Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Integration of remote sensing and spatial conservation prioritization approaches for aiding large-area,… Powers, Ryan Paul 2015

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

Item Metadata

Download

Media
24-ubc_2015_may_powers_ryan.pdf [ 13.92MB ]
Metadata
JSON: 24-1.0167183.json
JSON-LD: 24-1.0167183-ld.json
RDF/XML (Pretty): 24-1.0167183-rdf.xml
RDF/JSON: 24-1.0167183-rdf.json
Turtle: 24-1.0167183-turtle.txt
N-Triples: 24-1.0167183-rdf-ntriples.txt
Original Record: 24-1.0167183-source.json
Full Text
24-1.0167183-fulltext.txt
Citation
24-1.0167183.ris

Full Text

INTEGRATION OF REMOTE SENSING AND SPATIAL CONSERVATION PRIORITIZATION APPROACHES FOR AIDING LARGE-AREA, MULTI-JURISDICTIONAL BIODIVERSITY CONSERVATION IN CANADA’S BOREAL FOREST  by Ryan Paul Powers BSc, University of Victoria, 2006 MSc, University of Calgary, 2009  A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF  DOCTOR OF PHILOSOPHY in THE FACULTY OF GRADUATE AND POSTDOCTORAL STUDIES (FORESTRY)  THE UNIVERSITY OF BRITISH COLUMBIA (VANCOUVER)  April 2015  © Ryan Paul Powers, 2015 ii  Abstract Remote sensing is an important complementary data source to enable cost effective monitoring and mapping of biodiversity indicators over large extents in a consistent and repeatable manner. As such, remote sensing is capable of supporting the information needs of modern biodiversity conservation efforts. However, a number of critical challenges and opportunities deserve greater attention. The aim of this research is to advance the use of remote sensing and other geospatial techniques for large-area, multi-jurisdictional conservation of Canada’s boreal forest. Outcomes of this dissertation contributed to progress in each of four research themes: (i) assessing biodiversity across broad areas, (ii) identifying areas of high conservation priority (iii) evaluating the efficacy of current and hypothetical reserve networks, and (iv) incorporating future vegetation variability in conservation planning. The overall research findings indicate the tremendous capacity of the Canadian boreal forest to provide suitable areas for conservation investment and demonstrate the usefulness of these coarse-scale approaches for guiding ongoing research aimed at boreal conservation planning. Key findings included: - Reserves that were restricted to only intact forest landscapes were less flexible and efficient (more costly) - Reserves using accessibility (distance from road and human settlement) as a cost surrogate were able to satisfy a range of conservation targets and compactness levels while remaining remote from human influence - Reserves (≥1000 km2; <10000 km2) were relatively less variable   iii  - Climate change impacts (estimated vegetation productivity variability) greatly influences the cost of reserve networks and the amount of area required to meet conservation targets - Conservation of more sites spread across locations with higher variable vegetation probability values, yet low cost (wilderness areas), proved most efficient   - Reserve networks optimized under “current” or “least change (B1)” conditions are unlikely to maintain their current representative targets in 2080          iv  Preface This research was initially proposed by Trisalyn Nelson and Nicholas Coops, and carried out and lead by Ryan Powers who (i) established the research objectives and questions, (ii) conducted the data analysis, (iii) wrote the papers and manuscript, and (iv) revised papers and communicated with academic referees. In Chapter 6, co-author Trisalyn Nelson processed the estimated Dynamic Habitat Index vegetation productivity for the years 2020, 2050, and 2080 that were then used to further assess estimated vegetation variability. Also in Chapter 6, co-author Vivitskaia Tulloch processed the Marxan with probabilities (MarProb) runs using inputs and the experimental design formulated by Ryan Powers.    A version of Chapter 2 (section 2.1) has been published in:  Powers, R.P., N.C. Coops, J.L. Morgan, M.A. Wulder, T.A. Nelson, C.R. Drever and S.G. Cumming. (2013). A remote sensing approach to biodiversity assessment and regionalization of the Canadian boreal forest. Progress in Physical Geography. 37(1), 36-62.    A version of Chapter 3 has been published in:  Powers, R.P., N.C. Coops, J.L. Morgan, M.A. Wulder, T.A. Nelson, C.R. Drever and S.G. Cumming. (2013). A remote sensing approach to biodiversity assessment and regionalization of the Canadian boreal forest. Progress in Physical Geography. 37(1), 36-62.    A version of Chapter 22 (section 2.2-2.4) and Chapter 4 has been published in:  Powers, R.P., N.C. Coops, T.A. Nelson, M.A. Wulder, and C.R. Drever. (2013). Integrating accessibility and intactness into large-area conservation planning in the Canadian boreal forest. Biological Conservation. 167, 371-379.     A version of Chapter 5 has been submitted for publication as:  Powers, R.P., N.C. Coops, T.A. Nelson, and M.A. Wulder. An evaluation of large-scale reserve design efficacy using a long Time-Series Earth observation dataset.    A version of Chapter 6 has been submitted for publication as:  Powers, R.P., N.C. Coops, V.J. Tulloch, S.E. Gergel, T.A. Nelson, and M.A. Wulder. A conservation assessment of Canada’s boreal forest in the face of climate change. v  Table of Contents Abstract .................................................................................................................................................................. ii Preface ................................................................................................................................................................... iv Table of Contents ................................................................................................................................................... v List of Tables ....................................................................................................................................................... viii List of Figures ....................................................................................................................................................... ix List of Acronyms ................................................................................................................................................... xi Acknowledgements .............................................................................................................................................. xii Dedication ........................................................................................................................................................... xiii 1. Introduction .................................................................................................................................................... 1 1.1. General background, objectives and chapter overview ......................................................................... 1 1.2. Remote sensing of biodiversity ............................................................................................................. 7 1.2.1. Land cover (direct) ........................................................................................................................... 9 1.2.2. Topography (indirect) ..................................................................................................................... 11 1.2.3. Vegetation productivity (indirect) .................................................................................................. 14 1.2.4. Disturbance (indirect) ..................................................................................................................... 14 1.2.5. Fragmentation (indirect) ................................................................................................................. 15 1.2.6. Snow cover (indirect) ..................................................................................................................... 16 1.3. Regionalization ................................................................................................................................... 17 1.4. Spatial conservation prioritization (SCP) ............................................................................................ 18 1.5. SCP and incorporating system dynamics and anticipated climate variability ..................................... 20 1.6. SCP and system design criteria and considerations ............................................................................ 25 2. Study area and data ....................................................................................................................................... 27 2.1. Study area............................................................................................................................................ 27 2.1.1. Land cover ...................................................................................................................................... 28 2.1.2. Topography (ruggedness) ............................................................................................................... 31 2.1.3. Vegetation productivity .................................................................................................................. 31 2.1.4. Snow cover ..................................................................................................................................... 33 2.1.5. Disturbance and fragmentation ....................................................................................................... 33 2.1.6. Species data .................................................................................................................................... 35 2.2. Environmental domains and species-at-risk distribution data ............................................................. 36 2.3. Naturalness surrogate (intact forest landscapes) ................................................................................. 39 2.4. Human “access” as a cost surrogate .................................................................................................... 40 3. A remote sensing approach to biodiversity assessment and regionalization of the Canadian boreal forest . 42 3.1. Introduction ......................................................................................................................................... 42 3.2. Methods .............................................................................................................................................. 49 vi  3.2.1. Cluster analysis ............................................................................................................................... 49 3.2.2. Stepwise regression ........................................................................................................................ 51 3.3. Results ................................................................................................................................................. 52 3.3.1. Correlations between indicators ...................................................................................................... 52 3.3.2. Cluster analysis ............................................................................................................................... 53 3.3.3. Cluster attribution with supplementary indicators .......................................................................... 58 3.4. Indicators as predictors of species richness ......................................................................................... 60 3.5. Discussion ........................................................................................................................................... 63 4. Integrating accessibility and intactness into large-area conservation planning in the Canadian boreal forest..   ...................................................................................................................................................................... 67 4.1. Introduction ......................................................................................................................................... 67 4.2. Methods .............................................................................................................................................. 71 4.2.1. Data ................................................................................................................................................. 71 4.2.2. Prioritization approach and analysis ............................................................................................... 71 4.3. Results ................................................................................................................................................. 74 4.3.1. Reserve efficiency, total area, and spatial prioritization ................................................................. 74 4.4. Discussion ........................................................................................................................................... 80 5. Evaluating reserve design efficacy in the Canadian boreal forest using time series AVHRR data .............. 85 5.1. Introduction ......................................................................................................................................... 85 5.2. Methods .............................................................................................................................................. 90 5.2.1. Data ................................................................................................................................................. 90 5.2.1.1. Remotely sensed data: time series of DHI productivity ......................................................... 90 5.2.1.2. Boreal stratification ................................................................................................................ 92 5.2.1.3. Candidate reserves based on mid-2000 conditions ................................................................ 95 5.2.2. Statistical analyses .......................................................................................................................... 95 5.3. Results ................................................................................................................................................. 96 5.3.1. Ecozone strata ................................................................................................................................. 96 5.3.2. Productivity strata ........................................................................................................................... 99 5.3.3. Land cover strata........................................................................................................................... 101 5.4. Discussion ......................................................................................................................................... 103 6. A conservation assessment of Canada’s boreal forest incorporating alternate climate change scenarios .. 111 6.1. Introduction ....................................................................................................................................... 111 6.2. Methods ............................................................................................................................................ 116 6.2.1. Conservation features ................................................................................................................... 116 6.2.2. Quantifying vegetation variability ................................................................................................ 116 6.2.3. Conservation targets and planning units ....................................................................................... 118 6.2.4. Prioritization approach and analysis ............................................................................................. 118 6.3. Results ............................................................................................................................................... 123 vii  6.3.1. Reserve efficiency, total area and proportion of VVP values ....................................................... 123 6.3.2. Spatial prioritization and site selection frequency ........................................................................ 126 6.3.3. Comparison of the proportion of VVP values under different climate/productivity conditions ... 133 6.4. Discussion ......................................................................................................................................... 135 6.4.1. Location and characterization of high priority areas..................................................................... 135 6.4.2. Influence and implications of different climate conditions on reserve VVP ................................ 137 6.4.3. Climate change adaptation considerations .................................................................................... 137 7. Conclusions ................................................................................................................................................ 140 7.1. Goal 1: Identifying critical habitat for conservation ......................................................................... 141 7.2. Goal 2: Improving boreal conservation assessments ........................................................................ 143 7.3. Limitations ........................................................................................................................................ 146 7.4. Future directions ............................................................................................................................... 148 References .......................................................................................................................................................... 150 Appendices ......................................................................................................................................................... 174         viii  List of Tables Table  1.1: Examples of studies which have used remotely sensed or other geospatial data to map or model environmental indicators related to biodiversity in the Canadian boreal forest. (adapted from Powers et al., 2013a) ............................................................................................................................................................. 12 Table  1.2: Marxan equation terms and definitions (Ball et al., 2009) .................................................................. 20 Table  1.3: Advantage and disadvantage overview of four methods for accounting for natural disturbance dynamics within conservation planning in intact areas. A description of these methods is provided in Figure  1.2 (adapted from Leroux and Rayfield et al., 2013) ........................................................................... 23 Table  2.1: Landscape pattern metrics used to characterize cluster groupings (adapted from Wulder et al., 2008b; Soverel et al., 2010) ........................................................................................................................................ 34 Table  2.2: Priority species based on threat status, geographic distribution and data availability ......................... 39 Table  3.1: A sample of approaches used to map or model biodiversity (species richness) using remotely sensed data in Canada for six general taxonomic classes ........................................................................................... 47 Table  3.2: Pearson’s correlation analysis of environmental indicators................................................................. 53 Table  3.3: Discriminant analysis of cluster input variables. ................................................................................. 53 Table  3.4: Description of the fifteen clusters and the relative indicator ranking. Rankings were derived from mean indicator values  per cluster and defined by the natural breaks (Jenks) classification scheme. ............. 57 Table  3.5: Description of the fifteen clusters with anthropogenic change and forest fragmentation indicator mean values..................................................................................................................................................... 59 Table  3.6: Mean species richness per cluster ........................................................................................................ 60 Table  3.7: Summary of stepwise regression - cluster input indicators and species richness (N=100) .................. 62 Table  4.1: Description of the commonly selected areas. Feature rankings were derived from mean indicator values from environmental domains (Powers et al., 2013a) and defined by the natural breaks (Jenks) classification scheme. ..................................................................................................................................... 80 Table  5.1: Strata and data sources for the stratification of the Canadian boreal ................................................... 92 Table  5.2: Number of years that DHI metrics significantly differ (±3 STD) from the ecozone baseline ............. 98 Table  5.3: Number of years that DHI metrics significantly differ (±3 STD) from the productivity strata baseline ...................................................................................................................................................................... 101 Table  5.4: Number of years that DHI metrics significantly differ (±3 STD) from the land cover strata baseline ...................................................................................................................................................................... 103 Table  6.1: MarProb term description and definitions ......................................................................................... 121 Table  6.2: List of scenarios identifying the method, software, probability, and targets used. In total there were eight assessments (4 scenarios × 2 representative targets). ........................................................................... 123 Table  6.3: Description of the commonly selected areas. Feature rankings were derived from median indicator values from 27 “DHI” variability maps, species richness (Powers et al., 2013b), accessibility cost (Powers et al., 2013b), and defined by the natural breaks (Jenks) classification scheme. .............................................. 131   ix  List of Figures Figure  1.1 Organization of thesis chapters (orange) and key data sources (grey) .................................................. 7 Figure  1.2: Overview of four methods for accounting for natural disturbance dynamics within conservation planning in intact areas. (1) This method optimizes solutions for areas represented by spatial catalysts of ecological processes (e.g., river network and intactness). (2) This method uses probabilistic models to assess the likelihood of a planning unit (i.e., individual grid or site) containing a target or feature (e.g., a given serial stage) over a specified timeframe. In this example, solutions are optimized to be representative of each serial stage over the specified timeframe. Darker shades indicate a higher probability that a planning unit contains an occurrence of a serial stage. (3) This method simulates natural disturbance over time to provide valuable information (e.g., the minimum reserve size for maintaining conservation features) for setting robust targets that account for ecological process (e.g., stand replacing wildfire). (4) This method first provides optimal solutions like those mentioned in (1; darker are better sites), then evaluates the efficacy of those solutions using (3) over a given timeframe. Solutions can then be modified based on any target deficiencies highlighted by the evaluation. (adapted from Leroux and Rayfield, 2013) ................................ 24 Figure  2.1: Study area (dark grey) encompassing the entire Canadian boreal forest as defined by Brant (2009) 28 Figure  2.2: HWL class decision tree. (adapted from DUC, 2010) ........................................................................ 30 Figure  2.3: (a) Spatial distribution of 15 environmental domains (Powers et al., 2013a). (b) Species richness of 16 threatened species. (c) Global Forest Watch Canada (GFWC) intact forest landscape (Lee et al., 2010) and current protected areas (IUCN I-IV). (d) Spatial distribution of access cost surrogate. ........................... 38 Figure  3.1: Spatial distribution of 15 environmental clusters within the study area, encompassing the Canadian boreal forest as defined by Brandt (2009) ....................................................................................................... 56 Figure  3.2: Normal probability plots of residuals for each multiple regression model: (A) bird, (B) butterfly, (C) mammal and (D) forest. .................................................................................................................................. 61 Figure  4.1: Total area prioritized and relative cost (reserve cost/total reserve cost) of best scenario solutions for three different representative area-based targets: (a) 15%, (b) 25%, and (c) 35%. Reserve cost is determined by cost (based on the area or accessibility cost surrogate) of prioritized areas. Total reserve cost refers to the sum of the cost for all candidate priority areas. .............................................................................................. 75 Figure  4.2: Best scenario solutions (top panel) and selection frequencies (bottom panel) for different targets (15–35%) for the same compactness level (moderately compact). (a) Area cost surrogate incorporated. (b) Access cost surrogate incorporated. (c) Access cost surrogate incorporated and prioritization restricted to intact forest landscapes only. Selection frequency is used to determine how often a specific candidate priority area (i.e., 5 km2 grid) is selected over the 500 runs, and provides an indication of its relative importance for an efficient reserve design. ..................................................................................................... 77 Figure  4.3: Areas commonly prioritized (>50%) in all scenario runs. Numbers correspond to area description in Table  4.1 ......................................................................................................................................................... 79 Figure  5.1: Map of (a) ecozones; (b) relative MODIS GPP productivity classes with unproductive (≤3000 kg C/m2/yr) in blue, productive (>3000 kg C/m2/yr; ≤7000 kg C/m2/yr) in pink, and highly productive (>7000 kg C/m2/yr) in red; and (c) MODIS UMD land cover classes with open shrubland in yellow green, evergreen needleleaf in dark green, and mixed forest in lime green. Ecozones in (a) are 1. Boreal Cordillera, 2. Taiga Cordillera, 3. Taiga Plains, 4. Southern Artic, 5. Boreal Plains, 6. Boreal Shield, 7. Hudson Plains, and 8. Taiga Shield. ................................................................................................................................................... 94 Figure  5.2: Reserve size (small, medium and large) score of the occurrence of major deviations (±3 STD) from the ecozone baseline means for the six dynamic habitat metrics for the years 1987-2006. Numbered ecozones are 1. Boreal Cordillera, 2. Taiga Cordillera, 3. Taiga Plains, 4. Southern Artic, 5. Boreal Plains, 6. Boreal Shield, 7. Hudson Plains, and 8. Taiga Shield. Values 0-14 represent the number of major deviations x  out of the 21 years. For example, a value of 10 represents 10 years the fall outside (±3 STD) the ecozone baseline mean. ................................................................................................................................................. 97 Figure  5.3: Comparison of 1987 to 2006 mean AVHRR cumulative annual fPAR and seasonal variation to baseline means of small reserves (<1000 km2) at (a) unproductive (<=3000 kg C/m2/yr), (b) productive (>3000 kg C/m2/yr; <=7000 kg C/m2/yr) and (c) highly productive (>7000 kg C/m2/yr) sites. The shaded area shows the ±3 standard deviation for the AVHRR DHI components (blue) and baseline (grey). .......... 100 Figure  5.4: Comparison of 1987 to 2006 mean AVHRR cumulative annual fPAR to baseline land cover means of small reserves (<1000 km2) at (a) mixed forest, (b) evergreen needleleaf and (c) open shrubland sites. The shaded area shows the ±3 standard deviation for the AVHRR DHI component (blue) and baseline (grey). 102 Figure  5.5: Comparison of 1987 to 2006 mean AVHRR cumulative annual fPAR to baseline South Artic ecozone means of (a) small reserves (<1000 km2), and (b) large reserves (> 4,000 km2; <=10,000 km2). The shaded area shows the ±3 standard deviation for the AVHRR DHI component (blue) and baseline (grey). 105 Figure  6.1: Map of 2080 VVP for IPCC climate change scenarios (a) B1 least extreme change, (b) A1B business as usual and (c) A2 most extreme change. Dark grey areas indicate currently protected areas (IUCN I-IV) .............................................................................................................................................................. 120 Figure  6.2: Number of prioritized sites (a) and relative cost (reserve cost/total reserve cost) of best scenario solutions (b) for three different scenarios with VVP and the one scenario based on current conditions. T15 and T25 represent the 15% and 25% representative targets respectively. In (a) the proportion of the reserve network`s VPP values was determine using natural breaks, where very dark teal represents low VVP sites (≤ 30%), or locations in 2080 likely containing similar levels of productivity as current conditions, and medium and light teal represent high (≥ 40%) and very high (≥ 45%) VVP respectively. ......................................... 125 Figure  6.3: Best scenario solutions for different targets (15 and 25%) using the same compactness level and planning unit cost. (a) A1B VVP incorporated. (b) A2 VVP incorporated. (c) A2 VVP incorporated. (d) Prioritization based on current conditions without VVP............................................................................... 128 Figure  6.4: Blue gradient maps represent selection frequencies for different targets (15 and 25%) using the same compactness level and planning unit cost. (a) A1B VVP incorporated. (b) A2 VVP incorporated. (c) B1 VVP incorporated. (d) Current conditions without VVP. Selection frequency is used to determine how often a specific planning unit or site (i.e., 10 km2 grid) is selected over the 200 runs, and provides an indication of its relative importance for an efficient reserve design. ................................................................................. 129 Figure  6.5: Sites commonly prioritized for scenarios with VVP. (a) Overlapping best solutions for the 25% target. (b) Areas frequently selected (> 95%) in the 200 scenario (2, 3 and 4) runs for the 15% target. (c) Areas frequently selected (> 95%) in the 200 scenario (2, 3 and 4) runs for the 25% target. There is 100% overlap between (a) and (b). Numbers correspond to area description in Table  6.3. .................................... 130 Figure  6.6: Box plot output per frequently selected areas (0-6; x-axis) and 2080 DHI variability (S) values (y-axis) under B1 (a), A1B (b) and A2 (c) conditions. Small boxes within the box plot indicate median DHI variability; boxes represent 25th and 75th percentile, and the entire range of the DHI values is indicated by horizontal markers outside the box plot. ....................................................................................................... 132 Figure  6.7: Comparison of the number of prioritized sites and the proportion of VVP values. T15 and T25 represent the 15% and 25% representative target respectively. The proportion of the reserve network`s VVP was determine using natural breaks, where very dark teal represents low VVP sites (≤ 30%), or locations in 2080 likely containing similar levels of productivity as current conditions, and medium and light teal represent high (≥ 40%) and very high (≥ 45%) VVP respectively. (a) Reserve network optimized for A1B VVP under A2 and B1 conditions. (b) Reserve network optimized for A2 VVP under A1B and B1 conditions. (c) Reserve network optimized for B1 VVP under A1B and A2 conditions. (d) Reserve network optimized for current conditions without VVP under B1, A1B and B2 future 2080 conditions. .................. 134   xi  List of Acronyms AAFC  Agriculture and Agri-Food Canada  AIC  Akaike Information Criterion AVHRR Advanced Very High Resolution Radiometer  BLM  Boundary Length Modifier CBFA  Canadian Boreal Forest Agreement  CCRS  Canadian Center for Remote Sensing  COSEWIC Committee on the Status of Endangered Wildlife in Canada, CV  Coefficient of Variation DHI  Dynamic Habitat Index DUC  Ducks Unlimited Canada  EOSD  Earth Observation for Sustainable Development of forests  FPAC  Forest Products Association of Canada  fPAR  Fraction of Photosynthetically Active Radiation FSC  Fractional Snow Cover GFWC  Global Forest Watch Canada GIS  Geographic Information System  GPP  Gross Primary Productivity GTOPO30 USGS Global 30-Arc Second Elevation Data Set HWL  Hybrid Wetland Layer IPCC  Intergovernmental Panel on Climate Change  MODIS Moderate resolution Imaging Spectrometer NARR  North American Regional Reanalysis NCEP  National Centers for Environmental Protection NDVI  Normalized Difference Vegetation Index NLWIS National Land and Water Information Service  PCIC  Pacific Climate Impacts Consortium  SCP  Spatial Conservation Prioritization  SRTM  Shuttle Radar Topographic Mission  UMD  University of Maryland  USGS  United States Geological Survey VVP  Vegetation Variability Probability     xii  Acknowledgements The long road of undertaking a doctoral dissertation, like many of life’s other great challenges, is almost always paved with the support of others, and to thank everyone who has helped along the way (wittingly or not) would certainly take a chapter in its own right. Intellectual debts are owed first and foremost to my brilliant supervisor Nicholas Coops, whose patient prodding and exacting criticisms (sweetened with encouragement) has been vital to seeing this project to (timely) completion. My deep thanks are also extended to Mike Wulder and the members of my supervisory committee, Trisalyn Nelson, John Nelson, and Sarah Gergel for their friendly advice and many invaluable suggestions over the years.  Obvious thanks must go to the University of British Columbia whose financial support (mostly in the form of Graduate Fellowships) was very helpful. Additional financial support was gratefully received from the Department of Forest Resources Management through a suite of scholarships, fellowships, awards and teaching assistantships. Additionally, this dissertation was made possible by NSERC funding (PGSD) and the generous support of Nicholas Coops as part of the ‘BioSpace: Biodiversity monitoring with Earth Observation data’ project’ and joint funding through the Ivey Foundation, the Nature Conservancy of Canada, Canadian Space Agency (CSA) Government Related Initiatives Program (GRIP), and the Canadian Forest Service (CFS) Pacific Forestry Centre (PFC).  Some pain is inevitable when carrying out a doctoral dissertation. Suffering, however, is optional. Thanks to my wife, whose enduring love and inspiration is like a soothing balm that helps me face even the most trying of days. I cannot hope to repay all the support that she has given me. Thanks also go to a supporting cast of family, friends and the IRSS gang, and many others for bringing joy into my life and keeping me (relatively) sane.    xiii  Dedication :To my Nazishka اي شب از روياي تو رنگين شده  سينه از عطر توام سنگين شده  روي چشم من گسترده خويش ه اي ب شاديم بخشيده از اندوه بيش همچو باراني كه شويد جسم خاك   لودگي ها كرده پاكآهستيم ز  اي تپش هاي تن سوزان من  من مژگانآتشي در سايه  اي ز گندمزارها سرشارتر  پربارتراي ز زرين شاخه ها  اي در بگشوده بر خورشيد ها در هجوم ظلمت ترديدها با توام ديگر ز دردي بيم نيست  خوشبختيم نيست جز درد ،هست اگر فروغ فرخزاد   1  1. Introduction 1.1. General background, objectives and chapter overview The Canadian boreal forest encompasses approximately 5.5 million square kilometers, most of which remains intact (~82%), making it one of the largest forest ecosystems in the world (Brandt, 2009). Canada’s boreal forest contains expansive forests, wetlands and lakes and provides essential habitat and sustains a diverse range of plants and animals as well as offering many essential ecosystem services vital for human and planetary health such as regulating regional and global climate, sequestering large amounts of carbon, and purifying water (Price et al., 2013). Increasingly, human activity, mostly in the form of land conversion (e.g., agriculture, forestry, mining and gas exploration, road construction) (Anielski and Wilson, 2009) continues to intensify and expand into the region. This expansion coupled with the anticipated increase in climate variability has the potential to expedite the loss of boreal biodiversity in Canada (Anielski and Wilson, 2009).   The creation of large protected areas from naturally functioning ecosystems that are largely without anthropogenic activity is viewed as an important option for the persistence of biodiversity and for allowing natural ecological and evolutionary processes to continue. At present, only 8.1% (449 178 km2) of Canada’s boreal forest is under some form of permanent protection. Though there is no consensus on what amount should be conserved, the current level of protection does not meet the generally regarded 12% minimum (Brundtland, 1987). Furthermore, existing protected areas are slightly biased to low productive areas, while 2  protected areas located at higher productivities are facing greater threats (Andrew et al., 2011a). Recognition of the Canadian boreal forest’s conservation shortfall and its ecological importance as well as its high conservation potential has resulted in a number of initiatives proposing the need for, or options to, expand the existing reserve networks in the region.   The present challenge is to design an expanded network of protected areas that complements those that already exist to ensure the long-term persistence of biodiversity (e.g., species richness) while at the same time addressing key socio-political and economic considerations. However, it is not possible to design meaningful protected areas without relevant spatial data that identifies the locations of strategic biodiversity areas. This is a particular problem for Canada’s boreal forest where its size and remoteness make it difficult to apply expensive conventional field-based techniques to acquire biodiversity data. However, remote sensing methods and imagery from spaceborne/airborne platforms offer a viable, cost-effective means of directly and indirectly characterizing aspects of biodiversity over large areas in a way that is explicit, repeatable and multi-scale (Kerr and Ostrovsky, 2003; Turner, 2003).   