1 10.5 do 12.3 70 2008 1888 120 204  sn h [m] c d [cm] cr [%] c.y. f.y. a 2_04_2 7.9 co 10.8 40 2008 1877 131 2_04_3 5.9 in 8.9 80 2009 1888 121 2_05_1 10.8 do 17.1 50 2008 1843 165 2_05_2 8.2 co 14.3 80 2009 1874 135 2_05_3 6.1 in 9.7 80 2009 1785 224 2_06_1 6.5 co 9.5 95 2008 1857 151 2_06_2 3.4 in 9.3 95 2008 1878 130 2_07_1 19.5 do 37.5 95 2009 1732 277 2_07_3 9.2 in 14.9 90 2008 1918 90 2_07_4 7.3 in 11.5 90 2008 1931 77 2_07_5 4.4 su 7.6 90 2008 1919 89 2_08_1 11.3 co 23.1 95 2008 1898 110 2_08_2 7.4 in 14.7 90 2008 1877 131 2_09_1 17.9 do 35.7 90 2009 1890 119 2_09_2 14.4 co 18.1 80 2008 1916 92 2_09_3 5.9 in 8.7 80 2008 1947 61 2_10_1 19.2 do 34.7 95 2008 1893 115 2_10_2 11 co 27.4 95 2009 1923 86 2_10_3 7.1 in 14.8 95 2009 1948 61 2_11_1 14.7 do 28.7 95 2008 1948 60 2_11_2 12.3 co 18.2 90 2008 1976 32 2_11_3 8.2 in 12.5 95 2008 1952 56 2_11_4 5.7 in 8.1 90 2008 1964 44 2_12_1 12.6 do 22.7 80 2009 1938 71 2_12_2 13 do 19.7 60 2008 1934 74 2_12_3 9.5 co 14 60 2008 1941 67 2_12_4 7.5 in 10 50 2008 1935 73 2_13_1 13.8 do 24.2 80 2008 1881 127 2_13_2 7.9 co 13.4 80 2008 1935 73 2_13_3 5.8 in 7.4 80 2008 1968 40 2_14_1 18.4 do 38.8 80 2008 1902 106 2_14_2 11.7 co 23.4 90 2008 1899 109 2_14_3 8.5 in 11.7 80 2008 1928 80 2_14_7 7.1 in 8.2 60 2009 1936 73 2_15_1 14.7 co 31 90 2008 1918 90 2_15_2 9 in 10.1 50 2009 1947 62 2_16_1 10.7 do 21.1 90 2009 1927 82 2_16_2 8.7 co 12.8 90 2008 1926 82 2_16_3 6.2 in 10.7 70 2008 1928 80 2_17_1 11.9 co 26.3 90 2008 1919 89 2_17_2 6.2 in 9.7 95 2008 1908 100 205  sn h [m] c d [cm] cr [%] c.y. f.y. a 2_19_1 11 do 24.6 95 2008 1917 91 2_19_2 9.9 co 19.8 80 2008 1943 65 2_19_3 8.4 co 10.6 90 2008 1943 65 2_19_4 5.9 in 8.1 80 2008 1937 71 3_01_1 12.3 do 16.3 70 2008 1956 52 3_01_2 9.2 co 12.6 95 2008 1949 59 3_01_3 8.9 co 11.8 65 2008 1752 256 3_01_4 7.4 in 7.7 50 2008 1954 54 3_02_1 19.5 do 33.3 70 2008 1972 36 3_02_6 19.1 do 26.8 70 2009 1850 159 3_02_3 13.5 co 17.9 90 2008 1849 159 3_02_2 16.9 co 24.3 90 2009 1826 183 3_02_7 13 co 19.4 65 2009 1824 185 3_02_4 6.8 in 12.3 95 2008 1857 151 3_03_1 18.3 do 22 90 2008 1815 193 3_03_6 17 do 30.2 80 2009 1812 197 3_03_2 13.8 co 17.2 80 2008 1824 184 3_03_3 8.1 in 11.1 90 2008 1839 169 3_03_4 4.1 su 8.4 30 2008 1836 172 3_04_1 17.3 do 22.2 10 2009 1792 217 3_04_7 18.5 do 37.9  2008 1826 182 3_04_2 15.5 co 17.7 80 2008 1914 94 3_04_8 13.22 co 15.3 40 2009 1823 186 3_04_3 9.2 in 9.1 60 2009 1835 174 3_05_1 11 do 18.9 90 2008 1945 63 3_05_2 9.4 co 14.9 80 2008 1943 65 3_05_3 10.2 co 13.4 80 2008 1941 67 3_05_4 6.9 in 7.4 50 2008 1949 59 3_06_1 9.8 co 20.5 99 2008 1943 65 3_06_2 9.6 co 13.4 80 2008 1948 60 3_06_3 7 in 9.5 80 2008 1956 52 3_07_1 22.3 do 30.2 95 2009 1834 175 3_07_2 14.3 co 20.5 95 2008 1839 169 3_07_4 8.3 co 12.1 70 2008 1949 59 3_07_3 12.4 in 11 70 2008 1839 169 3_07_5 6.4 in 7.3 90 2008 1948 60 3_08_5 18.7 do 27.4  2008 1825 183 3_08_2 16.6 co 19.5 80 2009 1852 157 3_08_3 10 in 9.9 80 2009 1827 182 3_09_1 9.3 co 12.2 80 2008 1943 65 3_09_2 7 in 8.3 80 2008 1944 64 206  sn h [m] c d [cm] cr [%] c.y. f.y. a 3_10_1 8.8 co 15.2 90 2008 1945 63 3_10_2 7.1 in 12.1 95 2008 1974 34 4_04_2 14.3 co 20.5 75 2008 1882 126 4_04_3 10.8 in 15.5 60 2008 1904 104 4_04_4 8.3 su 12 50 2009 1869 140 4_05_1 20.7 do 29.7 90 2009 1816 193 4_08_4 10 do 16 70 2008 1791 217 4_08_5 9.7 do 16.9 80 2008 1740 268 4_08_1 8.3 co 14.3 75 2008 1904 104 4_08_2 7.5 in 12.1 25 2008 1798 210 4_08_3 4.9 su 8 45 2008 1807 201 4_11_2 15.3 co 20.1 70 2008 1809 199 4_11_3 11.6 in 14.4 85 2009 1919 90 4_11_4 8.1 su 11 50 2008 1822 186 4_12_1 15.3 do 21.5 50 2008 1903 105 4_12_2 12.6 co 18 60 2008 1890 118 4_12_3 10 in 11.3 50 2008 1904 104 4_12_4 5.4 su 9.5 60 2008 1884 124 4_14_1 5.2 su 7.5 75 2008 1938 70 4_15_2 18.7 co 22.3 60 2008 1820 188 4_15_6 21.9 co 32.5 50 2009 1829 180 4_15_3 10.4 in 9.9 60 2008 1827 181 4_15_4 7.4 su 8.5 80 2008 1908 100 4_16_2 15.7 co 25.3 90 2008 1868 140 4_16_5 15.1 co 19.3 60 2009 1849 160 4_16_3 10.7 in 12.8 70 2009 1888 121 4_16_6 10.4 in 12.9 75 2009 1895 114 4_16_4 9 su 10 85 2009 1925 84 4_17_1 18.4 do 23 60 2008 1928 80 4_17_5 17.3 do 24.1 80 2009 1949 60 4_17_9 15.9 co 38 85 2008 1914 94 4_17_2 14.5 co 19.7 50 2008 1939 69 4_17_3 10.7 in 11.5 80 2009 1940 69 4_17_4 6.4 su 8.1 80 2008 1965 43 4_18_2 17.3 co 17.9 65 2008 1903 105 4_18_3 12.7 in 12.2 50 2009 1911 98 4_18_4 9.8 su 7.8 70 2009 1898 111 4_19_1 12.4 co 16 85 2008 1943 65 4_19_2 11.8 in 12.8 50 2008 1941 67 4_19_3 7.5 su 9.1 65 2009 1940 69 4_20_2 15 co 21.9 20 2008 1861 147 207  sn h [m] c d [cm] cr [%] c.y. f.y. a 4_20_3 12.7 in 14.4 60 2008 1905 103 4_20_4 8.9 su 10.4 45 2008 1933 75 4_21_1 22.6 do 32.7 70 2008 1909 99 4_21_6 19.5 co 19.1 60 2008 1925 83 4_21_3 13.2 in 13 75 2008 1939 69 4_21_4 12.3 su 8.2 70 2008 1934 74 4_22_1 25 do 27 75 2008 1895 113 4_22_2 15 co 15.6 50 2008 1930 78 4_22_3 13 in 11.5 60 2008 1920 88 4_22_4 9.3 su 10 60 2008 1937 71 4_23_1 22.2 do 31.4 75 2009 1821 188 4_23_6 17.9 co 20.7 90 2008 1829 179 4_23_3 10.7 in 16.5 75 2008 1876 132 4_23_4 8 su 10.4 60 2008 1828 180 4_24_5 19.2 do 20.3 60 2009 1898 111 4_24_2 18 co 26.2 50 2008 1891 117 4_24_6 15.95 co 23 70 2009 1815 194 4_24_3 14 in 17.6 70 2008 1896 112 4_24_4 7.7 su 8.6 85 2008 1923 85 4_25_1 19.8 do 25.1 60 2008 1803 205 4_25_2 17.5 co 22.1 70 2008 1826 182 4_25_3 10.5 in 11.9 65 2008 1848 160 4_25_4 6.1 su 7.5 70 2008 1851 157 5_02_1 3.5 su 9 100 2008 1970 38 5_03_6 11.2 co 13.2 45 2008 1931 77 5_03_3 9.6 in 13.6 65 2009 1941 68 5_03_5 9.2 in 12.1 70 2009 1935 74 5_03_7 8.3 in 15.3 50 2009 1928 81 5_03_4 7.5 su 12 70 2008 1944 64 5_04_1 11.9 co 42.3 95 2008 1934 74 5_04_2 10.1 in 16.4 80 2008 1958 50 5_04_3 6.1 su 11.3 75 2008 1971 37 5_05_1 18.5 do 34.1 70 2008 1869 139 5_05_2 14.4 co 20.9 60 2008 1867 141 5_05_3 11.5 in 14.5 40 2009 1878 131 5_05_4 7 su 7.2 40 2008 1883 125 5_06_1 17.1 do 22.5 60 2008 1867 141 5_06_9 19.77 do 35.7 45 2009 1877 132 5_06_2 13.6 co 16.1 90 2008 1928 80 5_06_3 9.2 in 13.8 95 2008 1882 126 5_06_4 7.8 su 8.1 85 2008 1905 103 208  sn h [m] c d [cm] cr [%] c.y. f.y. a 5_07_2 17.7 do 28.7 45 2009 1899 110 5_07_3 12.9 co 21.7 50 2008 1891 117 5_07_4 15.3 co 21.8 65 2009 1875 134 5_07_5 12.4 co 16.3 45 2009 1941 68 5_07_8 9.7 in 13 70 2008 1929 79 5_07_6 7.5 su 7.3 70 2009 1909 100 5_08_1 18.1 do 27.3 80 2009 1896 113 5_08_2 15.7 co 22.5 60 2008 1893 115 5_08_3 11.7 in 17.3 75 2008 1934 74 5_08_4 6.1 su 7.6 80 2008 1937 71 5_13_2 4.4 su 7.5 95 2008 1966 42 6_01_2 13.2 co 18 80 2008 1909 99 6_01_3 10.7 in 12.1 65 2008 1943 65 6_01_4 8.2 su 9.4 80 2008 1932 76 6_02_1 15.2 do 22.6 30 2008 1880 128 6_02_2 12.2 co 16.1 85 2008 1927 81 6_02_3 11.2 in 11.4 25 2008 1870 138 6_02_4 6.6 su 8.8 60 2008 1889 119 6_03_1 19 do 24.9 75 2008 1865 143 6_03_6 17.1 do 20.6 45 2009 1886 123 6_03_2 14.5 co 19 40 2008 1869 139 6_03_3 10.8 in 15.7 60 2008 1865 143 6_03_4 8.3 su 8.5 65 2009 1862 147 6_04_1 16.5 do 27.4 45 2008 1880 128 6_04_2 13.6 co 15.5 65 2008 1818 190 6_04_3 10.4 in 10.4 60 2008 1943 65 6_04_4 6.3 su 7 75 2008 1938 70 6_05_1 14.4 do 20.8 65 2008 1880 128 6_05_5 14 do 18.5 35 2009 1872 137 6_05_2 12.7 co 17.2 65 2008 1885 123 6_05_3 10.3 in 15.6 65 2008 1889 119 6_05_4 8.2 su 11.4 50 2009 1890 119 6_06_1 14.8 do 22.7 70 2008 1876 132 6_06_2 12 co 15.7 65 2008 1880 128 6_06_3 9.7 in 15.2 50 2008 1880 128 6_06_4 7.8 su 13 45 2008 1875 133 6_07_1 16.2 do 22.9 80 2008 1882 126 6_07_2 14.4 co 21.2 70 2008 1971 37 6_07_3 12.7 in 18.3 65 2008 1883 125 6_07_4 10.5 su 12.2 90 2008 1893 115 6_08_1 18.1 do 28.3 90 2009 1870 139 209  sn h [m] c d [cm] cr [%] c.y. f.y. a 6_08_2 15.6 co 23.8 45 2008 1894 114 6_08_3 10.5 in 12 65 2008 1887 121 6_08_4 6.5 su 10.5 55 2009 1879 130 6_09_1 15.4 do 26 85 2008 1869 139 6_09_2 14 co 20.5 80 2008 1909 99 6_09_3 10.4 in 13 70 2009 1889 120 6_09_4 6.2 su 9.5 60 2008 1891 117   Table B3: Vegetation identified at the 1-m 2  plot level (from 270 plots): Vernacular Name Scientific Name Database Code Yarrow Achillea millefolium Ach_mil Monkshood Aconitum delphiniifolium Aco_del Red Columbine Aquilega formosa Aqu_for Red Berry Arctostaphylos rubra Arc_rub Kinnikinnick (bearberry) Arctostaphylos uva-ursi Arc_uva Arnika Arnica cordifolia Arn_cor Aster Aster sp. Ast_spp Bog moss, Glow moss Aulacomnium sp. Aul_spp sedge Carex sp. Car_spp Paintbrush Castilleja sp. Cas_spp Moonshine cetraria Cetraria pinastry Cet_pin Reindeer Lichen Cladina sp. Clad_spp Pale-stalked broom moss Dicranum palidisetum Dic_pal Wet rock moss Dichodontium pellucidum Dic_pel Broom moss Dicranum sp. Dic_spp Blue Wildrye Elymus glaucus Ely_gla Fuzzy spike wild rye Elymus innovatus Ely_inn Crowberry Empetrum nigrum Emp_nig Fireweed Epilobium augustifolium Epi_aug euqisetum Equisetum sp. Equ_spp Subalpine Daisy Erigeron peregrinus Eri_per Fescue Festuca sp. Fes_spp Curled cetraria (lichen) Flavocetraria cucullata Fla_cuc Wild sweet pea Hedysarum mackenzii Hed_mac Clear moss Hookeria lucens Hoo_spp 210  Vernacular Name Scientific Name Database Code Meadow Barley Hordeum brachyantherum Hor_bra Bluebell Hyacinthoides non-scripta Hya_non Hooded tube lichen Hypogymnia physodes Hyp_phy Labrador Tea  Ledum groenlandicum Led_gro Twinflower Linnaea borealis Lin_bor Arctic Lupin Lupinus arcticus Lup_arc Monkey flower Mimulus sp. Mim_spp Field locoweed Oxytropis canpestris Oxy_can Freckled lichen (funny L) Peltigera aphthosa Pel_aph Doglichen  Peltigera canina Pel_can Sweet Coltsfoot Petasites sp. Pet_spp Red stemmed feather moss Pleurozium schreberii Ple_sch Haircap moss Polytrichum sp. Pol_spp Arctic Cinquefoil Potentilla nana Pot_nan Silverweed  Potentilla sp. Pot_spp Whiteveined wintergreen Pyrola asarifolia Pyr_asa Wintergreen  Pyrola picta Pyr_pic Wintergreen Pyrola sp Pyr_spp Pipecleaner moss Rhytidiopsis robusta Rhy_rob Prickley Rose (<15cm) Rosa acicularis Ros_aci Dwarf Nagoonberry Rubus arcticus Rub_arc Sanicule Sanicula sp. San_spp Saxifraga Saxifraga sp. Sax_spp Stonecrop Sedum sp. Sed_spp Soopolallie = shrub Shepherdia canadensis She_can Northern Goldenrod Solidago multiradiata Sol_mul Fat bog moss Sphagnum papillosum Sph_pap starwort  Stellaria sp. Ste_spp Dandelions Taraxacum officinale Tar_off Sidewalk moss Tortula ruralis Tor_rur Trillium Trillium grandiflorum Tri_gra Blood -spattered Beard Usnea sp. Usn_spp Dwarf Blueberry Vaccinium caespitosum Vac_cae Lingonberry Vaccinium vitis-idaea Vac_vit Blue violet Viola adunca Vio_adu   211  Table B4: Soil data with sn=sample number; RD=rooting depth [cm]; T1-4=texture layers (texture not shown in this table) [cm]; G=gravel [%]; C=cobble [%]; S=stones [%]; SM=soil moisture [%]; O=organic layer [cm]; elev.