Focal Species, Biodiversity and Island Biogeography on the Central Coast of British Columbia, Canada: a GIS Approach

Patricia L. Swan (1), Shelley M. Alexander (1), Paul C. Paquet (2), Chris T. Darimont (3)

(1) Department of Geography, University of Calgary, Alberta (CA)
(2) Faculty of Environmental Design, University of Calgary, Alberta (CA)
(3) Department of Biology, University of Victoria, British Columbia (CA)

The integration of Remote Sensing (RS) and Geographic Information Systems (GIS) in coastal wildlife habitat analysis and biodiversity assessment is an emerging field. This study used RS and GIS to examine the relationships among focal species, biodiversity, island biogeography and ocean characteristics along the British Columbia (BC) Central and North Pacific Coast in Canada.

The study region encompasses the island archipelago between the northern tip of Vancouver Island and Prince Rupert, BC (550 37' N, 1290 48'W), and covers approximately 29 700 km2, of which 19 300 km2 is land (Darimont and Paquet 2001). The area is bounded by the Coast Mountains to the east and the Pacific Ocean to the west. Coastal temperate rainforest dominates, forming part of the most extensive habitat of this type left in the world (MacDonald and Cook 1996). Climate is temperate and wet; most areas receive upwards of 350 cm of annual precipitation.

There are few human settlements consisting primarily of First Nations people. Human disturbance and development is minimal, resulting in a nearly pristine landscape. However, mainland logging has occurred and is proposed for areas throughout the study region. Oil and gas exploration also is expected to commence in the near future. The study region, combined with southeastern Alaska, supports the highest endemic species concentration for the temperate rainforest region of Pacific North America (Cook and MacDonald 2001). This region offers a unique and diverse landscape for the integration of RS and GIS in fragmentation research.

GIS and RS applications have been used to predict areas of species and biodiversity importance (Debinski et al. 1999). More recently, GIS has been applied in habitat selection modelling of grizzly bear (Ursus arctos)(McLellan and Hovey 2001), caribou (Rangifer tarandus)(Apps et al. 2001), and Rocky Mountain carnivores (Carroll et al. 2000). GIS has the ability to integrate empirical and spatial data used to build predictive models, which are particularly useful for conservation decisions. Habitats that support focal species and which are biologically diverse can be identified and protected through conservation initiatives.

Gray wolves (Canis lupus) are suitable focal species for this region as they display indicator, keystone and umbrella species attributes. They have been found to be sensitive to human modifications of landscapes and avoid developed and high road-density areas (Weaver et al. 1996). Human-related mortalities of coastal wolves in the study area are estimated to be 2.3% of the annual population (Darimont and Paquet 2001); however, this rate would likely increase if human disturbance of the landscape intensified. Coastal wolves are a far ranging species as evidenced by their presence on 40 of 42 islands surveyed in the area (Darimont and Paquet 2001). They are capable of swimming between islands but may be limited by island to island distance, wind, and water currents (Darimont and Paquet in press). From the limited research conducted, coastal wolves are thought to prefer larger, less isolated islands that are not topographically complex, which may be a function of prey density. Their presence is also expected near watersheds where they prefer to den, and on islands that support deer habitat.


Presence/absence data of focal species can be integrated easily into GIS analysis and used in conjunction with island biodiversity data to identify areas of biological importance in the coastal region. In this study, wolf presence will be determined by scat occurrence collected on random transects in spring, summer and fall between 1999 and 2003 (Darimont and Paquet 2001). Diversity of mammalian species was georeferenced to islands, based on surveys reported by Craig (1990).

Landscape biophysical data, including habitat characteristics and configuration metrics, were extracted from LANDSAT-7, a digital elevation model (DEM), and BC TRIM data. Solar insolation was derived from the DEM by use of ArcView Solar Analyst (ESRI 2003). Ocean tide and current characteristics were derived from Canadian Hydrographic Services digital tidal and marine charts using marine navigation software (Nobeltec 2003). Digital bathymetry data provided by Natural Resources Canada was used to develop a seafloor depth model. Prey density measures were interpolated from an existing black-tailed deer transect database (C. Darimont pers. comm.). Anthropogenic disturbances including forest cutblocks, roads, and development were extracted from LANDSAT-7, National Topographic Series (NTS) digital maps and BC TRIM data.

We postulated that distances between landmasses, characteristics of ocean waterways, physiographic features, and human disturbance influence biodiversity and focal species density and distribution. We examined these relationships at the watershed scale. Specifically, we compared use-density of wolves and biodiversity counts to a series of independent variables including island size and shape, distance to nearest island, species composition, average current speed, average sea depth, tidal flux, vegetative heterogeneity, greenness, wetness, percent canopy cover, topographic complexity, aspect (eastness, northness), slope, and elevation. Human disturbance variables included percent forest cutblocks, as well as road density and distance to residential areas.

Logistic regression and Akeike's Information Criterion (AIC) were used to assess which of the previous characteristics best predict focal species presence and biodiversity. A Receiver Operating Characteristic (ROC) curve was used to validate predictive models, using a subset of data that did not contribute to model development. We predict that larger, less isolated islands that are more productive and less disturbed will support larger populations of focal species and greater species diversity. Moreover, we anticipate that topographic features, measured at the watershed level, will limit wolf distribution and biodiversity. We also expect that average tidal displacement, current speed and sea depth of the intervening matrix may limit island colonization, resulting in detectable differences in endemic species diversity and focal species presence on islands. Predicting areas of focal species importance and isolating centres of high biodiversity are imperative for conservation decisions for the region, which are imminent due to increasing development pressure.

References: