Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Identifying priority conservation areas using systematic reserve selection and GIS at a fine spatial… Warman, Leanna Dawn 2002

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

Item Metadata

Download

Media
831-ubc_2002-0280.pdf [ 22.75MB ]
Metadata
JSON: 831-1.0090312.json
JSON-LD: 831-1.0090312-ld.json
RDF/XML (Pretty): 831-1.0090312-rdf.xml
RDF/JSON: 831-1.0090312-rdf.json
Turtle: 831-1.0090312-turtle.txt
N-Triples: 831-1.0090312-rdf-ntriples.txt
Original Record: 831-1.0090312-source.json
Full Text
831-1.0090312-fulltext.txt
Citation
831-1.0090312.ris

Full Text

IDENTIFYING PRIORITY C O N S E R V A T I O N A R E A S USING S Y S T E M A T I C R E S E R V E SELECTION A N D GIS A T A FINE SPATIAL S C A L E : A TEST C A S E USING T H R E A T E N E D V E R T E B R A T E SPECIES IN THE O K A N A G A N , BRITISH C O L U M B I A by Leanna Dawn Warman B.Sc, University of British Columbia, 1994  A THESIS SUBMITTED IN P A R T I A L F U L F I L L M E N T OF THE REQUIREMENTS FOR THE D E G R E E OF M A S T E R OF SCIENCE in THE F A C U L T Y OF G R A D U A T E STUDIES Department of Zoology  We accept this thesis as conforming to the required standard.  THE UNIVERSITY OF BRITISH C O L U M B I A November 2001 © Leanna D. Warman, 2001  In presenting this thesis in partial fulfilment of the requirements for an advanced degree at the University of British Columbia, I agree that the Library shall make it freely available for reference and study. I further agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the head of my department or by his or her representatives. It is understood that copying or publication of this thesis for financial gain shall not be allowed without my written permission.  Department of Zoology The University of British Columbia Vancouver, Canada  Date: 15 November 2001  ABSTRACT Biologists and wildlife managers recognise the need for systematic reserve selection techniques to conserve habitat for species. Geographic information systems (GIS) provide a tool that helps identify conservation areas using geographically referenced data. Mapping continuous geographical phenomena with discreet boundaries affects the spatial organisation of data. However, most published studies of systematic reserve selection techniques have been completed at only one level of data organisation and usually for large regions at coarse scales. In this thesis, I examined the effects of data organisation on reserve selection in the South Okanagan, a small region in British Columbia. I used the software program "C-Plan" with ArcView GIS to identify the minimum amount of area required to achieve explicit conservation targets that maintain species within the region. I evaluated the reserve selection technique using terrestrial ecosystem mapping (TEM) and species habitat models that predict the suitability of T E M polygons for each of twenty-nine threatened vertebrate species. C-Plan selected 37.2% of the region to represent habitat that maintains current population sizes of these threatened vertebrate species. Although habitat area targets were achieved, these priority sites were small and scattered throughout the region and were therefore not practical for implementation or viable for many species. I examined the effects of data organisation on priority site selection by altering three algorithm parameters: (1) size and shape of the unit used to map data and select sites, (2) type of species included in selections, and (3) quantity of the conservation target for each species. The spatial overlap of priority sets of sites that were identified for different values of each parameter was low. Therefore, the spatial distribution of priority conservation sites depends on values for these parameters. Data organisation also influenced the evaluation of existing protected areas in the region for maintaining the threatened vertebrate species. Both selection unit size and assignment of protection status to selection units, based on area of overlap with actual protected areas, resulted in different evaluations of reserve performance. I demonstrate that systematic reserve selection cannot be performed with data at only one spatial organisation unless the consequences are recognised.  ii  T A B L E OF CONTENTS Abstract Table of Contents List of Tables List of Figures List of Abbreviations Acknowledgements  ii iii viii xi xiii xiv  1.0 C H A P T E R 1: G E N E R A L I N T R O D U C T I O N  1  1.1 T H E C O N S E R V A T I O N CRISIS  1  1.2 S Y S T E M A T I C R E S E R V E SELECTION  1  1.3 R E S E A R C H OBJECTIVES  3  1.4 OUTLINE OF CHAPTERS  4  2.0 C H A P T E R 2: D A T A Q U A L I T Y ISSUES A N D R E Q U I R E M E N T S O F SYSTEMATIC RESERVE SELECTION  5  2.1 INTRODUCTION 2.1.1 The Conservation Issue 2.1.2 Conventional Reserve Selection Techniques 2.1.3 Systematic Reserve Selection Techniques 2.1.4 Systematic Reserve Selection Software 2.1.5 Biological Data Required for Reserve Selection Algorithms 2.1.6 Objectives  5 5 5 6 8 9 10  2.2 M E T H O D S 2.2.1 Study Area 2.2.2 Data Compilation and Description 2.2.2.1 Terrestrial Ecosystem Mapping 2.2.2.2 Species Occurrences 2.2.2.3 Wildlife Habitat Relationship Models 2.2.3 Identification of Conservation Targets 2.2.3.1 Population Estimates 2.2.3.2 Density Estimates 2.2.3.3 Life History Requirements 2.2.3.4 Conservation Targets Used in Analyses 2.2A Reserve Selection Algorithm  10 10 13 13 13 13 17 17 17 18 19 20  2.3 RESULTS 2.3.1 Reliability of Vertebrate Species Data  23 23  iii  2.3.1.1 Occurrence Data 2.3.1.2 Wildlife Habitat Relationship Model Reliability 2.3.2 Minimum Set of Sites for Conservation  23 24 26  2.4 DISCUSSION 2.4.1 Issues Related to Reserve Selection 2.4.1.1 Reliability of the Species Data 2.4.1.2 Viability of the Reserve Network 2.4.2 Conclusions and Recommendations  29 29 29 30 32  3.0 C H A P T E R 3: H O W D O E S T H E I D E N T I F I C A T I O N O F P R I O R I T Y C O N S E R V A T I O N SITES D E P E N D O N C R I T E R I A U S E D F O R CHOOSING THEM?  34  3.1 INTRODUCTION 3.1.1 Choosing Selection Unit Size and Shape 3.1.2 Choosing Indicators of Biodiversity 3.1.3 Choosing Conservation Targets 3.1.4 Objectives  34 35 36 37 38  3.2 M E T H O D S 3.2.1 Study Area and Data Sets 3.2.2 Reserve Selection Algorithm 3.2.3 Sensitivity of Priority Site Selection to Variation in Parameter Values 3.2.3.1 Selection Unit 3.2.3.2 Biodiversity Indicator 3.2.3.3 Conservation Target 3.2.4 Criteria for Comparing Priority Sets of Sites 3.2.5 Surrogacy of Priority Sets of Sites 3.2.5.1 Surrogacy of Vegetation Classes for Vertebrate Species 3.2.5.2 Surrogacy of Priority Sites for Actual Occurrences of Species  39 39 40 40 40 43 46 48 48 48 49  3.3 RESULTS 3.3.1 Variation in Selection Unit Size and Shape 3.3.2 Variation in Indicators of Biodiversity 3.3.3 Variation in Conservation Targets for Vertebrate Species 3.3.4 Selection Units Occurring in A l l Priority Sets of Sites 3.3.5 Surrogacy of Priority Sets of Sites 3.3.5.1 Surrogacy of Priority Sites for Vertebrate Species 3.3.5.2 Surrogacy of Priority Sites for Actual Occurrences of Species  49 49 52 54 57 58 58 60  3.4 DISCUSSION 3.4.1 The Problem of Scale 3.4.2 The Question of Surrogacy 3.4.3 The Problem of Identifying and Maintaining Viable Populations 3.4.4 The Use of Habitat Area Predictions and Presence Data in Reserve Selection  61 61 64 66 67  iv  3.5 CONCLUSIONS A N D R E C O M M E N D A T I O N S  68  4.0 C H A P T E R 4: S C A L E E F F E C T S O N P E R F O R M A N C E I N D I C E S U S E D T O EVALUATE RESERVE NETWORKS 71 4.1 INTRODUCTION 4.1.1 Measuring Performance of Reserve Networks 4.1.2 Congruence of Reserve Networks and Units Used for Evaluation 4.1.3 Population Viability within Reserve Networks 4.1.4 Objectives  71 71 72 72 73  4.2 M E T H O D S 4.2.1 Study Area and Data Sets 4.2.2 Methodology Issues 4.2.2.1 Selection Unit Size 4.2.2.2 Assignment of Protection Status to Selection Units 4.2.3 Evaluation of Reserve Networks 4.2.3.1 Efficiency 4.2.3.2 Effectiveness 4.2.3.3 Similarity and Redundancy 4.2.3.4 Connectivity 4.2.4 Reserve Selection Algorithm for Minimum Sets  74 74 74 74 75 75 76 78 78 79 79  . •  4.3 RESULTS 4.3.1 Efficiency of Reserve Networks 4.3.2 Effectiveness of Reserve Networks 4.3.2.1 All Protected Areas 4.3.2.2 Provincial Protected Areas 4.3.3 Redundancy in Reserve Networks 4.3.3.1 All Protected Areas 4.3.3.2 Provincial Protected Areas 4.3.4 Connectivity of Reserve Networks  79 79 82 82 83 84 84 88 89  4.4 DISCUSSION 4.4.1 Methodology for Evaluating Reserve Networks 4.4.2 Reserve Network Performance in the South Okanagan 4.4.3 Recommendations 4.4.3.1 All Protected Areas 4.4.3.2 Provincial Protected Areas  92 92 94 96 96 98  4.5 CONCLUSIONS  99  v  5.0 C H A P T E R 5: G E N E R A L C O N C L U S I O N  100  5.1 S U M M A R Y 5.1.1 Ecological Data Used for Reserve Selection 5.1.2 Scale of Reserve Selection 5.1.2.1 Selection Unit Size 5.1.2.2 Biodiversity Indicator 5.1.2.3 Conservation Target 5.1.3 Existing Reserve Network Performance  100 101 101 101 102 103 103  5.2 SPECIES S U R V I V A L PROBABILITY IN R E S E R V E N E T W O R K S 5.2.1 Species Interaction 5.2.2 Socio-economic Influence 5.2.3 Climate Change  104 104 105 105  5.3 I M P R O V E M E N T S TO SYSTEMATIC R E S E R V E SELECTION  105  5.4  107  CONCLUSIONS  LITERATURE CITED  108  Appendix I  Classification of the vegetation associations in the terrestrial ecosystem map of the South Okanagan region into fifteen plant communities and three human modified categories. 118  Appendix II  Vertebrate species occurrence data used to produce a final pooled occurrence database for the twenty-nine vertebrate species included in this thesis. 122  Appendix III  Conversion factor used to calculate the amount of habitat area in a polygon for each of the twenty-nine vertebrate species for the (a) 6-class rating scheme and (b) 4-class rating scheme of the wildlife habitat relationship models. 123  Appendix IV  Maps of the wildlife habitat relationship models that were based on habitat suitability ratings within the South Okanagan region for the twenty-nine vertebrate species included in this thesis. 124  Appendix V  Literature sources and raw data used to determine current population estimates in the South Okanagan region and density estimates in suitable habitat of the twenty-nine vertebrate species included in this thesis. 134  Appendix V I  Chi-square table with observed and expected number of occurrence records within 500 metres of (a) paved and (b) any (paved or gravel) roads.  VI  163  Appendix VII  Number of individuals and occurrence records per square kilometre of twenty-nine threatened vertebrate species relative to the habitat quality categories identified by the wildlife habitat relationship models.  164  Appendix VIII The (a) Chi-Square (%) and (b) Log-likelihood Ratio (G) values of the distributions of individuals and occurrence records in relation to habitat quality categories identified by the wildlife habitat relationship models for each of the twenty-nine vertebrate species, except American Bittern, Ferruginous Hawk and Short-eared Owl. The degrees of freedom (DF) for each W H R model is 3 unless specified otherwise. The probability that both the frequencies of individuals and occurrence records in each habitat quality category do not result from a uniform distribution is denoted as: *** fori°<0.01; ** forP<0.025; and * forP<0.05. 167 2  Appendix IX  List of Red and Blue Listed plant communities and vascular plant species used in the analyses of surrogacy of priority sites for actual occurrences of species (data provided by the Conservation Data Centre). 170  Appendix X  List of rare invertebrate species included in the analyses of surrogacy of priority sets of sites for actual locations of species (unpublished data provided by Dr. G. Scudder). 171  Appendix XI  Total area of the reserve network, redundant area, total area in the database, proportion of area selected and proportion of targets achieved in the reserve network for (a) A l l Protected Areas and (b) Provincial Protected Areas. Reserve networks were defined by their actual boundary, 2 k m hexagons, and 10 k m hexagons. Proposed and existing reserve networks included the minimum set, existing reserve network (ER), existing reserve network with complementary sites added (ER + CS), and existing reserve network with complementary sites added and redundant sites removed (ER + CS - RS). Minimum sets identified for (b) reflected the difference in total area of provincial land, which depended on how protection status was assigned to hexagons. 174 2  2  vii  LIST OF T A B L E S Table 2.1  The provincial status in British Columbia, COSEWIC status, current population estimate, density estimate in suitable habitat, and the number of sources for the density estimate for twenty-nine vertebrate species found in the South Okanagan region.  15  Table 2.2  Habitat capability and suitability rating schemes for three levels of knowledge of species habitat use (MELP 1999). 16  Table 2.3  Conservation targets based on current population and density estimates for twenty-nine vertebrate species used in the reserve selection analysis.  20  Conservation targets of vegetation classes and their associated range conditions and successional stages in the T E M database.  45  Table 3.1  Table 3.2  Area-based and site-based conservation targets for the twenty-nine vertebrate species included in this thesis. 47  Table 3.3  Comparison of priority sets of sites identified by C-Plan resulting from variation in (a) the type and size of selection unit, (b) the indicator of biodiversity included in analysis and (c) the type and size of conservation target.  51  Jaccard's similarity coefficients for priority sets of sites resulting from different selection units. The probability of obtaining each similarity coefficient by random was calculated from 1000 random selections of the equivalent total number of sites, or area for polygon selection units, and denoted as: *** for P < 0.001. A l l significant coefficients are higher than similarity values between randomly selected sets of sites and each of the minimum sets.  52  Table 3.4  Table 3.5  Jaccard's similarity coefficients for priority sets of sites resulting from different indicators of biodiversity. The probability of obtaining each similarity coefficient by random was calculated from 1000 random selections of the equivalent total number of sites and denoted as: * for P < 0.05; *** for P < 0.001. A l l significant coefficients are higher than similarity values between randomly selected sets of sites and each of the minimum sets. 54  Table 3.6  Jaccard's similarity coefficients for priority sets of sites resulting from different conservation targets. The probability of obtaining each similarity coefficient by random was calculated from 1000 random selections of the equivalent total number of sites and denoted as: NS for P > 0.05; * for P < 0.05; *** for P < 0.001. A l l significant coefficients are higher than similarity values between randomly selected sets of sites and each of the minimum sets. 57  viii  Table 3.7  Table 3.8  Table 3.9  Total area and percentage of spatial overlap of priority sets of sites resulting from variation in algorithm parameters.  58  Proportion of area-based conservation targets for Red and Blue Listed vertebrate species achieved in priority sets of sites resulting for different biodiversity indicators.  60  Mean proportion of occurrence records and standard deviation of Red and Blue Listed vertebrate, invertebrate and plant species located in priority sets of sites resulting for different biodiversity indicators.  60  Table 4.1  Total area (km ) of the A l l Protected Areas (APA) Reserve Network with Complementary Sites added (ER + CS), A P A Reserve Network with Complementary Sites added and Redundant Sites removed (ER + CS - RS), Minimum Set, and mean of 1000 Random Sets of sites. Each reserve network achieved the conservation targets for twenty-nine threatened vertebrate species. A l l values were significantly higher than random with probability values o f P < 0.0001. 80  Table 4.2  Effectiveness of the A l l Protected Areas (APA) Reserve Network, Minimum Set of equivalent total area as the A P A Reserve Network, and mean of 1000 Random Sets of sites of equivalent total area as the A P A Reserve Network. Effectiveness was calculated as the proportion of the conservation targets for twenty-nine threatened vertebrate species achieved in each set of sites. The probability of obtaining each effectiveness value by random was calculated from 1000 random selections of equivalent total area and denoted as: NS for P >0.05; * for P< 0.05; ** fori* < 0.01; *** f o r P < 0.001. 82  Table 4.3  Effectiveness of the Provincial Reserve Network, Minimum Set of equivalent total area as the Provincial Reserve Network, and mean of 1000 Random Sets of sites of equivalent total area as the Provincial Reserve Network. Effectiveness was calculated as the proportion of the conservation targets for twenty-nine threatened vertebrate species achieved in each set of sites. The probability of obtaining each effectiveness value by random was calculated from 1000 random selections of equivalent total area and denoted as: NS for P >0.05; ** for P< 0.01; *** for P< 0.001. 83  Table 4.4  Similarity of the A l l Protected Areas (APA) Reserve Network and Minimum Set of equivalent total area as the A P A Reserve Network. Similarity was calculated by Jaccard's Similarity Coefficient. The probability of obtaining each similarity coefficient by random was calculated from 1000 random selections of equivalent total area (i.e. Random Set) and denoted as: NS for P > 0.05; * for P< 0.05; ** for P< 0.01. 88  ix  Table 4.5  Similarity of the Provincial Reserve Network and Minimum Set of equivalent total area as the Provincial Reserve Network. Similarity was calculated by Jaccard's Similarity Coefficient. The probability of obtaining each similarity coefficient by random was calculated from 1000 random selections of the equivalent total area (i.e. Random Set) and denoted as: NS for P > 0.05; * for P < 0.05; *** for P< 0.001. 89  Table 4.6  Connectivity of the A l l Protected Areas (APA) reserve network identified by actual boundaries of the reserve network and by (a) 2 k m hexagons and (b) 10 km hexagons, where any portion of each hexagon is protected and where more than half of the hexagon is protected, the A P A reserve network with complementary sites added, the A P A reserve network with complementary sites added and redundant sites removed, and the minimum set. Connectivity was measured by the total number of patches, mean and median patch area and mean perimeter to area ratio. The relative ranking of connectivity values for reserve networks that achieved conservation targets for twenty-nine threatened vertebrate species (the last 5 reserve networks listed in each table) is denoted as for the best value and * for the worst value. 90 2  2  H  x  LIST OF FIGURES Figure 2.1 Vegetation associations in the terrestrial ecosystem mapping of the South Okanagan and Lower Similkameen Valleys. Large lake and urban polygons were excluded from the systematic reserve selection analyses. 12 Figure 2.2 The observed and expected number of occurrence records that are located within 500 metres of (A) paved roads and (B) any (paved or gravel) roads in the South Okanagan region.  25  Figure 2.3 The mean proportion of occurrence records of vertebrate species per square kilometre and standard deviation in habitat identified by the (a) 4-class habitat rating, (b) 4-class habitat rating, where high quality habitat is located outside of the region, and (c) 6-class habitat rating, for California Bighorn Sheep only, in the wildlife habitat relationship models. 27 Figure 2.4 Maps of (A) Species Richness, (B) Initial Irreplaceability, and (C) the Minimum Set of sites that achieves the conservation targets for twenty-nine threatened vertebrate species in the South Okanagan. A l l three analyses were based on data from the wildlife habitat relationship models (Appendix IV). 28 Figure 3.1 Selection units used to identify priority sets of sites for conservation: (a) 10 km hexagons, (b) 2 km hexagons, (c) 0.155 km hexagons and (d) T E M polygons. 2  2  2  Figure 3.2 Priority sets of sites for conservation identified by C-Plan for different 2  2  42  2  selection units: (a) 10 km hexagons, (b) 2 km hexagons, (c) 0.155 km hexagons and (d) T E M polygons.  50  Figure 3.3 Priority sets of sites for conservation identified by C-Plan for different biodiversity indicators consisting of: (a) Red and Blue Listed vertebrate species, (b) Red Listed vertebrate species, (c) vegetation classes, and (d) Red and Blue Listed vertebrate species and vegetation classes, collectively. 53 Figure 3.4 Priority sets of sites for conservation identified by C-Plan for different conservation targets for Red and Blue Listed vertebrate species consisting of: (a) minimum viable population estimates, (b) current population estimates (1.0-populations), (c) 50% of the current population estimates (0.5populations), and (d) 0.5-populations identified by species presence in each selection unit. 56 Figure 3.5 Area of overlap of priority sets of sites identified by C-Plan for variation in three algorithm parameters: (a) selection unit, (b) biodiversity indicators (species and vegetation classes), (c) conservation targets, and (d) all three parameters listed in a, b, and c.  xi  59  Figure 4.1 The existing reserve network in the South Okanagan consisting of provincially and privately owned land, which was identified by: (a) the actual boundaries of the reserve network, (b) 2 km hexagons that coincide with the 2  2  1  reserve network (2km -over), and (c) 10 km hexagons that coincide with the reserve network (10km -over). A hexagon was considered protected i f any portion of the hexagon coincided with actual boundaries of the existing reserve network. 77 2  Figure 4.2 Efficiency of the reserve networks identified by the (a) 2 k m hexagonal grid and (b) 10 k m hexagonal grid that occur on both private and provincial land. The reserve networks consist of the A l l Protected Areas (APA) reserve network with complementary sites added (ER + CS), A P A reserve network with complementary sites added and redundant sites removed (ER + CS RS), minimum set, and mean and standard deviation of 1000 random sets of sites. Efficiency is calculated as 1 - (XIT) where X is the total area (km ) required to achieve the conservation targets for twenty-nine threatened vertebrate species and T is the total area available in the database. A l l efficiency values were significantly higher than random selections of sites with probability values of P < 0.0001. 81 2  2  2  Figure 4.3 Redundancy of the existing reserve network identified by a 2 k m hexagonal grid after complementary sites were added that achieved conservation targets for twenty-nine threatened vertebrate species. The reserve network was identified by hexagons with two levels of protection: (a) where any portion of each hexagon coincided with actual boundaries of the existing reserve network (2km -over), and (b) where more than half of each hexagon coincided with actual boundaries of the existing reserve network (2km under). 85 2  2  2  Figure 4.4 Redundancy of the existing reserve network identified by a 10 km hexagonal grid after complementary sites were added that achieved conservation targets for twenty-nine threatened vertebrate species. The reserve network was identified by hexagons with two levels of protection: (a) where any portion of each hexagon coincided with actual boundaries of the existing reserve network (10km -over), and (b) where more than half of each hexagon coincided with actual boundaries of the existing reserve network (10km under). 86 2  2  Xll  LIST OF ABBREVIATIONS AND DEFINITIONS General: Fine scale = large scale (i.e. 1:2Q,000), high resolution, regional scale mapping Coarse scale = small scale (i.e. 1:500,000), low resolution, national scale mapping CDC = Conservation Data Centre COSEWIC = Committee on the Status of Endangered Wildlife in Canada GIS = Geographic Information System L R M P = Land and Resource Management Plan M A U P = Modifiable Areal Unit Problem M E L P = Ministry of Environment, Lands and Parks (recently restructured as Ministry of Water, Land and Air Protection) PAS = Protected Areas Strategy T E M = Terrestrial Ecosystem Map U B C = University of British Columbia SOCS = South Okanagan Conservation Strategy WHR = Wildlife Habitat Relationship (models)  Chapter 3: Conservation Target Abbreviations: 1.0-populations = 100% of the current population estimates (of threatened species in the South Okanagan region) 0.5-populations = 50% of the current population estimates (of threatened species in the South Okanagan region) M V P = Minimum Viable Population (estimates of threatened species in the South Okanagan region)  Chapter 4: Site Selection Abbreviations: A P A = A l l Protected Areas (consisting of privately and provincially owned land) CS = Complementary Sites ER = Existing Reserves RS = Redundant Sites  Chapter 4: Reserve Network Abbreviations: 2km -over = 2 k m hexagonal grid of the reserve network, where any portion of each hexagon coincides with the existing reserve network; over-estimates the total area of the actual reserve network 2  2  2km -under = 2 km hexagonal grid of the reserve network, where more than half of each hexagon coincides with the existing reserve network; under-estimates the total area of the actual reserve network 2  2  10km -over = 10 k m hexagonal grid of the reserve network where any portion of each hexagon coincides with the existing reserve network; over-estimate 2  2  10km -under = 10 km hexagonal grid of the reserve network where more than half of each hexagon coincides with the existing reserve network; under-estimate 2  2  Minimum Set = a set of sites that represents the minimum amount of area required to achieve the conservation targets for a region xiii  ACKNOWLEDGEMENTS This thesis would not have been possible without the support and patience of many people. I would like to thank my supervisor, Dr. Tony Sinclair, for providing guidance in researching my own ideas and exploring questions regarding the methodology of systematic reserve selection. Special thanks to my committee member, Dr. Brian Klinkenberg, for his creative ideas and programming expertise. I would also like to thank my committee members Dr. Geoff Scudder and Dr. Kathy Martin for their encouragement and helpful suggestions. The time, data, expertise and workspace provided by many personnel at the Ministry of Environment, Lands and Parks were invaluable and greatly appreciated. In particular, I would like to thank Orville Dyer and Tom Ethier for their generous support of this project and for their friendship. I am very grateful of the time and expertise that Glenna Boughton, Nathan Hollenbeck, and Mark Cudmore provided for GIS queries and data management, which were essential for setting up this project. I would like to thank the National Parks and Wildlife Service, New South Wales, Australia for providing complimentary C-Plan software and Tom Barrett for his invaluable technical support for the analyses in this thesis. Much of this thesis required programming in languages that I was not fluent in. Without the help of Alistair Blachford, Steve Martell, Nathaniel Newlands, and Dr. Nick Nicholls my level of frustration would have been much higher. I would also like to thank the Zoology Computing Unit personnel for their coveted time and suggestions related to computer problems, which also alleviated my level of frustration. I would like to thank Nancy Mahony, Susan Paczek and Doug Olson for contributing data for the analyses, and Alison Haney and Mike Sarell for data and help with interpreting analyses that used their data. In addition to Dr. Dave Forsyth's suggestions and assistance with C-Plan analyses, I would also like to thank him for furnishing the office. I am grateful for the encouragement, support, friendship and comic relief offered by many "hut dwellers", in particular Alistair Blachford, Dr. Bea Beisner, Dawn Cooper, Grant Hopcraft, and Maria Morlin. I would like to especially thank my parents, Jim and Dorothy Warman, for their continual encouragement, support, interest and their regard for my well being. And I am grateful for Dr. Lee Gass's wonderful supervisory skills during my undergraduate directed studies project, which motivated me to pursue a graduate degree and helped direct my research interests. Finally, I would like to thank the many organisations that contributed to the funding of this project and the individuals that helped arrange for the financial assistance (listed below), and Irene Wingate for managing the grants and performing countless other tasks that helped me daily. • Habitat Conservation Trust Fund - Dr. Ian McTaggart-Cowen • Environment Canada's Science Horizons Youth Internship Program - Dr. Kathryn Freemark • Friends of Ecological Reserves - Dr. Bristol Foster • Canadian Wildlife Service - Dr. Rob Butler • Ministry of Environment, Lands and Parks, in kind support - Tom Ethier • National Science and Engineering Research Council, operating grant - Dr. Tony Sinclair  xiv  1.0 CHAPTER 1: GENERAL INTRODUCTION 1.1 THE CONSERVATION CRISIS There is an escalating concern over the loss of global biological diversity and the rate at which this loss is increasing. The primary cause of the decline in biodiversity is attributed to the modification and destruction of natural ecosystems. Urbanisation, agriculture, forestry practices and other activities along with global warming have accelerated the rate of decay of natural habitat (Sinclair et al. 1995). Changes associated with human activity have increased the rate that species are going extinct to the present estimate of 100 times the natural background rate (Pimm and Lawton 1998). If the rate of habitat degradation continues, without protecting essential habitat for biodiversity, the rate of species extinction will continue to increase (Noss 1994). The number of species that require immediate protection to mitigate the potential for extinction is much greater than the resources available for their conservation (Myers et al. 2000). Furthermore, usually only a small part of the landscape within a region can be managed primarily for nature conservation (Pressey and Logan 1998). Therefore, we must identify priorities for conservation of biodiversity before the opportunities for reserve selection are limited by alternate land uses. However, it is difficult to determine how to maximise the number of species conserved when the resources for conservation are insufficient to protect all species (Camm et al. 1996, Faith and Walker 1996).  1.2 SYSTEMATIC RESERVE SELECTION Many current reserve networks have been protected using an ad hoc or opportunistic approach. Current protected areas tend to represent natural features disproportionately, thereby decreasing the potential for conservation of biodiversity and increasing the overall cost of conservation (Pressey 1994). Therefore, more effective and directed conservation planning techniques are necessary to minimise the loss of biodiversity. Although, the conservation of biodiversity must be the primary goal for an effective approach, practical plans must also identify the options available for reserve selection because of the limited resources available for conservation. Many different reserve selection techniques have been developed to systematically determine the best areas to conserve a desired species or subset of species (Margules and Redhead 1995, Faith and Walker 1996, Williams et al. 1996, Csutsi et al. 1997, Pressey et al.  1  1997). Some of the systematic reserve selection techniques have focused on the efficiency of selected sets of sites, in terms of the amount of area or cost of protecting selected sites, by identifying the minimum area required to conserve biodiversity (Pressey and Nicholls 1989a). Efficiency is a critical component of reserve selection because of the limited resources available for conservation. Because systematic reserve selection is efficient, it has been used as a practical management tool for identifying alternatives for conserving biodiversity when competing land uses are considered (NPWS 1999). The National Parks and Wildlife Service in New South Wales, Australia, has applied a systematic reserve selection technique in 1996 and successfully negotiated with the forest industry for areas that were required for conservation of native flora and fauna (Finkel 1998). Reserve selection techniques that identify the minimum area required for conservation of species are based on the principle of 'complementarity' (Vane-Wright et al. 1991). Complementarity is used to identify sites for conservation that complement previously selected sites (or reserve networks) by contributing the most new species or biological attributes that are not already represented in the selected sites (Vane-Wright et al. 1991, Faith and Walker 1996, Williams et al. 1996). The premise of complementarity is that by minimising the redundancy in the overall reserve network, resources available for conserving biodiversity will be used more effectively, and overall cost of conserving the full complement of biodiversity within a region, in terms of land and funds, is reduced (Pressey et al. 1993, Pressey 1994). Many systematic reserve selection techniques use geographic information systems (GIS) as a tool for analysis and display of spatial data. Although, GIS provides a powerful tool for spatial analyses, the accuracy of spatial data that represent continuous landscapes is affected by arbitrary boundaries used for mapping spatial phenomena (Heywood et al. 1998). The effects of cartography on systematic reserve selection have not been thoroughly investigated. Therefore, any evaluation of priority conservation areas identified by systematic reserve selection must consider both the ecology of biodiversity and how biodiversity is mapped. Scientific studies that have applied systematic reserve selection techniques to different regions identify that these techniques are efficient at achieving conservation goals with limited resources, defensible and flexible when encountering competing land uses, and accountable since decisions can be critically reviewed (Margules and Pressey 2000). 2  However, the recommendations from many of these scientific studies have not been implemented, in part, because of the scale of the reserve selection analyses. Systematic reserve selection has primarily been applied at coarse spatial scales such as nations, provinces or states, or other large regions (Williams et al. 1996, van Jaarsveld et al. 1998, Freemark et al. 2000, Brooks et al. 2001). Phenomena mapped at coarse spatial scales have low resolution (i.e. maps are at a small scale, such as 1:500,000). Therefore, systematic reserve selection must be applied at fine spatial scales (i.e. large map scales, such as 1:20,000) in regions where priority conservation sites can actually be protected and have their effectiveness for conserving biodiversity evaluated over time (da Fonesca et al. 2000). 1.3 R E S E A R C H  OBJECTIVES  The overall goal of this thesis was to explore a systematic reserve selection technique for conserving biodiversity at a fine spatial scale. I used C-Plan as the tool to explore systematic reserve selection because this software provided a refined algorithm for reserve selection at the beginning of this thesis (NPWS 1999). Through the exploration of the systematic reserve selection process, I identified priority conservation areas for threatened biodiversity located within the South Okanagan region. The South Okanagan and Lower Similkameen valleys constitute a small region in British Columbia, which is considered to contain one of Canada's four most endangered ecosystems (Schluter et al. 1995). Therefore, this thesis has considerable practical value for conservation within the region and the nation. My objectives were to: 1. Identify the amount of area required to conserve each of twenty-nine threatened vertebrate species within the region. These values are called 'conservation targets' and are a necessary component of systematic reserve selection. 2. Determine the minimum amount of area that achieves the conservation targets of each of twenty-nine threatened vertebrate species within the region. This selected set of sites represents the most efficient network of sites for conserving threatened biodiversity within the South Okanagan region. 3. Identify the variation in sets of sites chosen by a systematic reserve selection algorithm when parameter values are altered. It is important to understand the sensitivity of an algorithm to variation in parameter values, since there are many arbitrary decisions related to scale in systematic reserve selection and cartography.  3  4. Assess the performance of the current reserve network in the region for conserving threatened species. Compare the performance of the existing reserve network to reserves chosen systematically to investigate the differences between opportunistic and systematic reserve selection. 5. Identify additional sites that complement the current reserve network and achieve the conservation targets for species in the region. These complementary sites identify priorities for future conservation in the region. 1.4 O U T L I N E O F C H A P T E R S  In Chapter 2 of this thesis, I examine data quality issues related to reserve selection at fine spatial scales. M y approach is to identify the minimum amount of area required to conserve twenty-nine threatened vertebrate species in the South Okanagan using C-Plan. I describe the assumptions that are necessary to complete this analysis and the viability of the resulting reserve network for conserving threatened vertebrate species. The issues that I identify in this chapter are applicable to any reserve selection analysis, whether or not a systematic selection technique is used. In Chapter 3,1 apply some of the recommendations from Chapter 2 to reserve selection in the South Okanagan. I examine effects of altering three parameters of the C-Plan algorithm on reserve selection: selection unit size, type of biological data and conservation target values. I identify priorities for conservation in the region as the sites that are consistently selected in all variations of parameter values. In Chapter 4,1 examine the performance of the current reserve network in the South Okanagan for protecting threatened vertebrate species. Following the recommendations from Chapter 3,1 analyse the reserve network with two different selection units and determine whether the two units result in different assessments of reserve performance. Using C-Plan, I identify sites with the two selection unit sizes that complement the reserve network and achieve the conservation targets for the twenty-nine threatened species. Finally, in Chapter 5,1 provide a synthesis and summary of the results presented in this thesis. I recommend some improvements to the C-Plan algorithm that are essential for identifying reserve networks that maintain viable populations of species.  4  2.0 C H A P T E R 2: D A T A Q U A L I T Y ISSUES A N D R E Q U I R E M E N T S O F SYSTEMATIC RESERVE SELECTION 2.1  INTRODUCTION  2.1.1 The Conservation Issue In an attempt to minimise the loss of biological diversity, large amounts of time, money and effort have been focused on preserving natural habitats to maintain biodiversity or portions thereof in perpetuity (Sinclair et al. 1995). Management and conservation planning decisions are required immediately to protect the greatest biological diversity possible. Many different approaches have been developed to determine the best areas for conserving multiple species (Margules and Redhead 1995, Faith and Walker 1996, Williams et al. 1996, Csutsi et al. 1997, Pressey et al. 1997). The efficiency of the reserve selection approaches, in terms of their ability to identify a greater diversity of natural areas or species for conservation in a given amount of area, is essential (Pressey and Nicholls 1989a). The problem facing conservation is that the resources available for conservation are insufficient to save everything. Therefore, reserve selection techniques need to identify the optimum allocation of resources, in terms of land and funds, to save the maximum amount of habitat and species. 2.1.2 Conventional Reserve Selection Techniques Many of the existing reserve networks have been designated using ad hoc or opportunistic decisions for site selection (Pressey 1994). Reserve networks were usually selected because of their natural beauty, recreation value, historical value, or a lack of any obvious economic value (Pressey 1994, Csutsi et al. 1997). The cost-effectiveness of this approach is very low, since conservation of biodiversity was not the primary objective for creating reserves. Many species, communities and ecosystems are left without protection in regional reserve networks when an ad hoc approach is used, which often leads to an increase in the total cost of conserving regional biodiversity (Pressey 1994). One of the approaches developed with the conservation of species as the primary objective is the selection of areas with a high concentration of species, known as hotspots. Two types of hotspots have been used for identifying conservation areas: 'richness hotspots' and 'rarity hotspots' (Prendergast et al. 1993a, Williams et al. 1996). Richness hotspots are areas that have an exceptional concentration of species or high species richness. Rarity hotspots identify areas that are richest in the species with the most restricted ranges (Williams  5  et al. 1996). However, comparative studies of conservation planning techniques have demonstrated that hotspots are inefficient at representing regional biodiversity (Williams et al. 1996). Hotspots of species with similar ecological requirements (i.e. a particular taxonomic group) do not overlap hotspots of species with different ecological requirements (Prendergast et al. 1993a). Therefore, the hotspot approach identifies areas that are important for groups of similar species without identifying areas that may maintain diverse species. Gap analysis has been developed as another approach for comprehensive conservation of biological diversity (Flather et al. 1997) and has been implemented as a program to evaluate conservation of biodiversity in the United States (Scott et al. 1993). The overall goal of gap analysis is to identify areas with high species diversity that have species that are not included in the existing reserve networks (Scott et al. 1993). These areas are referred to as "gaps". Until recently, gaps could only identify areas that were species rich, but could not identify the overall value of a site or its priority for inclusion in the reserve network (Williams 1998). Therefore, reserve networks identified with this approach were inefficient because of the duplication of sites that were rich in similar species. 2.1.3 Systematic Reserve Selection Techniques The realisation that conventional methods of designating reserves would not contribute substantially to the conservation of global and local biodiversity led to the development of more systematic approaches to reserve selection. Systematic reserve selection techniques use scoring procedures or iterative algorithms that are explicit and repeatable. The procedure for multi-criteria scoring relies on the predetermined ranking of sites in order of value according to a combined score from a variety of criteria, such as diversity or rarity of species (or other biological attributes), and site size and naturalness (Pressey and Nicholls 1989a). When two sites have equal values for conservation, current scoring procedures select sites in the order in which they appear on the list, disregarding differences in attributes in the equally scored sites (Pressey and Nicholls 1989a). Therefore, even though this method takes into account a variety of criteria to determine the overall score of a site, the cost-effectiveness of the process is decreased as a consequence of the order sites are listed in the data set. Iterative algorithms are designed to identify minimum or near minimum solutions of complementary sites, which encompass all known or desired biological attributes in a region (Pressey and Nicholls 1989a). The process of identifying complementary sites is known as 'complementarity'. This process identifies additional sites that complement sites previously 6  selected by contributing the most new species or biological attributes not already represented in the reserve network (Vane-Wright et al. 1991, Faith and Walker 1996, Williams et al. 1996) . Complementarity is used to identify areas that represent the maximum number of species in the minimum amount of area. Therefore, instead of just focusing on species rich areas, this approach also considers the species in the hotspots and whether they have been adequately represented in alternate locations. The decision process works through a sequence that selects areas with the greatest number of unrepresented species until all species have been represented to their conservation target. Conservation targets are set at either a certain number of sites or total area required to represent each species in the selected sites. Complementarity may be used to identify the minimum amount of area required to achieve the conservation goals of a region using iterative algorithms based on species richness and rarity (Camm et al. 1996, Pressey et al. 1997). Iterative algorithms that are based on species richness identify sites with the greatest number of unrepresented species (i.e. species that are not represented to their conservation targets) and then add sites one at a time according to which site contains the most remaining unrepresented species. Algorithms based on species rarity identify sites that contain unique species and then add sites progressively according to which site contains the next rarest unrepresented species. Iterative algorithms are complete when the conservation target for each species is achieved or when the predetermined amount of area, representing as much of the conservation targets for each species as possible, is selected (Pressey andNicholls 1989a, Margules and Redhead 1995, Pressey et al. 1997). Iterative algorithms are used to identify either optimal or heuristic solutions for conservation. Optimal algorithms identify the best solution for conservation (Pressey et al. 1997) . However, since species can be conserved in different sets of sites, the contribution of adding any particular site is dependent on the sites that have already been chosen (Camm et al. 1996). Therefore, the only way to guarantee optimality is to identify all possible combinations of sites for conservation. This process is very time consuming, with computer running times of up to days or weeks depending on the data set and the hardware used to run the analysis (Pressey et al. 1997). Heuristic algorithms, which identify near optimal solutions, have much shorter computer running times (Pressey et al. 1997). These algorithms are heuristic because they make logical decisions on quantities computed during the course of the algorithm, and 7  selections are essentially unpredictable in advance (Pressey et al. 1997). Heuristic algorithms can give different solutions depending on the starting point of the algorithm, but generally selections occur in the same order for any given starting point (Pressey et al. 1997). This is a good approach for evaluating differences between sets of sites that represent species' conservation targets and their potential for protection. 