=elevation [m.a.s.l.]; slope [%] sn RD T1G T1C T1S T1SM T2D T2G T2S T2SM T3G T3C T3S T3SM T4G T4C T4S T4SM O elev. slope 1 37 5 0 0 9.6 5.0 5 0 9.6 5 0 0 18.4 5 0 0 18.4 3 954 10 2 45 0 0 0 8.0 5.0 5 0 8.0 10 5 0 10.4 10 5 0 10.4 0.9 947 11 3 20 0 0 0 14.7 7.7 20 0 10.5 20 20 0 10.5 20 20 0 10.5 0 962 17 4 63 0 0 0 10.2 18.0 0 0 10.2 10 40 0 10.5 10 40 0 10.5 0 967 12 5 42 4.9 0 0 13.1 29.0 50 0 6.0 50 0 0 13.4 50 0 0 13.4 1 965 21.5 6 75 0 0 0 14.8 34.0 40 30 10.7 50 0 30 18.7 50 0 0 18.7 0.9 955 14 7 90 0 0 0 27.4 6.0 0 0 28.3 40 0 0 15.2 30 0 0 15.2 0 957 26.5 8 27 0 0 0 25.5 10.0 15 0 20.0 40 0 0 9.2 40 0 0 9.2 3 952 8 9 47 0 0 0 47.7 7.0 0 0 36.2 0 0 0 27.4 0 0 0 20.8 0.5 945 3 10 64 0 0 0 39.0 15.0 0 0 37.8 10 0 0 33.7 10 0 0 33.7 0 935 5 11 47 0 0 0 32.4 6.0 0 0 43.7 0 0 0 37.4 0 0 0 37.4 2 988 6 12 52 0 0 0 16.5 4.0 0 0 23.3 0 0 0 23.3 0 0 0 27.9 0 933 11 13 83 0 0 0 13.1 4.5 0 0 17.2 0 0 0 22.2 0 0 0 24.9 0 955 24.5 14 25 0 0 0 51.0 3.3 0 0 51.0 0 0 0 51.0 0 0 0 51.0 5 887 3 15 36 0 0 0 25.5 5.0 0 0 35.7 0 0 0 37.4 0 0 0 37.4 0 899 6 18 23 0 0 0 0.0 0.0 0 0 0.0 0 0 0 0.0 0 0 0 0.0 10 920 2 19 27 0 0 0 0.0 0.0 0 0 0.0 0 0 0 0.0 0 0 0 0.0 27 931 4 20 15 0 0 0 5.3 3.7 10 0 6.0 10 0 0 6.0 10 0 0 6.0 0 943 23 21 20 0 0 0 0.0 0.0 0 0 0.0 0 0 0 0.0 0 0 0 0.0 20 943 23 22 17 0 0 0 0.0 0.0 0 0 0.0 0 0 0 0.0 0 0 0 0.0 17 966 27 23 23 0 0 0 30.1 4.0 0 0 30.1 0 0 0 30.1 0 0 0 30.1 9 1005 29 24 11 0 0 0 8.5 1.5 0 0 8.5 0 0 0 8.5 0 0 0 8.5 9 1064 71.5 25 36 0 0 0 1.5 4.0 0 0 4.3 0 0 0 17.2 0 0 0 16.0 0 906 3 26 55 0 0 0 13.9 6.5 0 0 13.9 0 5 10 10.9 10 10 30 27.8 2.5 910 6 27 32 10 0 60 51.8 5.8 10 60 51.8 10 0 60 51.8 10 0 60 51.8 7 924 6.5 28 40 0 0 0 8.4 16.0 0 0 42.8 30 10 30 18.3 30 10 30 18.3 6.5 959 23 29 34 0 0 0 0.6 8.8 5 55 17.9 5 0 55 17.9 5 0 55 17.9 4 943 29 212  sn RD T1G T1C T1S T1SM T2D T2G T2S T2SM T3G T3C T3S T3SM T4G T4C T4S T4SM O elev. slope 30 11 0 0 0 0.0 0.0 0 0 0.0 0 0 0 0.0 0 0 0 0.0 24.5 984 39.5 31 13 0 0 0 42.7 0.8 0 0 42.7 0 0 0 42.7    42.7 13 1031 44 32 42 0 0 0 18.1 5.0 0 0 43.8 5 0 0 41.5 5 0 0 41.5 3 1095 57 33 52 50 0 0 21.3 3.0 0 0 27.2 30 0 40 38.6 30 0 40 38.6 0 1158 50 34 26 0 0 0 18.1 5.0 0 0 43.8 0 0 0 41.5 0 0 0 41.5 0 1154 70.5 35 83 25 0 0 8.9 37.0 10 0 10.4 35 5 0 7.2 35 5 0 7.2 0 763 18 36 48 0 0 0 36.5 25.0 0 0 14.0 0 0 0 19.8 0 0 0 19.8 2 754 12 37 83 0 0 0 24.7 30.0 0 0 24.7 0 0 0 14.4 0 0 0 16.3 2 767 10 38 17 5 5 0 47.8 1.5 5 0 47.8 5 5 0 47.8 5 5 0 47.8 4 765 3.5 39 45 0 0 0 22.2 11.0 0 0 17.1 0 5 0 17.1 0 5 0 17.1 0 777 18 40 94 0 0 0 13.8 54.0 0 0 13.8 0 0 0 7.7 0 0 0 7.7 0 775 20 41 73 0 0 0 17.6 44.0 0 0 17.6 0 0 0 17.1 0 0 0 17.1 1 773 9 42 24 0 0 0 33.4 3.5 0 0 33.4 0 0 0 33.4 0 0 0 33.4 4 770 4 43 51 0 0 0 16.4 2.0 0 0 13.8 0 0 0 18.5 0 0 0 18.5 1 769 9 44 66 0 0 0 23.7 19.7 0 0 9.6 0 0 0 9.6 0 0 0 9.6 2 786 17 45 27 0 0 0 21.7 8.5 0 0 21.7 10 0 0 17.6 10 0 0 17.6 0.1 676 0 46 58 0 0 0 9.0 49.0 0 0 9.0 10 0 0 22.0 10 0 0 22.0 0 668 24.5 47 45 0 0 0 50.7 15.7 30 0 36.2 30 0 0 36.2 30 0 0 36.2 4 667 10.5 49 40 0 0 0 11.8 5.0 0 0 11.8 0 0 0 12.1 70 0 0 4.8 0 659 11.5 50 58 0 0 0 6.8 12.7 5 0 17.7 5 0 0 17.7 5 0 0 17.7 0 670 19.5 51 39 0 0 0 27.5 9.8 80 0 10.5 80 0 0 10.5 80 0 0 10.5 1.5 677 7 53 51 0 0 0 32.8 4.0 0 0 25.5 10 0 0 35.4 10 0 0 35.4 0 676 7.25 54 46 0 0 0 19.7 12.3 70 0 19.1 70 0 0 19.1 70 0 0 19.1 0 681 43.5 55 39 0 0 0 16.0 13.0 0 0 14.0 0 0 0 14.0 0 0 0 14.0 0 693 13.5 56 73 0 0 0 20.4 18.7 0 0 9.7 0 0 0 9.7 0 0 0 9.7 0 697 5.5 57 55 0 0 0 15.8 4.0 5 0 11.7 5 0 0 11.7 5 0 0 11.7 0 691 3.5 58 51 0 0 0 21.6 13.0 0 0 26.0 40 0 0 14.4 40 0 0 14.4 1 689 5.5 59 51 0 0 0 25.5 25.0 0 0 32.9 0 0 0 42.7 0 0 0 42.7 0 689 5.5 60 67 0 0 0 21.3 18.5 0 0 21.3 0 0 0 21.3 0 0 0 21.3 0 680 0 61 11 0 0 0 25.3 11.1 0 0 25.3 0 0 0 25.3 0 0 0 25.3 0 681 10 213  sn RD T1G T1C T1S T1SM T2D T2G T2S T2SM T3G T3C T3S T3SM T4G T4C T4S T4SM O elev. slope 62 25 0 0 0 30.0 9.0 0 0 35.3 80 10 0 17.5 80 10 0 17.5 2 665 26.5 63 13 0 0 0 16.2 6.5 70 0 16.2 70 0 0 16.2 70 0 0 16.2 0 656 0 64 28 60 20 0 31.7 3.0 60 0 31.7 60 20 0 31.7 60 20 0 31.7 0 653 8.5 65 33 0 0 0 43.8 5.8 0 0 43.8 0 0 0 43.8 0 0 0 43.8 6 667 0 66 35 0 0 0 44.4 12.0 0 0 47.9 80 10 0 17.3 80 10 0 17.3 0 656 0 67 11 0 0 0 0.0 0.0 0 0 0.0 0 0 0 0.0 0 0 0 0.0 0 651 5 69 11 0 0 0 0.0 0.0 0 0 0.0 0 0 0 0.0 0 0 0 0.0 0 659 10 70 64 0 0 0 28.0 19.3 0 0 17.6 0 0 0 17.6 0 0 0 17.6 2.5 780 0 71 39 0 0 0 23.1 10.5 0 0 23.1 0 0 0 23.1 0 0 0 23.1 2 781 0.9 72 26 0 0 0 25.5 9.0 30 10 2.3 30 20 10 2.3 30 20 10 2.3 0 778 4 73 34 0 0 0 21.3 14.0 0 0 19.2 0 30 0 17.2 0 30 0 17.2 0 779 0 74 45 0 0 0 17.2 1.7 50 0 9.2 50 0 0 9.2 50 0 0 9.2 0 786 0 75 45 0 0 0 11.7 8.3 80 0 4.8 80 0 0 4.8 80 0 0 4.8 0 789 9 76 36 0 0 0 34.6 12.6 0 0 34.6 0 0 0 34.6 0 0 0 34.6 1.5 787 0 77 25 80 10 0 4.3 7.1 80 0 4.3 80 10 0 4.3 80 10 0 4.3 0.5 788 0 78 42 0 0 0 20.0 20.0 0 0 20.0 0 0 0 22.9 0 0 0 25.8 0 791 0 80 45 0 0 0 26.3 1.5 0 0 31.1 0 0 0 35.8 0 0 0 35.8 0 779 0 81 30 0 0 0 23.8 8.0 0 0 23.8 0 0 0 23.8 0 0 0 23.8 3 781 0 82 37 0 0 0 12.1 9.0 30 0 15.1 30 0 0 15.1 30 0 0 15.1 0 782 11.5 83 25 10 10 0 19.5 7.8 40 0 14.4 40 10 0 14.4 40 10 0 14.4 0 779 11 84 39 30 0 0 10.7 10.3 30 0 10.7 30 0 0 10.7 30 0 0 10.7 0 781 5 85 47 0 0 0 23.4 10.5 10 0 21.7 10 0 0 21.7 10 0 0 21.7 0 774 21.5 86 34 0 0 0 14.6 9.3 10 0 22.5 10 0 0 22.5 10 0 0 22.5 1 784 4 87 30 0 0 0 8.6 11.0 0 0 8.6 30 0 0 5.7 30 0 0 5.7 0 778 13 88 44 0 0 0 22.7 11.3 40 0 10.5 40 5 0 10.5 40 5 0 10.5 3 778 7 89 35 0 0 0 22.0 5.3 40 0 12.5 40 10 0 12.5 40 10 0 12.5 1 773 0 90 33 0 0 0 34.2 7.3 20 0 15.2 20 30 0 15.2 20 30 0 15.2 5 782 19.5 91 61 0 0 0 20.8 14.0 0 0 12.7 0 0 0 12.7 0 0 0 12.7 3 781 0  214  Appendix C: AHP calculations  The Analytic Hierarchy Process (AHP, Saaty 1977, 2001) is a method used for analyzing a problem in a systematic manner. It consists of three basic parts: (1) Decomposition of the problem (here the Yukon working group meetings are deployed); (2) Evaluation of the problem: judging and pair-wise element comparison at each hierarchy level (here the expert group is deployed); and, (3) Synthesis (here a bottom-up approach using priority relations is deployed).  The three steps are described in the following: Step 1 — decomposing the problem (or complexity) into constituent elements: The Yukon working group The AHP Hierarchy: The Yukon working group aided in the development of a static hierarchy of constituent elements (criteria and sub-features) that best described the two potentially conflicting criteria ‗functioning forest ecosystems‘ (labelled L2A in Figure 2.1), and ‗community economic sustainability and benefits‘ (labelled L2B in Figure 2.1). All the hierarchy elements originated from the Strategic Forest Management Plan (SFMP 2004). The terminology used in the SFMP was adopted in order to allow clear identification of the values chosen from that document (see Table A1). The Ratings Table: During the Yukon working group meetings, a ratings table was designed by the participants to help complete the AHP (i.e., judgment of the hierarchy elements; Figure 2.1). The ratings table (Table A2) consisted of three parts: (1) Objective-based values; (2) landscape-level planning approaches; and (3) stand-level actions. The three groups of elements 215  characterize the alternative and competing forest management strategies FMS1–5 (bottom level of the hierarchy in Figure 2.1) based on their respective importance. Landscape and stand scale tactics are based on forest management actions and operations that can be undertaken to achieve the respective FMS.  Step 2 — evaluating the elements of the problem: The SFM-Expert Group The AHP Judgments: The developed ratings table (Table A2) was completed by an independent SFM-expert group (12 UBC researchers) by assigning ratings of ‗low‘ for ‗least important‘ to ‗high‘ for ‗very important‘ to the elements within the AHP hierarchy. The qualitative ratings were then transformed into a ‗9-point Saaty scale‘ (Saaty 2001) in order to use numeric values to assign a rating to each element within the AHP hierarchy (see Figure 2.1). To calculate a score for each FMS, I arithmetically averaged the values of the corresponding elements of each of the hierarchies‘ thematic topics from the ratings table (Table A3). The resulting hierarchy box ratings were then used to generate pair-wise judgments for the AHP hierarchy (see Table A4).  