There are two main approaches used to inform the decision making process when addressing the conservation planning problem of prioritizing areas for conservation: expert opinion and the quantitative approach of spatial conservation prioritization (SCP) (Ferrier and Wintle, 2009). Expert driven conservation planning, for example, could include a representative reserve network based on expert derived ecozones. Traditionally, the Canadian landscape has been partitioned and described as a generalized, and nested levels of distinct areas (i.e. ecozones, ecoprovinces, ecoregions ) based on a number of similar landscape characteristics 3  (e.g., soil, wildlife, geological ) from a number of sources (e.g., maps and scientific reports) (ESWG, 1995). SCP is a contemporary branch of conservation biology that involves using spatial analysis of quantitative data to provide spatial information about conservation priorities. Here spatial analysis refers to the use of spatially oriented numerical models or statistical modeling methods that incorporate decision making theory to derive a suite of optimized spatial solutions from amongst the millions of potential alternatives (Ferrier and Wintle, 2009). A key advantage of the SCP approach is that it is unambiguous, efficient and repeatable, and thus its results are regarded as more scientifically credible than those derived from expert knowledge alone (Margules and Pressey, 2000 and Knight et al. 2006). In essence, it can provide a scientific and mathematical basis for selecting those protected areas or other conservation actions that  meet biodiversity conservation targets.    Fundamental to systematic conservation planning is the identification of sites that collectively represent the overall biodiversity of the region of interest (Margules and Pressey 2000; Ferrier et al. 2002; Trakhtenbrot; Kadmon, 2006). A more contemporary approach uses remote sensing derived biodiversity indicators to quantitatively group areas of like environmental characteristics. These environmentally distinct groups or clusters can be used to form a starting point from which to identify sets of priority areas for conservation that are representative of different environmental conditions. However, determining the amount and spatial location of areas to protect within these distinct clusters is not trivial, and can vary greatly depending on the available data (e.g., supplementary species data ), goals being considered (i.e., type of conservation actions) and the level of engagement expected (e.g., optimal solution versus scenario evaluation).  4   Despite the many recent advances in the field, conventional SCP approaches still remain inherently static and do not take into account system dynamics in the reserve designs (Possingham et al., 2009). For the expansive and largely intact Canadian Boreal, whose vegetation species and structure is strongly influenced by large natural disturbances such as stand replacing fire, this represents a particular problem. Accordingly, conservation targets set within a static reserve network may not be preserved through time due to the changes brought about by system dynamics and anticipated climate impacts. In support of improving protected area effectiveness, whereby conservation targets are likely to persist, the use of archival remote sensing data to examine and evaluate reserve designs represents a critical area of new boreal conservation research.  Understanding how remote sensing can contribute to conservation planning also represents an important avenue for new research. Remote sensing of environmental conditions and land cover is increasingly viewed as a key approach for assessing biodiversity, particularly for conservation (Kerr and Ostrovsky, 2003; Turner et al., 2003; Pettorellie et al., 2005; Bunchanan et al., 2008). There have also been a number of studies that have used medium spatial resolution (≥30 m) remote sensing indicator(s), typically just a single indicator, to successfully measure various characteristics of biodiversity within certain geographic regions of the Canadian boreal forest (e.g., Goetz et al., 2005; Wulder et al., 2008a; Zhang et al., 2009; Temini et al., 2010). These studies focus primarily on key drivers of diversity such as primary productivity, habitat structure (e.g., topography), and land cover. However, to more 5  effectively assess boreal biodiversity, a review and development of additional biodiversity indicators pertinent to northern ecoregions (e.g., snow cover and disturbance) is warranted.  The main research objective of this dissertation is to advance the use of remote sensing and other geospatial techniques for large-area, multi-jurisdictional biodiversity conservation. This research, which focuses on the entire Canadian boreal forest, contributed to progress in each of five research questions:   (1) Which remote sensing biodiversity indicators are useful for assessing boreal biodiversity? (2) How can these biodiversity indicators be used to provide a spatially contiguous classification of the Canadian boreal forest based on environmental similarity?  (3) How can boreal environmental domains and spatial datasets be used to identify areas of high conservation priority?  (4) How can long time-series Earth observation datasets be used to evaluate and improve the efficacy of boreal reserves? (5) How can incorporating threat probabilities based on future productivity variability and return-on-investment principles be used to constructively inform on-going boreal conservation efforts?  Chapter 2 provides a detailed description of the general study area and key data sources that constitute the basis of this thesis. Following this description, Questions 1 and 2 are addressed in Chapter 3, while questions 3, 4, and 5 are the focus of Chapters 4, 5, and 6 respectively. 6  Specifically, Chapter 3 investigates a suite of spatially explicit and remote sensing derived biodiversity indicators (e.g., vegetation production, topography, seasonality) to characterize boreal biodiversity and demonstrates the utility of these indicators to classify the boreal forest into environmentally unique regions (i.e., environmental domains). Chapter 4 integrates these environmental domains with SCP tools in conjunction with other highly relevant boreal biodiversity data (e.g., species-at-risk) to locate areas which have high conservation value. Chapter 5 presents an approach for improving reserve design efficacy by evaluating how reserve design considerations contribute to the long-term protection value of new candidate reserves. Chapter 6 examines the use of threat probabilities based on future productivity variability to assess park performances under different climate scenarios. Lastly, Chapter 7 synthesizes the major findings of the previous chapters and discusses potential applications and recommendations for future research directions.   The interconnection between the chapters and the integration of key data sources are illustrated in Figure 1.1.   7   Figure 1.1 Organization of thesis chapters (orange) and key data sources (grey)  The remainder of the introduction includes a review of remote sensing for biodiversity assessment, and  provides the conservation context for integrating spatial datasets with SCP approaches for conservation planning within the dynamic, and relatively intact Canadian boreal forest. 1.2. Remote sensing of biodiversity  Biodiversity is a broad concept that can encompass many facets (e.g., number of genes, traits, abundances, distributions, community structure, and ecosystems). That being said, this research defines biodiversity according to Turner et al. (2003) who state “species, certain species characteristics (e.g., distribution and quantity per unit area) and, in the more general sense, as ecological communities and species assemblages” (Turner et al., 2003). There are two general approaches used by remote sensing to assess biodiversity: direct and indirect 8  (Turner et al., 2003). Land-cover maps are examples of direct diversity indicators, and they are used to assess features of biodiversity such as species composition or abundance (Turner, 2003). Alternatively, indirect diversity indicators measure those physical environmental variables that affect features of biodiversity and typically include climate, topography, vegetation production, and disturbance metrics (Turner et al., 2003).  The effective and economic use of remote sensing for assessing biodiversity is closely tied to the spatial resolution of the sensor being used. Here spatial resolution refers to the grain size or the finest unit (i.e., pixel) that a remote sensor can define (Turner et al., 2001; Fassnacht et al., 2006). In general, local level studies typically use more detailed, but expensive, high-spatial resolution (i.e., <5 m) remotely sensed data (e.g., Quickbird, aerial photos, and IKONOS data). Medium spatial resolution (i.e., 10-100 m) data from sensors such as TM/ETM+, SPOT, ASTER are used primarily for regional level studies. Lastly, coarse spatial resolution data (e.g., MODIS and AVHRR) with spatial resolutions of 100m and greater used for regional and global level studies. Due to the large spatial extent of Canada’s boreal forest, the focus will be on coarse spatial resolution sensors (≥100 m) based on amenability and appropriateness for a national or continental assessment of biodiversity and the availability of satellite imagery.  In a review of remote sensing for national biodiversity monitoring in Canada, Duro et al. (2007) recommended four key indicators: vegetation productivity, disturbance, land cover, and the physical environment (e.g., topography). Using these broad categories as the basis, it is beneficial to judiciously consider additional indicators that may be important regionally. 9  For instance, in Canada’s boreal forest, seasonal snow cover influences climate and hydrological processes (Pulliainen, 2006) and can be used to gauge food resources and animal habitat throughout the year (Coops et al., 2009a). Additionally, the boreal peatlands warrant special conservation attention as they host a major proportion of Canada’s plant biodiversity and provide habitat to many animals, some of which are restricted to those environments (Warner and Asada, 2006). As wetlands are not always well identified by standard land cover remote-sensed products, there is a need for specialized indicators of their presence. In other words, a wetland specific land cover map is required to adequately identify them.  Adapting from Duro et al. (2007), this research identifies and develops a set of direct and indirect environmental variables designed to map biodiversity in the Canadian boreal forest. Due to the large spatial extent of the boreal, this research focused on freely available, ecologically meaningful products derived from coarse spatial resolution sensors (> 100 m resolution, but primarily 1 km), as appropriate for national or continental assessments. The “primary” coarse and medium spatial resolution remotely sensed data explored in this thesis measure (1) land cover, (2) topography, (3) vegetation productivity, (4) disturbance, (5) fragmentation and (6) snow cover. The following subsections 1.2.1 to 1.2.6 provide background on these respective measures.  1.2.1. Land cover (direct) Land-cover maps derived from remote sensing differentiate broad plant communities. This mapping provides an opportunity for directly assessing plant biodiversity, or indirectly 10  through species distribution models having land cover as a covariate. Recently, numerous broadscale studies have successfully employed remote sensing to provide land-cover type information on Canada and the boreal forest in general (Table 1.1). For instance, a Canada-wide landcover map of forested areas with 23 landcover classes was produced using Landsat-7 ETM+ data (25 m) meeting a target overall accuracy of 85% in tested areas (Wulder et al., 2007, 2008a). Steyaert et al. (1997) presented a land cover classification approach that used multitemporal 1 km NOAA-11 AVHRR Normalized Difference Vegetation Index (NDVI) composites (1 km) to classify a 508,199 km2 region of the central boreal ecosystem for the purpose of serving as a basis for regional land cover mapping, fire disturbance assessment and land cover. In a Canada-wide study, Cihlar et al. (1997) applied two different classification approaches to 1 km NOAA-11 AVHRR composites to classify the region into 17 land cover classes according to the IGBP classification scheme (Belward, 1996) with an overall classification accuracy ranging from 57% (all pixel) to 89% (> 80% pixel purity).  Direct mapping of vegetation species assemblages has also been used to derive information related to regional biodiversity across the Canadian boreal. For example, forest species composition of the western subarctic treeline was mapped using 1 km AVHRR data with a canopy reflectance model, an approach that uses reflectance and transmittance values of different canopy components as well as known structural relationships to estimate species composition and predict change over time (Olthof and Pouliot, 2010). The results revealed a high spatial correlation (r 0.85 to 0.98) between the five treeline zones and recently validated MODIS vegetation data. Similarly, a 250 m categorical land-cover product (Latifovic et al., 2008) was an important covariate in models of boreal-wide distributions of 103 songbird species (Cumming et al., 2014). In general, land-cover maps play an important role in 11  conservation planning and reserve design by providing both direct and indirect indices of biodiversity. As indicators of habitat type they can be directly used in setting conservation targets for specific habitat types. 1.2.2. Topography (indirect)  Topography generates environmental gradients that influence regional biodiversity by shaping species distribution patterns and productivity and influencing natural disturbances (Swanson et al., 1998; Dorner et al., 2002). Numerous topographic indicators and variables derived from elevation data (Table 1.1) are inputs to ecological mapping (MacMillan et al., 2004). For example, Temini et al. (2010) successfully used a DEM in conjunction with microwave and vegetation data to model soil wetness for the Peace Athabasca Delta in central Canada. Their findings showed a 70% correlation between the soil wetness map and observed precipitation data. Anderson and Ferree (2010) found that topographic data was a useful predictor of species diversity across the Maritime Provinces and southeastern Quebec. Specifically, range in elevation, paired with geological variables and latitudinal position, was able to predict species diversity with high certainty (adjusted R2 of 0.94).  12  Table 1.1: Examples of studies which have used remotely sensed or other geospatial data to map or model environmental indicators related to biodiversity in the Canadian boreal forest. (adapted from Powers et al., 2013a) Aspect of relevance to biodiversity  Region Study reference Summary of Methodology Remotely sensed/mapped data used Land Cover: General Land Cover Forested ecozones of Canada Wulder et al. (2008a) Unsupervised hyperclustering approach used to classify Canada into 23 unique land cover classes Landsat ETM+ (Over 480 scenes) Forest Cover  Western subarctic treeline Olthof and Pouliot (2010) GeoSail canopy reflectance model used to map tree species composition and estimate/predict change over time AVHRR spectral and NDVI  Geology South-eastern Canada  Anderson and Ferree (2010) Geological variables, elevation data and latitude used to predict species diversity for a region of the boreal and hemi-boreal zones  Species inventory, geologic maps, USGS DEM, climate data Fragmentation:  Pattern Indices Forested ecozones of Canada Wulder et al. (2008b) Landscape pattern metrics were calculated over several spatial scales to represent fragmentation of forest patches in Canada Land cover classification of Landsat ETM+ Fragmentation (Intact Patches) Canada (forested) GFW 2010, Lee et al. (2006) A visual interpretation and GIS buffering approach was used to identify unbroken forested regions larger than 1000 hectares  Existing GIS data, Landsat TM and ETM+, ASTER imagery (reference data) Topography:     Elevation Peace Athabasca Delta  Temini et al. (2010)  A topography-based soil wetness index was modeled from SRTM DEM data, LAI data from MODIS and passive microwave data AMSR-E Microwave data, LAI MODIS, DEM (SRTM) Wetland Classes Canada Li and Chen (2005)  Rule-based (object-based, decision-tree classification) mapping method used to classify wetlands in sites in eastern Canada. Landsat ETM+, Radarsat SAR, DEM and derivatives Landforms  South-western Yukon Giles (1998) Discriminant analysis of slope profiles and SPOT imagery used to map slope units or geomorphological landform classes SPOT imagery and DEM      13  Table 1.1: (continued…) Aspect of relevance to biodiversity  Region Study reference Summary of Methodology Remotely sensed/mapped data used Climate: Climate Northern boreal range Zhang et al. (2009) Patterns of change in water balance derived from climate data, and modeled from an algorithm based on evapotranspiration AVHRR NDVI, MODIS land cover, climate data Snow Cover Northern Hemisphere Dye (2002) Temporal variability in snow cover variables were analyzed using regression techniques  AVHRR snow cover charts (NOAA) Productivity: Vegetative productivity  North American boreal Goetz et al. (2005) Autoregressive modeling used to analyze temporal trends in the spatial distributions of photosynthetic activity over 22 years NDVI data (AVHRR), land cover, temperature, wildfire extent maps Forest Age Southern Ontario Zhang et al. (2004) Forest stand age was modeled using a short-wave vegetation index and change indicator based on NDVI data over time  SPOT (Vegetation data), AVHRR (NOAA), Landsat TM Dynamic habitat index Canada Coops et al. (2008) Productivity indices used to model habitat diversity over time and cluster analysis used to map regional biodiversity fPAR MODIS data Disturbance:     Fire events Canadian boreal ecozones Burton et al. (2008) A hierarchical analysis was used to map the spatial variability of fire events and severity across the boreal ecozones of Canada  Canadian Large Fire Database, AVHRR NDVI data, Landsat TM, land cover classification Anthropogenic and natural disturbances West-central Alberta Rocky Mts. Linke et al. (2009) Dynamic disturbance maps created using spatial overlay analysis of object-based analysis and manual digitizing methods  Landsat TM and ETM+ (EWDI), DEM, existing GIS data and maps      14  1.2.3. Vegetation productivity (indirect) Vegetation productivity, measured as rate of gross or net carbon fixation, is directly linked to biodiversity. More productive areas provide more available energy and energy pathways than less productive areas, hence supporting larger populations of species and higher species diversity (Walker et al., 1992). Vegetation productivity derived from remote sensing can be considered a predictor of species richness and can identify regional biodiversity ‘hotspots’ (Rocchini et al., 2007). A suite of remotely sensed indicators related to vegetation productivity have been developed and applied to the boreal forest (Table 1.1). NDVI derived from AVHRR data was used by Goward et al. (1985) to analyse seasonal vegetation patterns for North America. More recently, Coops et al. (2008) used productivity variables from MODIS data, including measures of the fraction of available incoming energy used by vegetation (fPAR) as the basis of a Dynamic Habitat Index (DHI). DHI, combined with indicators of topography and land cover, effectively modeled overall habitat diversity and regional biodiversity across Canada. 1.2.4. Disturbance (indirect) The main natural disturbances in the Canadian boreal forest are wildfire and insect outbreaks. Flooding, storms, pathogens and animal activity like that of moose (Alces alces) or beaver (Castor canadensis) (Engelmark, 1999; Esseen et al., 1997; Kuuluvainen, 2002) are also widespread, but of secondary and local importance in influencing the structure and composition of boreal ecosystems. Each process has its characteristic duration, frequency, intensity and spatial extent (Coops et al., 2007). Disturbances affect biodiversity through altering landscape 15  structure and ecosystem function (Turetsky and Louis, 2006), which can also change habitat characteristics and impact species’ population dynamics (Kuuluvainen, 2002). Given the long duration and diffuse spatial extent of some disturbances such as gradual infestations of insect defoliators, accurate remotely sensed indicators can be difficult to develop (Coops et al., 2007). Stand-replacing disturbances, such as those caused by fire or harvesting, are more readily identified and can be successfully monitored using remote sensing technology. A Canada wide study by Li et al. (2000), developed an algorithm for AVHRR satellite data (1 km spatial resolution) to monitor fire within the boreal. In addition to providing a consistent nationwide fire database, the study also demonstrated that it was possible to use a remote sensing approach to monitor fires in near real time. Mildrexler et al. (2009) evaluated disturbance in woody ecosystems across North America with an improved disturbance detection algorithm using MODIS satellite data for the years 2002–2006. Their method could detect the location and extent of wildfire, identify areas with downed trees resulting from large hurricanes and identify large logging disturbances. 1.2.5. Fragmentation (indirect) Fragmentation, often defined as the breaking apart of contiguous areas of habitat (Fleishman and MacNally, 2007), represents another important aspect of landscape spatial pattern that can affect biodiversity (Fahrig, 2003). Habitats can become fragmented through disturbances brought about by both natural processes, such as fire and insect outbreaks, and anthropogenic activity, such as logging or road construction (Linke et al., 2007). Fragmentation can be viewed as a state indicating the juxtaposition of land-cover conditions over a land base, or can inform on process when considered over time. As with disturbance, remote sensing technologies have been 16  employed to monitor fragmentation across Canada (Soverel et al., 2010; Wulder et al., 2008b, 2009).   To date, the focus of most anthropogenic disturbance studies in boreal forest areas has been on quantifying the impact of fragmentation (e.g., Hobson and Bayne, 2000; Meddens et al., 2008; Linke et al., 2005) or land cover variation (e.g.,  Gillanders et al., 2008; Potapov et al., 2011; Schroeder et al., 2011) using Landsat (TM and ETM+) imagery. For example, in the context of Alberta’s boreal forest, the magnitude and diffuse nature of anthropogenic activities (e.g., oil and gas) have greatly contributed to contemporary landscape changes in the region, and have raised a number of concerns regarding its impacts on forest ecosystem structure (e.g., fragmentation) and condition (Schneider and Dyer, 2009). Specifically, cumulative impacts associated with the expansion/intensification of oil and gas related activities (e.g., long linear seismic lines used for oil and gas exploration, roads, and the operation of large equipment) can increase regional environmental damage by fragmenting the landscape and compacting the soil or damaging the vegetative mat (Severson-Baker, 2004). Using Alberta’s oil sands region as a case study, Powers et al. (2015) demonstrated the usefulness of remote sensing using SPOT 5 imagery for identifying these oil and gas related disturbances, and highlight its potential for allowing continued monitoring (e.g., fragmentation impacts) and/or aiding conservation planning.  1.2.6. Snow cover (indirect) Biodiversity is highly influenced by climate (Gaston, 2000).  Specifically, climatic factors act as regional and global determinants of biodiversity by limiting species establishment and occurrence, and by regulating vegetation seasonality (Sarr et al., 2005). Climatic factors are 17  highly predictive of spatial patterns in species richness at higher latitudes (Currie and Paquin, 1987). Compared to other types of variables, climate factors are, in aggregate, the most important determinants of bird species richness (Hawkins et al., 2003) and bird species’ distributions (Cumming et al., 2014) in the boreal forest. Of particular relevance to seasonally snow-covered areas like the Canadian boreal are climatic variables pertaining to snow, ice cover or albedo. In the case of mammals, snow can sometimes play an essential role in seasonal habitat requirements. For example, Copeland et al. (2010) used MODIS snow-cover composites in conjunction with interpolated temperature data to develop a more accurate map of the wolverine’s (Gulo gulo) current circumboreal range. Thus, by including snow cover, we can potentially describe the habitat of this species that, though widely distributed, occurs at very low densities and is of conservation concern. 1.3. Regionalization  Ensuring that new reserves are representative of the overall biodiversity of the region is a fundamental component of systematic conservation planning and biodiversity management (Margules and Pressey, 2000). To meet this requirement, remotely derived biodiversity indicators are often employed to classify regions into areas with similar biodiversity characteristics, especially if there is a shortage of available species distribution data (Trakhtenbrot and Kadmon, 2005). These groupings, typically labeled as regionalizations, ecoregions, environmental domains or clusters are associated with a range of different environmental conditions, which, in theory, should be representative of species diversity (Mackey et al., 1988; Belbin, 1993; Belbin, 1995; Trakhtenbrot and Kadmon, 2005). As previously mentioned, the Canadian landscape has traditionally been defined as a series of nested 18  groupings (i.e., ecozones, ecoprovinces, and ecoregions) based on landscape similarities (e.g., soil, wildlife, and geologic) derived from varied map and scientific (reports) sources (ESWG, 1995). However, while such stratifications are useful, their reliance on expert judgment and the subjective human eye introduces a certain level of bias and makes it more challenging to compare areas and situations in a reproducible way. 1.4. Spatial conservation prioritization (SCP) Increased climate variability via anthropogenic induced climate change (global) and resource extraction (local) has the potential to greatly impact biodiversity. In recognition of these potential threats, many international and national conservation initiatives stress the need for immediate action to maintain (or improve upon) current levels of biodiversity protection and stem further loss (Klien, 2010). While it is important to be aware that there is a broad range of planning processes for addressing planning decisions on how land is used or managed to sustain biodiversity persistence, reserve systems, such as protected area networks, are the cornerstone of most conservation strategies (Soulé, 1991). As such, the protection of biodiversity is typically tied to the limited financial resources available for establishing these areas; thus, it is prudent to effectively and efficiently prioritize areas that are potential candidates for conservation.  Biodiversity conservation planning tools such as quantitative spatial conservation prioritization (SCP) are intended to provide decision support for various conservation efforts. For example, SCP can be used to (spatially) inform managers, stake holders and conservation planners when addressing conservation planning decisions that involve a spatial choice (e.g., where to locate a national park, stewardship efforts, biosphere reserve). SCP uses quantitative techniques, 19  algorithms (e.g., simulated annealing) and spatially explicit data (e.g., species distributions, biodiversity indicators) to determine where to efficiently and effectively allocate conservation effort and resources (Moilanen et al., 2009a).  At present, there is a suite of reliable and publically available SCP software tools developed to handle many kinds of conservation planning problems (Moilanen et al., 2009a). Examples of popular SCP software include the following: Marxan (Posingham et al., 2000), Zonation (Moilane et al., 2009b), and C-Plan (Pressey et al., 2009). Of these, Marxan is the most widely used SCP software for terrestrial and marine conservation planning. The freely available Marxan software was originally developed to solve a minimum-set reserve design problem (Cocks and Baird, 1989 and Ball et al., 2009) via identifying areas that meet biodiversity targets at an efficient cost and low boundary length (i.e., high compactness). The boundary length modifier (BLM) can be used to alter the structural connectivity of the reserve system. In this respect, a larger BLM contributes to a greater importance being placed on the reserve system’s compactness (i.e., lower the boundary length between candidate priority areas) than cost efficiency. Marxan’s optimization algorithm is called simulating annealing, and it has the added advantage of being able to provide quick and good solutions to a wide range of different sized problems. The mathematical problem Marxan evaluates (see Table 1.2 for terms) can be defined as (Ball et al., 2009): 𝑚𝑚𝑚𝑚𝑚𝑚𝑚𝑚�𝑥𝑖𝑐𝑖𝑁𝑠𝑖+ 𝑏��𝑥𝑖𝑁𝑠ℎ(1− 𝑥ℎ)𝑐𝑐𝑖ℎ𝑁𝑠𝑖 conditional on meeting all target objectives 20  �𝑥𝑖𝑟𝑖𝑖 ≥ 𝑇𝑖  ∀ 𝑗𝑁𝑓𝑖 and xi  is either 0 (i.e., not protected/selected) or 1 (i.e., protected/selected) 𝑥𝑖 ∈ {0,1} ∀ 𝑚 Any missing or unmet targets of features in the final reserve design configuration will result in an additional penalty. This ensures that all targets are met.  Table 1.2: Marxan equation terms and definitions (Ball et al., 2009) Term Description rij The amount of the target feature in site i ci Cost of site i Ns The number of sites Tj The target objective for target feature j B Boundary multiplier which affects the cost of the reserve system relative to the level of compactness cvih The cost of the boundary shared between planning units i and h  1.5. SCP and incorporating system dynamics and anticipated climate variability  SCP approaches have only just started considering the dynamic nature of ecosystems and are inherently static. Given the Canadian boreal forest’s dynamic landscape and anticipated climate impacts (Price et al., 2013), future boreal conservation assessment and planning targets based on static representation (fixed areas or features), which do not account for this dynamism, may not represent the best use of conservation resources (Andrew et al., 2014). The long-term protection value of new candidate protected areas rests on their ability to maintain biodiversity over time; thus, it is important to better understand how altered landscape properties impact these protected 21  areas and what protected area design criteria (e.g., location, compactness, threat, size ) improve long-term conservation effectiveness. For example, reserve size represents an important reserve design criteria for incorporating natural disturbances (Pickett and Thompson, 1978; Baker, 1992), which, over time, can potentially alter the landscape structure and function of reserves. Simply put, accounting for system dynamics and climate change impacts in a well-designed protected area network will, in the long run, increase the preservation, or at least reduce the endangerment, of boreal species and habitat diversity in current and anticipated future conditions (e.g., altered patterns of land use or shifts in biological distributions due to changes in climate variability and disturbance regimes).   While a few studies have attempted to incorporate landscape dynamics with SCP approaches for assessing boreal conservation effectiveness and planning (e.g., Leroux et al., 2007b, 2007c; Rayfield, 2008), the geographic extents evaluated (within the boreal) are constrained to small regions; thus, it is difficult to generalize to the boreal as a whole. As a result, accounting for system dynamics and climate variability in SCP still represents an area requiring active research. Planning frameworks using SCP approaches can play a vital role in determining how to best expand the level of protection (e.g., expanded/new protected areas) within the Canadian boreal forest, but only by accounting for its dynamism.  A recent overview by Leroux and Rayfield et al. (2013) has identified four emerging methods that are likely best suited for handling natural dynamics in conservation planning of large, relatively intact regions like the Canadian boreal forest: (1) spatial catalysts and site-selection tools; (2) probability theory and site-selection tools; (3) spatial simulation models; and (4) spatial simulation models and site-selection tools. A summary of these methods is provided in Figure 1.2 and Table 1.3. Of these methods, Leroux 22  and Rayfield et al., 2013 conclude that a combination of probabilistic models and SCP tools (e.g., Marxan or Benchmark Builder) has significant potential for accounting for the Canadian boreal’s dynamism, mirroring the approach developed in this dissertation.        23  Table 1.3: Advantage and disadvantage overview of four methods for accounting for natural disturbance dynamics within conservation planning in intact areas. A description of these methods is provided in Figure 1.2 (adapted from Leroux and Rayfield et al., 2013) Method  Study references Advantages Disadvantages Spatial catalysts + site-selection Cowling et al., (2003); Klein et al., 2009; Anderson and Ferree, (2010) Minimal data requirements relative to other   methods Designed for use with user-friendly site-selection   tools Useful for processes with clear spatial footprints   (e.g., hydrological flow) Maps surrogates of processes as opposed to actual    process dynamics Requires validation Assumes processes have clear spatial patterns and    simple temporal dynamics     Probabilistic models + site-selection tools Game et al., (2008); Drechsler et al., (2009); Lourival et al., (2011) Modest data requirements relative to other   methods Provides an analytical solution to simple   conservation planning problems Designed for use with user-friendly site-selection   tools Enables a priori consideration of ecological   Processes Simple probabilistic models may not be able to    capture complex process dynamics Requires familiarity with probabilistic models such    as Markov chain models  Not fully integrated with all site-selection tools –    SCP tools with this capability are not yet publicly     available.     Spatial simulation models Leroux et al., (2007a); Rayfield et al., (2008); Ban et al., (2012) Explicit modelling of stochasticity Potential for detailed estimates of process size and   frequency Potential to explicitly model interacting processes Potential to analyse transient dynamics High data requirements relative to other methods Computationally demanding for large extents Not well integrated with other conservation    planning tools Software is not user-friendly and may require some    basic programming knowledge     Spatial simulation models + site-selection tools Leroux et al., (2007a); Anderson, (2009); Saucier, (2011) Same advantages as spatial simulation models Allows for adaptive site-selection through time Same disadvantages as spatial simulation models Not fully integrated with site-selection tools    24   Figure 1.2: Overview of four methods for accounting for natural disturbance dynamics within conservation planning in intact areas. (1) This method optimizes solutions for areas represented by spatial catalysts of ecological processes (e.g., river network and intactness). (2) This method uses probabilistic models to assess the likelihood of a planning unit (i.e., individual grid or site) containing a target or feature (e.g., a given serial stage) over a specified timeframe. In this example, solutions are optimized to be representative of each serial stage over the specified timeframe. Darker shades indicate a higher probability that a planning unit contains an occurrence of a serial stage. (3) This method simulates natural disturbance over time to provide valuable information (e.g., the minimum reserve size for maintaining conservation features) for setting robust targets that account for ecological process (e.g., stand replacing wildfire). (4) This method first provides optimal solutions like those mentioned in (1; darker are better sites), then evaluates the efficacy of those solutions using (3) over a given timeframe. Solutions can then be modified based on any target deficiencies highlighted by the evaluation. (adapted from Leroux and Rayfield, 2013)    25  1.6. SCP and system design criteria and considerations  SCP approaches are commonly used in conservation planning for the purpose of identifying areas of high conservation value that should be targeted for protection (i.e., strict reserve). Conservation actions for strict protection can include many levels of complexity with respect to system design criteria. To improve effectiveness and add realism, spatial planning frameworks have incorporated criteria such as costs associated with conservation action (e.g., costs of land acquisition; Ando et al., 1998; Cameron et al., 2006; Moilanen and Arponen, 2011; Schneider et al., 2011), reserve connectivity (Moilanen et al. 2005) or compactness (Ball et al. 2009), and species relationships (Rayfield et al. 2009). For example, accounting for costs, if care is taken to reduce elements of uncertainty, can facilitate a better, more efficient use of conservation funds (Carwardine et al., 2010). Likewise, a highly connected/compact reserve system is an important reserve design criteria since it has the potential to increase the likelihood of persistence for some species (Davies et al., 2000; Margules and Pressey, 2000) and are often easier to implement (i.e., purchase land or manage) than highly fragmented/unconnected systems. However, increasing the system’s compactness also increases the cost of the respective conservation action (e.g., land acquisition). Imposing a size restriction or criteria to generate large reserves will reduce the reserve design flexibility to include, for example, less desirable areas (e.g., high economic cost) that would otherwise be avoided in a series of smaller reserves. In general, an efficient (i.e., cost effective) reserve system is important for achieving conservation goals, and necessitates the balancing of trade-offs between biodiversity gain and economic cost. These trade-offs are typically evaluated by comparing different reserve spatial configurations (i.e., conservation 26  scenarios) and their system respective efficiency and target objective performance (e.g., Carvalho et al., 2011; Game et al., 2011; Schneider et al., 2011).   In addition to these system design criteria, how conservation targets are specified represents a critical reserve design consideration. Setting conservation targets are often a criticized component of SCP, where, for instance, targets are considered subjective, arbitrary, set too low, and/or fail to address biodiversity persistence (Carwardine et al., 2009). While it is inevitable that some subjectivity will be introduced when formulating targets, it is also possible to improve their objectivity by accounting for factors that may impact biodiversity persistence and conservation goals. For example, targets can improve objectivity by accounting for (i) natural rarity, where more emphases is placed on species at risk, and (ii) compositional distinctiveness, where a representative amount of distinct biodiversity-indicative environmental domains are protected.       27  2. Study area and data This chapter describes the key datasets commonly used within the thesis. Each section explains the specific details on how the data was used and processed.   2.1. Study area  The study area (Figure 2.1) is the Canadian boreal forest (~5.37 million km2) as described by Brandt (2009) excluding the southern transitional hemiboreal subzone (includes much of British Columbia), which is considered temperate in North America and not formally recognized as boreal (Brandt, 2009). Situated primarily in the northern latitudes, the region is principally forested (~58%) and dominated by cold tolerant forest types within the genera Larix, Abies, Picea or Pinus as well as Betula and Populus (Brandt, 2009). Water features such as lakes and rivers, as well as wetlands are also common throughout the boreal forest (Wulder et al., 2008). Stand replacing fire and insect infestation are the dominant natural disturbances on the landscape (Kurz et al., 1992; Fleming et al., 1998).    