2.1.4 Systematic Reserve Selection Software Various organisations have developed software, which can be used to identify efficient networks of reserves, but the functionality of the software varies. One of the software programs, called W O R L D M A P , was designed to perform specialist biological analyses for unlimited numbers of species (or other biological attributes), to explore biodiversity data for research (Williams 2001). Complementarity is an option in W O R L D M A P that can be used to select reserve networks. This software package has incorporated attributes of commercial geographic information systems, but cannot be used interactively with GIS. Another program, called SITES, is a customised ArcView project (a GIS) that facilitates designing and analysing alternative regionally representative systems of nature reserves while incorporating spatially explicit reserve design strategies (Andelman et al. 1999). There are two algorithms that can be used to identify reserve networks. The first algorithm is called 'simulated annealing', which evaluates alternative complete reserve systems at each step and approximates an optimal solution, and the second is a 'greedy heuristic' algorithm, which selects the best site at each step. The heuristic algorithm selects sites that reduce the total cost of the reserve network, but has not explicitly included the concept of complementarity. The National Parks and Wildlife Service in New South Wales, Australia, has recently produced CPlan, which can be used to identify conservation areas in landscapes subject to the effects of human development. C-Plan is a decision-support system that, together with a geographic information system (GIS), maps options for achieving explicit regional conservation goals using complementarity. C-Plan is similar to another software program that was developed in Australia, called C O D A , which currently does not interface with GIS (Bedward 1999). The C-Plan software provided a refined procedure for reserve selection and was therefore used in this thesis. The heuristic algorithm used in C-Plan selects 'sites', which are areas of land or water used for analysis and display, in an iterative manner based on the extent and number of 'features' that are located within a site. Features are defined as important attributes for 8  conservation. They can be physical (unique environmental units defined by geology and terrain), biological (vegetation types, species, populations), or cultural (Aboriginal or historical sites). Features are used to calculate the 'irreplaceability' values of a site, identify conservation targets and select priority sites for conservation. Conservation targets can be calculated as a percentage of the total amount of each feature in a region (i.e. 12% of each feature), or as a defined amount of total area or feature occurrence within the region. Irreplaceability provides an objective measure that is used to identify sites that are essential for achieving explicit conservation targets. The approach used in C-Plan identifies the minimum amount of area that maintains the maximum amount of regional biological diversity. The approach can be used to supplement existing reserve networks with sites that are complementary, distinguish between irreplaceable and flexible sites for conservation by evaluating the rarity or importance of biological features, and provide alternatives for conservation planning in the presence of competing land uses (Williams et al. 1996, Csutsi et al. 1997). The great advantage of this planning approach is the objective process for setting priorities for which habitats, species or other categories should be conserved, using biological, social and economic data. 2.1.5 Biological Data Required for Reserve Selection Algorithms The conservation solutions produced by systematic reserve selection algorithms depend on the reliability of the biological data used in the analyses. The biological data usually provide an index of regional 'biodiversity'. The most commonly accepted measure of biodiversity is 'species diversity', since species are biologically natural units with little taxonomic uncertainty and can be grasped intuitively by most people (Caughley and Gunn 1996, Myers et al. 2000). The most accurate data for reserve selection analyses are identified by systematic surveys of entire regions that record both presence and absence of species. However, these types of data are not available for most species. Predictive models of species distributions or probability of habitat use have been developed to identify potentially important areas for species that have not been surveyed. Although predictive data are limited by assumptions required to identify species' habitat, these data are useful for reserve selection analyses when systematic surveys of a region are not available. Reserve selection not only depends on the reliability of the data, but also on the conservation targets that are identified to maintain species over the long term (Cabeza and Moilanen 2001). Population viability assessments, which identify the minimum population 9  size that prevents extinction of a species, require data on population demography and genetic variation, environmental stochasticity, and natural catastrophes (Shaffer 1981, Soule 1987). However, this information is not known for most species and, therefore, general estimates of minimum viable populations have been proposed. These estimates range from fifty individuals to populations in the low thousands (Franklin 1980, Soule 1987, Caughley and Gunn 1996). Conservation decisions based on these general estimates must be used cautiously because of the uncertainty associated with them. 2.1.6 Objectives Most ecological reserves, provincial parks and wildlife management areas in the South Okanagan, British Columbia have been selected using an ad hoc or opportunistic approach. This approach has led to an uneven representation of natural features within protected areas, thereby decreasing the potential for conservation of biological diversity within the region (Pressey 1994). If the current rate of habitat loss continues, without protecting the essential habitats required to maintain populations, many species will be extirpated from the region (Noss 1994). Therefore, the overall goal of this study was to identify priority areas for conservation of threatened biodiversity in the South Okanagan. I used C-Plan to identify conservation areas using predictive models of habitat suitability for twenty-nine threatened vertebrate species. The objectives of this chapter were to (1) assess the reliability of the species data consisting of (a) occurrence records and (b) predictive models of habitat suitability, (2) identify viable or appropriate conservation targets for each species in the region, and (3) identify the sites that represent the minimum amount of area required to maintain twenty-nine threatened vertebrate species in the region. The results of this chapter identify some of the issues related to using a systematic reserve selection algorithm at fine spatial scales. 2.2 M E T H O D S 2.2.1 Study Area The South Okanagan and Lower Similkameen Valleys located in south central British Columbia (BC) are considered to contain one of Canada's four most endangered ecosystems (Schluter et al. 1995). The study area, which is defined by the South Okanagan Conservation Strategy, encompasses these two valleys and extends from the United States border north to Okanagan Mountain Provincial Park and east to west from Anarchist Mountain to the  10  Ashnola River (Hlady 1990, Figure 2.1). This area is commonly referred to as Canada's "pocket desert" because the microclimate of the valley is hot and dry. Much of the natural habitat in these valleys has been converted to urban developments and agricultural areas (Bryan et al. 1994, Schluter et al. 1995). Because of the rapid loss of habitat (Hume 1997), this region has many threatened species. The Committee on the Status of Endangered Wildlife in Canada (COSEWIC) has identified thirty-one species in the South Okanagan region that are "at risk" (COSEWIC 2001). In addition to the national listing of species at risk, the British Columbia Conservation Data Centre (CDC) identifies Red and Blue Listed species, which are considered to be provincially at risk. Red Listed taxa are species that have been, or are being, evaluated as endangered or threatened and Blue Listed taxa are species that are vulnerable and at risk of becoming endangered or threatened (Harper et al. 1994). If the rate of habitat loss continues without protecting essential habitats required for their survival, the extirpation of many of these species is expected. The South Okanagan region was chosen for this study because of the large number of endangered and vulnerable species, the imminent threat of habitat loss from human habitation, and the identification of the region as a priority for fine scale analysis within Canada (Freemark et al. 2000). Another reason for selecting this region is that the British Columbia government has developed a detailed regional vegetation map and extensive data sets of Red and Blue Listed species occurrences and habitat.  11  Figure 2.1. Vegetation associations in the terrestrial ecosystem mapping of the South Okanagan and Lower Similkameen valleys. Large lake and urban polygons were excluded from the systematic reserve selection analyses.  12  2.2.2 Data Compilation and Description 2.2.2.1 Terrestrial Ecosystem Mapping The Ministry of Environment, Lands and Parks (MELP) (recently restructured as the Ministry of Water, Land and Air Protection) provided the data for this project. The data consist of geo-referenced digital data sets that are compatible with Environmental Systems Research Institute (ESRI) geographic information systems. The 1:20,000 scale terrestrial ecosystem mapping (TEM) developed by M E L P in 1989 classifies the regional landscape into 111 vegetation associations based on climate, physiography, surficial material, soil and vegetation (Lea et al. 1991, Appendix I). Ongoing corrections and updates to the T E M map were completed in 1997 with the help of M E L P habitat biologists. The vegetation associations were mapped as polygons of irregular shape and size. There are 10,192 polygons in the T E M for this study area, with a total area of 1770.13 km . The large lakes that occur 2  along the valley bottom and polygons that are completely urban were not considered as potential sites for conservation in the reserve selection analyses, since they cannot realistically be conserved or restored (Figure 2.1). In total 10,125 T E M polygons that range in size from a portion of a small riparian island at 2.16E-04 km to an agricultural area at 72.36 km , with a total area of 1568.21 km (88.6% of the region) were used in the analyses. 2  2.2.2.2 Species Occurrences The species occurrence data consist of actual locations of species, representing species presence only, that were identified by a range of observers including the general public, naturalists and both government and non-government biologists. Four databases were joined to produce one pooled occurrence database for the twenty-nine vertebrate species included in this study, with precision of occurrence that ranges from plus or minus 100 metres to plus or minus 1000 metres (Appendix II). The data were obtained from the Conservation Data Centre, M E L P and from the Centre for Applied Conservation Biology at the University of British Columbia. Duplicate records were excluded from the final database. The occurrence database identifies the number of individuals in each record. In this chapter, a database entry is referred to as an 'occurrence record' and number of individuals per record as 'individuals'. 2.2.2.3 Wildlife Habitat Relationship Models Predictions of habitat suitability were identified by wildlife habitat relationship (WHR) models, which identify the amount of habitat available within each T E M polygon for the life history requirements of individual species found in the study area (Verner et al. 1986). The 13  W H R models identify habitat suitability for each species, which was based on the probability of current habitat use in each T E M polygon, and provide an indication of habitat quality. Lower habitat quality ratings were included in the models because these habitats are important for species' population dynamics and distribution. Habitat quality ratings were determined using species' distribution and habitat preferences from published and unpublished reports, and from local biologists and naturalists. There were twenty-nine WHR models developed for selected Red and Blue Listed vertebrate species located in the study area (Warman et al. 1998, Table 2.1). Standardised habitat quality ratings in BC were produced at three scales depending on the amount of information known about a species and the scale of the mapping (Table 2.2). The ratings were based on the quality of habitat relative to the best habitat in the province and measured as a percentage of the best quality habitat (MELP 1999). A six class rating scheme was used when there was a detailed level of knowledge about habitat use of a species, a four class rating scheme when there was an intermediate level of knowledge and a two class rating scheme when there was limited knowledge. Two class rating schemes did not provide reliable data for making conservation decisions. Therefore, only species models based on six or four class rating schemes were included in this thesis. Habitat rated as 'nil' is not likely to be used by a species except for potential transition between habitat patches.  14  CD  " >, in  jai  o  cn oo  CD  co  *t  00 (D  8 8 •t^ cn roc™™™™ccc™  O »> O  o c o o o o c c c o Z L U Z Z Z Z L U U J U J Z  •siii  0)D)D)^=;  o  to  to  o  o  O O  8  rorocoroCro O  W  11  o o  = >3 ro(DO) 0O.  pj  Q)  0oj<D(D0  cu C D  5  ro  o  ro  OCO  O  £  " i  JD OJ  Q.  TO  a> CO _ ™ ro ul ro m  . l a -ro  CO  ro  5  co  ro  CQ  cu ro co < ._> ro § z C5 i£ 5 CO S 1 1 £ S  a> „  OJ  Q. o  —  CD  ro o O CQ  QCO  <D LO  oo ro m co  Si I -  T -  s  oo ^ U  >>  - a>  CM  ~  ID  S o O  CO  ro00  O  ^  O  ro  W  M  o  o  —~ .</)-o"• _oo —  Q) 3 a) J  a>  ra I  in  co co C CD o00 o oj T -  a)" *•"  T-  co . 6 6  »-  ID ^ _ g  =  w cr 3 rn CD C D  O  ro ro tu o ™ O J O I D J  C  C  <D  §  CO  E (0 E  W  J  e c  Li  0  IS  .-ti  00 c"  co  ro00  T_>  m  _ J  c c  to TO 0)  o  s 1i i §s  3 -C to CQ O LL  c S  § c  T3  to  c  o o  I-. tu N  £  i2 O Q.co  TJ;  5. c o. c  O  to O O  3 "to  o a. t  15  o. ro  JO  CO O  §1  5 .__ CO _1  5  CD co  < <  CQ  E 00  w  co  00 00 CD 0  •§ 3 C to .o  .c  S  o o —  O  QJ C  o  S a  CO  o  to  in  ro CO CO ro  S op  m § § <2 v  a:  o ro  w  00 0  I '5 ™ 8 £ ii 3 5  T -  CL  c o  E cu •j= o  O <D  1 -s i I  •s tD  to 3 O C O  £ S co C  -  <D ^ Q. ^  to o u  O O  •5 S  T T  in  c -  o 00 m u.  aeu  ^  kett,  Tt  <D  s °°§  ro ro  ro CQ CQ „ C D  CQ 0 . O  O  T3 •8 >< 00 3 (0 CU ,V CD TD  -a o o  _  .S> OJ Q. OJ cn a. c o  <», o Qm  co E •!2 tS  o I-  ro ro co  o  CD C0CQCQCQCQC0CQCQ__.  CO  £  O CD  m  E roE  in •o v CD  •-  CM •a 0 0 0 O j a > ( D < D ( D O J 0 C D < D < D < D < D ( D 0 CQ CQ CQ CO CO CQ  (1)  E E o o  o o  to  zzcocozzzzzcoco  o </)  c o  in  to  - * ~ r o r o r o r o r o r o r o r o r o r o r o r o r o r o CO o O O O _ a . O O O O O a . O 1 — CO  ro  z  in in in  o  No  §3  o  ;us  05 (/) 0)  No  (0  sn:  (A  Coi  O Q.  O  Tt  O O O O O O O o O O O O O O O O ° O O O O O O o O O C M O t D O O O o O O O O O O O O C O O T f O O I ^ - T - O o O CM T< M O t - o m m i f ) i r ) C M ( N C N ^ - C M C O - — >- m N T OO  o o if) CM  aeu  U  o  o O  •^t co  ^- (O oo ^ oo co CM eg T - CM  :us  is iMl  1^  No  .1 So  oi o  CM CO  o  CO CD  ro  topF  •S  1  Spec  E -° :£ • f f  to ,o>  CD  §  *  s  i5  2  w .to -a  e(0  tD  o o  CD  I o?  co .o  g £  c  c  o  3g  CO . D to  to OJ  00 c  OJ to  to  I  c -ft  J2 to 3 C  to E  g to  I  (1)  11I |•e  co  2  LU  O  UJ  O  O  E  CO  c -  Table 2.2. Habitat capability and suitability rating schemes for three levels of knowledge of species habitat use (MELP 1999). % of Provincial Best* 100-76% 75-51% 50 - 26% 25 - 6% 5-1% 0%  Detailed Knowledge of Habitat Use (6-class) High Moderately High Moderate Low Very Low Nil  Intermediate Knowledge of Habitat Use (4-class) High Moderate Low Nil  Limited Knowledge of Habitat Use (2-class) Habitat Useable Likely No Value  * "Provincial Best" is the provincial benchmark habitat for a species against which all other habitats for that species are rated.  I modified the WHR models to work with C-Plan. The original ratings of habitat use for each polygon were converted to habitat area in each polygon based on the total area of habitat in the polygon and the habitat quality rating (Appendix III). For some species many polygons were intersected by boundaries of a migration distance buffer, which was based on known migration distances from critical habitat, such as breeding, denning and hibernation areas. The migration distance buffer was created by buffering polygons that contained critical habitat for a species by the distance that the species migrates from critical habitat to foraging areas. The habitat quality of polygons that contained potential foraging habitat, but were located outside of the migration distance buffer, was rated lower than foraging habitat located within the buffer. In the original models, polygons that were intersected by the migration distance buffer were assigned more than one rating depending on the number of intersections. To eliminate the division of polygons by the migration distance buffer in this thesis, polygons were assigned the highest rating that occurred in the polygon and the intersection boundary was removed. The wildlife habitat relationship models (Appendix IV) are also referred to as vertebrate species models in this chapter. The vertebrate species occurrence database was merged with the T E M map to determine the number of occurrences located in each habitat quality rating of the WHR models. Analyses were run using the awk programming language in UNIX to identify the reliability of the WHR models based on the number of occurrence points in each polygon of a particular habitat quality ranking. Model reliability was measured by the density of species occurrences in each habitat quality category for each species' WHR model. A model was considered reliable if there was a greater density of occurrences in areas that were considered high quality habitat than in lower quality habitat, which was determined by chi-square and log-likelihood 16  tests. 2.2.3 Identification of Conservation Targets 2.2.3.1 Population Estimates Determining viable population sizes of the twenty-nine species included in this project was not feasible, since there is little known about their life history requirements and demography. Furthermore, since the South Okanagan region is small compared to some of the distributions of the threatened species, it may not be possible to maintain viable populations within the region. Therefore, population viability analyses would have been built on many assumptions and would not have provided reliable estimates. Instead, current population estimates of species within the South Okanagan region and within British Columbia were obtained from literature to determine conservation targets (Appendix V). Current population estimates of species in the Okanagan and British Columbia were obtained from the literature for 16 of the 29 vertebrate species (Table 2.1). The population estimates for the remaining 13 species were based on a combination of estimates from outside of British Columbia, on best guesses given the amount of available habitat within the study area, and on general minimum viable population estimates (Table 2.1). 2.2.3.2 Density Estimates Density estimates were used to calculate the area of habitat required to maintain the current population of each species. These areas were used as the conservation targets for each species. Density, territory and home range sizes for the twenty-nine vertebrate species were obtained from literature, with preference given to studies that occurred in closer proximity to the South Okanagan, since the general trend for all twenty-nine species was a decrease in density with an increase in latitude (Appendix V). The final density estimate for each species was calculated as the mean value from literature sources that provided the 'best' density estimates of South Okanagan populations. The best density estimates were identified from studies that directly measured density and were near or within the South Okanagan. When these types of studies were not available for a particular species, the best density estimates were identified from studies that measured territory or home range size, even if the populations studied were not near the South Okanagan region. The final density estimate for each species was calculated as the mean value from literature sources that provided the best density estimates of South Okanagan populations (Table 2.1). Mean estimates of densities were used if available. However, some studies 17  provided only ranges for density measurements. In those cases, mean density was calculated by averaging the minimum and maximum value in the range. This methodology does not actually calculate the mean value and, therefore, may have resulted in inaccurate estimation of densities. When density estimates of a species were unavailable, territories and home ranges were used. Territories were assumed to give an accurate measure of density, since they are considered to be "an exclusive area". This means that a territory is defined by the degree to which it is used exclusively by its occupant and does not lie in the mechanism (overt defence or any other action) by which the territory becomes identified with its occupant (Pitelka 1959). Density estimates that rely on home ranges are less accurate because many species are gregarious and the degree of overlap of individual home ranges is unknown. Home ranges were used to determine densities for only five of the twenty-nine species, Great Blue Heron , 1  Sandhill Crane, Pallid Bat, Spotted Bat, and Rubber Boa (Appendix V). Because of the uncertainty in home range overlap, I used the assumption that another individual of the same species uses 50% of the home range. Therefore, only half of the reported home range area was used in calculations for density. This area was then treated as a territory. When the data available for a species were limited, it was necessary to use studies that occurred farther away from the study area to calculate density. I included data for a different subspecies or species to calculate the density estimate of eight species (Table 2.1). These other species were of similar body size or mass to the South Okanagan species (Appendix V). Although there were potential errors in these density estimates, they are currently the best available for estimating the area required to maintain a population or group of individuals of the threatened vertebrate species included in this study. 2.2.3.3 Life History Requirements W H R models identify habitat for specific life history requirements, such as breeding habitat, foraging habitat and hibernating habitat (Warman et al. 1998). However, it was not possible to identify conservation targets, in terms of the amount of area required for each life history requirement, because of a lack of detailed information on the ecology of each species. Therefore, i f a polygon contained habitat for more than one life history requirement, the overall habitat area for all life requisites within the polygon was combined. The conservation  ' Latin names and authors of each vertebrate species are listed in Table 2.1.  18  targets were set at representing the overall habitat area that was required for each species within the region. This presented a potential problem if the habitat for 'critical' life history requirements was not represented in the sites selected by the algorithm. Critical life history habitat, which can consist of habitat for breeding, hibernation, denning, or roosting, depending on the species, is where management is necessary because the extent of the habitat in the region is limited (Warman et al. 1998). Therefore, polygons that contained critical life history habitat, according to the WHR models for each species, were identified in the database as having critical life history habitat 'present' within the polygon, but the area of critical habitat was not recorded. Because the area of critical habitat required by a species could not be identified as a conservation target using this database, conservation targets for critical life history habitat were set at a certain number of sites. The conservation target for the number of sites that contained critical habitat was an arbitrary decision based on the estimated population size and ecology of the species (Table 2.3). The number of sites was generally set at 10% of the current population size. However, there were exceptions for some species based on their ecology. 2.2.3.4 Conservation Targets Used in Analyses  The conservation target for each species was identified as the amount of area required to maintain its current population in the South Okanagan based on the density of individuals in suitable habitat (Table 2.3).  19  Table 2.3. Conservation targets based on current population and density estimates for twenty-nine vertebrate species used in the reserve selection analysis.  Species  Critical Habitat (#)  3.01 1.20 36.36 7.48 60.00 30.00 4.89 2.61 4.93 1.16 15.04  N/A 20 20 8 3 30 N/A N/A N/A 10 20 N/A 30  2  Grasshopper Sparrow Great Blue Heron Short-eared Owl American Bittern Ferruginous Hawk Turkey Vulture Canyon Wren Lark Sparrow Bobolink Prairie Falcon Sandhill Crane Yellow breasted Chat Lewis' Woodpecker Long-billed Curlew S a g e Thrasher Western Screech Owl White-headed Woodpecker Brewer's Sparrow California Bighorn Sheep Pallid Bat Spotted Bat Townsend's Bat Blotched Tiger Salamander Great Basin Spadefoot T o a d Rubber Boa Racer 2  Western Rattlesnake Night Snake Gopher Snake £  Target (km )  £  0.93 13.16 38.64 2.76 2.44 181.82 5.78 34.25 22.73 3.10 142.86 30.00 30.00 15.02 5.06 6.06 16.67 7.87  30 6 N/A 20 100 50 10 10 20 30 30 50 50 50 50 50  Individuals (#) 200 200 200 80 6 30 200 200 140 20 200 100 300 170 60 10 200 1000 500 100 100 200 30 pops. 30 pops. 500 500 500 500 500  Populations are selected using a field that identifies individual breeding sites and the area-based target includes terrestrial habitat  2.2.4 Reserve Selection Algorithm Near-optimal solutions for conservation were determined using a heuristic 'Minset' algorithm in C-Plan that identified the minimum amount of area required to achieve the conservation targets for each vertebrate species. The algorithm uses complementarity to select sites that minimise redundancy in the minimum set. C-Plan identifies the conservation value of a site by its 'irreplaceability'. Irreplaceability is the likelihood that a given site will need to be protected to achieve the conservation targets or, conversely, the extent to which options for achieving these targets are reduced if the site is not protected (Pressey et al. 1994). A simple example of irreplaceability can be explained as follows: if a species range is confined to one  20  site, then the site is 100% irreplaceable for conserving that species. Likewise, if a species range covers ten sites and the conservation target is to conserve the species in only one site, then there are ten options for conservation of the species and each site would have a low irreplaceability value. However, the calculation of irreplaceability is considerably more complex. The calculation for irreplaceability is based on a predictive approach that estimates the expected frequency distribution of the area selected by all possible site combinations of a predetermined size for a given species (Ferrier et al. 2000). The expected distribution is used to estimate the total number of these site combinations that achieve the conservation target for the species, which in turn is used to estimate the number of times that a particular site of interest is included in the site combinations. The irreplaceability of a site (site x) for an individual species is expressed as a proportion of the estimated total number of site combinations where site x is included in a set of sites. Irreplaceability (Irr) is calculated as: (-^jr_ included  ^x _ removed)  (R-x _ included  ^j-^  ^-x _ excluded )  where i ^ j n c i u d e d is the number of representative combinations that include site x, R _e c\uded is x  X  the number of representative combinations that do not include site x and ^ r e m o v e d is the number of representative combinations that include site x but would still be representative i f site x were removed (i.e. combinations where site x is redundant). To determine the irreplaceability value for each site based on multiple species, the proportion of combinations that achieve targets for individual species are multiplied together. For example, if the proportion of combinations that include site x and achieve the conservation targets for three species are 0.4 for Species A , 0.5 for Species B, and 0.2 for Species C, then the proportion of combinations that achieve all three species' targets is 0.4*0.5*0.2=0.04.  i?^j iuded n c  is  calculated by multiplying this value by the number of possible combinations of sites that include site x. The final values range from 0 to 1, where 0 represents a low irreplaceability value and 1 represents a high irreplaceability value. The rule sequence in the Minset algorithm used to identify conservation sites was: 1.  Select sites with the highest irreplaceability value.  2.  If there is a choice after Rule 1, select the site with the highest 'summed irreplaceability' 21  value. Summed irreplaceability is the sum of the irreplaceability values for each individual species in a site (NPWS 1999). The values can range from 0 to n, where the upper limit of n is defined by the number of species that are located in a site. For example, if there are 10 species in a site with irreplaceability values equal to 1, then the summed irreplaceability value for this site is equal to 10. If there is a choice after Rule 2, select the site with the "sum of the highest weighted proportion of contributing habitat area to conservation targets for each species". This measure is calculated as the sum of the contributing habitat areas for each species located in site x, expressed as a percentage of the conservation target for each species that can actually be achieved within the region and weighted by the frequency of species i (i.e. number of sites that species i occurs in) within the region (NPWS 1999). Contributing habitat area within each site is calculated based on the amount of remaining habitat required to achieve the conservation targets for each species (i.e. if the habitat area in site x is greater than the area required to achieve the conservation target for species i, then the contributing habitat area in site x is decreased to equal the area required to achieve the conservation target). The sum of the weighted proportion of contribution measure (SumWPrC) is calculated as:  SumWPrC =100* JT  r  x  \^  where  F, _ ntributing_area x  CO  100  F r  ix _ contributi ngarea _ available_target  (2)  _ frequency J J  is the habitat area in site x that is required to achieve the  conservation target for species i, F/avaiiabiejarget is the total habitat area available in all sites that is required to achieve the conservation target for species / and ^ f r e q u e n c y is the number of available sites that contain species i. For example: •  Species A and Species B occur in three sites (Sites X , Y and Z) with a total available 2  2  habitat area in all three sites of 10 km for Species A and 15 km for Species B; •  Site X and Y have equal irreplaceability and summed irreplaceability values, but Site Z_has a lower irreplaceability value than Site X and Y ;  •  Site X and Y have different amounts of habitat for Species A and B, Site X has 2 km 2  2  2  for Species A and 5 km for Species B and Site Y has 6 km for Species A and 3 km for Species B; 22  The conservation target for both species is 8 km ; SumWPrC value for Site X is: (f ^,....2  2km'  SumWPrCx =  10 km'  •x  ,™ > 100  f  3 sites  5  k  m  2  1  x-  15 km  Q  0  \Y  3 sites  x l 0 0 = 1777.78  SumWPrC value for Site Y is: ff  SumWPrC, = • 4.  *,„2 6km'  ^lOkm  •x  inn 100  f  A  3sites j  3  k  m  2  ^15 km  l  2  f  j  0  \ \  3 sites  x 100 = 2666.67  The algorithm selects Site Y .  If there is a choice after Rule 3, select the site with the highest initial irreplaceability value (i.e. the irreplaceability value calculated before the first iteration of the algorithm).  5.  If there is a choice after Rule 4, select the site with the highest initial summed irreplaceability value (i.e. the summed irreplaceability value calculated before the first iteration of the algorithm).  6.  If there is a choice after rule 5, select the first site in the database list. This rule was not applied in the Minimum Set analysis, but is a required tie-breaking rule in C-Plan. Once the algorithm has selected a site, the irreplaceability value of each unselected site  is recalculated based on the species that have not been represented in the selected site(s) and the Minset rules are applied to the database iteratively. This procedure continues until all species have been conserved to their conservation target value or until the predetermined allowable total area for conservation is reached. 2.3 R E S U L T S 2.3.1 Reliability of Vertebrate Species Data 2.3.1.1 Occurrence Data The distribution of records of species occurrence indicates that the data were not compiled from systematic surveys of the entire study area. Most records (68.6%) were within a kilometre of paved roads (695 km , 39% of the study area) and 97.9% of the records were within a kilometre of any (paved or gravel) road (1541 km , 87% of the study area). Paved roads are primarily located below 500 metres in elevation in the study area, except for mountain passes. The 587.8 km area below 500 metres, approximately one third of the study area, contains 51.2%) of the observation records. Occurrence records were located within 500 metres of roads significantly more than 23  Occurrence records were located within 500 metres of roads significantly more than expected from a uniform distribution of points for paved roads (% = 595.76, DF = 1, .PO.OOl) 2  and any (paved or gravel) roads (x = 360.58, DF = 1, PO.001) (Figure 2.2, Appendix VI). 2  Observed frequencies of occurrences within 500 metres of paved roads were greater than expected from a uniform distribution of points. Observed frequencies of occurrences within 200 metres of any (paved or gravel) road, which consisted of 51.1% of the study area, were greater than expected from a uniform distribution of points. However, at distances greater than 200 metres the observed values were less than expected. The non-uniform distribution of observation records could either be a result of biased sampling or because of the lower elevation habitat preferences of the vertebrate species. Without a systematic survey that identifies both the presence and absence of each species, neither hypothesis can be rejected. 2.3.1.2 Wildlife Habitat Relationship Model Reliability The accuracy of the occurrence records is affected by the location of roads as well as by the precision of the recorded location. The highest precision value in the occurrence data was plus or minus 100 metres. Therefore, precision of occurrence records is an issue when the geographical coordinates of species occurrences are overlaid on the T E M vegetation polygons. A n occurrence record mapped in a particular polygon may actually be located in an adjacent polygon as a result of errors in location. A 50 metre distance on either side of each polygon boundary was identified to determine the number of records potentially affected by this problem. The area within 50 metres of the polygon boundaries was 914.99 km , which is 2  equal to 51.7% of the study area. There were 1075 occurrence records located in this area. Therefore, 63.7% of the occurrence records could potentially be mapped in the wrong polygon. Because of these potential biases and errors in the available occurrence records for the region, wildlife habitat relationship models that were based on the literature (i.e. not on the occurrence records) to predict the habitat suitability of sites that were not adequately surveyed, were used in the reserve selection analyses. The occurrence records provided the only data available to examine the reliability of the predictions in the W H R models. Therefore, the reliability of the wildlife habitat relationship models was assessed using the vertebrate species occurrence records, even though the records are potentially biased by the location of roads and may be mapped in the wrong polygons.  24  a) Distance from paved roads 350 • Observed B Expected  300 250 200 150 100 50 0 0-100  "D  o o  100-200 200-300 300-400 400-500  cu  E  b) Distance from any road  3  1200 • Observed H Expected  1000 800 600 400 200 0 0-100  100-200 200-300 300-400 400-500 Distance (metres)  ure 2.2. The observed and expected number of occurrence records that are located within 500 metres of (a) paved roads and (b) any (paved or gravel) roads in the South Okanagan region.  25  Because there is potential for error when counting or estimating the number of individuals, depending on observer experience (Boulinier et al. 1998), the number of occurrence records were used to assess the reliability of the WHR models, disregarding the number of individuals that were recorded in each occurrence record (Figure 2.3). The majority of the occurrence data for California Bighorn Sheep were collected using radiocollared individuals. The number of records for radio-collared bighorn sheep per k m was 2  determined by counting an individual only once per polygon, to eliminate potential sampling biases, which result in unequal number of observations of radio-collared sheep. Predictions of habitat suitability for each vertebrate species from the wildlife habitat relationship models were reasonable, given the limitations of the occurrence data, since densities of occurrence records for all taxonomic groups were highest in high quality habitat (Figure 2.3, Appendix VII). The densities of occurrence records were significantly higher within the habitat quality categories identified by the wildlife habitat relationship models than expected from a uniform distribution of points for each species (chi-square or log-likelihood ratio tests, P O . 0 5 , Appendix VIII). Neither chi-square nor log-likelihood ratio tests could be performed for three species because there were not enough occurrence records. The species were American Bittern with 1 record in high quality habitat, Ferruginous Hawk with 2 records in low quality habitat, and Short-eared Owl with 4 records in moderate quality habitat. 2.3.2 Minimum Set of Sites for Conservation The W H R models identify that the highest species richness areas were in the southern portion of the region (Figure 2.4, Appendix IV). Initial irreplaceability values calculated by C-Plan were influenced not only by species richness of the region, but also by the conservation target for each species relative to habitat available for selection within the region (Table 2.3). At each iteration of the algorithm, irreplaceability was recalculated based on species that were not represented in selected sites. The resulting minimum set of sites representing enough habitat to maintain current population sizes of the twenty-nine threatened vertebrate species consisted of 583.99 km (37.2% of the region) (Figure 2.4). The mean 9  patch size of the sites in the minimum set was 1.34 km and the median was 0.37 km (SD=6.10).  26  9  a) 4-class habitat rating 1.0 • Amphibians (2 spp 0.8  -I  El Reptiles (5 spp.) • Birds (13 spp.) E3 Bats (3 spp.)  0.6  0.4 -I CD  O O CU  0.2 0.0  -I High  Moderate  Low  Nil  DC  M — o c g o Q. O c  b) 4-class habitat rating (best habitat is outside region) 1.0 • Birds (5 spp.) 0.8  CO  cu  0.6  0.4 ^ 0.2 X 0.0 Moderate  Low  Nil  c) 6-class habitat rating 1.0 0 California Bighorn Sheep  cu  0.8  Q.  s CO  T3  8  CD  c g  t. o Q_ O  0.6  0.4 0.2 0.0 Very  High  Moderate  Low  High  Very  Nil  Low  W H R Model Habitat Rating  Figure 2.3. The mean proportion of occurrence records of vertebrate species per square kilometre and standard deviation in habitat identified by the (a) 4-class habitat rating, (b) 4-class habitat rating, where high quality habitat is located outside of the region, and (c) 6-class habitat rating, for California Bighorn Sheep only, in the wildlife habitat relationship models.  27  28  2.4 D I S C U S S I O N 2.4.1 Issues Related to Reserve Selection The process of identifying priorities for conservation, using systematic reserve selection software at fine spatial scales, has inherent difficulties. The research in this chapter identified that the reliability of species occurrence data must be determined, since the sites selected by the reserve selection algorithm are dependent on data quality. Reliability of selected sites is increased when detailed ecological data on species or biological attributes are included in reserve selection analyses. However, these data are usually unavailable (Cabeza and Moilanen 2001). Alternatively, predictive models of suitable habitat for each species can provide useful data for conservation planning in areas that have not been adequately surveyed. However, the accuracy of the reserve selections using these types of data cannot be determined analytically, since propagation of errors in both the assumptions of W H R models and in the vegetation base map through the reserve selection algorithm are not well understood (Stoms 1994). Field checking of selected sites is then required to obtain empirical values of reliability for protecting suitable habitat for each species. 2.4.1.1 Reliability of the Species Data Potential errors in predictions of species occurrence has fueled the argument that conservation areas should be selected on the basis of observed records of occurrence rather than on predictions, because of the added uncertainty to reserve selection (Araujo and Williams 2000). However, in the South Okanagan the reserve selection algorithm would be forced to select sites close to roads if observed records of occurrence were used (Figure 2.2). Another potential bias in occurrence data results from preferential visits by observers to relatively pristine areas (Prendergast et al. 1993b). In the South Okanagan, the Vaseux Lake Wildlife Area draws many naturalists and biologists because of wildlife known to occur in the area. The number of potential observers is higher in these pristine areas, amplifying the number of occurrence records compared to those in less frequented areas. Therefore, an atlas or inventory effort of species occurrence that alleviates the gaps in available data is an important preliminary step in design of reserve selection studies. Wildlife habitat relationship models tend to overestimate potential habitat for species within a region (Stoms 1994). Since the models are predictive, they cannot determine whether a particular site is currently or typically occupied by that species (Stoms 1994).  29  Correlations between the W H R models and density of occurrences were reasonably good in high quality habitat, however in lower quality habitat correlations were less obvious (Figure 2.3). The density of occurrence records did not decrease with a decrease in habitat quality for reptiles, birds and, in particular, bats. The fine scale of bat habitat, consisting of crevices in rocks for maternity roosts, day roosts and hibernation, could not be mapped at the 1:20,000 scale of T E M , which may be the reason for observed high bat densities in lower quality habitat. Therefore, to increase reliability of habitat models for bats, finer mapping resolution of cliffs, rock outcrops and talus is required. Because of these errors in predicting habitat quality, higher absolute numbers of species are calculated with W H R models than what the region can actually maintain. Although 'experts' on each species reviewed the assumptions and habitat maps of the W H R models used in this thesis, each species map should be groundtruthed (i.e. verified with field surveys) to determine the reliability of the models (Warman et al. 1998). There are three additional reasons why the species maps must be ground-truthed. First, the terrestrial ecosystem map, providing the base map for the W H R models, was developed over a decade ago. Agricultural and urban areas have expanded over this time period and the map may not depict current conditions of the landscape. Secondly, the assumptions of habitat quality in the W H R models were based on the characteristics of each vegetation unit. A n individual vegetation unit may have different characteristics, depending on its location within the region, that were not described and the accuracy of the mapping is unknown. Many species require a particular plant species or sub-species for their survival. If there is variation in the plant species composition of a vegetation unit, then the required plant species may not be present at every location where that the vegetation unit is mapped. Thirdly, the scale of the mapping may not identify the vegetation that each species requires. Paczek (in prep) identified five grassland bird species in the South Okanagan that were associated with plant species that could not be mapped at the 1:20,000 scale of the T E M . Thus, the potential for overestimating the amount of habitat for a species within a region would decrease i f each species map were ground-truthed. 2.4.1.2 Viability of the Reserve Network Wildlife habitat relationship models predict where species could occur, but do not predict the amount of habitat that is required to maintain species over time. When information necessary to conduct population viability assessments is not available, as was the 30  case in the South Okanagan, conservation targets need to be estimated using alternative methods. There is potential for error with any estimate of population viability, but it increases when the biology of the species is not known or considered (Shaffer 1981). Although current population sizes of the twenty-nine threatened species were used as the conservation target for reserve selection, the populations are probably not viable because the designation by the province as either Red or Blue listed implies that their habitat or population numbers are decreasing. Therefore, the reserve network identified in this chapter will likely not maintain viable populations of the twenty-nine vertebrate species within the region. The minimum set of sites required to achieve the conservation targets in the South Okanagan resulted in many habitat patches that were too small (median=0.4 km , SD=6.1) to 2  maintain life requisites of most species. Therefore, the size of the selection units (sites) used to identify areas for reservation was too small for the systematic reserve selection procedure. There are at least two potential solutions to this problem. One solution is to include a rule in the algorithm to select sites that aggregate into large enough patches to maintain viable populations of species. This rule would break ties between sites by selecting the site that is closest to previously selected sites. This component of the algorithm is currently being developed for C-Plan, but was not available for present analyses. Another potential solution would be to increase the size of the selection units. This could be done by aggregating the polygons in the T E M into larger polygons that consist of similar types of vegetation, or by overlaying a uniform grid over the region. However, both methods decrease map resolution, which decreases the reliability of sites identified by the reserve selection algorithm (Stoms 1994, Heywood et al. 1998). The most difficult problem to solve is how to identify a network of sites that achieves the habitat and spatial requirements for each individual species. Systematic reserve selection identifies sites that achieve the overall conservation targets for species. However, the algorithms are not able to accommodate species-specific spatial requirements or consider interactions between species. The spatial configuration of selected areas is important for species, such as reptiles and amphibians, which cannot migrate long distances. They migrate seasonally between their denning and/or hibernation sites to foraging and breeding areas. If large distances separate the two habitat types, then these sites do not aid in conservation of these species. There are potential negative interactions between species when competition and predation are considered. If a site is selected for two species where either of these 31  interactions can occur, then the site may be inadequate at conserving both species (Witting et al. 2000, Schoener et al. 2001). Although reserve selection algorithms have been developed to avoid species-specific conservation, it is still necessary to include species-specific requirements in the process if viable reserve networks are to be achieved. 