216  Table C1:  Corresponding SFMP elements (on the left), and AHP hierarchy elements (on the right). Round brackets represent the AHP nodes and keywords, with L1 being level 1 of the AHP hierarchy, and FMS1–5 representing the bottom-level alternative forest management strategies (but also see Figure 2.1). SFMP elements AHP hierarchy elements SFMP (L1; SFM) Sustainable Forest Management CATT Functioning Forest Ecosystems (L2A; Environment) Functioning Forest Ecosystems Community Sustainability and Benefits (L2B; Economy) Community sustainability/benefits Support the ecosystem's ability to maintain natural processes (L3AA; Nat. Process) Maintain/enhance natural processes Maintain, restore or enhance forest ecosystem function (including: regeneration & succession, species & ecosystem diversity) (L3AB; Diversity) Maintain/enhance ecosystem diversity Promote a forest industry within the region that is appropriately scaled to resource capacity as guided by forest plans and the socially acceptable level of harvest as defined and recommended by forest management planning (L3BA; Industry) Promote a forest (timber) industry within the region Enable and encourage forest-based activities that stimulate employment opportunities (L3BB; Employment) Enable/encourage forest activities stimulating employment opportunities Maintain forest productivity in areas subject to harvest practices (L4AA1; Productivity) Maintain forest productivity Support the ecosystem's ability to maintain natural processes (L4AA2; Succession) Maintain/enhance/restore forest succession Area & severity of insect/fire/disease disturbance and succession patterns afterwards (as compared against natural range) (L4AA3; Disturbance) Maintain natural disturbances Protect fish and wildlife populations and their habitats, including species and species-at-risk, and biological distinctive or unique features (L4AB1; Habitat) Protect fish/wildlife populations/habitat Maintain naturally occurring quantity and quality of water (L4AB2; Water) Maintain naturally occurring quantity/quality of water Ensure appropriate wildlife movement corridors between important habitats and key landscape features (L4AB3; Connectivity) Ensure wildlife corridors btw. important habitat/key landscape features Strengthen local timber harvesting and processing capacity and the benefits to local business and entrepreneurs through resource certainty by using tenure options appropriate to the region and allocation criteria based on value added, conversion and utilization rates, local training and hiring, local benefits and best practices (L4BA1; Timber) Strengthen local timber harvesting/process capacity  (L4BA2; Biomass) Promote biomass harvesting Support and integrate through forest harvest planning, commercial wilderness tourism values and revenue generating activities and opportunities (L4BB1; Tourism) Tourism  217  SFMP elements AHP hierarchy elements Respect the rights and interests of trappers and outfitters and, where appropriate, support their revenue generating opportunities throughout the planning region (L4BB2; Hunting) Trapping/hunting Number of people employed in forest-based activities broken down by category of forest sector activity (e.g., timber primary production, value-added processing, non-timber sectors such as wilderness tourism, trapping, etc.) (L4BB3; NTFP) Non-timber forest products SFMP priority (FMS1) Manage to support/enhance a sustainable forest industry supposed to representing SFMP (FMS2) Managing for multiple values SFMP key issue (FMS3) Managing for fire risk reduction SFMP priority "wildlife maintenance"  (FMS4) Managing for wildlife SFMP priority "promote forest renewal" (FMS5) Managing for carbon economy 218  Table C2: Ratings table for SFM-experts, with two standard questions: How important is each of the Objective-based values for the respective forest management strategy (FMS)? How important are the management tactics at landscape and stand scale under each FMS? Forest Management Strategies: timber (FMS1), multiple values (FMS2), fire (FMS3), wildlife (FMS4), carbon (FMS5). Verbal ratings: l=low, ml=medium-low, to h=high. In round brackets is listed the class (cultural, environmental, economic); in square brackets is listed the Element Number which will be used for further processing in Table A3. Ratings category Forest Management Strategies SFM-expert ratings Objective based values timber multiple fire wildlife carbon Provide for sustainable harvest non-timber forest products (cultural) [16] h h m mh mh Maintain/enhance hunting/fishing meat hunting (cultural) [4] m m m mh mh Respect the rights and interests of local trappers  (=local importance) (cultural) [5] mh mh h mh h Protect species at risk (environmental) [6] mh h mh h mh Maintain/enhance recruitment/population ∞ habitat (environmental) [7] mh mh mh h mh Maintain/enhance habitat (environmental) [8] mh mh h h h Conserve biodiversity (wildlife/fish/plants) (environmental) [9] m h h h h Maintenance of forest ecosystem health (pests) (environmental) [13] h h mh h h Maintenance of forest ecosystem resilience (environmental) [14] h h mh h h Promote timber harvesting (economic) [1] mh m mh l mh Promote harvest for biomass (economic) [2] mh m mh l m Provide recreation/tourism opportunities (economic) [3] mh m mh m mh Maintain carbon stocks and enhance carbon sequestration (economic) [10] mh m m m h Manage for conservation credits (economic) [11] m m m mh h Maintain forest productivity  (economic) [12] h h m mh h Enable and encourage forest based activities that stimulate employment opportunities (economic) [15] h h m m m  Tactics at landscape scale timber multiple fire wildlife carbon Promote a diversity of seral stages [17] mh h mh h Mh Promote sustainable harvest levels [18] h h m m mh Employ a variety of harvest and regeneration methods [19] mh mh m m m Protect areas of important wildlife habitat [20] m h mh h m Provide appropriate wildlife movement corridors between important wildlife habitats [21] m h mh h m Maintain natural disturbance regimes [22] mh mh m mh m Minimize fragmentation of habitat [23] mh mh m h m Utilize prescribed burning for maintenance of seral stages [24] m m m m ml Promote a variety of harvest block sizes [25] m m m m m Establish restrictions for seasonal access of roads[26] ml m mh mh m 219  Ratings category Forest Management Strategies SFM-expert ratings Objective based values timber multiple fire wildlife carbon Establish (triad) zoning [27] m m m m mh Establish intensive forest mgmt zone [28] mh m mh ml m Establish extensive (multi use) forest mgmt zone [29] mh mh mh m mh Establish conservation mgmt zone [30] m mh mh h mh Establish fuel abatement zone(s) [31] m mh mh mh m  Tactics at stand scale timber multiple fire wildlife carbon Utilize clearcuts [32] m ml m l ml Utilize shelterwood system for regeneration [33] m m m M m Utilize seed trees for regeneration [34] mh mh mh mh mh Utilize group selection [35] m m m m m Utilize prescribed burning (site prep) [36] m m mh ml ml Utilize mechanical site prep [37] mh m m ml m Reduce regeneration lag [38] mh m m m m Utilize fuel reduction treatments [39] m m h m m Apply natural regeneration [40] m h mh h mh Planting on harvest blocks [41] mh m m mh mh Small block sizes [42] ml mh ml m m Large block sizes [43] mh m m ml m Plant willow on harvest blocks (stand conversion) [44] m ml ml ml mh Encourage aspen regeneration (stand conversion) [45] m ml ml ml mh Plant pine [46] m ml ml ml m Single species reforestation (spruce) [47] m ml ml ml m Mixed species reforestation [48] mh mh mh mh m Utilize grouped retention [49] mh mh m m mh Utilize dispersed retention [50] m m ml m m Control invasive species [51] h mh mh h mh Ensure retention of coarse woody debris [52] m m m mh m Do nothing (outside of harvest blocks within a planning area) [53] l l l m l   220  Table C3: From ratings to judgments – to calculate the Saaty judgments (see Table A4), the Saaty transformed ratings from Table A2 are been averaged according to their thematic relation. Left column shows the AHP hierarchy pair (first hierarchy element decides the thematic topic), and left column lists the Element Numbers from Table A2. FMSi = FMS1-5. L2A-L1 6, 7, 8, 9, 11, 13, 14, 17, 20, 21, 22, 23, 24, 25, 30, 40, 51, 52 L3AA-L2A 7, 8, 9, 13, 14, 17, 20, 21, 22, 23, 26, 30, 40, 51, 52 L3AB-L2A 6, 7, 8, 9, 11, 14, 17, 20, 21, 22, 23, 24, 25, 48, 51 L4AA1-L3AA 12, 13, 17, 19, 24, 33, 36, 42, 48, 51 L4AA2-L3AA 13, 14, 17, 22, 24, 35, 40, 48 L4AA3-L3AA 17, 22, 29, 52, 53 L4AB1-L3AB 6, 7, 8, 9, 11, 13, 14, 20, 21, 23, 30 L4AB2-L3AB 7, 8, 9, 13, 14 L4AB3-L3AB 9, 14, 20, 21, 23 L2B-L1 1, 2, 3, 4, 9, 12, 13, 15, 16, 17, 18, 22 L3BA-L2B 1, 12, 13, 14, 15, 17, 18, 19, 33, 39, 40, 42, 45, 48, 51 L3BB-L2B 3, 4, 5, 8, 9, 15, 16, 20, 21, 23, 29 L4BA1-L3BA 1, 12, 13, 15 L4BA2-L3BA 2, 10, 13, 17, 18, 19, 22, 24, 38, 44, 45 L4BB1-L3BB 3, 6, 7, 8, 9, 13, 16, 19, 20, 29, 48 L4BB2-L3BB 4, 5, 7, 8, 9, 16, 20, 21, 29 L4BB3-L3BB 4, 5, 7, 8, 9, 13, 16, 20, 21, 29 FMSi-L4AA1 12, 13, 17, 19, 24, 33, 36, 42, 48, 51 FMSi-L4AA2 13, 14, 17, 22, 24, 35, 40, 48 FMSi-L4AA3 17, 22, 29, 52, 53 FMSi-L4AB1 6, 7, 8, 9, 11, 13, 14, 20, 21, 23, 30 FMSi-L4AB2 7, 8, 9, 13, 14 FMSi-L4AB3 9, 14, 20, 21, 23 FMSi-L4BA1 1, 12, 13, 15 FMSi-L4BA2 2, 10, 13, 17, 18, 19, 22, 24, 38, 44, 45 FMSi-L4BB1 3, 6, 7, 8, 9, 13, 16, 19, 20, 29, 48 FMSi-L4BB2 4, 5, 7, 8, 9, 16, 20, 21, 29 FMSi-L4BB3 4, 5, 7, 8, 9, 13, 16, 20, 21, 29  221  Table C4: Pair-wise comparison judgments. Values are the result of the arithmetic average according to Table A3.     