28   Figure 2.1: Study area (dark grey) encompassing the entire Canadian boreal forest as defined by Brant (2009) 2.1.1. Land cover The MODIS Land cover product includes five categorical maps derived from observations collected over a period of a year at a 1 km spatial resolution (USGS, 2011). I selected the University of Maryland (UMD) classification which was based on data from 2004. In total there were 14 classes, which include five forest classes (e.g., evergreen needleleaf forest, and evergreen broadleaf forest), and two shrubland classes.   The Ducks Unlimited Canada hybrid wetland layer (HWL) was also used to identify the locations of water and wetlands. The HWL is a multi-source product that classifies Canada’s 29  landscape, minus the arctic, into the three general classes of water, wetland and upland (DUC, 2010). Two freely available and national datasets were used to derive the HWL: land cover (Wulder et al., 2008a; GC/AAFC, 2009) and CanVec (GC/NRCan/ESS/MIB/CTI-S, 2011). The land cover component was generated from Landsat imagery and covers Canada’s forested and agricultural areas and the Northern Territories. Forest cover data was produced by the Earth Observation for Sustainable Development of forests (EOSD) project, which classified the forested areas of Canada (23 land cover classes) using over 480 Landsat-7 Enhanced Thematic Mapper Plus (ETM+) images and covering over 80% of Canada (Wulder et al., 2008a). The agricultural coverage was created by the National Land and Water Information Service (NLWIS) of Agriculture and Agri-Food Canada (AAFC). Land cover information for the Northern Territories was provided by the Canadian Center for Remote Sensing (CCRS). CanVec is a topographic product in vector format produced and maintained by Natural Resources Canada. To create the HWL, the EOSD rasters (land cover) and rasterized CanVec datasets were processed to a common raster format (Lambert Conformal Conic projection) and combined into a single hybrid layer at a 25 m spatial resolution (DUC, 2010) using the combine tool (available in the Spatial Analyst Toolbox for ArcGIS 9.3.1) . The combined raster was then classified into the final HWL class values using a hierarchical classification scheme (DUC, 2010, Figure 2.2). I exported the HWL wetland class to its own raster layer and resampled it from a 25 m to a 1 km spatial resolution, where each cell represents the percentage of water and wetland.      30  CanVec = 1 OREOSD = 20(Open water classes)NoYesHWL = 1(Water)CanVec = 2 OREOSD = {40, 80-83}(Wetland classes)NoYesHWL = 2(Wetland) Derive class value from EOSD, such that x ≥ 3HWL = x(Upland classes) Figure 2.2: HWL class decision tree. (adapted from DUC, 2010)      31  2.1.2. Topography (ruggedness) I used two sources of remotely sensed topographic data: (i) the 90 m NASA Shuttle Radar Topographic Mission (SRTM) and (ii) the USGS Global 30-Arc Second Elevation Data Set (GTOPO30). The SRTM was initiated in 2000 by NASA and the United States National Geospatial Intelligence Agency to acquire digital elevation data for ~80% of the earth’s surface between approximately ±60° latitude (Farr and Kobrick, 2000). The GTOPO30 dataset (~1 km spatial resolution) was utilized for that portion of the study area above 60° North latitude. The coefficient of variation (CV) of elevation (i.e., ruggedness) was used to better differentiate topography between different environments across the boreal. The CV was computed at a 1 km neighbourhood distance for each pixel in both elevation datasets (i.e., a 10 and 1 pixel neighbourhood distance for SRTM and GTOPO30 respectively). Here the neighbourhood was rectangle (centred at each pixel); thus, included overlapping blocks. The SRTM CV values were then resampled to a 1 km spatial resolution and combined with the GTOPO30 dataset. High CV values indicate areas with extreme changes in elevation, whereas low CV values represent areas with minimal elevation differences. 2.1.3. Vegetation productivity MODIS sensors provide a global coverage at a 1 km spatial resolution, and are subjected to processing by an expert science team that include advanced geo-location, atmospheric correction and cloud screening (Justice et al., 2002). Vegetation productivity was measured by the Dynamic Habitat Index (DHI) of Coops et al. (2008) using the MODIS monthly fPAR product. The DHI is comprised of three indicators of vegetation dynamics derived from six years (2000-2005) of 32  monthly MODIS data: annual primary productivity, annual minimum cover, and seasonal greenness. Annual primary productivity was calculated by summing monthly fPAR observations over each of the six years of data (2000-2005). These six components were then averaged to create a long-term annual productivity indicator. Similarly, the annual minima of monthly fPAR observations were summed to calculate the annual minimum cover for each year. Areas that maintain some degree of vegetation over the year will have positive minimum cover values, while those that do not (e.g., areas predominately snow covered) will have values near or equal to zero (Coops et al., 2008; Coops et al., 2009b). These six annual minimum cover components were then averaged to produce a long-term annual minimum cover indicator. Seasonal greenness was computed by first dividing the annual standard deviation of monthly fPAR observations by the annual mean value to obtain an annual CV. They were then averaged over the 6 years of data. Areas with variable climate or limited productivity will have high seasonal greenness values, while those that are consistently productive (e.g., evergreen forests), or experience less extreme climate conditions will have low values (Coops et al., 2008; Coops et al., 2009b). Annual primary productivity can be interpreted as the summed greenness over a year, annual minimum cover as the landscape’s capacity to sustain populations over a year (Schwartz et al., 2006), and seasonal greenness as the combined assessment of climate, topography and land use (Coops et al., 2009b).  33  2.1.4. Snow cover MODIS snow cover products have been available since 2000, at different spatial and temporal resolutions (Hall and Riggs, 2007; Riggs et al., 2006). I used the global, 0.05° (~5 km) spatial resolution monthly Terra snow cover product MOD10CM (Hall et al., 2006). This product estimates a mean monthly snow cover extent by averaging the daily fractional snow cover (FSC) extents of the MODIS product MOD10C1 (Riggs et al., 2006). To assess seasonal snow cover across the study area, I computed two average monthly values for each cell for the years 2000 to 2010 during (i) the autumn (September to November) and (ii) spring (March to May) months. These two datasets were then resampled to a 1 km spatial resolution. 2.1.5. Disturbance and fragmentation  To characterize landscape condition and the imprint of disturbances, I used six metrics of landscape pattern and an anthropogenic disturbance index. The 1 km spatial resolution pattern metrics were calculated by Wulder et al. (2008b) using the 25 m spatial resolution EOSD cover product from the year 2000. I selected metrics of the total abundance and spatial configuration of forest patches within 1 km cells (Table 2.1). Based on a review of the literature, these selected metrics were found to be particularly well suited to characterizing fragmentation across large areas (e.g., Riitters et al., 2002; Kupfer, 2006).     34  Table 2.1: Landscape pattern metrics used to characterize cluster groupings (adapted from Wulder et al., 2008b; Soverel et al., 2010) Fragmentation  Description Edge density (m/ha.) The amount of forest edge (m/ha) in the analysis unit. Larger values indicate more edge habitat and more forest fragmentation (McGarigal et al., 1995; Li et al., 2005). Mean Patch Size (ha.) The average size of the forest patch within a 1 km cell. A smaller average forest patch size is considered indicative of a more fragmented forest (McGarigal et al., 2002). Number of Forest Patches The number of forest patches within 1 km cell. The more forest patches there are, the more fragmented the forest is considered to be (Turner et al., 2001; McGarigal et al., 2002) Proportion of Forested Patches (%) The proportion of all landscape patches that are forest; this metric links fragmentation with cover type (Wulder et al., 2009). Relative Area (%) Proportion of analysis unit that is occupied by forest (Turner et al., 2001). Standard Deviation of Patch Size (ha.) A measure of the absolute variation in patch size within 1 km. The mean patch size can obscure the presence of very large or very small patches (McGarigal and Marks, 1995; Cumming and Vervier, 2002).  The number of forest patches and the proportion of all patches that are forested, combined with the proportional area of forest in the landscape (the relative area metric) are, in combination, indices of relative forest fragmentation (Wulder et al., 2008b). For instance, if a landscape contains many distinct patches of which a high proportion is forested, then the forested area is divided into many patches separated by non-forest, indicative of forest fragmentation (Wulder et al., 2008b). Alternatively, a relatively low proportion of forest patches would imply that the landscape is fragmented with respect to non-forested habitats, but that the forested area it does contain is not necessarily fragmented forest. The six landscape metrics represent different types and spatial arrangement of forested and non-forested habitat patches in total and relative to each other. The number of forest patches, and the mean and standard deviation of forest patch size collectively describe the total area and the size distribution of forest patches within the landscape (Soverel et al., 2010). While not spatially explicit (McGarigal and Marks, 1995), edge density can also measure landscape fragmentation (Li et al., 2005).  35  To measure the degree of anthropogenic disturbance, I used the Global Forest Watch Canada’s Landsat (30 m spatial resolution) derived combined anthropogenic change mapping datasets (1990 to 2001) for areas within the provinces of Nova Scotia (Cheng and Lee, 2009), Saskatchewan and Manitoba (Stanojevic et al., 2006a), Ontario (Cheng and Lee, 2008), Québec (Stanojevic et al., 2006b) and British Columbia (Lee and Gysbers, 2008). It should be noted that these datasets do not cover the entire Canadian boreal, but are, to the best of our knowledge, the most complete dataset of its kind for the study area. These datasets identify and buffer human-caused forest disturbances such as road construction, agriculture, reservoirs and clearcuts. They were converted from their native polygon form to a 1 km spatial resolution raster. The anthropogenic index for each cell measures the percentage area impacted by human activity.  2.1.6. Species data To quantify species richness I used NatureServe’s digital distribution maps of mammals (Patterson et al., 2007) and birds (Ridgely et al., 2007) of the western hemisphere (version 3.0, available at http://www.natureserve.org/getData/animalData.jsp). These maps are primarily based on the US Defense Mapping Agency’s (DMA) Digital Chart of the world basemap (1:1,000,000 scale) and were derived from data sources dating from the years 1980-2000 (birds) and 1981-1999 (mammals). Mammal and bird species richness ranged from 0 to 16 and 0 to 91 respectively per 1 km2 cell. U.S. Geological Survey range maps of tree species in North America (http://esp.cr.usgs.gov/data/atlas/little/; Little, 1999), derived from source maps at 1:10,000,000 scale, were used to produce a tree species richness layer that ranged in value from 0 to 45. Lastly, a butterfly species richness layer (5 km spatial resolution) was created using the butterfly specimen and observation dataset provided by the Canadian Biodiversity Information Facility 36  (http://www.cbif.gc.ca/). The species maps were summed to produce a species richness layer for each taxon, which was then intersected with the study area grid. 2.2. Environmental domains and species-at-risk distribution data I used two suites of data to represent biodiversity: environmental domains and distributions for 16 at-risk species. Environmental domains (Figure 2.3a) were generated in a previous study (Powers et al., 2013a) by classifying the boreal forest into 15 domains based on productivity, seasonality (snow cover), and land cover similarity. Seasonal greenness (Coops et al., 2008), a vegetation productivity index, was the most important indicator for discriminating among the environmental domain groups. Spanning from east to west along latitudinal gradients, the 15 domains represent regions of environmental uniqueness with spatial and attribute detail that is appropriate for large area conservation planning (Coops et al., 2009). The five northernmost domains are dominated by open shrub vegetation and can be characterized as having high seasonality and low productivity environments. In contrast, the 8 southern domains have a relatively low seasonality and high productivity and are dominated by coniferous and mixed forest. The two central domains are mostly coniferous forest and open shrubland and experience relatively moderate productivity and seasonality.    Prioritizing reserve areas using only data on environmental domains increases the chance of missing or underrepresenting at-risk species (Margules and Pressey, 2000). As such, I identified 16 species of fauna (Table 2.2; Figure 2.3b) based on threat status, geographic distribution and data availability. For instance, the conservation of Wolverine (Gulo gulo) and American Marten 37  (Martes americana), which are both widely distributed and require large habitat patches, could potentially benefit a variety of other species. The American Marten is also a commonly used indicator species for mature boreal forest biodiversity (Thompson, 1991; Buskirk and Powell, 1994), and its home ranges have been used to investigate whether dynamic (floating) protected areas can maintain old growth in Québec’s boreal forest (Rayfield et al., 2008). The Rusty Blackbird (Euphagus carolinus), a boreal breeding song-bird and one of the 11 at-risk listed bird species considered, is widespread across the North American boreal forest, but has experienced sustained population declines, range retractions and local extirpation over the last 40 years (COSEWIC, 2006).   38   Figure 2.3: (a) Spatial distribution of 15 environmental domains (Powers et al., 2013a). (b) Species richness of 16 threatened species. (c) Global Forest Watch Canada (GFWC) intact forest landscape (Lee et al., 2010) and current protected areas (IUCN I-IV). (d) Spatial distribution of access cost surrogate.             39  Table 2.2: Priority species based on threat status, geographic distribution and data availability. Distribution maps are illustrated in Appendix A.  Common Name Scientific name Status (COSEWIC) American Marten Martes americana atrata Threatened Wolf Canis lycaon Special Concern Wolverine Gulo gulo Endangered; Special Concern Woodland Caribou Rangifer tarandus caribou  Endangered; Threatened; Special Concern Grey fox Urocyon cinereoargenteus Threatened  Barrow's Goldeneye Bucephala islandica Special Concern Chimney Swift Chaetura pelagica Threatened Harlequin Duck Histrionicus histrionicus Special Concern Rusty Blackbird Euphagus carolinus Special Concern Sprague's Pipit Anthus spragueii Threatened Yellow Rail Coturnicops noveboracensis Special Concern Whooping Crane Grus Americana Endangered Peregrine Falcon  Falco peregrinus anatum/tundrius Special Concern Burrowing Owl Athene cunicularia Endangered Piping Plover  Charadrius melodus circumcinctus/melodus Endangered Ferrugionous Hawk Buteo regalis Threatened  2.3. Naturalness surrogate (intact forest landscapes) To define locations where the forest ecosystems can be considered largely natural, I identified intact forest areas larger than 5,000 hectares using Global Forest Watch Canada’s (GFWC) Intact Forest Landscapes dataset, which maps an interpretation of forest landscape fragments that meet a series of size and access rules (Lee et al., 2010; Figure 2.3c). GFWC defines intact areas as natural ecosystems areas (e.g., forest, tundra, wetland ) within forest ecozones where no human influence can be visually detected via Landsat satellite images (Lee et al., 2010). A buffer was applied to human disturbances depending on the infrastructure type, with spatially smaller or linear infrastructure (e.g., power lines, oil and gas pipelines, camps ) receiving a 500 m buffer and spatially larger infrastructure (e.g., populated places, oil and gas plants, mines ) having a 40  buffer width of 1000 m. The combined buffered features in conjunction with areas with substantial waterbodies (>400 000 ha) and forest landscape fragments containing a large waterbody component were not considered intact. By these criteria and definition, there was approximately 4 272 805 km2 of intact forest landscapes within the Canadian boreal forest. 2.4. Human “access” as a cost surrogate I used human “access” as a cost surrogate (Figure 2.3d) since areas closer to human settlements or main road networks have a greater present potential for resource extraction than remote or roadless areas (Naidoo et al., 2006); thus, inclusion of accessible areas into new protected areas is more limited, especially if they contain altered habitat (Andrew et al., 2012). That being said, in many instances the more southerly and accessible areas may warrant additional conservation emphasis as they generally contain greater diversity and have the potential to constrain disturbance activity to localized areas. From a social perspective, there are fewer land-use conflicts in areas with limited human access, particularly in the northern boreal (Sanderson et al., 2002), which makes addressing issues surrounding multiple competing land uses simpler. As such, reserve design solutions involving remote areas often garner higher social acceptance and therefore are more likely to be implemented. Access was calculated using distance to main roads and human settlements across a 1 km spatial grid (Wulder et al., 2011) derived from Statistics Canada’s 2008 road network (Statistics Canada, 2008) and the circa 2000 Version 2 DMSP-OLS Night-time Lights Time Series cloud-free composites (NOAA, 2011). The inverse of these distances were summed, and then rescaled to a cost surrogate ranging from 0-1000 for ease of interpretation.  Higher values were assigned to areas with close proximity to human influence 41  (e.g., road or settlement); thus, encouraging the preferential prioritization of more remote, low cost areas.    42  3. A remote sensing approach to biodiversity assessment and regionalization of the Canadian boreal forest 3.1. Introduction The Canadian boreal forest has marked ecological, cultural and economic importance. Extending over 5.5 million km2 (Brandt, 2009), Canada’s boreal provides many essential ecosystem services such as regulating regional and global climate, sequestering large amounts of carbon, and purifying water. It is one of the world’s largest contiguous forests, making up 25 percent of the world’s remaining large frontier forests (Lee et al., 2003). The Canadian boreal’s expansive forests, wetlands and lakes also offer essential habitat for a diverse range of plants and animals, including over 90 endangered species such as the whooping crane [Grus americana], woodland caribou [Rangifer tarandus] and grizzly bear [Ursus arctos horribills]. The boreal forest also supports continental populations of many bird species, with 450 bird species directly depending on it and up to five billion birds migrating to the region annually (NRCan, 2011). Forest cover is dominated by a number of cold tolerant tree species from the genera such as fir [Abies spp], spruce [Picea spp], tamarack [Larix laricina], poplar [Populus spp] and pine [Pinus spp].   The boreal forest contains an abundance of natural resources both renewable, such as productive soils and timbers stocks and non-renewable, such as mineral deposits and energy reserves. For natural resources such as timber, only a subset of the boreal is subject to harvesting, most commonly in the Canada’s southern boreal forest region. Mining, and oil/gas 43  exploration/extraction are of more regional importance, but are of notable economic benefit, with an estimated potential net market value at $14.5 billion CAD (est. 2002) per annum (Anielski and Wilson, 2009). Human activity, often in the form of land conversion (e.g., agriculture, mining, road construction), continues to expand into Canada’s boreal forest. Depending on opportunity and market forces, this on-going expansion, coupled with the impacts of climate variability, has the potential to expedite the loss of biodiversity and habitat within the region. Much of the northern reaches of the boreal forest are subject to limited management activities, with forest productivity generally too low and transportation costs too high to support industrial harvesting (Wulder et al., 2007). Further, fire suppression is not practiced in the northern boreal, with few ignitions leading to large areas burned annually, on average approximately 2 million ha (Stocks et al., 2003). In southern areas of the boreal, where human access is more of an issue, a wide range of anthropogenic disturbances are present, including land use changes associated with expansion of urban areas and agricultural activities. In these areas, active forest management occurs where forests following harvest are replanted and monitored to ensure sufficient natural regeneration. Thus, harvested forests can follow a successional trajectory and return to a pre-harvest state. In areas where fire rates are low, whether due to climate or effective fire suppression, forest harvesting and other industrial activities can be the primary agents of landscape level disturbance (Schneider et al., 2003).   In order to conserve and protect the boreal ecosystem and safeguard against biodiversity loss, the creation of additional protected areas is an important option. Approximately 9.8% (977,621 km2) of Canada, comprising 8.1% of the boreal (448,178 km2), is currently under some form of protection. While the appropriate level of protection is a source of much discussion, this figure 44  falls short of the national protection target of 12% often considered as a minimum standard (Brundtland, 1987). Recognition of the uniqueness of the boreal, in terms both of its ecological value and remaining high conservation potential, has inspired debate around the appropriate level and type of protection. For instance, the Canadian Boreal Initiative promotes the protection of 50% of Canada’s boreal forest through a series of large interconnected protected areas (CBI, 2011). The governments of Québec and Ontario are developing plans to protect up to 50% of the more northern areas of boreal forest within their jurisdictions (FNP, 2011; Plan Nord, 2011) mostly beyond the northern limit of commercial forestry operations. Due to existing land use and the complexity of the trade-offs required, the potential for substantial new areas of protected wilderness in more southern areas is limited. Andrew et al. (2012) indicate that much of the boreal, due to the lack of access and low productivity, is essentially functioning as though under protection, that is, the areas are de facto protected areas. While this point can be debated, the need for protection of more southerly locations at greater risk to conversion and as harbours of greater diversity, may merit more emphasis.   Area based protection targets may lead to the protection of more remote and low productivity environments at the expense of more spatially limited, yet at greater risk, ecosystems. Consideration of value-based, as well as area-based, protection targets and strategies is warranted (Andrew et al., 2011). In 2010, the Canadian boreal forest Agreement (CBFA) was signed between the Forest Products Association of Canada (FPAC) and prominent environmental organizations, whereby a key commitment was made to collaboratively scope and encourage the completion of new protected areas within the boreal (CBFA, 2011). CBFA recommendations for the expansion of protected areas are given due consideration by the relevant governments 45  (provincial, federal, and territorial), who have ultimate stewardship responsibilities and oversee park legislation and policy. These initiatives highlight the growing recognition of the boreal’s importance and urgency for protection.  While the actual extent of the Canadian boreal forest to be conserved remains uncertain, the identification and monitoring of biodiversity components represents a logical first step to defining and maintaining an effective future protected area network. Presently, protected areas in Canada areas are slightly biased to the protection of low productive environments (Andrew et al., 2011) typically found in the northern boreal or alpine regions. However, assessing biodiversity presents a challenge for Canada’s boreal forest given its size, remoteness and difficult terrain (e.g., wetlands, topographically complex / mountainous). Conventional field-based techniques to acquire biodiversity data are logistically difficult and expensive. Remote sensing methods and imagery from spaceborne/airborne platforms offer a practicable, and cost-effective means of directly and indirectly characterizing certain aspects of biodiversity over large areas that is spatially explicit, repeatable and multi-scale (Kerr and Ostrovsky, 2003; Turner et al., 2003). Land cover maps are examples of direct diversity indicators; these are used to model and predict features of biodiversity such as species composition or abundance (Turner et al., 2003). Indirect diversity indicators measure physical environmental variables that affect species distributions and communities and typically include climate, topography, vegetation productivity, and disturbance metrics (Turner et al., 2003; Duro et al., 2007).   46  Over the last decade, there has been an increasing recognition of the potential and applicability of remote sensing derived indicators for assessing biodiversity, particularly for conservation planning applications (Kerr and Ostrovsky, 2003; Turner et al., 2003; Pettorelli et al., 2005; Buchanan et al., 2008). In the Canadian boreal forest, numerous studies have successfully used indicators to map or model species diversity or richness within taxa (Table 3.1). For instance, Coops et al. (2009b) used remote sensing derived estimates of productivity, land cover, and topography to predict bird species richness across the province of Ontario, Canada. They found that the environmental descriptors were useful for predicting bird species richness, where land cover was indicated as being the driving variable of species richness while vegetation productivity and energy indicators played a pivotal role in defining the amount of species within different habitat types. In a Canada-wide study, Kerr et al. (2001) investigated butterfly species richness and its relationship to energy and climate data versus remotely derived heterogeneity data. The authors found that > 90% of the variability in species richness was explained by land cover heterogeneity combined with secondary effects of climate.  These studies provide a basis for supporting the use of remote sensing technology for assessing biodiversity over large areas such as the Canadian boreal forest. However, determining which biodiversity indicator(s) to use is not trivial, particularly if the goal is to monitor the spatially and temporally complex mosaic of flora, fauna and ecological processes that make up the boreal region.   47  Table 3.1: A sample of approaches used to map or model biodiversity (species richness) using remotely sensed data in Canada for six general taxonomic classes General taxonomy Region Study reference Methodology Data used Butterfly  Canada Kerr et al. (2001) Land cover data and environmental variables compared to observed patterns of butterfly species richness data within a GIS Land cover maps derived from AVHRR and SPOT imagery, productivity and climatic variables Amphibian  Global Gallant et al. (2007) Land use, spatial population data and general land cover classification data was used to model and assess global patterns of amphibian species distribution and change over time Spatial land use and land cover data Fish  Canada Chu et al. (2003) Freshwater fish biodiversity was estimated using species presence/absence data to model species richness from environmental variables for watersheds throughout Canada Watershed and disturbance maps, climate data Bird Ontario Coops et al. (2009b) Decision tree approach used to estimate bird species richness based on land cover, productivity and elevation indicators Land cover and vegetation productivity (MODIS), DEM (SRTM) Mammal  Canada Hawkins and Porter (2003) Species range maps used to derive measures of diversity from environmental variables using multiple regression and spatial autocorrelation analysis  Climate, topography, land cover and historical data Plant Central Saskatchewan Warren and Collins (2007) Plant richness estimated at pixel and stand level using regression modeling with distance to ridgeline, time since fire, canopy species type and canopy density variables Landsat TM, SIR-C SAR imagery, DEM, map of time-since-fire    48  Ensuring that new reserves contribute optimally to the representation of regional biodiversity is a fundamental design goal of systematic conservation planning (Margules and Pressey, 2000). To meet this goal, remotely derived biodiversity indicators are often employed to classify regions into areas with similar biodiversity characteristics, especially if there is a shortage of species distribution data (Trakhtenbrot and Kadmon, 2005). These groupings, typically labeled as regionalizations, ecoregions, environmental domains or clusters are associated with unique combinations of environmental conditions, which in theory should be representative of species diversity (Mackey et al., 1988; Belbin, 1993; Belbin, 1995; Trakhtenbrot; Kadmon, 2005). These groups will be referred to as clusters from this point forward. In targets-based approaches to conservation planning (Carwardine et al., 2009), the objective is to delineate a network of protected areas that encompasses a minimum proportional area of each cluster within a protected areas network. Supplementary indicators that describe the state of these distinct clusters, such as forest fragmentation, road density, or other measures of disturbance can also be used to verify the uniqueness and assess the suitability of areas for inclusion in conservation networks (Wulder and Franklin, 2007).   The goals of this research are to: review a variety of spatially explicit and remotely derived biodiversity indicators, assess their suitability for characterizing biodiversity within the Canadian boreal forest, and explore their potential uses in conservation planning. To achieve these goals I (i) compare a variety of biodiversity indictors based on freely available datasets, primarily from the MODerate resolution Imaging Spectrometer (MODIS), (ii) evaluate the utility of biodiversity indicators for a classification of the boreal into environmentally distinct clusters, (iii) determine 49  if supplementary indicators are useful for further describing and understanding the developed clusters.  3.2. Methods Here, I briefly outline the available datasets used in this chapter, followed by the classification (cluster analysis) and statistical analysis sections. Eight remotely derived indicator datasets were used in the classification (encompassing land cover, topography, vegetation production, and snow cover – see sections 2.1.1 to 2.1.4 ) to create the clusters and seven indicators were used post-hoc to characterize the classified clusters with respect to habitat configuration (including aspects of fragmentation) and anthropogenic disturbance (see section 2.1.5). In addition, four species richness datasets (tree, mammal, bird, and butterfly) were employed to assess the utility of remotely derived indicators for characterizing richness patterns and to provide additional description of the classified regions (see section 2.1.6). All datasets were assembled into a common 1 km spatial resolution grid that contained 4,604,910 pixels. 3.2.1. Cluster analysis Before classifying the boreal into clusters, I first examined the data distributions of the indicators (UMD land cover, wetland, ruggedness, vegetation production (DHI), spring snow cover, and autumn snow cover) and assessed their correlation structure. The indicator values within the 4,604,910 1 km cells were then converted into a table format and classified using a clustering procedure in PASW (Predictive Analytics SoftWare) Statistics 18 software [SPSS, Inc., 2010, Chicago, IL, version 18.0.2]. A cluster is a set of cells or locations, not necessarily spatially 50  contiguous, which share a range of distinct environmental conditions as described by the indicator variables. The clustering procedure imposes a tradeoff between precision and generality, which is determined largely by the number of clusters generated. Forming too few clusters means that considerably different kinds of environments are not distinguished. Forming too many clusters makes it difficult to identify trends or to describe environmental uniqueness in a useful way. I adopted the two-step approach of Zhang et al. (1996), which is able to handle large datasets with both continuous and categorical variables.   The first step involves a pre-clustering of the cells into a manageable number of sub-clusters, in this case 100. Pre-clustering is achieved through PASW’s two-step cluster classification, a sequential clustering approach that evaluates cells in sequence and uses a distance or nearness criterion to determine if the cell should be joined with the previous cluster or initiate a new cluster of its own (SPSS, 2001). Once I had the initial pre-clusters, a hierarchical cluster classification was applied in PASW to recursively merge the pre-clusters into a user specified number of clusters. Fifteen clusters were selected as they represent a level of organizational detail useful for aiding large area conservation planning within the boreal and commensurate the fifteen expert derived terrestrial ecozones regularly used in Canada (see Coops et al., 2009c for further justification).    Both the cluster analysis steps use a distance measure as a criterion for merging clusters; I used the log-likelihood distance option provided in the PASW Statistics 18 software, which is compatible with continuous and categorical (i.e., land cover) variables. The contribution of each 51  indicator to the prediction of the cluster group membership was examined using a stepwise discriminant analysis in the Statisitica software [StatSoft, Inc., Tulsa, OK, version 7.1]. The final clusters were brought into a geographic information system (GIS) format (polygon) for visualization and analysis via linking the classified table to the 4,604,910 1 km cells (i.e., the entire Canadian boreal forest). 3.2.2. Stepwise regression Using a sample consisting of the mean pre-clusters values (n = 100), I also verified that the indicators were potentially useful predictors of species richness. Specifically, I constructed four multiple regression models in R software [R Development Core Team, 2010, Vienna, Austria, version 2.12.1] using mean pre-cluster species richness values as response variables and the mean pre-cluster environmental indicators values as independent variables. Environmental indicators (land cover, topography, vegetation production (DHI), and snow cover) were first assessed with a Pearson’s correlation analysis before being included in the models, with precedence given to uncorrelated indicators with, as defined in the discriminant analysis, the largest contribution to the clusters’ creation. Models were constricted by forward stepwise selection, which sequentially adds predictive variables until the Akaike information criterion (AIC), a measure of relative model fit, is minimized (Akaike, 1973; 1974). To accomplish this, the lm linear model and stepAIC functions in the R software were used. I inferred relative importance of the selected covariate from their order of selection and their contribution to the models’ coefficient of determination (R2). 52  3.3. Results 3.3.1. Correlations between indicators A Pearson’s correlation analysis of the environmental indicators (1 km spatial resolution) confirm the presence of some intercorrelation, indicating a certain amount of redundant information (Table 3.2). The highest correlations were -0.82 between annual primary productivity and seasonal greenness and 0.73 between the spring and autumn seasonal snow cover. These fell below the correlation threshold of 0.90 proposed by Kaufman and Rousseeuw (2005) and by Mooi and Sarstedt (2011), who indicated that variables correlated above this point are problematic and may over be represented in the classification. Below this level, correlated pairs of indicators contain useful information for the cluster analysis; thus, none of the indicators were removed. The majority of pairwise correlations were significant (p < 0.05), but of low magnitude, whether positive or negative. There were few cases of significant (p < 0.05) positive or negative correlations between seasonal snow cover and productivity indicators (e.g., 0.45 to 0.60).       53  Table 3.2: Pearson’s correlation analysis of environmental indicators Variable Ruggedness  (CV) Spring  Snow Cover Autumn  Snow Cover Annual  Primary Productivity Annual Minimum  Cover Seasonal Greenness Wetland Ruggedness (CV) 1.00 0.19 0.08 -0.07 -0.70 -0.05 0.08 -0.26 Spring Snow Cover  1.00 0.73 0.05 0.60 -0.16 Autumn Snow Cover   1.00 -0.57 0.07 0.45 -0.14 Annual Primary Productivity    1.00 0.05 -0.82 0.05 Annual Minimum Cover     1.00 -0.24 0.07 Seasonal Greenness      1.00 -0.06 Wetland      1.00 3.3.2. Cluster analysis To determine the relative importance of each indicator for discriminating between the final fifteen clusters discriminant analysis was used with results shown in Table 3.3. Seasonal greenness proved to be the most important indicator, followed by the wetland and ruggedness indicators. Annual primary productivity was of least importance. Table 3.3: Discriminant analysis of cluster input variables. Variable Overall indicator importance Seasonal Greenness 1 Wetland 2 Ruggedness (CV) 3 Spring Snow Cover 4 Annual Minimum Cover 5 Autumn Snow Cover 6 Annual Primary Productivity 7  54  The clusters are mapped in Figure 3.1 and primarily span from east to west, with a pronounced latitudinal gradient consistent with the relative levels of certain climate-related indices (Table 3.4). That is, most clusters capture the latitudinal similarity in landscape conditions as expected from energy, climate, and vegetation drivers. There are two large clusters (7, 11) that occupy the middle of the boreal forest. Cluster 11 corresponds to the extensive area of wetlands known as the Hudson Bay Lowlands and contains the largest wetland component (75.6%) and is dominated by the “open shrub” MODIS class. Cluster 7 is the largest cluster, spanning the entire longitudinal span of the boreal and covering 20.9% of the study area. It is moderately productive, moderately seasonal and dominated by the “evergreen needleleaf forest” class.   