2.4.2 Conclusions and Recommendations The polygons in the terrestrial ecosystem mapping did not provide selection units for reserve selection that were large enough to maintain the life requisites of species within selected sites. Because the size of the selection unit affects the organisation of species data and ultimately the spatial solutions, different selection units will likely result in different selections of sites that achieve the same conservation targets (Stoms 1994). A selection unit size that is appropriate for conservation in the study area should be identified, along with the variation in reserve selection that results from changes in the selection unit size. In this chapter, the threatened vertebrate species were used as surrogates for biodiversity in the region. Reserve selection algorithms identify different conservation areas depending on the species included in the analyses (Prendergast et al. 1993 a). If the overall goal for conservation in the South Okanagan is to protect all biodiversity in the region, the adequacy of using threatened vertebrate species as surrogates will need to be assessed. Other species or groups of species may provide better surrogates and could decrease the number of species, and subsequently the amount of biological data, that need to be considered in the conservation planning process (Freemark et al. 2000, Cabeza and Moilanen 2001). Therefore, careful consideration needs to be given to the species that are included or excluded from the reserve selection technique. Conservation targets for each species are difficult to define for the South Okanagan because of the geopolitical boundaries of the region. Many species interact with populations that are located outside the region. Therefore, it may not be possible or necessary to represent viable populations of some threatened species within the region. Since definite conservation targets cannot be determined until population viability assessments are completed, conservation targets should be set at conservative values and altered to determine the variability of the location and amount of area in the reserve network that achieves the targets. The existing reserve network in the South Okanagan was selected in an opportunistic manner, which increases the overall cost of representing biodiversity in reserve networks (Pressey 1994). The existing reserve network will likely form the foundation for additional 32  protected areas. Therefore, the adequacy of the existing reserve network for conserving biodiversity in the region should be determined. Systematic reserve selection algorithms can be used to supplement the existing reserve network with complementary sites that represent species that are not adequately protected (Margules and Pressey 2000). The relative costs of supplementing the existing reserve network versus starting from scratch can be assessed in terms of the amount of land required to achieve conservation targets for each species (Pressey 1994). If opportunistic acquisition and protection of habitat continues, the overall cost may inhibit the protection of all regional biological diversity in a reserve network. Therefore, supplementing the existing reserve network with complementary sites provides an economical strategy. The region is undergoing rapid development and agricultural conversion, which can drastically alter the predicted habitat quality values for species in the wildlife habitat relationship models. Therefore, any future conservation plans for the South Okanagan region that are based on the reserve network identified in this chapter need to consider the current condition of the landscape. Although ecological data are limited for threatened vertebrate species in the South Okanagan and the terrestrial ecosystem map may be outdated, the lack of good quality and detailed data should not be used as an excuse to delay habitat acquisition or protection. The reserve network identified by this project provides a reasonable direction for conserving threatened vertebrate species in the South Okanagan region i f the issues identified in this chapter are heeded.  33  3.0 CHAPTER 3: HOW DOES T H E IDENTIFICATION OF PRIORITY CONSERVATION SITES DEPEND ON CRITERIA USED FOR CHOOSING THEM? 3.1 INTRODUCTION The impacts of biodiversity loss on functional ecosystems are not completely understood. However, most scientists agree that natural areas must be protected to decrease the rate of biodiversity decline (Pressey and Tully 1994, Church et al. 1996, Arcese and Sinclair 1997, Margules and Pressey 2000). Systematic reserve selection procedures have been developed as a tool to aid in identifying priority conservation areas in light of limited knowledge and understanding of ecological systems (Pressey 1994, Pressey et al. 1997, Brooks et al. 2001). The advantage of systematic procedures is that conservation goals and rules are explicit and the selection process is repeatable (Bedward et al. 1992). However, there are uncertainties in systematic reserve selection related to the criteria used to select areas for reservation. Systematic procedures require decisions on various aspects of scale that include both spatial and ecological relations. Values assigned to parameters that are associated with scale influence spatial analyses of ecological data (Stoms 1994). Therefore, it is important to determine the sensitivity of systematic reserve selection techniques to variation in parameter values. In this chapter, effects of variation in parameter values on priority site selection are evaluated using a reserve selection algorithm called C-Plan (NPWS 1999). Parameters that have a large influence on the priority site selections in the reserve selection algorithm are (1) selection units, which provide the sites that are available for selection, (2) flora and fauna that are included in the algorithm as indicators of biodiversity, and (3) conservation targets or goals for flora and fauna within a region. Selection units provide the framework for compiling data on the occurrence and distribution of flora and fauna within a region (Pressey and Logan 1998). Determining the size and shape of selection units that are appropriate for conservation within a region can be arduous, since there is no strong theoretical basis for using a specific selection unit (Stoms 1994, Pressey and Logan 1998). However, i f the probability of implementation is considered, selection unit size and shape should attempt to minimise the associated costs of reservation (Pressey and Logan 1998).  34  3.1.1 Choosing Selection Unit Size and Shape Priority conservation sites identified by reserve selection algorithms are strongly influenced by the size and shape of the selection unit (Stoms 1994, Pressey et al. 1999). Small selection units achieve conservation targets for flora and fauna within a region more efficiently, in less area, than larger units (Pressey and Logan 1995). However, very small selection units can be indistinguishable by indices of conservation value and isolated units may not be large enough to conserve the flora and fauna within them. Amalgamating small selection units during the selection process to produce areas that are large enough to maintain flora and fauna prevents the isolation of small areas. However, preferentially selecting units that are in close proximity to previously selected units results in a greater total area than priority sites that have not been amalgamated (Bedward et al. 1992, Nicholls and Margules 1993, Lombard et al. 1995). Because there are no strict guidelines for choosing a selection unit, justification in published studies for using a particular selection unit have been based on different reasons (Lombard et al. 1995, Price et al. 1995, Kiester et al. 1996, Pressey et al. 1997). The problem is that each method of justification alters the selection unit and consequently the relative conservation values within the region, which ultimately influences the selection and configuration of priority areas for conservation. Shape of the selection unit also influences priority site selection, although this may be less crucial than selection unit size if the boundaries are altered during implementation (Pressey and Logan 1998). Polygons provide a selection unit of irregular shape and size with boundaries that are normally defined by physical features on the landscape. Selection units with irregular area present a problem in reserve selection because large units tend to have more habitat for each species occurring in that unit and potentially have a greater number of species than smaller selection units (Rickets et al. 1999). This problem is apparent when a large selection unit with low quality habitat for species is selected over a smaller unit with higher quality habitat. Grids consist of cells of uniform shape and size, with boundaries that are not related to landscape features (Fotheringham 1989, Stoms 1994). Most published systematic reserve selection studies used grid cells as selection units, which minimise the effect of unit area on algorithm selections (Nicholls and Margules 1993, Lombard et al. 1997, Pressey and Logan 1998, White et al. 1998, Freemark et al. 2000). Although grid cells decrease the accuracy of ecological data because of the arbitrary boundary placed on the landscape (Dobson et al. 1997, Heywood et al. 1998, Rickets et al. 1999), they provide a 35  simple method for mapping data and evaluating the effects of selection unit variability on reserve selection. 3.1.2 Choosing Indicators of Biodiversity Most decisions in reserve selection studies were based on locations of flora and/or fauna, which provide an indicator of the biodiversity present within the region. The most accepted measure of biodiversity are species, since they provide a biologically natural unit with little taxonomic uncertainty and can be grasped intuitively by most people (Caughley and Gunn 1996, Myers et al. 2000). Detailed information on locations and potentially life history requirements is usually only known for endangered and threatened flora and fauna, since comprehensive surveys of biodiversity are expensive and time-consuming (Ferrier and Watson 1997, Howard et al. 1998, Balmford and Gaston 1999, Ricketts et al. 1999). These species have the greatest risk of extinction or extirpation and, therefore, provide a reasonable starting point for identifying priority sites for conservation. Protected areas that are selected based on habitat requirements of endangered species and more easily studied taxa are assumed to provide habitat for other taxonomic groups (Pearson and Cassola 1992, Scott et al. 1993, Ricketts et al. 1999, Margules and Pressey 2000). However, there have been opposing results from studies on species congruence among different taxa. Studies have found that there is little congruence between species rich areas of different taxa, between hotspots of rare species of different taxa, and between species hotspots and sites identified using systematic reserve selection (Prendergast et al. 1993a, Williams et al. 1996, Dobson et al. 1997, Lawton et al. 1998, van Jaarsveld et al. 1998). In contrast, other studies have found that hotspots for plants and vertebrates coincide, combined distributions of well-studied taxa provide reasonable surrogates for biodiversity, and systematic reserve selection based on one taxonomic group represent important sites for other taxa despite low spatial congruence between taxa (Howard et al. 1998, Ricketts et al. 1999, Freemark et al. 2000, Myers et al. 2000). The overall goal of these studies was to determine whether one or a combination of taxonomic groups would behave as a surrogate for other taxa, but not whether endangered or threatened species, which are a subset of a region's biodiversity, are surrogates for a larger group of organisms. Therefore, it is important to determine if systematic reserve selection techniques can successfully identify conservation areas for the biodiversity of a region when only a subset of the biodiversity is used to identify priority conservation sites.  36  Since the location of populations of species is a function of their habitat distribution, the location of distinct vegetation types or ecosystems may provide a surrogate or indicator for species diversity (Scott et al. 1993, Ricketts et al. 1999). If this relation is true, a relative estimate of biodiversity could be assessed in regions based on vegetation when there is limited information on species occurrences and habitat requirements (Pressey and Logan 1995). Reserve planning at broad regional scales has used environmental surrogates ranging from landform-vegetation classes to environmental domains (Pressey and Nicholls 1989b, Bedward et al. 1992, Awimbo et al. 1996), with the assumption that most species will be represented in the reserve network if the variation in the environmental surrogate is represented (Nicholls and Margules 1993). There are some sound theoretical reasons as to why environmental variables are good predictors of the spatial distribution of species (Austin 1985) and good empirical evidence that provides support (Austin et al. 1990, Wessels et al. 1999). In contrast, there is empirical evidence, although limited, that abiotic environmental variables do not correlate with species distributions (Ferrier and Watson 1997, Currie 1991). The congruence of sites identified to represent ecosystems with those identified for species has not been examined at fine spatial scales with more precise mapping of vegetation structural stages. Therefore, at fine scales, ecosystems or vegetation types may provide useful surrogates for conserving biodiversity. 3.1.3 Choosing Conservation Targets The goal of most conservation programs is to conserve viable populations of species for a specified period of time into the future (Shaffer 1981, Soule 1987). Minimum viable populations are calculated using population viability analysis, which is based on the demography and genetic variation of populations, as well as environmental stochasticity and natural catastrophes (Shaffer 1981, Soule 1987). However, for most species this information is not known. Therefore, general estimates of minimum viable populations have been proposed. These estimates include values that range from fifty individuals to populations that consist of individuals that number in the low thousands (Franklin 1980, Soule 1987, Caughley and Gunn 1996). Conservation decisions that are based on these general estimates must be used cautiously, since it is not known whether any of these estimates actually represent viable populations of species. In addition, the total area of the region influences the conservation target that can be achieved for a species. The South Okanagan study area is small in comparison to some of the distributions of species included in this thesis, and is defined in the 37  south by a political boundary rather than an ecological boundary. Because many of the species are migratory and interact with individuals across the border, the confines of the region may prevent the conservation of viable populations. These two problems demonstrate that choosing conservation targets for systematic reserve selection can be somewhat arbitrary. Therefore, it is necessary to determine the sensitivity of reserve selection procedures to variation in conservation target values. The identification of conservation targets for species also depends on how the species distribution data were recorded. Often reserve selection studies are forced to use categorical data of species presence or absence because of the lack of more detailed habitat suitability data (van Jaarsveld et al. 1998, Freemark et al. 2000, Brooks et al. 2001). Conservation targets based on presence-absence data are constrained to representing a specified number of selection units instead of representing the amount of area required to conserve a species. Most conservation targets in these studies select at least one selection unit for each species in the priority sites identified by the reserve selection algorithm (Pressey et al. 1997, Dobson et al. 1997, Freemark et al. 2000). However, it is erroneous to assume that because a particular species occurs in a selection unit that a viable population can be maintained in that unit (Dobson et al. 1997). Conservation targets based on presence-absence data that are set at a particular number of units usually will underestimate the amount of land necessary to conserve species with large area requirements. Therefore, at fine spatial scales, data that identify the amount of area available for a species in each selection unit may be more appropriate for identifying priority conservation sites. 3.1.4 Objectives The overall purpose of this chapter was to identify the variation in the priority conservation sites selected by a systematic reserve selection algorithm when the values for selection units, biodiversity indicators and conservation targets were altered. The research objectives were to (1) determine if priority conservation sites that are selected by a systematic reserve selection algorithm are congruent across a range of selection unit sizes and shapes, biodiversity indicators and conservation targets, (2) identify the "best" selection unit size, i f any, for selecting conservation sites within the region, (3) evaluate the surrogacy of the priority conservation areas for flora and fauna that were not included in the initial algorithm, and (4) identify priority conservation sites that occur in all priority sets of sites resulting from variation in parameter values. The results may determine whether systematic reserve 38  selection techniques are useful for conservation at fine spatial scales of mapping given the uncertainty in assigning values for algorithm parameters. Although the objectives of this chapter are based on using a systematic reserve selection algorithm, the results have implications for any reserve selection plan whether or not a systematic selection algorithm is applied. 3.2 M E T H O D S 3.2.1 Study Area and Data Sets The study area is defined by the South Okanagan Conservation Strategy, and has been described in Chapter 2 of this thesis (Figure 2.1, Hlady 1990). The 1:20,000 scale terrestrial ecosystem map produced by M E L P provided the base map for reserve selection analyses and has been described in Chapter 2 (Lea et al. 1991). The data used to determine the surrogacy of priority sets of sites for actual locations of species are Red and Blue Listed vertebrate (Harper et al. 1994) and plant species occurrences (Appendix IX) maintained by the Conservation Data Centre (CDC) and by the Penticton M E L P office, and rare invertebrate inventories (G.G.E. Scudder, unpubl. data). The vertebrate and plant species occurrence data consisted of known locations of species that have been identified by a range of observers that include the general public, naturalists and both government and non-government biologists. Therefore, accuracy and reliability of these data are variable. The rare invertebrate inventory data were obtained by general collecting and from pitfall traps located in a range of sites in the southern portion of the study area (Appendix X). A l l determinations were by entomologists at the University of British Columbia (UBC). The ecological data that identify the conservation value of individual selection units consisted of habitat suitability identified by wildlife habitat relationship models (Verner et al. 1986) and distributions of broad vegetation classes. These data are referred to as vertebrate species and vegetation classes, respectively, in this chapter. There are 97 vegetation associations in the T E M database that have been grouped into fifteen broad vegetation classes (Holm 1998). The fifteen vegetation classes used in the reserve selection analyses include a classification for cliffs, rock outcrops, talus, ponds and shallow open water habitats (Appendix I). There are twenty-nine WHR models for Red and Blue Listed vertebrate species  39  located in the study area (Appendix IV), which predict the amount of habitat available within each T E M polygon for each species (Warman et al. 1998) and are described in Chapter 2.  3.2.2 Reserve Selection Algorithm Systematic reserve selection software, called C-Plan, recently developed by the National Parks and Wildlife Service in New South Wales, Australia identifies priority conservation areas in landscapes that are subject to effects of human development (NPWS 1999). Priority conservation sites are identified using the heuristic 'Minset' algorithm in CPlan. The Minset algorithm uses complementarity methodology to identify the near minimum amount of area required to achieve conservation targets for vertebrate species and vegetation classes in the region. The Minset algorithm rule sequence used to identify priority conservation sites is described in Chapter 2 of this thesis. The procedure included a check for redundant sites after each set of ten iterations of the algorithm and identified whether the removal of initially selected sites cause the representation of any vertebrate species or vegetation class to fall below their conservation target. If none fall below target, then the selection unit is removed from the priority sites and is available for selection at later stages in the algorithm.  3.2.3 Sensitivity of Priority Site Selection to Variation in Parameter Values There are three input parameters in the C-Plan Minset algorithm that influence priority conservation site selections. A l l of the calculations in the algorithm are based on the matrix of selection units, biodiversity indicators (vertebrate species and vegetation classes) that occur in each selection unit, and conservation targets for each vertebrate species and vegetation class (NPWS 1999). Selection units can be any area of land or water and are used for both reserve selection and display. Conservation targets are identified for each vertebrate species and vegetation class and are based on the conservation goal for the region. To examine the variability in priority sites identified by the algorithm, values for each parameter were altered while keeping values for the other two parameters constant.  3.2.3.1 Selection Unit Sensitivity of the C-Plan algorithm to variation in size and shape of the selection unit was examined by comparing the distribution of priority sites that resulted from each variation. Selection unit variation included in the analyses consisted of T E M polygons and three sizes of hexagons. T E M polygons provide 10,125 selection units that range in size from a portion of a 2  2  small riparian island at 2.16E-04 km to an agricultural area at 72.36 km , when urban and 40  large lake polygons are excluded from the database. Hexagons have commonly been used as grid cells for reserve selection because they provide a systematic, hierarchical, equal-area sample unit that minimises spatial distortion (White et al. 1992) and they have smaller perimeter to area ratios than square grid cells of the same size. The smallest hexagon size used in the analyses was 0.155 km , which is the mean size of the T E M polygons when large 2  lake and urban polygons are excluded. This hexagon size results in 10,935 selection units for the study area (Figure 3.1). The intermediate hexagon size was 2 km , resulting in a total of 2  963 selection units for the study area (Figure 3.1). The 2 km hexagon represents the mean 2  size (2.33 km ) of existing areas that are either provincial parks or are managed for conservation by M E L P and private organisations. This value was calculated using spatial data of land ownership for the region obtained from M E L P in October 2000. Polygons identified in the land ownership data could not be used as selection units because the data were incomplete for the study area. The largest hexagon size was 10 km , which results in 2  233 selection units for the study area (Figure 3.1). The 10 km hexagon size was chosen 2  because Ferruginous hawks occur at 0.1 individuals per square kilometre, which is the lowest density of the twenty-nine vertebrate species included in the analyses. Therefore, I assumed that 10 km is required to support the life requisites of one hawk. Because Ferruginous hawks are not known to be territorial, it is not possible to estimate accurately the amount of contiguous area required by exclusively by one hawk. This assumption does not consider the area requirements of a mating pair of hawks. However, the conservation target was never set at representing just one hawk in the region. Because Ferruginous hawks are wide ranging, it 2  ,  2  was assumed that the 10 km patch, in combination with other 10 km patches, whether or not they are adjacent, would provide enough habitat to maintain breeding pairs of hawks. Hexagons were created using Arclnfo 8.0.2 set at double precision and a program that was developed in Arc Macro Language (Lewis 2000). The hexagons were merged with the T E M polygons using the Union command in Arclnfo and the vertebrate species and vegetation class data located in the polygons were used to populate the hexagonal databases using the awk programming language in UNIX.  41  42  Sensitivity of the reserve selection algorithm to variation in selection unit size and shape was determined by using all twenty-nine vertebrate species with conservation targets that represent habitat area for fifty percent of the current population estimates of each species within the region. 'Fifty percent of the current population estimates' is referred to as '0.5populations' in the remainder of this chapter. The target was arbitrarily set at 0.5-populations because the identification of priority areas that represent one hundred percent of the current population estimates of each species requires 37.2% of the area within the region. This proportion of the study area was too large to adequately compare variation in resulting sets of priority sites. 'One hundred percent of the current population estimates' is referred to as ' 1.0populations' in the remainder of this chapter. 3.2.3.2 Biodiversity Indicator The research and management focus of the provincial government has been on Red and Blue Listed species in the study area. Therefore, WHR models for twenty-nine Red and Blue Listed vertebrate species were used to identify priority conservation sites. Although this information provides the largest set of data on biological diversity found in the region, managers are often forced to look at subsets of these data. The Species at Risk Act (Bill C-5), introduced by the federal government in 2001, focuses primarily on endangered species identified by COSEWIC, which include provincial Red Listed species. Therefore, one of the variations in the algorithm included WHR models for only Red Listed species to determine i f the subset of species can adequately represent the habitat area requirements of Blue Listed species. There are eleven Red Listed species and eighteen Blue Listed species (Table 2.1). The conservation targets represented the habitat area that maintains the 0.5-population of each of the twenty-nine vertebrate species within the region. Since ecosystems or vegetation types may provide surrogate data for biodiversity, they provide another indicator of biodiversity that can be used to explore variation in priority sites selected by C-Plan (Ricketts et al. 1999). Because ecosystems are difficult to define, the fifteen vegetation classes identified for the region (Table 3.1, Appendix I), which provide an approximation for ecosystems (Pressey and Nicholls 1989b), were used to identify priority sites. Since guidelines for identifying conservation targets for ecosystems have not been developed (Pressey and Logan 1995), targets for vegetation classes were identified as 27%> of the existing area of each vegetation class in the T E M database. The sum of these targets for vegetation classes approximates the sum of the area targets that represent habitat for the 0.543  populations of the twenty-nine vertebrate species. Vegetation class targets were calculated for each range condition of shrubsteppe and grassland and for each successional stage of forest. There were three exceptions to the target calculation of vegetation classes (Table 3.1). Two of the exceptions were for antelope brush and grassland vegetation classes. The existing area of excellent and good range conditions, which are desired in conservation plans, was less than the area of poor range condition in the T E M database for these two vegetation classes. The third exception is for the riparian vegetation class. The existing area of the riparian shrub-herb successional stage in the T E M database was greater than the riparian old growth successional stage. Shrub-herb stages are not considered to be high priorities for conservation in the region, since they are generally products of forest harvesting. Therefore, conservation targets for excellent and good range conditions and the old growth successional stage of these three vegetation classes were adjusted to represent greater than 27% of their existing area in the T E M database. The targets for undesirable range conditions and successional stages were decreased to represent less than 27 percent of the existing area. Overall, representation of vegetation classes, disregarding range condition and successional stage, was set at 27% of the existing area for each vegetation class in the database. The efficiency of identifying priority sites for diverse indicators of biodiversity was examined by including both the vegetation classes and the Red and Blue Listed vertebrate species in the algorithm simultaneously. The resulting priority sites were compared with priority sites that represent vegetation classes and vertebrate species independently. The selection unit used for these analyses was the 10 km hexagon. This size of selection unit provides the most biologically relevant unit for conservation, which is especially important when a selection unit is isolated from other priority conservation sites. This unit provides a contiguous area that is large enough to provide habitat to support the life requisites of all threatened vertebrate species considered in this thesis.  44  Table 3.1. Conservation targets of vegetation classes and their associated range conditions and successional stages in the TEM database. Vegetation Class  Range Condition  Grassland Grassland Grassland Grassland  Poor Fair Good Excellent Poor Fair Good Excellent  Antelope Brush Antelope Brush Antelope Brush Antelope Brush Big Sagebrush Big Sagebrush Big Sagebrush Vasey's Sagebrush Vasey's Sagebrush Ponderosa Ponderosa Ponderosa Ponderosa Ponderosa  Engelmann Engelmann Engelmann Engelmann Engelmann  (Lodgepole (Lodgepole (Lodgepole (Lodgepole  Spruce Spruce Spruce Spruce Spruce  Broad-leafed Broad-leafed Broad-leafed Broad-leafed Broad-leafed  and and and and and  3.73* 4.29 3.73** 0.24**  1.14 1.41  6.44  Shrub/Herb Pole/Sapling Young Mature Old-growth  9.31 1.32 17.55 57.32 38.66  Shrub/Herb Pole/Sapling Young Mature Old-growth  3.72 1.95 12.41 26.97 5.70 0.04 1.51 2.35 0.48  Shrub/Herb Young Mature Old-growth Fir Fir Fir Fir Fir  Shrub/Herb Pole/Sapling Young Mature Old-growth  Communities Communities Communities Communities Communities  Shrub/Herb Pole/Sapling Young Mature  -  Riparian Vegetation Riparian Riparian Riparian Riparian Riparian  0.57* 9.52 42.59 0.57**  Fair Good  Pine) Pine) Pine) Pine)  Sub-alpine Sub-alpine Sub-alpine Sub-alpine Sub-alpine  2  15.85 25.18  Pine Pine Pine Pine Pine  Spruce Spruce Spruce Spruce  Conservation Target (km )  Poor Fair Good  Douglas-fir Douglas-fir Douglas-fir Douglas-fir Douglas-fir Hybrid Hybrid Hybrid Hybrid  Successional Stage  Shrub/Herb Pole/Sapling Young Mature Old-growth  Vegetation Vegetation Vegetation Vegetation Vegetation  -  Wetlands Open water (small lakes, ponds) Rock Outcrops Cliffs Talus  0.02 0.06 0.56 1.71 0.16 0.11 1.71 3.35 0.16 0.11 0.60* 0.75 3.50 5.86 0.60** 2.10 2.07 2.07 23.32 2.38 17.16  Exceptions: * Adjusted to less than 27% of the total existing area in TEM database. Adjusted to greater than 27% of the total existing area in TEM database. 45  3.2.3.3 Conservation Target Conservation targets identify the conservation value of individual selection units during the selection process (Margules and Pressey 2000). Four different conservation targets were used to investigate the variability in priority sets of sites produced by the C-Plan algorithm. The targets included three area-based targets that represent the area of species habitat within a selection unit and one site-based target that represents species presence within a selection unit. The three area-based targets represent habitat area for (1) the current population estimates of the twenty-nine Red and Blue Listed species (1.0-populations), (2) fifty percent of the current population estimates for the twenty-nine species (0.5-populations) and (3) general estimates of minimum viable population (MVP) size for the twenty-nine species (Table 3.2). The WHR models were used to identify the amount of suitable habitat in each selection unit for each of the twenty-nine species. Since minimum viable population estimates are potentially in the low thousands (Soule 1987), the M V P conservation target used in this chapter for amphibians and reptiles was 1000 individuals. Amphibians and reptiles are resident species with localised ranges within the region. Migratory species, consisting of birds and mammals, interact with populations outside of the region and are distributed over a larger area compared to resident species. The study area is not large enough to maintain populations of 1000 individuals of Red and Blue Listed birds and mammals. Therefore, M V P conservation targets for migratory species represent habitat for 500 individuals. The amount of habitat required to maintain life requisites for each of the population estimates was calculated using known densities of each of the species in suitable habitat (Table 2.1). The density estimates were also used to identify site-based targets that represent species presence within a selection unit. The relative importance of a selection unit for a species, in terms of the amount of habitat that is suitable within a unit, is lost when identifying the conservation value of a selection unit by species presence within a unit. Therefore, the assumption for site-based targets was that the entire area of a selection unit provides habitat for each of the species that are present in the unit. The 10 km hexagons were used as selection units in these analyses. The indicator of biodiversity for each analysis was the twenty-nine Red and Blue Listed vertebrate species. The number of individuals of a species potentially occurring in a selection unit, based on the assumption that the entire 10 k m of a 2  unit provides suitable habitat for each species, was calculated using the density estimates (Table 2.1). 46  M o  CO 3 "D ">  CO CO  CN  o  00 Q. O Q.  00 00  00  TJ  c  CO  CD O CD O. CO  0)  CO "CD  C3)  -Q  c -*.  o| ro  co CO Q. O Q.  co 3  -g 0. > OJ  -0  to  CD  TJ  C  CO CD  CO  >  CN LO CD  CN LO  CN  o  CO CJ)  col CN  LO CN  Tfr  LO  s  O O O  CO  LO  _g  CO CO CO CO  CO o d  CO  o  " 3 3 CO  ro  00 CL o CL  O  o  3  .21 •g  ro  o  TJ CD CO  J2 ro C  s  o oi  T>  J3  ro CD  1  ro  >  CD TJ  C  c ro  CL  o o  LO  CD  ci  CD CO  cn 00  o  CO LO  CO CD  d l  O)  CO  CO  c TJ  CD  0  tt  tn  ro  CD  c 3 .21 •g  I  JS  CO Q. O CL  o o CN  O o CO CO  > TJ C  ro 3 •g CO > CD  CL  £  TJ  in  0.  CN  CO CO CD CO  CO CO CO  LO CN Tt  CO  CO  CD CO CN  £ro  o o  Tt  C « CD CD  3 __  d co  !2 £  CD CL  ro  J2 ro  E  TJ  CO CD O  CM CO  E 5 co c c CO  CD C  3  3  3 CO  5» CO CO CO  to  to  to CO  •B  CO 3  CD  E E Io e E co • C <e .o CD <0  3  55  C CO o  O .C .O)  E  c  to  JD  o  EH CD  3 OQ  CO  O  to  e §  1 c o  I  o Q  to 3  C CO  o  c E  •s > s o CO  o c  E  CD  CO CO .3  £  CD  E 3  2  O .o  5  c co 3  o o 47  c CO  E *§  "to o  .<o co  c  CD 13 CO  c CO  o  to S O  C  s-3 to  co  JS  "CD  C| CD CO  I 3 c •E o I o O  E  JO  o o  to 3  C  •5  CD CO  *  2 «  °  o  5 1 ,2 co 4— —  3  -g '>  3 .O  u.  C CO CO >> CD T J  • CO  0 5  E co co c  E CM§  Jr.  B  o  O C  CT ®  L^TJ  °  2  S -2 CD y CO i_ Ji_ CD § 2 O CD ro CT _ * - co E CD CO 7 3 2 _ i_ 3 O CO <fi CD  "2  00  C  £__ 0-*— S g o > 'ro ro CD CO co .to  1  3  OlS  CD Q . E o  3  CL O  O o_ 0. * m a u at  3.2.4 Criteria for Comparing Priority Sets of Sites Priority sets of sites resulting from variation in values of the algorithm parameters were evaluated with the following criteria: 1. Total Area: total area of priority sites; 2. Number of Selection units: number of selection units in priority sites; 3. Number of Patches: number of isolated, contiguous areas (patches) of priority sites; 4. Patch Area: mean and median area of priority site patches, which was calculated using Patch Analyst 2.1 for ArcView 3.1; 5. Mean Perimeter to Area Ratio: perimeter and area of each patch was calculated using Patch Analyst 2.1 for ArcView 3.1; 6. Conservation Targets Achieved: percentage of conservation targets for species and vegetation classes that are achieved in priority sites. Similarity of priority sets of sites was measured using Jaccard's Similarity Coefficient, which provides a coefficient for binary (presence/absence) data. Values of the similarity coefficient range from 0 (no similarity) to 1 (complete similarity). The coefficient is calculated as A/(A+B+C) where A represents the sites that are present in two priority sets of sites, and B and C represent the sites that are present in one of the priority set of sites but are absent in the other (Krebs 1999). Coefficients that identify similarity between priority sites identified using different selection units were calculated using total area of overlap to represent the presence data (A) and total area that does not overlap as absence data (B+C). 3.2.5 Surrogacy of Priority Sets of Sites The surrogacy of priority sets of sites was examined using two methods. The first evaluation of surrogacy determined the proportion of conservation targets achieved for Red and Blue Listed species in priority sites identified for vegetation classes and priority sites identified for Red Listed species separately. The second evaluation determined the proportion of actual occurrences of vertebrate, invertebrate and plant species in priority sets of sites (Ferrier and Watson 1997). 3.2.5.1 Surrogacy of Vegetation Classes for Vertebrate Species Priority sites that were identified for Red Listed vertebrate species, vegetation classes and combined Red and Blue Listed vertebrate species with vegetation classes were examined for their degree of surrogacy for the twenty-nine Red and Blue Listed vertebrate species.  48  Surrogacy was measured as the percentage of conservation targets for Red and Blue Listed species that were achieved in each set of priority sites. 3.2.5.2 Surrogacy of Priority Sites for Actual Occurrences of Species Species occurrence data for Red and Blue Listed vertebrates, rare invertebrates (Appendix X) and plants (Appendix IX) were used to evaluate the surrogacy of priority sites for actual locations of species. The mean proportion of species occurrences that were located within sets of priority sites was calculated. These values allowed for a comparison of the number of species occurrences within each priority set between vertebrates, invertebrates and plants. 3.3 R E S U L T S 3.3.1 Variation in Selection Unit Size and Shape Conservation targets that represented habitat for the 0.5-population of each of the twenty-nine vertebrate species resulted in the selection of 16.6% of the total area of the region in the priority sites using T E M polygons as selection units (Figure 3.2). The most efficient selection unit, measured as function of the total area of priority sites, was the 0.155 k m  2  hexagon, with a total area of 239.01 km (Table 3.3a). Efficiency of the priority sets of sites, excluding sites identified by T E M polygons, decreased with an increase in selection unit size, but was not significantly correlated (r=0.925, P>0.05) (Table 3.3a). The 0.5-population conservation targets for vertebrate species were achieved in all priority sets of sites that resulted from variation in the selection unit. The most contiguous priority set of sites resulted from the 10 k m hexagon selection 2  unit, where there were eleven patches with a mean area of 32.73 km and the lowest mean 2  perimeter to area ratio of 0.97 km to 1 km (Table 3.3a). Contiguity, measured as a function of the number of patches, mean and median patch size, and mean perimeter to area ratio, decreased with a decrease in size of the hexagonal selection units (Table 3.3a). However, only median patch size was significantly correlated with selection unit size (r=1.000, PO.OOl). Priority sites identified using T E M polygon selection units had the lowest contiguity with a mean patch area of 0.92 km for 285 patches and a mean perimeter to area ratio of 9.06 (km:km ). Median patch sizes were always smaller than mean patch sizes (Table 3.3a).  49  50  CD  CD  V.  i - •=  >  OlU  CDro-K  o o o  o ci o  o d o  o o o  CD O  CD ro 0- t -  a. h- £  CD S 5  N OJ  a> ro  CD  d  0)  ci  o  o  ci  o  o  o  o  o  CD O  ci ci  T -  •I ct: E  CN  r-:  CL  OJ  CD  o CD  5  r^-  T - co •<»•  CD  ro cOrO  ?CD ro E S «  ci  w  _ CD C0c<P =5 j= E  ci  D  U ^  o o  o o  o o  ci  ci ci  T-  CN CM  N  CO  S <°<C hCD -c fc  ro ^  CM CO  00 •<t  ci  CO  ci  CD N  CN OJ CD  ci  CL  Q>  O  CD  OJ N  £  c  iS o co > 1<  CD  CO ^  T—  T—  o &o £w£| CD  ro  2 _ .t: _ E co  oc o 2 =*•  CL  1—  — r  O  O  O  O  CO CO CN  CO CO CO  o  LO CN  U J  c o '•e o  N  CO OJ  o  Q.  CM CO  <  6  E  o  O o  ci CO CO  CN LO  T—  o o •<r  o CO  CN lO  T—  ci CO CN  CO T - T T - C M CM  O  O  CO n  CO CO CL  O  o o  ci  ci  CO  cco  CM  y—  o LO  csi cvi CM CN  CO  CD  T—  T - CO n ^  o  iCO  O O ci  T—  00 CO  CO  Tt-  ci  ci  d  ci  _  CO 5 CD JO CO CM  CO CO CM  CO CO CN  CO CO CN  t-  <  OJ 00  CO CO  o  O O ci  o o  o o  CD CO  o T—  CD  d  «2  z »)  .t;  O) t  co o CO CM  co  O O ci  w  H  CL  ro  2 _  H  o  T—  S  .t; _ E o iS c o  OO O J  CD  OJ OJ  O  CD  o  H  O  ct  CO CN  o o  CL  .2ro"g  CO CO CO CO CO CO CM CN CM  co ro co S _  CO  o  O O  ro  3 Z  <  CO  ci  CD  T -  CO i-  roCCOO  o £ °  T—  ci  CM  b OT— ro ^ CCMO  CD CM  CM  CN lO  o  CO  t!  — OJ  h~  CO CO  OJ CD CO  fOJ  o  CL  ro  8 "5  T—  ^—  CD ^  CD  3  2  .2 ro •c 9>  fc ro  CD  LO LO S CO CM CN  T—  ci  o.  0 0 CD o CO o CO CN LO CN CN CO  CO  o  —  ro COCN CD O  CL  E o u  CO  o o  „  CD  CD CO  ci  o  «= - _ o o  ro  0. (1)  o  CD  oo ci  2 15 -  CL  ro 9i E  CD  <= -  CO CO L O 1-  o  ci  o  c <^  ^  o  ci  05  O  00  ^  <»••=_ c row • i cc: E  O l O O O )  c  CD  ro „  E  ii  0-1-  » (D 00 o  N  oo CD > Ol.l) CN  CDro  & O CD -i= _ C COCM  _  la: E CL  o  O O ci  T—  CO CN  OJ 00  ci  d  co OJ CO CO  c  ro O  o c  o  CD CD X OJ  hE I  CN  o  E j£  <  o  DJ  c  ro  V  CD  ™  X  oi  J  °  '^1  ™r-  LO  _  CO  E CN  LO  I  LU CJ — I T-  O  I  QJ  .a a 'C O cu T3  <  > O  C O  CO  CQ  0  <=a  CO  Q)  0  -*—' CO -4—'  OJ ° 3 w  CO CO  OJ 5]  O OT C  o  T3  CO 0  CD  m  °S  0 oo  ct: a: > cc  51  > OJ  W co "O 0 CD o  c  I  X.  CL CL  CO O  c  CO  CO  ^—  1  c  o  p  c  Lw U  CO  £  OJ CO  c  o O  >  CO  co  o 0. 1 o  CO  o 1  CL LO  o  0 o  c~  sei  &o  CD > DJ.CD  pul ion pul ion  c O  0 CL CO  0 o 0 CL C O  Priority sets of sites that were most similar in site location resulted from the 10 k m  2  hexagon and 2 km hexagon selection units, with a coefficient of similarity of 0.344 (Table 3.4). Priority sets of sites that were least similar resulted from the 10 k m hexagon and T E M 2  polygon selection units, with a similarity coefficient of 0.190. The coefficients of similarity decreased as the difference in selection unit area increased (Table 3.4). A l l of the similarity coefficients were significantly higher than similarity values between randomly selected sets of sites and each of the minimum sets. However, all of the priority sets of sites shared less than 35% of their total area with priority sites identified for different selection units.  Table 3.4. Jaccard's similarity coefficients for priority sets of sites resulting from different selection units. The probability of obtaining each similarity coefficient by random was calculated from 1000 random selections of the equivalent total number of sites, or area for polygon selection units, and denoted as: *** for P < 0.001. All significant coefficients are higher than similarity values between randomly selected sets of sites and each of the minimum sets. Selection Unit 10 km Hexagon 2 km Hexagon 0.16 km Hexagon TEM Polygon 2  2  10 km Hexagon 2  1 000„.  2 km Hexagon  0.16 km Hexagon  TEM Polygon  0.344*** 1.000  0.252*** 0.339***  0.190*** 0.217***  1.000  0.283*** 1.000  2  -  2  2  -  3.3.2 Variation in Indicators of Biodiversity The total area of priority sites that represent habitat for 0.5-populations of the twenty2  2  nine Red and Blue Listed species using the 10 km hexagon selection unit was 360.0 km (Table 3.3b). The area required to represent eleven Red Listed species was 310.0 km , which was only 50.0 km less than priority sites identified for Red and Blue Listed species. The total area of priority sites identified for range conditions and successional stages of the vegetation classes was 480.0 km (Table 3.3b). These priority sites were only one site fewer than priority sites identified for twenty-nine Red and Blue Listed species and fifteen vegetation classes, simultaneously (Figure 3.3). The summed area of priority sites identified separately for vertebrate species and vegetation classes was 730.0 k m \ Therefore, it is more efficient to identify priority sites for vertebrate species and vegetation classes simultaneously, rather than separately, using complementarity (Table 3.3b).  52  CO CD  'o  0  a SC  '  CO 03  rr" t= 35 ..  0  O  OT  CD—I  .£  0  W = W CQ  o ro CO O  ro* o  0  =5 B >.  C  ro 2 -  .t;  2  .2  0 co  T3 W  o ro 1Q O 0 JS 0  t£  ro . •  O ^  ®  o  CL. CO  0  cp ro c -Q 0 0  c >  .2 T3 *i  0  £3 .  0 —' >>  . o OT oj Q 0  0  -  ~~' CO  TT  OT  CD  o ,9-  ® W  & OT 0 oto CD -H  OT 2 >.-Q  CO 0 -ti  «  o i l • 0  ^ OT CO ~ i  0 Z3 D) U_  53  C  -2 CD  0 > T  c  CO  Priority sites identified for vegetation classes required more area than priority sites that were identified for vertebrate species in a greater number of patches (Table 3.3b). Median patch sizes were always less than mean patch sizes. Priority sets of sites that were most similar in site location resulted from the priority sites identified for Red and Blue Listed species and Red Listed species, with a coefficient of similarity of 0.523 (Table 3.5). The lowest coefficients of similarity were between priority sets of sites identified for dissimilar indicators of biodiversity (i.e. between vegetation classes and vertebrate species). Coefficients of similarity were highest for priority sets of sites identified for similar biodiversity indicators. A l l of the similarity coefficients were significantly higher than similarity values between randomly selected sets of sites and the minimum sets. However, all priority sets of sites, except one, shared less than 50% of their total area with priority sites identified for different biodiversity indicators (Table 3.5).  Table 3.5. Jaccard's similarity coefficients for priority sets of sites resulting from different indicators of biodiversity. The probability of obtaining each similarity coefficient by random was calculated from 1000 random selections of the equivalent total number of sites and denoted as: * for P < 0.05; *** for P < 0.001. All significant coefficients are higher than similarity values between randomly selected sets of sites and each of the minimum sets.  