Hierarchy Elements Saaty Judgments Hierarchy Elements Saaty Judgments Hierarchy Elements Saaty Judgments L2A-L1 7.1 FMS2-L4AA1 7.0 FMS4-L4AA1 6.4 L2B-L1 6.8 FMS2-L4AA2 7.5 FMS4-L4AA2 7.5 L3AA-L2A 7.2 FMS2-L4AA3 5.8 FMS4-L4AA3 6.6 L3AB-L2A 7.0 FMS2-L4AB1 7.9 FMS4-L4AB1 8.8 L4AA1-L3AA 6.3 FMS2-L4AB2 8.2 FMS4-L4AB2 9.0 L4AA2-L3AA 6.9 FMS2-L4AB3 8.6 FMS4-L4AB3 9.0 L4AA3-L3AA 5.6 FMS2-L4BA1 8.0 FMS4-L4BA1 5.5 L4AB1-L3AB 7.5 FMS2-L4BA2 6.1 FMS4-L4BA2 5.2 L4AB2-L3AB 8.2 FMS2-L4BB1 7.7 FMS4-L4BB1 7.5 L4AB3-L3AB 7.5 FMS2-L4BB2 7.7 FMS4-L4BB2 7.9 L3BA-L2B 6.6 FMS2-L4BB3 7.8 FMS4-L4BB3 8.0 L3BB-L2B 7.0 L4BA1-L3BA 7.1 FMS3-L4AA1 5.8 FMS5-L4AA1 5.8 L4BA2-L3BA 5.9 FMS3-L4AA2 6.3 FMS5-L4AA2 6.3 L4BB1-L3BB 7.3 FMS3-L4AA3 5.0 FMS5-L4AA3 5.0 L4BB2-L3BB 7.3 FMS3-L4AB1 7.0 FMS5-L4AB1 7.4 L4BB3-L3BB 7.4 FMS3-L4AB2 7.8 FMS5-L4AB2 8.6   FMS3-L4AB3 7.0 FMS5-L4AB3 6.6 FMS1-L4AA1 6.6 FMS3-L4BA1 6.0 FMS5-L4BA1 7.5 FMS1-L4AA2 6.8 FMS3-L4BA2 5.2 FMS5-L4BA2 6.3 FMS1-L4AA3 5.4 FMS3-L4BB1 7.0 FMS5-L4BB1 7.0 FMS1-L4AB1 6.5 FMS3-L4BB2 7.2 FMS5-L4BB2 7.2 FMS1-L4AB2 7.4 FMS3-L4BB3 7.2 FMS5-L4BB3 7.4 FMS1-L4AB3 6.2 FMS1-L4BA1 8.5 FMS1-L4BA2 6.8 FMS1-L4BB1 7.0 FMS1-L4BB2 6.3 FMS1-L4BB3 6.6 222  Step 3 — synthesizing the AHP process to choose the best forest management strategy to solve the problem: The development of priority relations The AHP Priority Relations: The AHP structure, with its judgments and pair-wise comparisons for obtaining the weights of the criteria on all the levels including the alternatives at the bottom level, can be synthesized either top-down as prescribed by Saaty (2008), or bottom-up starting at the lowest hierarchy level, as shown by Kuusipalo and Kangas (1994). Both lead to the identification of the overall priorities for the five alternatives. Here, I present the latter approach as this shows the building of local priorities with the building of weights to reach the global and overall priorities at the top-level of the AHP hierarchy. A consistency check of all the pair-wise comparison matrices revealed a Consistency Ratio (CR) < 1% (according to Saaty (2001) a CR < 10% is acceptable). Diagonalization (i.e., the transformation of a square matrix to a diagonal matrix) of the 11 pair-wise comparison 5x5 matrices (e.g., first matrix in Table A5) for the five alternatives FMSi, i = 1,…,5, corresponding to the 11 cover criteria on level L4 resulted in the identification of local priorities (Table 2.1). Diagonalization of the comparison matrices for the L3 and L2 criteria and sub-criteria (e.g., second matrix in Table A5) resulted in the respective weights, also listed in Table 2.1. The local priorities pi-L4xxj were normalized to one: (1) i L xxj i=1 p = 1 5 4 , for all xx = AA, AB, BA, BB, and all j = 1,2,3 or j = 1,2 as applicable to the 3- or 2-clusters in L4 (Figure 2.1). These priorities are linearly combined with the weights L4xxj L3xx w  of the 223  covering criteria in the four clusters xxj of L4 to obtain the global priorities for the L3 elements as shown in the priority relation (2): (2) i-L4xxj L4xxj L3xx n p  = p wi-L3xx j xxxx xx xx  , where for xx = {AA, AB, BB}, nxx = 3, and for xx = BA, nxx = 2. In total, there were four such priority relations resulting in four global priorities ( pi-L3xx ); each of these priority vectors was normalized to one. In continuing the AHP synthesis, I moved one level up in the hierarchy (L2, Figure 2.1) and use priority relation (3) to calculate the global priorities for the two L2 criteria: (3)  i L A i L AA L AA L A i L AB L AB L A p p w + p w2 3 3 2 3 3 2     . Finally, reaching L1 of the AHP hierarchy (Figure 2.1), I built the priority relation (4) to obtain the overall priorities: (4) p p w  +  p w i L i L A L A L i L B L B L1 2 2 1 2 2 1      224  Table C5: Pair-wise comparison matrices (symmetric reciprocal matrices) for all five levels of the AHP hierarchy totalling 18 pair-wise comparisons.         L1: 1 2x2 matrices        L1 L2A L2B       L2A 1 1.0336       L2B 0.9675 1           L2: 2 2x2 matrices        L2A L3AA L3AB  L2B L3BA L3BB   L3AA 1 1.0343  L3BA 1 0.9380   L3AB 0.9669 1  L3BB 1.0661 1           L3: 3 3x3 matrices, 1 2x2 matrix       L3AA L4AA1 L4AA2 L4AA3  L3BA L4BA1 L4BA2  L4AA1 1 0.9226 1.1367  L4BA1 1 1.2015  L4AA2 1.0839 1 1.2320  L4BA2 0.8323 1  L4AA3 0.8797 0.8117 1           L3AB L4AB1 L4AB2 L4AB3  L3BB L4BB1 L4BB2 L4BB3 L4AB1 1 0.9157 1.0039  L4BB1 1 0.9983 0.9803 L4AB2 1.0920 1 1.0963  L4BB2 1.0017 1 0.9820 L4AB3 0.9961 0.9122 1  L4BB3 1.0201 1.0183 1 225  L4: 11 5x5matrices:            L4AA1 FMS1 FMS2 FMS3 FMS4 FMS5  L4BA1 FMS1 FMS2 FMS3 FMS4 FMS5 FMS1 1 0.9429 1.1379 1.0313 1.1379  FMS1 1 1.0625 1.4167 1.5455 1.1333 FMS2 1.0606 1 1.2069 1.0938 1.2069  FMS2 0.9412 1 1.3333 1.3333 1.0667 FMS3 0.8788 0.8286 1 0.9063 0.9063  FMS3 0.7059 0.7500 1 1.0909 0.8000 FMS4 0.9697 0.9143 1.1034 1 1.1034  FMS4 0.6471 0.7500 0.9167 1 0.7333 FMS5 0.8788 0.8286 1.0000 0.9063 1  FMS5 0.8824 0.9375 1.2500 1.3636 1              L4AA2 FMS1 FMS2 FMS3 FMS4 FMS5  L4BA2 FMS1 FMS2 FMS3 FMS4 FMS5 FMS1 1 0.9000 1.0800 0.9000 1.0800  FMS1 1 1.1194 1.3158 1.3158 1.0870 FMS2 1.1111 1 1.2000 1.0000 1.2000  FMS2 0.8933 1 1.1754 1.1754 0.9710 FMS3 0.9259 0.8333 1 0.8333 1.0000  FMS3 0.7600 0.8507 1 1.0000 0.8261 FMS4 1.1111 1.0000 1.2000 1 1.2000  FMS4 0.7600 0.8507 1.0000 1 0.8261 FMS5 0.9259 0.8333 1.0000 0.8333 1  FMS5 0.9200 1.0299 1.2105 1.2105 1              L4AA3 FMS1 FMS2 FMS3 FMS4 FMS5        FMS1 1 0.9310 1.0800 0.8182 1.0800        FMS2 1.0741 1 1.1600 0.8788 1.1600        FMS3 0.9259 0.8621 1 0.7576 1.0000        FMS4 1.2222 1.1379 1.3200 1 1.3200        FMS5 0.9259 0.8621 1.0000 0.7576 1   226   L4AB1 FMS1 FMS2 FMS3 FMS4 FMS5  L4BB1 FMS1 FMS2 FMS3 FMS4 FMS5 FMS1 1 0.8161 0.9221 0.7320 0.8765  FMS1 1 0.9059 1.0000 0.9277 1.0000 FMS2 1.2254 1 1.1299 0.9299 1.0741  FMS2 1.1039 1 1.1039 1.0241 1.1039 FMS3 1.0845 0.8851 1 0.7938 0.9506  FMS3 1.0000 0.9059 1 0.9277 1.0000 FMS4 1.3662 1.0754 1.2597 1 1.1975  FMS4 1.0779 0.9765 1.0779 1 1.0779 FMS5 1.1408 0.9310 1.0519 0.8351 1  FMS5 1.0000 0.9059 1.0000 0.9277 1              L4AB2 FMS1 FMS2 FMS3 FMS4 FMS5  L4BB2 FMS1 FMS2 FMS3 FMS4 FMS5 FMS1 1 0.9024 0.9487 0.8222 0.8605  FMS1 1 0.8261 0.8769 0.8028 0.8769 FMS2 1.1081 1 1.0513 0.9111 0.9535  FMS2 1.2105 1 1.0615 0.9718 1.0615 FMS3 1.0541 0.9512 1 0.8667 0.9070  FMS3 1.1404 0.9420 1 0.9155 1.0000 FMS4 1.2162 1.0976 1.1538 1 1.0465  FMS4 1.2456 1.0290 1.0923 1 1.0923 FMS5 1.1622 1.0488 1.1026 0.9556 1  FMS5 1.1404 0.9420 1.0000 0.9155 1              L4AB3 FMS1 FMS2 FMS3 FMS4 FMS5  L4BB3 FMS1 FMS2 FMS3 FMS4 FMS5 FMS1 1 0.7209 0.8857 0.6889 0.9394  FMS1 1 0.8462 0.9167 0.8250 0.8919 FMS2 1.3871 1 1.2286 0.9556 1.3030  FMS2 1.1818 1 1.0833 0.9750 1.0541 FMS3 1.1290 0.8140 1 0.7778 1.0606  FMS3 1.0909 0.9231 1 0.9000 0.9730 FMS4 1.4516 1.0465 1.2857 1 1.3636  FMS4 1.2121 1.0256 1.1111 1 1.0811 FMS5 1.0645 0.7674 0.9429 0.7333 1  FMS5 1.1212 0.9487 1.0278 0.9250 1 227  Table C6: Priorities and consistency checking. CI = (lambda max-n)/(n-1); CR=100*(CI/ACI); Eigenvalue (showing Lambda max) and Eigenvectors are generated with the matrix calculator (http://www.bluebit.gr/matrix-calculator/default.aspx). The ACI value varies according to the size of the matrix (Saaty 1980). Eigenvalue Eigenvector normalized CI  CR  Eigenvalue Eigenvector normalized CI  CR 2 0.7187 0.50826721  0.69532 0.49173279  1.41402 1   L2A     L2B 2 0.71892 0.50842639   2 0.68413 0.484001  0.69509 0.49157361    0.72936 0.515999  1.41401 1    1.41349 1   L3AA     L3BA 3 0.58235 0.337425979   2 0.76862 0.545771  0.63119 0.365724914    0.6397 0.454229  0.51232 0.296849107    1.40832 1  1.72586 1 0 0   L3AB     L3BB 3 0.56035 0.323819375   3 0.57318 0.330939  0.61191 0.353615265    0.57413 0.331488  0.55818 0.322565359    0.58467 0.337573  1.73044 1 0 0  1.73198 1 0 0  228  Eigenvalue Eigenvector normalized CI  CR  Eigenvalue Eigenvector normalized CI  CR  L4AA1     L4BA1 4.98 -0.46722 0.2096   5.00 -0.5295 0.239691  -0.49554 0.2223    -0.4899 0.221765  -0.40286 0.1807    -0.3738 0.169193  -0.45306 0.2032    -0.3487 0.157862  -0.41059 0.1842    -0.4672 0.21149  -2.2293 1 -0.0047 -0.4239  -2.20914 1 0.0002 0.0205  L4AA2     L4BA2 5 0.43921 0.1971   5 0.51301 0.230769  0.48801 0.2190    0.45829 0.206154  0.40668 0.1825    0.38989 0.175385  0.48801 0.2190    0.38989 0.175385  0.40668 0.1825    0.47197 0.212307  2.22859 1 0 0  2.22305 1 0 0  L4AA3 5 0.43185 0.1942  0.46384 0.2086  0.39986 0.1799  0.52781 0.2374  0.39986 0.1799  2.22322 1 0 0 229  Eigenvalue Eigenvector normalized CI  CR  Eigenvalue Eigenvector normalized CI  CR  L4AB1     L4BB1 5.00 -0.38233 0.1719   5 0.43111 0.192983  -0.47191 0.2122    0.4759 0.213033  -0.41464 0.1865    0.43111 0.192983  -0.5186 0.2332    0.4647 0.208019  -0.43618 0.1962    0.43111 0.192983  -2.2237 1 4E-05 0.0036  2.23393 1 0 0  L4AB2     L4BB2 5 -0.40263 0.1805   5 0.38873 0.174312  -0.44615 0.2000    0.47057 0.21101  -0.42439 0.1902    0.44329 0.198778  -0.48968 0.2195    0.4842 0.217122  -0.46792 0.2098    0.44329 0.198778  -2.2308 1 0 0  2.2301 1 0 0  L4AB3     L4BB3 5 0.36664 0.1658   5 0.39799 0.1784  0.50856 0.2299    0.47036 0.2108  0.41395 0.1872    0.43418 0.1946  0.53222 0.2406    0.48242 0.2162  0.39029 0.1765    0.44624 0.2000  2.2117 1 0 0  2.23119 1 0 0 230  Appendix D: Soil clustering  Figure D1: Clustering of the 90 ecological research plots into five site types (mesic to sub-hygric), showing the main soil characteristics (texture, rooting depth, coarse fragment) used in the TACA analysis.  231  Appendix E: LANDIS-II modules and parameters Species Parameters: LandisData     ―Species‖ Table E1: LANDIS species parameters. l.=longevity; s.m.=sexual maturity; s.t.=shade tolerance; f.t.=fire tolerance; e.d.=effective distance; m.d.=maximum distance; v.r.p.=vegetative reproduction probability; p-f.r.=post-fire regeneration.      Seed Dispersal  Sprout Age Name l. s.m. s.t. f.t. e.d. m.d. v.r.p. min. max. p-f.r.  pinucont 300 8 1 2 50 3000 0 0 0 serotiny  abielasi 250 20 4 2 100 3000 0 0 0 none  picemari 270 30 4 3 100 200 0.1 30 180 serotiny  piceenge 600 40 4 2 100 200 0 0 0 none betupapy 200 15 1 2 200 5000 0.5 10 100 resprout poputrem 200 15 1 2 1000 5000 0.9 10 100 resprout  tsugmert 500 20 3 3 500 500 0 0 0 none  piceglau 300 25 3 2 30 200 0 0 0 none  popubals 200 20 1 2 1000 5000 0.9 10 100 resprout  larilari 180 45 1 3 25 40 0 0 0 none   Ecoregions Parameters: LandisData     ―Ecoregions‖ Table E2: LANDIS ecoregions. Map Code refers to the LANDIS output in GIS. A=Aihishik, B=Bison, H=Haines Junction ecoregions. Active Map Code Description no 0 non-active (waterbodies, wetland, roads, cities) yes 6 alpine yes 7 non-productive >=6%plants and <10%trees (grass, shrubs) yes 11 A_subhygric yes 13 A_subxeric yes 14 A_submesic yes 15 A_mesic yes 21 B_subhygric yes 22 B_xeric yes 23 B_subxeric yes 24 B_submesic yes 25 B_mesic 232  Active Map Code Description yes 31 H_subhygric yes 32 H_xeric yes 33 H_subxeric yes 34 H_submesic yes 35 H_mesic   Initial Communities Parameters: LandisData     ―Initial Communities‖ Table E3: LANDIS initial communities. Map Code Species Age Corhorts 1 piceglau 10 2 piceglau 10 20 3 piceglau 10 20 30 4 piceglau 10 20 30 80 140 5 piceglau 10 20 70 80 6 piceglau 10 20 70 80 100 7 piceglau 10 20 70 80 90 8 piceglau 10 20 80 110 9 piceglau 10 20 80 90 100 110 120 10 piceglau 10 20 80 90 110 130 11 piceglau 10 20 90 12 piceglau 10 30 50 60 80 13 piceglau 10 30 60 90 14 piceglau 10 30 70 80 90 100 15 piceglau 10 30 80 90 16 piceglau 10 30 80 90 100 110 17 piceglau 10 60 80 90 18 piceglau 20 30 40 50 100 120 130 160 170 280 19 piceglau 20 30 40 60 70 100 120 130 20 piceglau 20 30 40 60 70 90 120 130 21 piceglau 20 30 40 70 90 100 22 piceglau 20 30 60 70 100 23 piceglau 20 30 60 70 120 130 24 piceglau 20 30 60 90 100 140 25 piceglau 20 30 60 90 130 26 piceglau 20 30 70 80 90 27 piceglau 20 30 70 80 90 120 233  Map Code Species Age Corhorts 28 piceglau 20 30 70 90 110 29 piceglau 20 30 80 130 30 piceglau 20 30 90 100 120 31 piceglau 20 30 90 100 120 160 170 32 piceglau 20 30 90 100 140 180 33 piceglau 20 30 90 120 140 180 34 piceglau 20 300 70 100 35 piceglau 20 40 50 80 100 120 130 36 piceglau 20 40 60 80 37 piceglau 20 40 70 120 130 38 piceglau 20 40 70 80 100 130 39 piceglau 20 40 80 100 40 piceglau 20 40 80 90 100 41 piceglau 20 40 90 100 120 130 42 piceglau 20 50 60 90 110 130 43 piceglau 20 50 90 100 130 44 piceglau 20 50 70 80 100 110 130 45 piceglau 20 50 90 100 150 46 piceglau 20 50 90 100 150 180 47 piceglau 20 60 80 90 48 piceglau 20 70 100 140 170 200 49 piceglau 20 70 110 50 piceglau 20 70 80 130 170 51 piceglau 20 70 90 100 52 piceglau 20 80 110 150 180 260 53 piceglau 20 80 150 180 54 piceglau 20 80 90 170 180 250 55 piceglau 30 40 50 60 100 140 56 piceglau 30 50 70 90 110 130 140 57 piceglau 30 50 90 58 piceglau 30 50 90 100 120 59 piceglau 30 60 100 120 60 piceglau 30 60 80 110 120 61 piceglau 40 70 90 110 130 160 62 piceglau 40 60 90 110 130 63 piceglau 50 60 90 64 piceglau 60 70 65 piceglau 30 90 100 110 66 piceglau 40 80 100 120 130 67 piceglau 30 80 120 130 68 piceglau 30 80 69 piceglau 40 70 110 234  Map Code Species Age Corhorts 70 piceglau 30 90 120 160 180 190 230 71 piceglau 40 90 100 150 180 200 220 72 piceglau 40 50 90 150 180 190 220 73 piceglau 40 90 120 150 74 piceglau 40 90 120 150 190 220 75 piceglau 50 80 100 76 piceglau 60 90 120 140 77 piceglau 40 90 100 130 78 piceglau 60 90 120 160 79 piceglau 70 90 120 170 190 80 piceglau 40 90 130 170 190 200 81 piceglau 40 90 100 170 180 82 piceglau 50 90 120 130 170 83 piceglau 50 80 110 120 84 piceglau 40 80 110 130 85 piceglau 40 100 110 86 piceglau 40 80 110 130 180 190 87 piceglau 40 80 110 130 180 190 200 88 piceglau 40 80 110 140 89 piceglau 60 80 110 130 90 piceglau 70 100 150 160 91 piceglau 70 100 130 92 piceglau 40 90 100 170 93 piceglau 60 90 110 180 190 94 piceglau 60 90 110 150 180 95 piceglau 50 90 110 160 200 97 piceglau 60 80 120 130 160 190 220 98 piceglau 40 80 110 140 160 180 230 99 piceglau 50 90 110 160 100 poputrem 10  piceglau 10 101 piceglau 10 20  poputrem 10 20 102 piceglau 10 20 80  poputrem 10 20 60 103 piceglau 10 20 80 90 110  poputrem 60 70 110 104 piceglau 10 20 90  poputrem 10 20 70 105 piceglau 10 20 90 120  poputrem 10 20 60 110 106 poputrem 20 50 90 130 235  Map Code Species Age Corhorts  piceglau 30 60 90 110 120 107 piceglau 20 70 90 100 150 160  poputrem 10 20 90 120 108 poputrem 50 110 130  piceglau 40 70 80 100 120 109 poputrem 50 90 120  piceglau 40 60 80 120 110 poputrem 50 90 130  piceglau 20 70 80  90 110 111 poputrem 60 110  popubals 60 70 90 120 112 poputrem 70 110 120  piceglau 40 70 90 110 120 130 150 113 poputrem 70 110 140  piceglau 70 90 110 120 114 poputrem 60 110 120  piceglau 30 70 80 100 120 130 150 115 poputrem 70 110 150  piceglau 60 70 90 110 120 120 pinucont 30  piceglau 80 110 120 130 130 piceglau 50 90 100 130 140 170  popubals 30 60 120 131 piceglau 50 60 90 110 130 150  popubals 30 60 90 120 132 piceglau 60 90 110 130 160  popubals  70 90 130 133 piceglau 50 70 80  90 100 120 160 170  popubals 60 90 140 134 popubals 10  piceglau 10 135 popubals 10 20  piceglau 10 20 136 piceglau 50 70 80  90 100 120 160 170  popubals 50 60 90 130 140 popubals 10 20  poputrem 10 20 141 popubals 50 90 130 150  poputrem 50 90 130 150 142 poputrem 60 110 120  popubals 60 70 90 120 150 143 poputrem 50 110 120 140 236  Map Code Species Age Corhorts  popubals 50 70 90 120 150 160 144 poputrem 70 110 120  popubals 60 90 140 145 poputrem 40 50 60 70 80 90 130 140  popubals 60 70 150 tsugmert  220 250 300 400 420  piceenge  200 230 300 460 480 151 tsugmert  10 50 150 190  piceenge  20 80 120 130 152 tsugmert  220 250 300 400 410  piceenge  200 230 300 460 470 500 200 piceglau 60 90 110 130 150 201 piceglau 40 70 90 110 130 150 180 190 202 piceglau 40 70 90 110 130 203 piceglau 40 70 90 110 120 204 piceglau 40 70 90 110 130 150 180 190 220 205 piceglau 50 70 90 110 140 206 piceglau 50 70 90 110 140 150 207 piceglau 30 70 90 110 140 160 208 piceglau 40 70 90 110 130 150 170 190 210 209 piceglau 30 70 90 110 140 170 210 piceglau 30 70 90 110 140 170 211 piceglau 30 70 90 110 130 180 212 piceglau 50 70 90 110 130 160 213 piceglau 40 60 90 110 130 160 214 piceglau 60 90 110 130 160 180 215 piceglau 70 90 100 130 216 piceglau 70 90 100 130 150 217 piceglau 50 90 100 130 140 218 piceglau 50 90 100 130 140 170 219 piceglau 60 90 100 130 140 220 piceglau 30 80 110 221 piceglau 40 60 80 110 120 160 222 piceglau 40 70 80  90 130 170 223 piceglau 40 70 80  90 130 224 piceglau 40 70 80  90 100 120 225 piceglau 40 70 80  90 100 120 160 227 piceglau 40 70 80  90 120 140 228 piceglau 50 70 80 90 100 120 150 229 piceglau 70 80  90 100 120 130 230 piceglau 70 80  90 100 120 231 piceglau 60 80  90 100 237  Map Code Species Age Corhorts 232 piceglau 50 70 80 100 120 130 150 160 233 piceglau 50 70 80 100 120 130 150 160 180 234 piceglau 40 60 130 150 235 piceglau 50 60 130 150 160 236 piceglau 70 60 130 140 237 piceglau 40 70 80 100 120 130 150 160 170 238 piceglau 30 70 80 100 120 130 150 160 239 piceglau 70 80 100 120 160 180 240 piceglau 60 80 100 120 140 241 piceglau 60 80 100 120 140 150 242 piceglau 70 80 100 120 130 243 piceglau 50 90 100 120 160 244 piceglau 30 70 80 100 120 130 150 245 piceglau 40 70 90 100 120 130 140 246 piceglau 40 70 90 100 120 130 140 150 247 piceglau 30 80 90 100 120 140 160 248 piceglau 30 80 90 100 140 150 249 piceglau 50 80 90 100 140 150 160 170 250 piceglau 30 80 90 100 140 251 piceglau 40 80 90 100 140 150 160 252 piceglau 60 80 90 100 140 150 180 253 piceglau 40 60 80 90 100 130 254 piceglau 50 80 90 100 140 150 170 255 piceglau 40 80 90 100 140 256 piceglau 70 80 90 120 130 257 piceglau 50 80 90 120 258 piceglau 40 60 80 90 110 259 piceglau 40 60 80 90 110 130 260 piceglau 50 60 80 90 110 130 150 261 piceglau 40 70 80 100 120 262 piceglau 40 70 80 263 piceglau 50 60 80 110 120 264 piceglau 50 60 80 110 120 140 265 piceglau 60 80 110 120 140 150 266 piceglau 30 60 80 110 120 140 150 267 piceglau 30 60 80 110 130 268 piceglau 40 60 100 120 130 140 269 piceglau 50 60 100 150 270 piceglau 50 60 100 140 271 piceglau 50 60 100 140 160 272 piceglau 40 50 60 100 140 160 170 273 piceglau 50 60 100 130 238  Map Code Species Age Corhorts 274 piceglau 40 50 60 100 120 275 piceglau 50 70 110 130 140 276 piceglau 40 70 120 130 140 277 piceglau 50 70 110 130 140 170 278 piceglau 50 70 110 130 140 160 279 piceglau 50 70 110 150 280 piceglau 40 50 70 110 130 140 160 281 piceglau 40 60 70 90 110 120 150 282 piceglau 40 50 70 90 110 130 283 piceglau 50 70 90 110 130 140 160 284 piceglau 30 70 90 110 120 140 150 285 piceglau 30 70 90 100 120 130 286 piceglau 40 50 70 100 110 130 140 287 piceglau 40 70 90 110 130 150 288 piceglau 30 70 90 120 140 289 piceglau 40 50 80 130 290 piceglau 50 90 130 300 poputrem 10 301 poputrem 10 20 302 poputrem 40 50 60 303 poputrem 40 80 130 160 304 poputrem 70 80 90 130 305 poputrem 60 70 306 poputrem 60 80 100 307 poputrem 60 110 308 poputrem 70 100 320 popubals 10 20 400 picemari 60 70 80 100 401 picemari 20 50 60 120 200 402 picemari 100 120 160 450 betupapy 50 60 150 451 betupapy 70 90 100 130 452 betupapy 20 50 60 120 453 betupapy 20 30 50 100 110 500 larilari 70 100 140 160  abielasi 50 90 140 170  pinucont 40 80 130 180 501 larilari 20 70  abielasi 20 50 60 100 130  pinucont 30 60 70 90 110 150 502 larilari 90 100 120 150  abielasi 80 90 140 170 239  Map Code Species Age Corhorts  pinucont 100 130 190 220 260  240  Module Biomass Succession: LandisData     ―Biomass Succession v2‖  Table E4: LANDIS Biomass per ecoregion. S.H.