The southernmost clusters (1-6, 10 and 12) are characterized by high productivity and lower seasonality. Their land cover is dominated by a combination of the “evergreen needleleaf” and “mixed” forest classes. The most distinct of these (cluster 3) is found in the Atlantic maritime and is defined by highly productive evergreen and mixed forest on drier, highly variable terrain, as one would expect of a largely managed forest land base (Wulder et al., 2008a). Both clusters 1 and 12 contain a large wetland component of 70.9% and 67.4% respectively.   The northernmost clusters (8, 9, 13, 14 and 15) have low productivity and high seasonality. Their vegetation is dominated by the “open shrubland” class. Cluster 14 represents the largest northern cluster (16.3% of the total study area). It spans the northern limit of the boreal forest, except where interrupted by Hudson’s Bay and the adjoining lowlands. Clusters 13 and 15 are distinguished by a relatively large annual minimum cover and wetland components respectively. 55  Cluster 8 is distinct in that it mostly corresponds with alpine areas above the tree limit within the MacKenzie mountain range, and some apparent areas of tundra.  56   Figure 3.1: Spatial distribution of 15 environmental clusters within the study area, encompassing the Canadian boreal forest as defined by Brandt (2009) 57  Table 3.4: Description of the fifteen clusters and the relative indicator ranking. Rankings were derived from mean indicator values  per cluster and defined by the natural breaks (Jenks) classification scheme.  Cluster General cluster location Ruggedness  Spring snow cover  Autumn snow cover Annual primary productivity Annual minimum cover Seasonal greenness Wetland  UMD land cover (vegetation type) Cluster% of boreal 1 Southern Boreal Forest Low Low Low High Med Low High Evergreen Needleleaf forest & Mixed Forest 4.9 2 Southern Boreal Forest Low Low Low High Med Low Med Mixed Forest & Evergreen Needleleaf forest 11.5 3 Southern Boreal Forest High Med Low Med Low Med Low Evergreen Needleleaf forest & Mixed Forest 4.7 4 Southern Boreal Forest Low Med Low High Med Low Med Evergreen Needleleaf forest & Mixed Forest 5.1 5 Southern Boreal Forest Med Med Low High Med Low Low Evergreen Needleleaf forest 4.2 6 Southern Boreal Forest Low Low Low High High Low Med Evergreen Needleleaf forest 1.4 7 Mid-latitude Boreal Forest Low High Med Med Low Med Med Evergreen Needleleaf forest 20.9 8 Northern Boreal Forest Med High High Low Med High Low Open Shrubland 3.5 9 Northern Boreal Forest High High Med Med Med High Low Open Shrubland 8.8 10 Southern Boreal Forest High Med Low High Med Low Low Evergreen Needleleaf forest 0.98 11 Mid-latitude Boreal Forest Low Med Low Med Med Med High Open Shrubland 9.1 12 Southern Boreal Forest Low Low Low High Med Low High Evergreen Needleleaf forest 0.94 13 Northern Boreal Forest Low High High Med High High Med Open Shrubland 3.6 14 Northern Boreal Forest Low High High Low Med High Med Open Shrubland 16.3 15 Northern Boreal Forest Low High High Low Med High High Open Shrubland 3.5   58  3.3.3. Cluster attribution with supplementary indicators Table 3.5 gives a summary of the fragmentation and anthropogenic change present within each of the fifteen clusters. The Canadian boreal is predominantly forested (Brandt, 2009; Wulder et al., 2008a); thus, clusters typically had a large forest cover component. In this case, there was a mean and maximum forest cover of 63% and 91% respectively. However, there were five clusters (8, 9, 13, 14 and 15) with less than 49% forest cover. These occur in the northernmost part of the boreal and are dominated by the “open-shrub” class. All five of these clusters had a high level of forest fragmentation typified by small and few forest patches with a high edge density. This is indicative of the presence of isolated patches of forest within a matrix of shrubs and tundra vegetation as is found in these regions. With the exception of cluster 8, which primarily occupies the alpine, the fragmentation present in these areas is related to a lessened climatic suitability for forests with increasing latitude and the concurrent presence of wetlands and lakes (Wulder et al., 2008b; Wulder et al., 2011). Thus, this fragmentation of forest is a natural state of these northern areas.  Only six clusters experienced an average value greater than 10%. The largest of these are clusters 2 and 3, with mean anthropogenic footprint of 57.5% and 53.9% respectively. Interestingly, areas with large anthropogenic change do not appear to coincide with areas that have high forest fragmentation. The southern clusters areas all have a large forest component that ranges from 54 to 91.1% forest cover.   59  Table 3.5: Description of the fifteen clusters with anthropogenic change and forest fragmentation indicator mean values Cluster Anthropogenic change (%) Standard deviation of patch size (ha.) Relative area (%) Proportion of forest patch (%) Number of forest patches Mean patch size (ha.) Edge density (m/ha.) 1 8.75 16.78 70.01 39.96 4.81 42.21 79.36 2 57.56 14.01 58.82 46.70 5.01 32.40 69.73 3 53.91 16.52 75.22 38.88 4.92 44.21 80.57 4 27.52 15.07 86.90 44.09 2.45 63.35 51.36 5 22.93 12.63 91.15 48.33 1.89 73.12 38.04 6 12.19 12.98 89.35 42.19 2.06 71.97 41.98 7 16.67 15.15 69.22 42.08 5.35 38.22 85.31 8 0.14 5.96 23.61 43.57 7.86 6.05 72.36 9 4.56 11.65 48.42 46.42 8.79 19.32 99.66 10 7.31 12.75 90.46 40.27 1.93 70.97 41.57 11 2.01 14.96 54.06 45.75 6.52 23.74 89.98 12 5.47 15.80 82.90 36.81 2.93 60.52 58.76 13 1.71 7.21 32.58 57.10 12.97 7.66 116.19 14 0.28 11.28 47.01 47.76 10.42 14.22 125.48 15 0.01 5.76 26.79 54.59 10.92 5.62 94.80     60  The highest species richness was found in the southernmost clusters (Table 3.6). Here clusters 2, 6, 7 and 12 had the highest species richness averages for butterfly (0.18 species), tree (22.9 species), mammals (~12 species) and bird (~72 species) respectively. Conversely, northern clusters possess less species richness. For example, the northernmost cluster (13), which is situated near the northern tree limit or northern boreal extent, had the lowest tree species richness average (7.54 species). Table 3.6: Mean species richness per cluster                    3.4. Indicators as predictors of species richness Based on the variables selected as important by the discriminant analysis and with low intercorrelation, three indicators (spring snow cover, wetland, and annual minimum cover) were selected for assessing the relationship with biodiversity (tree, mammal, bird, and butterfly). These models were highly explanatory (R2 ≥ 0.84) for tree and bird species richness, and were Cluster  Butterfly Tree Mammal Bird 1 0.05 21.52 12.17 68.88 2 0.18 22.47 10.95 68.80 3 0.07 19.54 9.97 51.35 4 0.10 21.81 11.65 65.16 5 0.08 22.90 11.89 67.12 6 0.11 22.79 12.46 71.29 7 0.05 16.57 11.16 52.74 8 0.07 9.23 11.12 47.39 9 0.05 11.45 11.05 43.37 10 0.06 19.95 11.19 53.10 11 0.01 14.52 11.19 54.90 12 0.06 22.35 12.37 71.86 13 0.02 7.54 10.50 37.65 14 0.02 9.66 10.94 39.98 15 0.00 7.74 10.53 37.78 61  moderately explanatory for butterfly species richness (R2 = 0.61), but explained relatively little of the variance in mammal species richness (Table 3.7). The relative importance of the variables differed among taxa, however spring snow cover was important in all models of all four taxa. In the butterfly model, one case fell outside of the ± 3 times sigma (i.e., standard deviation of the residual) limit and was deemed an outlier and removed. Normal probability plots of all four models indicated that the relationships were approximately linear, and suggest a normal distribution of the residuals (Figure 3.2). However, there were more errors observed in the models explaining butterfly and mammal species richness.  Figure 3.2: Normal probability plots of residuals for each multiple regression model: (A) bird, (B) butterfly, (C) mammal and (D) forest.  62  Tree species richness was mostly explained by spring snow cover (Table 3.7). Spring snow cover accounts for 89.3% of the variance in tree species richness, and the inclusion of wetland and annual minimum cover resulted in an additional 1.8% and 1.0% variance explained respectively.   Butterfly species richness was explained by spring snow cover and wetland. Spring snow cover accounts for 41.2% of the variance, with wetland explaining an additional 20.2% of the variance. Mammal and bird species richness were only explained by spring snow cover (Table 3.7). Table 3.7: Summary of stepwise regression - cluster input indicators and species richness (N=100) Tree species (R2 = 0.92; p-value = 0.00)  Butterfly species (R2 = 0.61; p-value = 0.00) Indicator Beta Multiple R-Square R-Square Change  Indicator Beta Multiple R-Square R-Square change Spring Snow Cover -0.96 0.89 0.89  Spring Snow Cover -0.75 0.41 0.41 Wetland -0.12 0.91 0.01  Wetland -0.46 0.61 0.20 Annual Minimum Cover -0.10 0.92 0.01                  Mammal species (R2 = 0.22; p-value = 0.00)  Bird species (R2 = 0.84; p-value = 0.00) Indicator Beta Multiple R-Square R-Square Change  Indicator Beta Multiple R-Square R-Square change Spring Snow Cover -0.47 0.22 0.22  Spring Snow Cover -0.91 0.84 0.84  In terms of the total variance explained for each taxon, the remotely derived indicators explained the most of the variance in the tree (92.6%), bird (84.0%) and butterfly (61.4%) taxa. A lesser proportion (22.6%) of mammal species richness was also explained. Overall, the stepwise multiple regression analysis indicated that the spring snow cover explained the most variance within each species richness type. Prediction of species richness was not markedly improved by including the annual minimum cover indicator.  63  3.5. Discussion In this research I utilized a variety of freely available broad-scaled (1 km) remotely derived pan-boreal indicators for the characterization and monitoring of biodiversity within the Canadian boreal. Though choosing specific indicators for assessing biodiversity can be considered subjective, our selection of indicators was guided by many past studies (e.g., Duro et al., 2007). Our results indicate that metrics of seasonality such as the spring snow cover, explained much of the variance in the species richness of three taxa of boreal flora (tree species) and fauna (birds and butterflies). This result is supported by others such as Hurlbert and Haskell (2003) who examined seasonal bird species richness across North America and remote sensing derived production NDVI at different spatial (≤ 20,000 to 80,000 km2 grid cells) and temporal (seasonal) scales. Their findings show that seasonal NDVI in conjunction with habitat heterogeneity information (i.e., topography) could explain the majority of bird species richness (69%), approximately 15% less than this study’s seasonal snow cover. They were also able to capture the seasonal dynamics of migrant species and demonstrate its importance for determining their numbers and proportions within breeding communities. This is particularly true in in northern latitudes and in the boreal regions, where the migration of breeding species is dictated by the seasonal variation in available energy.  For example, Ivits et al. (2011) demonstrated that remotely sensed total biomass, a measure of seasonal vegetation change over an area, is strongly correlated to species breeding in northern Europe and boreal regions located in Finland, Russia, Sweden, and Norway. As a result, I conclude that indicators of seasonal snow cover may be indicative of the annual dynamics of a landscape and provide insight into for example, the production of food availability, which will not be the same between winter and summer.  64   Hawkins and Porter (2003) identified potential evapotranspiration, a measure of current climate or energy input, as the strongest predictor of mammal and bird species richness in Canada, explaining 76% and 82% of the variance respectively. Like the other taxa, spring seasonal snow cover explained the most variance for mammal species richness, however in the case of mammals it was relatively lower at 22.67%. Considering the variety of boreal mammalian species types (e.g., Wolverine, Northern River Otter, Caribou, and Gray Wolf) used to quantify the species richness variable and their diverse life histories and associated behaviours, it is likely that the importance of spring snow cover as a habitat feature varies greatly. As a result, while spring snow cover may be a useful predictor of certain species (e.g., Gray Wolf [Canis lupus] and Wolverine [Gulo gulo]), it may be less effective for mammalian species richness overall. In the case of butterflies, Kerr et al. (2001) was able to explain more variation for species richness in Canada than this study (> 90% vs 61.49%). This could be explained by the sampling density and number of butterfly records within the boreal forest compared to Canada as a whole. Highest species richness for butterflies occurs in the very south of Canada, away from the boreal study area and as a result, the lower sampling densities within the boreal may reduce the predictive capacity of the regression. In addition, this study did not incorporate landscape heterogeneity, the variable identified as the strongest predictor of butterfly species richness for Canada (Kerr et al., 2001).   In this chapter I applied a quantitative regionalization approach (i.e., cluster analysis) in conjunction with key remotely sensed pan-boreal indicators to delineate the boreal forest into 65  fifteen clusters. An important advantage of this quantitative approach is that it minimizes the internal environmental variability within the cluster groupings. Consequently, this approach can result in a robust and consistent stratification for monitoring and have been shown to produce more representative ecosystem descriptions than qualitative delineation approaches (Leathwick et al., 2003). Since a quantitative cluster analysis forms nested or hierarchical clusters, it can also be flexibly used to produce classes at different levels of detail, depending on the application requirements, while maintaining functional continuity among levels (Leathwick et al.; 2003; Lugo et al., 1999). To date, many national/continental quantitative regionalization studies have demonstrated the potential of such approaches for addressing a host of conservation planning problems over a range of diverse environments, for example, in Australia (Mackey et al., 2008) and New Zealand (Leathwick et al., 2003) to Europe (Metzger et al., 2005) and Canada (Coops et al., 2009c). Within the boreal forest, the clusters defined within the paper could form a basis for an appropriate stratification, highlighting areas of unique conservation value and ultimately be used to help identify any deficiencies in current park networks.   When I specify the cluster analysis to produce fifteen clusters, it resulted in clusters of varied size, averaging around 6.6% of the Canadian boreal, with each cluster relating to a distinct set of environmental conditions. Seasonal greenness and the wetland indicators were the most important for discriminating between cluster groups. The spatial delineation of the clusters reflects a latitudinal gradient. Accordingly, seasonal greenness is highly related to climate conditions, meaning cluster boundaries are also defined largely in part by north to south vegetation productivity gradients. Likewise, there were areas within Canada’s boreal that contain a large wetland component, which, with its distinctive attributes, explains why these areas were 66  differentiated.  The similarities between the clusters generated in this study and the 14 Canada wide clusters generated by Coops et al. (2009c) are visually apparent. Both exhibit a latitudinal gradient and span from east to west and the northern clusters have a similar size and distribution. Apart from the different extents, there are a greater number of clusters from this study located in the southern boreal and the Hudson’s bay region. The additional clusters generally nest within the broader clusters of Coops et al. (2009c).  Clusters located in the northern reaches of the Canadian boreal are sparsely forested, leading to higher levels of forest fragmentation than found nationally. Similar findings were also noted by Wulder et al. (2008b), who, when employing the same landscape pattern (fragmentation) metrics, found that sparsely forested ecozones (Taiga Shield, Taiga Cordillera) were characterized by numerous small forest patches juxtaposed with low vegetation (e.g., shrub), wetlands and lakes. For those areas where data was available, the majority of anthropogenic activity, such as timber harvesting and road construction occurs in the southern boreal regions. However, these areas are also characterized as having a large forest cover component, which acts to offset the relatively small (spatially) anthropogenic disturbances. This suggests that anthropogenic disturbances may not be effectively generalized within large spatial clusters, but rather should be reported in smaller spatial units or evaluated in regional analyses (Wulder et al., 2008b).  67  4. Integrating accessibility and intactness into large-area conservation planning in the Canadian boreal forest 4.1. Introduction It is anticipated that biodiversity across the world’s boreal forest will be increasingly threatened by change, including altered disturbance regimes, the variable intensification and expansion of human activity such as land conversion and resource extraction (e.g., mineral, energy, timber) mostly in Russia and Canada’s southern boreal forest extent, and increasing global climate variability (Lee et al., 2006; Cyr et al. 2009; Bradshaw et al., 2009). Although the protection of large intact areas is seen as an important option for conservation efforts (Bradshaw et al., 2009), based on global conservation targets that consider 10-12% a minimum standard (e.g., IUCN, 1993; Coad et al., 2009), the boreal forest is under protected (Schmitt et al., 2009) at approximately 8.5% (Coad et al., 2009).  In Canada, approximately 8.1% (448 178 km2) of the boreal forest is under some form of permanent protection, with a slight bias towards low productivity environments (Andrew et al., 2011) typically found in the more northern regions or at higher elevations. However, since as much as 80% of the Canadian boreal forest is free of human disturbance and may be considered de facto protected (Andrew et al., 2012), a unique opportunity exists for implementing comprehensive conservation strategies. In the North American context, the temperature changes in the boreal ecoregion over the next 60 years (up to 2070) are projected to be relatively minor 68  compared to other regions globally (Beaumont et al., 2011). Similarly by 2100, climate driven changes in the global boreal biodiversity are expected to be less than those triggered by other dominant drivers of change (e.g., land use and nitrogen deposition) in other biomes such as savanna, Mediterranean, and alpine (Sala et al., 2000). The implications for Canada are that protected areas established in the boreal forest under current conditions are likely to retain conservation target contributions and relevancy at least in the short term until such adaptation and mitigation responses are needed. However, it is also important to recognize that the Canadian boreal forest covers an extensive area and that future climate induced disturbances will be both highly spatially variable and will have impacts that are difficult to accurately predict. Likewise, expected changes in biodiversity over the next century (up to the year 2100) can also vary greatly, and are sensitive to the degree of interaction between drivers of biodiversity change (e.g., land use, climate, nitrogen deposition, biotic exchange, and atmospheric CO2)  (Sala et al., 2000). Nonetheless, recognition of the conservation potential within the Canadian boreal forest has generated serious debate surrounding the expansion of current protected areas to include substantial new areas, with some initiatives advocating over 50% conservation of the boreal forest (CBI, 2005). As such, designing an expanded comprehensive protected area network that meets current needs should be the priority by way of complementing those protected areas that already exist and providing a basis upon which to build future protected areas as needed.  Systematic conservation planning (Margules and Pressey, 2000) focuses principally on finding cost-effective solutions to conservation problems by achieving conservation targets for the least cost. Cost of conservation can be assessed in a variety of ways, financial or otherwise, including area in reserve or costs related to acquisition, management, transaction, damage or forgone 69  opportunities (Naidoo et al., 2006).  To date, most research in conservation planning has focused on issues around ensuring adequate representation of at-risk species or representation of biodiversity elements including habitat types, species assemblages, and ecosystems. (Church et al., 1996; Cabeza and Moilanen, 2001; Onal and Briers, 2006). One approach to ensuring biodiversity is represented in conservation planning is to use environmental domains (i.e., coarse-filter, ecological regionalization) to provide an indication of the types of environmental conditions present in the landscape. These domains should in theory represent the range of species diversity that can be supported by the landscape (Mackey et al., 1988; Belbin, 1993, 1995; Trakhtenbrot and Kadmon, 2005). The environmental domain approach has been successfully applied in a number of studies for different environments, such as in Australia where environmental domains produced from a continental classification (Mackey, 2008) provided biological data in spatial conservation prioritization studies (Carwardine et al., 2010; Klein et al., 2009a,b). This approach was also applied by Coops et al. (2009) to highlight the most unique domains across Canada with 40 (classes) and 14 (classes) level classifications.    Recently, the attention of conservation planning has shifted towards incorporating spatially explicit information about economic costs (Faith et al., 1996; Stewart and Possingham, 2005; Richardson et al., 2006;  Schneider et al., 2011), by informing on a key limiting factor which has been shown to increase the effectiveness and efficiencies of conservation initiatives (Naidoo et al., 2006). For instance, Schneider et al. (2011) investigated the incorporation of spatial distribution of biological data (coarse-filter) and economic costs (foregone resource opportunities) in conservation planning to determine any conservation gain and found that the 70  efficiency of conservation solutions is improved by minimizing cost across all ecosystem representation targets.   Biological and cost considerations may not be sufficient by themselves to ensure the long-term persistence of biodiversity at regional and continental scales (Soulé and Sanjayan, 1998; Possingham et al., 2000). For example, whether protected areas are capable of supporting broad-area and long-term ecological processes and withstanding change largely depends on their size and wilderness quality (Soulé et al., 2004). Over time, the dynamic nature of the Canadian boreal forest may alter the landscape structure of protected areas. Consequently, size represents an important reserve design consideration for incorporating natural disturbance (Baker, 1992); whereby a minimum reserve size or dynamic area (Pickett and Thompson, 1978) could be used help perpetuate the viability of species and ecological processes. Thus, in the interest of maintaining biodiversity, it is beneficial to judiciously consider three key aspects: biological representativeness, the size and quality of the protected areas as well as cost considerations when planning conservation investment (Klein et al., 2009b).    Based on the three considerations outlined above, and with recent conservation initiatives in mind, I provide a case study of conservation planning for the Canadian boreal forest.  I apply spatial conservation planning tools to assess three scenarios with varied levels of reserve sizes and different conservation targets for environmental domains and at-risk species. Two of the approaches preferentially prioritized areas away from human influence (i.e., wilderness), and one prioritized intact forest landscapes. I then evaluate the trade-offs between reserve size and 71  relative reserve costs associated with the establishment of large reserves (i.e., > 6,480 km2 ), that expand by 10% intervals from a minimal target of 15% to areas that encompasses a more substantial 25% and 35% of the boreal forest. To meet our objectives I (i) determined if there was any conservation efficiency (i.e., reduced relative reserve cost) gained by using an accessibility cost surrogate instead of an area cost surrogate which has typically been used in past conservation planning efforts, (ii) evaluated how reserve compactness influences relative reserve cost and total area, and (iii) examined the effects of using the accessibility cost surrogate and forest landscape intactness on the areas selected for conservation.  4.2. Methods 4.2.1. Data Here I used a product (Landsat derived) by Global Forest Watch Canada’s to identify intact forest ecosystems (see section 2.3 for details). In addition, human “access” was used as a cost surrogate and applied to all planning units (see section 2.4 for details).  4.2.2. Prioritization approach and analysis  Since the vast majority of the Canadian boreal forest (>92%) is public land under provincial and federal jurisdiction (Wulder et al., 2007), Canada has an opportunity for expanding its protected area network to include additional large areas. As such, I have focused our analyses on new protected areas, and did not consider other conservation actions such as stewardship, zoning options or habitat restoration. The Canadian boreal forest was partitioned into 5 × 5 km grids and 72  the freely available decision-support tool Marxan (Ball and Possingham, 2000) was used to determine which areas should be prioritized. Marxan was originally developed to solve a minimum-set reserve design problem, and uses a simulated annealing algorithm capable of rapidly identifying a suite of good solutions at a minimal cost, typically “net present value” or “area”.  I used Marxan for three planning scenarios with different representative biodiversity targets ranging from 15-35% of the boreal area with different levels of compactness. Targets for both the at-risk species and environmental domains were area weighted, where; for instance, a target of 15% meant that a minimum of 15% of the boreal’s area would be protected based on the relative areas. In the case of environmental domains, a portion of each domain was given a target value (in area), based on their relative size, for a total area equal to 15%, 25%, or 35% of the Canadian boreal forest. Similarly, a portion of each species’ range was assigned a target value (in area) based on its relative size. The area weightings (for the environmental domains and species-at-risk respectively) can be defined as: 𝑇𝑖 = ��𝑓𝑖∑ 𝑓𝑖𝑁𝑓𝑖� × 𝑅� − 𝐶𝑖 Where 𝑇𝑖 is the target objective for conservation feature 𝑗. The terms 𝑓𝑖 and  𝑓𝑖 are the areas of features 𝑗 and 𝑚 respectively. The number of features, 𝑁𝑓, is either 15 clusters or 16 species-at-risk. The representative target, 𝑅, is calculated as the area of either 15%, 25% or 35% of the Canadian boreal forest. Lastly 𝐶𝑖 represents the amount of feature 𝑗 located in existing protected areas.    73  Scenario 1 provided a baseline, and did not consider intactness or access, but instead used “area” (km2) of the reserve network as a cost surrogate; thus, larger reserves are considered more costly than smaller reserves. In scenario 2 I used access as a cost surrogate to preferentially select remote areas removed from human influence into the reserve network. Likewise, scenario 3 used the access cost surrogate, but prioritization was restricted to areas identified as GFWC intact forest landscapes. Candidate reserve areas (5 x 5 km grid cells) that reside (> 50% overlap) in protected areas (IUCN status I-IV, Fig. 1c) were not considered for prioritization; however, their contribution towards biodiversity targets was accounted for. The average size for individual reserves, in all scenarios, was greater than 6,480 km2  (0.12% of boreal forest) a minimum reserve size calculated by Leroux et al. (2007a) for maintaining key ecological processes and function in a dynamic boreal landscape dominated by massive stand replacing fire disturbances. Reserve size and compactness was controlled using a boundary length modifier (BLM) parameter to alter connectedness of the reserve system. In this respect, a larger BLM places greater importance on the reserve system’s compactness than cost efficiency (Ball and Possingham, 2000) and will result in a more spatially compact reserve shape. Reserve compactness has important economic implications, as the cost of management often scales more closely with the reserve’s boundary length than its area (Possingham et al, 2000). Further, compact reserves can improve local persistence by reducing edge effects (Murcia, 1995; Cabeza and Moilanen, 2001).  Through examining the degree of spatial clustering (boundary length vs. area) in relation to the BLM values (Stewart and Possingham, 2005), I selected three different BLM weightings (0.57, 3.40 and 7.00) to assess trade-offs in moderately compact, compact and very compact reserve designs at different conservation targets.    74  In total, 500 Marxan runs (separate iterations) were generated for each scenario configuration (target and compactness level) and the best solutions (27 in total, one for each 500 run sets)  were used to evaluate trade-offs associated with relative cost, compactness and target representation. Specifically, we compared the performance between the best solutions of the three scenarios to identify differences in the total area prioritized (km2) and reserve efficiency (i.e., relative reserve cost). Relative reserve cost was calculated as the cost (based on either the area or access cost surrogate) of those areas selected for prioritization over the total cost of all areas. Best solutions and selection frequencies were also used to illustrate how compactness, cost and intactness influence the spatial distribution of prioritization. (see Appendix B for more details on scenario constrains) 4.3. Results 4.3.1. Reserve efficiency, total area, and spatial prioritization  All 27 best solutions (Figure 4.1) were able to satisfy target requirements. With the exception of two cases, the reserve efficiencies, determined by relative reserve cost, across the different levels of compactness were very similar for the 3 target levels, differing between 1.9-5.7% of relative cost. The most efficient-lowest cost solution was achieved using access as the cost surrogate (scenario 2), particularly for moderately compact reserve solutions whereas the area cost surrogate (scenario 1) was more costly. Conversely, using access as the cost surrogate, but restricting candidate priority areas to only intact areas (scenario 3) typically resulted in the least cost-efficient solution, especially when compactness and target levels increased. Most notably, the highest relative cost was observed in the compact and very compact reserve solutions at the 75  35% target, where scenario 3 was approximately 14 and 18% more expensive than scenario 2 at these compactness levels. It is also important to note that increasing compactness resulted in a large cost increase. Specifically, the largest cost differences between the levels of compactness were approximately 10.6%, 9.8% and 22.2% for the 15-35% targets respectively.   Figure 4.1: Total area prioritized and relative cost (reserve cost/total reserve cost) of best scenario solutions for three different representative area-based targets: (a) 15%, (b) 25%, and (c) 35%. Reserve cost is determined by cost (based on the area or accessibility cost surrogate) of prioritized areas. Total reserve cost refers to the sum of the cost for all candidate priority areas.  Similar to what was observed in the reserve efficiency comparison, the largest total priority area differences between the scenarios occurred in the last two cases for the compact and very compact reserve solutions at the 35% target (Figure 4.1). Here scenario 3 contained approximately 654 600 km2 (12.2% of the boreal forest) and 788 500 km2 (14.7% of the boreal forest) more area than scenarios 1 and 2. Aside from these two cases, there were only slight area differences observed between the scenarios for each respective compactness level, ranging from approximately 12,725-202,000 km2 (0.2-3.8% of the boreal forest) and with scenario 3 consistently selecting the most total area. In contrast, when comparing the area differences 76  between the levels of compactness, they differed by as much as 7.1%, 6.3%, and 16.3% of the boreal forest for the 15-35% targets respectively.   In Figure 4.2, I looked at prioritization distribution by comparing the selection frequency and best scenario solutions for each target level (15-35%) at a moderate compactness (average reserve sizes most comparable to current IUCN status I-V protected areas). Scenario 1 (Figure 4.2a) had a tendency to prioritize more in the western boreal forest (e.g., Northwest Territories, Yukon and northern British Columbia) for areas surrounding the MacKenzie Mountain Range, Coast Mountains and Great Bear Lake. While these areas are characterized by lower productivity and high seasonality, they do contain more of the at-risk species, as indicated by the distribution maps. Areas located in southwestern boreal forest were also more commonly prioritized in scenario 1. Both scenarios 2 and 3 (Figure 4.2b, c) selected more areas in northern Québec and in Ontario and Québec along the northern portion of the Great Lakes - St. Lawrence drainage basin. These are low access areas that contain fewer of the at-risk species (Table 2.2). However, scenario 3 preferentially prioritized much more areas in the central boreal forest, principally across the southern extent of the GWFC’s intact forest landscape (Lee et al., 2010). 77   Figure 4.2: Best scenario solutions (top panel) and selection frequencies (bottom panel) for different targets (15–35%) for the same compactness level (moderately compact). (a) Area cost surrogate incorporated. (b) Access cost surrogate incorporated. (c) Access cost surrogate incorporated and prioritization restricted to intact forest landscapes only. Selection frequency is used to determine how often a specific candidate priority area (i.e., 5 km2 grid) is selected over the 500 runs, and provides an indication of its relative importance for an efficient reserve design.  78  Figure 4.3 illustrates candidate areas commonly selected (>50%) in all scenario, target and compactness levels, and are described in Table 4.1. These ecologically diverse areas span across the boreal forest and are mostly situated adjacent to current protected areas. Located near the Parc national des Lacs-Guillaume-Delisle-et-à-l’Eau-Claire, the northernmost candidate area (Figure 4.3, 1), for instance, is characterized by low productivity and high seasonality (Coops et al., 2008), which can be attributed to its predominately low to high subarctic ecoclimate and the cool summers and very cold winters. The more or less poorly drained area supports open, very stunted conifer stands (e.g., Black spruce (Picea mariana) and tamarack (Larix laricina)) and shrubs.  In contrast, the eastern maritime area (Figure 4.3, 2) is highly productive and experiences low seasonality (with respect to DHI seasonality index; Coops et al., 2008). The land cover is dominated by dense, intermediate stands of mixed and evergreen needleleaf forest. Commonly selected areas within the distinctive region known as the James Bay Lowlands (Figure 4.3, 4) contain the largest wetland component and are dominated by vegetation consisting of sedge, mosses and lichens. These areas experience cooler summers and cold winters, and are characterized by moderate productivity and seasonality.       79   Figure 4.3: Areas commonly prioritized (>50%) in all scenario runs. Numbers correspond to area description in Table 4.1         80  Table 4.1: Description of the commonly selected areas. Feature rankings were derived from mean indicator values from environmental domains (Powers et al., 2013a) and defined by the natural breaks (Jenks) classification scheme.  Commonly selected areas Ecoregion locations Spring snow cover  Annual minimum primary productivity Annual minimum cover Seasonal greenness Wetland  UMD land cover (vegetation type) 1 New Québec Central Plateau & Southern Ungava Peninsula High Low Med High Med Open Shrubland 2 Maritime Barrens Low High Med Low Med Mixed Forest & Evergreen Needleleaf forest 3 Central Laurentians & River Rupert Plateau High Med Low Med Med Evergreen Needleleaf forest 4 Abitibi Plains & James Bay Lowlands Med Med Med Med High Open Shrubland 5 Lac Seul Upland High Low Med High Med Open Shrubland 6 Interlake Plain Low High Med Low Med Mixed Forest & Evergreen Needleleaf forest 7 Tazin Lake Upland High Med Low Med Med Evergreen Needleleaf forest 8 Mid-Boreal Uplands High Low Med High Med Open Shrubland  4.4. Discussion In this study I applied a SCP approach to prioritize large conservation areas across the entire Canadian boreal forest using environmental domains and at-risk species. The results are consistent with the findings of other studies (e.g., Schneider et al., 2011; Klein et al., 2009a), and indicate that compactness increases the reserve cost and amount of area prioritized. Specifically, our findings suggest compactness, which determines reserve size, has more influence over reserve cost and total area than the choice of cost surrogate (i.e., access and area) and intactness (i.e., reserve condition). Moreover, our findings  suggest the trade-offs between the importance 81  of reserve size and minimizing conservation resources will require careful consideration when designing an expanded system of protected areas. For instance, larger reserves are more effective than smaller reserves for maintaining long-term persistence of area-sensitive and extinction-prone species (Burkey, 1995; Ferraz et al., 2003) and minimizing management problems (White, 1987). Conceivably, a large reserve size, with respect to the largest expected disturbance size, can also reduce the risk of damage to surrounding human-occupied lands caused by spread of disturbances outside the reserve (Baker, 1992).   Additionally, I found that the area surrogate (scenario 1) was less efficient than the access surrogate (scenario 2) for minimizing relative reserve costs, especially for less compact reserve designs, where fewer size constraints meant that the spatially variable access surrogate could be used more effectively. Alternatively, since the area surrogate is uniform across the landscape, places with higher concentrations of biodiversity elements are preferentially selected. Given the diffuse spatial extent of biodiversity elements across the boreal forest, there were few high priority areas (i.e., distinct areas that were frequently selected)  in scenario 1 and solutions were not spatially concentrated. Having few high priority areas does not imply that the low priority areas merit less conservation value, but rather that it is more likely that there is a greater amount of spatial options available for conservation investment capable of satisfying conservation targets (Carwardine et al., 2007; Game et al., 2011). Similar to what was found in a study by Carwardine et al. (2008), the area surrogate was able to effectively minimize reserve area because, irrespective of the cost surrogate used, roughly the same amount of area was required to satisfy the area-based representative targets. While minimizing area is desirable for reducing 82  conservation costs associated with, for example, acquisition and management costs, the usefulness of area as a surrogate in this study was considerably undermined by the outcome that the access surrogate also minimizes area.    