Biodiversity Indicator Red & Blue Listed Species Red Listed Species Vegetation Classes R&B Spp. & Veg. Classes  Red & Blue Listed Species  Red Listed Species  Vegetation Classes  0.523*** 1.000  0.151* 0.197***  1.000  -  -  R&B Species & Vegetation C l a s s e s  1.000  -  -  0.328*** 0.231*** 0.406*** 1.000  3.3.3 Variation in Conservation Targets for Vertebrate Species Area-based conservation targets used in the analyses that represent habitat for three different population sizes of the twenty-nine Red and Blue listed vertebrate species were (1) 1.0-populations (current population estimates), (2) 0.5-populations (50% of current population estimates) and (3) general minimum viable population (MVP) estimates. When the conservation target was doubled from representing habitat for 0.5-populations of the vertebrate species to representing habitat for 1.0-populations, the number of priority sites selected more than doubled with an increase from 36 hexagons to 89 hexagons, respectively (Figure 3.4, Table 3.3c). Priority sites that represent M V P estimates of the Red and Blue Listed species required 231 out of 233 hexagons in the region. Only 82.8% of the M V P 54  conservation targets for Red and Blue Listed species were achieved in these priority sites. There was not enough habitat area to achieve the M V P conservation targets for five of the Red and Blue Listed species within the region. These species were the American Bittern, Ferruginous Hawk, Sandhill Crane, White-headed Woodpecker and Western Screech Owl. However, the American Bittern would likely have sufficient habitat if the large lake polygons were included in priority site selections. When site-based targets, based on species presence in a selection unit, were used to represent habitat for 0.5-populations of the Red and Blue Listed species, only ten hexagons were selected in the priority sites (Figure 3.4). This result is based on the assumption that the entire area of the 10 k m selection unit provides habitat for a species present in the selection 2  unit. The ten sites were selected by the last rule in the Minset algorithm, which selects units that have equal conservation value in the order that they occur in the database. This is because there were two or more selection units with equal irreplaceability and summed irreplaceability values at each iteration of the algorithm. The ten priority sites selected using site-based targets represent 72.4% of the 0.5-population area-based conservation targets for Red and Blue Listed species. Species that did not achieve their 0.5-population area-based conservation targets in the ten sites were the American Bittern, Bobolink, Ferruginous Hawk, Long-billed Curlew, Sandhill Crane, Short-eared Owl, White-headed Woodpecker and Townsend's Big-eared Bat. Priority sites that represented habitat for M V P estimates of Red and Blue Listed species required 99.1% of the total area in the region and formed one patch (Table 3.3c). The total area of priority sets of sites, and mean and median patch size decreased with a decrease in the conservation target values. The median patch size was smaller than the mean patch size, except in the priority sites identified for the M V P conservation target (Table 3.3c).  55  B  2 <D JD -C  ja -a -c 0 o UZ CD  I°  ~ 0 giS  12  0  CD  s * .2  •K CO  O  TJ  LO  O  —' CD Z3  S« CD £  ^jP. to Q)  .3 °- c CL CO O O 0) '43  * si 3  O  • CL  0  a: 9- tu  to CO  i_ o o "•^ ~—' 0  JS jo §  0 to 0° ro"toE ' o c O 0 0 a.  5 I .2 o  Q  Ira  •43  CL CL O O  0  o E Q. 0 0  -l_  * S 0 "D E  11"  O CD  E 0  o to LO  IT  CD -Q  « 3 "O o CL 0 o  C LO  vP  L-,  Etc  o A. co ^~3 .CO 0 o  ro 9-  D_  6 1^  to  >,|° •° ® xT  CD ro £  to  T3 C 0 O T3  o  0  w  0 c  ro ro to  5 o £ c a- o  CO c  3 .2  g-I O.  C  ro  s  co  ^"1  =5  CL  0 to  O 0 CL i— i  I  -* O CL  o E o  O £.  fe c~  o .£ to  E a>  0^2  ^  to  v * r CO O O 0 W Dl c  CD  Si II co  0 .E o  " I ro  ® I  & <2 3 ™ 58 Q.2 °a c u  CL to ^ 0 C  E.i C  ^  CD C  co  3 0 b  co n • ^ co o  Q.  CO  CL  LL.  56  The priority sets of sites identified for targets representing habitat for 1.0-populations and 0.5-populations of vertebrate species were most similar in location, with a similarity coefficient of 0.402 (Table 3.6). Coefficients of similarity were highest for priority sets of sites that had similar conservation targets. The lowest coefficients of similarity were between priority sites identified for site-based targets and all three priority sets of sites identified for area-based targets. The coefficient was highest between priority sites identified for site-based targets and priority sites identified for 1.0-population targets with a value of 0.076 (Table 3.6). There were only two similarity coefficients that were significantly higher than similarity values between randomly selected sets of sites and the minimum sets. The remaining four similarity coefficients were not significantly different from comparisons of random selections of sites.  Table 3.6. Jaccard's similarity coefficients of priority sets of sites resulting from different conservation targets. The probability of obtaining each similarity coefficient by random was calculated from 1000 random selections of the equivalent total number of sites and denoted as: NS for P > 0.05; * for P < 0.05; *** for P < 0.001. All significant coefficients are higher than similarity values between randomly selected sets of sites and each of the minimum sets. Conservation Targets MVP Estimate 1.0-Population 0.5-Population Species Presence  Minimum Viable Pop. Estimate  1.0-Population  0.5-Population  1.000 -  0.385/VS 1.000 -  0.156A/S 0.402*** 1.000 -  Species Presence 0.043/VS 0.076* 0.045/VS 1.000  3.3.4 Selection Units Occurring in All Priority Sets of Sites The amount of spatial overlap of priority sets of sites that result from variation in each of the algorithm parameters was determined by calculating total area of overlap (Figure 3.5). The greatest amount of spatial overlap was 70.00 km , which was measured in priority sets of z  sites that included different biodiversity indicators in the analyses (Table 3.7). The smallest spatial overlap, with a total area of 20 km , was calculated for the four priority sets of sites that had different conservation target values. Since the site-based conservation targets were identified using presence data, whereas all other priority sets of sites were identified for habitat area requirements of species, spatial overlap of the three priority sets of sites for areabased targets was also calculated, excluding priority sites identified for site-based targets. The total area of spatial overlap of priority sets of sites that had different area-based 57  conservation targets was 35.00 km . There was no spatial overlap when priority sets of sites 2  resulting from variation in all three parameters were examined collectively. However, when priority sites identified for site-based targets were excluded from the calculation of spatial overlap, there was an area of 9.90 km that was common to the remaining eleven priority sets of sites (Table 3.7).  Table 3.7. Total area and percentage of spatial overlap of priority sets of sites resulting from variation in algorithm parameters. Total Area (km ) of Percent Overlap of Overlap of Priority Sites Priority Sites 2  Parameter Examined  Parameter Variation  Selection unit Biodiversity Indicator Conservation Target Conservation Target All Parameters All Parameters  All Variations All Variations All Variations Site-based target excluded All Variations Site-based target excluded  44.22 70.00 20.00 35.00 0.00 9.90  2.82 3.00 0.86 1.50 0.00 0.63  3.3.5 Surrogacy of Priority Sets of Sites 3.3.5.1 Surrogacy of Priority Sites for Vertebrate Species Priority sites identified solely for vegetation classes represented 93.1 % of the areabased conservation targets for 0.5-populations of Red and Blue Listed species (Table 3.8). Conservation targets for the White-headed Woodpecker and Sandhill Crane were not achieved in priority sites that represented vegetation classes. Priority sites identified for Red Listed species also represented 93.1% of the conservation targets for Red and Blue Listed species, but in approximately two-thirds (64.6%) of the area identified in priority sites for vegetation classes. Area-based targets for American Bittern and Sandhill Crane, which are both Blue Listed species, were not achieved in priority sites that represented Red Listed species.  58  0  '% a >- 0 £ E?  C CO CO" _Q "ti  co j D G> ~ co O  I°" C-2  .2 -D -5.2  £ o 0 0  -  C  _0 CO 0  o-  E s;  w  —  0  CO CO  ^"55  CD =3 0 Q - o CCOD =  <  «I  X 0  2.  0  CO CD  5. Q. *E i _c •c S o _  0  E? ro  0 co CD CD  _CpTj  9  CD  CD * * CD  = &  0 0  0  .2 ro ro ^ > 0 l_ w  E  CJ ^ II  0  >l  c  O  M=  O JD CD  O JD  C D  is  ^ o o  0 co  — 1 CO  J C  CO o™ E?~ 5 iS B C c o .2 £ E RO  0  "CD O  ro  « 0  c o  CD  *2 ^ 0 JD CO  r «  CO  o  1  ro o  0  T3 C  !2  §  -j=  "> 0 CD 0  >> -w  -ti CO  'co I—  0 > TD  >  •si  to ro "5  g  It CO  0 >, o -ti 0 CO  O- co  O o Q. CO CD 2 CO  T C  o >^ o w ro 0 0  C  o o 0 0  TJ  > •-  'c v  *  <  T5  o  in ~  W  0 i— 3  CD  L L  59  W  Table 3.8. Proportion of area-based conservation targets for Red and Blue Listed vertebrate species achieved in priority sets of sites resulting for different biodiversity indicators.  Priority Sets of Sites Red & Blue Spp. & Veg. Classes Vegetation Classes Red & Blue Listed Species Red Listed Species  Total Area (km )  Proportion of Targets Achieved for Red & Blue Listed Species  490.00 480.00 360.00 310.00  1.00 0.93 1.00 0.93  2  3.3.5.2 Surrogacy of Priority Sites for Actual Occurrences of Species In each of the priority sets of sites, Red and Blue Listed vertebrate occurrences were represented at higher proportions compared to invertebrate and plant occurrences (Table 3.9). This result is not unexpected, since all priority sets of sites, except for the set of sites identified solely for vegetation classes, were identified based on W H R models of habitat suitability for threatened vertebrate species. Priority sites that included Red and Blue Listed species as biodiversity indicators had the highest proportions of vertebrate occurrences and priority sites that included vegetation classes as biodiversity indicators had the highest proportions of plant occurrences (Table 3.9). Proportions of vertebrate and invertebrate occurrences in priority sets of sites were not correlated with decreasing total area (r=0.540 and 0.534 respectively, P>0.05). Proportions of plant occurrences in priority sets of sites decreased with a decrease in total area, however, the correlation was not significant (r=0.779, P>0.05). Therefore, the biodiversity indicator used in reserve selection may have a direct effect on proportions of actual species occurrences in priority sites. Table 3.9. Mean proportion of occurrence records and standard deviation of Red and Blue Listed vertebrate, invertebrate and plant species located in priority sets of sites resulting for different biodiversity indicators.  Total Area (km ) Priority Sets of Sites 490 Red & Blue Spp. & Veg. Classes Vegetation Classes 480 Red & Blue Listed Species 360 310 Red Listed Species 2  60  Vertebrate 0.55 ± 0.23 0.38 ± 0.23 0.52 ± 0.26 0.22 ± 0.20  Invertebrate 0.16 ±0.32 0.08 ± 0.22 0.14 ±0.31 0.01 ± 0.07  Plant 0.24 ± 0.36 0.21 ±0.30 0.20 ± 0.33 0.07 ± 0.20  3.4 DISCUSSION 3.4.1 The Problem of Scale Systematic reserve selection procedures provide a relatively quick and reasonable method for assessing conservation priorities within a region. However, the scale of both the spatial data and biological data influence the spatial distribution and number of priority sites selected by a systematic reserve selection algorithm (Table 3.3, Figures 3.2, 3.3, and 3.4). Although the intended conservation targets were achieved in each of the priority sets of sites, except in the sites identified for minimum viable populations of vertebrate species, differences in the spatial distribution and number of priority sites are important considerations for reserve selection. As a consequence of the low level of spatial congruence between priority sets of sites resulting from variations in parameter values relating to scale, priority sites identified in published studies for regional conservation should not be implemented unless the effects of scale have been assessed. In this chapter, the effect of scale on priority sets of sites identified by a systematic reserve selection algorithm at a fine scale of mapping was investigated. Systematic reserve selection procedures have usually been applied at coarse scales, with published examples of selection units that range in size from 9 km to greater than 11,000 k m (Nicholls and 2  2  Margules 1993, Williams et al. 2000). In these studies, selection unit size was selected arbitrarily (Nicholls and Margules 1993, Lombard et al. 1995, Wessels et al. 1999, Freemark et al. 2000, Williams et al. 2000), politically (Pressey et al. 1997, Ando et al. 1998), biologically (Price et al. 1995), for technical reasons (Strittholt and Boerner 1995) or as a compromise between management and ecology (Kiester et al. 1996, Lombard et al. 1997). However, none of the studies examined the consequences of their decision for selection unit size on priority site selections. Selection unit sizes used in this study represent the range of decisions that have been made in previous studies. The 10 km hexagon selection unit was 2  chosen for both ecological and management reasons, the 2 km hexagon selection unit was 2  chosen for political reasons, the 0.155 km hexagon selection unit for arbitrary reasons and 2  the polygon selection unit for biological reasons. Each of these selection units resulted in different priority sets of sites, with a maximum similarity in location of priority sites of 34.4% (Table 3.4). Therefore, recommendations from studies that identify priority conservation sites should consider the reasons for using a particular selection unit size.  61  Systematic reserve selection procedures that have been applied at coarse spatial scales recognise that boundaries of priority sites identified using large selection units are to be refined on the ground before reservation (Freemark et al. 2000, Brooks et al. 2001). At fine spatial scales, such as the South Okanagan region, polygons that map precise boundaries of vegetation can be identified. The area of these selection units is variable, but can be quite small. Small units are usually too small to refine boundaries on the ground, since refinement of boundaries implies reduction of area. The results in this chapter demonstrate that selection unit shape also has an effect on priority site selection (Table 3.3a, Figure 3.2), even though it has been suggested otherwise (Pressey and Logan 1998). Precise boundaries, although more ecologically correct, may not provide the best unit for reservation because of the effect of size and shape on priority site selection. Therefore, systematic reserve selection procedures may be better suited for coarse scale analyses at provincial or national scales, to highlight areas that require more detailed research on fine scale conservation priorities (Nicholls and Margules 1993). Selection units identified at coarse spatial scales are usually large enough to maintain viable populations of species even after refinement of boundaries. Schonewald-Cox (1983) predicted that existing reserves of 10 to 100 square kilometres in size could maintain viable populations of small mammals and that reserves of 10,000 to 100,000 square kilometres could maintain viable populations of larger mammals. Stoms (1994) determined that the size of selection unit should vary with the ecoregion and recommended that the selection unit for the Intermountain Sagebrush ecoregion in Idaho should be 100 km to 600 km in size. Although lower elevations of the South Okanagan region are predominately sagebrush ecosystems and the largest mammal included in the analyses was California Bighorn Sheep, decisions on selection unit size cannot rely solely on these broad recommendations without considering the scale of the regionally mapped data (Fotheringham 1989). The total area of the South Okanagan region is 1770 km , which is smaller than the recommendation for maintaining viable populations of large mammals, and the region would consist of only 3 to 18 homogeneous selection units if the recommendations for sagebrush ecosystems were followed. Therefore, it is essential that habitat complexity of the region and scale of the data available for reserve selection are examined before size and shape of the selection unit is decided.  62  Since a priority site that is selected by the algorithm could be isolated from other priority sites, selection unit size must be large enough to provide adequate habitat area to maintain life requisites of species (Pressey and Logan 1995). Large selection units (i.e. hexagons) result in fewer and larger patches of priority sites, with lower mean perimeter to area ratios, than small selection units of similar shape, which increases the contiguity of the reserve system (Table 3.3a). Another advantage of using large selection units is that it is more difficult politically to increase the size of a reserve after it has been identified and implemented than it is to decrease size. However, large selection units tend to over-represent conservation targets for flora and fauna in priority sites, which leads to higher cost for implementation of the reserve system (Pressey and Logan 1998). If selection units and final reserves are to be manageable and valuable for species in the long term, then overrepresentation of conservation targets for regional flora and fauna is inevitable (Pressey and Logan 1998). Although polygons in the terrestrial ecosystem mapping provide the most accurate map of potential habitat in the region, most of them are too small to maintain species if the polygon was isolated from surrounding habitat. Priority sites identified using these polygons were a collection of widely scattered, tiny cells of a minimum combined area that provided one of the most efficient sets of sites for conservation but were too small to persist (Table 3.3a, Figure 3.2, Pimm and Lawton et al. 1998, Pressey and Logan 1998). Smaller selection units can be used when procedures are available in reserve selection algorithms to amalgamate priority sites that are too small to maintain life requisites of species. Amalgamating procedures that preferentially select units that are in close proximity to previously selected sites result in greater total areas of priority sites than algorithms that exclude amalgamating procedures (Bedward et al. 1992, Nicholls and Margules 1993, Lombard et al. 1995). However, if the goal of conservation is to maintain biodiversity in perpetuity, then the efficiency of the reserve network should be a secondary consideration in systematic reserve selection. Therefore, the 10 km hexagon, which is the largest selection unit used in this chapter, provides the best unit for conserving biodiversity within the South Okanagan region. The minimum patch size (10 km ) in priority sites identified with this selection unit is large enough to provide contiguous habitat for life requisites of all of the species included in these analyses. When a procedure is available in the reserve selection algorithm that amalgamates priority sites during the selection process, smaller selection units, 63  which contain more precise data, may be more appropriate for identifying priorities in the South Okanagan region.  3.4.2 The Question of Surrogacy In British Columbia, provincially designated Red and Blue Listed species, although not recognised as legally endangered, provide a reasonable starting point for identifying conservation priorities (Dobson et al. 1997). Selecting complementary areas for threatened species first and then selecting additional sites to represent more common species is almost as efficient as selecting areas for all species at once, disregarding their conservation status (Brooks et al. 2001). However, this result depends on the 'nestedness' of species or biodiversity indicators within a region (Pressey et al. 1999). If the species are found in the same areas, their distributions are nested, and the algorithm can identify complementary sites more efficiently than for species with dissimilar distributions. When the biodiversity indicator was restricted to Red Listed vertebrate species, thereby increasing the proportion of unique or infrequent species in the data set, the total area of priority sets of sites increased relative to the number of species included in the algorithm (Table 3.3b). The decrease in efficiency of priority site selections, as a result of an increase in the proportion of rare features, has been found in other studies (Lombard et al. 1995, Pressey et al. 1999). As a consequence of the inefficiency, seventeen out of the nineteen Blue Listed species were represented to their conservation targets in priority sites identified solely for Red Listed species (Table 3.8). This result concurs with findings of Brooks et al. (2001) that priority sites identified for threatened species represent a greater diversity of species. When data are not available for vertebrate species requirements, or for even a subset of the biodiversity within a region, such as the threatened species, vegetation classes can be used as a first approximation for identifying priority conservation sites using a systematic reserve selection technique. In the South Okanagan region, 93.1% of the conservation targets for vertebrate species were achieved in priority sites identified for vegetation classes (Table 3.8). Therefore, vegetation classes provide a reasonable surrogate for threatened vertebrates in this region. The efficiency of priority sites, which is a function of total area, was lower for priority sites identified for vegetation classes than those identified for vertebrate species (Table 3.3b). However, initial site selections using vegetation can be revised as data become available on locations and habitat requirements of vertebrate species.  64  Proportions of vertebrate, invertebrate and plant species occurrences in priority sets of sites provide a measure to determine the surrogacy of priority sites based on predictive models for actual locations of species. Unfortunately, the available occurrence data were not collected systematically for the entire region and, therefore, can provide only a rough estimation of the surrogacy of priority sites for species occurrences (Ferrier and Watson 1997). The differences in the proportions of species occurrences in priority sets of sites illustrate that priority sites identified using WHR models provide better surrogates for vertebrate species than for invertebrate and plant species (Table 3.9). This is not surprising since the W H R models were based on habitat requirements of threatened vertebrate species. However, low proportions of invertebrate and plant species occurrences in priority sets of sites indicate that these species would not be adequately conserved. Individual selection units with actual locations of rare invertebrates and plants could be manually selected as priority sites until their distributions within the region are determined. The efficiency of the C-Plan algorithm when selecting priority sites for diverse sets of features was demonstrated in this chapter by the selection of sites for vertebrate species and vegetation classes simultaneously. The total area of priority sites identified for vertebrate 2  2  species and vegetation classes simultaneously was 490 km , which was only 10 km more than the total area of priority sites identified for vegetation classes alone (Table 3.3b), and 240 km less than the combined area of priority sites selected for vegetation classes and vertebrate 2  species separately. Although locations of priority sites were different (Table 3.5), only one additional site was required to achieve the conservation targets for the combination of vertebrate species and vegetation classes compared to the number of sites identified for vegetation classes alone. Therefore, systematic reserve algorithms, based on a complementarity selection process, are useful for identifying efficient sets of priority sites for diverse biodiversity indicators within a region. However, these priority sites may not have the highest conservation value for vertebrate species or vegetation classes, since priority sets of sites that were identified for vertebrate species and vegetation classes separately are more likely to have higher conservation values for each biodiversity indicator. When diverse biodiversity indicators are considered in reserve selection, complementarity may actually select marginal areas of species ranges where species distributions overlap (Gaston et al. 2001). Therefore, efficiency of priority sites may need to be compromised when selecting  65  reserves for diverse biodiversity indicators by preferentially selecting sites that have high conservation value for each biodiversity indicator on their own.  3.4.3 The Problem of Identifying and Maintaining Viable Populations Many of the Red and Blue Listed vertebrate species are on the peripheries of their ranges, with the majority of their range in the United States (Banfield 1974, Green and Campbell 1984, Peterson 1990, Nagorsen and Brigham 1993, Gregory and Campbell 1999). This presents a problem in the identification of species that are important within a region, such as the South Okanagan (Erasmus et al. 1999, Freemark et al. 2000). None of the species included in this thesis are considered endangered in the United States, where there is formal legislation for protection of endangered species (WDFW 2000), although populations of some species, such as the Brewer's Sparrow, are declining throughout their range (Sauer et al. 1997). In Canada, and consequently the South Okanagan region, conservation decisions are forced to occur within area defined by a geopolitical boundary that follows the 49 parallel. th  Therefore, many of the rare and threatened species in Canada are peripheral populations of species at the northern extent of their ranges (Freemark et al. 2000). Peripheral populations may be important to the conservation of a species because of genetic variation provided by populations at the limit of species ranges, their role in ecosystem function, and the increase in public awareness that may prevent these species from becoming globally endangered (Hunter and Hutchinson 1994). The geopolitical boundary of the South Okanagan study area presents difficulties in identifying conservation targets that maintain viable populations of threatened species. The conservation of minimum viable populations of peripheral species implies that there are separate populations of these species north of the 49 parallel. Since many of the species th  included in this study interact with conspecifics south of the international border, the minimum viable population estimates identified in this study, of 500 to 1000 individuals, are not realistic for most of the species in the South Okanagan region. In addition, most of the M V P estimates for threatened species are greater than any known population sizes of these species in the region before European settlement (Appendix V). M V P estimates identified in this chapter were very general because the required information on demography and genetic variation to perform population viability analyses was not known for any of the threatened species (Shaffer 1981, Soule 1987). Minimum viable population estimates were chosen only to illustrate how conservation targets affect selection of priority sites. The most realistic 66  conservation goal for this region is to identify priority areas that maintain current or, if known, historical population estimates of these species. Priority sites that achieve the conservation targets for current estimates of population size (1.0-populations) requires approximately 37.2% of the land area in the Okanagan region, primarily in lower elevations of the valley and, unfortunately, most of this habitat is owned privately. However, empirical evidence is necessary to refine calculations of the amount of area required to conserve 1.0populations of threatened species before implementing a reserve network in the region based on the findings in this chapter. 3.4.4 The Use of Habitat Area Predictions and Presence Data in Reserve Selection Predicted probabilities of occurrence of species are used infrequently in systematic reserve selection (Margules and Nicholls 1987, Margules and Stein 1989) because of added uncertainty in priority site selections from erroneous predictions (Araujo and Williams 2000). Since the species data used in this thesis were based on predictions from wildlife habitat relationship models for Red and Blue Listed species, priority sites that result may need revision when detailed information on actual distributions of species are known (Flather et al. 1997). The need for revision is also recognised because of the low proportions of actual species occurrences located in each priority set of sites. The mean proportion of vertebrate occurrences in the priority sites that represent habitat for 0.5-populations of Red and Blue Listed vertebrate species was 0.52 (Table 3.9). Although there are biases in the occurrence data, this value illustrates that priority sites identified for 0.5-populations exclude areas where threatened vertebrate species are known to occur. Priority sites are only as reliable as the data that are used for the analyses and, therefore, it is important to improve the quality of databases on which reserve selection algorithms operate (Margules et al. 1994). However, until actual occurrences of species have been systematically surveyed for an entire region, predicted probabilities of occurrence provide a reasonable starting point for assessing conservation priorities within a region (Nicholls 1989). Site-based conservation targets were based on predictions from wildlife habitat relationship models, which identify the presence of species within selection units. The area of suitable habitat within a selection unit was not identified. Therefore, the assumption that the entire area within a selection unit provided suitable habitat for a species present within the unit was necessary to determine conservation targets, in terms of the number of sites required to conserve each species. However, this assumption is inaccurate since much of the habitat 67  area within a selection unit may not be suitable for a particular species (Dobson et al. 1997). For this reason, site-based targets were included in the analyses to determine the effects of information loss on priority site selections. The similarity of priority sites selected using sitebased targets compared with area-based targets was low, since there were far fewer sites selected for the site-based targets (Table 3.6). Similar irreplaceability values were calculated for many selection units at each iteration of the algorithm using site-based conservation targets because of nested species distributions. Consequently, priority site selections were influenced by the order in which selection units were listed in the database, which was the last tie-breaking rule of the algorithm. This selection problem may be entirely related to scale, and not occur at coarse scales. However, most published studies have not explicitly identified which rules were applied as tie-breaking rules during priority site selection. The differences in selection of priority sites using the two different types of targets has implications for studies that are done at coarse spatial scales. The result identifies that the calculation of conservation value for a particular selection unit is different using a site-based target from a target that identifies the relative value of each selection unit for species (Polasky et al. 2000). Area-targets allow for a 'ranking' in conservation value of selection units based on the amount of suitable habitat located within each unit. By limiting data to categorical accounts of species presence or absence within selection units, the relative conservation value of each site for a species is lost. The consequence of using presence data in reserve selection algorithms is that sites are chosen based on the geographical concentration of species or 'species richness', but not on habitat quality. Therefore, priority conservation sites identified using systematic reserve techniques that are based on presence data do not identify the total area that must be protected to conserve species (Dobson et al. 1997).  3.5 CONCLUSIONS AND RECOMMENDATIONS Systematic reserve selection procedures have developed in response to the urgency of decisions required for conservation and lack of efficient methods of identifying reserves given that resources available for conservation are limited (Pressey 1994). Although systematic reserve selection provides a technique that is both explicit and repeatable, priority sites identified for conservation must be considered cautiously because of low congruence of priority sets of sites resulting from variation in values of algorithm parameters.  68  Selection unit size and shape had a large influence on priority sites identified by the algorithm. Although smaller selection units allow for the identification of landscape elements more accurately than do larger units, selection units should be large enough to provide habitat to conserve viable populations of species in regions where the spatial scale is appropriate. Although large units may contain areas of unsuitable habitat, they provide a buffer against uncertainty in priority site selection. However, i f an amalgamation procedure is included in reserve selection, smaller selection units may identify priority conservation sites more accurately. Conservation plans should include data for as much of the biodiversity of a region as possible. However, if occurrence or distribution data for all species are not available, reserves identified for a subset of the biodiversity, that is not restricted to an individual taxonomic group, may provide a reasonable surrogate for identifying priority sites for all biodiversity in the region. At fine spatial scales, predictions of habitat suitability for species provide useful data for identifying the relative conservation value of selection units. Known locations of species can be incorporated into the reserve selection procedure as data become available. These data help identify realistic conservation targets for each species, which should consider both the biology of species and boundaries of the region. The values identified for each of the algorithm parameters in this chapter represent some of the decisions that will need to be made by scientists and managers when identifying priority areas for conservation. These values were used to evaluate the variation in priority sets of sites selected by the C-Plan algorithm and should not be used as final decisions for identifying priority conservation areas without careful consideration of their implications. Unfortunately, the implications cannot be assessed a priori since the scale of the region and the biodiversity indicators included in systematic reserve selection will affect the results. Although efficiency and the associated priority sets of sites of minimum area are a practical necessity for management, the ecology of species cannot be overlooked in the process. The cautious recommendations from this chapter are that the 9.9 km area that occurs in all of the priority sets of sites (excluding priority sites identified for the site-based conservation targets) should be considered first for conservation in the South Okanagan (Figure 3.5). After selection unit boundaries of this 9.9 km area are refined on the ground 2  and selected in C-Plan, overlapping priority sites that result from further variation in algorithm parameters could be identified in the remaining unreserved area. 69  The utility of systematic reserve selection algorithms will increase when the consequences of choosing particular parameter values on priority site selections within a region are understood. The effect of rule sequence or type of rules used in a heuristic minimum set algorithm has not been investigated in this chapter. However, there is evidence that variation in algorithm rules produces variation in priority sites selected (Pressey et al. 1999). Although systematic reserve selection algorithms are valuable for quickly exploring the options available for conservation within a region, the findings in this chapter demonstrate that there are no simple solutions for conserving biodiversity.  70  4.0 C H A P T E R 4: S C A L E E F F E C T S O N P E R F O R M A N C E I N D I C E S U S E D T O EVALUATE RESERVE NETWORKS 4.1 I N T R O D U C T I O N It is widely acknowledged that global biodiversity is declining and that aquatic and terrestrial areas must be protected to minimise the loss of species (Pressey and Tully 1994, Church et al. 1996, Arcese and Sinclair 1997, Margules and Pressey 2000). However, most existing reserve networks have been selected in an ad hoc manner, with goals other than simply protecting biodiversity, and have been constrained by geopolitical factors and financial limitations (Soule and Simberloff 1986, Pressey and Tully 1994, Lombard et al. 1995, Freitag et al. 1998, Nantel et al. 1998, Rodrigues et al. 1999). Since options for conservation are decreasing as alternative land uses increase, the protection of biodiversity in these reserve networks is critical (Pressey et al. 1993, Williams et al. 1996, Margules and Pressey 2000). Therefore, existing reserve networks must be evaluated to determine the degree of protection of regional biodiversity and to identify new areas that complement reserve networks with habitat for species that are not adequately protected. 4.1.1 Measuring Performance of Reserve Networks Recognising that the resources available for conservation are limited, systematic reserve selection procedures were developed that minimise overall cost of reserve networks by identifying the minimum amount of area required to achieve explicit regional conservation goals (Pressey et al. 1993). Systematic procedures identify efficient reserve networks called 'minimum sets', that represent the most biodiversity in the least amount of area (or cost). Efficiency is an index that has commonly been used for evaluating reserve network performance. The index is calculated as 1 - (X/T), where X is the total amount of area required to achieve the conservation goal for the region and T is the total area in the region (Pressey and Nicholls 1989a). The efficiency of a reserve network that achieves the conservation goal for a region is high when the area of the reserve network is small relative to the total area in the region. Conversely, efficiency is low if the reserve network contains a large amount of area relative to the total area in the region. Many evaluations have identified that existing reserve networks are composed of greater area (i.e. efficiency is low) than minimum sets (Rebelo and Siegfried 1992, Pressey et al. 1993, Williams et al. 1996, Freitag et al. 1998, Nantel et al. 1998, Margules and Pressey 2000). However, reserve networks with  71  low efficiency can still be 'effective' if regional biodiversity is adequately represented in the network. The index of 'effectiveness' measures the proportion of species represented in the reserve network, and is calculated as N/K where N is number of species that are represented in the reserve network and K is the total number of species in the region. Although efficiency is necessary because of limited resources available for conservation, effectiveness may be a more important index for evaluating reserve networks (Araujo 1999, Rodrigues et al. 1999). 4.1.2 Congruence of Reserve Networks and Units Used for Evaluation Many published studies of reserve networks have used regular grids to map occurrences of species within a region using a geographic information system and have evaluated the adequacy of reserve networks using the same grid (Rebelo and Siegfried 1992, Lombard et al. 1995, Williams et al. 1996, Nantel et al. 1998). Because grids are usually placed arbitrarily on the landscape, the grid cells do not necessarily coincide with natural or geopolitical units of a region (Dobson et al. 1997, Rickets et al. 1999, Rodrigues et al. 1999). Therefore, the units used for reserve network evaluation are usually different than the actual reserve boundaries. A n arbitrary decision is required for assigning protection status to a grid cell based on the proportion of the grid cell that coincides with actual protected areas (Freitag et al. 1998, Rodrigues et al. 1999). Both the assignment of protection status to a grid cell and grid cell size affect reserve network evaluation. Both of these issues are related to the scale of the data and mapping within the region. Although there are no general rules for determining the appropriate scale to conduct evaluations of reserve networks (Stoms 1994, Rodrigues et al. 1999), many published evaluations were completed at one particular scale of mapping and did not explicitly consider the effects of scale on reserve network evaluations. 4.1.3 Population Viability within Reserve Networks The notion of population viability is absent from many studies that have either evaluated existing reserve networks or identified minimum sets (Williams et al. 1996, Virolainen et al. 1999). Many of these studies are based on representing occurrences of species in one or more grid cells, rather than the habitat area that would maintain viable populations. At coarse spatial scales with very large grid cells, representing occurrences of each species may be adequate because large areas of habitat are likely located in each grid cell (Kirkpatrick 1983). However, at fine spatial scales with small grid cells, representing species occurrences without considering the amount of area required to maintain viable populations will lead to the extirpation of species from a region (Soule and Simberloff 1986). Isolated 72  grid cells will not be able to maintain viable populations of every species at fine spatial scales. Therefore, the connectivity of the entire reserve network must be considered when evaluating reserve network performance (Margules et al. 1994, Williams et al. 1996, Margules and Pressey 2000). 4.1.4 Objectives A thorough evaluation of existing reserve networks should measure performance at multiple spatial scales with multiple indices that are important for maintaining viable populations of species (Stoms 1994, Williams et al. 1996). In this chapter, four indices were used to evaluate the existing reserve network in the South Okanagan with the goal of representing habitat to maintain current population sizes of twenty-nine threatened vertebrate species. Although current population sizes of many of these threatened species are probably not viable, population viability assessments were not possible because data required for the assessments were not available (Soule and Simberloff 1986). A confounding factor in the assessment of population viability is that many of the threatened species in the South Okanagan are part of populations that extend outside of the region. Therefore, estimates of current population size of threatened species in the South Okanagan provide a reasonable goal for evaluating the reserve network until population viability assessments can be completed (Rodrigues et al. 2000). The overall purpose of this study was to evaluate the adequacy of the existing reserve network in the South Okanagan for conserving threatened vertebrate species. The relative performance of the existing reserve network was determined by comparing performance indices with minimum sets and randomly selected networks of sites. The research objectives of this chapter were to (1) identify the effect of scale on reserve network performance indices by (a) altering grid cell size and (b) altering the threshold value that is used to assign protection status to a grid cell, which is identified by the proportion of the grid cell that coincides with actual protected areas, (2) determine the 'best' scale, i f any, for reserve network evaluation in the South Okanagan and evaluate reserve network efficiency, effectiveness, redundancy and connectivity at this scale, and (3) identify grid cells that complement the existing reserve network and achieve the regional conservation goals. This study may determine whether published evaluations of reserve networks are reliable and comparable since both the scale of the data and performance indices used for evaluation affect the results. 73  4.2 M E T H O D S 4.2.1 Study Area and Data Sets The study area is defined by the South Okanagan Conservation Strategy and has been described in Chapter 2 of this thesis (Hlady 1990). The 1:20,000 scale terrestrial ecosystem map developed by M E L P provided the base map for reserve selection analyses and has been described in Chapter 2 (Lea et al. 1991). The ecological data that identified the conservation value of individual polygons consisted of wildlife habitat relationship models (Verner et al. 1986). There are twenty-nine W H R models for Red and Blue Listed vertebrate species (Harper et al. 1994) located in the study area (Appendix IV), which predict the amount of habitat available within each T E M polygon for each species (Warman et al. 1998) and are described in Chapter 2. These data are referred to as W H R models and wildlife habitat models, interchangeably, in this chapter. 4.2.2 Methodology Issues 4.2.2.1 Selection Unit Size Two sizes of hexagonal grids were used as selection units in the analyses. 'Selection unit' and 'site' are equivalent to the term grid cell and are used interchangeably in the remainder of this chapter. Hexagons have commonly been used as selection units for reserve selection because they provide a systematic, hierarchical, equal-area unit that minimises spatial distortion (White et al. 1992), and they have a smaller perimeter to area ratio than •  2  square units of the same size. The smallest hexagon size used in the analyses was 2 km , which resulted in a total of 963 selection units for the study area. The 2 km hexagonal grid represents the mean size (2.33 km ) of current areas that are either provincial parks or managed for conservation by the provincial government and private organisations. This value was calculated using spatial data of land ownership for the region obtained from M E L P in October 2000. Polygons identified in the land ownership data could not be used as selection units because the data were incomplete for the study area. The largest hexagon size was 10 km , which resulted in 233 selection units for the study area. The 10 km hexagon size was chosen because Ferruginous hawks occur at a density of 0.1 individuals per square kilometre, which is the lowest density of the twenty-nine vertebrate species included in the analyses. Therefore, 10 km of contiguous habitat is required to support the life history requirements of 2  one hawk.  74  The W H R models were used to identify the amount of suitable habitat in each selection unit for each of the twenty-nine species. The conservation target was identified as the habitat area required to conserve current population estimates of each species within the region. Current population estimates were compiled from various studies and sources (Table 2.1, Appendix V). Population estimates were not available for amphibians and reptiles in the study area. Therefore, populations for these taxa were estimated to be 500 individuals, since general minimum viable population estimates range from 50 to more than 1000 (Franklin 1980, Soule 1987, Caughley and Gunn 1996). Density estimates were calculated from various published and unpublished studies that occurred outside of the study area for eight of the twenty-nine species (Table 2.1, Appendix V). The density estimates were used to determine the amount of habitat area required to maintain current populations of the vertebrate species. 4.2.2.2 Assignment of Protection Status to Selection Units Hexagons were assigned protection status based on the proportion of the hexagon that coincided with the actual boundaries of the existing reserve network. Protection status was determined for each hexagon in the 2 km and 10 km grids with two threshold proportions. The two proportions used for assignment of protection status were (1) i f any portion of the hexagon (greater than 0%) coincided with the actual boundaries of the existing reserve network, the hexagon was considered protected, and (2) i f more than half of the hexagon (greater than 50%) coincided with the actual boundaries of the existing reserve network, the hexagon was considered protected. The first threshold value over-estimates total area of the reserve network and the second value under-estimates the total area. 4.2.3 Evaluation of Reserve Networks  The existing reserve network was identified using two grid sizes and two threshold values for assigning protection status to the grid cells. Therefore, I evaluated the existing reserve network on four different versions of the reserve network. The four versions of the existing reserve network consisted of a: (1) 2 km hexagonal grid of the reserve network, 2  where any portion of each hexagon was protected, (2) 2 km hexagonal grid of the reserve 2  network, where more than half of each hexagon was protected, (3) 10 km hexagonal grid of 2  the reserve network, where any portion of each hexagon was protected, and (4) 10 km hexagonal grid of the reserve network, where more than half of each hexagon was protected. 