=Shade Class; A=Aishihik, B=Bison, H=Haines Junction; x=xeric, sx=subxeric, sm=submesic, m=mesic, sh=subhygric, np=non-productive; alp=alpine. S.C. A- sm B- sm H- sm B- x H- x A- sx B- sx H- sx A- m B- m H- m A- sh B- sh H- sh np alp 1 5 5 5 5 5 5 5 5 5 5 5 5 5 5 100 100 2 30 30 30 30 30 30 30 30 30 30 30 30 30 30 100 100 3 45 45 45 45 45 45 45 45 45 45 45 45 45 45 100 100 4 60 60 60 60 60 60 60 60 60 60 60 60 60 60 100 100 5 95 95 95 95 95 95 95 95 95 95 95 95 95 95 100 100  Table E5: LANDIS establishment probabilities (historic). spp=species; A=Aishihik, B=Bison, H=Haines Junction; x=xeric, sx=subxeric, sm=submesic, m=mesic, sh=subhygric, np=non-productive; alp=alpine. All species establishment probabilities (including the climate change ones are derived from TACA modeling). spp A- sm B-sm H- sm B-x H-x A-sx B-sx H-sx A-m B-m H-m A-sh B-sh H-sh np alp pinucont 0.519 0.449 0.392 0.349 0.353 0.502 0.449 0.392 0.519 0.449 0.392 0.519 0.449 0.392 0 0 abielasi 0 0 0.262 0 0.039 0 0 0.114 0 0 0.281 0 0 0.281 0 0 picemari 0.46 0.471 0.509 0.28 0.132 0.273 0.405 0.395 0.477 0.504 0.509 0.477 0.504 0.509 0 0 piceenge 0.363 0.15 0.1 0.051 0 0.128 0.117 0 0.515 0.15 0.119 0.515 0.15 0.119 0 0 betupapy 0.034 0 0.3 0 0.084 0 0 0.222 0.051 0 0.3 0.051 0 0.3 0 0 poputrem 0.502 0.428 0.349 0.305 0.308 0.484 0.394 0.349 0.502 0.428 0.349 0.502 0.428 0.349 0 0 tsugmert 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 piceglau 0.585 0.51 0.569 0.31 0.321 0.466 0.444 0.549 0.585 0.51 0.569 0.585 0.51 0.569 0 0 popubals 0.135 0.144 0.226 0.055 0 0.078 0.055 0.036 0.237 0.191 0.442 0.27 0.254 0.48 0 0 larilari 0.385 0.373 0.471 0.13 0.037 0.203 0.29 0.145 0.52 0.437 0.546 0.52 0.437 0.546 0 0  241  Table E6: LANDIS establishment probabilities (2020s). spp A- sm B-sm H- sm B-x H-x A-sx B-sx H-sx A-m B-m H-m A-sh B-sh H-sh np alp pinucont 0.52 0.465 0.444 0.36 0.33 0.434 0.441 0.432 0.52 0.465 0.444 0.52 0.465 0.444 0 0 abielasi 0 0 0.241 0 0.007 0 0 0.057 0.025 0.006 0.354 0.036 0.036 0.373 0 0 picemari  0.46 0.46 0.48 0.268 0.062 0.247 0.397 0.297 0.554 0.523 0.505 0.554 0.523 0.505 0 0 piceenge 0.079 0.165 0.013 0.029 0 0.02 0.118 0 0.241 0.183 0.046 0.27 0.212 0.058 0 0 betupapy  0.172 0.116 0.244 0.019 0.044 0.095 0.073 0.152 0.214 0.165 0.265 0.214 0.165 0.265 0 0 poputrem 0.492 0.502 0.351 0.355 0.233 0.395 0.466 0.332 0.492 0.502 0.351 0.492 0.502 0.351 0 0 tsugmert 0.001 0 0.033 0 0 0 0 0.022 0.002 0 0.034 0.002 0.001 0.035 0 0 piceglau 0.535 0.498 0.537 0.334 0.174 0.345 0.441 0.433 0.57 0.527 0.543 0.57 0.527 0.543 0 0 popubals 0.165 0.233 0.117 0.038 0 0.102 0.142 0.012 0.32 0.316 0.329 0.447 0.388 0.474 0 0 larilari 0.266 0.313 0.322 0.078 0.006 0.128 0.223 0.055 0.485 0.379 0.492 0.514 0.424 0.516 0 0  Table E7: LANDIS establishment probabilities (2050s). spp A- sm B-sm H- sm B-x H-x A-sx B-sx H-sx A-m B-m H-m A-sh B-sh H-sh np alp pinucont 0.495 0.432 0.453 0.291 0.327 0.398 0.396 0.416 0.518 0.45 0.459 0.518 0.45 0.459 0 0 abielasi 0.113 0.156 0.212 0 0.006 0.068 0.098 0.052 0.237 0.195 0.308 0.321 0.26 0.4 0 0 picemari  0.418 0.447 0.441 0.247 0.059 0.236 0.372 0.24 0.563 0.505 0.492 0.575 0.53 0.504 0 0 piceenge 0.222 0.266 0.052 0.06 0 0.134 0.192 0 0.454 0.294 0.079 0.539 0.33 0.097 0 0 betupapy  0.185 0.097 0.263 0.021 0.045 0.1 0.061 0.155 0.266 0.146 0.295 0.279 0.171 0.308 0 0 poputrem 0.431 0.399 0.355 0.246 0.243 0.319 0.35 0.316 0.462 0.424 0.361 0.462 0.424 0.361 0 0 tsugmert 0.028 0.001 0.006 0 0 0.005 0.001 0.003 0.04 0.002 0.006 0.041 0.003 0.007 0 0 piceglau 0.477 0.472 0.558 0.308 0.177 0.302 0.414 0.405 0.556 0.513 0.583 0.562 0.532 0.59 0 0 popubals 0.135 0.239 0.097 0.05 0 0.085 0.139 0.013 0.254 0.309 0.28 0.45 0.42 0.515 0 0 larilari 0.245 0.359 0.284 0.1 0.006 0.153 0.271 0.051 0.459 0.412 0.445 0.54 0.487 0.562 0 0   242  Table E8: LANDIS establishment probabilities (2080s). spp A- sm B-sm H- sm B-x H-x A-sx B-sx H-sx A-m B-m H-m A-sh B-sh H-sh np alp pinucont 0.447 0.457 0.415 0.3 0.272 0.358 0.414 0.365 0.507 0.474 0.427 0.513 0.485 0.427 0 0 abielasi 0.114 0.219 0.195 0.073 0.006 0.072 0.15 0.051 0.18 0.254 0.333 0.352 0.336 0.435 0 0 picemari  0.363 0.415 0.392 0.264 0.057 0.22 0.344 0.214 0.508 0.475 0.469 0.58 0.51 0.494 0 0 piceenge 0.212 0.261 0.086 0.097 0 0.125 0.211 0.013 0.327 0.285 0.119 0.448 0.328 0.132 0 0 betupapy  0.134 0.159 0.284 0.075 0.044 0.075 0.103 0.163 0.206 0.215 0.339 0.265 0.234 0.358 0 0 poputrem 0.434 0.459 0.398 0.299 0.226 0.334 0.405 0.332 0.512 0.465 0.411 0.519 0.478 0.411 0 0 tsugmert 0.01 0.006 0.003 0.002 0.001 0.008 0.005 0.002 0.012 0.007 0.004 0.015 0.008 0.006 0 0 piceglau 0.466 0.486 0.509 0.282 0.132 0.307 0.415 0.352 0.582 0.529 0.589 0.631 0.552 0.601 0 0 popubals 0.169 0.311 0.089 0.058 0 0.105 0.2 0.012 0.308 0.385 0.236 0.551 0.504 0.548 0 0 larilari 0.233 0.383 0.241 0.13 0.006 0.137 0.293 0.06 0.349 0.434 0.422 0.558 0.526 0.576 0 0   243  Module Biomass Harvest: LandisData     ―Biomass Harvest‖ ManagementAreas C:\..\..\management.gis <<map delineating MA-boundaries Stands      C:\..\..\stand.gis <<map delineating forest stand boundaries Table E9: LANDIS harvesting scenario M3 (development of ‗firebreaks‘ = fuel treatment: replace current coniferous stands with aspen). StandRanking: MaxCohortAge.Plant poputrem. Mgmt Area    Harvest Area     Begin Time   End Time 116 100% 41 50 118 100% 41 50 104 100% 10 40 128 100% 10 40 100 100% 10 40 111 100% 10 40 134 100% 10 40 132 100% 10 40 155 100% 10 40 117 100% 10 40 151 100% 20 50 133 100% 20 50 154 100% 20 50 150 100% 41 50 114 100% 20 50 136 100% 20 50 113 100% 20 50 126 100% 20 50 139 100% 20 50 153 100% 20 50 115 100% 20 40 142 100% 41 50 157 100% 20 20 156 100% 30 40 120 100% 30 40 119 100% 30 40 125 100% 41 50 135 100% 30 40 152 100% 40 50 123 100% 40 50 129 100% 40 50 138 100% 40 50 244  Mgmt Area    Harvest Area     Begin Time   End Time 127 100% 40 50 124 100% 40 50 112 100% 41 50 122 100% 11 20 100 100% 40 50 103 100% 50 50 110 100% 50 50 121 100% 50 50 143 100% 50 50 137 100% 50 50 141 100% 50 50 101 100% 50 50 140 100% 50 50 131 100% 50 50 130 100% 50 50 102 100% 50 50   Module Dynamic Fuel System: LandisData     ―Dynamic Fuel System‖ HardwoodMaximum 15 DeadFirMaxAge 40 Table E10: LANDIS fuel types. Fuel Type Base Fuel Age Range Species 2 Conifer 101 to 600 pinucont piceglau picemari piceenge abielasi larilari tsugmert 4 Conifer 41 to 100 pinucont piceglau picemari piceenge abielasi larilari tsugmert 3 Conifer 0 to 40 pinucont piceglau picemari piceenge abielasi larilari tsugmert 8 Deciduous  11 to 200 betupapy poputrem popubals 9 Conifer 0 to 200  piceglau betupapy poputrem popubals pinucont piceenge abielasi larilari tsugmert picemari 10 Conifer 0 to 200  piceglau betupapy poputrem popubals pinucont piceenge abielasi larilari tsugmert picemari 11 Conifer 0 to 200  piceglau betupapy poputrem popubals pinucont piceenge abielasi larilari tsugmert picemari 12 Conifer 0 to 200  piceglau betupapy poputrem popubals pinucont piceenge abielasi larilari tsugmert picemari 17 Open - - 245  Fuel Type Base Fuel Age Range Species 18 Deciduous  0 to 200 poputrem betupapy popubals     Module Fire System: LandisData     ―Dynamic Fire System‖ GroundSlopeFile         C:\..\..\azi.gis <<topographic map showing slopes UphillSlopeAzimuthMap   C:\..\..\upslope.gis <<topographic map showing aspects  Table E11: LANDIS fire seasons. Season Name leaf Satus Prop of Fires Percent Curing DayLength Proportion Spring LeafOff 0.19 100 1 Summer LeafOn 0.73 0 1 Fall LeafOff 0.08 90 1  Table E12: LANDIS fire region parameters: e.c.=Eco Code; e.n.=Eco Name (A=Aishihik, B=Bison, HJ=Haines Junction matrix, E=between B and C, C=Canyon, H=Haines Junction, S=wildlife sanctuary); mu=average; sigma=standard deviation; m.f.s.=maximum fire size; FMC=Fine Fuel Moisture Code, with h.=high, l.=low, p.=proportion; o.f.=open fuel; #F.= number of fires.      FMC Spring FMC Summer FMC Fall e.c. e.n. mu si. m.f.s. h. l. p. h. l. p. h. l. p. O.F. #F. 1 A 3.61 2.28 10000 85 100 0.5 90 120 0.5 120 120 0.5 17 0.33 2 B 3.61 2.28 10000 85 100 0.5 90 120 0.5 120 120 0.5 17 0.226 3 HJ 3.61 2.28 10000 85 100 0.5 90 120 0.5 120 120 0.5 17 0.288 4 E 3.61 2.28 10000 85 100 0.5 90 120 0.5 120 120 0.5 17 0.104 5 C 3.61 2.28 10000 85 100 0.5 90 120 0.5 120 120 0.5 17 0.004 6 H 3.61 2.28 10000 85 100 0.5 90 120 0.5 120 120 0.5 17 0.038 7 S 3.61 2.28 10000 85 100 0.5 90 120 0.5 120 120 0.5 17 0.001   246  Module Biomass by Age Output LandisData     ―Biomass AgeClass‖ Table E13: LANDIS ageclass output. spp ageclass 1 ageclass 2 ageclass 3 pinucont <40 40-100 >100 abielasi <40 40-100 >100 picemari  <40 40-100 >100 piceenge <40 40-100 >100 betupapy  <40 40-80 >80 poputrem <40 40-80 >80 tsugmert <40 40-100 >100 piceglau <40 40-100 >100 popubals <40 40-80 >80 larilari <40 40-100 >100   Module Biomass Reclass: LandisData     ―Reclass Biomass Output‖ Table E14: LANDIS forest types. Forest Type Species Composition Aspen poputrem MixedWood piceglau poputrem popubals Deciduous poputrem betupapy popubals SprucePine pinucont piceglau Sprucefir abielasi piceglau Birch betupapy BWBS piceglau picemari Tamarack larilari Pine pinucont BalsamPop popubals Blackspruce picemari Hemlock  tsugmert Engelspruce  piceenge Subfir abielasi Novelconifer piceenge tsugmert abielasi pinucont picemari larilari NovelMix piceenge tsugmert abielasi pinucont picemari larilari betupapy SpruceBirch  piceglau betupapy 247  Appendix F: Point fire ignitions   Figure F1: Digital Elevation Model (DEM) showing the CATT region with the study landscape delineated in white. Point fire ignitions are shown as white dots. DEM: dark = low elevation, white =higher elevation. 