The realistic possibility for new large priority areas in the southern boreal forest is severely impeded by existing land use and the complexity of trade-offs required (Andrew et al., 2012; Powers et al., 2013b). Unique to the Canadian boreal forest are vast areas far removed from anthropogenic use or disturbance that, when viewed collectively, can still provide considerable flexibility in meeting conservation goals (Sanderson et al., 2002; Andrew et al., 2012), and where the expansion of large intact areas for prioritization are much more likely to be implemented. Based on these ideas, I used two approaches (scenarios 2 and 3) with area-based representative targets to emphasize the prioritization of more remote, yet potentially less productive environments. Encouragingly, both approaches were able to meet the conservation targets for varied levels of compactness across large remote portions of the Canadian boreal forest. High priority areas (Figs 3 and 4) are distant from roads and human settlements. I believe such an approach is useful to conservation planners and stakeholders addressing priorities for biodiversity conservation at the boreal-wide scale by identifying naturally intact landscapes large enough to maintain persistence of biodiversity and ecological functions under current climate conditions.    In spite of the boreal forests capacity to support a range of reserve configurations, it is important to note that restricting prioritization to only intact forest landscapes (scenario 3) reduced the 83  reserve efficiency by limiting the amount candidate areas in the more productive southern boreal forest that were available for meeting conservation targets. As a result, much of the available intact southern boreal forest was prioritized in scenario 3, which implies better differentiation between candidate areas for conservation investment, but less flexibility in potential reserve design. Likewise, increasing target amount and the compactness level further reduces reserve flexibility in intact forest landscapes, and increases conservation costs and total area selected.   One concern is that the static systematic conservation planning approach to designing a fixed reserve, such as the one used in this study, does not explicitly account for the dynamic nature of ecological systems (Cabeza and Moilanen, 2001; Moilanen and Cabeza, 2002; Leroux et al., 2007b). Particularly for the relatively intact Canadian boreal forest, where large-area stand replacing wildfire and insect outbreaks are still the primary agents of disturbances. As noted in Stocks et al. (2002) approximately 2 million ha are impacted annually by large wildfires, with some years reaching more than 7 million ha burned. However, if reserves are sufficiently large as to accommodate periodic natural disturbances, it should enhance resilience to disturbance events and maintaining biodiversity (Carroll et al., 2010). Further, such reserves would support a wide range of seral stages, thus providing much more area for habitat-specialist species (Berg et al., 1994; Bradshaw et al., 2009).  Similarly, an anticipated period of climate change raises the question as to whether or not static protected areas established to conserve a fixed representative target of ecosystems and species can cope with greater climate variability and a changing disturbance regime (Lemieux et al., 2011). Changes to both mean climate conditions and climate variability have been shown to 84  impact the geographic distributions of plant and animal species (Zimmermann et al, 2009;  Parmesan and Yohe, 2003) and increased climate variability may have a greater influence on species and ecosystems than before (Beaumont et al., 2011). To meet long-term conservation targets, some adaptive response(s) will need to be taken by conservation planners and stakeholders to improve the robustness and resilience of protected areas to the effects of climate change (see Lemieux et al., 2011). Furthermore, the success of any adaptation and mitigation management efforts and their associated costs is highly influenced by the quality of the initial selection of protected areas (Rodrigues et al., 2000). Thus, establishing a large comprehensive protected area network in the boreal forest using these criteria, which is to some degree resilient to disturbance events, within remote areas across a variety of environmental conditions and habitat types can represent a prudent hedging strategy for reducing vulnerability to future climate change impacts, while providing the needed flexibility for future adaptation and mitigation options. The boreal ecoregion is projected to maintain a monthly temperature pattern within a range previously experienced (1961-1990) until at least 2070 (Beaumont et al., 2011), which supports the establishment of robust protected areas to meet current short-term conservation needs rather than those based solely on projected future conditions.   85  5. Evaluating reserve design efficacy in the Canadian boreal forest using time series AVHRR data 5.1. Introduction Protected areas are a vital component of biodiversity conservation and ecological sustainability. Recognition of the uniqueness of the Canadian boreal forest, in terms of both its ecological value and high remaining conservation potential (Powers et al., 2013a, Andrew et al., 2012a), has triggered a number of initiatives to expand current protected areas in this region (e.g., FNP, 2011; Plan Nord, 2011; CBI, 2005). As such, the present challenge is to design a more extensive protected area network that is realistic given the nature of the landscape, while complementing those protected areas that already exist. Systematic conservation planning (Margules and Pressey, 2000) is commonly used to develop plans that (i) help guide where (spatially) conservation investment (e.g., reserves) should be placed to efficiently meet conservation objectives, and (ii) to help prioritize candidate locations. Systematic planning is not restricted to a particular spatial scale and is typically used to guide conservation decisions both regionally and nationally (e.g., Klien et al., 2009; Rayfield et al., 2008; Leroux et al., 2007). However, even though many advances in techniques have been developed, the methods and data (e.g., conservation features) employed by most systematic conservation plans are largely based on a static view of biodiversity (Pressey et al., 2007).  86  Given the degree of anticipated changes in climate and disturbance regimes for the boreal forest (Dale et al., 2001; Fleming et al., 1998; Kurz et al., 1995), it is becoming more important to better understand (i) how changes in landscape properties can impact the effectiveness of candidate reserves, and  (ii) what reserve design (e.g., reserve compactness and connectedness) or adaptation considerations better accommodate climate and disturbance impacts and enable more effective long-term conservation (e.g., representation) of species and ecosystems. For example, reserve size represents an important reserve design consideration for incorporating natural disturbances (Baker, 1992), which, over time, can potentially alter the landscape structure and function of reserves. To address these questions, one approach would be to evaluate the effectiveness (e.g., ability to maintain initial conditions) of a range of candidate reserves before they are implemented.  A key requirement for evaluating reserve design traits is understanding how biodiversity varies both spatially and temporally. When biodiversity monitoring is required over large areas, characteristics such as species richness cannot be characterized by detailed field based measures alone and there is benefit to proxy measures that can be more easily captured for large areas and multiple time periods. Remotely sensed measures of vegetation productivity, and hence available energy, have been shown to be strong predictors of biodiversity (Waide et al., 1999; Mittelbach  et al., 2001; Hawkins et al., 2003a,b; Hurlbert and Haskell, 2003; Evans et al., 2005; Coops et al., 2008, 2009a,b; Latta et al., 2009; St-Louis et al., 2009) as well as useful for providing reliable estimates of broad scale biodiversity patterns and community composition (Kerr and Ostrovsky, 2003; Turner et al., 2003; Pettorelli et al., 2005; Buchanan et al., 2008). In principle, the amount of available energy and energy pathways in a system increases with productivity; 87  thus, highly productive areas typically support greater species richness and diversity (Walker et al., 1992). While current research that utilize such remotely sensed measures generally support the species-energy hypothesis (Bonn et al., 2004; Storch et al., 2005; Waring et al., 2006; Rowhani et al., 2008), the mechanisms that give rise to its positive relationship are still not fully understood and have inspired debate (Currie et al., 2004; Evans et al., 2005; Storch et al., 2005). Despite this lack of consensus; however, there is widespread agreement that productivity measures are a major determinant of biodiversity (Hawkins et al., 2003a; b; Field et al., 2009).  Remote sensing offers an efficient means of monitoring and assessing the state of vegetation productivity over large extents in a consistent and repeatable manner (Foody and Cutler, 2003; Kerr and Ostrovsky, 2003; Turner et al., 2003). For instance, NDVI and fraction of Photosynthetically Active Radiation (fPAR) are two examples of remotely derived vegetation metrics for monitoring and modeling vegetation dynamics over time. Time series of the widely used NDVI (Rouse et al., 1973), an empirical-based measure of “greenness” (Coops et al., 2008), have been applied in a variety of studies to assess trends in productivity since the early 1980s (e.g., Myneni et al., 1997; Kawabata et al., 2001; Slayback et al., 2003; Tateishi and Ebata, 2004; Pouliot et al., 2009). fPAR is a physically-based measure of photosynthetic activity and, while not as widely used as NDVI, provides a link to the energy used during photosynthesis, and is more directly associated with vegetation productivity.  In the past decade, fPAR has been used to construct an integrated index called the dynamic habitat index (DHI) applied, to date, in Australia (Mackey et al., 2004; Berry et al., 2007), Canada (Coops et al., 2008), and the United States (Coops et al 2009b) to assess habitat and forage conditions. This integrated index comprises three annual fPAR metrics (cumulative greenness, seasonality, and minimum cover) 88  based on ecological theory, and provides more comprehensive description of the vegetation dynamics than a single remote sensing metric (Coops et al., 2014). In the Canadian context, DHI metrics, derived from freely available Moderate-resolution Imaging Spectroradiometer (MODIS) (Justice et al., 1998), have been shown to be effective at representing broad-scale biodiversity patterns (Andrew et al., 2012b; Coops et al., 2009a; b) and community composition (Andrew et al., 2011a) as well as useful for evaluating potential productivity biases and productivity trends in existing protected areas (Andrew et al., 2011b; Coops et al., 2014). For example, Coops et al. (2009 a; b) examined the effectiveness of DHI as a predictor of breeding bird species richness in the United States and Ontario. Results indicated that DHI was able to successfully estimate bird species richness, explaining as much 75% of the variation for certain guilds. In Ontario, Michaud et al., (2014) used DHI to estimate moose occurrence and abundance, and found that the DHI indicators, particularly greenness, were able to successfully estimate occurrence.     MODIS data are available from 2000 onward and can be integrated with longer data archives available from the Advanced Very High Resolution Radiometer (AVHRR; Cracknell, 1997), whose data record begins in 1981, with operational considerations typically resulting in an initiation date of post-1985 (Pouliot et al., 2009; Fontana et al., 2012). The AVHRR-derived NDVI has been applied in both regional (Tucker et al., 2001) and global (Kidwell, 1990; Tucker et al., 2001; De Jong et al., 2012) studies on vegetation dynamics, and the utility of AVHRR NDVI time-series data have been well established (Myneni et al., 1995; Tucker et al., 2001; Zhou et al., 2001; Nemani et al., 2003; Fontana et al., 2012). Many studies that applied this remote sensing index to the study of assessing vegetation changes have found differences in the way regional climate trends affect vegetation dynamics over a marked range of ecosystems and 89  spatial scales (Myneni et al., 1995; 1997; Nemani et al., 2003; Zhao & Running, 2010; De Jong et al., 2012; Chen et al., 2014).  For example, Myneni et al. (1997) utilized AVHRR remotely derived estimates of productivity to predict global plant growth in northern high latitudes. The authors identified a general increasing trend in photosynthetic activity in the region, with the north-western portion of Canada experiencing the largest NDVI increase in North America. In a recent review of long-term AVHRR NDVI vegetation studies, Pouliot et al. (2009) developed a new 1 km data record for the years 1985-2006 to evaluate and compare NDVI trends across Canada. The comparison of trend analysis supported the positive greening trend in the north, but also found some inconsistencies, particularly in the south of the country due to land cover change. These results both indicate that AVHRR data sets are uniquely suited for monitoring long-term vegetation trends in the boreal forest based on its high latitude and level of intactness (~80% intact).   The goal of this study is to explore the capacity of DHI productivity metrics derived from a long time-series AVHRR dataset (1987-2007) to assess how reserve design configurations (size and location) impact the efficacy of candidate reserves based on mid-2000 conditions. AVHRR measures were initiated from 1987 due to fragmented AVHRR data coverage for the years 1981-1984 (Fontana et al., 2012) and poor data coverage over large portions of western Canada during 1985 and 1986. These gaps prevented consistent data processing at a continental scale for these periods. To focus our analysis I utilize a series of previously generated reserve designs (Powers et al., 2013a) that are based on MODIS productivity data from 2000 to 2005 representing mean conditions in the early to mid-2000s, and other biodiversity data. This 2000 to 2005 epoch formed the baseline period for the analysis. I then assess the ability of the reserves to maintain 90  initial DHI baseline levels though time under a natural disturbance regime across various boreal ecozones, productivity strata, and land cover types. Specifically, I compared the stability of the longer-term AVHRR DHI metrics from 1987-2007, to the averaged 2000-2005 AVHRR DHI values to establish how often, and under what conditions, reserve conditions differed from those with which they were designed in the reserve selection process to represent. I hypothesize that over the 21 year period that (i) larger reserves will be more stable than small reserves (< 1,000 km2); (ii) reserves located in productive environments will experience more variability; (iii) predominately forested reserves will be more stable than reserves dominated by shrub land cover, and (iv) predominately evergreen forested reserves will be more stable than reserves dominated by mixed forest land cover. 5.2. Methods 5.2.1. Data 5.2.1.1. Remotely sensed data: time series of DHI productivity The fPAR based DHI was computed from remotely sensed imagery to assess vegetation productivity and identify changes in habitat and forage conditions within candidate reserves. For this research the AVHRR archive (1987-2007) over Canada (see Latifovic et al., 2005) was used to derive the index components for each year at a 1 km spatial resolution. The AVHRR data record comprised of overlapping observations from all satellites of the NOAA series and were processed using a new methodology developed by Fontana et al. (2012) to enable improved geolocation and ortho-rectification accuracy (efficiency rate >90%). DHI was calculated based 91  on the April to September period to avoid very low fPAR values associated with northern hemisphere seasonality and snow contamination.   The index has three fPAR components representing different aspects of vegetation productivity: (a) the cumulative annual fPAR; (b) the annual minimum greenness and (c) seasonal variation of the greenness. Cumulative annual fPAR or annual cumulative greenness provides an indication of the annual productive capacity of a landscape (Berry et al., 2007) and is strongly associated with species richness (Connell and Orias,1964). This integrated annual estimate of greenness was calculated by summing monthly fPAR observations for each year. Annual minimum greenness describes a site’s base level of cover within a year and provides a measure of the landscape’s ability to sustain sufficient levels of greenness and permanent resident species year-round (Coops et al., 2009b). Positive values indicate that some degree of vegetation was maintained over the year, while predominately snow-covered areas, for example, will have values near or equal to zero (Coops et al., 2008, 2009b). Seasonal variation of greenness is an integrated measure linked to local climate, topography, and land use (Coops et al., 2014) and relates information about a site’s annual variation in productivity. Annual variation in productivity is calculated as the coefficient of variation (CV) of fPAR estimates. High CV values indicate extreme seasonal changes in vegetation cover or climate conditions, and typically characterize habitats at high elevations or areas with seasonal winter snow cover and spring vegetation green-up. Areas with low CV values indicate habitats with less variation in seasonal vegetation cover, such as irrigated pastures, barren land or highly productive evergreen forests.            92  5.2.1.2. Boreal stratification Three different datasets were used to stratify the boreal forest at a 1 km spatial resolution: ecozone, productivity (GPP), and land cover (Figure 5.1; Table 5.1) and were chosen to represent boreal conditions in early to mid-2000s that coincide with the MODIS and AVHRR temporal overlap. Land cover data for Canada’s boreal were obtained from the MODIS global land cover product (MOD12Q1), and includes five categorical maps derived from observations collected over a period of a year (LP DAAC, 2010). I selected the University of Maryland (UMD) classification, which was based on data from 2004 and consists of 14 general biome types such as evergreen needle-leaf forest, savanna, and grassland. The three main land cover classes were used to stratify the boreal forest: mixed forest, evergreen needle-leaf, and open shrubland. Table 5.1: Strata and data sources for the stratification of the Canadian boreal  Strata  Dataset Reference Land cover MODIS MOD12Q1 1 km land cover product Friedl et al., 2010; LP DAAC, 2010 Ecological Units Canadian Ecozones Ecological Stratification Working Group, 1995 Productivity MODIS Net Primary Production MOD17A3 1 km product Zhao and Running, 2010; LP DAAC, 2011  The Ecozone stratification of Canada (Ecological Stratification Working Group, 1995) represents the highest level of a nested ecoregion hierarchy and defines large discrete regions base on similar geology, soil, topography, vegetation, climate, land use, hydrology, and wildlife. The majority of the boreal forest is located within eight of Canada’s 15 terrestrial ecozones (Wulder et al., 2008); these were used for stratification: Boreal Shield, Boreal Plains, Hudson Plains, Taiga Shield, Taiga Plains, Southern Artic, Boreal Cordillera, and Taiga Cordillera. Lastly, the productivity strata was defined using the annual MODIS gross primary productivity (GPP) product MOD17A3 (Zhao and Running, 2010; LP DAAC, 2011). Because there are inter-annual 93  variation and long-term trends present within the dataset (Zhao and Running, 2010), an average of the GPP products was taken instead of a single year. GPP estimates of a single year are unlikely to produce an accurate representation of long-term forest productivity (Bolton et al., 2013). The 11-year average of annual GPP (2000-2011) was compiled (available at ftp://ftp.ntsg.umt.edu/pub/MODIS/NTSG_Products/MOD17/MOD17A3/) and then stratified into three relative productivity classes. 94   Figure 5.1: Map of (a) ecozones; (b) relative MODIS GPP productivity classes with unproductive (≤3000 kg C/m2/yr) in blue, productive (>3000 kg C/m2/yr; ≤7000 kg C/m2/yr) in pink, and highly productive (>7000 kg C/m2/yr) in red; and (c) MODIS UMD land cover classes with open shrubland in yellow green, evergreen needleleaf in dark green, and mixed forest in lime green. Ecozones in (a) are 1. Boreal Cordillera, 2. Taiga Cordillera, 3. Taiga Plains, 4. Southern Artic, 5. Boreal Plains, 6. Boreal Shield, 7. Hudson Plains, and 8. Taiga Shield.  95  5.2.1.3. Candidate reserves based on mid-2000 conditions  Candidate reserves were generated in chapter 4 (Powers et al., 2013a) by partitioning the boreal forest into 5 × 5 km grids and using the freely available spatial conservation prioritization tool Marxan (Ball and Possingham, 2000) to identify potential areas for prioritization. Conservation targets were set using 15 environmental domains based on remotely derived boreal specific biodiversity indicators (Powers et al., 2013b) and 16 at-risk species to represent biodiversity. A long-term DHI index representing an average of the three components for the years 2000-2005, derived from MODIS fPAR data, was used in constructing the environmental domains. Environmental domains, typically labelled as coarse filters, ecological regionalizations, ecoregions, or clusters, are associated with unique combinations of environmental conditions, which in theory should be representative of species diversity (Belbin, 1993, 1995; Mackey et al., 1988; Trakhtenbrot and Kadmon, 2005). The 6-year DHI index was used to establish a vegetation productivity baseline of DHI differences between varied sized candidate reserves located in a number of land cover, productivity, and ecozone strata. Reserve size and compactness levels were adjusted using the boundary length modifier (BLM) parameter, whereby a larger BLM places greater emphasis on reserve compactness than cost efficiency (Ball and Possingham, 2000). In total, I used 738 individual reserves, which were categorized by size: small (≤1,000 km2), medium (>1,000 km2; ≤4,000 km2), and large (>4,000 km2; ≤10,000 km2). 5.2.2. Statistical analyses The three AVHRR DHI metrics were resampled to a 5 × 5 km cell to match the reserve planning units (cell) and then aggregated to individual reserves using simple averaging. To account for the internal variability, the coefficient of variation (standard deviation / mean) of the DHI metrics 96  were also calculated, making six DHI values in total for each reserve. I then assessed whether the reserves experienced any major changes in DHI values from their baseline means (2000-2005) over the 21 year period. Significant differences between annual DHI values and baseline means were assessed per reserve size and stratification because each might respond differently to changes (e.g., disturbances and climate change impacts). In total there were 54 assessments made (six DHI values × three reserve sizes × three stratifications). Here, differences were defined as annual DHI values (± 3 SD) outside the baseline means (± 3 SD). Significant annual DHI differences were then summed for each reserve size and stratification to indicate the reserves’ ability to maintain conservation targets (i.e., initial conditions). Strata and reserve sizes with large sums indicated high temporal variability. 5.3. Results 5.3.1. Ecozone strata Analysis from section 5.2.2 confirmed that over the 21 year period there were many instances where the six DHI components were different (±3 STD) from the base-line means (2000-2005). Half of the eight ecozones contained reserves that experienced moderate to high temporal variability (≥8 years different from baseline mean out of 21 years); however, there were distinct differences in the proportion of variability between the reserve sizes and ecozone. Figure 5.2 and Table 5.2 provide an overview of the reserve temporal variability stratified by ecozone and size.  97   Figure 5.2: Reserve size (small, medium and large) score of the occurrence of major deviations (±3 STD) from the ecozone baseline means for the six dynamic habitat metrics for the years 1987-2006. Numbered ecozones are 1. Boreal Cordillera, 2. Taiga Cordillera, 3. Taiga Plains, 4. Southern Artic, 5. Boreal Plains, 6. Boreal Shield, 7. Hudson Plains, and 8. Taiga Shield. Values 0-14 represent the number of major deviations out of the 21 years. For example, a value of 10 represents 10 years the fall outside (±3 STD) the ecozone baseline mean.98  Table 5.2: Number of years that DHI metrics significantly differ (±3 STD) from the ecozone baseline Ecozone Reserve Size Cumulative Seasonality Minimum Cum. (CV) Sea. (CV) Min. (CV)  SM 0 1 0 0 5 0 Boreal Cordillera (1) MED 0 0 0 0 0 3  LG 6 3 2 0 11 1  SM 8 9 14 2 5 11 Taiga Cordillera (2) MED 0 1 1 0 0 1  LG 0 0 0 0 0 1  SM 0 0 0 0 0 0 Taiga Plains (3) MED 0 1 0 0 1 0  LG 0 0 0 0 0 0  SM 9 14 7 1 3 8 Southern Arctic (4) MED 0 2 0 0 0 3  LG 0 1 3 0 0 2  SM 3 1 3 0 0 0 Boreal Plains (5) MED 2 1 4 0 1 0  LG 5 0 3 0 0 0  SM 0 8 1 1 2 4 Hudson Plains (6) MED 0 0 0 0 1 0  LG 0 2 0 0 1 0  SM 6 3 2 0 5 1 Boreal Shield (7) MED 5 3 6 0 3 4  LG 4 3 5 0 0 2  SM 1 1 0 0 4 0 Taiga Shield (8) MED 4 1 0 0 3 0  LG 5 4 0 0 3 0   The DHI productivity metrics indicate that small reserves (≤1,000 km2) in the Southern Artic and Taiga Cordillera had relatively high levels of temporal variability in annual minimum greenness, annual minimum greenness (CV), cumulative annual fPAR and seasonal variation. Likewise, the Hudson Plains had a moderate amount of variability in seasonal variation and annual minimum greenness for small reserves. Both Boreal Plains and Taiga Shield experienced minor temporal variability. The Boreal Shield had moderate temporal variability for cumulative annual fPAR, seasonal variation, and annual minimum greenness for all reserve sizes. Large reserves (> 4,000 km2; ≤10,000 km2) within the Boreal Cordillera experienced the greatest temporal variability in seasonal variation (CV) and a moderate amount in cumulative annual fPAR. It is also important 99  to note that all reserve sizes in the Taiga Plains, an area dominated by low-lying plains, remained stable throughout the 21 year period for all DHI productivity metrics.  5.3.2. Productivity strata  The most productive strata (>7000 kgC m-2 yr-1) extends across the southern boreal border. Overall, reserves located in this stratum were the most variable over the 21 year period (Table 5.3; Figure 5.3). Reserves that are contained within the productive strata (>3000 kgC m-2 yr-1; ≤7000 kgC m-2 yr-1) experience moderate variability for the DHI metrics (Table 5.3). Situated in the northernmost boreal extent and in the northern portions of the Rocky Mountains, the least productive strata (≤3000 kgC m-2 yr-1) has long and cold winters and short and cool summers. Very low temperatures and low precipitation (e.g., ~250 mm per year) combine to reduce vegetation development and encourages only smaller plants. The large greening that occurs in the majority of this stratum during the spring and summer also acts to increase the cumulative annual fPAR and results in a high seasonality. All reserves located in this stratum experience a moderate to high amount of variability related to seasonal variation and cumulative annual fPAR, with the highest amount occurring within small reserves (Table 5.3).     100   Figure 5.3: Comparison of 1987 to 2006 mean AVHRR cumulative annual fPAR and seasonal variation to baseline means of small reserves (<1000 km2) at (a) unproductive (<=3000 kg C/m2/yr), (b) productive (>3000 kg C/m2/yr; <=7000 kg C/m2/yr) and (c) highly productive (>7000 kg C/m2/yr) sites. The shaded area shows the ±3 standard deviation for the AVHRR DHI components (blue) and baseline (grey).     101  Table 5.3: Number of years that DHI metrics significantly differ (±3 STD) from the productivity strata baseline   Productivity Strata Reserve Size Cumulative Seasonality Minimum Cum. (CV) Sea. (CV) Min. (CV)  SM 7 9 4 0 2 4 Unproductive MED 3 6 2 0 0 1  LG 5 6 1 0 0 1  SM 5 4 2 1 0 3 Productive MED 5 5 4 0 0 4  LG 4 2 3 0 0 4  SM 8 3 6 0 4 1 Highly Productive MED 10 7 10 2 0 9  LG 10 3 9 0 0 3  5.3.3. Land cover strata The variability of DHI metrics for each of the reserve sizes of the land cover classes (mixed forest, evergreen needle-leaf and open shrubland) is described in Table 5.4, and indicates that open shrub typically had the highest levels of variability. Specifically, small reserves (≤1,000 km2) dominated by open shrub land cover had the highest variability for the cumulative annual fPAR and seasonal variation (Figure 5.4). Overall, reserves that were dominated by evergreen needle-leaf forests were slightly less variable in the DHI metrics over the 21 year period than both the open shrub and mixed forest land cover types. However, large reserves (> 4,000 km2; ≤10,000 km2) in the evergreen needle-leaf strata experience a high amount of variability in cumulative annual fPAR and annual minimum greenness. Similar to what was observed with the Ecozone and Productivity strata, all reserves had a relatively low amount of variability for the annual minimum greenness, cumulative annual fPAR (CV) and seasonality (CV) DHI metrics. 102   Figure 5.4: Comparison of 1987 to 2006 mean AVHRR cumulative annual fPAR to baseline land cover means of small reserves (<1000 km2) at (a) mixed forest, (b) evergreen needleleaf and (c) open shrubland sites. The shaded area shows the ±3 standard deviation for the AVHRR DHI component (blue) and baseline (grey).    103  Table 5.4: Number of years that DHI metrics significantly differ (±3 STD) from the land cover strata baseline Land Cover Strata Reserve Size Cumulative Seasonality Minimum Cum. (CV) Sea. (CV) Min. (CV)  SM 4 1 2 0 2 0 Mixed Forest MED 6 4 4 0 0 0  LG 9 5 7 0 0 2  SM 3 1 3 0 0 4 Evergreen Needle-leaf MED 6 2 6 1 0 5  LG 3 1 1 0 0 0  SM 10 9 6 1 2 5 Open Shrubland MED 2 5 1 1 0 1  LG 5 2 1 0 0 1  5.4. Discussion In this paper, I have presented an approach utilizing remotely sensed data to assess the efficacy of candidate reserves in boreal ecosystems. Specifically, I used a long-term AVHRR data record to evaluate reserve interannual variability in the dynamic habitat index (DHI), which is closely linked to habitat and forage conditions. Reserve efficacy was gauged by how well initial DHI conditions (2000-2005 baseline) of various reserve sizes stratified by ecozone, productivity, and land cover were maintained through time. Reserves that consistently deviate (±3 STD) from the DHI baseline means over the 21 year period likely experienced highly variable habitat and food supply, which can impact the species distribution and abundance within the reserves. Overall, our results indicate that the remotely derived DHI productivity metrics were able to show differences in the proportion of temporal DHI variability between the three stratifications (ecozone, productivity, and land cover) and three reserve sizes (small, medium, large).   104  As previously discussed, the Canadian boreal can be differentiated along a north/south latitudinal gradient, with five ecozones relating northern conditions (i.e., Taiga Cordillera (2); Taiga Plains (3); Southern Artic (4); Hudson Plains (6) and Taiga Shield (8)) and three relating conditions in the south (i.e., Boreal Cordillera (1), Boreal Plains (5) and Boreal Shield (6)). Our results indicate that only small reserves (<1000 km2) within the two most northerly ecozones (ecozones 2 and 4) experienced a notable amount of temporal variability (>10 years of 21 outside ±3 STD of baseline) with respect to seasonality, cumulative annual fPAR, minimum annual greenness and minimum annual greenness (cv), suggesting that reserve size is particularly important for maintaining overall productivity levels in these regions. The numerous and large deviations from the DHI baseline mean indicate that these small reserves are likely sensitive to changes in productivity. Interpretation of these findings should take into account that small reserves contain fewer samples (less area) than large reserves; therefore, can be more affected by a sampling bias.  However, this bias is likely to be small since the DHI responses are aggregated by strata and reserve size, and small reserves still represent a sample of tens of thousands of km2. Similarly, small reserves within the Hudson Plains (ecozone 6), an area that is poorly drained and has an extensive wetland component, experienced some moderate temporal variability related to annual seasonal greenness. In this case, wetlands such as bogs, fens, swamps and marshes represent a diverse range of landscape conditions (e.g., treed, shrubby, mossy ) depending on the state of local moisture, nutrient and hydrodynamic regimes (Smith et al., 2007). Given the influence of these fluctuating regimes (e.g., water availability is often variable) over landscape conditions and variable snow and ice conditions, it is expected that there would be some variability in the reserve DHI productivity over time.   105  In the context of the southern boreal, large reserves (> 4,000 km2; ≤10,000 km2) within the mountainous Boreal Cordillera (ecozone 1) had a high degree of temporal variability (≥8 years outside baseline means) related to seasonal annual greenness (cv), but maintained a consistent seasonal annual greenness. Overall, bigger reserves (i.e., ≥1000 km2; ≤10,000 km2) typically have a greater range of DHI productivity values than small reserves, thus can be less susceptible to deviating (± 3 STD) from baseline conditions in intermittent environments (Figure 5.5). Generally, reserves large enough to accommodate periodic natural disturbances tend to be more resilient to disturbances and better able to maintain biodiversity (Carroll et al., 2010). Furthermore, such large reserves can also provide more area for habitat-specialist species by supporting a wide range of seral stages (Berg et al., 1994; Bradshaw et al., 2009). Mammal diversity and abundance, for example, varies considerably between successional stages (Fisher and Wilkinson, 2005), thus including many different stages represents an important reserve design consideration for boreal conservation.   Figure 5.5: Comparison of 1987 to 2006 mean AVHRR cumulative annual fPAR to baseline South Artic ecozone means of (a) small reserves (<1000 km2), and (b) large reserves (> 4,000 km2; <=10,000 km2). The shaded area shows the ±3 standard deviation for the AVHRR DHI component (blue) and baseline (grey).  106   Our results demonstrate that out of the three land cover strata, small reserves dominated by open shrub, areas primarily in the northern mid latitude boreal regions and in the mountainous areas within the southern Yukon and British Columbia (e.g., ecozones 1, 2, 4, 6, and 8), had the greatest amount of temporal variability in DHI values. These findings suggest that small reserves in this stratum have an inconsistent level of vegetation production, which will likely affect the ability of these reserves to continuously support food supply and habitat for species. Further, I have shown that reserves dominated by mixed forests, specifically large reserves, are slightly more temporally variable in DHI than those reserves dominated by evergreen needle-leaf. This was not surprising since evergreen needle-leaf forests maintain foliage cover throughout the year, thereby allowing them to sustain a consistently high level of productivity that typically results in a higher cumulative annual fPAR and low seasonality (Coops et al., 2008). It is likely that greater differences would be observed if the DHI metrics were calculated using the November-March months.    Our results also confirmed that reserves located in highly productive (>7000 kgC m-2 yr-1) areas experienced greater temporal variability in DHI productivity than in less productive environments (≤3000 kgC m-2 yr-1). In unproductive areas (e.g., high latitudes and elevations), vegetation productivity is constrained by very cold annual temperatures and plant available moisture, which results in a limited growing period (Kimball et al., 2006). In the southern, more productive boreal regions a longer, yet still variable, growing period allows for greater vegetation development and a higher annual cover, which, depending on environmental 107  conditions, can potentially lead to greater annual variation in DHI between years than less productive regions.   It is important to note that the candidate reserves used in this study were based partly on MODIS DHI productivity data (2000-2005) and not data derived from the AVHRR satellite sensor. While the timeframe of the baseline is only six years, it is consistent with the datasets used in current remote sensing efforts related to biodiversity monitoring or highlighting unique geographical areas across Canada and the Canadian boreal landscape with similar ecological features (e.g., Coops et al., 2009a; Andrew et al., 2011b; Powers et al., 2013a). Compared to AVHRR, the MODIS sensor provides better spatial and spectral resolution (Gallo et al., 2005), and improved atmospheric corrections and geo-registration. As a consequence, MODIS products, such as NDVI and DHI, have been regularly used since the 2000s to assess spatial and temporal changes in vegetation condition. However, the overall agreement of productivity indices between AVHRR and MODIS is high (e.g., van Leeuwen, 2006; Ji et al., 2008). Here, a newly processed historical AVHRR data set by Fontana et al. (2012) was used to derive DHI across Canada’s boreal. The improved geometric accuracy of this new data set is in agreement with the standards outlined by the Global Climate Observing System (GCOS) (WMO, 2006; 2011); thus, facilitates a more precise examination of changes between years and better cross-sensor (AVHRR) continuity. It should also be noted that processed NDVI data using this AVHRR data set were shown to be highly correlated to NDVI MODIS for spatially and temporally (2001 to 2005) overlapping areas across a variety of biomes and ecoregions (Fontana et al., 2012).   108  The problem of evaluating the effectiveness of protected areas is complex, and made more challenging by climate change. Addressing this challenge will necessitate the use of accessible, spatially explicit and long-standing information sources or models. As such, inherent uncertainties will arise when assessing protection, be it from estimated species distribution models, climate change models, emission projections or observations derived from biodiversity proxies. Using temporal variability in DHI productivity as a measure of conservation effectiveness relies on the assumption that protected areas that remain constant through time will experience less severe ecological and biodiversity changes and constitute a better conservation investment. This assumption may not always be the case. It is wholly possible that protected areas that undergo large changes in DHI productivity may retain their conservation value, or even become more important, but different with respect to what habitats and species they are able to support. Nevertheless, incorporating reserve design considerations that facilitate consistent productivity values, while it does not always guarantee a better investment from a biodiversity perspective, it does represent a less risky one (with respect long-term protection value). Sites with stable vegetation productivity, and hence stable energy, generally maintain similar habitat conditions and resources (e.g., food supply and biomass), and are likely to support similar levels of biodiversity. Moreover, sites with stable productivity and predicable environments may contain important conditions for the occurrence of high species diversity (Fjeldså et al., 1997; Rowhani et al. 2008). Remote sensing based results by Rowhani et al. (2008), for example, offer some credence to this hypothesis. Specifically, Rowhani et al. (2008) observed a decrease in avian richness across the conterminous U.S. with decreasing energy availability and increasing energy variability, which suggests that a greater amount of avian species reside in more stable and productive environments. In essence, conservation of sites with 109  low variability in productivity or energy can help reduce uncertainty in achieving long-term conservation goals. The immense size and remoteness of Canada’s boreal forest provides a unique conservation opportunity and unprecedented flexibility in potential reserve designs. Therefore, given the current and anticipated dynamism of Canada’s boreal forest (Price et al., 2013), it seems sensible that boreal conservation planning aim to reduce uncertainties associated with change impacts if possible.    