2 ' 2 The four reserve networks are referred to as: (1) 2km -over (estimate), (2) 2km -under 75  2  2  (estimate), (3) 10km -over (estimate), and (4) 10km -under (estimate) in the remainder of this chapter, respectively. The existing reserve networks, identified by the four different versions, were separated into two categories by land ownership: (1) A l l Protected Areas, and (2) Provincial Protected Areas (Figure 4.1). The " A l l Protected Areas" category was a combination of provincially and privately owned protected areas in the study area. Provincially owned protected areas were considered to have a higher degree of protection than privately owned areas and were therefore analysed separately. This assumption was not because management of biodiversity is enhanced in provincial protected areas, but because these are public lands and are more likely to be managed in the long term to maintain a relatively natural state than privately owned protected areas. The four performance indices used to evaluate the four versions of the existing reserve network for both " A l l Protected Areas" and "Provincial Protected Areas" were efficiency, effectiveness, redundancy and connectivity. 4.2.3.1 Efficiency Efficiency of the existing reserve network was calculated as 1 - (X/T) where X is the total amount of area required to achieve the conservation targets for each of the twenty-nine vertebrate species in the region and T is the total area in the region (Pressey and Nicholls 1989a). To determine the efficiency of the existing reserve network, it was necessary to add complementary sites until the conservation targets for each of the twenty-nine vertebrate species were achieved. Reserve network efficiency was compared to a minimum set that achieved the conservation targets for each of the twenty-nine vertebrate species and with the mean of 1000 random sets of sites that achieved the conservation targets for each species. Random sets of sites were generated without replacement in the programming language of SPlus 2000. Efficiency was comparable between different versions of the reserve network only when all conservation targets were achieved. Therefore, efficiency could be calculated only for the " A l l Protected Areas" reserve network with complementary sites added, since the conservation targets for all twenty-nine species could not be achieved exclusively on provincial land in the study area.  76  CD  TJ  s  2  TJ  u CD I Q-  e  J3  O  •o  L T C  (/) o  0  CD  X CD  o  iiji  c o t? o CL >* c  TJ C  CO  03 CU  TJ  S O  a> &  o i—  CO  c o  CD "55 0) CD o  CL  CD TJ  >  TJ C  CD § _ CD CD J C co  'o o ^5 8CO c 0 CO .E > op c CL C o  "5 8  CO  c?5  X CD  a.  re 'o c §  ca  CO  c • o -ccn CD co > _ x o  * <D  2 c  co  CO CD  0  CD  •g  m  X  >  •2 tj c—  CL  CD CO  LU T J  O "Cl  0  | i r &?C<I 10 r £ C £= ^ CO O •  ^  T  J -£  o  %  _  CO ^  CD  C CD  CO " ' CD  CO CD & ' CD  0  CD CO .9: 0  CD  1  &  2 1 <-  0  CD 0 * - 0 C CO J C 0 L.  0  35  0  32 X  j? F «^  c  CD  UJ T J  0  i  cn  g  1 1 l  l  c: ±= _  I 8 §5 - a" £ x £ o 0 ~ co J C g  i_  77  0 TJ  ~  °  2  f  0  4.2.3.2 Effectiveness Effectiveness of the existing reserve network was defined as the proportion of species in the reserve network that are represented at their conservation targets. The use of effectiveness as a performance index of reserve networks was introduced by Rodrigues et al. (1999). They calculated effectiveness as 1 -  r , where T gap  gap  measures the proportion of species that were  not represented at their conservation target in the reserve network, along with the proportion of each target that was not achieved. In this chapter, effectiveness was calculated as N/K where N is the number of species that are represented to their conservation target and K is the total number of species included in the analyses (K=29 in all analyses). This calculation was based on an "all or none" concept and did not consider the proportion of each species' conservation target that was achieved in the reserve network. The conservation targets in this chapter were identified as the habitat area required to maintain populations of each species. Therefore, a species was considered protected only if its entire conservation target was achieved in the reserve network, since it would likely not persist in the region otherwise. Effectiveness of the existing reserve network was compared with the minimum set of equivalent total area as the reserve network and the mean of 1000 selections of random networks of equivalent total area as the existing reserve network. The probabilities of obtaining the effectiveness values of the existing reserve network and minimum set by chance were calculated using the 1000 sets of random sites. 4.2.3.3 Similarity and Redundancy After complementary sites were added to the network to achieve the conservation targets for each of the twenty-nine vertebrate species, sites that contained more habitat for species than required to achieve the conservation targets (redundant sites) were removed from the network of sites. Redundant sites are defined as sites that do not cause conservation targets for any species to fall below target when the site is removed from the network. Similarity of the existing reserve network with both the minimum set and 1000 random sets of sites was measured using Jaccard's Similarity Coefficient, which provides a coefficient for binary (presence/absence) data. Values of the similarity coefficient range from 0 (no similarity) to 1 (complete similarity). The coefficient is calculated as A/(A+B+C) where A represents the sites that are present in two priority sets of sites, and B and C represent the sites that are present in one of the priority set of sites but are absent in the other (Krebs 1999). 78  4.2.3.4 Connectivity Connectivity of the A l l Protected Areas reserve network was determined using four indices: (1) total number of isolated, contiguous areas (patches) of sites, (2) mean patch area, (3) median patch area, and (4) mean perimeter to area ratio. These indices were calculated with Patch Analyst 2.1 for ArcView 3.1. Connectivity was not evaluated for Provincial Protected Areas because most of the provincially owned land was selected in each analysis to achieve as many of the conservation targets for the twenty-nine vertebrate species as possible, and therefore had small variation in connectivity values. However, the conservation targets could not be achieved exclusively on provincially owned land. 4.2.4 Reserve Selection Algorithm for Minimum Sets Systematic reserve selection software, developed by the National Parks and Wildlife Service in New South Wales, Australia, called C-Plan, works with a GIS to identify conservation areas in landscapes that are subject to the effects of human development (NPWS 1999). Near-optimal solutions for conservation are identified using a heuristic 'Minset' algorithm in C-Plan that identifies the minimum amount of area required to achieve the conservation targets for the twenty-nine vertebrate species. This algorithm uses a complementarity approach to select sites that minimise the amount of redundancy in the minimum set. The rule sequence in the Minset algorithm used to identify conservation sites is described in Chapter 2 of this thesis. The algorithm included an optional check for redundant sites after each set of ten iterations, and was used to identify redundant sites in the existing reserve network. The procedure calculates whether the removal of initially selected sites cause the representation of any species to fall below their conservation target. If there are no species that fall below target then the selection unit is considered a redundant site and is removed from the network of sites. The redundant site becomes available for selection at later stages in the algorithm. 4.3 R E S U L T S 4.3.1 Efficiency of Reserve Networks The A l l Protected Areas reserve network with complementary sites added was slightly less efficient at achieving conservation targets for the twenty-nine threatened vertebrate species than the minimum set, but was more efficient than a randomly selected set of sites (Figure 4.2). The difference between efficiency values of minimum sets and existing reserves  79  that had complementarity sites added was not large, with a maximum difference of 0.09 between the minimum set and 10km -over reserve network (Figure 4.2b). This is likely a consequence of the use of complementarity to select additional sites that achieved the species' conservation targets in the reserve network. Even when redundant sites were removed from the reserve network, after complementary sites were added, the minimum set was more efficient for both the 2 km and 10 km grids. Efficiency values of the different versions of 2  2  the existing reserve network with complementary sites added, as well as the minimum sets, were significantly better than random (P < 0.0001). Randomly selected sets of sites required 2.3 times the area of the minimum sets to achieve conservation targets for the vertebrate species in both grid sizes (Table 4.1). After adding complementary sites, reserve networks that were identified by hexagons where any portion of each hexagon was protected (2km 2  over and 10km -over) were less efficient than reserve networks that were identified by hexagons that were more than half protected (2km -under and 10km -under). 2  2  Efficiency values for each of the reserve network versions identified by the 2 k m and 2  10 k m hexagonal grids were very similar (Figure 4.2). However, the total area of reserve 2  2  2  networks identified by the 10 km hexagonal grid was on average 208 km greater than the area of reserve networks identified by the 2 km grid (SD = 81.19). Therefore, the efficiency index may not be sensitive enough to determine the actual cost of the reserve network, in terms of the amount of area that needs to be reserved, when the scale of the analyses are altered. Table 4.1. Total area (km ) of the All Protected Areas (APA) Reserve Network with Complementary Sites added (ER + CS), APA Reserve Network with Complementary Sites added and Redundant Sites removed (ER + CS - RS), Minimum Set, and mean of 1000 Random Sets of sites. Each reserve network achieved the conservation targets for twenty-nine threatened vertebrate species. All values were significantly higher than random with probability values of P < 0.0001. 2  Reserve Networks  ER + C S  ER + C S - RS  Minimum Set  Random Set  2 k m - over (estimate)  870.00  842.00  -  -  2 k m - under (estimate)  790.00  776.00  -  -  -  -  744.00  1693.35  1090.00  1080.00  -  -  930.00  930.00  -  -  -  -  890.00  2044.63  2  2  2 k m - protection omitted 2  10 k m - over (estimate) 2  10 k m - under (estimate) 2  10 k m - protection omitted 2  80  a) Reserve network identified by a 2 km hexagonal grid 1.00  • Any Portion of Hexagon Protected • More than Half of Hexagon Protected  0.80  0.60  0.40  0.20  J  0.00 ER + C S  ER + C S - RS  Minimum Set  Random Set  b) Reserve network identified by a 10 km hexagonal grid 1.00  • Any Portion of Hexagon Protected • More than Half of Hexagon Protected  0.80  0.60  0.40  iBBi jjjpi 0.20  0.00 ER + C S  ER + C S - R S  Minimum Set  Random Set  Reserve Networks  re 4.2. The efficiency of the reserve networks identified by the (a) 2 km hexagonal grid and (b) 10 km hexagonal grid that occur on both private and provincial land. The reserve networks consist of the All Protected Areas (APA) reserve network with complementary sites (ER + CS), the APA reserve network with complementary sites and redundant sites removed (ER + CS RS), the minimum set and the mean and standard deviation of 1000 random sets of sites. Efficiency is calculated as 1 - (XIT) where X is the total area (km ) required to achieve the conservation targets for twenty-nine threatened vertebrate species and T is the total area available in the database. All efficiency values are significantly higher than random selections of sites with probability values of P < 0.0001. 2  2  2  81  4.3.2 Effectiveness of Reserve Networks 4.3.2.1 All Protected A re as  Effectiveness of the existing reserve networks identified by hexagons where any portion of each hexagon was actually protected (2km -over and 10km -over) was greater than the 2  2  effectiveness of the existing reserve network identified by its actual boundaries (Table 4.2). These reserve networks had larger total areas than the actual reserve network. Effectiveness of the hexagonal reserve networks where greater than half of each hexagon was protected 2 2 (2km -under and 10km -under) was less than the effectiveness of the existing reserve network identified by its actual boundaries. These reserve networks had smaller total areas than the actual reserve network. Table 4.2. Effectiveness of the All Protected Areas (APA) Reserve Network, Minimum Set of equivalent total area as the APA Reserve Network, and mean of 1000 Random Sets of sites of equivalent total area as the APA Reserve Network. Effectiveness was calculated as the proportion of the conservation targets for twenty-nine threatened vertebrate species achieved in each set of sites. The probability of obtaining each effectiveness value by random was calculated from 1000 random selections of equivalent total area and denoted as: NS for P> 0.05; * for P < 0.05; ** for P < 0.01; *** for P < 0.001. Versions of the Existing Reserve Network  A P A Existing Reserve Network  Minimum Set  Random Set  -  -  Actual Boundary  0.724  2 k m - over (estimate)  0.862/VS  0.897**  0.844  2 k m - under (estimate)  0.586*  0.793***  0.656  10 k m - over (estimate)  0.897NS  0.897A/S  0.871  10 k m - under (estimate)  0.345*  0.759***  0.478  2  2  2  2  Effectiveness values for each version of the A l l Protected Areas reserve network were less than minimum sets of equivalent total area, except for the 10km -over reserve network, 2  which had the same effectiveness value as the minimum set (Table 4.2). Effectiveness of reserve networks identified by a hexagonal grid where any portion of each hexagon was protected (2km -over and 10km -over) was better than random, but these values were not significant. Effectiveness of hexagonal reserve networks where more than half of each 2 2 hexagon was protected (2km -under and 10km -under) were significantly less than random. Eight species were not protected at their conservation targets in the existing reserve network identified by actual reserve boundaries: Townsend's Big-eared Bat, American Bittern, Bobolink, Ferruginous Hawk, Long-billed Curlew, Sandhill Crane, Short-eared Owl, 82  and White-headed Woodpecker. Three species were consistently not represented at their target values in each version of the existing reserve network, and in the minimum and random sets that were of equivalent total area as the existing reserve network. These species were Sandhill Crane, Ferruginous Hawk, and White-headed Woodpecker. The latter two species require large areas of land to maintain their current populations and Sandhill Crane habitat was restricted in the database by excluding large lakes from the analyses. 4.3.2.2 Provincial Protected Areas Effectiveness of the 2km -over and 10km -over provincial reserve networks was better 2  2  than the effectiveness of the provincial reserve network identified by its actual boundaries (Table 4.3). Effectiveness of the 2km -under and 10km -under provincial reserve networks was less than the effectiveness of the existing reserves identified by their actual boundaries. This trend was similar to the effectiveness values of reserve networks that were identified for A l l Protected Areas.  Table 4.3. Effectiveness of the Provincial Reserve Network, Minimum Set of equivalent total area as the Provincial Reserve Network, and mean of 1000 Random Sets of sites of equivalent total area as the Provincial Reserve Network. Effectiveness was calculated as the proportion of the conservation targets for twenty-nine threatened vertebrate species achieved in each set of sites. The probability of obtaining each effectiveness value by random was calculated from 1000 random selections of equivalent total area and denoted as: NS for P > 0.05; ** for P < 0.01; *** for P< 0.001. Versions of the Existing Reserve Network  Provincial Existing Reserve Network  Minimum Set  Random Set  -  -  0.759***  0.759***  0.721  2 k m - under (estimate)  0.379/VS  0.690***  0.456  10 k m - over (estimate)  0.862**  0.897***  0.778  10 k m - under (estimate)  0.345A/S  0.655***  0.396  Actual Boundary  0.517  2 k m - over (estimate) 2  2  2  2  Similar to the A l l Protected Areas reserve networks, effectiveness values for each version of the provincial reserve network were less than minimum sets of equivalent total area, except for the 2km -over reserve network, which had the same effectiveness value as the 2  minimum set (Table 4.3). Effectiveness of hexagonal reserve networks, where any portion of 2  2  each hexagon was protected (2km -over and 10km -over), was significantly better than random selections of sites. Effectiveness of hexagonal reserve networks, where more than  83  half of each hexagon was protected (2 km -under and 10km -under), was less than random, but these values were not significant. Fourteen species were not protected at their conservation targets in the provincial reserve network identified by actual reserve boundaries: California Bighorn Sheep, Pallid Bat, Townsend's Big-eared Bat, Great Basin Spadefoot Toad, Blotched Tiger Salamander, American Bittern, Bobolink, Brewers Sparrow, Ferruginous Hawk, Grasshopper Sparrow, Long-billed Curlew, Sandhill Crane, Short-eared Owl, and White-headed Woodpecker. Three species were consistently not represented at their target values in each version of the provincial reserve network, and the minimum and random sets of equivalent total area as the existing reserve network. These species were Sandhill Crane, Ferruginous Hawk, and Whiteheaded Woodpecker. These are the same three species that were not represented in the different versions of the A l l Protected Areas reserve network, and the minimum and random sets of equivalent total area as the A l l Protected Areas reserve networks. 4.3.3 Redundancy in Reserve Networks 4.3.3.1 All Protected Areas There were redundant sites in the versions of the existing reserve network identified 2  2  2  by both 10 km hexagons and 2 km hexagons. A n area of 28 km , which was 7% of the 2km -over reserve network, was redundant after complementary sites were added to the A l l 2  Protected Areas reserve network. This percentage was smaller than the 10% redundancy in the 2km -under reserve network, with an area of 14 km (Figure 4.3). The redundancy in the existing reserve networks identified by the 2 k m hexagonal grid was greater than the 1% 2  7  7  redundancy in the 10km -over and 0%> redundancy in the 10km -under reserve networks (Figure 4.4). Reserve networks identified by hexagons, where any portion of each hexagon 2  2  was protected (2km -over and 10km -over), had larger percentages of redundant sites than hexagonal reserve networks where more than half of each hexagon was protected (2km under and 10km -under). This result is not unexpected since the total area and number of 2  sites considered as protected is larger when the threshold value used to assign protection status to a hexagon (i.e. the proportion of overlap with actual protected areas) is low than when using a high threshold value.  84  (a) A hexagon was considered "protected" when any portion of the hexagon coincided with the existing reserve network. There were 14 redundant sites in this reserve network.  LEGEND g g ^ Redundant Site ggg >0% to 50% Protected [(b) only]  (b) A hexagon was considered "protected" when more than half of the hexagon coincided with the existing reserve network. There were 7 redundant sites in this reserve network. There were 64 complementary sites that were already "protected" at greater than 0% and less than or equal to 50%.  1 = first complementary site selected by the algorithm 2 = second complementary site selected by the algorithm N  r~~"] Not Reserved | Current Reserve • Complementary Site  0  10 Kilometers  A  Figure 4.3. Redundancy of the existing reserve network identified by a 2 sq. km hexagonal grid after complementary sites were added that achieved conservation targets for twenty-nine threatened vertebrate species. The reserve network was identified by hexagons with two levels of protection: (a) where any portion of each hexagon coincided with actual boundaries of the existing reserve network (2km2-over), and (b) where more than half of each hexagon coincided with actual boundaries of the existing reserve network (2km2-under).  85  (a) A hexagon was considered "protected" when any portion of the hexagon coincided with the existing reserve network. There was 1 redundant site in this reserve network.  (b) A hexagon was considered "protected" when more than half of the hexagon coincided with the existing reserve network. There were no redundant sites in this reserve network. There were 31 complementary sites that were already "protected" at greater than 0% and less than or equal to 50%.  Figure 4.4. Redundancy of the existing reserve network identified by a 10 sq. km hexagonal grid after complementary sites were added that achieved conservation targets for twenty-nine threatened vertebrate species. The reserve network was identified by hexagons with two levels of protection (a) where any portion of each hexagon coincided with actual boundaries of the existing reserve network (10km2-over), and (b) where more than half of each hexagon coincided with the existing reserve network (10km2-under).  86  When the threshold value used to assign protection status to a hexagon was high (i.e. greater than fifty percent of the hexagon coincided with actual protected areas), fewer hexagons were considered protected. Some of the hexagons that were not considered protected at this threshold value overlapped actual protected areas, but the percentage of overlap was less than or equal to fifty percent. There was an area of 128 k m that was already 2  protected at less than or equal to fifty percent in the complementary sites added to the 2km 2  under reserve network, which was 20% of the totalarea of the complementary sites (Figure 4.3). A slightly larger percentage (39%) or 310 km ) of the complementary sites added to the 2  10km -under reserve network were already protected at less than or equal to fifty percent 2  (Figure 4.4). The first two complementary sites selected by the algorithm for the 2km -under 2  and 10km -under reserve networks were protected at less than or equal to fifty percent. The first two complementary, sites were identified because these sites have the highest irreplaceability values and are therefore priorities for conservation within the region (Figures 4.3 and 4.4). The selection order of the remaining complementary sites was not identified because the order will change if either of the first two sites is not protected as part of the existing reserve network. The similarity of existing reserve networks to minimum sets of equivalent total area, which was measured as sites that are common to both networks, was significantly greater than random for hexagonal reserve networks, where any portion of each hexagon was protected 2 2 (2km -over and 10km -over) (Table 4.4). Similarity values of hexagonal reserve networks, 7 7 where hexagons were more than half protected (2km -under and 10km -under), and minimum sets of equivalent total area were not significantly different from random. The lack of significance in similarity values may be a result of the lower number of sites that were considered as protected in reserve networks where more than half of each hexagon was protected.  8 7  Table 4.4. Similarity of the All Protected Areas (APA) Reserve Network and Minimum Set of equivalent total area as the APA Reserve Network. Similarity was calculated by Jaccard's Similarity Coefficient. The probability of obtaining each similarity coefficient by random was calculated from 1000 random selections of equivalent total area (i.e. Random Set) and denoted as: NS for P > 0.05; * for P < 0.05; ** for P < 0.01.  Versions of the Existing Reserve Network 2 km - over (estimate) 2 km - under (estimate) 2  2  10 km - over (estimate) 10 km - under (estimate) 2  2  All Protected Areas 0.148* 0.039A/S 0.248** 0.000A/S  Random Set 0.119 0.036 0.172 0.024  4.3.3.2 Provincial Protected Areas Four redundant sites were identified in the 2km -over reserve network on provincially 2  owned land (Appendix XI). There were no redundant sites in the existing reserve networks 2  2  identified by the 10 km hexagonal grid or in the 2km -under reserve network. Minimum sets identified by the 10 km and 2 km hexagonal grids required 99% of the provincially owned land in the study area to achieve conservation targets for 90%> of the vertebrate species. Since most of the study area was selected by the minimum set, redundancy in the provincial existing reserve network is minimal. A n area of 74 km was already protected at less than or equal to fifty percent in the 2  complementary sites added to the 2km -under reserve network, which was 10%> of the total area of the complementary sites (Figure 4.4). A slightly larger percentage (16%> or 180 km ) of the complementary sites added to the 10km -under reserve network were already protected at less than or equal to fifty percent. This trend was similar to the redundancy values of reserve networks and percentages of complementary sites that overlap actual protected areas that were calculated for A l l Protected Areas. Similarity values of hexagonal reserve networks where any portion of each hexagon was protected (2km -over and 10km -over), and minimum sets of equivalent total area were significantly better than random (Table 4.5). Similarity values of the existing reserve network identified by hexagonal grids where hexagons were more than half protected (2km -under and 10km -under), and minimum sets of equivalent total area were less than random. However, 2  only one of these similarity values was significant.  88  Table 4.5. Similarity of the Provincial Reserve Network and Minimum Set of equivalent total area as the Provincial Reserve Network. Similarity was calculated by Jaccard's Similarity Coefficient. The probability of obtaining each similarity coefficient by random was calculated from 1000 random selections of the equivalent total area (i.e. Random Set) and denoted as: NS for P > 0.05; * for P < 0.05; *** for P < 0.001. Versions of the Existing reserve Network  Provincial Protected Areas  Random Set  0.216* 0.018*  0.173 0.068  0.361*** 0.000/VS  0.209 0.051  2 km - over (estimate) 2  2 km - under (estimate) 2  10 km - over (estimate) 2  10 km - under (estimate) 2  4.3.4 Connectivity of Reserve Networks There was no consistent trend in the indices of connectivity for each of the versions of the existing reserve network and minimum sets (Table 4.6). However, the versions of the existing reserve network with complementary sites added generally had fewer patches, a greater mean and median patch area, and a lower mean perimeter to area ratio than minimum sets. The mean perimeter to area ratio of the existing reserve network identified by its actual boundaries was two orders of magnitude larger than the ratios of reserve networks identified by hexagonal grids. The 2km -over reserve network with complementary sites added and redundant sites removed had the smallest number of patches, largest mean patch area, one of the largest median patch areas, and lowest mean perimeter to area ratio (Table 4.6a). Therefore, this reserve network had the highest connectivity. The minimum set identified by 2 km hexagons had the lowest mean perimeter to area ratio and one of the lowest median patch areas. 2 2 Similar to the 2km -over reserve network, the 10km -over reserve network with complementary sites added and redundant sites removed had the smallest number of patches. However, the 10km -over reserve network with complementary sites added without removing 2  redundant sites had an equal (smallest) number of patches (Table 4.6b). The latter network also had the largest mean and median patch area and the mean perimeter to area ratio of this network was only slightly larger than the smallest ratio, with a difference of 0.01. Therefore, the reserve network that had the highest connectivity was the 10km -over reserve network 2  with complementary sites added. The minimum set had the lowest mean patch area, highest mean perimeter to area ratio and one of the highest number of patches. 89  CN  8  0 0 co  ro  o TJ 0  CO  x , ° jet  cn  i E  .2  cz  CM  IS CD  O + LU  O  CM 00  XJ XJ 10 DT ro o JZ XJ CO  £ o C CD CD x CO  CD  CD  CD  £ ° to  a ccr-  TJ CO c ro o 0 ro CL to  CO  O  •c ro  12 §  ro co 0) <  c  co  s  E  E  CD  x>  0  II o > t l  co  8^  TJ 0 >  LU CD CD  O w  CO  o  O  LO  O CD  o o  O O  O CD  CO  CO Cjj  LO CO LO  «  CO  CO  TJ  " £ ™ ro < -3 S3£ o TJ £ ^ ^CO o c: 0o o XJ ro S a c s»< ° 0 J3  o o  LO  o co  o o  CM  1^ CM  u.  .9> £ ^  -Q -o  CM O  i  to  o £  O O  CM  CO CO  JZ  J  co  CD  O O  CO  CM  ro  0 "SCCOD ^ ro W £-E CO  LO  to  ro 0 4= ^O 0 TJ - CO-e<= £ XJ 2 C CD ^  CM LO  CD  a>  ^ °- 0 '.Z. E  "o 2 T: Z CD  CL .to  o o  SI  5  co  oT  c  CD  35  £ to  co T - c  CD  CO  0  O O  o >CO O) $ o I- 0 SQ, ro * P CD £ c x 3 < < JD £ 0C -1 O 0w 0 XJ  TJ o CD  * r  CM  *!  T:  > — 0 J= 3R ro 0 ro 0 0 TJ 0 > o £ «0 °^ CO E .2 0 0 0 o & >. CC 0 Q. c o | O ° £ < ro 0 o ^ !£ ro 0S: 0 .£ CC 0 0 E O2 XJ . 0 0 CL 0 -2 > 0 cz 0 oW OE 0 0 o 0 CP CL 0 0 -b •!=! 0 -S - * 0 £ Z TJ  ci  o  d  d  o o  o o  o o  CO  CD  1^  d  £ 0  in 0  c  ^  F  CN  a: i  - E ° L  1§  CO  CO  LO  0  co  <  CM  CD CO CO  to  w  ^-  to  o o  00 CO CO LO  o  o o 5  o  o  o o  o  o  CD oo  oo  d  CM  CO  CO  U  D)  ^  L  o  O  To  CO 0  to  JO  -CZ ~ ,  E  If  — I  §  I ro  cu  co  0  to  S  0  CO  x>  0 a:  0  co  t  0 XJ o > 0 0 XJ CC | CO  §2 c3 &<  10 ®£  E ro CO< !C 0 CD £ 0 £ 0 F= JC £ co 0 j a co  0 CO 0  to  co  o c 0 ro  £  co  c o  CD CO X 0 JC CN  CC Ol c  X LU 'x o £J: TCOJ LU  .  0  M— O  in  c g U) i  >  c D O  0  ro  > o 0  ^—'  c  E 0  E  0 0  CO  CM,  in 0 £ 0  in 0  CC  cn c  ca LU ro Ti <  CO 0 £ 0  in 0  CC  cn c  X LU  E E  'c  2  P o o  O I (OCM  £ 0  m 0  CC u>  w  X LU 90  0  1?  i " C J  •c c  ro *—*  c  CO  EJ E 0  Em O  CD  = c CO  E E 0CN  0  0  Q . CD ?  E  ° T U3  CO 0  CC  TJ  CD  rf U  W  II CO  X LU  S C M £ 0 (0  ^ :  Q-CM^ O <0  CJ  S  co CO  1  S i TE L0  CJ  c I  0  = E CL Cl ^  0  co -c>,  CO ^ CO >^  0  to  < I CN  to  CO  in  m  CO 0  co -a  MTJ CD U) 0  0  CC  CC  Qi co  X LU  1  0  cn  ?!  X LU  E  "35 E  CM  O + LU  CU CL c  CM CO  O  CO  CO  CO  d  d  J~ ^  00  o  1^  o  O o d co  o o d co  o  O  CD CO  CM CO CM  CM CO CM  CD  C D C 0D 0  C D C D CO  o o d  o o d co  o  O  o o d  o o d  C CO O C O C CO O  o o d  co  cu o  CM  CO c £L Jij oLU c CD CO COCO  o o d  o o  o o d  O O  O O  O  CM  =5 £  CD < O M  CM  m  C M C N CM  CD CO  lO  CO  C CO D CD O E  CO 0. TJ  CD 0) CD c o CO CO D o C T)  o o  C N C D CN  Q.  CM  o d  0 0 CO  CD  d  d  d  O  d  TJ  CO  CD £  cu  co ™ oo CD { CC 1 CO 1—  o  CO  If)  o o d oo  o o  o o d  o o d  C 0D 0  CD  o o d co o  CD  o  CO  OJ  CD  CO  •4—>  o  — I  0)  co  c Co D  c o aj  CO >>  co -c>  CO  CO  CO  C CO D £ CU TcuJ C 0D ) > c CD CU o cc TJ £ CD E CO CDCDl£ o C C o CD X CO CD LU ( 0 cu E 00 CD o £ x cu D LU £ £> co C O cu coD E C D TC J JZ cc CD c CD CC c o COQ c c o "co "co CD I t) X c CD < LU UJ > X o  CO  M—  CO  LU  ro c CD cu  CO  E  _gj , ~ .  OJ  cz o IS I Crv, E Q.  Q. CD  i§  O  I  ^  «? o T-  i«  cj  B  CD  E „ CD - C -  f-s  O ZJ  u  "5  If CD™ CD  Q.O  E ~  o (0  0> CO co C CO D o^ CU T J £ — £ i CD CU -o CD CU ry CC CD _C I  CfltN  co™ co CO DE CO o CD •a So CO T J £ C D C O CD M CC DJ CD CC "55 C  0  J CO  CO  "55  X  LU  x  CC co LU ? ! 1 E  x LU  91  1 E X  LU  4.4 D I S C U S S I O N 4.4.1 Methodology for Evaluating Reserve Networks The estimated performance of the existing reserve network in the South Okanagan for protecting threatened biodiversity varies with the methodology used to conduct the evaluation. Both selection unit size and assignment of protection status to selection units affect the measurements of the efficiency, effectiveness, redundancy and connectivity of the existing reserve network. Generally, efficiency and redundancy of the reserve network that included all of the protected areas were lower for the 10 km selection unit size, and 2  connectivity was higher than for the 2 k m selection units. Effectiveness was generally higher 2  for the 2 km selection unit size. This result can be explained by the scale effect outlined in the Modifiable Areal Unit Problem (MAUP), which is defined as "a problem arising from the imposition of artificial units of spatial reporting on continuous geographical phenomena resulting in the generation of artificial spatial patterns" (Heywood et al. 1998). Size of the selection unit affects the results because the data are aggregated into artificial boundaries at different levels of spatial resolution. The aggregation of data into large selection units results in more homogeneous selection units and a decrease in the precision of the mapped data (Heywood et al. 1998). Because the relations between variables change with the use of different selection units, the reliability of published reserve network evaluations is a concern. Redundancy is lower in reserve networks that are identified with larger selection units since there are fewer sites available for selection. As a consequence of the fewer sites mapped within the region, the probability that each site is irreplaceable for a particular species increases, thereby decreasing the redundancy of each site within the region. Larger selection units also contain more area than smaller units. Therefore, the effectiveness and connectivity of the reserve network are higher. Although efficiency values were very similar between large and small units (Figure 4.2), larger units have a tendency to over-represent the conservation targets for species in reserve networks as a consequence of the increased area (Pressey and Logan 1995). Therefore, in terms of the total area of the reserve network, larger units are less efficient than smaller units of selection. The effect of the assignment of protection status to a hexagon is also explained by the scale effect in M A U P . The aggregation of actual protected areas into hexagons requires an arbitrary decision on how to assign the level of protection to a hexagon (Rodrigues et al.  92  1999). To determine the adequacy of the existing reserve network using hexagons or some other grid system, it is necessary to identify a hexagon as either protected or not protected. Assigning the protection status of a hexagon as protected when any portion of a hexagon overlaps the actual protected areas results in the inclusion of more area into the hexagonal reserve network than is bounded by actual protected areas (Table 4.6). Hexagons that are considered protected if more than half of the hexagon overlaps current protected areas result in a network that has less area than the actual reserve network. Therefore, effectiveness and redundancy of the existing reserve network are higher and efficiency is lower when comparing reserve networks that are identified by hexagons that have a low proportion of overlap with actual protected areas, to networks where hexagons have a high proportion of overlap (Freitag et al. 1998). These artifacts are a concern when evaluating the effectiveness of the existing reserve network. By excluding protected areas that do not coincide with more than half of an arbitrary grid cell, the evaluation of effectiveness of the reserve network results in a value that is less than the actual reserve network. The problem with evaluating reserve networks using a regular grid system, where only a portion of each grid cell coincides with actual protected areas, is that it is not known whether the habitat of the desired species actually occurs within the protected portion of the hexagon. The assumption that the entire hexagon is protected results in the inclusion of areas that are not actually protected. The 10km -over (estimate) reserve network was the most optimistic 2  reserve network because of the extra area that was considered protected in each selection unit. The evaluation of this network identified that it was effective at conserving habitat for twentyfive threatened vertebrate species included in this thesis (one fewer species than the minimum set of equivalent total area). Although this reserve network with complementary sites added was much more efficient than random sets of sites, it was not as efficient as the minimum set, 9  even with redundant sites removed from the network (Figure 4.2b). In contrast, the 10km under (estimate) reserve network was the most pessimistic, because it excluded more than 40% of the total area of the actual protected areas. The most pessimistic reserve network was not nearly as effective as the minimum set, and was significantly less effective than randomly selected networks of sites (Table 4.2). A n evaluation of the reserve network in the South Okanagan that was based on the scale that provided an optimistic view would conclude that the existing reserve network is satisfactory, whereas an evaluation based on the scale that provided a more pessimistic view would conclude that the existing reserves are inadequate. 93  Therefore, it is possible to bias results of reserve network evaluation based on the scale of the analyses. Selection unit size and the method used to assign protection status to a selection unit have an effect on the evaluation of the performance of reserve networks. The arbitrary decisions required to assign protection status to a selection unit that result from the mismatch of the selection unit boundaries with land ownership data can be avoided by using geopolitical units of the region (Pressey et al. 1993). However, a mismatch between units cannot be avoided entirely in the South Okanagan, since geopolitical units dissect natural units (polygons of vegetation associations) used to develop the wildlife habitat models. Geopolitical units could provide a useful comparison to evaluations produced with hexagonal grids, however, they were not available for this chapter. Although performance index values 9  9  of the 2 km hexagonal grid were closer in value to the actual reserve network than the 10 km grid, it is not possible to generalise the findings in this chapter to results expected when other data (i.e. other taxa or regions, occurrence data versus WHR models, etc.) are used. However, two recommendations arise from the results in this chapter. The first is that the performance of reserve networks should be determined with multiple indices at different scales (i.e. selection unit size and assignment of protection status), reports should include explicit methodology and rigorous comparisons of alternative versions of the reserve network (Pressey and Tully 1994, Heywood et al. 1998, Cabeza and Moilanen 2001). The second recommendation is that the performance of reserve networks identified by their actual boundaries should be used as comparisons to evaluations of reserve networks identified as grids. Future studies could determine the selection unit size and shape that provides the best approximation to the actual boundaries of the reserve network and how to assign protection status to selection units to get performance values that are close to the values of the actual reserve network. 4.4.2 Reserve Network Performance in the South Okanagan I have based the overall evaluation of the reserve network in the South Okanagan on the analyses using the 2 km hexagonal grid because this hexagon size provided performance values that were most similar to the actual reserve network boundaries. The existing reserve network identified by the 2 km hexagonal grid did not represent all of the conservation targets for the twenty-nine threatened vertebrate species. Therefore, efficiency could be measured only when complementary sites were added to the network that achieved the 94  conservation targets for each species. The efficiency of the existing reserve network with complementary sites added was less than the minimum set (Figure 4.2). This result confirms the general opinion that existing reserve networks are inefficient and inadequate at representing regional biodiversity (Pressey and Tully 1994, Lombard et al. 1995, Freitag et al. 1998, Rodrigues et al. 1999). However, the total area of the reserve network with complementary sites added was significantly less than random selections of sites that achieved the conservation targets for each species (Table 4.1). Although the existing reserve network was selected for reasons other than simply representing biodiversity, its efficiency is much better than a randomly selected set of sites. Therefore, reserves that have been chosen on an ad hoc basis may still be valuable for representing biodiversity within a region (Rodrigues et al. 2000). These results are encouraging, since existing reserves will likely form the foundation of future protected areas (Rodrigues et al. 1999). The effectiveness of the A l l Protected Areas and Provincial Protected Areas reserve networks, at representing conservation targets for each of the twenty-nine threatened vertebrate species, was also encouraging since both values are better than randomly selected networks of sites. The effectiveness values were also comparable to values calculated for minimum sets of equivalent total area, even though the minimum sets were more effective overall. The range in the number of species protected in reserve networks, as a result of the threshold proportion used to assign protection status to a hexagon, overlapped the range in the number of species represented in minimum sets of equivalent total area (Tables 4.2 and 4.3). However, the range in effectiveness of the existing reserve network implies that there is potential for unreliable evaluations of reserve effectiveness. Freitag et al. (1998) found that the effectiveness of a reserve network increased with a decrease in the threshold proportion used to assign protection status to a grid cell. Therefore, a general conclusion about the effectiveness of existing reserve networks may not be possible because of the effects of scale. Redundancy in the A l l Protected Areas reserve network with additional complementary sites that achieved the conservation targets for the twenty-nine vertebrate species also varied with the assignment of protection status (Figure 4.3). However, some of the complementary sites added to hexagonal reserve networks where more than half of each hexagon was protected were already protected at less than or equal to fifty percent. Although similarity, in terms of the number of sites common in both the existing reserve network and minimum set of equivalent total area, was quite low, it was significantly better than random (Tables 4.4 and 95  4.5). Therefore, both the similarity between the existing reserve network and minimum set, as well as the addition of complementary sites that already had some degree of protection, indicates that the ad hoc decisions used to protect areas in the South Okanagan have included the representation of biodiversity as effectively and efficiently as possible, given the constraints inherent in protecting areas. The low similarity values between the minimum set and existing reserve network are partially explained by the difference in their connectivity. The connectivity of the existing reserve network identified by hexagons, where any portion of each hexagon was protected, was better than minimum sets for most connectivity indices (Tables 4.6). The minimum set did not have the best value for any of the connectivity indices, with the mean perimeter to area ratio consistently ranking the worst out of all the versions of the existing reserve network that was represented by hexagonal grids. Unfortunately, the algorithm used to produce the minimum set cannot identify sites that are in close proximity to previously chosen sites. Therefore, although the Minset algorithm can identify sites that achieve a desired total amount of area, the algorithm cannot identify the appropriate configuration of a reserve network (Nicholls and Margules 1993, Williams et al. 1996, Pressey et al. 1997). A n 'adjacency rule' that preferentially selects sites that are closer to already reserved sites is currently being developed to increase the connectivity of minimum sets. The inclusion of an adjacency rule has been found to decrease the efficiency of the reserve network (Bedward et al. 1992, Nicholls and Margules 1993, Lombard et al. 1995). However, this result may depend on the distribution of the species in the region, since the adjacency rule did not significantly effect the efficiency of reserve selection for species with overlapping distributions (Freitag et al. 1998). Once adjacency has been incorporated into the algorithm, evaluations of existing reserve networks may be more positive, since efficiency values will be more similar to minimum sets and sites identified by minimum sets will be more practical for implementation. 4.4.3 Recommendations 4.4.3.1 A11 Protected A reas The performance of the existing reserve network depends on the index used for evaluation. If efficiency is the only measurement used for the evaluation of the existing reserve network in the South Okanagan, the network is insufficient. However, the effectiveness of the South Okanagan reserve network was comparable to the minimum set and 96  the connectivity was better. The redundancy of the existing reserve network is less useful as an index, since this index reflects the values of efficiency and effectiveness, and because it is not likely that current protected areas will be traded for areas that are more complementary and less redundant. In addition, some redundancy in the reserve network is desirable as insurance that rare species will not be lost because of stochastic environmental disturbances (Soule and Simberloff 1986, Rebelo and Siegfried 1992, Margules et al. 1994). The similarity of the reserve network with the minimum set may not be an important comparison either, since the patches of sites selected by the minimum set may be too small to conserve species (Rodrigues et al. 2000, Cabeza and Moilanen 2001). The inclusion of an adjacency rule in systematic reserve selection algorithms may increase the importance of this comparison in the future. The South Okanagan study area is small relative to the distributions of most of the threatened vertebrate species included in these analyses. Therefore, population viability is considered over the entire study area rather than for individual contiguous patches of sites. The avian and mammalian species included in this study can move from one side of the valley to the other relatively effortlessly (Warman et al. 1998). Therefore, individual patches of sites will not maintain viable populations of these species exclusively within them. The amphibians and reptiles included in these analyses migrate only a few kilometres from breeding sites. When data become available to identify the requirements for viable populations of these species and an adjacency rule is included in the algorithm, future analyses should identify viable habitat patches for these taxa. The performance of the reserve network will need to be monitored over time to account for the movement of populations of species in response to climatic change and human encroachment (Margules et al. 1994, Margules and Pressey 2000). Contextual data for each selection unit, such as road density, physical barriers to dispersal, timber resources and other threats that might influence the selection and implementation of a reserve network, will also need to be considered in future analyses (Margules and Pressey 2000). It is also important to emphasise that the data used for the analyses in this chapter are based on predictions of habitat use. Therefore, the data should be verified before protecting any of the sites that were identified using the complementarity algorithm in this chapter. However, the existing reserve network identified by actual boundaries of the protected areas is not likely to maintain viable populations of any of the twenty-nine species in the 97  South Okanagan. Most of the patch sizes in the existing reserve network are too small to maintain all of the life requisites of many of the threatened species within the protected areas (Table 4.6). To increase the viability of the existing reserve network, areas surrounding each of the protected areas should be included in the reserve network either as formally protected areas or as buffers around protected areas that are managed for wildlife (Soule and Simberloff 1986). The evaluations using hexagonal grids may help to identify priority areas where these efforts should be focused. In addition to increasing individual patch sizes, the existing reserve network must be augmented with complementary sites to include areas for the eight species that are not protected at a level that will maintain their current populations in the South Okanagan. Although the first two sites selected by the complementarity algorithm were identified as priorities for conservation within the region, the location of the sites varies depending on the scale of the evaluation. Therefore, it is not possible to recommend a particular site as being the most important for conservation within the region. 