248  Appendix G: Seral stage distribution Table G1: Seral stage distribution. Means and standard deviations are given in hectares. year fire climate early seral mid seral mid-late late seral 0   14125  754  11990  10135    Mean Std Dev Mean Std Dev Mean Std Dev Mean Std Dev 20 0 0 84 6 14468 0 5846 0 6690 0  0 1 83 6 14468 0 5846 0 6690 0  1 0 1061 743 13672 755 5343 380 5979 627  1 1 857 501 13833 488 5424 431 6225 385 50 0 0 72 7 14460 6 1127 0 21501 0  0 1 66 6 14459 6 1127 0 21501 0  1 0 1776 910 13603 562 869 238 19898 906  1 1 2389 806 12813 1070 811 241 19542 780 80 0 0 25 4 156 6 14472 0 22532 0  0 1 25 4 149 6 14472 0 22532 0  1 0 1315 648 2559 1478 12328 1113 20209 1117  1 1 2155 685 2814 787 11428 1004 19768 696 150 0 0 69 4 432 9 546 8 36127 0  0 1 78 6 409 10 540 9 36127 0  1 0 1947 1095 2906 1336 2349 1399 29314 1962  1 1 2296 1003 3653 1536 2408 857 28122 1790  249  Appendix H: Projected climate change CATT     Figure H1: Projected climate change CATT. Above: Precipitation and average temperature. Below: Effective precipitation.  0 5 10 15 20 25 30 0 1 2 3 4 2020s 2050s 2080s p ro je c te d  c h a n g e in  p er c en ta g e p ro je c te d  c h a n g e in  d eg re e C el si u s precipitation high CC precipitation low CC Tmean high CC Tmean low CC -10 -8 -6 -4 -2 0 2020s 2050s 2080s p er c en ta g e ef fe c ti v e p re c ip it a ti o n effective precipitation high CC effective precipitation low CC 250  Appendix I: Harvesting distribution  Functions F1–F3 describe the approximate distribution of harvesting over time (Figure F1). The mathematical integration (e.g., ‗area underneath the curve‘) is always close to 5,910 hectares. F1 (M1, harvesting over 150 years): 2.9496*x 3 -44.983*x 2 +89.914*x+561.67 F2 (M2, harvesting over 100 years): -0.11009235*x 6 +3.5514519*x 5 - 44.64402394*x 4 +276.800319*x 3 -874.534419*x 2 +1273.34583*x F3 (M3, harvesting over 50 years): -412.71*x 4 +4870.6*x 3 - 19596*x 2 +31090*x-14902   Figure I1: Harvesting distribution for LANDIS modelling. M1-3 are the three LANDIS harvesting scenarios.  0 500 1000 1500 2000 2500 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 h ar v es ti n g  i n  h ec ta re s decades M1 M2 M3 251  Appendix J: Indicators listed in the SFMP Table J1: Excerpt of the Strategic Forest Management Plan for the CATT (SFMP 2004) listing all Objectives (O) and respective Indicators (I) for the two Goals used in the thesis (‗ecology‘ and ‗economy‘). SFMP elements and explanations Goal A Functioning Forest Ecosystems O1 Maintain, restore or enhance forest ecosystem function (including: regeneration & succession, species &ecosystem diversity) I1 Variation in forest mosaic (pattern, composition, structure, age-class distribution) I2 Level of abundance and diversity of wildlife populations I3 Extent and diversity of key ecosystem features O2 Support the ecosystem's ability to maintain natural processes I1 Area & severity of insect/fire/disease disturbance and succession patterns afterwards (as compared against natural range) I2 Area & severity of human-caused disturbances and succession patterns afterwards (level of success in mimicking normal range of variability of natural disturbances) I3 Negative impacts of re-introduced species (e.g., bison) I4 Woody debris O3 Protect fish and wildlife populations and their habitats, including species and species-at-risk, and biological distinctive or unique features I1 Change in abundance and distribution of forest dependent species classified as species-at-risk I2 Change in abundance of fish and wildlife species that play key ecosystem roles I3 Change in productivity of selected species (e.g., moose, whitefish) I4 Change in landscape and quality of suitable habitat (land and water) for selected or valued species or species-at-risk I5 Status of unique or distinctive biological features (e.g., important staging and nesting areas, areas of high wildlife concentrations) and activities or arrangements to protect them I6 Conservation management and protection arrangements (e.g., pans, programs, agreements, habitat and species designation, etc.) O4 Ensure appropriate wildlife movement corridors between important habitats and key landscape features I1 Status of areas of suitable habitat for existing and potential movement corridors, considering factors such as connectivity, fragmentation and habitat quality 252  SFMP elements and explanations I2 Habitat monitoring arranges and reports O5 Maintain naturally occurring quantity and quality of water I1 Percentage of forest managed primarily for water protection (e.g., riparian buffers) I2 Water data (flow, temperature, turbidity, etc. and current conditions compared with stream-specific historic information and values) O6 Maintain forest productivity in areas subject to harvest practices I1 Change in productive forest land base over time I2 Mean annual increment of average forest productivity of the planning region (m3/ha/year) I3 Amount of harvested area with significant soil compaction, displacement, erosion, loss of organic matter, etc. O7 Establish an information base of the best available scientific, local and traditional knowledge and experience to guide forest management and planning I1 An established information base designed and shared by CAFN, Yukon and Parks Canada Agency I2 Information from the information base used in forest operations and development plans O8 Integrate monitoring with harvesting activities and utilize monitoring to assess the condition of the forest, harvest levels, management activities and their socio-economic and environmental effects I1 Reporting from scheduled monitoring activities that include community participation and multi-agency involvement O9 Implement an explicitly defined adaptive management strategy in response to the results of monitoring programs I1 An established adaptive forest management strategy with a clear methodology and consistent procedures that can be replicated over time to provide comparison of results and changes Goal B Community Sustainability and Benefits O1 Manage forest uses and developments consistent with ecosystem capacity and long-term sustainability I1 Lower level forest management plans that are consistent and complaint with the SFMP I2 An established adaptive management strategy with a clear methodology and consistent procedures that can be replicated over time to provide comparison of results and changes O2 Enable and encourage forest-based activities that stimulate employment opportunities I1 Number of people employed in forest-based activities broken down by category of forest sector activity (e.g., timber primary production, value-added processing, non-timber sectors such as wilderness tourism, trapping, etc.) I2 Investment in training to promote best practices related to SFM 253  SFMP elements and explanations I3 Information about job satisfaction for forest-based workers (survey) O3 Optimize the use of the forest land base for commercial timber management where appropriate and desirable. This may involve practicing a full range of forest management activities from "intensive" stand management on specific areas where the focus is on providing for continued timber production, to full integration where multiple values are to be managed euqitably, to areas of no-harvest where "other" values and uses restrict forestry operations I1 Proportion of area commercially harvested relative to the land base available for timber production I2 Amount and proportion of forest land harvested and regenerated in a manner that meets their assigned values I3 Distribution of commercial harvest patterns relative to the land base available for timber production I4 Volume of merchantable wood left on the site after harvest I5 Total value of value-added forest product manufacturing I6 Information about the effectiveness of silvicultural treatments O4 Promote a forest industry within the region that is appropriately scaled to resource capacity as guided by forest plans and the socially acceptable level of harvest as defined and recommended by forest management planning I1 Commercial timber allocations that are consistent with forest management plans I2 High social acceptance of the "woods operation" measured by the number of complaints per year O5 Strengthen local timber harvesting and processing capacity and the benefits to local business and entrepreneurs through resource certainty by using tenure options appropriate to the region and allocation criteria based on value added, conversion and utilization rates, local training and hiring, local benefits and best practices I1 Number of locally owned operations I2 Number of local people employed I3 Operational assessments of value-added I4 Operational assessments of Conversion and utilization rates I5 Operational assessments of Local training and hire I6 Operational assessments of Local benefits I7 Operational assessments of Best practices O6 Respect the rights and strengthen the traditional use of forest resources by CAFN citizens I1 Status of important traditional use areas I2 Level of consumption and use of traditional foods and other products 254  SFMP elements and explanations I3 Level of participation in traditional use activities O7 