It is becoming clear in the literature that reserves do not remain static through time, and that their environmental conditions will likely undergo changes (e.g., Dockerty et al., 2003; Lovejoy and Hannah, 2005; Gaston et al., 2006; Arau´ joet al., 2004; Lemieux et al., 2011a), ultimately resulting in altered vegetation composition or density. This is particularly important for the relatively intact boreal forest, since it still experiences an active natural disturbance regime such as large-area stand replacing wildfire and insect outbreaks (Price et al. 2013). However, determining the timing, location and manner in which boreal reserve conditions will be impacted by anticipated climate variability and changing disturbance regimes is not straight forward (Lemieux et al., 2011b) and remains, in general, a major challenge for systematic conservation planning (Margules and Pressey, 2000; Scott et al., 2001; Gaston et al., 2002, Carvalho et al., 2011). Nonetheless, steps can be taken to evaluate the effectiveness of candidate reserves and identify current shortfalls and possible future vulnerably; thereby enabling better informed conservation planning (e.g., Leroux et al., 2007; Rayfield et al., 2008).  A key advantage of evaluating reserves using remotely derived data like vegetation productivity is that it addresses a major limitation in evaluating the effectiveness (e.g., changing condition) of 110  reserves by enabling objective and consistent assessments across sites and through time. Vegetation metrics like DHI and NDVI from long time-series earth observation datasets can be used to assess how well reserves maintain conservation targets under current and past conditions. As such, this approach could be useful for developing more robust long-term conservation targets by, for example, highlighting potential uncertainty in reserve designs associated with change impacts and testing potential mitigation strategies before implementation.    111  6. A conservation assessment of Canada’s boreal forest incorporating alternate climate change scenarios 6.1. Introduction Predictions of future climate, under varying climate change scenarios largely agree that Canada’s boreal forest ecosystems will experience substantial warming (Plummer et al., 2006), and face multiple direct and indirect effects, from more frequent large wildfires and extreme droughts, to potential shifts in ecosystem state (Price et al., 2013). Climate change impacts that affect habitat, and subsequently species distribution, will have potentially important implications for boreal biodiversity. This added uncertainty over the nature and alterations in comparison to present conditions of future habitats and distributions poses a considerable challenge to long-term biodiversity conservation planning in the region (Andrew et al., 2013; Lemieux et al., 2011). As such, proactive managing of uncertainty for future boreal biodiversity conservation necessitates the use of planning approaches that incorporate anticipated climate change impacts.  In general, the establishment of protected areas or reserves is the primary tool used for managing areas for conservation objectives (Margules and Pressey, 2000), and will likely play a central role in any future boreal biodiversity conservation effort. However, climate change impacts to boreal forests, such as changes in vegetation productivity, can affect the ability of these reserves to maintain current levels of food supply and biomass, which may affect long-term conservation targets, and will likely influence their efficacy. Thus, to be effective in supporting the long-term persistence of boreal biodiversity, updated conservation planning approaches will need to 112  consider the impact of these impending changes, as well as including measures of uncertainty and incorporation of boreal forest dynamics, when formulating new and expanded reserve designs. In essence, a risk-averse approach would entail targeting areas with high conservation value and low uncertainty for conservation (Moilanen et al., 2006).    Systematic conservation planning (Margules and Pressey, 2000; Margules and Sarkar, 2007) is a framework for guiding where (spatially) conservation efforts should be directed to meet conservation objectives at the least cost, and is typically employed to identify the location of areas that should be targeted for conservation (i.e., protected areas or reserves). At present, most methods and data used by conventional systematic conservation planning approaches have stemmed from a static interpretation of biodiversity (Pressey et al., 2007). However, existing planning tools and emerging methods are increasingly being used to incorporate broad-scale ecological processes and account for the dynamic nature of ecosystems as well as anticipated climate change impacts. As highlighted in a review by Leroux and Rayfield (2013), these advanced approaches include spatially explicit simulation models (e.g., Leroux et al., 2007; Rayfield et al., 2008), spatial optimization algorithms (e.g., Lourival et al., 2011), and spatial probabilistic theory (e.g., Drechsler et al., 2009; Game et al., 2011).  In the context of conserving remaining boreal wilderness areas, Leroux et al. (2007), for example, demonstrated how spatially explicit simulation models of patch dynamics and fire could be used to iteratively evaluate the ability of different competing hypothetical reserve designs to maintain their initial conservation targets over time. In a boreal-wide (Canada) study by Powers et al. (2013b), SCP approaches (i.e., Marxan) were used to evaluate the trade-offs 113  between different reserve design scenarios consisting of varied sized reserves and optimizations for both (i) wilderness/intact areas and (ii) areas with low human access. These hypothetical reserves were then evaluated in a later study (Powers et al., in press) using time-series earth observation data to determine, based on a comparison of different locations and reserve sizes, any sustained large deviations from the initial reserve productivity baseline, which could indicate a change in reserve species composition and diversity. Results demonstrated that landscape dynamics, as determined by stability of productivity, could not only be accommodated by simply establishing large reserves, but that, in many locations, productivity in small-medium sized reserves also remained stable through time; thus, constitute a worthwhile (i.e., less risky) conservation investment. Using a 200-year simulation of boreal forest dynamics, Rayfield et al., (2008) examined static and dynamic (i.e., floating; updated/re-located every 50 years) protected areas for the conservation of American Martin (Martes americana) habitat within boreal Québec (Canada). Results indicated that, over the 200-year period, the dynamic protected areas safeguarded more high quality home ranges than static reserves.             Combining probability theory with site-selection approaches (i.e., spatial conservation prioritization) is particularly helpful for allowing planners to proactively incorporate ecosystem dynamics and anticipated effects of climate change impacts at the onset of the planning process (Game et al., 2008, 2009, 2011; Lourival et al., 2011). For example, in a climate change adaption study by Game et al. (2011), probabilities based on differences between current and future conditions were used to preferentially select/identify protected areas in locations likely containing climate change refugia (i.e., areas where current environmental attributes closely match those of their projected future conditions; Saxon et al., 2005). While probability theory 114  combined with site-selection tools have been singled out as being particularly adept at accounting for the boreal forest’s dynamism (Leroux and Rayfield, 2013), to the best of our knowledge no study has applied this method to addressing boreal-wide conservation planning in Canada.  Over the last decade, researchers have provided many solutions and recommendations for incorporating system dynamics and/or climate change adaptation into conservation planning (e.g., Lemieux and Scott, 2005; Heller and Zavaleta, 2009; Game et al., 2010; Lemieux et al., 2011; Groves et al., 2012). Large and connected reserves, for example, are commonly prescribed as key reserve design components for protecting shifting patterns of biodiversity under both climate change and landscape dynamics (Bradshaw et al., 2009; Heller and Zavaleta, 2009; Andrew et al., 2013; Powers et al., 2013b). Using these recommendations as a guideline, I focus on three strategies for accounting for system dynamics and climate change effects that are well suited for a boreal-wide (Canada) conservation assessment and likely to meet long-term biodiversity conservation objectives: strategy (i) conserving large inaccessible/wilderness areas (>6480 km2), strategy (ii) protecting environmental connectivity, and strategy (iii) protecting stable habitat (stable productivity).   The creation of large protected areas from naturally functioning ecosystems that are largely without anthropogenic activity (strategy i) is viewed as an important option for maintaining the persistence of biodiversity and for allowing natural ecological and evolutionary processes to continue (Burkey, 1995; Ferraz et al., 2003). Such large reserves are possible in the Canadian boreal forest, one of the few remaining places on Earth that still possesses large tracks of remote 115  intact areas with minimal anthropogenic disturbance. Furthermore, the creation of large reserves that are environmentally diverse (i.e., provide connectivity between habitats) and robust to change, may help alleviate or buffer against some of the conservation uncertainty associated with climate change such as changing habitat conditions and species distributions. Specifically, reserves based on strategies (ii and iii) could prove more robust to climate change impacts and have sufficient environmental heterogeneity to accommodate changing habitats within the reserves. Using the entire Canadian boreal forest as a case study, our objective is to demonstrate how these three adaptive strategies can be constructively integrated with probability theory with site selection approach to provide a proactive boreal conservation assessment. Here, I ask the questions: (i) where are high priority areas and how are they characterized?; (ii) how is the overall productivity variance of sites within reserve networks influenced by different climate conditions and what are the implications for reserve networks optimized for one estimated climate scenario should another occur? To accomplish this, I use a probabilistic prioritization method to design a reserve network using predicted productivity variability (2080) based on a range of different climate scenarios incorporating varying climate change outcomes to maximize future representation of environmental conditions in locations that are stable under climate change and large enough to accommodate the Canadian boreal forest’s dynamism.   116  6.2. Methods 6.2.1. Conservation features Conservation features included (i) at-risk species and (ii) environmental domains, the details of which are provided in Chapter 2, section 2.2.  6.2.2. Quantifying vegetation variability  In this study, variance in productivity was determined by the spatial and temporal variability of an area’s vegetation productivity over a given time. In this context, areas with high productivity variability represent a riskier investment for achieving long term conservation targets. To quantify productivity variance I used future productivity maps by Nelson et al. (2014) briefly described below. These mapping and productivity predictions were carried out using boreal ecodistricts (Ecological Stratification Working Group, 1995) as the spatial unit (n = 592; ~ 100 ha).  Future vegetation productivity maps were produced using three sources of data: AVHRR DHI, historical climate data, and scenarios of future climate. As in Chapter 5, AVHRR DHI time series data processed by Fontana et al. (2012) was used (see section 5.2.1.1 for more details). Historical climate data (point; 32 km grid spacing) was derived from climate datasets (1987-2007) supplied and modeled by the Pacific Climate Impacts Consortium (PCIC) and the National Centers for Environmental Protection (NCEP) North American Regional Reanalysis (NARR) respectively. NARR climate variables (precipitation, maximum temperature, minimum 117  temperature, and mean annual growing season index) were (i) interpolated (1 km) to spatially integrate with the DHI data, and (ii) summarized annually. Lastly, three future climate scenarios from the Intergovernmental Panel on Climate Change (IPCC) were used: B1, A1B, and A2. These scenarios represent a range of possible climate change outcomes from least extreme change (B1), to business as usual (A1B), and most extreme climate change (A2).       To map future vegetation productivity, Nelson et al. (2014) utilised random forest derived regression trees from boot-strapped samples to quantify the relationship between historical climate and the DHI components. Three 1987-2007 climate-productivity models were produced, one each for annual cumulative greenness, seasonal variation in greenness, and minimum annual cover respectively. Model fits were then used with the IPCC climate change scenarios (A1B, A2, B1) to forecast DHI values for the years 2020, 2050 and 2080 (Nelson et al., 2014). In total, there were 27 future vegetation productivity maps created (3 climate change scenarios × 3 DHI components × 3 different future dates).   These future productivity maps were used to assess productivity variance associated with climate change and its impact on the productivity variability. The 2020, 2050 and 2080 modelled DHI values (27 maps) were used to calculate and map the temporal productivity variability (S) of each “ecodistrict” up to 2080. One productivity variability map was created for each DHI component, where variability was defined as S = mean / standard deviation (Tilman 1999; Lehman and Tilman 2000). For example, the mean estimated DHI cumulative productivity for 2020, 2050 and 2080 of each ecodistrict was divided by the standard deviation of the three cumulative productivity values of the same ecodistrict. Lower S values means that an ecodistrict 118  is likely to experience a greater productivity change by 2080, whereas ecodistricts with larger values suggests that the region has a higher chance of holding productivity levels.  6.2.3. Conservation targets and planning units The target representation for both the 16 at-risk species and 15 environmental domains was set at 15% and 25% of the Canadian boreal forest’s extent (see Chapter 4, section 4.2.2 for more details).  Here I partitioned the Canadian boreal forest differently than Chapter 4, in that I used 10 × 10 km grids/planning units instead of 5 × 5 km. Reducing the solution space from 190,000 to 50,529 planning units was necessary to obtain tractable and optimal solutions using the newer modified version of Marxan called Marxan with probabilities (hereafter referred to as MarProb; Watts et al., in review). In addition to finding near optimal reserve solutions, MarProb is also tasked with accounting for uncertainties (e.g., uncertainty associated land cover map accuracy, anthropogenic and natural threats) associated with each planning unit’s contribution towards meeting conservation targets. Thus, MarProb has less capacity to handle such large solution spaces as the standard Marxan. As before, I used the cost surrogate of “human access” to define the cost of each planning unit, where locations closer to human activities are deemed less cost effective (see Chapter 2, section 2.4 for more details). 6.2.4. Prioritization approach and analysis  I used a combined probabilistic and site selection approach to identify hypothetical boreal protected areas in locations that are likely to maintain consistent vegetation productivity. First, productivity probabilities that correspond to predicted DHI variability were assigned to each of 119  the 50,529 planning units (10 × 10 km). Specifically, the 2080 ecodistrict DHI productivity variability maps for each climate scenario (A1B, A2, and B1) were averaged across each planning unit. For each climate scenario, a composite of the DHI 2080 variability components was created and normalized to a scale from 0 to 1 (90th percentile). A low probability indicates a greater likelihood of a planning unit’s productivity being less variable through time (Figure 6.1). Hereafter I refer to this probability as vegetation variability probability (VVP). In total there were three 2080 VVP maps created – one for each climate scenario. 120   Figure 6.1: Map of 2080 VVP for IPCC climate change scenarios (a) B1 least extreme change, (b) A1B business as usual and (c) A2 most extreme change. Dark grey areas indicate currently protected areas (IUCN I-IV)  121  Second, VVP values were used as inputs in MarProb to identify reserve networks optimized to reduce productivity variance in each of the three climate scenarios (A1B, A2 and B1). In addition to MarProb incorporating the “uncertainty” probability of a feature h occurring in the planning unit i (phi), it also has a “certainty target” (Ch) which can be used to set the target confidence (e.g., 70%, 80% or 90% chance of meeting a conservation target). The main difference between Marxan and MarProb is the inclusion this new term in the objective function (see Table 6.1 for term description):     𝑤� 𝐹ℎ𝐻(𝑆ℎ)(𝑆ℎ𝐶ℎ)𝑁ℎℎ=1   Table 6.1: MarProb term description and definitions  Term Description W Probability weighting that can be applied to place greater importance to this term relative to others in the objective function (e.g., optimizing for cost or compact boundary) Fh A penalty applied to reserve solutions should they not satisfy the target amounts of all features at the specified likelihood (ℎ = 1. . . 𝑁ℎ) Sh The shortfall representing the difference between the estimated probability of meeting the conservation target and the certainty target H 1 when S ≤ 0 and zero if S ≥ 0   To test the influence of different VVP on site prioritization, we compared the outcomes of reserve planning scenarios with and without the inclusion of VVP, using the conservation planning software Marxan 2.43 (Ball et al., 2009). Marxan aims to minimize the objective function, a combination of the cost of reserves and the boundary length (a penalty that is applied using the boundary length modifier (BLM)), subject to meeting representation targets. We used the same parameters (representation targets, cost, and BLM) across all scenarios (Table 6.2). The conventional Marxan was used to provide baseline reserve solutions for comparison, and did not include VVP. I used MarProb in scenarios 2, 3 and 4 to include A1B, A2 and B1 VVP respectively. As in Chapter 4, sites that reside (>50% overlap) in protected areas (IUCN status I-122  IV) were not considered for prioritization; however, their contribution towards biodiversity targets was accounted for. To help ensure that reserves were likely to accommodate the boreal’s dynamism (conservation strategy i), the BLM was set to 2.3, resulting in reserves much larger than the suggested 3,000 km2 (~10 times the size of the average disturbance event; Wiersma et al., 2005) or 6,480 km2 (accommodates the largest expected disturbance event; Leroux et al., 2007) minimum dynamic reserve size. The certainty target (Ch) in scenarios 2-4 was set to 0.9 or 90% confidence that the representative targets protected within the reserve network solution will still be protected in 2080.    Each scenario had 200 runs or separate sets of reserve design solutions. To explore the trade-offs associated with each scenario’s efficiency and performance, I used the best solution (out of 200) of each scenario to evaluate relative cost and the amount variability. Best solutions and selection frequencies were also used to identify common high priority areas between the scenarios. The distribution of DHI vegetation variability probabilities values for areas frequently selected (>95%) in scenarios with probabilities were also presented graphically using box plots, showing median, range and 25th to 75th percentiles for each area. Lastly, I tested how a reserve network optimized for one climate scenario performs under different climatic conditions by examining the changes in the best solution VVP values. Specifically, I examined how the proportion of VVP values (low, med, high) changed. For example, to what degree does the proportion of VVP values for sites optimized for B1 change under A1B and A2 climate conditions?     123  Table 6.2: List of scenarios identifying the method, software, probability, and targets used. In total there were eight assessments (4 scenarios × 2 representative targets).  Scenario  1 Baseline scenario, Marxan, 15% as well as 25% representative targets   2 Probabilistic; MarProb; B1 VVP; certainty target of 0.9;  15% and 25% representative targets 3 Probabilistic; MarProb; A1B VVP; certainty target of 0.9; 15% and 25%representative targets 4 Probabilistic; MarProb; A2 VVP; certainty target of 0.9; 15% and 25%representative targets  6.3. Results 6.3.1. Reserve efficiency, total area and proportion of VVP values All eight best solutions from the scenarios with and without VVP values (Figure 6.2a) were able to meet representative target requirements (15% and 25%) at the certainty target (Ch) of 0.9 (scenarios 2-4). In general, there were very minor differences in reserve efficiencies, determined by relative reserve cost of management, between scenarios 3 and 4. The most efficient lowest cost solution was achieved by the baseline scenario 1, where probabilities were not included. In contrast, scenarios 3 and 4, which included VVP values for business as usual (A1B) and most change (A2) climate scenarios, had the least cost-efficient best solutions. Here scenario 2, which includes VVP for the least change (B1) climate scenario, is approximately 26% and 30% less expensive than scenarios 3 and 4 for the 15% and 25% targets respectively. Overall, the inclusion of VVP resulted in a less cost efficient solution. Compared to scenario 1, scenario 2 represented a moderate 25% and 26% increase in relative reserve cost for the 15% and 25% targets. Most notably, in this case very large cost differences were observed between scenarios 3 and 4, which were approximately 45% and 48% more expensive than scenario 1 relative reserve costs for 15% and 25% targets respectively. 124  Similar to what was observed in the reserve efficiency comparison, there were few differences between scenarios 3 and 4 (Figure 6.2b). However, with respect to the proportion of VVP values, scenario 4 contained approximately 10,320 km2 (18%) and 21,550 km2 (24%) more high VVP sites than scenario 3. In addition, the scenario 4 also contained 16,980 km2 (20%) more high VVP sites than scenario 3 for the 15% target. Together, scenarios 3 and 4 represented the largest total priority areas with 121,560 km2 and 119,860 km2 for the 15% target and 209,410 km2 and 211,140 km2 for the 25% target respectively. Comprised of mainly low VVP sites, scenario 2 contained approximately 30,000 km2 (25%) and 56,000 km2 (27%) less total priority area than scenarios 3 and 4 for the 15% and 25% targets respectively. Likewise, the baseline scenario 1 represented the fewest priority sites of all scenarios with approximately 50,000 km2 (42%) and 90,000 km2 (43%) less area than scenarios 3 and 4 for the 15% and 25% targets respectively.       125    Figure 6.2: Number of prioritized sites (a) and relative cost (reserve cost/total reserve cost) of best scenario solutions (b) for three different scenarios with VVP and the one scenario based on current conditions. T15 and T25 represent the 15% and 25% representative targets respectively. In (a) the proportion of the reserve network`s VPP values was determine using natural breaks, where very dark teal represents low VVP sites (≤ 30%), or locations in 2080 likely containing similar levels of productivity as current conditions, and medium and light teal represent high (≥ 40%) and very high (≥ 45%) VVP respectively.  126  6.3.2. Spatial prioritization and site selection frequency In Figure 6.3 and Figure 6.4, I looked at prioritization distribution by comparing the selection frequency and best scenario solutions for each target level (15% and 25%). In general, the solutions (Figure 6.3) for scenarios 1 and 2 was made up of smaller reserves than scenarios 3 and 4. As illustrated in Figure 6.4, the frequency distributions between scenarios with 2-4 were very similar. The main differences in frequency distribution occurred between scenario 1 (baseline) and those that included VVP (scenarios 2-4), whereby more sites were selected at greater frequencies in and around the western boreal forest, Hudson Bay lowlands, and Newfoundland.       Figure 6.5a shows the overlapping locations for the best solutions (scenarios with VVP) at the 25% target. These overlapping sites span across the boreal forest, encompassing a portion of each regionalization, and represent a collection of sites selected under a range of forecasted climate/productivity conditions. Similarly, we found  seven sites that were frequently selected (≥95%; over the 200 runs) in these scenarios 2-4 (Figure 6.5), and are described in Table 6.3. It should be noted that there is a 100% overlap between frequently selected areas for the 15% (Figure 6.5b) and 25% (Figure 6.5c) targets.   Overall, the seven frequently selected areas primarily consisted of sites with medium to high DHI productivity variability under A1B and A2 climate conditions (Table 6.3). However under B1 climate conditions many of these seven areas contained low DHI productivity variability. The distribution and range of the DHI vegetation variability values (not VVP) are summarized for each of the seven sites and presented in box plot format (Figure 6.6). The box plots in Figure 6.6 shows that almost all frequently selected areas (0-6) had moderate to high productivity 127  variability (< 50) for each DHI component, and indicates, with the exception of B1 DHI (Figure 6.6a), little to no relationship between the vegetation variability of a site and its selection frequency (i.e., it’s relative importance). Based on the summary from Table 6.3, it is likely that the selection frequency is related to the site’s low cost (accessibility) and species richness.   128    Figure 6.3: Best scenario solutions for different targets (15 and 25%) using the same compactness level and planning unit cost. (a) A1B VVP incorporated. (b) A2 VVP incorporated. (c) A2 VVP incorporated. (d) Prioritization based on current conditions without VVP.   129    Figure 6.4: Blue gradient maps represent selection frequencies for different targets (15 and 25%) using the same compactness level and planning unit cost. (a) A1B VVP incorporated. (b) A2 VVP incorporated. (c) B1 VVP incorporated. (d) Current conditions without VVP. Selection frequency is used to determine how often a specific planning unit or site (i.e., 10 km2 grid) is selected over the 200 runs, and provides an indication of its relative importance for an efficient reserve design.    130   Figure 6.5: Sites commonly prioritized for scenarios with VVP. (a) Overlapping best solutions for the 25% target. (b) Areas frequently selected (> 95%) in the 200 scenario (2, 3 and 4) runs for the 15% target. (c) Areas frequently selected (> 95%) in the 200 scenario (2, 3 and 4) runs for the 25% target. There is 100% overlap between (a) and (b). Numbers correspond to area description in Table 6.3.                    131  Table 6.3: Description of the commonly selected areas. Feature rankings were derived from median indicator values from 27 “DHI” variability maps, species richness (Powers et al., 2013b), accessibility cost (Powers et al., 2013b), and defined by the natural breaks (Jenks) classification scheme.  Commonly selected areas Ecoregion locations Annual cumulative greenness variability Annual minimum cover variability Seasonal greenness variability Species richness Cost UMD land cover (vegetation type) 0 James Bay Lowlands B1: Med A1B: Med A2: Med B1: Med A1B: High A2: Med B1: High A1B: High A2: High Med Med Evergreen Needleleaf forest; Open Shrublands 1 Riviere Rupert Plateau B1: Med A1B: High A2: High B1: Med A1B: High A2: High B1: High A1B: High A2: High Med Med Evergreen Needleleaf forest; Open Shrublands 2 Mecatina River B1: Low A1B: High A2: High B1: Med A1B: Med A2: High B1: High A1B: High A2: High Med Med Evergreen Needleleaf forest 3 Southern Ungava Peninsula & New Quebec Central Plateau B1: Medium A1B: High A2: High B1: Low A1B: High A2: High B1: Med A1B: High A2: High Med Low Open Shrublands 4 Ungava Bay Basin B1: Low A1B: High A2: High B1: Low A1B: Med A2: Med B1: Med A1B: High A2: High Med Low Open Shrublands 5 Selwyn Lake Upland & Kazan River Upland B1: Low A1B: Med A2: High B1: Low A1B: High A2: High B1: High A1B: High A2: High Med Low Open Shrublands 6 Coppermine River Upland B1: Low A1B: Med A2: Med B1: Low A1B: Med A2: High B1: Med A1B: High A2: High Med Low Open Shrublands                  132                                            Figure 6.6: Box plot output per frequently selected areas (0-6; x-axis) and 2080 DHI variability (S) values (y-axis) under B1 (a), A1B (b) and A2 (c) conditions. Small boxes within the box plot indicate median DHI variability; boxes represent 25th and 75th percentile, and the entire range of the DHI values is indicated by horizontal markers outside the box plot. 133  6.3.3. Comparison of the proportion of VVP values under different climate/productivity conditions  Figure 6.7 illustrates (i) the number of prioritized sites and (ii) the proportion of VVP values for those sites. Here I evaluate how the proportion VVP values of a reserve network optimized for one climate scenario changes when under different climate conditions. In both Figure 6.7a and Figure 6.7b, the proportion of low VVP values markedly increases under B1 (least change) conditions, resulting in a very small proportion of sites with high and very high VVP values. Specifically, under B1 conditions the overall proportion of low VVP sites increased by 64% and 40% in the A1B (business as usual) optimized reserve network (Figure 6.7a), and by 65% and 69% in the A2 (most change) optimized reserve network (Figure 6.7b) for the 15% and 25% targets respectively.   When the A1B optimized reserve network is under A2 conditions, there was a moderate 15% increase in the overall proportion of high VVP sites and a slight 7.5% overall decrease in low and very high VVP sites for the 15% target (Figure 6.7a). For the 25% target, there was a moderate 21% and 14% increase in the overall proportion of high and very high VVP sites and a moderate 35% overall decrease in low VVP sites (Figure 6.7a). There were only minor changes (~10%) between the overall proportions of VVP values for the A2 optimized reserve under A1B conditions (Figure 6.7b).    Under A1B and A2 conditions, there was a large increase in the overall proportion of B1 sites with high and very high VVP and, subsequently, a substantial decrease in the proportion of low VVP sites (Figure 6.7c). Specifically, the overall proportion of low VVP B1 sites decreased by 134  approximately 68% and 69% under A1B conditions and by 79% and 75% under A2 conditions for 15% and 25% targets respectively.   It is interesting to note that the reserve network based on current conditions (Figure 6.7d) contained only moderate 16% and 18% proportion of high VVP sites under B1 conditions for 15% and 25% targets, but very large proportions of high and very high VVP sites under A1B and A2 conditions. Approximately 78-85% of sites within reserves based on current conditions have high and very high VVP under A1B and A2 conditions for 15% and 25% targets.    Figure 6.7: Comparison of the number of prioritized sites and the proportion of VVP values. T15 and T25 represent the 15% and 25% representative target respectively. The proportion of the reserve network`s VVP was determine using natural breaks, where very dark teal represents low VVP sites (≤ 30%), or locations in 2080 likely containing similar levels of productivity as current conditions, and medium and light teal represent high (≥ 40%) and very high (≥ 45%) VVP respectively. (a) Reserve network optimized for A1B VVP under A2 and B1 conditions. (b) Reserve network optimized for A2 VVP under A1B and B1 conditions. (c) Reserve network optimized for B1 VVP under A1B and A2 conditions. (d) Reserve network optimized for current conditions without VVP under B1, A1B and B2 future 2080 conditions.      135  6.4. Discussion 6.4.1. Location and characterization of high priority areas In this study I applied a site selection approach using future climate and their respective VVP to provide a Canadian boreal conservation assessment. Specifically, I developed systematic conservation plans that incorporated three proactive and adaptive conservation strategies based on sound ecological principles to identify hypothetical reserves that are (i) large (≥ 6,480 km2) and low cost or low human access (wilderness) areas, (ii) representative of environmental condition and provide environmental connectivity, and (iii) are located in places where climate change impacts, like changes in vegetation productivity, are attenuated (i.e., protecting climate refugia; Saxon, 2008). Overall, results indicate that the inclusion of VVP in the site prioritization process greatly increases the cost and size of reserves. In other words, to meet conservation targets with 90% certainty under modeled productivity levels will require much larger (total area) and less efficient (costly) reserve networks. Others have also found that more area is typically required to meet boreal conservation targets when additional design criteria (e.g., connectivity, minimum reserve size, and wilderness areas) are incorporated in prioritization approaches (e.g., Beazley et al., 2005; Leroux et al., 2007a; Powers et al., 2013b). Furthermore, these findings are supported by the literature suggesting that rapid climate change will likely necessitate the protection of more area than required under static conditions to reduce the risk of under-representing current and future conservation targets (Hannah et al., 2007, Andrew et al., 2013).        136  In general, our assessment of the trade-offs between site VVP and cost implies that it is expensive to conserve low risk or low VVP areas and that it is more efficient to conserve more areas of higher VVP that are more cost effective (i.e., in this case, remote areas). Under A1B and A2 conditions, there were fewer low VVP sites available for selection, explaining why more planning units were required (as compared to the baseline scenarios 1 and scenario 2 with B1 conditions), to achieve conservation targets with a 90% certainty. However, similar costs and total areas of the networks chosen in scenario 3 and 4 were unexpected. Specifically, it was anticipated that scenario 4 would produce a larger and more costly reserve network since it included VVPs based on A2 (most change) conditions. While similar in cost and area, the proportion of high and very high VVP sites within the scenario 4 reserve network was moderately greater than scenario 3. In spite of this difference in overall reserve VVP, representation targets were still met by scenario 4 and likely attributed to its large total area (compared to scenario 1).   That there were only a few frequently selected regions with low DHI variability also implies that low cost wilderness areas are relatively important for optimized reserve design solutions. Frequently selected regions (as in Figure 6.6) for scenarios with VVP, despite having moderate to high DHI variability values, are primarily located in wilderness areas (low cost) containing a moderate amount of species richness. Given that all these scenarios share the same cost and conservation features per planning unit, it is understandable that these areas were frequently selected.  137  6.4.2. Influence and implications of different climate conditions on reserve VVP Reserve networks designed for current climatic conditions may not perform well under different climate conditions. Findings from the comparison of the proportion of VVP values suggests that reserve networks based on current conditions or incorporate projected B1 VVP may not be adequate to achieve representative targets under the influence of A1B and A2 conditions. In essence, the relative variability of sites that make up the reserve network may influence its overall efficacy. Subsequently, additional planning units are likely required to compensate for the large increase in the reserve network’s proportion of high and very high VVP sites.  6.4.3. Climate change adaptation considerations Conserving climate refugia (places less affected by anticipated climate impacts) can be considered an important hedging strategy against climate impacts by protecting species and habitats marginalized by ecological changes in other areas. However, this form of climate change adaption is by no means entirely comprehensive. One concern is that this approach relies upon the assumption that areas with relatively constant climate induced impacts will experience less severe ecological changes (Game et al., 2011). That said, how important an area is for biodiversity may not necessarily be reduced if greatly impacted by climate change, but rather it could retain its importance or become even more important, but different with respect to the habitats and species it supports (Groves et al., 2012). The protection of climate refugia alone, while it can certainly assist some ecosystems’ ability adapt to changes, it will not guarantee its viability (Groves et al., 2012). Hence, it seems wise that climate change adaptation approaches 138  include other strategies and criteria, such as those proposed here, when identifying important conservation areas.    Likewise, an inherent limitation with this adaptation approach is its reliance on modeled climate change projections, in this case projections of expected DHI productivity conditions, and the inherent uncertainties associated with those projections. Such models; therefore, are not viewed as literal truth, but as means of providing spatial and categorical insight into potential trends (Andrew et al., 2013). Based on this rationale, assessing a range of potential conditions will likely provide a more valuable insight and be of greater use for aiding scenario development and planning exercises. Another important concern is that this approach will almost always introduce additional costs into conservation decisions, thus these costs should be justifiable.   