4.4.3.2 Provincial Protected Areas In British Columbia, the government has implemented a Protected Areas Strategy (PAS) to protect 12% of the provincially owned land base (LUCO 2001). As of 2000, this goal has been achieved. In the South Okanagan, the Provincial Protected Areas are located primarily at higher elevations, protecting forest rather than shrub-steppe where most of the threatened species occur. The Provincial Protected Areas cannot maintain the current population sizes of fourteen of the twenty-nine threatened species included in this thesis. There has been an addition of 156 km to provincial protected areas in the South Okanagan as a result of the approval of the Okanagan-Shuswap Land and Resource Management Plan (LRMP) in January 2001, which has not been included in the analyses in this chapter. However, even if all of the provincial land in the South Okanagan study area was protected, current population sizes of three of the vertebrate species could not be maintained (Table 4.3). Not only has the 12% goal been criticised because it is too small to maintain viable populations of species (Soule and Sanjayan 1998), but the land base that the Protected Areas Strategy has to work with may be inadequate to prevent extinction of many threatened species. Fortunately, many non-government organisations are aware of the inadequacies of PAS and are managing private lands for threatened wildlife.  98  4.5 C O N C L U S I O N S While the existing reserve network in the South Okanagan identified by actual boundaries of protected areas is inadequate for protecting the threatened biodiversity included in this thesis, evaluations of the reserve network using hexagonal grids at different scales produced both positive and negative evaluations of performance. The results in this chapter demonstrate that the scale of analyses influences the evaluation of reserve networks. Many published studies that identify or evaluate reserve networks analysed data at only one scale. Therefore, the reliability of these studies is low. Scale effects must be considered in any analysis that is affected by M A U P , especially if general conclusions are desired (i.e. on the performance of current protected areas). The relative performance of existing reserve networks has often been compared to the efficiency of minimum sets. Minimum sets of sites identify efficient conservation networks, however, their connectivity can be low when compared with existing reserve networks and, thus, may not provide an appropriate configuration for reserves. It may not be useful to compare the efficiency of current protected areas with minimum sets until sites can be aggregated into viable patches of habitat for species. However, the effectiveness of the existing reserve network in the South Okanagan was comparable to a minimum set of equivalent total area, which means that the minimum set was not much better at protecting threatened biodiversity than the current protected areas in the South Okanagan. Therefore, the ad hoc process of conservation inherent in existing reserve networks may provide a reasonable starting point for protecting biodiversity and may not be as flawed as generally thought.  99  5.0 C H A P T E R 5: G E N E R A L C O N C L U S I O N 5.1 S U M M A R Y Effective reserve selection approaches that consider the ecology of species and the landscape, as well as identify options available for conservation, are essential to minimise the loss of biodiversity. Systematic, scientific conservation planning approaches based on a complementarity selection process have been developed and successfully implemented as management tools to prioritise conservation efforts (Margules and Redhead 1995, Faith and Walker 1996, Williams et al. 1996, Csutsi et al. 1997, Pressey et al. 1997). The principle of complementarity allows for the identification of areas that provide the greatest contribution of unprotected biodiversity to an existing network of protected areas (Vane-Wright et al. 1991). Since the resources available for conservation are limited, complementarity is an essential component of effective reserve selection. However, it is necessary to apply systematic reserve selection approaches to current conservation crises to evaluate the success of the selected reserves over time (Cabeza and Moilanen 2001, Brooks et al. 2001). Scientific studies of systematic reserve selection have usually been completed at coarse spatial scales (i.e. 1:500,000), and recommend that fine spatial scale (i.e. 1:20,000) analyses should be performed in areas that were identified as priorities for conservation at coarse scales (Nicholls and Margules 1993, Freemark et al. 2000, Brooks et al. 2001). Selection units at coarse scales are often large enough to maintain populations of species within a particular unit and reserve design considerations can be applied within each unit during implementation. However, at fine scales the selection units are small and the consequences of reserve design must be considered during the selection process (Soule and Simberloff 1986, Cabeza and Moilanen 2001). In addition to the effect of regional scale, other components of systematic reserve selection techniques influence the selection of minimum sets of sites (Pressey and Logan 1995, Pressey et al. 1999, Rodrigues et al. 1999, Virolainen et al. 1999). In this thesis I used C-Plan, systematic reserve selection software, as a tool for investigating the process of reserve selection using complementarity at the fine spatial scale of the South Okanagan region (NPWS 1999). The South Okanagan region in British Columbia has been identified by a national systematic reserve selection study as a priority for a fine scale analysis of biodiversity (Freemark et al. 2000). By examining the effects of components of systematic reserve  100  selection on areas that are identified as priorities for conservation within a region, we will have a better understanding of the success of these techniques for conserving biodiversity. 5.1.1 Ecological Data Used for Reserve Selection The first component of reserve selection that I investigated was the reliability of the ecological data used in reserve selection. The most complete data sets available for determining conservation priorities within the South Okanagan region were related to the location of threatened vertebrate species. The occurrence data for threatened vertebrate species were not collected systematically throughout the region and observations were significantly closer to roads than expected from a uniform distribution of occurrence. These data represented a biased sample of the habitats used by threatened vertebrate species. Therefore, twenty-nine wildlife habitat relationship models that predicted the amount and location of suitable habitat for threatened species within the region were used for reserve selection (Verner et al. 1986, Warman et al. 1998). These habitat models have not been verified in the field and, therefore, the predictions from the models and the areas selected by the complementarity algorithm that were based on these predictions may be inaccurate. The success of systematic reserve selection depends on the data used in the analyses and, therefore, future work should focus on developing an unbiased sample of species occurrence in the region (Freitag and van Jaarsveld 1998, Brooks et al. 2001). 5.1.2 Scale of Reserve Selection 5.1.2.1 Selection Unit Size The second component of systematic reserve selection that has a large effect on conservation priorities is the spatial scale of the data. In this thesis, I investigated the effect of altering selection unit size and shape on the sites selected by the reserve selection algorithm. Different sizes of selection units resulted in reserve networks that were dissimilar (i.e. the algorithm did not select overlapping sites), but they were more similar to each other than to a random selection of sites. The variability in the sites selected by the reserve selection algorithm brings the reliability of published reserve selection studies that were performed at only one spatial scale into question. Because results can be altered depending on the unit chosen for mapping ecological data, the results of systematic reserve selection can be biased by an a posteriori choice of scale. The use of different scales to bias results could be detrimental to biodiversity if, for instance, urban developers initiated systematic reserve selection studies to justify locations for urbanisation. Conversely, if the selection unit size is 101  chosen a priori, the potential variability in the selected sites is unknown and the reserves that are identified at one unit size may not provide the best reserve network to conserve viable populations of species overall. 5.1.2.2 Biodiversity Indicator The data used to identify regional biodiversity affects the sites selected by the reserve selection algorithm. I used twenty-nine threatened vertebrate species from four taxonomic groups as an indicator for biodiversity within the South Okanagan region. These species are on the provincial Red and Blue Lists and are either endangered or vulnerable to extinction, respectively (Harper et al. 1994). The data for the South Okanagan were more extensive than data that were available for other regions in British Columbia. Therefore, it is important to understand how reserve selection changes when the data available for a region are limited. As a comparison to the sites identified for the threatened vertebrate species, I identified a set of sites that represented habitat for a subset of the threatened species (eleven Red Listed species), a set of sites for vegetation classes, and a set of sites for twenty-nine threatened vertebrate species and vegetation classes collectively. The similarity between these sets of sites was significantly better than the similarity between these sets and randomly selected sets of sites. However, the two most similar sets of sites, which were identified for the twenty-nine threatened vertebrate species and the subset of threatened species, were only 52% similar. Since the data used as the biodiversity indicator for a region influences reserve selection, all reliable species-based information should be included in conservation assessments (Gaston 1996, van Jaarsveld et al. 1998, Freemark et al. 2000). Even though the similarity between the different sets of sites was low, the sites identified for both the subset of threatened vertebrate species (ten Red Listed species) and the vegetation classes represented 93%) of the habitat required to maintain all twenty-nine threatened species. Therefore, in the South Okanagan region a subset of species found in the region could be used as a surrogate to identify a reasonable network of sites that conserves all regional biodiversity. If the data on species distributions were limited, vegetation classes could also be used to identify a reasonable network of sites. However, the set of sites selected to represent vegetation classes consisted of a greater total area than the sites selected to represent all twenty-nine vertebrate species. It is important that systematic reserve selection is based on the best available data within a region, but that it is not impeded by the lack of comprehensive biological data (Freemark et al. 2000, Polasky et al. 2000). Data that act as surrogates for 102  regional biodiversity should be used when more detailed data are limited because conservation decisions must be made before the options available for conservation are diminished. However, surrogates should be chosen based on their ecology rather than on political reasons, such as the ten Red Listed species, because of their influences on reserve selection. 5.1.2.3 Conservation Target Perhaps the most important component of systematic reserve selection is the conservation of viable populations of species within selected complementarity sites. Conservation targets used in reserve selection should, therefore, identify habitat that maintains minimum viable populations of each species. However, I could not identify minimum viable populations of threatened species included in this thesis. The detailed data on demography of species and effects of environmental stochasticity on species, required for population viability analysis (Shaffer 1981, Soule 1987), were not available. The South Okanagan region is small relative to the distribution of many of the threatened vertebrate species and many species interact with populations outside of the region. Therefore, viable populations that are isolated from other populations likely do not occur in this region. Because minimum viable population sizes are difficult to determine, it is important to understand how conservation targets affect the selection of sites in systematic reserve selection. In this thesis, I altered the conservation target for each threatened vertebrate species. Most of the resulting sets of complementarity sites had low similarity and were not significantly different from randomly selected sets of sites. The sets of sites selected for conservation targets that represented a certain number of sites for each species had the lowest similarity values when compared to sites that represented the habitat area requirements of each species. Representing species in a certain number of sites cannot guarantee that there is an adequate amount of habitat area in each site. Therefore, conservation targets must be chosen carefully and should be based on the habitat area required to maintain each species. 5.1.3 Existing Reserve Network Performance Existing reserve networks will likely provide the foundation for future protected areas, even though they have been selected for reasons other than simply conserving biodiversity (Rodrigues et al. 1999, Margules and Pressey 2000). Therefore, it is important to determine the level of biodiversity protection in existing reserve networks so that future reserves complement existing reserve networks and avoid redundancy. Because of the effects of scale on reserve selection, I chose to evaluate the existing reserve network in the South Okanagan 103  with two selection unit sizes in this thesis. The two units resulted in opposing evaluations of reserve network performance. The larger unit produced results that concur with the general belief that existing reserve networks are inadequate and inefficient at protecting biodiversity (Rebelo and Siegfried 1992, Pressey et al. 1993, Williams et al. 1996, Freitag et al. 1998, Nantel et al. 1998, Margules and Pressey 2000). The evaluation of the reserve network using the smaller unit, which provided results that were more representative of the actual reserve network, determined that the reserve network was satisfactory, but needed to be augmented with complementary sites to maintain populations of the threatened species. I evaluated the existing reserve network with only two different selection units. Therefore, a more rigorous testing procedure of different selection units should be performed to determine the unit that optimally represents the actual reserve network. However, the conflicting results demonstrate that scientific studies evaluating the performance of existing reserve networks may be unreliable if they were completed at only one spatial scale and only measured efficiency (total area that achieved the conservation goal) of the reserve network. 5.2 S P E C I E S S U R V I V A L P R O B A B I L I T Y 5.2.1 Species Interaction Reserves are usually identified to conserve species, however, it is difficult to identify all of the factors that affect species survival probabilities in reserve networks (Witting et al. 2000, Cabeza and Moilanen 2001). Although, population viability analyses aid in the identification of factors that have large influences on species survival (Boyce 1992), it is difficult to include these factors in systematic reserve selection. There are two reasons why it is difficult to incorporate interspecific and intraspecific processes in reserve selection algorithms. The first is that the biological and ecological knowledge may not be adequate to model the interdependence between species and processes (Goldstein 1999), and the second is that the interdependence may be difficult to solve computationally (Polasky et al. 2000). However, two factors may be relatively easy to include in reserve selection algorithms: (1) the dependence of species on the presence of other species (i.e. keystone or prey species) for their survival (Witting et al. 2000), and (2) the exclusion of a species based on the presence of another competing or predator species (Witting et al. 2000, Schoener et al. 2001). Both of these issues could be included in the reserve selection algorithm by selecting sites for a particular species where a keystone-species is present or competitor species is absent.  104  Systematic reserve selection may not be able to identify viable multi-species reserve networks unless species survival probability is considered in the algorithms. 5.2.2 Socio-economic Influence Species survival probability will also be affected by socio-economic factors (Ando et al. 1998, Brooks et al. 2001). It is possible to incorporate socio-economic data in reserve selection (NPWS 1999), although they were not included in this thesis. The likelihood of implementing a reserve network can be determined when socio-economic data that identify the relative costs of reserving each site are included in reserve selection (Faith and Walker 1996, Margules and Pressey 2000). Costs can be evaluated as the monetary value of the land, human population within a site, or as the amount of restoration that needs to take place for a site to be valuable to species. Alternative conservation scenarios that result from different constraints, such as land tenure restrictions and political agendas, can be used to identify the success of each scenario in achieving conservation targets and their associated costs of implementation. Reserve selection that does not consider socio-economic constraints within a region will likely not be implemented. 5.2.3 Climate Change Ecosystems naturally change over time and consequently the amount and type of species present in an ecosystem also changes (Holling 1986). Therefore, a reserve network identified using data from one point in time may not conserve the species that the reserve network was intended for in the future. This is especially a concern given the implications of global warming, which predicts that species will be forced to move to cooler latitudes and altitudes as temperatures increase (Hughes 2000). Although dispersal can only be predicted for most species, systematic reserve selection should consider the location of potential dispersal corridors and select sites for reservation that avoid dispersal barriers. Arid ecosystems, such as the South Okanagan, are predicted to be one of the most responsive ecosystem types to increased atmospheric CO2 and global warming (Smith et al. 2000). Therefore, the reserve networks that I have identified in this thesis need to be augmented with potential corridors for species dispersal through and out of the region. 5.3 I M P R O V E M E N T S T O S Y S T E M A T I C R E S E R V E S E L E C T I O N During the course of this thesis I have found that systematic reserve selection does not consider individual species requirements. Although systematic reserve selection has been  105  developed as a multi-species conservation tool to avoid species specific conservation plans (Lombard 1993), the selection of sites for a particular species must consider each site in context to other sites previously selected for that species. For example, a particular species' overall conservation target could be achieved in a set of sites selected by systematic reserve selection, however the habitat for this species could be spread throughout the region in each of the selected sites. The reserve network would likely be unsuccessful at conserving this species unless the species was highly mobile and could persist in isolated patches of habitat (Bedward et al. 1992, Araujo and Williams 2000). In accordance with this hypothetical example, Gaston et al. (2001) have found that the complementarity technique selects areas of ecological transition where species are at the margin of their ranges. Therefore, reserve network success should be evaluated for each individual species that the reserve network was intended for. This evaluation is necessary to determine if the sites selected by complementarity have an appropriate spatial distribution to protect viable populations of each species. I did not evaluate spatial distribution of the reserve network in the South Okanagan for each species because of the fine spatial scale of the region and the ability of many of the species included in this thesis to move freely throughout the region and beyond. Although the evaluation of species specific criteria is important in reserve networks, it may provide reliable results only at coarse scales where gaps between sites prevent species movement. Reserve design is an important component of species survival and, therefore it must be incorporated into systematic reserve selection (Bedward et al. 1992). Reserve design principles include overall connectivity of the reserve network, edge effects, size and replication (Shafer 1999, Margules and Pressey 2000). At fine spatial scales, such as the South Okanagan, the connectivity of the reserve network is critical to the persistence of species. Although the C-Plan algorithm is currently unable to aggregate sites, it is an essential component of systematic reserve selection. Other more recent versions of reserve selection software, such as SITES and W O R L D M A P , have included aggregation as an option for selecting sites (Andelmann et al. 1999, Williams 2001). The aggregation of sites should be based on patch size requirements of individual species, although this is not the method used currently in these software programs. The usefulness of systematic reserve selection techniques would increase dramatically if design principles and species specific criteria were incorporated in the algorithm.  106  5.4 C O N C L U S I O N S Many organisations are interested in identifying areas that are important for conserving biodiversity. Their often conflicting agendas result in areas that do not overlap, and lead to competition among organisations for limited funds (Margules and Pressey 2000). These organisations must work together towards a common goal i f conservation is to be successful within a region, nation or globally (Margules and Pressey 2000). Systematic reserve selection provides a promising approach for multi-species conservation given that the resources available for conservation cannot conserve everything. However, the sites identified by systematic reserve selection are only as good as the data that are used in the analyses. Because of the uncertainty in the data and in the complementarity selection process, sites identified by systematic reserve selection should not be considered as final reserves, but as priorities for further research and refinement (Nicholls and Margules 1993). Although this approach is still being developed, it is an important tool for identifying conservation areas, since it is difficult to scientifically identify a reserve network that maintains viable populations of every species (Noss and Cooperrider 1994). The sites identified in this thesis for the South Okanagan region need refinement as better data become available and connectivity is incorporated into the selection process. I used current population estimates of species within the region as conservation targets in the reserve selection analyses because these provided the most realistic regional goals for conservation. Reserve selection analyses that do not consider viability of populations and that are based on predictive models of species' habitat may provide misleading priorities for conservation. Therefore, the maps of reserve networks identified in this thesis may not identify enough habitat or the essential habitat to conserve each species within the region. However, the analyses in this thesis identify areas that need further investigation for conservation because many of the threatened species are not adequately protected in the existing reserve network. The options available for conservation in the South Okanagan are being depleted as natural areas are rapidly being converted to agricultural and urban areas (Bryan et al. 1994). This thesis is integral to the conservation of threatened species in the South Okanagan, as long as the assumptions that were necessary to perform the reserve selection analyses are valid and are acknowledged by scientific researchers and wildlife managers.  107  LITERATURE CITED Andelman, S., I. Ball, F. Davis and D. Stoms. 1999. Sites 1.0: A n Analytical Toolbox for Designing Ecoregional Conservation Portfolios. Manual prepared for The Nature Conservancy, Arlington, V A . 42pp. Ando, A., J. Camm, S. Plasky, A . Solow. 1998. Species distributions, land values and efficient conservation. Science 279: 2126-2128. Araujo, M . B . 1999. Distribution patterns of biodiversity and the design of a representative reserve network in Portugal. Diversity and Distributions 5:151-163. Araujo M . B . and P.H. Williams. 2000. Selecting areas for species persistence using occurrence data. Biological Conservation 96:331-345. Arcese, P. and A.R.E. Sinclair. 1997. The role of protected areas as ecological baselines. Journal of Wildlife Management 61:587-602. Austin, M.P. 1985. Continuum concept, ordination methods and niche theory. Annual Review of Ecology and Systematics 16:39-61. Austin, M.P., A.O. Nicholls, and C R . Margules. 1990. Measurement of the realized qualitative niche: environmental niches of five eucalyptus species. Ecological Monographs 60:161-177. Awimbo, J.A., D.A. Norton and F.B. Overmars. 1996. A n evaluation of representativeness for nature conservation, Hokitika Ecological District, New Zealand. Biological Conservation 75:177-186. Balmford, A . and K.J. Gaston. 1999. Why biodiversity surveys are good value. Nature 398:204-205. Banfield, A.W.F. 1977. The Mammals of Canada. University of Toronto Press, Toronto. Bedward, M . 1999. CODA: Conservation Options and Decision Analysis; A software package for nature conservation planning. http://members.ozemail.com.au/~mbedward/coda/coda.html. Sydney, Australia. Bedward, M . , R.L. Pressey and D.A. Keith. 1992. A new approach for selecting fully representative reserve networks: addressing efficiency, reserve design and land suitability with an iterative analysis. Biological Conservation 62:115-125. Boulinier, T., J.D. Nichols, J.R Sauer, J.E. Hines and K . H . Pollock. 1998. Estimating species richness: The importance of heterogeneity in species detectability. Ecology 79:10181028. Boyce, M . 1992. Population viability analysis. Annual Review of Ecology and Systematics 23:481-506.  108  Brooks T., A . Balmford, N . Burgess, J. Fjeldsa, L.A. Hansen, J. Moore, C. Rahbek and P. Williams. 2001. Toward a blueprint for conservation in Africa. Bioscience 51:613-624. Bryan, A.D., M.J. Sarell and J.A. Luoma. 1994. A national treasure; natural areas of the South Okanagan and Lower Similkameen valleys. Brochure prepared for the Ministry of Environment, Lands and Parks, Victoria, BC. Cabeza, M . and A . Moilanen. 2001. Design of reserve networks and the persistence of biodiversity. Trends in Ecology and Evolution 16:242-248. Camm, J.D., S. Polasky, A . Solow and B. Csuti. 1996. A note on optimal algorithms for reserve site selection. Biological Conservation 78:353-355. Caughley, G. and A . Gunn. 1996. Conservation Biology in Theory and Practice. Blackwell Science Inc. Cambridge, Massachusetts. Church, R.L., D.L. Stoms and F.W. Davis. 1996. Reserve selection as a maximal covering location problem. Biological Conservation 76:105-112. COSEWIC. 2001. Canadian Species At Risk: May 2001. Committee on the Status of Endangered Wildlife in Canada. Environment Canada. 31pp. Csutsi, B., S. Polasky, P.H. Williams, R.L. Pressey, J.D. Camm, M . Kershaw, A.R. Kiester, B Downs, R. Hamilton, M . Huso and K. Sahr. 1997. A comparison of reserve selection algorithms using data on terrestrial vertebrates in Oregon. Biological Conservation 80:83-97. Currie, D.J. 1991. Energy and large-scale patterns of animal- and plant- species richness. American Naturalist 137:27-49. da Fonesca, G.A.B., A . Balmford, C. Bibby, L. Boitani, F. Corsi, T. Brooks, C. Gascon, S. Olivieri, R.A. Mittermeier, N . Burgess, E. Dinerstein, D. Olson, L . Hannah, J. Lovett, D. Moyer, C. Rahbek, S. Stuart, P. Williams. 2000. ... following Africa's lead in setting priorities. Nature 405:393-394. Dobson, A.P., J.P. Rodriguez, W . M . Roberts and D.S. Wilcove. 1997. Geographic distribution of endangered species in the United States. Science 275:550-553. Erasmus, B.F.N., S. Freitag, K.J. Gaston, B.H. Erasmus and A.S. Van Jaarsveld. 1999. Scale and conservation planning in the real world. Proceedings of the Royal Society of London; Series B 266:315-319. Faith, D.P. and P.A. Walker. 1996. Integrating conservation and development: incorporating vulnerability into biodiversity-assessment of areas. Biodiversity and Conservation 5:417-429. Ferrier, S., R.L. Pressey and T.W. Barrett. 2000. A new predictor of the irreplaceability of areas for achieving a conservation goal, its application to real-world planning, and a research agenda for further refinement. Biological Conservation 93:303-325. 109  Ferrier, S. and G. Watson. 1997. A n evaluation of the effectiveness of environmental surrogates and modelling techniques in predicting the distribution of biological diversity. Biodiversity Group, Environment, Australia. 193pp. Finkel, E. 1998. Software helps Australia manage forest debate. Science 281:1789-1791. Flather, C.H., K.R. Wilson, D.J. Dean, and W.C. McComb. 1997. Identifying gaps in conservation networks: of indicators and uncertainty in geographic-based analyses. Ecological Applications 7:531-542. Fotheringham, A.S. 1989. Scale-independent spatial analysis, in Goodchild, M . and S. Gopal (eds) The Accuracy of Spatial Data Bases. Taylor and Francis, London. pp221-228. Franklin, I.R. 1980. Evolutionary change in small populations, in Soule, M . E . and B.A. Wilcox (eds) Conservation Biology: A n Evolutionary-Ecological Perspective. Sinauer, Sunderland, M A . ppl35-150. Freemark, K.E., H.Moore, D.Forsyth, A.R. Sinclair, T.Barrett, D.White and R.Pressey. 2000. Identifying sites for conserving biodiversity across Canadian landscapes using a complementarity approach, in Brunckhorst, D.J. and D.A. Mouat (eds) Landscape Futures: Examining Innovations and Advances in Research for Natural Resource Planning and Management Across Regional Landscapes that Might Lead us Towards a Sustainable Future. Proceedings of the 1st International Symposium on Landscape Futures. U N E S C O Institute for Bioregional Resource Management, Univ. of New England, Armidale, NSW, Australia. Freitag, S., A.O. Nicholls and A.S. van Jaarsveld. 1998. Dealing with established reserve networks and incomplete distribution data sets in conservation planning. South African Journal of Science 94:79-86. Freitag, S. and A.S. van Jaarsveld. 1998. Sensitivity of selection procedures for prioritiy conservation areas to survey extent, survey intensity and taxonomic knowledge. Proceedings of the Royal Society of London, Series B 265:1475-1482. Gaston, K.J. 1996. Biodiversity - congruence. Progress in Physical Geography. 20:105-112. Gaston, K.J., A.S.L. Rodrigues, B.J. van Rensburg, P. Koleff and S.L. Chown. 2001. Complementarity representation and zones of ecological transition. Ecology Letters 4:49. Goldstein, P.Z. 1999. Functional ecosystems and biodiversity buzzwords. Conservation Biology 13:247-255. Green, D . M . and R.W. Campbell. 1984. The Amphibians of British Columbia. Handbook No. 45. British Columbia Provincial Museum, Victoria, BC. Gregory, P.T. and R.W. Campbell. 1999. The Reptiles of British Columbia. Royal British Columbia Museum, Victoria, B C .  110  Harper, B., S. Cannings, D. Fraser and W.T. Munro. 1994. Terms of endangerment: provincial lists of species at risk, in Harding, L.E. and E. McCullum (eds) Biodiversity in British Columbia: Our Changing Environment. Environment Canada, Canadian Wildlife Service, Canada, pp 16-22. Heywood, I, S. Cornelius and S. Carver. 1998. A n Introduction to Geographical Information Systems. Prentice-Hall Inc., Upper Saddle River, NJ. 279pp. Hlady, D.A. 1990. South Okanagan Conservation Strategy (1990-19951 Policy and Planning Branch. Unpublished Report. Victoria, BC. Holling, C S . 1986. The resilience of terrestrial ecosystems: local surprise and global change. in Clark, W.C. and R.E. Munn (eds) Sustainable Development of the Biosphere. Cambridge University Press, Cambridge. pp291-317. Holm, M . (ed.). 1998. Habitat Atlas for Wildlife at Risk: South Okanagan and Lower Similkameen. Unpublished Report. B C Environment, Victoria, B C . Howard, P.C., P. Viskanic, T.R.B. Davenport, F.W. Kiyenyi, M . Baltzer, C.J. Dickinson, J.S. Lwanga, R.A. Matthews, and A. Balmford. 1998. Complementarity and the use of indicator groups for reserve selection in Uganda. Nature 394:472-475. Hughes, L . 2000. Biological consequences of global warming: is the signal already apparent? Tree 15:56-61. Hume, M . 1997. Okanagan's human sprawl competes with rare dryland life, in The Vancouver Sun: World of Wonders. ElO. August 2, 1997. Hunter Jr., M . L . and A. Hutchinson. 1994. The virtues and shortcomings of parochialism: Conserving species that are locally rare, but globally common. Conservation Biology 8:1163-1165. Kiester, A.R., J.M. Scott, B. Csuti, R.F. Noss, B. Butterfield, K . Sahr and D. White. 1996. Conservation prioritization using GAP data. Conservation Biology 10:1332-1342. Kirkpatrick, J.B. 1983. A n iterative method for establishing priorities for the selection of nature reserves: an example from Tasmania. Biological Conservation 25:127-134. Krebs, C.J. 1999. Ecological Methodology. Benjamin/Cummings, Menlo Park, C A . 620pp. Land Use Coordination Office (LUCO). 2001. Protected Areas Strategy Update (brochure). Government of British Columbia, Victoria, BC. Lawton, J.H., D.E. Bignell, B. Bolton, G.F. Bloemers, P. Eggleton, P.M. Hammond, M . Hodda, R.D. Holt, T.B. Larsen, N . A . Mawdsley, N.E. Stork, D.S. Srivastava, and A . D . Watt. 1998. Biodiversity inventories, indicator taxa and effects of habitat modification in tropical forest. Nature 391:72-76.  111  Lea, E.C., R.E. Maxwell and W.L. Harper. 1991. Biophysical Habitat Units of the South Okanagan Study Area. Unpublished report for the Ministry of Environment, Lands and Parks. Penticton, BC. Lewis, A . 2000. Created Hex.ami to generate a tessellation of hexagons. Manager, Information Coordination and Analysis, Great Barrier Reef Marine Park Authority. Townsville, Queensland. 29/2/00. Unpublished. Lombard A.T. 1993. Multi-species conservation, advanced computer architecture and GIS: where are we today? in News and Views. South African Journal of Science 89:415418. Lombard, A.T., A.O. Nicholls and P.V. August. 1995. Where should nature reserves be located in South Africa? A snake's perspective. Conservation Biology 9:363-372. Lombard, A.T., R . M . Cowling, .R.L. Pressey and P.J. Mustart. 1997. Reserve selection in a species-rich and fragmented landscape on the Agulhas Plain, South Africa. Conservation Biology 11:1101-1116. Margules, C.R., I.D. Cresswell and A.O. Nicholls. 1994. A scientific basis for establishing networks of protected areas. Systematics and Conservation Evaluation 50:327-350. Margules, C R . and A.O. Nicholls. 1987. Assessing the conservation value of remnant habitat 'islands': Mallee patches on the Western Eyre Peninsula, South Australia, in Saunders, D.A., G.W. Arnold, A . A . Burbidge and A.J.M. Hopkins (eds) Nature Conservation: The Role of Remnants of Native Vegetation. Surrey Beatty and Sons Pty Ltd. Chipping Norton, NSW, Australia. Margules, C.R., A.O. Nicholls, and M . B . Usher. 1994. Apparent species turnover, probability of extinction and the selection of nature reserves: a case study of the Ingleborough limestone pavements. Conservation Biology 8:398-409. Margules, C R . and R.L. Pressey. 2000. Systematic conservation planning. Nature 405: 243253. Margules, C R . and T.D. Redhead. 1995. Guidelines for using the BioRap methodology and tools in Series: BioRap, rapid assessment of biodiversity priority areas. CSIRO, Australia. Margules, C R . and J.L. Stein. 1989. Patterns in the distributions of species and the selection of nature reserves: an example from Eucalyptus forests in South-eastern New South Wales. Biological Conservation 50:219-238. Ministry of Environment, Lands and Parks (MELP). 1999. British Columbia Wildlife Habitat Rating Standards (V2.0). Unpublished Report. Victoria, B.C. 97pp. Myers, N . , R.A. Mittermeier, C.G. Mittermeier, G.A.B. da Fonseca and J. Kent. 2000. Biodiversity hotspots for conservation priorities. Nature 403:853-858.  112  Nagorsen, D.W. and R . M . Brigham. 1993. The Bats of British Columbia. U B C Press, Vancouver, BC. Nantel, P, A . Bouchard, L. Brouillet and S. Hay. 1998. Selection of areas for protecting rare plants with integration of land use conflicts: a case study for the west coast of Newfoundland, Canada. Biological Conservation 84:223-234. National Parks and Wildlife Service (NPWS). 1999. C-Plan: Conservation Planning Software: User Manual. C-Plan Version 2.2. New South Wales, Australia. Nicholls, A.O. 1989. How to make biological surveys go further with generalised linear models. Biological Conservation 50:51-75. Nicholls, A.O and C R . Margules. 1993. A n upgraded reserve selection algorithm. Biological Conservation 64:165-169. Noss, R.F. 1994. The Wildlands Project: Land Conservation Strategy, in Futures by Design. New Society Publishers. Gabriola Island, BC. pp. 127-165. Paczek, S. In Prep. Effects of fine scale and landscape level habitat features on sagebrush breeding birds of the South Okanagan and Similkameen Valleys, British Columbia. M.Sc. Thesis, University of British Columbia, Vancouver, BC. Pearson, D.L. and F. Cassola. 1992. World-wide Species Richness Patterns of Tiger Beetles (Coleoptera: Cicindelidae): Indicator Taxon for Biodiversity and Conservation Studies. Conservation Biology 6:376-391. Peterson, R.T. 1990. A Field Guide to Western Birds. Third edition. Houghton Mifflin, Boston. Pimm, S.L. and J.H. Lawton. 1998. Planning for biodiversity. Science 279:2068-2069. Pitelka, F.A. 1959. Numbers, breeding schedule and territory in pectoral sandpipers of Northern Alaska. Condor 61:233-264. Polasky, S., J.D. Camm, A.R. Solow, B. Csuti, D. White and R. Ding. 2000. Choosing reserve networks with incomplete species information. Biological Conservation 94:110. Prendergast, J.R., R . M . Quinn, J.H. Lawton, B.C. Eversham and D.W. Gibbons. 1993a. Rare species, the coincidence of diversity hotspots and conservation strategies. Nature ' 365:335-337. Prendergast, J.R., S.N. Wood, J.H. Lawton and B.C. Eversham. 1993b. Correcting for variation in recording effort in analyses of diversity hotspots. Biodiversity Letters 1:3953. Pressey, R.L. 1994. A d hoc reservations: forward or backward steps in developing representative reserve systems? Conservation Biology 8:662-668. 113  Pressey, R.L., C.J. Humphries, C R . Margules, R.I. Vane-Wright and P.H. Williams. 1993. Beyond Opportunism: Key principles for systematic reserve selection. Tree 8:124-128. Pressey, R.L., I.R. Johnson and P.D. Wilson. 1994. Shades of irreplaceability: towards a measure of the contribution of sites to a reservation goal. Biodiversity and Conservation 3:242-262. Pressey, R.L. and V.S. Logan. 1995. Reserve coverage and requirements in relation to partitioning and generalization of land classes: analyses for Western New South Wales. Conservation Biology 9:1506-1517. Pressey, R.L. and V.S. Logan. 1998. Size of selection units for future reserves and its influence on actual versus targeted representation of features: a case study in western New South Wales. Biological Conservation 85:305-319. Pressey, R.L. and A.O. Nicholls. 1989a. Efficiency in conservation evaluation: scoring versus iterative approaches. Biological Conservation 50:199-218. Pressey, R.L. and A.O. Nicholls. 1989b. Application of a numerical algorithm to the selection of reserves in semi-arid New South Wales. Biological Conservation 50:263278. Pressey, R.L., H.P. Possingham and J.R. Day. 1997. Effectiveness of alternative heuristic algorithms for identifying indicative minimum requirements for conservation reserves. Biological Conservation 80:207-219. Pressey, R.L., H.P. Possingham and J.R. Day. 1999. Effects of data characteristics on the results of reserve selection algorithms. Journal of Biogeography 26:179-191. Pressey, R.L. and S.L. Tully. 1994. The cost of ad hoc reservation: A case study in western New South Wales. Australian Journal of Ecology 19:375-384. Price, O., J.C.Z. Woinarski, D.L. Liddle and J. Russell-Smith. 1995. Patterns of species composition and reserve design for a fragmented estate: monsoon rainforests in the northern territory, Australia. Biological Conservation 74:9-19. Rebelo, A . G . and W.R. Siegfried. 1992. Where should nature reserves be located in the Cape Floristic Region, South Africa? Models for the spatial configuration of a reserve network aimed at maximizing the protection of floral diversity. Conservation Biology 6:243-252. Ricketts, T.H., E. Dinerstein, D . M . Olson and C. Loucks. 1999. Who's where in North America? Patterns of species richness and the utility of indicator taxa for conservation. Bioscience 49(5):369-381. Rodrigues, A.S.L., R.D. Gregory and K.J. Gaston. 2000. Robustness of reserve selection procedures under temporal species turnover. Proceedings of the Royal Society of London, Series B 267:49-55.  114  Rodrigues, A.S.L., R. Tratt, B.D. Wheeler and K.J. Gaston. 1999. The performance of existing networks of conservation areas in representing biodiversity. Proceedings of the Royal Society of London, Series B 266:1453-1460. Sauer, J.R., J.E. Hines, G. Gough, I. Thomas, and B.G. Peterjohn. 1997. The North American Breeding Bird Survey Results and Analysis. Version 96.3: www.mbr.nbs.gov/bbs. Patuxent Wildlife Research Center, Laurel, M D . Schluter, A., T. Lea, S. Cannings and P. Krannitz. 1995. Antelope-brush Ecosystems. Wildlife Branch. B C Environment, Victoria, BC. Schoener, T.W., D.A. Spiller and J.B. Losos. 2001. Predators increase the risk of catastrophic extinction of prey populations. Nature 412:183-186. Schonewald-Cox, C M . 1983. Conclusions: Guidelines to management: A beginning attempt. in Schonewald-Cox, C M . , S.M. Chambers, B. MacBryde and W.L. Thomas (eds) Genetics and Conservation: A Reference for managing Wild Animal and Plant. The Benjamin/Cummins Publishing Co. Inc., Menlo Park, CA. Scott J.M., H. Anderson, F. Davis, S. Caicoo, B. Csuti, T . C Edwards Jr., R. Noss, J. Ulliman, C. Groves, R.G. Wright. 1993. Gap analysis: a geographic approach to protection of biological diversity. Wildlife Monographs 123:1-41. Shafer, C L . 1999. National park and reserve planning to protect biological diversity: some basic elements. Landscape and Urban Planning 44:123-153. Shaffer, M . L . 1981. Minimum population sizes for species conservation. Bioscience 31:131134. Sinclair, A.R.E., D.S. Hik, O.J. Schmitz, G.G.E. Scudder, D.H. Turpin, N . C Larter. 1995. Biodiversity and the need for habitat renewal. Ecological Applications 5:579-587. Smith, S.D., T.E. Huxman, S.F. Zitzer, T.N. Charlet, D . C Housman, J.S. Colemand, L . K . Fensternaker, J.R. Seemann and R.S. Nowak. 2000. Elevated C 0 2 increases productivity and invasive species success in an arid ecosystem. Nature 408:79-82. Soule, M . E . (ed.) 1987. Viable Populations for Conservation. Cambridge Univ. Press, Cambridge. Soule, M . E . and M . A . Sanjayan. 1998. Conservation targets: do they help? Science 279:2060-2061. Soule, M . E . and D. Simberloff. 1986. What do genetics and ecology tell us about the design of nature reserves? Biological Conservation 35:19-40. Stoms, D . M . 1994. Scale dependence of species richness maps. Professional Geographer 46:346-358.  115  Strittholt, J.R. and R.E.J. Boerner. 1995. Applying biodiversity gap analysis in a regional nature reserve design for the edge of Appalachia, Ohio (U.S.A.). Conservation Biology 9:1492-1505. van Jaarsveld, A.S., S. Freitag, S.L. Chown, C. Muller, S. Koch, H. Hull, C. Bellamy, M . Kruger, S. Endrody-Youga, M.W. Mansell and C.H. Scholtz. 1998. Biodiversity assessment and conservation strategies. Science 279:2106-2108. Vane-Wright, R.L, C.J. Humphries and P. H . Williams. 1991. What to protect? - systematics and the agony of choice. Biological Conservation 55:235-254. Verner, J., M . L . Morrison and C.J. Ralph (eds). 1986. Wildlife 2000: Modeling Habitat Relationships of Terrestrial Vertebrates. University of Wisconsin Press, Madison. 470pp. Virolainen, K . M . , T. Virola, J. Suhonen, M . Kuitunen, A. Lammi, and P. Siikamaki. 1999. Selecting networks of nature reserves: methods do affect the long-term outcome. Proceedings of the Royal Society of London, Series B 266:1141-1146. Warman, L, S. Robertson, A. Haney and M . Sarell. 1998. Habitat Capability and Suitability Models for 34 Wildlife Species; Using Terrestrial Ecosystem Mapping (1:20,000) in the South Okanagan and Lower Similkameen Study Area and Forest Cover Mapping (1:20,000) in the Penticton Forest District. Unpublished report for the Ministry of Environment, Lands and Parks, Penticton, BC. Washington Department of Fish and Wildlife (WDFW). 2000. Species of Concern. Product of Wildlife Resource Data Systems. http://www.wa.gov/wdfw/wlm/diversty/soc/concern.htm. Last Updated: June 21, 2000. Unpublished Internet Report. Wessels, K.J., S. Freitag, and A.S. van Jaarsveld. 1999. The use of land facets as biodiversity surrogates during reserve selection at a local scale. Biological Conservation 89:21-38. White, D., A.J. Kimerling and W.S. Overton. 1992. Cartographic and geometric components of global sampling design for environmental monitoring. Cartography and Geographic Information Systems 19:5-22. White, D., E . M . Preston, K . E . Freemark and A.R. Kiester. 1998. A hierarchical framework for conserving biodiversity, in Klopatek, J.M. and R.H. Gardner (eds) Landscape Ecological Analysis: Issues and Applications. Springer-Verlag, New York. Williams, P.H. 1998. Key sites for conservation: area-selection methods for biodiversity in Mace, G . M , A. Balmford and J.R. Ginsberg (eds) Conservation in a Changing World. Cambridge University Press, Cambridge, U K . Williams, P. 2001. Biodiversity and W O R L D M A P . http://www.nhm.ac.uk/science/projects/worldmap/worldmap/info.htm. The Natural History Museum, Cromwell Road, London, U K .  116  Williams, P, D. Gibbons, C. Margules, A. Rebelo, C. Humphries and R. Pressey. 1996. A comparison of richness hotspots, rarity hotspots and complementary areas for conserving diversity of British birds. Conservation Biology 10:155-174. Williams, P., M . Wisz, C. Rahbek, J. Moore, L. Hansen, N . Burgess, T. Brooks and A . Balmford. 2000. Quantitative methods at the Congolese forest priority-setting workshop. WWF-US, Libreville, Gabon. Witting, L , J. Tomiuk, and V . Loeschcke. 2000. Modelling the optimal conservation of interacting species. Ecological Modelling 125:123-143.  117  Appendix I. Classification of the vegetation associations in the terrestrial ecosystem map of the South Okanagan region into fifteen plant communities and three human modified categories. 