Provide for a sustainable domestic harvest of wood, meat, fish, berries and other forest products I1 Level of harvest and effort by residents I2 Accessible wood harvest areas for local use I3 Areas designated as primary personal use areas (e.g., community woodlot) I4 Level of satisfaction with harvest opportunities O8 Respect the rights and interests of trappers and outfitters and, where appropriate, support their revenue generating opportunities throughout the planning region I1 Volume of fur harvested I2 Number of active concessions I3 Number of compensation claims I4 Number of non-resident hunting licenses issued I5 Number of total client days I6 Fewer use conflicts (by survey) I7 Demonstrated consultation requirements and working relationships between operators and  trappers and outfitters (e.g., specified consultation requirements during planning phases and as a condition of timber permits) O9 Support and integrate through forest harvest planning, commercial wilderness tourism values and revenue generating activities and opportunities I1 Total revenue generated by tourism lodges and business in the region I2 Cumulative access impacts on pristine values O10 Where appropriate, increase the amount of diversity of recreational forest-based activities I1 Land and resource base available for selected recreational activities (e.g., hunting, fishing, angling, backcountry travel, ect.) I2 Level of activity (participation) in selected recreational activities I3 Hunter/angling effort surveys I4 Level of satisfaction with recreational opportunities (by community surveys) I5 Sustainability of selected resources or features (e.g., fishing holes) O11 Protect known cultural and historic sites for current and future generations I1 Status of identified cultural and historic sites O12 Maintain or enhance visual quality of viewscapes and forest aesthetics within the region 255    SFMP elements and explanations I1 Number of visual quality objectives I2 Information from community surveys about viewscape values and views about how these are being addressed I3 Percentage of valued viewscapes that has been cut or significantly affected by natural disturbances (fire, insects, storms) 256  Appendix K: Calculation of indices used in this thesis  Spruce bark beetle: Endemic beetles such as Dendroctonus rufipennis can attack and kill individual trees. When the beetle population exceeds the endemic level, it can successfully mass attack and overcome younger and green trees (Garbutt 2004). Favorable forest conditions (e.g., high host susceptibility) for such mass attacks include mature, even-aged (e.g., white spruce) stands in contiguous patches. Normally, the beetle first attacks the largest (older) trees in a stand, and then moves to smaller trees as the infestation worsens (Matsuoka et al. 2001). Spruce beetle- related tree mortality is a function of: elevation and variation in mean temperature during July and August (favouring beetle reproduction); forest age and site quality (influencing forest susceptibility); maximum wind speed (promoting beetle dispersal); winter mortality of beetles (warmer winters imply more survivors for the following year); and high densities of susceptible trees (more favourable host conditions (Safranyik et al. 1990, Fettig et al. 2007)). Mott (1963) defined the term susceptibility as ―the probability of a forest being attacked‖ by a biological disturbance agent [e.g., spruce bark beetle, aspen leaf miner], and the term vulnerability ―as the probability of tree mortality resulting from a given level of attack.‖  Spruce bark beetle (Dendroctonus rufipennis) risk in this study was calculated based on (1) white spruce seral stage biomass output from LANDIS modeling (output derived from the succession module), and (2) on edaphic site conditions (Figure 3.2 in Chapter 3). (1) The seral stages that were considered (2–4) ranged from 40 to greater than 200 years of age. Seral stage 1 was not considered in the beetle risk calculation because according to Allen et al. (2006), trees with a diameter at breast height (dbh) of less than 15 cm are not susceptible to beetle attack. Seral stage 2 was subdivided into two risk classes, a lower risk ―1‖ and a higher risk ―3‖, with 30 257  t/ha as a biomass threshold since it represented the 15 cm dbh threshold (based on empirical data, and Table 2 in Hogg and Wein 2005). Seral stage 3 was assigned the highest risk class (i.e., ―3‖ as all trees are assumed to be above 15cm in dbh). Seral stage 4 (i.e., trees >200 years) was assigned a risk value between moderate and high (i.e., risk ―2.5‖) as the oldest white spruce trees are likelier to survive beetle attack (Doak 2004). In a next step I multiplied the assigned risk with a physiological modifier. (2) I considered site edaphic conditions from xeric to sub-hygric via a physiological modifier which reflected tree-susceptibility to successful beetle attack (i.e., a xeric soil weakens the spruce due to drier site conditions. For example Hogg and Wein 2005 state that spruce regeneration on drier sites is slowed relative to wetter sites; I found (in Chapter 3) that white spruce best performed on moister sites, and it is known that drought stress may cause a growth decline in white spruce (Driscoll et al. 2005). Thus weights of 1.3 and 1.2 were assigned to xeric and sub-xeric sites, respectively, to increase beetle vulnerability (i.e., increasing the risk of beetle attack); a weight of 0.7 and 0.6 was assigned to mesic and sub-hygric sites, respectively, to decrease the beetle risk (i.e., the trees were more fit to withstand an attack), and a weight of 1 was assigned to trees on mesic sites (e.g., no modification of the risk or susceptibility thereof).  Leaf miner (Phyllocnistis populiella) risk was calculated based on (1) aspen age and (2) stand type (i.e., pure aspen stand, mix aspen-spruce stands) derived from biomass output of the LANDIS succession modeling (Table J1). (1) The critical threshold for aspen age was set at 20 years, as this is when it is most vulnerable (Yukon FHR 2009). (2) We assumed that higher mixedwood density would reduce susceptibility to leaf miner attacks, as has been shown for other pests. We considered densities of mixed aspen and white spruce stands ranging from low 258  density (300 stems per hectare) to high density (540 stems per hectare) white spruce stands, which respectively reflected a gradient of high risk (high relative aspen density) to low risk stands (low relative aspen density). Table K1: Leaf miner risk matrix based on aspen age and white spruce density (ρ) [stems per hectares].Thresholds (i.e., 300, 540) are based on empirical data: mean and SD of plots with At and Sw were considered. S2-4 = seral stage 2-4.  Aspen leaf miner (Phyllocnistis populiella) risk (matrix presented in Table 4.1) was calculated using aspen stand age (>20; ≤20 years of age) and white spruce seral stage [stems/ha]; this is based on biomass data [t/ha] derived from LANDIS modeling. The stems per hectare calculation was as follows: V  biomass) boleby  divided biomass Sw ground above of (ratio biomass) stage (seral hectare)per  stems stage (seral                   H H HDBH 2 1.3- )5.0( 3 (0.45)(1.355) biomass) stage (seral hectare)per  stems stage (seral   With seral stage biomass [t*ha -1 ] ratio of above ground white spruce (Sw) biomass dry weight divided by bole biomass dry weight equal to 1.355, derived from Table 2 in Yarie and Van Cleve (1983); white spruce density ρ as 0.45[t*m-3]; seral stage bole volume V [m3] = (π/3*r2*H), radius r [m] calculated by applying the Theorème de Thalès / harmonische Teilung, stem height H [m]; and DBH at 1.3 [m] off the ground surface. Seral stages are: S2 40-100 years, S3 100-200, S4 >200 (Table J2).  Pure aspen stand Mixed stand low Sw ρ Mixed stand medium Sw ρ Mixed stand high Sw ρ At-age [years] (independent of ρ) 0<(S2+S3+S4)<300 300<(S2+S3+S4)<540 540<(S2+S3+S4) At<=20 H  H  M  L At>20 H  M L  L 259   Table K2: White spruce height and diameters. Empirical forest stand data from blocks 3-6 was used (mean height H, and mean DBH for seral stages S2-4).  S2 S3 S4 DBH 0.146 0.194 0.240 H 15.0 17.8 19.0  The numbers from Table J2 were derived from regression functions based on age/DBH and age/H of the cored and dated tree samples. The majority of the study landscape‘s white spruce forests was reflected by the samples in blocks 3–6 (e.g., similar topography, climate). Note: I ignored S1 assuming it was like an open stand (based on personal field observations); hence it is like a pure aspen stand.  I set thresholds for aspen cover (see Table 4.1) for high, medium, and low density of Sw stems in mixed At/Sw stands based on the biomass output of the modeled E0 scenario representing original baseline environmental conditions (for forest coverage/types see table 4.2): low Sw density reflected in the 1 st  quartile (<300 [stems/ha]), high density reflected in the 3 rd  quartile (>540), and the medium range between 300 and 540.  Forest types: Forest composition was assessed using the distribution of the dominant forest types (white spruce and aspen) with the addition of mixed forests, Spruce-Pine stands and Black 260  and White Boreal Spruce Forests (BWBS). Maximum stand age was combined with forest type (both being outputs from LANDIS) to measure the change in forest structure and composition under different management scenarios (i.e., the Desired State or the Management State).

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.24.1-0072761/manifest

Comment

Related Items