Despite the mentioned concerns and limitations, however, there is an important distinction/advantage between the site selection with probabilities (e.g., climate projections) approach used here and other climate change adaptation approaches (e.g., model simulations). Specifically, unlike the other approaches, reserve design solutions are not guided solely by climate projections, but also meet conservation targets and are optimized for efficiency (minimize cost). Thus, reserve solutions, while likely less efficient, are no worse and potentially much better than solutions produced using a conventional site selection approach that does not consider climate change impacts (Game et al., 2011). Although this may constitute as a ‘no-regret’ conservation strategy (Game et al., 2011), limited funds and resources are a reality in the conservation planning process and will likely necessitate trade-offs (Stewart and Possingham, 139  2005). For example, it might not be feasible to develop solutions that incorporate all the various elements related to climate change.   Establishing priorities capable of addressing the complexities of conservation planning in Canada’s boreal forest (e.g., biological systems, anticipated climate change, system dynamics, and limited/realistic funding and resources) remains a critical challenge (Lemieux et al., 2011) and represents a major avenue for continued research. Results from this research have many helpful implications for conservation planners and stakeholders, and can provide a useful framework for determining how to best expand existing protected areas in the Canadian boreal. Specifically, including elements of climate change impacts and ecologically driven criteria (large reserves, wilderness, and representativeness) with this approach can result in identifying areas for more effective conservation that, in the long run, will likely have important implications for conservation investment and help improve the persistence of boreal biodiversity and ecological systems.          140  7. Conclusions Canada’s boreal forest is one of the world’s largest remaining intact wilderness areas, with a land area of nearly 5.5 million km2 in size. Scientists recognize that maintaining biodiversity and associated ecosystem services requires monitoring (Kerr et al., 2003; Turner et al., 2003), and, given the vastness of Canada’s boreal forest with its diverse ecosystems, climate, land cover, topography and disturbance regimes, Canada will need to include approaches that enable frequent broad scale assessments.  Remote sensing is an important complementary data source to enable cost effective monitoring and mapping of biodiversity indicators over large extents in a consistent and repeatable manner. As such, remote sensing is capable of supporting the information needs of modern biodiversity conservation efforts (Kerr et al., 2003; Turner et al., 2003; Pettorellie et al., 2005; Buchanan et al., 2009). However, a number of critical challenges and opportunities deserve greater attention. The focus of this PhD research aims to advance the use of remote sensing and other geospatial techniques for large-area, multi-jurisdictional biodiversity conservation. The plan calls for progress in each of four research themes:   (i) Assessing biodiversity across broad areas,  (ii) Identifying areas of high conservation priority,  (iii) Evaluating the efficacy of current and hypothetical reserve networks, and  (iv) Incorporating system dynamics and climate change impacts in a Canadian boreal-wide conservation assessment  Two scientific goals as they relate to the four research themes are briefly summarized in the following sections. 141  7.1. Goal 1: Identifying critical habitat for conservation The first phase this research (Chapter 3, theme i) involves reviewing and developing a suite of spatially explicit and remotely derived biodiversity indicators (e.g., vegetation production, topography, seasonality ) for characterizing biodiversity unique to Canada’s boreal forests. This research (theme i) is designed to unfold in two parts: (a) identify the most suitable biodiversity indicators, and (b) to explore their potential uses for identifying areas of unique environmental diversity. The use of a quantitative cluster analysis approach (theme ib) phase is important, since it will allow me to use remotely derived biodiversity indicators to produce environmental domains that are continuous in nature, thereby enabling a more consistent and robust classification than the traditionally used ecoregions, which are restricted to spatially discrete areas and can subsequently contain more internal variability within groupings (Coops et al., 2009). These unique regions, typically labeled as regionalizations, environmental domains or clusters are associated with a range of different environmental conditions, which in theory should be representative of species diversity (Mackey et al., 1988; Belbin, 1993; Trakhtenbrot and Kadmon, 2005) and can be used to assess, for example, deficiencies in current reserve networks and systematic conservation planning of future reserves.   Of the three main biodiversity indicators tested, seasonality, as defined by spring snow cover, was shown to explain the most variance of species richness (bird, tree and butterfly) and, therefore, indicates that it could be a key indicator of biodiversity. I found that the 15 clusters generated from the cluster analysis were representative of a range of environmentally distinct conditions, and that seasonal greenness, along with wetland land cover, were the most important 142  indicators for differentiating between the clusters. The addition of forest fragmentation indicators provided useful information for describing the forest extent (composition) and spatial characteristics (configuration) of the clusters; however, anthropogenic disturbance was not associated with higher levels of fragmentation. I conclude that remotely derived indicators in conjunction with the quantitative cluster analysis and attribution, with both improved interpretability and added information content, have the capacity to provide useful insights for conservation planning in the Canadian boreal forest.  With respect to future conservation planning, once environmental domains are identified they can be incorporated with systematic conservation tools in conjunction with highly relevant biodiversity data (e.g., boreal specific biodiversity indicators and species at risk), for assessing potential reserves that could be important for conservation. The purpose of this research (theme ii) was to incorporate the remotely derived environmental domains with the Spatial Conservation Prioritization (SCP) approach to simulate the wide range of conservation scenarios necessary for testing the performance of this conservation approach and identifying potential areas of high conservation value. SCP tools such as Marxan can be used to help determine where (spatially) conservation investment should be prioritized. This method works by finding near optimal solutions to conservation problems by achieving conservation targets for the least cost, which can include a variety of factors such as area or economic costs associated with land acquisition, management, human accessibility, and forgone opportunity.   It is widely accepted that large reserves created from naturally functioning ecosystems can play a key role in maintaining long-term persistence of biodiversity and the continuance of ecological 143  and evolutionary processes. Such large reserves are possible in the Canadian boreal forest, one of the few remaining places on earth that still possesses large tracks of remote intact areas with minimal anthropogenic disturbance. Key aspects of this case study (Chapter 4, theme ii) include the incorporation of (i) boreal specific environmental domains to represent biodiversity, (ii) a naturalness surrogate to help single out intact forest landscapes for conservation, and (iii) human access as a cost surrogate to preferentially prioritize areas more removed from human influence. Our analysis of the three conservation prioritization scenarios applied here reveals that the level of compactness strongly affects reserve efficiency and that restricting prioritization to only intact forest landscapes reduces reserve design flexibility and efficiency. Furthermore, the application of an accessibility cost surrogate, as demonstrated in this research (Chapter 4; scenarios 2 and 3), was effective for preferentially prioritizing remote areas over a suite of conservation targets and compactness levels. These findings suggest that (i) the abundant intact areas within the Canadian boreal forest provide suitable areas for conservation investment and (ii) this coarse-scale prioritization approach is useful for aiding conservation planning at a boreal-wide scale. To determine what reserve characteristics stay relevant over time, an extension of this study could explore how conservation targets of varied reserve design configurations (size and shape) affect the long-term efficiency and efficacy of reserve systems. This research was carried out in Chapter 5 (theme iii), and was based on historic remotely derived biodiversity surrogates, in this case vegetation productivity (Fontana et al., 2012). 7.2. Goal 2: Improving boreal conservation assessments Despite the many recent advances in the field, conventional SCP approaches still remain inherently static and do not take into account system dynamics in reserve designs. Furthermore, 144  reserve networks are almost exclusively evaluated on representation and are rarely assessed for reserve effectiveness (Gaston et al., 2006; Gaston et al., 2008; Andrew et al., 2011). Changes in landscape properties can impact reserves, thus it is vitally important to understand what reserve design considerations (e.g., reserve compactness, threat, and connectedness) are necessary to ensure effective long-term conservation. For example, reserve size represents an important reserve design consideration for incorporating natural disturbances (Pickett and Thompson., 1978; Baker, 1992), which, over time, can potentially alter the landscape structure and function of reserves. Accounting for system dynamics and uncertainties associated with climate change impacts in conservation planning will, in the long run, help to ensure the preservation of species, habitat diversity and ecological systems in current and anticipated future conditions (e.g., altered patterns of land use or shifts in biological distributions by changes in climate variability and disturbance regimes). The purpose of this research (themes iii and iv) was to explore how remote sensing and other relevant spatial datasets can be used to assess what reserve design considerations are important for maximizing conservation potential under an active disturbance regime and in the face of climate change. To accomplish these aims, spatial simulation modelling tools (e.g., CONSERV) can be used to incorporate system dynamics in conservation planning by simulating, for example, patch dynamics and fire, insect infestation, and climate variability, and using these simulations to evaluate the efficacy of existing and competing hypothetical reserve networks (Leroux et al., 2007).   This research  (Chapter 5; theme iii) used a different approach where metrics (e.g., vegetation productivity) from long time-series earth observation datasets were used to determine how well reserves maintained conservation targets based on whether or not reserve values deviate from 145  long-term baseline means. Results (theme iii) showed that small reserves (<1000 km2) at high elevations, high latitudes, intermittent environments (wetlands) or dominated by open shrub experienced the greatest amount of temporal variability. In addition, larger reserves (≥1000 km2; <10000 km2) were stable under these same conditions. Reserves located at the most productive sites (>700 kg C/m2/yr) were also found to experience greater boreal variability than at areas with lower productivity.   Results from Chapter 6 (theme iv) demonstrated that the incorporation of climate change impacts, specifically estimated DHI productivity, greatly influences the cost of reserve networks and the amount of area required to meet conservation targets. Findings suggest that it was more efficient to conserve more sites spread across locations with higher VVP values, yet low cost (wilderness areas). Furthermore, low cost locations were also frequently prioritized despite having, in general, high productivity variability. Importantly, however, results also imply that reserve networks optimized under current or least change (B1) conditions will not maintain their representative targets.         In summary, the strength of approaches used in Chapter 4 (theme iii) and 5 (theme iv) is that they provide an objective and consistent means of evaluating reserve performance across different geographic areas and through time. Furthermore, highlighting uncertainty associated conservation planning (e.g., vegetation productivity variability and shifts in species distributions) can help guide conservation practitioners develop more robust long-term conservation targets in new reserves and allow for the testing of potential mitigation strategies prior to implementation. 146  7.3. Limitations While this dissertation presents some novel approaches and informative case studies to help guide decisions concerning priorities for expanding and assessing protected areas in Canada’s boreal, some limitations/concerns are worth noting: - Historical AVHRR data (1987-2007) was used in the evaluation of hypothetical reserves (Chapter 5) and in the conservation assessment of Canada’s Boreal (Chapter 6).  There is known error in the AVHRR record due to differences among sensors, which could affect interpretations of the results. However, the AVHRR data record used in this research, which comprised of overlapping observations from all satellites of the NOAA series, were processed using a new methodology developed by Fontana et al. (2012) to enable improved geolocation and ortho-rectification accuracy (efficiency rate >90%). The improved geometric accuracy of this new and highly processed data set is in agreement with the standards outlined by the Global Climate Observing System (GCOS) (WMO, 2006; 2011), and allows for effective cross-sensor (AVHRR) continuity.  - From a remote sensing perspective, one of the major themes of this dissertation was to develop an approach that provided an objective and consistent means of evaluating reserve efficacy across different geographic areas and through time. This objective was achieved by evaluating reserves based on variability in vegetation productivity (DHI). However, as with any study where modelling or proxies are used, our findings should be considered in the light of underlying assumptions and simplifications. One 147  of these assumptions is that variation in productivity is a meaningful measure for assessing how well biodiversity is being maintained in boreal reserves. As emphasized throughout the thesis, there is substantial empirical evidence that suggests areas with higher vegetation productivity, or more energy, are typically more biodiverse. Remotely derived measures of biodiversity have been shown to be good predictors of biodiversity (e.g., birds, moose, tree and butterfly) and, while this positive relationship and driving mechanisms are not yet fully understood, these measures do provide strong empirical evidence supporting the species-energy hypothesis. That said, the conservation of reserves that maintain consistent productivity, while it does not guarantee a better conservation investment, it does represent a less risky one that is more likely to maintain similar habitat conditions, resources (e.g., food supply and biomass), and existing levels of biodiversity. Moreover, sites with stable productivity and predicable environments may contain important conditions for the occurrence of high species diversity.  - Another assumption is that reserve cost, defined as human 'access', is adequately quantified using distance from road and light source (NOAA Night-time Light Time Series cloud-free composites). A concern is that our estimate of human presence, though based on highly processed data sets, may not reflect or well represent future conditions due to rapid economic development in some regions, particularly in the southern boreal forest or resource rich Alberta. Similarly, since only human access was used as the cost surrogate for reserve protection, not all economic or opportunity costs and benefits have been included in our analysis. Obviously reserves provide 148  many conservation and societal benefits in addition to incurring a suite of costs, whether in purchase of land or missed development opportunities. The challenging and complex task of quantifying of these benefits and costs under current and anticipated future conditions was beyond the scope of this research. However, the implications of our findings in terms of the trade-offs between reserve efficiency (cost effectiveness) and meeting conservation objectives, represents a conservative outlook (cost wise) that minimizes land-use conflict – a major conservation deterrent.  7.4. Future directions Future work will focus on three main areas:  - The first focuses on the need for flexible management and improving integrated or multiple land-use planning. While not explicitly investigated in this thesis, it is likely that off-reserve conservation will play an increasingly greater role in achieving conservation objectives. Economic activity (e.g., oil and gas, and mining) in Canada’s boreal has typically been limited by accessibility and its remoteness, however, it is expected to intensify and expand in the near future. Given this reality, it seems sensible to consider incorporating locations capable of supporting both species and economic activities (e.g., selective logging) in the conservation planning process whenever possible. This strategy would be especially useful in the more productive south, where the conservation potential is greatest and land-use conflict risk is high due to the increased presence of human activities and economic interests. As in Chapters 4 and 6, I suggest that spatial conservation prioritization approaches (SCP) be used to help guide the selection of 149  suitable locations for off-reserve conservation. SCP approaches like Marxan with Zones, for example, allow for multiple zoning and specific targets, thereby giving conservation planners and resource managers the flexibility needed to accommodate a variety of conservation and economic/anthropocentric objectives.  - The second area of research will be to further incorporate uncertainty in boreal conservation, particularly as it relates to climate change impacts and vulnerability assessments. Conserving biodiversity with protected areas (national parks, nature reserves or multiple-use conservation areas) in the face of climate change requires improved adaptation strategies based on ecological theory and mitigation. Specifically, such conservation measures will have to accommodate not only potential species range shifts, but changes in climate variability and disturbance regimes. Promising examples could include one or a combination of the following approaches: projected species distribution models, reserve selection algorithms or SCP, projected productivity or energy, and simulations of projected disturbance.    - The third area of research will be to assess conservation strategies based on (i) other species richness hypotheses and (ii) a more comprehensive species dataset. For example, the intermediate disturbance hypothesis (IDH; Grime, 1973; Connell, 1978) is an important hypothesis in ecology with respect understanding patterns of species richness (globally) in both terrestrial and aquatic ecosystems (Catford et al, 2011; Shea et al., 2004). The conservation strategies applied in this dissertation aim to prioritize areas that exhibit stable levels of productivity; thereby, reducing conservation risk. In contrast, IDH 150  postulates that conditions following a disturbance are conducive to colonization by both native and alien species (Catford et al., 2011), and that species richness is maximized at intermediate frequencies and/or intensities of disturbance (Hutchinson, 1961; Grime, 1973; Connell, 1978). Conservation strategies that focus on identifying/assessing these intermediate areas for conservation, thus, represent an interesting avenue for biodiversity conservation research. Lastly, this dissertation included only 16 species-at-risk in its Marxan analysis. It would be beneficial to investigate the inclusion of more species that are well-known indicators of boreal forest fauna (e.g., Blackburnian Warbler (Dendroica fusca), Harris’s Sparrow (Zonotrichia querula) and Broad-winged Hawk (Buteo platypterus)). In addition to NatureServe range data (www.natureserve.org), an excellent source of occurrence data on boreal birds is available from the boreal avian modelling (BAM) project (www.borealbirds.ca).  The Boreal Songbird Initiative (www.borealbirds.org) and Cornell Lab of Ornithology (www.allaboutbirds.org) also provide excellent lists of focal indicator species, and could be used to focus the species that form conservation targets.       150  References Akaike, H. (1973) Information theory and an extension of the maximum likelihood principle. In: Petrov BN and Csáki F (eds) Proceedings Second International Symposium on Inference Theory. Budapest: Académiai Kiadó, 267–281.  Akaike, H. (1974) A new look at the statistical model identification. IEEE Transactions on Automatic Control 19(6): 716–723.  Anderson, M.G. and Ferree, C.E. (2010) Conserving the Stage: Climate Change and the Geophysical Underpinnings of Species Diversity. PLoS ONE 5, e11554.  Andrew, M. E., Wulder, M. A., & Cardille, J. A. (2014). Protected areas in boreal Canada: a baseline and considerations for the continued development of a representative and effective reserve network 1. Environmental Reviews, 22(2), 135-160.  Andrew, M.E., Wulder, M.A. and Coops, N.C. (2012) De facto protected areas of the Canadian boreal forest. Biological Conservation 146(1): 97–107.  Andrew, M.E., Wulder, M.A., Coops, N.C., and Baillargeon, G. (2012b). Beta‐diversity gradients of butterflies along productivity axes. Global Ecology and Biogeography 21(3): 352–364.  Andrew, M.E., Wulder, M.A. and Coops, N.C. (2011a). How do butterflies define ecosystems? A comparison of ecological regionalization schemes. Biological Conservation 144: 1409–1418.  Andrew, M.E., Wulder, M.A. and Coops, N.C. (2011b) Patterns of protection and threats along productivity gradients in Canada. Biological Conservation 144(12): 2891–2901.  Anielski, M. and Wilson, S. (2009) Counting Canada’s Natural Capital: assessing the real value of Canada’s boreal ecosystems. The Canadian Boreal Initiative and The Pembina Institute. Available at: http://www.borealcanada.ca/documents/BorealBook_CCNC_09_enFINAL.pdf, Accessed on: Jan. 25, 2015.  Arau´ jo, M.B., Cabeza, M., Thuiller, W., Hannah, L. and Williams, P.H. (2004). Would climate change drive species out of reserves? An assessment of existing reserve-selection methods. Global Change Biology 10: 1618–1626.  Baker, W. L. (1992) The landscape ecology of large disturbances in the design and management of nature reserves. Landscape Ecology 7(3): 181–194.  151  Ball, I.R. and Possingham, H.P., 2000. Marxan (V1.8.2): Marine Reserve Design Using Spatially Explicit Annealing User Manual, Brisbane. Available at: http://www.uq.edu.au/marxan/, Accessed on: Jan. 25, 2015.  Beaumont, L.J., Pitman, A., Perkins, S., Zimmerman, N.E., Yoccoz, N.G. and Thuiller, W. (2011) Impacts of climate change on the world's most exceptional ecoregions. Proceedings of the National Academy of Sciences of the United States of America 108(6): 2306–2311.  Beazley, K., Smandych, L., Snaith, T., MacKinnon, F., Austen-Smith Jr, P., & Duinker, P. (2005). Biodiversity considerations in conservation system planning: map-based approach for Nova Scotia, Canada. Ecological Applications, 15(6), 2192-2208.  Belbin, L. (1993) Environmental representativeness: regional partitioning and reserve selection. Biological Conservation 66(3): 223–230.  Belbin, L. (1995) A multivariate approach to the selection of biological reserves. Biodiversity and Conservation 4(9): 951-963.  Belward, A., Ed. The IGBP-DIS Global 1Km Land Cover Data Set “DISCover”: Proposal and Implementation Plans. 1996. IGBP-DIS Working Paper 13. International Geosphere Biosphere Programme. European Commission Joint Research Center, Ispra, Varese, Italy.  Berg, Å., Ehnström, B., Gustafsson, L., Hallingbäck, T., Jonsell, M. and Weslien, J. (1994) Threatened plant, animal, and fungus species in Swedish forests: distribution and habitat associations. Conservation Biology 8: 718–731.  Bergerud, A.T. and Manuel, F. (1968) Moose damage to balsam fir-white birch forests in central Newfoundland. Journal of Wildlife Management 32: 729–746.  Berry, S., Mackey, B. and Brown, T. (2007). Potential applications of remotely sensed vegetation greenness to habitat analysis and the conservation of dispersive fauna. Pacific Conservation Biology 13: 12–127.  Bonn, A., Storch, D., & Gaston, K.J. (2004). Structure of the Species-Energy Relationship. Proceedings: Biological Sciences, 271: 1685–1691.  Bolton, D.K., Coops, N.C. and Wulder, M.A. (2013) Measuring forest structure along productivity gradients in the Canadian boreal with small-footprint Lidar. Environmental monitoring and assessment 185(8): 1–18.  Bradshaw, C.J.A., Warkentin, I.G., Sodhi, N.S. (2009) Urgent preservation of boreal carbon stocks and biodiversity. Trends in Ecology & Evolution 24(10): 541–548.  Brandt, J.P. (2009) The extent of the North American boreal zone. Environmental Reviews 17: 101–161.  152  Brandt, J.P., Flannigan, M.D., Maynard, D.G., and Thompson, I.D. (2013). An introduction to Canada's boreal zone: ecosystem processes , health , sustainability , and environmental issues. Environmental Reviews. 21, 207-226.  Brundtland Commission (1987) (World Commission on Environment and Development) Our Common Future. New York: Oxford University Press.  Buchanan, G.M., Nelson, A., Mayaux, P., Hartley, A. and Donald, P.F. (2008) Delivering a Global, Terrestrial, Biodiversity Observation System through Remote Sensing. Conservation Biology 23(2): 499–502.  Burkey, T.V. (1995) Extinction rates in archipelagos: implications for populations in fragmented habitats. Conservation Biology 9: 527–541.  Burton, P.J., Parisien, M-A., Hicke, J.A., Hall, R.J. and Freeburn, J.T. (2008) Large fires as agents of ecological diversity in the North American boreal forest. International Journal of Wildland Fire 17(6): 754–767. Buskirk, S.W., Powell, R.A. (1994) Habitat ecology of fishers and American martens, in: Buskirk, S.W., Harestad, A.S., Raphael, M.G., Powell, R.A. (Eds.), Martens, sables, and fishers biology and conservation. Cornell University Press, Ithaca, pp. 283–296. Cabeza, M. and Moilanen, A. (2001) Design of reserve networks and the persistence of biodiversity. Trends in Ecology & Evolution 16: 242–248.  Cairnes, A.L. and Telfer, E.S. (1980) Habitat use by four sympatric ungulates in boreal mixedwood forest. Journal of Wildlife Management 44: 849–851.  Canadian Boreal Initiative (CBI) (2011) Mandate. Available at: http://www.borealcanada.ca/, Accessed on: Jan 25, 2015.  Canadian Boreal Initiative (CBI) (2005) The Boreal in the Balance: Securing the Future of Canada’s Boreal Region, A Status Report. Canadian Boreal Initiative, Ottawa, ON. Citation in text: CBI, 2005.  Canadian Boreal Forest Agreement (CBFA) (2011) The Agreement. Available at: http://canadianborealforestagreement.com/, Accessed on: Jan. 25, 2015.  Carroll, C., Dunk, J.R. and Moilanen, A. (2010) Optimizing resiliency of reserve networks to climate change: multispecies conservation planning in the Pacific Northwest, USA. Global Change Biology 16(3): 891–904.  Carvalho, S.B., Brito, J.C., Crespo, E.G., Watts, M.E. and Possingham, H.P. (2011) Conservation planning under climate change: Toward accounting for uncertainty in predicted species distributions to increase confidence in conservation investments in space and time. Biological Conservation, 144(7): 2020–2030.  153  Carwardine, J., Klein, C.J., Wilson, K.A., Pressey, R.L. and Possingham, H.P. (2009) Hitting the target and missing the point: target-based conservation planning in context. Conservation Letters 2(1): 4–11.   Carwardine, J., Rochester, W.A., Richardson, K.S., Williams, K.J., Pressey, R.L. and Possingham, H.P. (2007) Conservation planning with irreplaceability: does the method matter? Biodiversity and Conservation 16: 245–258.  Carwardine, J., Wilson, K.A., Hajkowicz, S.A., Smith, R.J., Klein, C.J., Watts, M. and Possingham, H.P. (2010) Conservation Planning when Costs Are Uncertain. Conservation Biology 24(6): 1529–1537.  Carwardine, J., Wilson, K. A., Watts, M., Etter, A., Klein, C. J. and Possingham, H. P. (2008) Avoiding costly conservation mistakes: the importance of defining actions and costs in spatial priority setting. PLoS One 3(7): e2586.  Catford, J. A., Daehler, C. C., Murphy, H. T., Sheppard, A. W., Hardesty, B. D., Westcott, D. A., ... and Hulme, P. E. (2012). The intermediate disturbance hypothesis and plant invasions: Implications for species richness and management. Perspectives in Plant Ecology, Evolution and Systematics 14(3): 231-241.  Chen, B., Xu, G., Coops, N. C., Ciais, P., Innes, J. L., Wang, G., ... and Liu, Y. (2014). Changes in vegetation photosynthetic activity trends across the Asia–Pacific region over the last three decades. Remote Sensing of Environment 144: 28–41.  Cheng, R. and Lee, P.G. (2008) Urban Sprawl and Other Major Land Use Conversions in Ontario’s Greenbelt From 1993 to 2007: A Change Analysis Project Using Satellite Imagery. Report to the David Suzuki Foundation and the Greenbelt Foundation. Edmonton, Alberta: Global Forest Watch Canada. Available at: http://www.globalforestwatch.ca/ON-greenbelt-2008/ONgreenbelt-change-GFWC2008_LR.pdf, Accessed on: Feb. 5, 2011.  Cheng, R. and Lee, P.G. (2009) Recent (1990-2007) Anthropogenic Change within the Forest Landscapes of Nova Scotia. Edmonton, Alberta: Global Forest Watch Canada. Available at: http://www.globalforestwatch.ca/change_analysis/NS/GFWC_NS-change-2009_LR_WEB.pdf, Accessed on: Feb. 5, 2011.  Chu, C., Minns, C.K. and Mandrak, N.E. (2003) Comparing regional assessment of factors impacting freshwater fish biodiversity in Canada. Canadian Journal of Fisheries and Aquatic Sciences 60(5): 624–634.  Church, R.L., Storms, D.M. and Davis, F.W. (1996) Reserve selection as a maximal covering location problem. Biological Conservation 76: 105–112.  Coad, L., Burgess, N.D., Bomhard, B. and Besancon, C. (2009) Progress on the Convention on Biological Diversity’s 2010 and 2012 Targets for Protected Area Coverage. A technical report for the IUCN international workshop “Looking to the Future of the CBD Programme 154  of Work on Protected Areas”, Jeju Island, Republic of Korea, 14-17 September 2009. UNEP-WCMC, Cambridge UK.  Connell, J. (1978). Diversity in tropical rain forests and coral reefs. Science 199: 1302–1310  Coops, N. C., Fontana, F. M., Harvey, G. K., Nelson, T. A., & Wulder, M. A. (2014). Monitoring of a National-Scale Indirect Indicator of Biodiversity Using a Long Time-Series of Remotely Sensed Imagery. Canadian Journal of Remote Sensing 40(3): 179–191.  Coops, N.C., Waring, R.H., Wulder, M.A., Pidgeon, A.M. and Radeloff, V.C. (2009b) Bird diversity: a predictable function of satellite-derived estimates of seasonal variation in canopy light absorbance.  Coops, N.C., Wulder, M.A., Duro, D.C., Han, T. and Berry, S. (2008) The development of a Canadian dynamic habitat index using multi-temporal satellite estimates of canopy light absorbance. Ecological Indicators 8(5): 754–766.  Coops, N.C., Wulder, M.A., Iwanicka, D. (2009) An environmental domain classification of Canada using earth observation data for biodiversity assessment. Ecological Informatics 4(1): 8–22.  Coops, N.C., Wulder, M.A. and Iwanicka, D. (2009a) Demonstration of a satellite-based index to monitor habitat at continental-scales. Ecological Indicators 9(5): 948–958.  Coops, N.C., Wulder, M.A. and Iwanicka, D. (2009b) Exploring the relative importance of satellite-derived descriptors of production, topography and land cover for predicting breeding bird species richness over Ontario, Canada. Remote Sensing of Environment 113(3): 668–679. Coops, N.C., Wulder, M.A. and Iwanicka, D. (2009c) An environmental domain classification of Canada using earth observation data for biodiversity assessment. Ecological Informatics 4(1): 8–22.  Coops, N.C., Wulder, M.A. and White, J.C. (2007) Identifying and describing forest disturbance and spatial pattern: Data selection issues and methodological implications. In: Wulder MA and Franklin SE (eds) Understanding forest disturbance and spatial pattern: Remote sensing and GIS approaches. Boca Raton, Florida, USA: Taylor and Francis, 31–61.  Copeland, J.P., McKelvey, K.S., Landa, A., Persson, J., Inman, R.M., Krebs, J., et al. (2010) The bioclimatic envelope of the wolverine (Gulo gulo): do climatic constraints limit its geographic distribution? Canadian Journal of Zoology 88(3): 233–246.  Connell, J.H. and Orias, E. (1964). The ecological regulations of species diversity. The American Naturalist 98: 399–414.  155  COSEWIC, (2006) COSEWIC assessment and status report on the Rusty Blackbird Euphagus carolinus in Canada. Committee on the Status of Endangered Wildlife in Canada. Ottawa. vi + 28 pp. Available at: www.sararegistry.gc.ca/status/status_e.cfm, Accessed on: Mar. 2011.   Cracknell, A. (1997) The advanced very high resolution radiometer. London: Taylor & Francis Ltd.  Cumming, S. and Vernier, P. (2002) Statistical models of landscape pattern metrics, with applications to regional scale dynamic forest simulations, Landscape Ecology 17(5): 433–444.  Cumming, S. G., Stralberg, D., Lefevre, K. L., Sólymos, P., Bayne, E. M., Fang, S., ... & Song, S. J. (2014). Climate and vegetation hierarchically structure patterns of songbird distribution in the Canadian boreal region. Ecography, 37(2), 137-151.   Culbert, P. D., Radeloff, V. C.,  St-Louis, V.  Flather, C. H., Rittenhouse, C. D., Albright, T. P., and Pidgeon, A. M. (2012). Modelling broad-scale patterns of avian species richness across the Midwestern United States with measures of satellite image texture. Remote Sensing of Environment 118: 140–150.  Currie, D.J., Mittelbach, G.G., Cornell, H.V., Field, R., Guegan, J.F., Hawkins, B.A., Kaufman, D.M., Kerr, J.T., Oberdorff, T., O'Brien, E., & Turner, J.R.G. (2004). Predictions and tests of climate-based hypotheses of broad-scale variation in taxonomic richness. Ecology Letters 7: 1121–1134.  Currie, D.J. and Paquin, V. (1987) Large-scale biogeographical patterns of species richness of trees. Nature 329: 326–27.  Cyr, D., Gauthier, S., Bergeron, Y. and Carcaillet, C. (2009) Forest management is driving the eastern North American boreal forest outside its natural range of variability. Frontiers in Ecology and the Environment 7(10): 519–524.  Dale, V.H., Joyce, L.A., McNulty, S., Neilson, R.P., Ayres, M.P., Flannigan, M.D. and Wotton, B. M. (2001) Climate Change and Forest Disturbances Climate change can affect forests by altering the frequency, intensity, duration, and timing of fire, drought, introduced species, insect and pathogen outbreaks, hurricanes, windstorms, ice storms, or landslides. BioScience 51(9): 723–734.  De Jong, R., Verbesselt, J., Schaepman, M. E., & de Bruin, S. (2012). Trend changes in global greening and browning: Contribution of short-term trends to longer-term change. Global Change Biology 18: 642–655.  156  Dockerty, T., Lovett, A. and Watkinson, A. (2003) Climate change and nature reserves: examining the potential impacts, with examples from Great Britain. Global Environment Change 13: 125–135.  Dorner, B., Lertzman, K. and Fall, J. (2002) Landscape pattern in topographically complex landscapes: issues and techniques for analysis. Landscape Ecology 17(8): 729-743.  Drechsler, M., Lourival, R., & Possingham, H. P. (2009). Conservation planning for successional landscapes. Ecological Modelling, 220(4), 438-450.  Ducks Unlimited Canada (DUC) (2010) Hybrid Wetland Layer User Guide.   Duro, D.C., Coops, N.C., Wulder, M.A. and Han, T. (2007) Development of a large area biodiversity monitoring system driven by remote sensing. Progress in Physical Geography 31(3): 1–26.  Dye, D.G. (2002) Interannual variability and trends in the annual snow-cover cycle in Northern Hemisphere land areas, 1972-2000. Hydrological Proceeding 16(15): 3065–3077.  Ecological Stratification Working Group (1995) A National Ecological Framework for Canada. Agriculture and Agri-Food Canada, Research Branch, Centre for Land and Biological Resources Research and Environment Canada, State of the Environment Directorate, Ecozone Analysis Branch, Ottawa/Hull.  Engelmark, O. (1999) Boreal forest disturbances. In: Walker LR (ed.) Ecosystems of the world: Ecosystems of Disturbed Ground 16. Amsterdam: Elsevier Science, 161–186.  Esseen, P.A., Ehnström, B., Ericson, L. and Sjöberg, K. (1997) Boreal forests. Ecological Bulletins 46: 16–47.  Evans, K.L., Warren, P.H., & Gaston, K.J. (2005). Species-energy relationships at the macroecological scale: a review of the mechanisms. Biological Reviews 80: 1–25.  Fahrig, L. (2003). Effects of habitat fragmentation on biodiversity. Annual review of ecology, evolution, and systematics, 487-515.  Faith, D.P., Walker, P.A., Ive, J.R. and Belbin, L. (1996) Integrating conservation and forestry production: exploring trade-offs between biodiversity and production in regional land-use assessment. Forest Ecology Management 85, 251–260.  Far North Plan (FNP) (2011) About the Far North. Available at: http://www.mnr.gov.on.ca/en/Business/FarNorth/2ColumnSubPage/266506.html, Accessed on: Jan 25, 2015.   Farr, T.G. and Kobrick, M. (2000) Shuttle radar topography mission produces a wealth of data. Eos, Transactions American Geophysical Union 81(48): 583–585. 157   Ferraz, G., Russell, G.J., Stouffer, P.C., Bierregaard, R.O.J., Pimm, S.L., Lovejoy, T.E. (2003) Rates of species loss from Amazonian forest fragments. Proceedings of the National Academy of Sciences of the United States of America 100: 14069 –14073.  Ferrier, S., & Wintle, B. A. (2009). Quantitative approaches to spatial conservation prioritization: matching the solution to the need. Spatial conservation prioritization: Quantitative methods and computational tools, 1-15.  Fjeldså, J., Ehrlich, D., Lambin, E., and Prins, E. (1997). Are biodiversity 'hotspots' correlated with current ecoclimatic stability? A pilot study using the NOAA-AVHRR remote sensing data. Biodiversity and Conservation 6: 401–422.  Fisher, J.T. and Wilkinson, L. (2005) The response of mammals to forest fire and timber harvest in the North American boreal forest. Mammal Review, 35(1): 51–81.  Fleishman, E. and Mac Nally, R. (2007) Measuring the response of animals to contemporary drivers of fragmentation. Canadian Journal of Zoology 85(10): 1080–1090.  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.  Foody, G.M. and Cutler, M.E.J. (2003) Tree biodiversity in protected and logged Bornean tropical rain forests and its measurement by satellite remote sensing. Journal of Biogeography 30: 1053–66.  Fontana, F., Coops, N.C., Khlopenkov, K.V., Trishchenko, A.P., Riffler, M. and Wulder, M.A. (2012) Generation of a novel 1km NDVI data set over Canada, the northern United States, and Greenland based on historical AVHRR data. Remote Sensing of Environment 121: 171–185.  