1. Antelope-brush AE Antelope-brush - evening-primrose southerly aspect, coarse sand AN Antelope-brush - needle-and-thread grass coarse-textured soils ANc Antelope-brush - needle-and-thread grass coarse-textured soils, cool aspect ANf Antelope-brush - needle-and-thread grass coarse-textured soils, fan ANw Antelope-brush - needle-and-thread grass coarse-textured soils, warm aspect 2. Big Sagebrush SN Big sagebrush - needle-and-thread grass coarse-textured soil SNf Big sagebrush - needle-and-thread grass coarse-textured soil, fan SS Big sagebrush - selaginella very shallow soil SW1 Big sagebrush - bluebunch wheatgrass mesic, lacustrine SWm Big sagebrush - bluebunch wheatgrass mesic, morainal If moderate or dense shrubs then the following ecosystem units are in the Big Sagebrush plant community. WAc Bluebunch wheatgrass - arrow-leaved balsamroot deep soil, cool aspect WAw Bluebunch wheatgrass arrow-leaved balsamroot deep soifwarm aspect WBc Bluebunch wheatgrass Sandberg's bluegrass deep soil, cool aspect WBw Bluebunch wheatgrass Sandberg's bluegrass deep soil, warm aspect WF Bluebunch wheatgrass Idaho fescue coarse-textured soil WFf Bluebunch wheatgrass Idaho fescue coarse-textured soil, fan WJ Bluebunch wheatgrass junegrass mesic WSc Bluebunch wheatgrass selaginella shallow soil, cool aspect WSw Bluebunch wheatgrass selaginella shallow soil, warm aspect 3. Vasey's Big Sagebrush VK Vasey's big sagebrush - Kentucky bluegrass grassland VKc Vasey's big sagebrush - Kentucky bluegrass grassland, cool aspect VKw Vasey's big sagebrush - Kentucky bluegrass grassland, warm aspect 4. Dry Grassland FW Idaho fescue - bluebunch wheatgrass steep, warm aspect If sparse shrubs then the following ecosystem units are in the Dry Grassland plant community. WAc Bluebunch wheatgrass - arrow-leaved balsamroot deep soil, cool aspect WAw Bluebunch wheatgrass • arrow-leaved balsamroot deep soil,warm aspect WBc Bluebunch wheatgrass • Sandberg's bluegrass deep soil, cool aspect WBw Bluebunch wheatgrass • Sandberg's bluegrass deep soil, warm aspect WF Bluebunch wheatgrass • Idaho fescue coarse-textured soil WFf Bluebunch wheatgrass - Idaho fescue coarse-textured soil, fan WJ Bluebunch wheatgrass -junegrass mesic WSc Bluebunch wheatgrass - selaginella shallow soil, cool aspect WSw Bluebunch wheatgrass - selaginella shallow soil, warm aspect  118  5. Ponderosa Pine Forest PA PAf PB PF PS PSc PSw PW PWc PWf YS  Ponderosa pine - antelope-brush coarse-textured soil Ponderosa pine - antelope-brush coarse-textured soil, fan Ponderosa pine - water birch moist fan Ponderosa pine - Idaho fescue warm aspect, deep soil Ponderosa pine - selaginella shallow soil Ponderosa pine - selaginella shallow soil, cool aspect Ponderosa pine - selaginella shallow soil, warm aspect Ponderosa pine - bluebunch wheatgrass mesic Ponderosa pine - bluebunch wheatgrass mesic, cool aspect Ponderosa pine - bluebunch wheatgrass mesic, fan Yellow pine - saskatoon fan  6. Douglas-fir Forest DAd DAs DP DPc DPf DT DTc DTs DW DYd DYs LP LPc LPf LPs PPw PPd PPs SP  Douglas-fir - heart-leaved arnica warm aspect, deep soil Douglas-fir - heart-leaved arnica warm aspect, shallow soil Douglas-fir - pinegrass mesic Douglas-fir - pinegrass mesic, cool aspect Douglas-fir - pinegrass mesic, fluvial Douglas-fir - twinflower mesic Douglas-fir - twinflower mesic, cool aspect Douglas-fir - twinflower mesic, shallow soil Douglas-fir - bluebunch wheatgrass warm aspect, deep soil Douglas-fir - yarrow warm aspect, deep soil Douglas-fir - yarrow warm aspect, shallow soil Lodgepole pine - pinegrass mesic Lodgepole pine - pinegrass mesic, cool aspect Lodgepole pine - pinegrass mesic, fluvial Lodgepole pine - pinegrass mesic, shallow soil Ponderosa pine - pinegrass warm aspect Ponderosa pine - pinegrass warm aspect, deep soil Ponderosa pine - pinegrass warm aspect, shallow soil Common snowberry - pinegrass moist  7. Hybrid Spruce LL LLc LLs SD  Lodgepole pine - arctic lupine mesic Lodgepole pine - arctic lupine mesic, cool aspect Lodgepole pine - arctic lupine mesic, shallow soil Spruce - red-osier dogwood moist gully  8. Sub-alpine Fir - Engelmann Spruce FG FGc SF SFc SFw SG FP LF LFc LFs  Subalpine fir - grouseberry mesic Subalpine fir - grouseberry mesic, cool aspect Subalpine fir - falsebox shallow soil Subalpine fir - falsebox shallow soil, cool aspect Subalpine fir - falsebox shallow soil, warm aspect Spruce - black gooseberry moist Subalpine fir - pinegrass deep soil Lodgepole pine - falsebox mesic Lodgepole pine - falsebox mesic, cool aspect Lodgepole pine - falsebox mesic, shallow soil  119  9. Riparian Forest RM BS BD CD  Western redcedar - Douglas maple riparian Paper birch - common snowberry moist, gully Water birch - red-osier dogwood swamp Black cottonwood - red-osier dogwood floodplain  10. Broad-leafed Forest ASg ASp HA OS  Trembling aspen - common snowberry moist, gully Trembling aspen - common snowberry moist, floodplain Black hawthorn copse Oregon-grape - saskatoon gully  11. Wetlands BE CB CT GB SB SE WR  Beach Summer-cypress - bentgrass meadow Common cattail marsh Gravel bar Silverweed - bulrush meadow Sedge wetland Woolly sedge - rush marsh  12. Rock Outcrops RO ROc ROw  Rock outcrop - lichen Rock outcrop - lichen, cool aspect Rock outcrop - lichen, warm aspect  13. Cliffs CLc CLw CMc CMw  Cliff Cliff Cliff Cliff  high, cool aspect high, warm aspect moderate, cool aspect moderate, warm aspect  14. Talus TAc TAw SOc SOw  Talus, cool aspect Talus, warm aspect Saskatoon - mock-orange talus, cool aspect Saskatoon - mock-orange talus, warm aspect  15. Open Water (lakes, ponds, streams) LA OW OWa PO ST  Large lake Shallow open water Shallow open water, alkaline Pond Stream (includes seasonal gravel bars)  120  16. Urban UR Urban GC Golf Courses GP Gravel Pit SL Sewage Lagoon TC Transportation Corridor MI Mines 17. Agriculture CF Cultivated Fields CO Cultivated Orchards CV Cultivated Vineyards PM Moist Pasture PD Dry Pasture FL Feed lot 18. Barren BA Barren  121  Appendix II. Vertebrate species occurrence data used to produce a final pooled occurrence database for the twenty-nine vertebrate species included in this thesis.  1. • • • •  •  Conservation Data Centre Observation Data records range from 1966 to 1998 394 records were used in the analyses precision of the records ranged from plus or minus 100 metres (S precision) to plus or minus 1000 metres (M precision) many of the records contain information for more than one occurrence (i.e. observations from different years occur in the same record), which were not associated with an individual Universal Transverse Mercator (UTM) coordinate; therefore the data from C D C records with multiple U T M ' s that could not be separated into the number of individuals per coordinate were included in the analyses as having one or more individuals per U T M coordinate. records with multiple U T M ' s that contained detailed enough comments to determine the number of individuals that were observed in each U T M were separated into individual records  2. Ministry of Environment. Lands and Parks Observation Data • 2 databases were obtained in 1997; one database was maintained in (a) Kamloops and the other in (b) Penticton • precision of the records ranged from plus or minus 100 metres (S precision) to plus or minus one kilometre (M precision). a) 29 records from the Kamloops database were used in the analyses • records range from 1997 to 1998 b) 1421 records from the Penticton database were used in the analyses • records range from 1910 to 1995 3. Centre for Applied Conservation Biology Observation Data . • 10 records from 1998 reported by Nancy Mahony  122  Appendix III. Conversion factor used to calculate the amount of habitat area in a polygon for each of the twenty-nine vertebrate species for the (a) 6-class rating scheme and (b) 4-class rating scheme of the wildlife habitat relationship models. (a) 6-Class Rating Scheme Very High High Moderate Low Very Low Nil  % Quality Compared to Best Habitat in the Province 100-76 75-51 50-26 25-6 5-1 0  Conversion Factor Used to Calculate the Amount of Habitat Area in each Polygon 0.875 0.625 0.375 0.150 0.025 0  4-Class Rating Scheme High Moderate Low Nil  % Quality Compared to Best Habitat in the Province 100-76 75-26 25-1 0  Conversion Factor Used to Calculate the Amount of Habitat Area in each Polygon 0.875 0.500 0.125 0  (b)  The conversion factor was multiplied by the habitat area located in a polygon. The converted area was summed for the polygon to calculate the total amount of area in each polygon for each species. For example, the calculations for a 10 k m polygon that is composed of 60% Habitat Type Y and 40% Habitat Type Z, where Habitat Y is rated as high quality and Habitat Z is rated as low quality in a 4-class rating scheme for Species Xare: 2  Habitat Y: 0.875 * 6 k m = 5.25 k m Habitat Z: 0.125 * 4 k m = 0.50 k m Total Habitat Area in Polygon for Species X= 5.75 km 2  2  2  2  123  2  Appenix IV. Maps of wildlife habitat relationship models that were based on habitat suitability ratings within the South Okanagan region for the twenty-nine vertebrate species included in this thesis.  124  125  126  127  128  1 2 9  130  131  133  Appendix V. Literature sources and raw data used to determine current population estimates in the South Okanagan region and density estimates in suitable habitat of the twenty-nine vertebrate species included in this thesis. 1 2 3 4 5 6 7 8 9 10 11 12  13 14 15 16 17 18 19 20 21  22 23 24  25 26 27 28 29  Svedarsky, W.D. 1992. Observations on nesting of the American Bittern in Northwest Minnesota. Prairie nat. 24:241-250. Palmer, R.S. (ed.). 1962. Handbook of North American birds. Vol. 1. Yale Univ. Press, New Haven. Bent, A.C. 1926. Life histories of North American marsh birds. US National Mus. Bull. 135:75-84. Cannings, R.A. et al. 1987. Birds of the Okanagan Valley, British Columbia. Royal British Columbia Museum. Victoria, BC. Versall, D.B. 1940. Notes on nesting habits of the American bittern. Wilson Bull. 52:207-208. Campbell, R.W. et al. 1990. Birds of British Columbia: Volume 1, Introduction, Loons through Waterfowl. Royal British Columbia Museum. Victoria, BC. Gibbs, J.P. et al. 1998. American Bittern (Botaurus lentiginosus). The Birds of North America. No. 18. Manci, K.M. and D.H. Rusch. 1988. Indices to distribution and abundance of some inconspicuous waterbirds at Horizon Marsh. J. Field Ornithol. 59:67-75. Provost, M.W. 1947. Nesting of birds in the marshes of Northwest Iowa. Am. Midi. Nat. 38:485-503. Johnson, H. and T.A. O'Neil. 2001. Wildlife-Habitat Relationships in Oregon and Washington. Oregon University Press. Dunning, J.B. Jr. 1984. Body weights of 686 species of North American birds. West. Bird Band. Assoc., Monogr. No. 1, 38pp. Dechant, J. A. et al. 1999. Effects of management practices on grassland birds: American Bittern. Northern Prairie Wildlife Research Center, Jamestown, ND. 13 pp. http://www.npwrc.usgs.gov/resource/literatr/grasbird/ambi/ambi.htm (Version 17FEB2000). Azure, D.A. 1998. Aspects of American Bittern ecology in Northwest Minnesota. M.S. thesis. Univ. of ND, Grand Forks, ND. 139pp. Brininger, W.L., Jr. 1996. The ecology of the American Bittern in Northwest Minnesota. M.S. thesis. St. Cloud State Univ., St. Cloud, MN. 70pp. Brown, M., and J. J. Dinsmore. 1986. Implications of marsh size and isolation for marsh bird management. J. of Wildl. Manage. 50:392-397. Daub, B.C. 1993. Effects of marsh area and characteristics on avian diversity and nesting success. M.S. thesis. Univ. of MI, Ann Arbor, MI. 37pp. Eaton, E.H. 1910. Birds of New York. Univ. State of New York. Albany. Blankespoor, G.W. 1980 Prairie restoration: effects of nongame birds. J. of Wildl. Manage. 44:667-672. Rotenberry, J.T. and J.A. Wiens. 1980. Habitat structure, patchiness and avian communities in North American steppe vegetation: a multivariate analysis. Ecol. 61:1228-1250. BC Government. 1997. Species and Plant Community Accounts for Identified Wildlife. Forest Practices Code. Vol. 1., Victoria, BC. Koford, R.R. 1999. Density and fledging success of grassland birds in Conservation Reserve Programfieldsin North Dakota and west-central Minnesota, ppl87-195 in Vickery, P.D. and J.R. Herkert (eds). Ecology and conservation of grassland birds of the Western Hemisphere. Stud, in Avian Biol. 19. Herkert 1994. The effects of habitat fragmentation on midwestern grassland bird communities. Ecol Appl. 4:461-471. Vandamme, L.M. 1999. Status Report on the Bobolink Dolichonyx oryzivorus in British Columbia. Ministry of Environment Lands and Parks, Victoria, B.C. Dechant, J. A. et al. 1999. Effects of management practices on grassland birds: Bobolink. Northern Prairie Wildlife Research Center, Jamestown, ND. 23pp. http://www.npwrc.usgs.gov/resource/literatr/grasbird/bobo/bobo.htm (Version 17FEB2000). Wiens, J. A. 1969. An approach to the study of ecological relationships among grassland birds. Ornith. Monogr. 8:1-93. Raim, A. J. 1975. Territory and mating system in the Bobolink, Dolichonyx oryzivorus. M.A. thesis. West. MI Univ. 90pp. Martin, S.G. 1967. Breeding biology of the Bobolink. M.S. thesis. Univ. of WI, Madison, WI. 122pp. Helzer, C. J. 1996. The effects of wet meadow fragmentation on grassland birds. M.S. thesis. Univ. of NE, Lincoln, NE. 65pp. Martin, S.G. and G.A. Thomas. 1995. Bobolink Dolichonyx oryzivorus. Birds of North America. 176:1-24.  134  30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60  Bollinger, E.K. 1988. Breeding dispersion and reproductive success of Bobolinks in an agricultural landscape. Ph.D. dissertation. Cornell Univ., Ithaca, NY. 189pp. Krannitz, P. Unpublished Data. Canadian Wildlife Service, Delta, BC. Rotenberry, J.T., and J.A. Wiens. 1989. Reproductive biology of shrubsteppe passerine birds: geographical and temporal variation in clutch size, brood size, and fledging success. Condor 91:1-14. Sarell, M.J. and K.P. McGuinness. 1996. Status of the Brewer's sparrow in British Columbia. Wildlife working report no WR-77. Ministry of Environment, Lands and Parks. Wildl. Branch. Victoria, B.C. Willing, R.M. 1970. The role of song in the behaviour and evolution of three species of spizellid sparrow: sonograph and field playback studies. Unpubl. M.Sc. Thesis, Dep. Biol. Sci., SFU, Burnaby, BC. 115pp. Harvey, D.H. 1992. The distribution, density and habitat of Brewer's Sparrows Spizella breweri in the South Okanagan Valley of British Columbia. Ministry of Environment Lands and Parks. Paige, C. et al. 1999. Wings Info Resources / Species Information and Management Abstracts: Brewer's Sparrow. The Nature Conservancy, http://www.tnc.org/wings/wingresource/brsp.html Dobler, F.C. et al. 1996. Status of Washington's shrub-steppe ecosystem: extent, ownership, and wildlife/vegetation relationships. Wash. Dept. of Fish and Wildl., Wildl. Manage. Prog., Olympia, WA. 39pp. Royal British Columbia Museum. 1996. Endangered Species Web Site. http://rbcml.rbcm.gov.bc.ca/End_Species/index_es.html Mahony, N. 1998. Unpublished Data. Centre for Applied Conservation Research, UBC, Vancouver, BC. Wiens, J.A. 1991. Ecological similarity of shrub-desert avifaunas of Australia and North America. Ecol. 72:479-495 BC Government. 1999. Managing Identified Wildlife: Procedures and Measures. Forest Practices Code. Vol. 1., Victoria, BC. Medin, D.E. 1990. Birds of a shadscale (Atriplex confertifolia) habitat in east central Nevada. Great Basin Natur. 50:295-298. Wiens, J.A. et al. 1985. Territory size variations in shrubsteppe birds. Auk 102:500-505. Petersen, K. L. et al. 1986. Growth of nestling Sage Sparrows and Brewer's Sparrows. Wilson Bull. 98:535546. Cannings, R.J. 1992. Status Report on the Canyon Wren Catherpes mexicanus in Canada. COSEWIC, Ottawa, Ont. Preston, A. 1990. Canyon Wren, Sage Thrasher, White-headed Woodpecker, Gray Flycatcher and Grasshopper Sparrow in the South Okanagan. Unpubl. Field Rep. BC MoE, Penticton. Jones, S.L. and J.S. Dieni. 1995. Canyon Wren {Catherpes mexicanus). The Birds of North America. 197. Mirsky, E.N. 1976. Ecology of co-existence in a wren-wrentit-warbler guild. Ph.D. diss. Univ. of Calif. Los Angeles. Cannings, R.J. 1995. Status of the Canyon Wren in British Columbia. Wildl. Bull. No. B-75. Ministry of Environment, Lands and Parks, Victoria, BC. Gilligan, J. et al. 1994. Birds of Oregon. Cinclus Publications, McMinnville, OR. 330 pp. Schmutz, J.K. 1984. Artificial nests for Ferruginous and Swainson's Hawks. Short Comm. J. Wildl. Manage. 48:1009-1013. Schmutz, J.K. and D.J. Hungle. 1989. Populations of Ferruginous and Swainson's Hawks increase in synchrony with ground squirrels. Can. J. of Zool. 67:2596-2601 Canadian Ferruginous Hawk Recovery Team. 1994. National Recovery Plan for the Ferruginous Hawk. Recovery of Nationally Endangered Wildlife, Environ. Can., Ottawa, Ont. Palmer, R. S. (ed). 1988. Handbook of North American birds. Vol. 5. Yale Univ. Press, New Haven. 465pp. Schmutz, J.K. et al. 1980. Coexistence of three species of hawks (Buteo spp.) in the prairie-parkland ecotone. Can. J. Zool. 58:1075-1089. Lardy, M.E. 1980. Raptor inventory and Ferruginous Hawk breeding biology in southeastern Oregon. Univ. of ID, Moscow, M.S. thesis., 53pp. Olendorff, R.R. 1973. Ecology of nesting birds of prey of Northeast Colorado. U.S. Intl. Biol. Prog., Grassland Biome Tech. Rept. 211. 233 pp. Gilmer D.S. and Stewart, R.E. 1983. Ferruginous Hawk populations and habitat use in North Dakota. J. Wildl. Manage. 47:146-157. Lokemoen, J.T. and H.F. Duebbert. 1976. Ferruginous Hawk nesting ecology and raptor populations in northern South Dakota. Condor 78:464-470. Plumpton, D.L. and D.E. Andersen. 1998. Anthropogenic effects on winter behavior of Ferruginous Hawks. J. Wildl. Manage. 62:340-346  135  61  62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80  81 82 83 84 85 86 87 88 89 90 91  Dechant, J. A. et al. 1999. Effects of management practices on grassland birds: Ferruginous Hawk. Northern Prairie Wildl. Research Center, Jamestown, ND. Jamestown, ND. http://www.npwrc.usgs.gov/resource/literatr/grasbird/ferhawk/ferhawk.htm (Version 17FEB2000). Janes, S.W. 1985. Habitat selection in raptorial birds. ppl59-188 in M.L. Cody (ed). Habitat selection in birds. Academic Press, New York, NY. Leary, A.W. 1996. Home ranges, core use areas, and dietary habits of Ferruginous hawks in southcentral Washington. M.S. Thesis, Boise State Univ., Boise, ID. 72pp. Richardson, S.A. 1995. Second Draft. Washington State Recovery Plan for the Ferruginous Hawk. Wash. Dept. of Fish and Wildl. Kirk, D.A. and C. Hyslop. 1998. Population status and recent trends in Canadian raptors: A review. Biol. Cons. 83:91-118. Schmutz, J.K. 1995. Updated status report on the Ferruginous Hawk in Canada. COSEWIC, Ottawa. Richardson, S.A. 1996. Washington state recovery plan for the Ferruginous hawk. Wash. Dept. Fish and Wildl., Olympia, Wash., 63pp. Bechard, M.J. 1995. Ferruginous Hawk. The Birds of North America, 172. Bob Altman Personal communication. Janes, S.W. 1983. Status, distribution, and habitat selection of Grasshopper Sparrow in Morrow County, Oregon. Murrelet 64:51-54. Swanson, D.A. 1996. Nesting ecology and nesting habitat requirements of Ohio's grassland-nesting birds: A literature review. Ohio Fish and Wildl. Rep. 13:3-60. Good, E.E. and C A . Dambach. 1943. Effect of land use practices on breeding bird populations in Ohio. J. of Wildl. Manage. 7:291-297. Smith, R.L. undated. Ammodramus savannarum: Grasshopper Sparrow. U.S. National Mus. Bull. 237. Plate 41. Vickery, P. D. 1996. Grasshopper Sparrow. The Birds of North America, 239. Ohio Division of Wildlife Unpublished Data Delany, M.F. et al. 1995. Territory size and movements of Florida Grasshopper sparrows. J. Field Ornithol. 66:305-307. Ducey, J. and L. Miller. 1980. Birds of an agricultural community. Nebraska Bird Rev. 48:58-68. Cannings, R.J. 1995. Status of the Grasshopper sparrow in British Columbia. Wildl. Bull. No. B-77. Ministry of Environment, Lands and Parks, Victoria, BC. Samson, F.B. 1980. Island biogeography and the conservation of nongame birds. Trans. North Am. Wildl. Nat. Resour. Conf. 45:245-251. Dechant, J. A. et al. 1998. Effects of management practices on grassland birds: Grasshopper Sparrow. Northern Prairie Wildlife Research Center, Jamestown, ND. Jamestown, ND: http://www.npwrc.usgs.gov/resource/literatr/grasbird/grasshop/grasshop.htm (Version 17FEB2000). Herkert, J.R. 1991. An ecological study of the breeding birds of grassland habitats within Illinois. Ph.D. thesis. Univ. of IL, Urbana, IL. 112pp. Butler, R.W. 1995. Great Blue Heron, Ardea herodias. The Birds of North America, 25. Bayer, R.D. 1978. Aspects of an Oregon estuarine Great Blue Heron population. pp213-218 in Sprunt IV, A., J.C. Odgen and S. Winkler (eds) Wading Birds. Natl. Audubon Soc. Res. Rep. No. 7, New York. Butler, R.W. 1992. A review of the biology and conservation of the Great Blue Heron (Ardea herodias) in British Columbia. Can. Wildl. Serv. Tech. Rep. No. 154, Delta, BC. Bayer, R.D. and E. McMahon. 1981. Colony sizes and hatching synchrony of great blue herons in coastal Oregon. Murrelet 62:73-79. Gibbs, J.P. and L.K. Kinkel. 1997. Determinants of the size and location of Great Blue Heron colonies. Colon. Waterbirds. 20:1-7. Gibbs, J.P. 1991. Spatial relationships between nesting colonies and foraging areas of Great Blue Herons. Auk 108:764-770. Martin, J. W. and J.R. Parrish. 2000. Lark Sparrow (Chondestes grammacus). The Birds of North America, 488. Rotenberry, J.T. and J.A. Wiens. 1980. Temporal variation in habitat structure and shrubsteppe bird dynamics. Oecologia 47:1-9. Walcheck, K.C. 1970. Nesting bird ecology of four plant communities in the Missouri River breaks, Montana. Wilson Bull. 82:370-382. Stewart, R.E. and H.A. Kantrud. 1972. Population estimates of breeding birds in North Dakota. Auk. 89:766788.  136  92 93 94 95 96 97 98 99 100  101 102 103 104 105 106 107 108 109  110 111  112 113 114 115 116 117 118 119 120  Fitch, H.S. 1958. Home ranges, territories and seasonal movements of vertebrates of the natural history reservation. Univ. Kansas Mus. Nat. Hist. Misc. Publ. 11:63-326. Dunning, J.B. Jr. 1992. CRC handbook of avian body masses. CRC press, Boca Raton, FL. Byers, C.J. and U. Olsson. 1995. Sparrows and Buntings: a guide to the sparrows and buntings of North America and the World. Houghton Mifflin, Boston. Tobalske, B.W. 1997. Lewis' Woodpecker. (Melanerpes lewis). The birds of North America, 284. Thomas, J.W. et al. 1979. Snags pp60-77 in Thomas, J.W. (ed). Wildlife Habitat in Managed Forests: The Blue Mountains of Oregon and Washington. US Dept. Agric, For. Serv. Agric. Handbook 553. Siddle, C. and G. Davidson. 1991. Status of the Lewis' Woodpecker {Melanerpes lewis) in British Columbia. Second Draft. Ministry of Environment Lands and Parks. Victoria, B.C. Anonymous. 1995. Draft. Species Notes (Latest revision). Lewis' Woodpecker {Melanerpes lewis). BC Environment, Victoria, B.C. Sousa, P.J. 1983. Habitat Suitability Index Models: Lewis' Woodpecker. Fish and Wildl. Serv., U.S. Dept. of the Int., Washington, D.C. Smith, M.R. et al. 1997. Breeding birds of Washington State: Location data and predicted distributions. Vol. 4 in Cassidy, K.M., C.E. Grue, M.R. Smith, and K.M. Dvornich (eds). Washington State Gap Analysis - Final Rep. Seattle Audubon Soc. Publ. in Zool. No. 1, Seattle, 538pp. Tobalske, B.W. 1996. Scaling and muscle composition, wing morphology and intermittent flight behaviour in woodpeckers. Auk. 113:151-177. Sadler, D.A.R. and W.J. Maher. 1976. Notes on the Long-billed curlew in Saskatchewan. Auk 93:382-384. Allen, J.N. 1980. The ecology and behaviour of the Long-billed curlew in southeast Washington. The Wildlife Society, Wildl. Monogr. No. 73. 65pp. Ohanjanian, I. A. 1992. Numbers, distribution and habitat dynamics of Long-billed Curlews in the East Kootenay. Ministry of Environment, Cranbrook, B.C. Pampush, G.J. and R.G. Anthony. 1993. Nest success, habitat utilization and nest-site selection of Long-billed Curlews in the Columbia Basin, Oregon. Condor 95:957-967. Chapman, K. 1995. Draft. Species Notes (Latest revision). Long-billed Curlew {Numenius americanus). BC Environment, Victoria, B.C. DeSmet, K.D. 1992. Status report on the Long-billed curlew {Numenius americanus) in Canada. COSEWIC, Ottawa. Jenni, D.A. et al. 1982. Behaviour ecology and habitat relationships of Long-billed curlews in western Idaho. Bureau of Land Management. Boise, ID. 234pp. Redmond, R.L. et al. 1981. An evaluation of breeding season census techniques for Long-billed Curlews {Numenius americanus) ppl97-201 in Ralph, C.J. and J.M. Scott (eds). Estimating Numbers of Terrestrial Birds. Stud, in Avian Biol. No. 6. Paton, P.W.C. and J. Dalton. 1994. Breeding ecology of Long-billed Curlews at Great Salt Lake, Utah. Great Basin Natur. 54:79-85. Redmond, R.L. 1989. Long-billed curlew, Numenius americanus. pp75-76 in Clark, T.W., A.H. Harvey, R.D. Dorn, D.L. Genter and C. Groves (eds). Rare, Sensitive, and Threatened Species of the Greater Yellowstone Ecosystem. Hooper, T.D. 1997. Status of the Prairie Falcon in the Chilcotin-Caribou Region, British Columbia. Wildlife Branch, Ministry of Environment Lands and Parks, Victoria, B.C. Webster, H., Jr. 1944. A survey of the Prairie Falcon in Colorado. Auk 61:609-616. Hickman, G.L. 1971. Prairie Falcons and Red-tailed Hawks rearing young in inactive Golden Eagle nests. Condor 73:490. Johnsgard, P.A. 1990. Hawks, eagles and falcons of North America. Smithsonian Inst. Press. Washington, DC. 403pp. Steenhof, K. et al. 1999. Long-term Prairie Falcon population changes in relation to prey abundance, weather, land uses and habitat conditions. Condor. 101:28-41. Ogden, V.T. 1973. Nesting density and reproductive success of prairie falcon in southwest Idaho. Univ. of Idaho, Moscow, M.S. thesis. Marzluff, J.M. et al. 1997. The influence of habitat, prey abundance, sex, and breeding success on the ranging behavior of Prairie Falcons. Condor 99:567-584. Beauvais, G. et al. 1992. Home range habitat use and behavior of prairie falcons wintering in east-central Colorado. J. of Raptor Res. 26:13-18. Kirk, D.A. and U. Banasch. 1996. Updated status report on the Prairie Falcon {Falco mexicanus) in Canada. COSEWIC, Ottawa, Ont.  137  121 Woodsworth, G and K. Freemark. 1981. Status report on the Prairie Falcon (Falco mexicanus). COSEWIC, Ottawa. 122 Schoener, T.W. 1968. Sizes of feeding territories among birds. Ecol. 49:123-141. 123 Craighead, J.J. and F.C. Craighead. 1956. Hawks, owls, and wildlife. Stackpole. Harrisburg. 443pp. 124 Newton, I. 1979. Population ecology of raptors. Buteo Books, Vermillion, S.D. 125 Squires, J.R. et al. 1993. Home range size and habitat-use patterns of nesting prairie falcons near oil developments in northeastern Wyoming. J. of Field Ornithol. 64:1-10. 126 Steenhof, K. 1998. Prairie falcon, Falco mexicanus. The Birds of North America, 346. 127 Paige, C, et al. 1999. Wings Info Resources / Species Information and Management Abstracts: Sage Thrasher. The Nature Conservancy, http://www.tnc.org/wings/wingresource/sath2.htm. 128 Cannings, R.J. 1992. A Management Plan for the Sage Thrasher Oreoscoptes montanus in British Columbia. Ministry of Environment Lands and Parks. 129 Millikin, R.L. 1996. Update on the Sage Thrasher recovery project. Canadian Wildlife Service Report. 130 Blood, D.A. 1995. Sage Thrasher: Wildlife at Risk in British Columbia. Brochure. Ministry of Environment Lands and Parks, Victoria, B.C. 132 Cannings, R.J. 1992. Status report on the Sage Thrasher {Oreoscoptes montanus) in Canada. COSEWIC, Ottawa. 133 Reynolds, T.D. 1981. Nesting of the Sage Thrasher, Sage Sparrow, and Brewer's Sparrow in southeastern Idaho. Condor 83:61-64. 134 Melvin, S.M. et al. 1990. Population estimates nesting biology and habitat preferences of interlake Manitoba Canada Sandhill Cranes Grus canadensis. Can. Field Natur. 104:354-361. 135 Walkinshaw, L.H. 1965. Sandhill Crane studies on Banks Island, NWT. Blue Jay 23:66-72. 136 Carlisle, M.J. 1979. Distribution and habitat of Sandhill cranes in Athabina region of central Alberta. M.Sc. Thesis, North Dakota State Univ., Fargo, ND, 54pp. 137 Boise, C M . 1977. Breeding biology of the Lesser Sandhill Crane on Yukon-Kuskokwim Delta, Alaska. M.Sc. Thesis, Univ. of Alaska, Fairbanks, Alaska. 79pp. 138 Campbell, R.W. et al. 1988. Wildlife habitat handbook for the Southern Interior Ecoprovince - Vol. 2: Species notes for selected birds. Harcombe, A.P. (ed.). Ministry of Environment, Ministry of Forests, and Wildlife Habitat Canada. Wildlife habitat research WHR-29, Wildlife Report No. R-16. 139 Tacha, T.C. et al. 1992. Sandhill Crane. The Birds of North America, 31. 140 Stern, M.A. et al. 1987. Nesting ecology and productivity of Greater Sandhill Cranes at Sycan Marsh, Oregon. Proc. 1985 Crane Workshop; pp249-256. WR 206. 141 Bennett, A.J. 1989. Movements and home ranges of Florida Sandhill Cranes. J. Wildl. Manage. 53:830-836. 142 Cooper, J.M. 1996. Status of the Sandhill Crane in British Columbia. Wildl. Bull. No. B-83. Ministry of Environment, Lands and Parks. Wildlife Branch. Victoria, B.C. 143 Nesbitt, S.A. and K.S. Williams. 1990. Home range and habitat use of Florida Sandhill Cranes. J. Wildl. Manage. 54:92-96. 144 Safina, C. 1993. Population trends, habitat utilization, and outlook for the future of the Sandhill Crane in North America: A review and synthesis. Bird Pop. 1:1-27. 145 Lewis, J.C. 1977. Sandhill Cranes (Grus canadensis). pp5-43 in Anderson, G . C (ed). Management of migratory shore and upland gamebirds in North America. Int. Assoc. Fish and Game Agencies, Wash., DC. 146 Riley, J.L. 1982. Habitats of Sandhill Crianes in the southern Hudson Bay lowland, Ontario. Can. Field Nat. 96:51-55. 147 Cadman, M.D. and A.M. Page. 1994. Status report on the Short-eared Owl (Asioflammeus)in Canada. COSEWIC. Ottawa. 148 Beske, A. and J. Champion. 1971. Prolific nesting of Short-eared Owls on Buena Vista Marsh. Passeng. Pig. 33:99-103. 149 Holt, D.W. and S.M. Leasure. 1993. Short-eared Owl. The Birds of North America, 62. 150 Holt, D.W. 1992. Notes on Short-eared Owl, Asioflammeus,nest sites, reproduction and territory sizes in coastal Massachusetts. Can. Field Nat. 106:352-356. 151 Clark, R.J. 1975. A Field Study of the Short-eared Owl, Asioflammeus(Pontoppidan), in North America. Wildl. Monogr. 47:1-67. 152 Gronland, S. and H. Mikkola. 1969. Suopollon ekologiasta lapuan Alajoella. Suomenselan Linnut 4:68-76. 153 Holt, D.W. and S.M. Melvin. 1986. Population dynamics, habitat use, and management needs of the Shorteared Owl in Massachusetts. Summ. of 1985 research. MA Div. Fish, and Wildl., Nat. Heritage Prog. 57pp. 154 Lockie, J.D. 1955. The breeding habits and food of Short-eared Owls after a vole plague. Bird Study 2:53-69.  138  155 Village, A. 1987. Numbers, territory size and turnover of Short-eared Owls in relation to vole abundance, Ornis Scandinavica 18:198-204. 156 Pitelka, F.A et al. 1955. Ecological relations of jaegers and owls as lemming predators near Barrow, Alaska. Ecol. Monogr. 25:85-117. 158 Rodriguez, E.R. and B.T. Brown. 1990. Density and habitat use of raptors along the Rio Bavispe and Rio Yaqui Sonora, Mexico. J. Raptor Res. 24:47-51. 159 Kirk, D.A. and M.J. Mossman. 1998. Turkey Vulture {Cathartes aura). The birds of North America, 339. 160 Stewart, E.R. and C S . Robbins. 1958. Birds of Maryland and the district of Columbia. N. Am. Fauna 62. 161 Coleman, J.S. 1985. Home range, habitat use, behaviour and morphology of the Gettysburg vultures. M.Sc. Thesis. Virginia Polytechic Inst, and State Univ., Blacksburg, VA. 162 Coles, V. 1944. Nesting of the Turkey Vulture in Ohio caves. Auk 61:219-227. 163 Coleman, J. S. and J.D. Fraser. 1989. Habitat use and home ranges of Black and Turkey Vultures. J. Wildl. Manage. 53:782-792. 164 Tenney, M.K. 1986. A life history study of Turkey Vultures along the Root River Valley in Fillmore County, Minnesota, M.S. thesis, St. Cloud, Univ., St. Cloud, MN. 165 Santana, E.C. et al. 1986. Home range and activity patterns of Turkey Vultures in Puerto Rico. Carib. J. Sci. 22:175-177. 166 Kirk, D.A. and J.E.P Currall. 1994. Habitat associations of migrant and resident vultures in central Venezuela. J. Avian Biol. 25:327-337. 167 Campbell, R.W. et al. 1990. Birds of British Columbia: Vol. 2, Nonpasserines, diurnal birds of prey through woodpeckers. Royal British Columbia Museum. 168 Penak, B.L. 1986. Status report on the Eastern screech owl (Otus asio) in Canada. COSEWIC Ottawa, Ont. 169 Allaire, P.N. and D.F. Landrum. 1975. Summer census of screech owls in Breathitt County. Kentucky Warbl. 51:23-29. 170 Cink, C L . 1975. Population densities of screech owls in northeastern Kansas. Kansas Ornith. Soc. Bull. 25:1316. 171 Nowicki, T. 1974. A census of screech owls {Otus asio) using tape recorded calls. Jack Pine Warbl. 52:98-101. 172 Gehlbach, F.R. 1995. Eastern screech owl {Otus asio). The Birds of North America, 165. 173 DeGeus, D.W. and J.B. Bowles. 1991. Relative abundance of Eastern Screech Owl in Southcentral Iowa township. J. Iowa Acad. Sci. 98:91-92. 174 Lynch, P.J. and D.G. Smith. 1984. Census of Eastern Screech Owls {Otus asio) in urban open space areas using tape recorded song. Amer. Birds. 38:388-391. 175 Kirk, D.A. 1995. Status Report on the Western Screech-Owl {Otus kennicottii) in Canada. COSEWIC, Ottawa, Ont. 176 Johnson-Duncan, E.E. et al. 1988. Small nesting raptors as indicators of change in the southwest desert. pp232-236 in Natl. Wildl. Fed. Sci. Tech. Ser. No. 111. 177 Gehlbach, F.R. 1994. The eastern screech owl life history, ecology, and behaviour in suburbia and the country side. Texas. A. and M. Univ. Press College Station. 178 BC Environ. 1996. Managing Identified Wildlife Guidebook 1.0, Kamloops Forest Region. Ministry of Environment, Lands and Parks and Ministry of Forests. Internal Government Review Draft. 179 Hayward, G.D. 1983. Resource partitioning among six forest owls in the River of No Return Wilderness, Idaho. M.Sc. Thesis. Univ. of Idaho, Moscow. 180 Smith, D.G. and R. Gilbert. 1984. Eastern screech owl home range and use of suburban habitats in southern Connecticut. J. Field Ornithol. 55:322-329. 181 Anderson, R. 1996. Western Screech Owl, Otus kennicottii. Unpubl. Rep Prep, for the Int. Columbia Basin Ecosyst. Manag. Proj. 182 Garrett, K. L. et al. 1996. White-headed Woodpecker {Picoides albolarvatus). The Birds of North America, 252. 183 Mannan, R.W. and E.C. Meslow. 1984. Bird populations and vegetation characteristics in managed and old growth forests, Northeastern Oregon. J. Wildl. Manage. 48:1219-1238. 184 Dixon, R.D. 1995. Density, nest site and roost site characteristics, home range, habitat use and behaviour of White-headed Woodpeckers: Deshutes and Winema National Forest, OR. OR Dept. of Fish and Wildl. Nongame Rep. 93-3-01. 185 Servheen, G. et al. undated. A Species Conservation Assessment and Strategy for White-headed Woodpecker {Picoides albolarvatus). Idaho State Conservation Effort. 186 Dixon, R.D. 1995. Ecology of White-headed Woodpeckers in the central Oregon Cascades. M.Sc. Univ of Idaho, Moscow.  139  187 Raphael, M.G. and M. White. 1978. Snags, wildlife and forest management in the Sierra Nevada. Cal-Neva Wildl. 1978:23-41. 188 Turner, J.G. et al. 1993. Resident bird counts: old growth mixed coniferous redfirforest transition. J. Field Ornithol. 64:60. 189 Cannings, R.J. 1995. Status of the White-headed woodpecker (Picoides albolarvatus) in British Columbia. Wildl. Bull. No. B-80. 190 Bull, E.L. 1977. Specialized habitat requirements of birds: snag management, old growth and riparian habitat. pp74-82 in DeGraaf, R.M. (tech coord), Proceedings of the workshop on nongame bird habitat management in the coniferous forests of the Western United States. US Forest Serv. Gen. Tech. Rep. PNW-64 191 Kilgore, B.M. 1971. Response of breeding bird populations to habitat changes in giant sequoia forest. Am. Midi. Nat. 85:135-152. 192 Raphael, M.G. and M. White. 1984. Use of snags by cavity nesting birds of the Sierra Nevada. Wildl. Monogr. 86:1-66. 193 Cannings, R.J. 1992. Status report on the White-headed woodpecker {Picoides albolarvatus) in Canada. COSEWIC, Ottawa. 194 Milne, K. and S. Hejl. 1989. Nest site characteristics of White-headed Woodpeckers. J. Wildl. Manage. 53:50-55. 195 Carter, S. 1996. White-headed Woodpecker, Picoides albolarvatus. Unpubl. Rep. Prep, for the Int. Columbia Basin Ecosyst. Manage. Proj. 196 Thompson, C F . and V. Nolan, Jr. 1973. Population biology of the Yellow-breasted Chat {Icteria virens I.) in southern Indiana. Ecolog. Monogr. 43:145-171. 197 Cannings, R.J. 1995. Status of the Yellow breasted Chat in British Columbia. Wildl. Bull. No. B-81. BC Environment, Victoria, B.C. 198 Dennis, J.V. 1958. Some aspects of the breeding ecology of the Yellow-breasted Chat {Icteria virens). Bird Band. 29:169-183. 199 Brewer, R. 1955. Size of home range in eight bird species in southern Illinois swamp thicket. Wilson Bull. 67:140-141. 200 Cadman, M.D. and A.M. Page. 1994. Status report on the Yellow breasted chat (Icteria virens) in Canada. COSEWIC, Ottawa. 201 Gibbard and Gibbard Environ. Consultants. 1992. Yellow breasted Chat survey. Unpubl. Rep. BC Min. of Environ., Lands and Parks. Penticton, B.C. 202 Burhans, D.E. and F.R. Thompson, III. 1999. Habitat patch size and nesting success of Yellow-breasted Chats. Wilson Bull. 111:210-215. 203 Van Dyke, W.A. et al. 1983. Wildlife habitat in managed rangelands. The Great Basin of Southeastern Oregon: Bighorn Sheep. USDA Forest Serv. Gen. Tech. Rep. PNW 159 PNW Forest and Range. Experimental Station, Portland OR. 37pp. 204 Van Dyke, W. 1978. MS thesis. OR St. Univ. 205 Woodgerd, W. 1964. Population dynamics of bighorn sheep on Wild horse Island. J. Wildl. Manage. 28:381391. 206 Demarchi, R.A. 1965. An ecological study of the Ashnola bighorn winter ranges. Unpubl. Master's Thesis. UBC. Vancouver, BC. 103pp. 207 Smith T.S. et al. 1991. A habitat evaluation procedure for rocky mountain bighorn sheep in the intermountain west. Great Basin Natural. 51:205-225. 208 Coggins, V.L. 1980. Present status of Rocky Mountain Bighorn sheep in northeast Oregon. Biennial Symp. of the Northern Wild Sheep and Goat Counc. 2:90-104. 209 Blood, D.A. 1963. Some aspects of behaviour of a bighorn herd. Can. Field Nat. 77: 77-94. 210 Krausman P.R. and R.C. Etchberger. 1995. Response of desert ungulates to water project in Arizona. J. Wildl. Manage. 59:292-300. 211 McQuivey, R.P. 1978. The desert bighorn of Nevada. Biol. Bull. No. 6. NV Dept. of Wildlife, Reno 81pp. 212 Wehausen, J.D. 1983. Sierra Nevada bighorn sheep: history and population ecology. National Park Res. Studies Unit, U. of Calif, Davis. 243pp. 213 Berger, J. 1990. Persistence of different-sized populations: an empirical assessment of rapid extinctions in bighorn sheep. Cons. Biol. 1:91-98. 214 Holl, S.A. 1982. Evaluation of bighorn sheep habitat. Desert Bighorn Sheep Council Trans. 26:47-49. 215 Geist, V. 1975. On the management of mountain sheep: theoretical considerations. pp77-98 in Trefethen, J.B. (ed). The Wild Sheep of Modern North Amer. Boone and Crockett Club, Alexandria, Virginia. 302pp.  140  216 Trefethan, J.B. 1975. The wild sheep in modern North America : proceedings of the Workshop on the Management Biology of North American Wild Sheep, Univ. of Montana, Missoula, June 18-20, 1974, Nat. Audubon Soc, Wildl. Manage. Inst. 217 Demarchi, R.A. et al. 2000. Status of California Bighorn Sheep in British Columbia. Ministry of Environment, Lands and Parks, Wildlife Branch, 2000. Wildlife-Bulletin B-98. Victoria, BC. 218 Geist, V. 1971. Mountain sheep: behavior and evolution. Univ. Chicago Press, Chicago, IL. 219 Kopec. 1982. Proceedings of the biennial symposium of the Northern Wild Sheep and Goat Council, 3. 220 Chapman, T.E.A. 1999. South Okanagan Valley California Bighorn Sheep Habitat Inventory. Unpubl. Final Rep., BC Environment, Penticton, BC. 221 Verts, B.J. and L.N. Carraway. 1998. Land Mammals of Oregon. Univ. of Calif. Press, Berkeley, CA, 668pp. 222 Johnson, R.L. 1983. Mountain goats and mountain sheep of Washington. Wash. Dept. of Game. 204pp. 223 Lewis, S. 1996. Low roost site fidelity in Pallid bats: associated factors and effect on group stability. Behav. and Sociobiol. 39:335-344. 224 Bell, G.P. 1982. Behaviour and ecological aspects of gleaning by the desert insectivorous bat Antrozous pallidas (Chiroptera: Vespertilionidae). Behav. Ecol. and Sociobiol. 10:217-223. 225 Robinson, M.F. and R.E. Stebbings. 1997. Home range and habitat use by the serotine bat, Eptesicus serotinus, in England. J. Zool. (London) 243:117-136. 226 Jones K E. et al. 1996. Distribution and population densities of seven species of bat in northern England. J. Zool. (London) 240:788-798. 227 O'Donnell, C F . et al. 1999. Rediscovery of short-tailed bats (Mystacina sp.) in Fiordland, New Zealand: Preliminary observations of taxonomy, echolocation calls, population size, home range, and habitat use. NZ J. Ecol. 23:21-30. 228 Chapman, K. et al. 1994. Status of the Pallid bat in British Columbia. Ministry of Environment, Lands and Parks. Wildl. Br. Victoria, BC, WR-61. 229 Dalquest, W.W. 1947. Notes on the natural history of the bat, Corynorhinus rafinesquii in California. J. Mamm. 28:17-30. 230 Alcorn, J.R. 1944. Notes on the winter occurrence of bats in Nevada. J. Mamm. 25:308-310. 231 Lewis, S.E. 1990. Preliminary observations of pallid bat maternity roost. Bat Res. News. 34:4-86. 232 Beck, A.J. and R.L. Rudd. 1960. Nursery conlonies in the pallid bat. J. Mammal. 41:266-267. 233 Lewis, S.E. 1994. Night roosting ecology of pallid bats {Antrozous pallidus) in Oregon. Am. Midi. Nat. 132:219-226. 234 Hermanson, J.W. and T.J. O'Shea. 1983. Antrozous pallidus. Mamm. Species 213:1-8. 235 Vaughan, T.A. and T.J. O'Shea. 1976. Roosting ecology of the pallid bat, Antrozous pallidus. J. Mammal. 57:19-42. 236 Balcombe, J.P. 1988. Status report on the Pallid Bat (Antrozouspallidus) in Canada. COSEWIC, Ottawa. 237 Nagorsen, D.W. and R.M. Brigham. 1993. Bats of British Columbia. Vol. 1. The Mammals of British Columbia. Royal BC Museum Handbook, UBC Press, Vancouver, BC. 238 Davis, R. 1969. Wing loading in pallid bats. J. Mamm. 50:140-144. 239 Grindal, S.C. et al. 1991. Evidence for a breeding population of pallid bats, Antrozous pallidus. (Chiroptera: Vespertilionidae), in British Columbia. Contrib. Nat. Sci. RBC Mus. 14:1-4. 240 Leonard, M.L. and M.B. Fenton. 1983. Habitat use by spotted bats {Euderma maculatum) (Chiroptera: Vespertilionidae): roosting and foraging behaviour. Can. J. Zool. 61:1487-1491. 241 Woodsworth, G . C et al. 1981. Observations of the echolocation, feeding behaviour and habitat use of Euderma maculatum (Chiroptera: Vespertilionidae) in southcentral BC. Can. J. of Zool. 59: 1099-1102. 242 Garcia, P.F.J, et al. 1995. Status of the Spotted Bat in British Columbia. Ministry of Environment, Lands and Parks. Wildl. Br, Victoria, BC. WR-75. 243 Balcombe, J.P. 1988. Status report on the Spotted Bat {Euderma maculatum) in Canada. COSEWIC, Ottawa. 244 Priday, J. and B. Luce. 1999. New distributional records for Spotted Bat {Euderma maculatum) in Wyoming. Great Basin Natural. 59:97-101. 245 Fenton, M.B. et al. 1987. The status of Spotted Bats {Euderma maculatum) in the Okanagan Valley, BC. Report for World Wildlife Canada and BC Nature Trust. 246 Collard, T.S. et al. 1990. Identification of the status and critical habitats of the Spotted Bat (Euderma maculatum), Pallid Bat {Antrozous pallidus) and Fringed Bat {Myotis thysanodes) in the South Okanagan Valley, BC. Unpubl. Rep. World Wildl. Fund. BC Env. and BC Hab. Conserv. Fund. 26pp. 247 Holroyd, S.L. et al. 1994. A survey of the bat fauna of the dry interior of British Columbia. Ministry of Environment, Lands and Parks, Wildl. Br. Victoria, BC, WR-63.  141  248 Perry, T.W. et al. 1997. New locality for Euderma maculatum (Chiroptera: vespertilionidae) in New Mexico. Southwest. Nat. 42:99-101. 249 Storz, J.F. 1995. Local distribution and foraging behaviour of the Spotted bat {Euderma maculatum) in northwestern Colorado and adjacent Utah. Great Basin Nat. 55:78-83. 250 Wai-Ping, V. and M.B. Fenton. 1989. Ecology of Spotted Bat {Euderma maculatum): Roosting and Foraging Behaviour. J. Mamm. 70:617-622. 251 Roberts, G. and A. Roberts. 1992. Grassland Biodiversity in the Chilcotin Bunchgrass subzones and adjacent Interior Douglas-fir subzone. BC Environment. Williams Lake, B.C. 252 Sarell, M.J. and K.P. McGuiness. 1993. Rare bats of the shrub-steppe ecosystem of Eastern Washington. Wash. Dept. of Wildl. Nongame Prog. 253 Speakman, J.R. et al. 1991. Minimum summer populations and densities of bats in Northeast Scotland near the northern borders of their distributions. J. Zool. (London) 225:327-345. 254 Barbour R. W. and W. H. Davis. 1969. Bats of America. Univ. of Kentucky, Lexington, KY. 286 pp. 255 Szewczak, J.M. et al. 1998. Bats of the White and Inyo Mountains of California-Nevada. Great Basin Nat. 58:66-75. 256 Dobkin D.S. et al. 1995. Springtime movements, roost use and foraging activity of Townsend's big-eared bat (Plecotus townsendii) in central Oregon. Great Basin Nat. 55:315-321. 257 Adam, M.D. et al. 1994. Foraging areas and habitat use of the Virginia Big-eared Bat in Kentucky. J. Wildl. Manage. 58:462-469. 258 Idaho State Conservation Effort. 1995. Habitat conservation assessment and conservation strategy for the Townsend's Big-eared bat. Draft unpubl. rep. no. 1. Boise, Id. 259 Parker, W.S. and M.V. Plummer. 1987. Population Ecology in Seigel, R.A., J.T. Collins and S.S. Novak (eds) Snakes: Ecology and Evolutionary Biology. MacMillan Publishing Company, NY. pp253-301. 260 Fitch, H.S. 1949. Study of snake populations in central California. Am. Midi. Nat. 41:513-579. 261 Nelson, K.J. and P.T. Gregory. 1992. A Survey of the Distribution, Biology and Population Trends of the Great Basin Gopher Snake, Pituophis melanoleucus deserticola, in British Columbia. Ministry of Envir. Lands and Parks. Victoria, B.C. 262 Turner, F.B. 1977. The dynamics of populations of squamates and crocodilians, in Gans, C.and D.W. Tinkle (eds). Biology of the Reptilia. vol. 7 pp 157-264, Academic Press, NY. 263 Parker, W.S. 1974. Comparative ecology of two colubrid snakes, Masticophis t. taeniatus (Hallowell) and Pituophis melanoleucus deserticola (Stejneger) in northern Utah. Thesis, Univ. of Utah, Salt Lake City. 264 Fitch, H.S. 1977. Resources of a snake community in prairie-woodland habitat of northeastern Kansas in Scott, N.J., Jr. (ed) Herpetological Communities: A Symposium of the Society for the Study of Amphibians and Reptiles and the Herpetologists' League. 265 Shewchuk, C.H. and H.L. Waye. 1995. Draft. Status Report for the Gopher Snake in British Columbia. Ministry of Envir. Lands and Parks. Victoria, B.C. 266 Parker, W.S. and W.S. Brown. 1980. Comparative ecology of two colubrid snakes, Masticophis t. taeniatus and Pituophis melanoleucus deserticola, in northern Utah. Milw. Public Mus. Publ. Biol. Geol. 7 104pp. WR 194 . 267 Diller, L.V. and R.L. Wallace. 1996. Comparative ecology of two snake species (Crotalus viridis and Pituophis melanoleucus) in southwestern Idaho. Herpetologica 52:343-360. 268 Diller, L.V. and R.L. Wallace. 1986. Aspects of the life history and ecology of the desert night snake, Hypsiglena torquata deserticola: Colubridae in southwestern Idaho. Southwest. Nat. 31:55-61. 269 Sarell, M. Personal Communication. 270 Lacey, H. et al. 1996. The occurrence of the Night Snake, Hypsiglena torquata, in British Columbia, with comments on its body size and diet. Can. Field-Nat. 110:620-625. 271 Rodriguez-Robles J.A. et al. 1999. Feeding ecology of the desert nightsnake, Hypsiglena torquata (Colubridae). Copeia. 1999:93-100. 272 Storm, R.M. and W.P. Leonard. 1995. Reptiles of Washington and Oregon. Seattle Audubon Society, Seattle, WA, 176 pp. 273 Brown, W.S. and W.S. Parker. 1984. Growth, reproduction and demography of the racer, Coluber constrictor mormon, in northern Utah, in Seigel, R.A., L.E. Hunt, J.L. Knight, L. Malaret and N.. Zuschlag, (eds). Vertbrate Ecology and Systematics: A Tribute to Henry S. Fitch. Univ. Kans. Mus. Nat. Hist. Spec. Publ. 10:13-40. 274 Hirth, H.F. et al. 1969. Dispersal of snakes from a hibernaculum in northwestern Utah. Ecol. 50:332-339. 275 Hirth, H.F. and A.C. King. 1968. Biomass densities of snakes in the cold desert of Utah. Herpetologica 24: 333-335.  142  276 Fitch, H.S. 1963. Natural history of the racer, Coluber constrictor. Univ. Kansas Publ. Mus. Nat. Hist. 15: 351-468. 277 Shewchuk, C H . and H.L. Waye. 1995. Draft. Status of the Western Yellow-bellied Racer in British Columbia. Ministry of Envir. Lands and Parks. Victoria, B.C. 278 Rosen, P C . 