Gallant, A.L., Klaver, R.W., Casper, G.S. and Lannoo, M.J. (2007) Global Rates of Habitat Loss and Implication for Amphibian Conservation. Copeia (4): 967–979.  Game, E.T., Watts, M.E., Wooldridge, S. & Possingham, H.P. (2008) Planning for persistence in marine reserves: a question of catastrophic importance. Ecological Applications, 18, 670–680.  Game, E.T., Bode, M., McDonald-Madden, E., Grantham, H.S. & Possingham, H.P. (2009) Dynamic marine protected areas can improve the resilience of coral reef systems. Ecology Letters, 12, 1336–1346.  Game ET, Groves CR, Andersen M, Cross M, Enquist CAF, Ferdana Z, Girvetz EH, Gondor A, Hall K, Higgins J, Marshall R, Popper K, Shafer SL (2010) Incorporating climate change 158  adaptation into regional conservation assessments. The Nature Conservancy, Arlington, Virginia.  Game ET, Lipsett-Moore G, Saxon E, Peterson N, Sheppard S (2011) Incorporating climate change adaptation into national conservation assessments. Global Change Biology (17),3150–3160.  Game, E.T., Lipsett-Moore, G., Hamilton, R., Peterson, N., Kereseka, J., Atu, W., Watts, M. and Possingham, H. (2011) Informed opportunism for conservation planning in the Solomon Islands. Conservation Letters 4: 38–46.  Gaston, K.J. (2000) Global patterns in biodiversity. Nature 405: 220–227.  Gaston, K.J., Charman, K., Jackson, S.F., Armsworth, P.R., Bonn, A., Briers, R.A., ... and Tratt, R. (2006) The ecological effectiveness of protected areas: the United Kingdom. Biological Conservation 132(1): 76–87.  Gaston, K.J., Pressey, R.L. and Margules, C.R. (2002) Persistence and vulnerability: retaining biodiversity in the landscape and in protected areas. Journal of Biosciences 27 (Suppl. 2): 361–384.  Giles, P.T. (1998) Geomorphological signature: Classification of aggregated slope unit objects from digital elevation and remote sensing data. Earth Surface Processes and Landforms 23(7): 581–594.  Global Forest Watch (GFW) (2010) Canada’s Intact Forest Landscapes. Edmonton, Alberta: Global Forest Watch Canada. Available at: http://datawarehouse.globalforestwatch.ca/, Accessed on: Jun 12, 2012.  Goetz, S.J., Bunn, A.G., Fiske, G.J. and Houghton, R.A. (2005) Satellite-observed photosynthetic trends across boreal North America associated with climate and fire disturbance. Proceedings of the National Academy of Sciences of the United States of America 102(38): 13521–13525.  Government of Canada / Agriculture and Agri-Food Canada (GC/AAFC) (2009) Land Cover for Agricultural Regions of Canada, Circa 2000. Available at:  http://www.geobase.ca/geobase/en/data/landcover/index.html, Accessed on: Jan 25, 2015.  Government of Canada / Natural Resources Canada / Earth Sciences Sector / Mapping Information Branch / Centre for Topographic Information – Sherbrooke (GC/NRCan/ESS/MIB/CTI-S) (2011) CanVec (Version 5.0). Available at: ftp://ftp2.cits.rncan.gc.ca/pub/canvec/.  Goward, S.N., Cruickshanks, G.D. and Hope, A.S. (1985) Observed relation between thermal emission and reflected spectral radiance of a complex vegetated landscape. Remote Sensing of Environment 18(2): 137−146. 159   Grime, J.P. (1973) Competitive exclusion in herbaceous vegetation. Nature 242: 344–347.  Groves, C. R., Game, E. T., Anderson, M. G., Cross, M., Enquist, C., Ferdana, Z., ... & Shafer, S. L. (2012). Incorporating climate change into systematic conservation planning. Biodiversity and Conservation, 21(7), 1651-1671.  Hall, D.K. and Riggs, G.A. (2007) Accuracy assessment of the MIODIS snow products. Hydrological Processes 21(12): 1534–1547.  Hall. D.K., Riggs, G.A. and Salomonson, V.V. (2006) updated monthly. MODIS/Terra Snow Cover Monthly L3 Global 0.05Deg CMG V005, [2000-2010]. Boulder, Colorado USA: National Snow and Ice Data Center. Digital media.  Hannah, L., Midgley, G., Andelman, S., Araújo, M., Hughes, G., Martinez-Meyer, E., ... & Williams, P. (2007). Protected area needs in a changing climate. Frontiers in Ecology and the Environment, 5(3), 131-138.  Hawkins, B.A., Field, R., Cornell, H.V., Currie, D.J., Guégan, J-F., Kaufman, D.M. et al. (2003a) Energy, water and broad-scale geographic patterns of species richness. Ecology 84(12): 3105–3117.  Hawkins, B.A., Porter, E.E., & Diniz, J.A.F. (2003b). Productivity and history as predictors of the latitudinal diversity gradient of terrestrial birds. Ecology 84: 1608–1623.  Hawkins, B.A. and Porter, E.E. (2003) Relative influences of current and historical factors on mammal and bird diversity patterns in deglaciated North America. Global Ecology and Biogeography 12(6): 475–481.  Hely, C., Flannigan, M., Bergeron, Y. and McRae, D. (2001) Role of vegetation and weather on fire behavior in the Canadian mixedwood boreal forest using two fire behavior prediction systems. Canadian Journal of Forest Research 31: 430–441.   Heller NE, Zavaleta ES (2009) Biodiversity management in the face of climate change: a review of 22 years of recommendations. Biological Conservation 142:14–32  Huenneke, L.F., Jackson, R.B., Kinzig, A, Leemans, R., Lodge, D.M., Mooney, H.A., Oesterheld, M., Poff, N.L., Sykes, M.T., Walker, B.H., Walker, M. and Wall, D.H. (2000) Biodiversity - Global biodiversity scenarios for the year 2100. Science 287(5459): 1770–1774.  Hurlbert, A.H. and Haskell, J.P. (2003) The effect of energy and seasonality on avian species richness and community composition. American Naturalist 161(1): 83–97.  Hutchinson, G.E. (1961) The paradox of the plankton. American Naturalist 95: 160  137–145.  Ivits, E., Buchanan, G., Whittaker, L. and Cherlet, M. (2011) European Farmland Bird Distribution Explained by Remotely Sensed Phenological Indices. Environmental Modeling and Assessment 16: 385–399. IUCN (The World Conservation Union) (1993) Parks for life—report of the 4th world conference on natural parks and protected areas. IUCN, Gland, Switzerland.  Ji, L., Gallo, K., Eidenshink, J. C., and Dwyer, J. (2008). Agreement evaluation of AVHRR and MODIS 16‐day composite NDVI data sets. International Journal of Remote Sensing 29(16): 4839–4861.  Justice, C. O., Vermote, E., Townshend, J. R., Defries, R., Roy, D. P., Hall, D. K., and Barnsley, M. J. (1998). The Moderate Resolution Imaging Spectroradiometer (MODIS): Land remote sensing for global change research. Geoscience and Remote Sensing, IEEE Transactions on 36(4): 1228–1249.  Justice, C.O., Townshend, J.R.G., Vermote, E.F., Masuoka, E., Wolfe, R.E., Saleous, N., Roy, D.P., Morisette, J.T., 2002. An overview of MODIS Land data processing and product status. Remote Sensing of Environment 831: 3–15.  Kaufman, L. and Rousseeuw, P.J. (2005) Finding Groups in Data: An Introduction to Cluster Analysis. Hoboken, New Jersey, USA: Wiley, 368.  Kawabata, A., Ichii, K. and Yamaguchi, Y. (2001) Global monitoring of interannual changes in vegetation activities using NDVI and its relationships to temperature and precipitation. International Journal of Remote Sensing 22: 1377–1382.  Kerr, J.T. and Ostrovsky, M. (2003) From space to species: ecological applications for remote sensing. Trends in Ecology & Evolution 18(6): 299–305.  Kerr, J.T., Southwood, T.R.E. and Cihlar, J. (2001) Remotely sensed habitat diversity predicts butterfly species richness and community similarity in Canada. Proceedings of the National Academy of Sciences of the United States of America 98(20): 11365–11370.  Kimball, J.S., McDonald, K.C. and Zhao, M. (2006) Spring thaw and its effect on terrestrial vegetation productivity in the western Arctic observed from satellite microwave and optical remote sensing. Earth Interactions 10(21): 1–22.  Klein, C., Wilson, K.A., Watts, M., Stein, J., Berry, S., Carwardine, J., Stafford Smith, D.M., Mackey, B. and Possingham, H. (2009a) Incorporating ecological and evolutionary processes into continental-scale conservation planning. Ecological Applications 19: 206–217.  Klein, C., Wilson, K.A., Watts, M., Stein, J., Carwardine, J., Mackey, B. and Possingham, H. (2009b) Spatial conservation prioritization inclusive of wilderness quality: A case study of Australia’s biodiversity. Biological Conservation 142: 206–217. 161   Kupfer, J.A. (2006) National assessments of forest fragmentation in the US. Global Environmental Change 16(1): 73-82.  Kurz, W.A., Apps, M.J., Stocks, B.J. and Volney, W.J.A. (1995) Global climate change: disturbance regimes and biospheric feedbacks of temperate and boreal forests (pp. 119-133). Oxford University Press, New York.  Kurz, W.A., Apps, M.J., Webb, T.M. and McNamee, P.J. (1992) The Carbon Budget of the Canadian Forest Sector: Phase I. Information Report NOR-X-326, Forestry Canada, Northwest Region, Northern Forestry Centre, Edmonton.  Kuuluvainen, T. (2002) Disturbance Dynamics in Boreal Forests: Defining the Ecological Basis of Restoration and Management of Biodiversity. Silva Fennica 36(1): 5–11.  Latifovic, R. et al. (2008) Land cover map of Canada 2005 at 250m spatial resolution. Natural Resources Canada/ESS/Canada Centre for Remote Sensing. Available at: ftp://ftp.ccrs.nrcan.gc.ca/ad/NLCCLandCover/LandcoverCanada2005_250m/, Accessed on: Jan. 25, 2015.  Latta, G., Hailemariam, T., and Barrett, T. (2009). Mapping and imputing potential productivity of Pacific Northwest forests using climate variables. Can. Journal of Forest Research 39: 1197–1207.   Latifovic, R., Trishchenko, A., Chen, J., Park, W., Khlopenkov, K., Fernandes, R., et al. (2005) Generating historical AVHRR 1 km baseline satellite data records over Canada suitable for climate change studies. Canadian Journal of Remote Sensing 31(5): 324–346.  Leathwick, J.R., Overton, J.McC. and McLeod, M. (2003) An environmental domain classification for New Zealand and its use as a tool for biodiversity management. Conservation Biology 17(6): 1612–1623.  Lee, P., Aksenov, D., Laestadius, L., Nogueron, R. and Smith, W. (2003) Canada’s Large Intact Forest Landscapes. Edmonton, Alberta: Global Forest Watch Canada. Available at: http://pdf.wri.org/gfw_canada_lifl_text_section.pdf, Accessed on: Jan. 25, 2015.  Lee, P. and Gysbers, J.D. (2008) Recent Anthropogenic Changes within the Inland Temperate Rainforest of British Columbia: Interim Report. Edmonton, Alberta: Global Forest Watch Canada. Available at: http://www.globalforestwatch.ca/change_analysis/BC/GFWC_change-BC-INTERIM_lowres.pdf, Accessed on: Feb. 5, 2011.  Lee, P., Gysbers, J.D. and Stonojevic, Z. (2006) Canada’s Forest Landscape Fragments: a First Approximation (A Global Forest Watch Canada Report).  Edmonton, Alberta: Global Forest Watch Canada. 97pp. Available at: http://www.globalforestwatch.ca/FLFs/GFWC-FLFs-firstapprox-150dpi.pdf, Accessed on: Feb. 5, 2011. 162   Lee, P.G., Hanneman, M., Gysbers, J.D., Cheng, R. and Smith, W. (2010) Atlas of Canada’s Intact Forest Landscapes. Edmonton, Alberta: Global Forest Watch Canada 10th Anniversary Publication #1. 74 pp. Available at: http://globalforestwatch.ca/pubs/2010Anniversary/01CanadaAtlasIFL/01CanadaAtlas_IFL_lowres.pdf, Accessed on: Mar. 19, 2012.  Lehman, C. L., & Tilman, D. (2000). Biodiversity, stability, and productivity in competitive communities. The American Naturalist, 156(5), 534-552.  Lemieux, C.J., Beechey, T.J. and Gray, P.A. (2011) Prospects for Canada's protected areas in an era of rapid climate change. Land Use Policy 28(4): 928–941.  Lemieux, C. J., Beechey, T. J., Scott, D. J. and Gray, P. A. (2011b) The state of climate change adaptation in Canada's protected areas sector. The Canadian Geographer/Le Géographe canadien 55(3): 301–317.  Leroux, S.J., Schmiegelow, F.K.A., Cumming, S.G., Lessard, R.B. and Nagy, J. (2007b) Accounting for system dynamics in reserve design. Ecological Applications 17: 1954–1966.  Leroux, S.J., Schmiegelow, F.K.A., Lessard, R.B. and Cumming, S.G. (2007a) Minimum dynamic reserves: A framework for determining reserve size in ecosystems structured by large disturbances. Biological Conservation 138: 464–473.  Leroux, S. J., & Rayfield, B. (2014). Methods and tools for addressing natural disturbance dynamics in conservation planning for wilderness areas. Diversity and distributions, 20(3), 258-271.  Li, J.H. and Chen, W.J. (2005) A rule-based method for mapping Canada’s wetlands using optical, radar and DEM data. International Journal of Remote Sensing 26(22): 5051–5069.  Li, Z., Nadon, S. and Cihlar, J. (2000) Satellite detection of Canadian boreal forest fires: Development and application of the algorithm. International Journal of Remote Sensing 21(16): 3057−3069.  Li, X., He, H.S., Bu, R., Wen, Q., Chang, Y., Hu, Y. et al. (2005) The adequacy of different landscape metrics for various landscape patterns. Pattern Recognition 38(12): 2626–2638.  Linke, J., McDermid, G.J., Laskin, D.N., McLane, A.J., Pape, A., Cranston, J. et al. (2009) A Disturbance-Inventory Framework for Flexible and Reliable Landscape Monitoring. Photogrammetric Engineering and Remote Sensing 75(8): 981–995.  Linke, S., Pressey, R.L., Bailey, R.C. and Norris, R.H. (2007) Management options for river conservation planning: Condition and conservation re-visited. Freshwater Biology 52(5): 918–938.  163  Little, L.Jr. (1999) Digital representation of Atlas of United States Trees. Reston, Virginia: U.S. Geological Survey. Available at: http://esp.cr.usgs.gov/data/atlas/little/, Accessed on: Jan 25, 2015.  Lourival, R., Drechsler, M., Watts, M. E., Game, E. T., & Possingham, H. P. (2011). Planning for reserve adequacy in dynamic landscapes; maximizing future representation of vegetation communities under flood disturbance in the Pantanal wetland. Diversity and Distributions, 17(2), 297-310.  Lovejoy, T.E., Hannah, L. (Eds.), 2005. Climate Change and Biodiversity. Yale University Press, New Haven.  Lugo, A.E., Brown, S.L., Dodson, R., Smith, T.S. and Shugart, H.H. (1999) The Holdridge life zones of the conterminous United States in relation to ecosystem mapping. Journal of Biogeography 26(5): 1025–1038.  Mackey, B., Berry, S. and Brown, T. (2008) Reconciling approaches to biogeographic regionalization: a systematic and generic framework examined with a case study of the Australian continent. Journal of Biogeography 35: 213–229.  Mackey, B.G., Bryan, J. and Randall, L. (2004) Australia’s Dynamic Habitat Template 2003. In: Proceedings MODIS Vegetation Workshop II, University of Montana.  Mackey, B.G., Nix, H.A., Hutchinson, M.F., MacMahon, J.P. and Fleming, P.M. (1988) Assessing representativeness of places for conservation reservation and heritage listing. Environmental Management 12: 501–514.  Margules, C.R. and Pressey, R.L. (2000) Systematic conservation planning. Nature 405: 243–253.  Margules, C., Sarkar, S., & Margules, C. R. (2007). Systematic conservation planning. Cambridge University Press.  Mackey, B.G., Berry, S.L. and Brown, T. (2008) Reconciling approaches to biogeographical regionalization: a systematic and generic framework examined with a case study of the Australian continent. Journal of Biogeography 35: 213–229.  Mackey, B.G., Nix, H.A., Hutchinson, M.F., Macmahon, J.P. and Fleming, P.M. (1988) Assessing representativeness of places for conservation reservation and heritage listing. Environmental Management 12(4): 502–514.  MacMillan, R.A., Jones, R.K. and McNabb, D.H. (2004) Defining a hierarchy of spatial entities for environmental analysis and modeling using digital elevation models (DEMs). Computers, Environment and Urban Systems 28(3): 175-200.  164  Margules, C.R. and Pressey, R.L. (2000) Systematic conservation planning. Nature 405: 243–253.  McGarigal, K., Cushman, S.A., Neel, M.C. and Ene, E. (2002) FRAGSTATS: Spatial Pattern Analysis Program for Categorical Maps. Massachusetts, Amherst: University of Massachusetts, Amherst. Available at: http://www.umass.edu/landeco/research/fragstats/fragstats.html, Accessed on: Jan. 25, 2015.  McGarigal, K. and Marks, B.J. (1995) FRAGSTATS: spatial pattern analysis program for quantifying landscape structure. Portland, Oregon: U.S. Department of Agriculture, Forest Service, Pacific Northwest Research Station. Available at: http://www.fs.fed.us/pnw/pubs/pnw_gtr351.pdf, Accessed on: Jan. 25, 2015.  Metzger, M.J., Bunce, R.G.H., Jongman, C.A., et al. (2005) A climate stratification of the environment of Europe. Global Ecology and Biogeography 14: 549–563.  Michaud, J.S., Coops, N.C., Andrew, M.A., Wulder, M.A., Brown, G.S., Rickbeil, G.J.M. (2014). Estimating moose (Alces alces) occurrence and abundance from satellite-derived environmental indicators. Remote Sensing of Environment DOI: 10.1016/j.rse.2014.06.005.  Mildrexler, D.J., Zhao, M. and Running, S.W. (2009) Testing a MODIS Global Disturbance Index across North America. Remote Sensing of Environment 113(10): 2103–2117.  Mittelbach, G.G., Steiner, C.F., Scheiner, S.M., Gross, K.L., Reynolds, H.L., Waide, R.B., Willig, M.R., Dodson, S.I., & Gough, L. (2001). What is the observed relationship between species richness and productivity? Ecology 82: 2381–2396.  Moilanen, A. and Cabeza, M. (2002) Single-species dynamic site selection. Ecological Applications 12: 913–926.  Moilanen, A., Runge, M. C., Elith, J., Tyre, A., Carmel, Y., Fegraus, E., ... & Ben-Haim, Y. (2006). Planning for robust reserve networks using uncertainty analysis. Ecological Modelling, 199(1), 115-124.  Mooi, E. and Sarstedt, M. (2011) A Concise Guide to Market Research: The Process, Data, and Methods Using IBM SPSS Statistics. Heidelberg, Germany: Springer.  Morneau, C. and Payette, S. (1989) Postfire lichen-spruce woodland recovery at the limit of the boreal forest in northern Quebec. Canadian Journal of Botany 67: 2770–2782.  Murcia, C. (1995) Edge effects in fragmented forest: implications for conservation. Trends in Ecology & Evolution 10: 58–62.  Myneni, R., Keeling, C., Tucker, C., Asrar, G. and Nemani, N. (1997) Increased plant growth in the northern high latitudes from 1981 to 1991. Nature 386: pp. 698–702.  165  Myneni, R. B., Hall, F. G., Sellers, P. J., and Marshak, A.L. (1995). The interpretation of spectral vegetation indexes. IEEE Transactions on Geoscience and Remote Sensing 33: 481–486.  Naidoo, R., Balmford, A., Ferraro, P.J., Polasky, S., Ricketts, T.H. and Rouget, M. (2006) Integrating economic costs into conservation planning. Trends in Ecology & Evolution 21: 681–687.  Nelson, T. A., Coops, N. C., Wulder, M. A., Perez, L., Fitterer, J., Powers, R., & Fontana, F. (2014). Predicting Climate Change Impacts to the Canadian Boreal Forest. Diversity, 6(1), 133-157.  NOAA (2011) DMSP-OLS Nighttime Lights Time Series, Version 2. Image and Data Processing by NOAA’s National Geophysical Data Center, DMSP Data Collected by US Air Force Weather Agency, Boulder, CO.  NASA Land Processes Distributed Active Archive Center (LP DAAC). (2010) MODIS 12Q1. USGS/Earth Resources Observation and Science (EROS) Center, Sioux Falls, South Dakota. Citation in text: MODI12Q1.  NASA Land Processes Distributed Active Archive Center (LP DAAC). (2011) MODIS 17A3. USGS/Earth Resources Observation and Science (EROS) Center, Sioux Falls, South Dakota.  Citation in text: MODI17A3  Natural Resources Canada (NRCan) (2011) Birds research in the boreal forest. Available at: http://cfs.nrcan.gc.ca/pages/254, Accessed on: Mar. 16, 2012.  Nemani, R. R., Keeling, C. D., Hashimoto, H., Jolly, W. M., Piper, S.C., Tucker, C. J., et al. (2003). Climate-driven increases in global terrestrial net primary production from 1982 to 1999. Science 300: 1560–1563.  Nightingale, J.M., Fan, W., Coops, N.C. and Waring, R.H. (2008) Predicting Tree Diversity across the United States as a Function of Modeled Gross Primary Production. Ecological Applications 18(1): 93–103.  Olthof, I. and Pouliot, D. (2010) Treeline vegetation composition and change in Canada’s western Subarctic from AVHRR and canopy reflectance modeling. Remote Sensing of Environment 114(4): 805–815.  Onal, H. and Briers, R.A. (2006) Optimal selection of a connected reserve network. Operations Research 54(2): 379–388.  Parmesan, C. and Yohe, G. (2003) A globally coherent fingerprint of climate change impacts across natural systems. Nature 421: 37–42.  166  Patterson, B.D., Ceballos, G., Sechrest, W., Tognelli, M.F., Brooks, T., Luna, L. et al. (2007) Digital Distribution Maps of the Mammals of the Western Hemisphere, version 3.0. Arlington, Virginia: NatureServe. Available at: http://www.natureserve.org/getData/mammalMaps.jsp, Accessed on: Feb. 5, 2011.  Pearson, S.M., Turner, M.G., Wallace, L.I. and Romme, W.H. (1995) Winter habitat use by large ungulates following fire in northern Yellowstone National Park. Ecological Applications 5: 744–755.  Pettorelli, N., Vik, J.O., Mysterud, A., Gaillard, J.M., Tucker, C.J. and Stenseth, N.C. (2005) Using satellite-derived NDVI to assess ecological responses to environmental change. Trends in Ecology & Evolution 20(9): 503–510.  Pickett, S.T.A. and Thompson, J.N. (1978) Patch dynamics and the design of nature reserves. Biological Conservation 13: 27–37.  Plan Nord (2011) Protection of the Environment. Available at: http://www.plannord.gouv.qc.ca/english/environment/index.asp, Accessed on: Nov. 16, 2012. Citation in text: Plan Nord, 2011.  Plummer, D. A., Caya, D., Frigon, A., Côté, H., Giguère, M., Paquin, D., ... & De Elia, R. (2006). Climate and climate change over North America as simulated by the Canadian RCM. Journal of Climate, 19(13), 3112-3132.  Possingham, H., Ball, I. and Andelman, S. (2000) Mathematical methods for identifying representative reserve networks, in: Ferson, S., Burgman, M. (Eds.), Quantitative methods for conservation biology. Springer-Verlag, pp. 291–305.  Pouliot, D., Latifovic, R., and Olthof, I. (2009). Trends in vegetation NDVI from 1 km AVHRR data over Canada for the period 1985–2006. International Journal of Remote Sensing 30(1): 149–168.  Powers., R.P., Coops, N.C., Morgan, J.L., Wulder, M.A., Nelson, T.A., Drever, C.R. and Cumming, S.G. (2013a) A remote sensing approach to biodiversity assessment and regionalization of the Canadian boreal forest. Progress in Physical Geography 37(1): 36–62.  Powers, R.P., Coops, N.C., Nelson, T., Wulder, M.A. and Ronnie Drever, C. (2013b) Integrating accessibility and intactness into large-area conservation planning in the Canadian boreal forest. Biological Conservation 167: 371–379.  Powers, R. P., Hermosilla, T., Coops, N. C., and Chen, G. (2015). Remote sensing and object-based techniques for mapping fine-scale industrial disturbances. International Journal of Applied Earth Observation and Geoinformation, 34, 51-57.  167  Powers, R.P., Coops, N.C., Nelson, T.A. and Wulder M.A. (in review). An evaluation of large-scale reserve design efficacy using a long Time-Series Earth observation dataset. Remote Sensing of Environment.    Pressey, R.L., Cabeza, M., Watts, M.E., Cowling, R.M. and Wilson, K.A. (2007) Conservation planning in a changing world. Trends in Ecology & Evolution 22(11): 583–592.  Price, D.T., Alfaro, R.I., Brown, K.J., Flannigan, M.D., Fleming, R.A., Hogg, E.H., ... and Venier, L.A. (2013) Anticipating the consequences of climate change for Canada’s boreal forest ecosystems. Environmental Reviews 21: 322–365.  Pulliainen, J. (2006) Mapping of snow water equivalent and snow depth in boreal and sub-arctic zones by assimilating space-borne microwave radiometer data and ground-based observations. Remote Sensing of Environment 101(2): 257–269.  Rayfield, B., James, P., Fall, A. and Fortin, M. J. (2008) Comparing static versus dynamic protected areas in the Quebec boreal forest. Biological Conservation 141(2): 438–449.  Richardson, E.A., Kaiser, M.J., Edwards-Jones, G. and Possingham, H.P. (2006) Sensitivity of marine-reserve design to the spatial resolution of socioeconomic data. Conservation Biology 20: 1191–1202.  Ridgely, R.S., Allnutt, T.F., Brooks, T., McNicol, D.K., Mehlman, D.W., Young. B.E. et al. (2007) Digital Distribution Maps of the Birds of the Western Hemisphere, version 3.0. Arlington, Virginia: NatureServe. Available at: http://www.natureserve.org/getData/birdMaps.jsp, Accessed on: Feb. 5, 2011.  Riggs, G.A., Hall, D.K. and Salomonson, V.V. (2006) MODIS Snow Products User Guide Collection 5. Available at: http://nsidc.org/data/docs/daac/modis_v5/dorothy_snow_doc.pdf, Accessed on: Jan. 25, 2015.  Riitters, K.H., Wickham, J.D., O’Neill, R.V., Jones, K.B., Smith, E.R., Coulston, J.W. et al. (2002) Fragmentation of continental United States Forests. Ecosystems 5: 815-822.  Rocchini, D., Ricotta, C. and Chiarucci, A. (2007) Using satellite imagery to assess plant species richness: the role of multispectral systems. Applied Vegetation Science 10(3): 325–331.  Rodrigues, A.S., Gregory, R.D. and Gaston, K.J. (2000) Robustness of reserve selection procedures under temporal species turnover. Proceedings of the Royal Society of London. Series B: Biological Sciences 267(1438): 49–55.  Rouse, J.W., Haas, R.H., Schell, J.A. and Deering, D.W. (1973) Monitoring vegetation systems in the Great Plains with ERTS. Proc. Third Earth Resources Technology Satellite-1 Symposium, Goddard Space Flight Center, NASA SP-351, Science and Technical Information Office, NASA, Washington, DC, pp. 309–317. 168   Sala, O.E., Chapin, F.S., Armesto, J.J., Berlow, E., Bloomfield, J., Dirzo, R., Huber-Sanwald, E., Rowhani, P., C. A. Lepczyck, M. Linderman, A. M. Pidgeon, V. C. Radeloff, P. Culbert, and E. Lambin. 2008. Variability in energy influences avian distribution patterns across the U.S.A. Ecosystems 11(6): 854–867.  Sanderson, E.W., Jaiteh, M., Levy, M.A., Redford, K.H., Wannebo, A.V. and Woolmer, G. (2002) The human footprint and the last of the wild. Bioscience 52: 891–904.  Sarr, D.A., Hibbs, D.E. and Huston, M.A. (2005) A hierarchical perspective of plant diversity. Quarterly Review of Biology 80(2): 187–212.  Saxon E, Baker B, Hargrove W, Hofman F, Zganjar C (2005) Mapping environments at risk under different global climate change scenarios. Ecology Letters, 8, 53–60.  Schmitt, C.B., Burgess, N.D., Coad, L., Belokurov, A., Besançon, C., Boisrobert, L., Campbell, A., Fish, L., Gliddon, D., Humphries, K., Kapos, V., Loucks, C., Lysenko, I., Miles, L., Mills, C., Minnemeyer, S., Pistorius, T., Ravilious, C., Steininger, M. and Winkel, G. (2009) Global analysis of the protection status of the world’s forests. Biological Conservation 142: 2122–2130.  Schneider, R., Dyer, S., 2009. Death by a Thousand Cuts: The Impacts of in situOil Sands Development on Alberta’s Boreal Forest. The Pembina Institute forAppropriate Development.  Schneider, R.R., Hauer, G., Farr, D., Adamowicz, W.L. and Boutin, S. (2011) Achieving Conservation when Opportunity Costs Are High: Optimizing Reserve Design in Alberta’s Oil Sands Region. PLoS ONE 6(8): e23254.  Schneider, R.R., Stelfox, J.B., Boutin, S. and Wasel, S. (2003) Managing the Cumulative Impacts of Land Uses in the Western Canadian Sedimentary Basin: A Modeling Approach. Conservation Ecology 7(1): 8–18.  Schwartz, C.C., Haroldson, M.A., White, G.C., Harris, R.B., Cherry, S., Keating, K.A. et al. (2006) Temporal, spatial and environmental influences on the demographics of grizzly bears in the Greater Yellowstone Ecosystem. Wildlife Monographs 161(1): 1–68.  Scott, D., & Lemieux, C. (2005). Climate change and protected area policy and planning in Canada. The Forestry Chronicle, 81(5), 696-703. Statistics Canada, 2008. 2008 Road Network File, Reference Guide 92-500-GWE. Statistics Canada, Ottawa, ON.  Scott, J.M., Davis, F.W., McGhie, R.G., Wright, R.G., Groves, C. and Estes, J. (2001) Nature reserves: Do they capture the full range of America’s biological diversity? Ecological Applications 11: 999–1007.  169  Severson-Baker, C., 2004. Environment and Energy in the North Primer Series. The Pembina Institute for Appropriate Development.  Shea, K., Roxburgh, S.H., and Rauschert, E.S.J. (2004). Moving from pattern to process: coexistence mechanisms under intermediate disturbance regimes. Ecology Letters 7: 491–508.  Slayback, D., Pinzon, J., Los, S. and Tucker, C. (2003) Northern hemisphere photosynthetic trends 1982–99. Global Change Biology 9: 1–15.  Smith, K.B., Smith, C.E., Forest, S.F. and Richard, A.J. (2007) A field guide to the wetlands of the boreal plains ecozone of Canada. Ducks Unlimited Canada, Western Boreal Office: Edmonton, Alberta.  Soulé, M.E., Mackey, B., Recher, H., Williams, J., Woinarski, J., Driscoll, D., Dennison, W. and Jones, M. (2004) The role of connectivity in Australian conservation. Pacific Conservation Biology 10: 266–279.  Soulé, M.E. and Sanjayan, M.A. (1998) Conservation targets: do they help? Science 279: 2060–2061.  Soverel, N., Coops, N.C., White, J.C. and Wulder, M.A. (2010) Characterizing the forest fragmentation of Canada’s national parks. Environmental Monitoring and Assessment 164: 481–499.  SPSS (2001) The SPSS Two Step Cluster Component. A scalable component enabling more efficient customer segmentation. USA: SPSS Incorporation. Available at: http://www.spss.ch/upload/1122644952_The%20SPSS%20TwoStep%20Cluster%20Component.pdf, Accessed on: Jan. 25, 2015.  Stanojevic, Z., Lee, P. and Gysbers, J.D. (2006a) Recent Anthropogenic Changes within the Boreal Plains Ecozone of Saskatchewan and Manitoba: Interim Report. Edmonton, Alberta: Global Forest Watch Canada. Available at: http://www.globalforestwatch.ca/change_analysis/change-analysis-SKMB-150dpi.pdf, Accessed on: Nov. 16, 2012.  Stanojevic, Z., Lee, P. and Gysbers, J.D. (2006b) Recent Anthropogenic Changes within the Northern Boreal, Southern Taiga, and Hudson Plains Ecozones of Québec: A Global Forest Watch Canada Report. Edmonton, Alberta: Global Forest Watch Canada. Available at: http://www.globalforestwatch.ca/change_analysis/change-analysis-QC-150dpi.pdf, Accessed on: Mar. 22, 2012.  Statistics Canada, 2008. 2008 Road Network File, Reference Guide 92-500-GWE. Statistics Canada, Ottawa, ON.  170  Stewart, R. and Possingham, H. (2005) Efficiency, costs and trade-offs in marine reserve system design. Environmental Modeling & Assessment 10: 203–213.  Stocks, B.J., Mason, J.A., Todd, J.B., Bosch, E.M., Wotton, B.M., Amiro, B.D., Flannigan,  M.D., Hirsch, K.G., Logan, K.A., Martell, D.L. and Skinner, W.R. (2002) Large forest fires in Canada, 1959–1997. Journal of Geophysical Research: Atmospheres 107(D1): 8149.  Stocks, B.J., Mason, J.A., Todd, J.B., et al. (2003) Large forest fires in Canada, 1959-1997. Journal of Geophysical Research: Atmospheres 108:1–12.  St-Louis, V., Pidgeon, A.M., Radeloff, V.C., Hawbaker, T.J., & Clayton, M.K. (2006). Highresolution image texture as a predictor of bird species richness. Remote Sensing of Environment 105: 299–312.  Storch, D., Evans, K.L., & Gaston, K.J. (2005). The species-area-energy relationship. Ecology Letters 8: 487–492.  Swanson, F.J., Johnson, S.L., Gregory, S.V. and Acker, S.A. (1998) Flood disturbance in a forested mountain landscape. BioScience 48(9): 681–689.  Tarnocai, C., Kettles, I.M. and Lacelle, B. (2000) Peatlands of Canada. Open File 3834. Ottawa, Ontario: Geological Survey of Canada.  Tateishi, R. and Ebata, M. (2004) Analysis of phenological change patterns using 1982–2000 Advanced Very High Resolution Radiometer (AVHRR) data. International Journal of Remote Sensing, 25, pp. 2287–2300.  Temini, M., Leconte, R., Chaouch, N., Sukumal, P., Khanbilvardi, R. and Brissette, F. (2010) A combination of remote sensing data and topographic attributes for the spatial and temporal monitoring of soil wetness. Journal of Hydrology 388: 28–40.  Thompson, I.D. (1991) Could marten become the spotted owl of eastern Canada? The Forestry Chronicle 67: 136–140.  Tilman, D. (1999). The ecological consequences of changes in biodiversity: A search for general principles 101. Ecology, 80(5), 1455-1474.  Trakhtenbrot, A. and Kadmon, R. (2005) Environmental cluster analysis as a tool for selecting complementary networks of conservation sites. Ecological Applications 15(1): 335–345.  Tucker, C.J.,Slayback, D.A., Pinzon, J.E., Los, S.O., Myneni, R.B., and Taylor, M.G. (2001). Higher northern latitude normalized difference vegetation index and growing season trends from 1982 to 1999. International Journal of Biometeorology 45: 184–190.  171  Tulloch, V. J., H. P. Possingham, S. D. Jupiter, C. Roelfsema, A. I. T. Tulloch, and C. J. Klein. 2013. Incorporating uncertainty associated with habitat data in marine reserve design. Biological Conservation 162:41-51.  Turetsky, M.R. and Louis, S.T. (2006) Disturbance in boreal peatlands. In: Wieder RK and Vitt DH (eds) Boreal Peatland Ecosystems. Berlin, Heidelberg: Springer-Verlag, 359–379.  Turner, M.G., Gardner, R.H. and O’Neill, R.V. (2001) Landscape Ecology in Theory and Practice: Pattern and Process. New York: Springer-Verlag.  Turner, W., Spector, S., Gardiner, N., Fladeland, M., Sterling, E. and Steininger, M. (2003) Remote sensing for biodiversity science and conservation. Trends in Ecology & Evolution 18(6): 306–314.  U.S. Geological Survey (USGS) (2011) Land Cover Type Yearly L3 Global 1km SIN Grid. Available at: https://lpdaac.usgs.gov, Accessed on: Jan. 25, 2015. van Leeuwen, W.J., Orr, B.J., Marsh, S.E. and Herrmann, S.M. (2006). Multi-sensor NDVI data continuity: Uncertainties and implications for vegetation monitoring applications. Remote Sensing of Environment 100(1): 67–81.  Waide, R.B., Willig, M.R., Steiner, C.F., Mittelbach, G., Gough, L., Dodson, S.I., Juday, G.P., Parmenter, R. (1999). The relationship between productivity and species richness. Annual Review of Ecology, Evolution, and Systematics 30: 257–300.   Walker, R.E., Stoms, D.M., Etes, J.E. and Cayocca, K.D. (1992) Relationships between biological diversity and multitemporal vegetation index data in California. Albuquerque, New Mexico, 3-8 March 1992. Bethesda: ASPRS (American Society for Photogrammetry and Remote Sensing) Pubns.  Warner, B.G. and Asada, T. (2006) Biological diversity of peatlands in Canada, Aquatic Sciences 68(3): 240–253.  Waring, R.H., Coops, N.C., Fan, W., & Nightingale, J.M. (2006). MODIS enhanced vegetation index predicts tree species richness across forested ecoregions in the contiguous USA. Remote Sensing of Environment 103: 218–226.  Warren, A.J. and Collins, M.J. (2007) A pixel-based semi-empirical system for predicting vegetation diversity in boreal forest. International Journal of Remote Sensing 28(1): 83–105.  White, P.S. (1987) Natural disturbance, patch dynamics, and landscape pattern in natural areas. Natural Areas Journal 7: 14–22.  Wiersma, Y.F., Beechey, T.J., Oosenbrug, B.M., and Meikle, J.C. 2005. Protected areas in northern Canada: Designing for ecological integrity. Phase I Report. CCEA Occasional Paper No. 16. Canadian Council on Ecological Areas. CCEA Secretariat, Ottawa, Canada.  172  WMO. (2006). Systematic observation requirements for satellite-based products for climate: Supplemental details to the satellite-based component of the “Implementation Plan for the Global Observing System for Climate in support of the UNFCC”. GCOS-107, WMO/TD-1338. Available at: http://www.wmo.int/pages/prog/gcos/Publications/gcos-107.pdf, Accessed on: Jan. 25, 2015.  WMO (2011). Systematic Observation Requirements for Satellite-based Products for Climate Supplemental details to the satellite-based component of the Implementation Plan for the Global Observing System for Climate in Support of the UNFCCC - 2011 Update. Available at: http://www.wmo.int/pages/prog/gcos/Publications/gcos-154.pdf, Accessed on: Jan. 25, 2015.  Wulder, M.A., Campbell, C., White, J.C., Flannigan, M. and Campbell, I.D. (2007) National circumstances in the international circumboreal community. The Forestry Chronicle 83(4): 539–556.  Wulder, M.A., Campbell, C., White, J.C., et al. (2007) National Context in the International Circumboreal Community. The Forestry Chronicle 83(4): 539–556.  Wulder, M.A. and Franklin, S.E. (eds) (2007) Understanding forest disturbance and spatial pattern: Remote sensing and GIS approaches. Boca Raton, Florida, USA: Taylor and Francis.  Wulder, M.A., White, J.C., Andrew, M.E., Seitz, N.E. and Coops, N.C. (2009) Forest fragmentation, structure, and age characteristics as a legacy of forest management. Forest Ecology and Management 258(9): 1938–1949.  Wulder, M.A., White, J.C. and Coops, N.C. (2011) Fragmentation regimes of Canada’s forests. The Canadian Geographer 55(3): 288–300.  Wulder, M.A., White, J.C., Cranny, M., Hall, R.J., Luther, J.E., Beaudoin, A., Goodenough, D.G. and Dechka, J.A. (2008) Monitoring Canada's forests. Part 1: Completion of the EOSD land cover project. Canadian Journal of Remote Sensing 34(6): 549–562.  Wulder, M.A., White, J.C., Han, T., Coops, N.C., Cardille, J.A., Holland, T. et al. (2008b) Monitoring Canada’s forests. Part 2: National forest fragmentation and pattern. Canadian Journal of Remote Sensing 34(6): 563–584.  Wulder, M.A., White, J.C., Magnussen, S. and McDonald, S. (2007) Validation of a Large Area Land Cover Product Using Purpose-Acquired Airborne Video. Remote Sensing of Environment 106(4): 480–491.  Zhang, K., Kimball, J.S., Mu, Q., Jones, L.A., Goetz, S.J. and Running, S.W. (2009) Satellite based analysis of northern ET trends and associated changes in the regional water balance from 1983 to 2005. Journal of Hydrology 379(1-2): 92–110.  173  Zhang, Q., Pavlic, G., Chen, W., Latifovic, R., Fraser, R. and Cihlar, J. (2004) Deriving stand age distribution in boreal forest using SPOT VEGETATION and NOAA AVHRR imagery. Remote Sensing of Environment 91: 405–418.  Zhang, T., Ramakrishnan, R. and Linvy, M. (1996) BIRCH: An Efficient Data Clustering Method for Very Large Databases. In: Jagadish HV and Inderpal Singh Mumick (eds), Proceedings of the 1996 ACM SIGMOD International Conference on Management of Data, Montreal, Canada, 4-6 June 1996, New York: ACM Press.  Zhao, M. and Running, S.W. (2010) Drought-induced reduction in global terrestrial net primary production from 2000 through 2009. Science 329: 940–943.  Zimmermann, N.E., Yoccoz, N.G., Edwards, T.C., Meier, E.S, Thuiller, W., Guisan, A., Schmatz, D.R. and Pearman, P.B. (2009) Climatic extremes improve predictions of spatial patterns in tree species. Proceedings of the National Academy of Sciences of the United States of America 106: 19723–19728.  Zhou, L., Tucker, C. J., Kaufmann, R. K., Slayback, D. A., Shabanov, N. V., and Myneni, R. (2001). Variations in northern vegetation activity inferred from satellite data of vegetation index during 1981 to 1999. Journal of Geophysical Research 106: 20069–20083.                  174  Appendices Appendix A Distributions maps for 16 at-risk-species as shown in Table 2.2.          175         176               177  Appendix B In Chapter 4 we devised a number of exploratory planning scenarios to determine potential (cost effective) reserve designs. Specifically, the exploratory reserve design scenarios were used to examine alternative compactness (BLM) levels when different cost objectives (area and accessibility) and area restrictions (intact areas only) are used. The below table summarizes the different scenario constraints, with results for the best reserve design solutions (see Chapter 4 for more details). Here T15, T25, and T35 represent the 15 %, 25 %, and 35 % targets respectively.       In Chapter 6 we used MarProb in three exploratory planning scenarios based on alternate climate change scenarios. Here we examined alternative compactness (BML: 0-10), certainty targets (0.5-0.95), and SPF (0-5) when different projected productivity variability based on alternate climate change conditions (B1, A1B, and A2) are used. Cost Objectives Scenario  Objective 1, area  Objective 2, access  Objective 3, intact   BLM Relative Cost  Planning Units  BLM Relative Cost Planning Units  BLM Relative Cost  Planning Units T15  0.57 0.14 26675  0.57 0.11 26969  0.57 0.12 26460 T15  3.40 0.14 27749  3.40 0.14 30009  3.40 0.16 31779 T15  7.00 0.19 33129  7.00 0.18 34538  7.00 0.22 41630 T25  0.57 0.24 46358  0.57 0.21 47555  0.57 0.23 48893 T25  3.40 0.25 47820  3.40 0.24 47698  3.40 0.27 52824 T25  7.00 0.27 52415  7.00 0.28 55362  7.00 0.31 59897 T35  0.57 0.34 66300  0.57 0.31 67448  0.57 0.37 74388 T35  3.40 0.37 69870  3.40 0.34 68478  3.40 0.48 94661 T35  7.00 0.37 69752  7.00 0.36 71638  7.00 0.53 101293 

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.24.1-0167183/manifest

Comment

Related Items