1991. Comparative ecology and life history of the racer (Coluber-constrictor) in Michigan. Copeia 1991:897-909. 279 Plummer, M.V. and J.D. Congdon. 1994. Radiotelemetric study of activity and movements of racers (Coluber constrictor) associated with a Carolina Bay in South Carolina. Copeia 1994:20-26. 280 Fitch, H.S. and H.W. Shirer. 1971. A radiotelemetric study of spatial relationships in some common snakes, Copeia 118-128. 281 Brown, W.S. and Parker, W.S. 1976. Movement ecology of Coluber constrictor near communal hibernacula. Copeia 225-242. 282 Macartney, J.M. et al. 1988. A tabular survey of data on movements and home ranges of snakes. J. Herpetol. 22:61-73. 283 Seigel, R.A. et al. 1987. Snakes: Ecology and Evolutionary Biology. McGraw-Hill, NY. 284 Nussbaum, R.A. et al. 1983. Amphibians and Reptiles of the Pacific Northwest. Univ. of Idaho Press, Moscow, ID. 332pp. 285 Preston, W.B. 1964. The importance of the facial pit of the Northern Pacific Rattlesnake, Crotalus viridis oreganus, under natural conditions in southern British Columbia, M.S. Thesis, UBC. 286 Blood, D.A. 1993. Western Rattlesnake; Wildlife at Risk in British Columbia. Brochure. BC Environment. Victoria, B.C. 287 Macartney, J.M. 1985. The ecology of the Northern Pacific Rattlesnake, Crotalus viridis oreganus, in British Columbia. M.Sc. Thesis, Univ. of Victoria. Victoria, B.C. 288 Charland, M.B. et al. 1993. Status of the Northern Pacific Rattlesnake in British Columbia. Ministry of Environment, Lands and Parks. Victoria, B.C. 289 Stark, M. Personal Communication. 290 King, M.B. and D. Duvall. 1990. Prairie rattlesnake seasonal migrations episodes of movement vernal foraging and sex differences. Animal Behav. 39:924-935. 291 Ruibal, R. et al. 1969. The terrestrial ecology of the Spadefoot Toad Scaphiopus hammondii. Copeia 1969:571584. 292 Waldick, R.C. et al. 1999 The consequences for amphibians of the conversion of natural, mixed-species forests to conifer plantations in southern New Brunswick. Can. Field Nat. 113:408-418. 293 Beebee, T.J.C. et al. 1996. Factors affecting population densities of adult Natterjack Toads Bufo calamita in Britain. J. Appl. Ecol. 33:263-268. 294 Denton, J.S. and T.J.C. Beebee. 1993. Density-related features of Natterjack toad (Bufo-calamita) populations in Britain. J. Zool. 229:105-119. 295 Leonard, W.P. et al. 1993. Amphibians of Washington and Oregon. Seattle Audubon Soc, Seattle, WA, 168pp. 296 Lampo. M. and P. Bayliss. 1996. Density estimates of Cane toads from native populations based on markrecapture data. Wildl. Res. 23:305-315. 297 Tyning, T.F. 1990. A guide to amphibians and reptiles. D.W. Stokes Nature Guides, Boston. 298 Stebbins, R.C. 1985. A field guide to western reptiles and amphibians: field marks of all species in western North America, including Baja California. Nat. Audubon Soc, Nat. Wildl. Fed, Boston. 299 Richardson, J. et al. Unpubl. Data. Centre for Applied Conservation Research, UBC, Vancouver, BC. 300 Shatford, J. Personal Communication. 301 Palis, J.G. 1997. Breeding migration of Ambystoma cingulatum in Florida. J. Herpetol. 31:71-78. 302 Kleerberger, S.R. and J.K. Werner. 1983. Post-breeding migration and summer movement of Ambystoma maculatum. J. Herpetol. 17:176-177. 303 Semlitsch, R.D. 1981. Terrestrial activity and summer home range of the mole salamander (Ambystoma talpoideum). Can. J. Zool. 59:315-322. 304 Van Buskirk, J. and D.C. Smith. 1991. Density-dependent population regulation in a salamander. Ecol. 72:1747-1756. 305 Whiteman, H.H. et al. 1994. Seasonal movement patterns in a subalpine population of the Tiger Salamander, Ambystoma tigrinum nebulosum. Can. J. Zool. 72:1780-1787. 306 Petranka, J. 1998. Salamanders of the United States and Canada. Smithsonian Institution Press, Canada. 307 Madison, D.M and F. Livingston, III. 1998. Habitat use during breeding and emigration in radio-implanted tiger salamanders, Ambystoma tigrinum. Copeia 1998:402-410.  143  •g  2  E  .s »  .ti o  1  - 3 0> CO CO oi E  3 IS  > £  5 5 £ 2  5  ra  2 CO CD ti) 1> < D  £E  =  O) CM S CD |  §  g  CO 0)s <p § CM > &  01  1-  UI h-  Ct LU  ^ II  8ro  £ I £l  ?!• nj8 'ml x |  ? £E ro ro — 1  §3 *o E o •£ w 5  CT O 0) .  CO —  f S3.  ro " 3 ro > O ~ lO ^ II «  11  c a) 3 '  oO •S A o  a. i=  °° 3£ a)  6 -a o o , o  .£  Q. in  . •! a= ee gg  cn  ro  —  CD O  _ El ro M cu —\ ro S\ E  ro  Eo |  j ? 6 T? i  O  CM  O  cr> in E <  O  a— O  c ro o  co  ~ c  01 144  5  s §•«, T3 S E ij A  Z)  O  LU  u  C 01 .•s o o <o  CO  .2 £ 2 «  o i  S  i  3 3 cy  > CO §  I Us  (0  Q. CO  ro  E |o ?  ro  ° "I ro *5  CD TO  <2 co £ x: 2 co  CM OJ E E  in  zo 5£  a  CD El  <  o  § E  o ^  .or e 111 CO  CD CO  ri Ej  1 8  _-a|  IT) LO m  C£ Ul E tl  o CD (0 o -CZ ro•^r JC OJ m D o Ir-  o  reed TE =25) nte n=1  Q  CC CD 5  o  OJ  X  c c\i  It c  <= «  °> 15 E .a>  a. o  CO  <  ca co  |cco  cu  o  O m  ?< 2 co  Ui|.  co :  5 => ;  _ g ai u  CD a>| ri 5  Bx CO  J)  o E  145  13  Q  Sf o  ow  ra to  J2  3  i  « >  S  E  CO CD  in P  cu cu ° CO  co a <U  co jL cf ^ E o co 2 CN (  or  o <° «> co  CC T I  "M cV C tJ]  E o  to" 8. a  sz  .1:  E  X  CO It c  ?  CD CO X CO E  c o O CD >»  E 2 * 3  UJ rco~ _c CO  let  DC 5>|I  ° o  X  o, co co to co >_ CD 0) N x: CO o  ha,  o  E <=5  •7 •M"  E 2  O  K  CN  dc  E  E  ha  > 0> Cfl r*C 03  CC  LU H  < x  55 o E _>* cu CD 5 E .5 2 •$< .2? cu cu 01 * c CL 3 O >  CD  a. o  c < oj « _ co 3 co co —  tn *A  in  cu  2 ° o  i i  £ CJ O to  CJ> CQ  5  cfl  < in!  146  <t  col  >[<  c o l < CO  l< "1  LU  z  OlAI  147  _j  OR  ME  BC N.Am.  |A. herodias A. herodias  A. herodias  A. herodias IA. herodias  LO CO  •aCM CO CM CO  X X  39 to 750 nests/ha in colonies 78900 colonies of 5 to 550 nests with avg. of 158.9 nesting colonies from 2-80 nests, n=29 colonies up to 39 nests in a single tree  CO  no  no no  no no  yes  yes  no  no  no  no  no  no  no  no  no  no  no  no  Used for Density Estimate no  no  CD CO  X  X  heronries up to 2 ha of forest  size of forest stands supporting colonies median of 108ha  CO CO  5-8 pounds  25 prs. in heronry in aspens  o  2.1-2.5 kq  >= 300 m buffer of no human activity around colonies  0.04  X  1,000- >10,000ha; Mean dist. flown from colony to principle foraging sites 2.3-6.5 km (in SD and BC); will forage up to 30 km from colony; HR.  X  333.3  CN CO  mean area marine 8.4 ha; n=32, freshwater marshes 0.6 ha; n=7; HR  O CO  BC BC  X CO  A. herodias A. herodias  <10kg  X  CO CO  N. Am.  E CO CM  A. herodias  o  OR  CO  NE  CO  min. area of grassland 10-30ha min. area was 8ha  CD  A. savannarum A. savannarum Ardea herodias  12.6  « !ai  1-1 Oha is min. amt. of contiguous grassland for viable breeding pop.  «  A. savannarum  5  E  avg. 0.9 ha; TER  £  E  9 fields occupied avg. 1lha (121 ha); TER  Te t  0.66ha male, n=11; TER  121.7 "E  °1  E  0.85ha male, n=73; TER  1.2 to 3.3 acres, avg. of 2.03 acres; TER 0.8ha male, n=22; TER  Density 8 birds/1 Oha, n=45 and 2.1 birds/1 Oha, n=14 4.3 birds/1 Oha to 42.1 birds/1 Oha; 1.7 males/10 ha to 15.8 males/1 Oha S3  240-1348ha required for viable pop. of 50 breedina pairs  Area Requirements  ">  spring pop. <100 prs in BC; min. pop. size is 10 ors in Okanaaan  Population Size  Individuals per Square Cited in Kilometre Source Source 50.5  CM CO CM  CO  BC  Mass (q)  Home range (HR) / Territory size (TER) CO  A. savannarum A. savannarum  Geo-graphic Latin Name location A. ND savannarum A. SD savannarum A. USA savannarum A. PA savannarum A. savannarum A. CT savannarum A. OH savannarum A. FL savannarum CM  to  CO CO CM  X  X  I  3  LU  U  C  01  a o  Si  S> £ l to  .2,  Elio  SS  <D ro  cu ro 2 w w«_5 o) C CO D C D £ CD ^ 00 (0  15  IT"  T>  o °- 2 t  10  E  IIS  01  o  S5  11 CC c  ce  E  r- in CD a) E CD .!5 E f 9> E  -c ro . O  2  2  CJ Q. LU  O  0)  S|  3 IH  col  E u i 1  0  ca ^  Si  is  1.2 B  g5  E o Z  5 2 E a> .5=1 co 10 to CM  .E  £  •  CO  2 o  CO 3 ol  CO 3=  E co  CO w |> . CO -5 c c a r Sl 0)xi E  CD d)  co » B  CO O CD 3, S  O  0>  8  ra  10  o £  • co "*r  CO  w  CO  E  Ji  CO  ^  h. M  0  II CM  c  r  CO CD  E  "> ra £ E CO  « cc m co .S? cr v co  ^  o o CD —  CD  S  . -  §^^ CZ  LO  8 <S E  CO  CO  ai '3  148  E 51 3 E!  ~C~0>  .ti o o <n  •g cr > (0  E g o8 CO Q CD W cv| OT 3 ^< CD CO O U C cn . . m CN £  co f3 fS E  i  ™ a S s. °  :  jS <? co CM .c o — o ^ CO *a-  s - •£ e d O  C  T  -  e  3  g  ra ST jo a  c3 co cd g ; «! ^ ra o>  o co  £ 2S W CD  I O CO  s g> = ra m  8 W COS 8  CME " "  ^ co a CO N CO <  -S c  E ^ \£ CD S E  ?& i  _ at cc UJ  3 co  a ??: E  3H  CM 0)  E 1  CO CM ' E T- C O CN W ¥ E o '  ai ^ x  m"  v ..  c T. - CM W || 0) c o  x  0) <;  g 5  JC C CO O CN T-  T-  " m lO  2  o  o ca ~ o 8  O OO O CD CD E E  149  £5=  3  UJ  u  O CO  « S 0) « n i = 3 0)  •a or c > (0 §  «  &2  <0 (NI  cu E  .  S3  ™ C  o  Z.  E  CN CO ^ — CD C N CN  E  O) co  • S c\i C co 2 a> 2 £ Si ll  cc  CN  r  CN  1  CO  >  0 C  <° £  CT>  v  II  ? «  c  ° °£ I 01 > CO  cb  JC  > CO CO C N  !6  CC LU I-  _ UJ cc S  c  J= r-  o  O  ha  o  o o h-  c jd to Q)  c to o  Q. CO CO  O  JC  tabi ana  DfOX.  I  o  J *  O  C O (A CO o0 0  m  OJ  JT  OJ  JC  -—'  CD  1  0)  o  - I " I: A  Oj  0)  CD  CO  ~ E E S  CD  E °  CO  LO  tn  ?6 . tO  to  CO oCL CN  I <1> C i  a)  Cfl  Q.  Q.  § o|  - OJ  5? »  * 2 8 *, I CO Q 1  CD  Ol  v £- E g  CD  22  5 3  o ooo o 150  C5  8  •a 2 E  D ° LU .*! o o <o  IS in f P? •~  II  »  .2 CD  x  £ o; CO T3  3  o  co cn • E  c  CC UJ H 2  ^  Qi _ I 75  c c -ol  a o  CO CO  ci  •p g CO CD  CO tO . ca Q . co Q co CM  0 CD  C  > I CQ  11  75 CC CO LU co r-  E o  X  t | . iCDi - i §i  CD CO  i—J~ —  • -  o> « 5 y o cl « w *; a  u  *i i b 3 2 •~ cncnLOLOcoP3 ~ -H cx p CD  'II to O cC Ito**-I jo"co ofo: p  j5  s •*  s -a JO CO  75 co 75  i i 5  a> I—  CD -P E I?! 75 i co y ;  s i  ^ o CO CD  as  CD CO c 13  •Si CM  00 3  £  2 75  CD y T3 To ,S 8  O  CO CQ o  6  CM CQ  c "O  to  CO a d d CO II CO II E Q a) a CD CO CO CO E -  3  Z ,„ CDT—LO' to to a) fi: co to to  E 4 o> d oo m As E coj  cn  CD . . - ~ ~ U) h00 CM  V ft £ S3  CD  E|  151  pi  S J  15 a g O OT  to i  cr c  E  E. 3 • E •ifO o.o .i C  to  o  I&= . -  c  re vlt_ i3 sCO  o! ro 2.: in C tM c to ™CO cu o E B  E  o  CO  S E  CD IX ce teUJ  CD CO* -o J= C CO CD D N -  E o  X  S E Cp CM •S CM ce x  0) 21 o  D) CJ» OuCM.CD« E « O 5I •g E o tCDJCCOo T—6 o " d o CD S X - to .. N 4 S 9 .2 "= E CM CM* E | 9 « E S « CM* £ ° . E 5g I JC CO OT} L! > — C O CD to • Q. — C c  CD  £-!  C CO COCMhJC O  ?  5  CM CO II i_-  E * «"S CO ' •f g O CO  §2  r~ o " o:  CD I E cz P to § £ co c  3E  QL  £  X  ce Ix  CO flj E £  r  1  o  t= a* C OOJ •-c to•  ra cD o m C oC o Jto CC Z O) D O t CZC  §1° o  - g col JO JC c a o  co a) CD CO ra E E& CO CO r- ccoo to  8°  OT ca 6 CO to o CO o b 3  CO  ro  ion  3  C ED  O EO E  152  ££3  0 X) "O o >« O) ™ CO c m co  o •> to  m  « f  J Z  io ^ tl  2  ^  cu m  ^ E (13 |P 2 <N Tl 1 •_ 2 g 3 ^  CNJ  OJ  to o  o o -o  -Q o  CO 5  CX  CO  QJ  lO y>  do  JO J 2  CC LU  co H  to ce  o  £ x  ? % 8 *2  -g CD  X a: ce LU  °I  |y cn c 2  r? <°  C M ^<t= hT-  C cb  ©  O  Lc ' 1^ ^  ro a  E o  X  CM  «  cu D)  ro GO ra cu  0) •o cu  <  X  6 E cc LU  JC CC LU  3  2a  X  E CC co" LU  JC  ' CO Dfl co  co LU JZ JZ:  cn  22 ro E S to 2 •*  CD  1  rt O) i f ri o CN ' -tg m CO O "O Q O 3  CU CU . •= IS "=> o  - °_ £  E  , - — o to O "> Q.% C°-O  •°  ctoa C.£Q' 2O t; 1 1  S '«> ° to "O -E  °to> E. . E r  4  ^ -H co ™c E E -o ro  ro  to o  CD CNi  CO CO LOl V T-  E  '  2  2> T7  CO ~  l> s o o CO  P CO  —  CO  E  O E  0- ca  a. ro Q- ro  153  I cc  :O  E <  •o g E  ti o o co  2 cr I > to  cn o N tO 0) CD  £  !_  p a  a 151 tu ' co cn ' E to  .<*- E E E "  J  s  OT  E  C  a>3g'so  ro cr  £ B f . to  £ £  Ti) Tu ^TO  5  E  TJ -o  CN  tO CC  Jz f p toa>  .5 « c  II p m c  CN CD  c x ? " E  to to  E o  5 S  5"•  P  a. a.  tu a  C  P  ^  T:  t to  C °toO C3D>"CD 3>  tN  L:  E CN £ E  CD •=  O  •£ °> 5t  £  o d  3 o 10  I8 SI  a o  2  53  II  P  a. cu  CO CO ^ Q> CO CO {vj  CO  1  £15  in OO " C  T-  _ cc  oo  tn g £ ^ ca co  CD tD CO 2  Oo O 3  154  d S6 S  P  3 °  I sCO £ JC (D 0) to  P  8» S  t-  D fieri ro .2*  •== tn  to  •2 9 5 ccl  JO irof °Qill "tu £ .5 9- to p o- * ) °O  « E  p <D  co  £ tu  P  E •& X  —  co co - o 9' fl) Q T3 2 ^ ID "  X  ^ xa  <D CO CN  d 8d 3  0) CM  Ef  T3 2 E  S S "=  "3  t J  LU  O 05 o en »  £  a>  2 mi: -3 3 0) •O CT C > 05 o  I Us "O „ cu o  5 °>> ^ E I T •s # 8 ©a •—  01 2 c O  o  OJ  S S  o •o o  S ra C  o  OJ  ?|  F m  e  CN  id  CO  CO  a i a •g  X cc  — cc «r x  ? Eg> E  E o X  £  E JC  *5 >  P  ro s» « ,  =  to  0  CC I  CNi  _  JC  ^  .  T-  T~  1  °& $2  -<-- CO cc  T T  X  •* E I d> E  g co  CO  co co  <D  ro o  1  tO  =  LU ro co  aL  -o  O m  08  o S ro  CL  cc  X  0 s0 s  155  O  CO  CC  5 co  > O  — CM  3 = x ra g E E c co  . S > E  e S£ « S =3 II A  T? CO  ra  CN  "cO ro CN CN* OX  E JC CQ  EE  £ te 1^ 8 LE £  C  EE  CD  C  $ o :> ro  3  2 ° o  ,  " 9  ro  CO  O. o  CO  CO CD  CC  .  o coco E roc™ . 0. X c 3 oi" S o CN CN E co E E2 E «5  <0  CL O CL  x:  0 ro 1  E to 0) cu E 5 ro io E  10  •° 00ra-° T~ JQ JQ in  ra  CN  •a E  —  E  CD £ •  w E  T3  CU ~ CO • £  co  CD C L CO - C L CO  CN  ' i5 3= c | I gl  -rT « -4-  o o  co CM h-  C C  TJ  gE  Jag i e ±i U  tfi  ro = to=  01  o  <n  01  <5£  est  ro  cz  ^  N  O  «  o  E  {Z  CO  O)  ci' F ° s o  <  n  a CO  E  •if  cl  e E a; S » K " aX -O  oi a d)  •f 1 W  Cfl  J=  O l CO  5\JO  - .2 °  CD T;,  2\ 6  E £ 'ro ro c ro 1b c& o  w  § « ? §  »- 3  ES^  c  o)  5 0)  r to j,  5 « 3 "Ol  s CO  t/)  0)  CO  •° ro ro T"  H  i] I  0  U) C  ro 8 —  ct: n  V  £ >-  CM  CD •_^ CN CO C CO CO  E -g t •? o  £ E  C c J CO  CD  T-  CO  CN  TJ  CO  CD  CO  CO  oi ^E to If £ O  TJ  Q.  CO  fit 3  O  E  e ~,  O O  c  'sz  CO  E  O) CO tu > oE o> cOJ E p i to o o w c  CO it. "O CD  Q.  O Q. CO  CO  CD CT CO  E  m  E  CO  2 " n" cS =, E  »  <g t  . - ro ~  c oi c CO  II  o.  a  5  " It  c:  CO CU  ^o 3  CO  "O  CO  Ol  CN  CO  CD  £ «  i?c» o 2 CO  C 0)  ro >  c O CO  jit:  CO  i ro oi 2 ro  i co O  •= a cS S  *~ co" OJ  3  f-  TJ  II  CO  CO  co oi  •£ co c  ro  CD  £ 1  f c a ro  o ?5  §»8 Cl  O  —  O  CD  O cn  CO ^  LU E  156  O  .>  CO  •8 £ CO to  1  52 o  CM  E  1  ^ s ° trS-  a  q ro S o a > o a CD  > III  E  O  %  CU  CN  CO  ^  t5  0> CM £ ci £ . to O to CL* 8 E co 5 ro to to c T?  *- A  iS  Q  Cj  — o  .ti o O  '  co 8 I15I o-* co 3  — I Q.  S jo .2 3 to To to ^  cu • N tO  o  6 ro  V CN  to  to  °>  S  111 .Eo.cn o.E?«>  3 "a EES  s °> to  1  cd £  Z. or CC ui  ri £ ,  to 5> i  CD CNI CO c  £ c  ro 9 j  tC oD  x: g>  .2 «=  S E E EH ?o w a 3  a)  - E co  ^  ° = to.£ ^ Ew B ro .£ <- tu &" d  E  ro  CD C = 31  O !  ?  CO  O  CO »  H"  ci  » E » E  o • - co ±i ro II S "o Q. CZ  CO  a. o S ~ cn  c »«t- cc cc S O fO o Q O CQ O caCOI CQOl . r CO JC O d mi S §• E  8°  CD  O  CQ  ro  co a; o o CQ  3  3 U tJ  28  157  O  CQ  o 3  TJ O CO  2 o" 5E~ > to  3 ra  E  §  E  C D KI  •^  CM  > £  •o Jc E o  ,«> a I tO CO 2a E2 t2to ' ro i  a  i O  '  ;• <- <° CD  «  5  t/) "~ O cfl  ffl  2 '--L  i ^ ro ff  £ • •c «  'c  ro  0. JC 2 d.1 o. o C. C CO NI E E .E £ .E; o o o. o  E E  o ^  "5 31 O E  is 3 g  o" o ™ o m  (0 c  1  — o"  o ro E  ^  CD  ro  n E > S E -o  X or a: UJ  £ S, CO  *~. S.  E o  1  X  ro^ E CNi  oJE i -3 P  ttl  a:  2^ C <N1  CO  c  E .2 ro E  a. o  8£  C9  D)  ro Ic O O  O co  Si  m  £  158  E  3 Q  o co o  CO jo o co to £  3 3 CD > co §  ? t r  p ro  £ C3° ° g . E co E to" E  CO  P  c  III cT- ^ 8  E p S 73 o w  P C o pO  d>  o ws 1 to ^  CO  cu" io  ro Qi  35  o  E p a) E  la:  CO o 4 P a: CD O) .b I CO>  X cc  V  a eg" £O  CO JZ CD CO  C-  P  in ol  «  o JS  Si £ go °  CM  ^ i ro  5 o V  E  Q  w  co" '« E -ol  T—  ^  «  CO (j)  E  -  2  E E  CO CO  p -2 P  § 2 it 5  E o  E  CO CD  CO  co E fi. jD  ro  O  CM  5. E CO o 3 4=  -ol  o 1>CD •a>| Z.  /_ "O  3 CD  o  £  £2 CD cu co rol ro ^c ro O CD T —  50  "I <? CO CM CM 0) « CO CD  I ?£ •P  E ro " E  s  o o o — O  co  ae °co>ro ro 3 cu a  g. g- co  ; § 8  i s S  !o  159  E  co  •o » E  .•a o o to  CO  CN  CN  CN  co  • « i 3  3  0)  p CO JC  c  cf 3 If-8 ^ o O  sei  CO  JC .  JO CD  0)  CC LU  CD  0>  °  CL  CO  . -  E to  ™  <°  CN  cc r  .  CO  CO CO J C CD  •i I  CO  CD  S. ™ I ^  s ^.  N  CM  to CL  CO CO  CD  C xt!;  JC  o o c ai = n cc ™ o< •o Lc |co 5 OC a > X to o I f E cc X - CO oS E -asl E'! ? « co o - o 2 E o co" co CO CD; ll OJ E C 33 COj E S\JTCO c •si CO  Q)  .  E  <D" "£>  CO  CD  c  CD  nfp  Eo X  X  E  TJ  CO  JC  s-E CO  CD  > C  0)  CD  <L  E  v  mE«  -  OJ CD  o E S E & 3 E JEC CO o CO  <D  «  o>  -O  CD  ra c O9 CD£  £ c E E  CD  CO  C  CD  P  tD  O  "  5 J= = J= p 5 s ~ E p — E E .E , , . - E •o cn g i: tfl t o J: IH « c , 5 E ^ o vo ^ jo E -q 9- E I § « sg  O  C  .3  Xi  01  CD  P  N  . E , " D  r  to  ID  CD  Ol V  E  "O  T3  t-  °  „ o 3  o o O 0)  —  co  CJ  m  ICQ  160  I  I  CO  <  C J CO  5  O o  cn  a o o co  iS } j 2  «  *3  3 3 0) 3 cr c > to § I S.2 in co E p  ^  CD  3  31  E  3  CO •= CD E  Ol CD i i  co  '1 ro  to  5 "> "D  . 31  i f ^.  Dl  to to N " TJ _ r « » „ •S <2 „- t< to " JO C D • 3 To 8 3  ro £  O , 4 lO T)  cj n i ;  a  E o  co  -3  o  CN  to  i; a T3 "D T3  to  ^  fJ  3  «  TJ ^  tz c c  -a ™ ~ ZT Z" "* ro ra ro ro ro ro  O CD S « CO P  -  3  to ra  sz sz sz  J2. ai 5 m a- 3 CM to ~  0 * 0 ( 0 0)0)  CO CO E .E to cz o .E, 8 1 « ts •* Q. to Q. CO ±= CD O  Sro*i izt *=\  d  2>  CO 3 JO  II B^  X cc  CC UJ X b  to  E  i»  *3  = I  JZ> to "tol 3  O CD O ^ (jj  ra 12 iS ^  RO  J3 cx OJ co  C  OJ  co  " " I S 5 gro 28 "8  o to — U) o Q. jU "O 0) . - c O J to E O J 2=" m CN - co £L  OJ  T _  ^  "2 w^ ro  TJ  JS  ^  ts «  = ^ "O T3 CO § co  C-  W OJ  3 3 CO  CD CM J Z  a?™ ^ S o-g to o co r: C.£ro2 2  3  CD >* —  ii Ol c IN-  c  §  o— « .E  CM  e  CO JD f 0) C O CD  3  E  P in  "J.  5 tD o —  3S 5  3  CD CM  CO  £  to"  "co oi E to  «S?  Ol O  S iB to co £ .E E ; g r  co E r i IE £ to  CM  Q. O  82 o  co  2  O  !=! O O  "  E  CO  jz  ,§ i  coO CQ 3  CO o_ E °. <- <o "to E o °. CD CO CO  8 a  161  ro ro _ E E E E O 3 wrato •9 o E c E ~ 5 < E  ro  =5  Q  o <o  1 5  S 2  •1 ,5T  O  E  ° O O o cv  c o cu o CO CO " _ " r 10 -1 C Cfl CO t -o p PO X)t/> "5 c E h c c: a 5 a Q  ^  E  g  to c — 3 CJ  CD CO  p  o CD m ro to E co i5  T-  Z  O  § J E "  && &  E  —  I  S  1  c  *  <2 §  3  £  3  CO O CL Q. I  X on  at  UJ»= CE II2 ° o c  1 1 CM  El  i-  E  . to CD 0> ro  JJ  *> CD •si J5 ™ i_ o Q CO CO CD «-  •; >* V C  o • £ CD  f If CO  +  -rt-  CO T3  o  CD a-. £ ~ ro c'  CD o £  D U O )  o '•• £ E Si  c " * E 3 co >, h oi C CO CN CO T l i s .i3 D F - i (Nt CO C 3 °- "O TO + CO E -D -D 2r c  g -fe CO -EC  -  to CD p S c: E o CO E D. °> c E CO CO C _j o o •g E CD O CD C\i E .E CL Q. CO  3  E  J= c  a o co  2°  o  E  E ^ E  B  E  »  DC 3 E- CCO CD  <  ~  C  CO  ,_ E CD ro S co h- c/)  E  E  c- d  1  162  3  Appendix VI. Chi-square table with observed and expected number of occurrence records within 500 metres of (a) paved and (b) any (paved or gravel) roads. (a) x = 595.764, DF = 1, P<0.001 2  Distance (m) from Paved Road 0-100 100-200 200-300 300-400 400-500 < 500 > 500  Observed Number of Records 328 160 135 109 96 828 859  Expected Number of Records 109.645 84.195 75.383 68.581 63.380 401 184 1285.816  Total Area (km ) 115.048 88.344 79.098 71.960 66.503 420.953 1349 177  Proportion of the Study Area 0.064994 0.049908 0.044685 0.040652 0.037570 0.237809 0.762191  Total Area (km ) 597.108 307.924 185.918 127.084 91.175 1309.209 460.921  Proportion of the Study Area 0.337325 0.173956 0.105031 0.071794 0.051508 0.739612 0 260388  2  (b) x = 360.585, DF = 1, PO.001 2  Distance (m) from Any (Paved or Gravel) Road 0-100 100-200 200-300 300-400 400-500 < 500 > 500  Observed Number of Records 969 325 163 96 37 1590 97  Expected Number of Records 569.067 293.463 177.187 121.116 86.893 1247.726 439 274  163  2  Appendix VII. Number of individuals and occurrence records per square kilometre of twentynine threatened vertebrate species relative to the habitat quality categories identified by the wildlife habitat relationship models.  Species  Habitat Quality Rating High  Ambystoma tignnum melanosticum  Number of Records 27  Moderate Low Nil  Spea intermontana  Number of Individuals 581  llllillll illlllllj.  Dolichonyx oryzivorus  Spizella biewen  Buteo rogalis  Ardea herodias  Ammodiamus savannarum  1.63E-02  295.785  6 76E-03  6 76E-03 3.03E-03  3 03E-03 7.85E-01  1 13E-01  Moderate  42  12  393.985  1.07E-01  3.05E-02  Low  71  14  479.057  1.48E-01  2.92E-02  Nil  1  1  535.343  1.87E-03  1.87E-03  97 6571  1.02E-02  1 02E-02  Moderate Low  llllillll ll|Sil|H  134.336 68 8409  ^ ^ ^ ^  1469 29  High  26  5  42.7196  6.09E-01  1 17E-01  Moderate  0  0  52.4255  0  0  Low  0  0  27.1283  0  0  Nil  7  5  1647.85  4.25E-03  3.03E-03  High  379  132  324 595  1 17E+00  4 07E-01  Moderate  60  143 99  4.17E-0I  1 81E-01  67 3571  7.42E-02  4 45E-02  ISiS •Bill  1234 19  1.38E-02  6 48E-03  70  111.958  1.16E+00  6.25E-01"  41  28  198.878  2.06E-01  1.41E-01  Low  0  0  15.3494  0  0  Nil  40  33  1443.94  2.77E-02  2.29E-02  Moderate  0  0  263 536  0  641.393  3.12E-03  3.12E-03  Higi  130  Moderate  isisiiii  Low  0  865 199  Moderate  26  6  255.697  1 02E-01  2 35E-02  Low  4  2  382.773  1.05E-02  5.23E-03  Nil  24  6  1131.66  2.12E-02  5.30E-03  High  illlPiM  4 84029  ••lilllii 2.70E-02  2 70E-02  Moderate Low  •"•:4. . *  SilliwSiiisiiiii iliipiiiii iiiiiiiii 17 35  148 346 '  '41.2713 •' " 7-.27E-02  2 42E-02  1575.67  2.54E-03  1 90E-03  61 5034  5.69E-01  2.76E-01 1.79E-02  Moderate  8  7  390.186  2.05E-02  Low  0  0  260.507  0  0  Nil  3  3  1057.93  2.84E-03  2.84E-03  High  16  11  344.46  4.64E-02  3 19E-02  lilllliil  Moderate Low  0 523004 28 9463 1396 2  Iliii^^Bi lll^pSIjJ 4 30E-03  2 86E-03  High  198  51  1005.86  1 97E-01 "' """5.07E-02  Moderate  20  14  183.358  1.09E-01  Low  0  0  100.939  0  0  Nil  11  8  479.973  2.29E-02  1.67E-02  14  7  62:  2 24E-02  1 12E-02  5  272 693  1.83E-02  1 83E-02  367.549  8.16E-03  5 44E-03  505.99  IMlIiBii  " High Falco mexicanus  2.71 E-02  361.743  Nil  Melanerpes lewis  6.65E-02  737 685 330 527  High  Numenius amencanus  1.43E+00  2  41  Nil Chondestes grammacus  Number of Records per km  2  2e4  """Low •  Catherpes mexicanus  2  Number of Individuals per km  High  High Botaurus Icntiginosus  Total Area (km ) 406 13  Moderate  ISillillll^SIIillt  Low Nil  0  !8|llSllll||I|  164  7.64E-02  Species  Habitat Quality Rating  Number of Individuals  Number of Records  2 468 .346 247 13 1 11  1  Total Area (km ) 2  Number of Individuals per km 2  15.4837 • 9E-01 Low .,(151.853 3 08E+00 Grus canadensis Nil 1602 79 2.16E-01 ^ ^ ^ ^ High 148 136.747 1 81E+00 Moderate 11 100.233 1.30E-01 Oreoscoptes montanus Low 1 80.0325 1.25E-02 Nil 9 1453.12 7.57E-03 Moderate;.704.151 7 10E-03 B B H 1 ' 0 Low 91 029 l l l l S i l l ^ B Asio flammcus Nil 0 974.949 iillllSlllSSlllli 1180.12 Moderate 45 18 3.81 E-02 1 1 Low 258.029 3.88E-03 Cathartes aura 1 1 Nil 331.98 3.01 E-03 High .411 089 49 1 58E-01 248 438 8 45E-02 Moderate • Picoides albolarvatus Low 7 08E-02 SlllllllSt 141 334 Nil 48 69 969.268 7.12E-02 High 1 1 26.721 3 74E-02 Moderate 0 0 128.894 0 Otus kennicottii macfarlanei Low 1 . 1 29.1772 3.43E-02 Nil 6 4 1585.34 3.78E-03 High 5 18E-01 11111111 38.6..•Moderate 12.1733 5 75E-01 Icteria vnens Low 57.8094 5 19E-02 Nil 28 15 1661 52 1 69E-02 29 442 084 High 46 1 04E-01 2 2 Moderate 552.933 3.62E-03 Antrozous pallidus Low 0 0 217.725 0 Nil 6 4 1.08E-02 557.386 64 27 '1038.87 6 16E-02 l l I ^ p i l l ^ B i l l l l i High 82 02 Moderate Corynorhinus townsendn 54 353.427 Low 104 2 94E-01 Nil 0 295.811 0 llSilSsisliittf* High 4 111 1006.65 3.22IE-6T 3.08E-02 Moderate 17 9 551.165 Euderma maculatum Low 1 1 63.1639 1.58E-02 Nil 4 4 149.15 2.68E-02 '17117 831 406 653 . 4.21 E+01 Very High 222 High 2776 405 175 6 85E+00 204 Moderate 7073 411.319 1.72E+01 Ovis canadensis 5 02E+00 ' - Low 683 '-',,136.15 cahfonvana • > , ,'~ " 64'.6603 • 1 52E+00 Very Low 98 346.171 Nil 28 2 36E+00 818 683 571 High 9 9 1.32E-02 1 1 377.78 2.65E-03 Moderate Charina bottae 384.071 2.34E-02 Low 9 9 1 1 324.706 3.08E-03 Nil 871 694 .127 • 1 46E-01 High ' ~ 109 4 25E-02 15 - 0:1 ->, Moderate ubei constnctoi Low 167.278 5 98E-03 1.51 E-02 Nil 331.056 High 170 "*8~93.618 3.31 E-01 296 21 423.697 5.90E-02 Moderate 25 Crotalus viridis oreganus Low 0 0 172.128 0 8 8 280.686 2.85E-02 Nil Moderate  llllllii  1  3  iHiliipi  165  r  Number of Records per km 2  6 46E-02'• "j 2 63E-02 2.50E-03 1.08E+00 1.10E-01 1.25E-02 6.19E-03 5.68E-03  1.53E-02 3.88E-03 3.01 E-03 .1.19E-01 3 62E-02 3 54E-02 4.95E-02 3.74E-02 0 3.43E-02 2.52E-03 4.40E-01 5 75E-01 5 19E-02 9.03E-03 6.56E-02 3.62E-03 0 7.18E-03 2 60E:02  l^loi^pil 1.53E-01 1.10E-01* 1.63E-02 1.58E-02 2.68E-02 2 04E+00 5 48E-01 4 96E-01 2.94E-01 2 01 E-01 8 09E-02 1.32E-6'2 2.65E-03 2.34E-02 3.08E-03 1 25E-01 3 7EE-02 5 98E-03 1.21 E-02 ' 1.90E-01 4.96E-02 0 2.85E-02  Species  Habitat Quality Rating Hign  Moderate • Hypsiglena torquata deserticola'^,. , '"'- Lbw J  Number of Individuals 6  Number of Records 6  2  469.211 0  ^^  ^ ^ ^  .. Nil Pituophis catenifer deserticola  Total Area (km ) 438.419 -  370.23492 268  Number of Individuals per km 2  Number of Records per km 2  1.37E-02  1 37E-02  4.26E-03  4.26E-03  lllli^^s 2.03E-03  2.03E-03  High  117  92  832.754  1.40E-01  1.10E-01  Moderate  33  31  424.761  7.77E-02  7.30E-02  Low  8  8  178.722  4.48E-02  4.48E-02  Nil  9  9  333.891  2.70E-02  2.70E-02  166  Appendix VIII. The (a) Chi-Square (x ) and (b) Log-likelihood Ratio (G) values of the distributions of individuals and occurrence records in relation to habitat quality categories identified by the wildlife habitat relationship models for each of the twentynine vertebrate species, except American Bittern, Ferruginous Hawk and Short-eared Owl. The degrees of freedom (DF) for each WHR model is 3 unless specified otherwise. The probability that both the frequencies of individuals and occurrence records in each habitat quality category do not result from a uniform distribution is denoted as: *** for P<0.01; ** for P<0.025; and * for P$0.05. 2  (a) Chi-Square Results Species Ambystoma tigrinum melanosticum  Spea intermontana  Spizella breweri brewen  Ardea herodias  Melanerpes lewis  Grus canadensis  Oreoscoptes montanus  Cathartes aura  Picoides albolaivatus  Antrozous pallidus  Corynorhinus townsendu  Euderma maculatum  Habitat Rating  Observed Number of Individuals 581 20 2  High Moderate Low Nil ^iBllllMJIIIiiiBili 284 High 42 Moderate Low 71 Nil 1 379 High Moderate Low Nil Moderate 26 4 Low 24 Nil High 198 Moderate 20 Low !ISl!l!|l!illlillIIIIIIIIII plfflSlIlli Nil Moderate 2 Low 468 Nil 346 High 247 Moderate 13 Low it Nil Moderate 45 Low 1 Nil 1 High 65 Moderate Low Nil 69 High 46 2 Moderate 0 Low Nil 6 64 High Moderate 104 Low Nil High 324 17 Moderate Low 1 4 Nil  Expected Number of Individuals 138.579 251.712 100.927 112 782 81.335 88.585 107.712 120.368 84.535 37.500 17 542 321.424 7.800 11.677 34.523 130.127 23.721 13.058 62.094 7.138 70.002 738 860 21.013 15.402 12.298 223.288 31.334 6.851 8815 38 319 23.158 13.174 90.349 13.486 16.868 6.642 17.004 98.597 7.784 33 543 28.075 . . j 766 107.734 12.346 29.154  167  2  X Value for Individuals  Observed Number of Records 27 12  1833.512***  660.371***  "13111111 41 12 14 1 '.-2 26  1336.510***  Expected Number of Records 9.636 17.503 7.018  X Value for Records  42.576***  7.842 13.896 15.135 18.403  73.179***  20.565 30.990 13.747 6.431  444 360'**  117.832 50.718***  See Log-likelihood Test Results  78.028***  See Log-likelihood Test Results  2475.424***  See Log- ikelihood Test Results 148  13.056 9.570 7.641  9  138.734  2643.030***  17.885***  1522.095***  See Log-likelihood Test Results 49  25.778 15.579 8.863  48  60.780  24 588'"  28 0 6 8 " '  98.612***  See Log-likelihood Test Results  160.133***  See Log-likelihood Test Results  190.819***  See Log-likelihood Test Results  Species BlllllIl^Bllll^^fcll Ovis canadensis californiana  Coluber  Expected Number of • Individuals Very High"' : 15.3S2 High 15.336 Moderate 15.569 Low 5.153 Very Low 2 447 Nil 13.103 High 127 73.867 Moderate 17 33.904 Low 1 14.175 Nil 5 28.054 High 296 166.090 78 749 Moderates ^ ^ E l l l l l l i Low; ••; • 31 992 Nil 52 169 117 Hkj 78 565 Moderate 33 40.073 Low 8 16.861 Nil 9 31.500 Habitat Rating  constrictor  vtalus viridis oieganus  Pituophis catenifer deserticola  Observed Number of Individuals  2  X Value for Individuals  Observed Number of Records  Expected Number of Records  2  X Value for Records  86.761*** DF=3  See Log-likelihood Test Results  77.837***  109 15 1 4 170  63.526 29.158 12.191 24.126 100 462 47.633 19.351 31 555 65.863 33.594 14.135 26.408  175^693***'  wmmommmm 92 40.780***  31 8 9  G Value for Individuals  Observed Number of Records 5  66.489***  80 608*** DF=2  24.710***  (b) Log-Likelihood Results Habitat Rating  Species  Observed Number of Individuals 26  Expected Number of Individuals 0.796 0.977 0 506 30 720 13.345* 23.706 1.830 172.118  Hign Moderate i p s i s i i i i i f 160.552*** [•Dolichonyx oryzivorus Low IfilllllSil DF=1 Nil High 130 Moderate 41 520.031*** Catherpes mexicanus Low 0 DF=2 Nil 40 Moderate S P P r.hi-cciiiare Test Results Low Ardea herodias Nil High 0 0.030 Moderate 4 0.922 19.336*** Ammodramus Low 3 0.256 DF=2 savannarum 4 9.792 Nil High 1.598 35 Moderate 10.140 • Chondestes 198.965*** Low , 6.770 DF=2 jrammacus ... o Nil " -' 27.492 High 4.281 16 Moderate 0 0.007 29.444*** Numenius americanus Low 0 0.360 DF=1 Nil 6 17.353  Melanerpes  lewis  Falco  mexicanus  Grus  canadensis  High ..-y Moderate Low Nil High Moderate Low Nil Moderate ""llloy^H Nil  •: 'SeeGhi-square Test Results 14 5 3 0  7.754 3.389 4.568 6.289  17.909*** DF=2  lIBlllilip liBiilll  •r  28 0 33 liillleilll illlllli^ I l l l B B 0 4 1 3 j^Rplil ^piiiiii lliliilil ii* 0 0 4 51 14' • " i i s i l S lilllllB 7 5 2 0  4  G Value for Records  : 241  •"""TO"  See Chi-square Test Results  168  Expected Number of Records  :  V  0.296 0.153 309 8.286 ' 14.718 1.136 106.860 2 022 3.027 8 950 0.022 0.670 0.187 7.121 0.938 5 952 • 3.974 16.137 2.919 0.004 0.245 11.831 41.482 " " 562 4.163 19.794 4.934 2.157 2.907 4.002 0.079 0.772 8 149  . 24.094*** : DF=1  257.221*** DF=2 6.593* DF=2  12.461*** DF=2  90.678*** DF=2  20.511*** DF=2  23.823*** DF=2  11.808*** DF=2 12.550*** DF=2  Species Cathartes  aura  Otus kennicottii macfarlanei  Icteria virens  Antrozous  pallidus  Corynorhinus townsendii  Euderma  maculatum  Ovis canadensis californiana  Charina  bottae  Hypsiglena torquata deserticola  Habitat Rating Moderate Low Nil High Moderate Low Nil High Moderate Low Nil High Moderate Low Nil High Moderate Low Nil High  Observed Number of Individuals  Expected Number of Individuals  G Value for Individuals  Observed Number of Records 18  See Chi-square Test Results 1 1 ^ 1 0 1 6 20 7  0.121 0.583 0.132 7.165 1.266 0.399  28  54.441  liiii  1 6.151* DF=2  0 1 4 ^•llllillll  116.037***  See Chi-square Test Results  See Chi-square Test Results  Moderate Low Nil Very High High Moderate Low Very Low Ni! High Moderate Low Nil High Moderate Low  9 1 9 1 6 2 0  7.723 4.268 4.339 3.669 2 229 2.386 1.882  Nil  1  2.503  See Chi-square Test Results  See Chi-square Test Results  15 29 2 0 4 27 0 54 0 111 9 1 4  IllSoli^i* i3iaiSl  169  10.382**  9.342*** DF=2  9 1 9 1 6 2 0 illii* * § •  Expected Number of Records 13.334 2 915 3.751 0.091 0.437 0.099 5.374 0 917 0.289 1 372 39.423 8.741 10.933 4.305 11.021 47 538 53 16.173 13.536 71.086 38.921 4.460 10.532 3.216 3.205 3.253 1.077 0.511 2.738 7.723 4.268 4.339 3.669 2.229 2 386 1 882 2 503  G Value for Records 6.019* DF=2  7.069* DF=2  119.627***  54.654*** DF=2  99 664"* DF=1  61.839***  20.350***  10.382**  9.342*** DF=2  Appendix X. List of Red and Blue Listed plant communities and vascular plant species used in the analyses of surrogacy of priority sites for actual occurrences of species (data provided by the Conservation Data Centre in 1998). Global Rank  Subnational Provincial List Rank  Latin Name  Common Name  ELYMUS SPICATUS - BALSAMHORIZA SAGITTATA  BLUEBUNCH W H E A T G R A S S BALSAM ROOT  -  S2S3  BLUE  F E S T U C A IDAHOENSIS - ELYMUS SPICATUS  FESCUE-BLUEBUNCH WHEATGRASS  -  S2  RED  MARSILEA VESTITA - SCIRPUS AMERICANUS  MARSILEA-SCIRPUS  -  S1  RED  -  S3?  BLUE  P S E U D O T S U G A MENZIESII - PINUS PONDEROSA DOUGLAS-FIR - PONDEROSA / F E S T U C A IDAHOENSIS PINE / IDAHO F E S C U E SALIX EXIGUA - SALIX AMYGDALOIDES  SANDBAR WILLOW - PEACH-LEAF WILLOW  S1  RED  ALLIUM VALIDUM  SWAMP ONION  G4  S1?  BLUE  ARTEMISIA LONGIFOLIA  LONG-LEAVED MUGWORT  G5  S2  RED  A S T E R FRONDOSUS  SHORT-RAYED A S T E R  G4  S1  RED  A S T R A G A L U S SPALDINGII SSP SPALDINGII  SPALDING'S MILK-VETCH  G3?T3?  S1  RED  ATRIPLEX A R G E N T E A SSP A R G E N T E A  SILVERY O R A C H E  G5T5  S1  RED  BOTRYCHIUM PARADOXUM  TWO-SPIKED  G2  S1  RED  CAMISSONIA ANDINA  G4  S1  RED  CAREX C O M O S A  ANDEAN EVENING PRIMROSE BEARDED S E D G E  G5  S1?  BLUE  CAREX SAXIMONTANA  ROCKY MOUNTAIN S E D G E  G5  S2S3  BLUE  CAREX XERANTICA  DRY-LAND S E D G E  G5  S2S3  BLUE  CASTILLEJA MINOR SSP MINOR COREOPSIS ATKINSONIANA  ANNUAL PAINTBRUSH  S1  ATKINSON'S COREOPSIS  G5T5 G5  S1  RED RED  CRYPTANTHA CELOSIOIDES  C O C K S C O M B CRYPTANTHA  G5  S1  RED  C U S C U T A PENTAGONA  FIVE-ANGLED DODDER  G5  S2S3  BLUE  C Y P E R U S ERYTHRORHIZOS  RED-ROOTED C Y P E R U S  G5  S1  RED  DELPHINIUM BICOLOR  MONTANA LARKSPUR  G4G5  S2S3  BLUE  ELEOCHARIS ATROPURPUREA  PURPLE SPIKE-RUSH  G4G5  S1  RED  ELODEA NUTTALLII  NUTTALL'S WATERWEED  G5  S2S3  BLUE  ERAGROSTIS PECTINACEA  TUFTED LOVEGRASS  G5  S1  RED  ERIOGONUM STRICTUM SSP PROLIFERUM  STRICT BUCKWHEAT  G5T?  S1  RED  GAURA COCCINEA  S C A R L E T GAURA  G5  S1  RED  GAYOPHYTUM RAMOSISSIMUM  HAIRSTEM GROUNDSMOKE  G5  S1  RED  GENTIANA AFFINIS  PRAIRIE GENTIAN  G5  S1?  BLUE  GILIA SINUATA  SHY GILIA  G5  SH  RED  HUTCHINSIA PROCUMBENS  HUTCHINSIA  G5  S1  RED  IPOMOPSIS MINUTIFLORA  SMALL-FLOWERED IPOMOPSIS  G2G3  S2  RED  JUNCUS REGELII  REGEL'S RUSH  G5  S2S3  BLUE  LEPIDIUM DENSIFLORUM VAR PUBICARPUM  PRAIRIE P E P P E R - G R A S S  G5T4  S1  RED  LINDERNIA DUBIA VAR ANAGALLIDEA  FALSE-PIMPERNEL  G5T4  S2  RED  MIRABILIS HIRSUTA  HAIRY UMBRELLAWORT  G5  S1  RED  MUHLENBERGIA GLOMERATA  MARSH MUHLY  G5  S2S3  BLUE  O R T H O C A R P U S BARBATUS  GRAND C O U L E E OWL-CLOVER  G2G4  S1  RED  P E C T O C A R Y A PENICILLATA  WINGED C O M B S E E D  G5  S1  RED  PHYSARIA DIDYMOCARPA VAR DIDYMOCARPA  COMMON TWINPOD  G5T4  S2S3  BLUE  G?T?  S1?  BLUE  MOONWORT  POLEMONIUM C A E R U L E U M SSP AMYGDALINUM TALL JACOB'S-LADDER POLYGONUM PUNCTATUM  DOTTED SMARTWEED  G5  S1?  BLUE  SCIRPUS FLUVIATILIS  RIVER BULRUSH  G5  S1?  BLUE  170  Appendix X. List of rare invertebrate species included in the analyses of surrogacy of priority sets of sites for actual locations of species (unpublished data provided by Dr. G. Scudder in 2000). PHYLUM Annelida  ORDER Rhynchobdellae  FAMILY Piscicolidae  GENUS Piscicola  SPECIES punctata  SPECIES AUTHOR (Verrill)  Arthropoda  Acariformes  Cymbaeremaeidae  Scapuleremaeus  kobauensis  Behan-Pelletier  Arthropoda  Acariformes  Eremaeidae  Eueremaeus  michaeli  Behan-Pelletier  Arthropoda  Acariformes  Unionicolidae  Koenikea  sp.n.  Arthropoda  Araneae  Agelenidae  Agelenopsis  Oklahoma  (Gertsch)  Arthropoda  Araneae  Agelenidae  Cicurina  sp. near intermedi  Chamberlin & Ivie  Arthropoda  Araneae  Clubionidae  Agroeca  pratensis  Emerton  Arthropoda  Clubionidae  Castianeira  Arthropoda  Araneae Araneae  Clubionidae  Castianeira  alteranda mimula  Gertsch Chamberlin  Arthropoda  Araneae  Dictynidae  Dictyna  borealis cavernosa  Jones  Arthropoda  Araneae  Dictynidae  Dictyna  coloradensis  (Chamberlin)  Arthropoda  Araneae  Dictynidae  Dictyna  reticulata  Gertsch & Ivie  Arthropoda  Araneae  Dictynidae  Dictyna  terrestris  Emerton  Arthropoda  Araneae  Dictynidae  Mallos  niveus  0. Pickard-Cambridge  Arthropoda  Araneae  Erigonidae  Eperigone  dentosa  0. Pickard-Cambridge  Arthropoda  Araneae  Erigonidae  Sougambus  bostoniensis  (Emerton)  Arthropoda  Araneae  Gnaphosidae  Drassyllus  saphes  Chamberlin  Arthropoda  Araneae  Gnaphosidae  Herphyllus  propinquus  (Keyserling)  Arthropoda  Araneae  Gnaphosidae  Micaria  foxi  Gertsch  Arthropoda  Araneae  Gnaphosidae  Micaria  laticeps  Emerton  Arthropoda  Araneae  Gnaphosidae  Micaria  utahna  Gertsch  Arthropoda  Araneae  Gnaphosidae  Nodocion  voluntarius  (Chamberlin)  Arthropoda  Araneae  Linyphiidae  Sougambus  bostoniensis  (Emerton)  Arthropoda  Araneae  Philodromidae  Ebo  parabolis  Schick  Arthropoda  Araneae  Pholcidae  Psilochorus  sp. nr. hesperus  Gertsch & Ivie  Arthropoda  Araneae  Salticidae  Habronattus  hirsutus  (Peckham & Peckham)  Arthropoda  Araneae  Salticidae  Habronattus  sansoni  (Emerton)  Arthropoda  Araneae  Salticidae  Pellenes  shoshoensis  Lowrie & Gertsch  Arthropoda  Araneae  Salticidae  Phidippus  purpuratus  Keyserling  Arthropoda  Araneae  Salticidae  Synageles  leechi  Cutler  Arthropoda  Araneae  Thomisidae  Misumenops  serrensis  Schick  Arthropoda  Araneae  Thomisidae  Thanatus  altimontis  Gertsch  Arthropoda  Coleoptera  Apionidae  Apion  proclive  LeConte  Arthropoda  Coleoptera  Melanophila  californica  Van Dyke  Arthropoda  Coleoptera  Buprestidae Curculionidae  Ceutorhynchus  opertus  W.J. Brown  Arthropoda  Coleoptera  Curculionidae  Cleonidius  longinasus  R.S. Anderson  Arthropoda  Coleoptera  Curculionidae  Tychius  semisquamosus  LeConte  Arthropoda  Coleoptera  Histeridae  Teretrius  montanus  Horn  Arthropoda  Coleoptera  Staphylinidae  Xylodromus  depressus  (Gravenhorst)  Arthropoda  Coleoptera  Tenebrionidae  Eleodes  extricatus extrica  (Say)  Arthropoda  Coleoptera  Tenebrionidae  Eleodes  nigrinus nigrinus  LeConte  Arthropoda  Diptera  Anthomyiidae  Pegomya  setibasis  Huckett  Arthropoda  Diptera  Asilidae  Comantella  pacifica  Curran  Arthropoda  Diptera  Asilidae  Crytopogon  ablautoides  Melander  Arthropoda  Diptera  Asilidae  Efferia  okanagana  Cannings MS name  Arthropoda  Diptera  Asilidae  Machimus  vescus  (Hine)  Arthropoda  Diptera  Asilidae  Myelaphus  lobicornis  (Osten Sacken)  Arthropoda  Diptera  Mydidae  Nemomydas  patherinus  (Gerstacker)  Arthropoda  Diptera  Simuliidae  Prosimulium  constrictistylum  Peterson  Arthropoda  Diptera  Tipulidae  Tipula  imbellis  Alexander  Arthropoda  Ephemeroptera  Baetidae  Baetes  parallelus  Banks  Arthropoda  Ephemeroptera  Leptophlebiidae  Leptophlebia  gravastella  (Eaton)  Arthropoda  Ephemeroptera  Siphlonuridae  Ameletus  sparsatus  McDunnough  Arthropoda  Grylloptera  Oecanthidae  Oecanthus  californicus  Saussure  Arthropoda  Heteroptera  Anthocoridae  Lyctocoris  rostratus  Kelton & Anderson  Arthropoda  Heteroptera  Cimicidae  Hesperocimex  coloradensis  List  171  PHYLUM Arthropoda  ORDER Heteroptera  FAMILY Lygaeidae  GENUS Eremocoris  SPECIES canadensis  SPECIES AUTHOR Walley  Arthropoda  Heteroptera  Lygaeidae  Usinger  Heteroptera  Lygaeidae  Gastrodes Heterogaster  intermedius  Arthropoda  behrensii  (Uhler)  Arthropoda  Lygaeidae  Malezonotus  grossus  Van Duzee  Arthropoda  Heteroptera Heteroptera  Miridae  Ceratocapsus  downsei  Knight  Arthropoda  Heteroptera  Miridae  Chlamydatus  brevicornis  Knight  Arthropoda  Heteroptera  Miridae  Chlamydatus  schuhi  Knight  Arthropoda  Heteroptera  Miridae  Parthenicus  brooksi  Kelton  Arthropoda  Heteroptera  Miridae  Parthenicus  sabulosus  Van Duzee  Arthropoda  Heteroptera  Miridae  Trigonotylus  antennatus  Kelton  Arthropoda Arthropoda  Heteroptera Heteroptera  Miridae Nabidae  Trigonotylus Hoplistoscelis  brooksi heidemanni  Kelton (Reuter)  Arthropoda  Heteroptera  Rhopalidae  Aufeius  impressicollis  Arthropoda  Heteroptera  Saldidae  loscytus  politus  Stal (Uhler)  Arthropoda  Homoptera  Cicadellidae  Aceratogallia  compressa  Arthropoda  Homoptera  Cicadellidae  Aceratogallia  okanagana  Hamilton (MS name)  Arthropoda  Homoptera  Cicadellidae  Aceratogallia  zacki  Hamilton (MS name)  Arthropoda  Homoptera  Cicadellidae  Ballana  callipera  De Long  Arthropoda  Homoptera  Cicadellidae  Hebecephalus  crassus  (De Long)  Arthropoda  Homoptera  Cicadellidae  Hyliaus  oregonensis  (Baker)  Arthropoda  Homoptera  Cicadellidae  Idiocerus  indistinctus  Hamilton  Arthropoda  Homoptera  Cicadellidae  Latalus  mundus  Beamer & Tuthill  Arthropoda  Homoptera  Membracidae  Platycotis  quadrivittata  (Say)  Arthropoda  Hymenoptera  Andrenidae  Andrena  fulvicrista  Viereck  Arthropoda  Hymenoptera  Andrenidae  Andrena  trizonata  Arthropoda  Hymenoptera  Apidae  Bombus  vosnesenskii  (Ashmead) Radoszkowski  Arthropoda  Hymenoptera  Bethylidae  Parasierola  breviceps  (Krombein)  Arthropoda  Hymenoptera  Chrysididae  Chrysis  montana  Aaron  Arthropoda  Hymenoptera  Chrysididae  Chrysis  rivalis  Bohart  Arthropoda  Hymenoptera  Chrysididae  Cleptes  speciosus  Aaron  Arthropoda  Hymenoptera  Chrysididae  Holopyga  hora  Aaron  Arthropoda  Hymenoptera  Crabronidae  Ectemnius  dilectus  (Cresson)  Arthropoda  Hymenoptera  Eumenidae  Euodynerus  auranus albivestri  (Bohart)  Arthropoda  Hymenoptera  Eumenidae  Euodynerus  cockerelli  (Cresson)  Arthropoda  Hymenoptera  Eumenidae  Pterocheilus  morrisoni  Cresson  Arthropoda  Hymenoptera  Halictidae  Dialictus  albohirtus  (Crawford)  Arthropoda  Hymenoptera  Larridae  Miscophus  evansi  (Krombein)  Arthropoda  Hymenoptera  Larridae  Tachysphex  similis  Rohwer  Arthropoda  Hymenoptera  Megachilidae  Anthidium  palliventre  Cresson  Arthropoda  Hymenoptera  Megachilidae  Anthocopa  copelandica  Cockerell  Arthropoda  Hymenoptera  Megachilidae  Heriades  cressoni  Michener  Arthropoda  Hymenoptera  Megachilidae  Megachile  gentilis  Cresson  Arthropoda  Hymenoptera  Megachilidae  Megachile  subnigra subnigra  Cresson  Arthropoda  Hymenoptera  Megachilidae  Osmia  pikei  Cockerell  Arthropoda  Hymenoptera  Megachilidae  Osmia  texana  Cresson  Arthropoda  Hymenoptera  Megachilidae  Osmia  unca  Michener  Arthropoda  Hymenoptera  Megachilidae  Stelis  montana  Cresson  Arthropoda  Hymenoptera  Mutillidae  Myrmosa  bradleyi  Roberts  Arthropoda  Hymenoptera  Mutillidae  Odontophotopsis  erebus  (Melander)  Arthropoda  Hymenoptera  Mutillidae  Pseudomethoca  athamus  (Fox)  Arthropoda  Hymenoptera  Mutillidae  Pseudomethoca  bequaerti  Mickel  Arthropoda  Hymenoptera  Nyssonidae  Didineis  nodosa  Fox  Arthropoda  Hymenoptera  Nyssonidae  Stictiella  tuberculata  (Fox)  Arthropoda  Hymenoptera  Nyssonidae  Synevrus  sp.  Arthropoda  Hymenoptera  Pemphredonidae  Ammoplanellus  apache  (Pate)  Arthropoda  Hymenoptera  Pemphredonidae  Ammoplanellus  lenape  (Pate)  Arthropoda  Hymenoptera  Pemphredonidae  Diodontus  leguminiferus  Cockerell  Arthropoda  Hymenoptera  Philanthidae  Eucerceris  vittatifrons  Cresson  Arthropoda  Hymenoptera  Pompilidae  Ageniella  accepta  (Cresson)  Arthropoda  Hymenoptera  Pompilidae  Ageniella  grisea  Townes  172  .  Hamilton (MS name)  PHYLUM Arthropoda Arthropoda Arthropoda Arthropoda Arthropoda Arthropoda Arthropoda Arthropoda  ORDER Hymenoptera Hymenoptera Hymenoptera Hymenoptera Hymenoptera Hymenoptera Hymenoptera  FAMILY Pompilidae Pompilidae Sierolomorphidae Sphecidae Sphecidae Sphecidae Tiphiidae  GENUS Agenioideus  SPECIES birkmanni  SPECIES AUTHOR (Banks)  Anoplius Sierolomorpa Ammophila Ammophila  depressipes nigrescens aberti extremitata  Banks Evans Haldeman  Podalonia Paratiphia  sonorensis ephippiata banksaria  Cresson (Cameron) Allen Sperry McDunnough  Geometridae Geometridae  Chlorosea Meris  Noctuidae Noctuidae Noctuidae Riodinidae  Apamea Copablepharon  centralis absidum  Heliothis Apodemia  paradoxus mormo mormo  Hemileuca Sialis  nuttalli velata  Neuroptera (sensu s Neuroptera (sensu s  Saturniidae Sialidae Chrysopidae Hemerobiidae  Meleoma Micromus  schwa rzi  Arthropoda Arthropoda Arthropoda  Neuroptera (sensu s Odonata  Hemerobiidae Coenagridae Libellulidae  californicus vivida collocata  Banks Hagen  Odonata  Sympherobius Argia Erythemis  Arthropoda  Orthoptera  Acrididae  Dissosteira  spurcata  Saussure  Arthropoda  Plecoptera  Capniidae  Bolshecapnia  milami  (Nebeker & Gaufin)  Arthropoda  Plecoptera  Cultus  Arthropoda  tostonus elongatus  (Ricker) (Hagen)  Arthropoda  Plecoptera Raphidioptera  Perlodidae Perlodidae  Raphidioptera Solpugida  Inocellidae Raphidiidae Eremobatidae  Negha Agulla  Arthropoda  Solpugida  Eremobatidae  Eremobates Eremobates  inflata crotchi sp.n.1  (Hagen)  Arthropoda Arthropoda Arthropoda  Solpugida  Eremobatidae  Hemerotrecha  denticulata  Muma  Arthropoda Mollusca  Solpugida Basommatophora  Eremobatidae Lymnaeidae  Hemerotrecha Fossaria  sp.n. truncatula  (Muller)  Mollusca  Basommatophora  Physidae  Physella  propinqua nuttalli  (Lea)  Arthropoda Arthropoda Arthropoda Arthropoda Arthropoda Arthropoda Arthropoda Arthropoda Arthropoda  Lepidoptera Lepidoptera Lepidoptera Lepidoptera Lepidoptera Lepidoptera Lepidoptera Megaloptera  Isogenoides  173  suffusaria  subanticus  (Smith) (Harvey) Ert (Felder & Felder) (Strecker) Ross (Banks) (Walker)  (Hagen)  Banks  sp.n.2  « -  •§  ro -5 cz  — °  CD CO  O O  .2 ro •e 0  co  CM  T=>  0)  o cz  X  OJ  o o  o o  CO  o o  o o  CM  d  r— o o co co co ho  d  o d d  d  CM CO CT) CD CM  d  d  T CD •<3-  O) O)  O) O) CO  d  d  o CO d d  d  E ^ x> o <u  > 5=  co  2 rooP'  "D  0  CD  CM  CO CD CM CM  CO CM  O  CM  £ ^  CN CM  CN  CN CN  »— 'I  •s  -3 0 t± CC CO XJ  o o d  0  CO IN 0 2 E  o < ^  o o d  CL  CL £  co — co  O O  O O  •<3-  O  O  t E ^  o  CD  CD CD  o  CO CO  O O  O O  O O  O  O  CM  CD CD in in CD CD  CD CD f -"3-  0 CO CL 0  "te 0 o— 0 b £ 0 CO 0  ro < o | 8 i= TR ro ro jP, x o. T3 TJ 0 0 J= CO > EP a I  o  O  5 5  o  CM 00  CM 00  o o  •<- r-- -^r -3" co CO CO T-  o o O OJ  r--  o o  O O  o o o> 00 o o  CD  r-.  o o  o o d  o o d  CO OJ  CO OJ  CO  0 CL  £  w  CO  X  CO CO cr: O i + co  o  T3 0  g.  c^  Z  CCo LU +  0 cC0 0 O CO CL 0 O CC  CO C0 O + cr:  CO  o  + cr:  LU  cr:  LU  LU  XJ  co  0 CO  ro a)  IT ro < T3  0  * -—»  O 0  o  DL  JZ. CO 0 cz Z .2 0 e  £ u  0 > co 0  ct: ro  o 0  -»—;  0  ro E  1  ro E  o  o  I CN  E  E  CO 0 0 T3 CZ zs  0 >  CN  CM  CM  174  0  0 ^—<  CO  CO  o  cc I  CO  + o cr: + LU cr: UJ  5  0 0 X0 "ro £ o . ° Z .3 0 p ro 0 o o ro roc0o E XJ si Oo (i0cz 0  O  a:  2  c  o o o  O  (/i  1/5  o o o  i»  m  ~ £ E fV CO cz t5 0 CNJ LU ^ ra ro 0 ro ^ co w J ro _ Tcz 4J C § ro E O — ZJ I— "O o 0 0 2 E a. 0 JZ!  o o co oo co o O  £> ro  0  "° — x l o ^ . § > 0 o 0 CO CO <g S Ct: co .E 2 ro CO I CO ro -o £-<co ro 0 ^ 0 ro O (O + 3 rofi-g cz ra? &0cz cn ro 3 ra J£ 0 by. - CO ro 0 ro 0 coQ- -a 0  o o o o co  1_  ra _ T J CI T o,  " co J= TD cz 0 ro cz g > ro zs X°J ro w — w" <= £ c | 2 0 ^ o 2f -o 0 0 ° O) roj= -o 2 x » « r CO S 0 x "O > 0 0 0 TJ ^ 0<N ^ ra -o"  E  TJ ^ W CZ -—• XJ 0  +  l N  o o o o o o co co co CO CO CO co CO CO CO CO CO  CD CO CM CM CO  Ol O) O)Ol O)  •2 % TJ CD 0 0 2 Q) co SCO T J 0 ^ co § =3 0 £ o ™ Q) cz n E c7o T J CO to — ° cr rz£0 0o cz o "cz'^ • Tz co cu ro O 0 £ CO jo o o < 0 O 0 co co ro ri-D £ x 0 0 0 0"O « ^ 1 ^ m CO CD Ez  CL CL <  o  l  •-  I ' l l CD ra co E c O  o o  LL c5<  .2 <u ^ ^_ 2 • e g o ~ CD o *: in  o  o o  O O  c  ~  -Q  CZ CO .2 0 0 t Dl > o != 0 O r- o  o c o  '•4—>  O  0  o L  QJ I  ^—  1  CO 0 0  > o  E XL  0 ro  CO 0 0  XJ  cz ZJ  E XL O  rz co n + J CJ  .2  CD OJ  ti  OJ >  o ;= < u  St 2  CL  O  O CO  <  'fe $  CD O  O  C D O O O O )  o> a> r— co N s o ci d o d o  CD  O CD  i  CD O CO CD  o  CO N O)  CD O ) CO CN CD  C OS 00 N C OCD  (D S S CN CO 0 0 CD C D  CD  in  CD  CO CD  o  o  o  O  d o d o  OJ  CN CN  o  o  o  O  O  o  CM  o  •tf  O CD  <f  O CD C DCO  d o d o  o  O  (U ,  •*-•  —  O  O O CD C D  O CD  co  CD  co  o  o  CN  CM C N  o  -a- -3- ^  CD CD CO 0 0 CO  d o d o  C Df co 0 0  C DCD  o  t  CD C O C O O O) O)  O  o  O CMCM CM  O  O  O O O CN CN CM C N CM CN  CD CO ' — '  o Q  -S  E  CZ -—•  •5 T3  i CO  w  CD * 5 Ctr oo CD  a> E <-> CD  ^  2  CD  CD  ^  CD  t  E^S <D co  o o  o o  O O  O O  O O  00  00 CO O) O ) CN CM  •sr  -*t  ^  ffi CM  o  O  CD 0)  O  0 0 CO  CN CD CO  o  o  o o  o o o o o CN  o o o o o o o d d  O O  O) O ) O) ^ ^J-  O O  o  O O  o o o  CD C D  O O  o d d  o  o  o  o  o o o 5z; o  o o o o  CD CD  CM  Q _ CD  o o  °5 ! O  r-  Z  C  o o  E  0 01  ro  3!  o o  C O CO O) O) O)  oi oi  CM  Ol  CO  o o  o o  -^J-  o o  CM  o o  8 g 8 8 2 ?? ?  CN  i - 0 ) O) i - CO 0 0  o o o o o o CM  o o Q  II  CM  CN  CD  CH c  "co x  CD  CO  LU T3  cl  2  co  o  cr: + cr:  CD  £ CD CD  co co  cr: X  o +  E ZJ  E  CO  cr: LU  Ct: UJ  o  + cr:  CO CH  +o  UJ  CH + m  cr: LU  •4—'  CD  UJ  E zs E  CO  CD  CD  CO  CH CD  I  CO  LU  CO CD O CO Q . CD  P  CD  CO  CD 5  CO CD  i  CU  > o  E CM  jD "(D  a)  ro E  CO  4—>  CO CD  CD CZ  L_  CD T3 CZ Z3  E  CD >  o  E  XL  CM  175  CO CD  o  

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.831.1-0090312/manifest

Comment

Related Items