BC’s Inland Rainforest – Conservation and Community

Conference Proceedings

A Conservation Area Design for the Inland Temperate Rainforest Based on Focal Species

Lance Craighead 1, Thomas Olenicki 2, Brent Brock 3, Justin Williams 4, and Baden Cross 5.

A Conservation Area Design for the Inland Temperate Rainforest of North America


Temperate rainforest, and in particular the inland temperate rainforest, has immense conservation value. The rainforest is also faced with increasing threats to biodiversity. We describe a new approach that relies primarily upon the use of focal species to define areas of greatest conservation need: both for secure habitat and movement of those species. Our approach assumes that almost all of the critical areas for maintenance of biodiversity, and the structure and function of ecological processes, can be identified and protected under the umbrella of habitat for an appropriate suite of focal species.

A Conservation Area Design (CAD) is a broad-scale conservation plan that utilizes a science-based architecture for identifying and prioritizing areas for sustainable conservation Sanjayan and others 2000). These types of plans are also called Reserve Designs. The overall objective is to serve four well-accepted goals of conservation: 1) represent ecosystems across their natural range of variation; 2) maintain viable populations of native species; 3) sustain ecological and evolutionary processes within an acceptable range of variability; and 4) build a conservation network that is resilient to environmental change. Much current conservation planning includes the generally accepted elements of representation, special elements, and focal species analysis (Noss and others 1997, 2001). However, most current CADs do not adequately identify core areas sufficient for long-term viability of focal species or networks of habitat for movement.

We feel that our Inland Rainforest CAD in particular provides adequate guidelines to maintain viable populations of native species. We feel that this approach meets the needs of focal species better than previous conservation plans that we have built upon. In so doing, this CAD should also adequately meet the other three goals of conservation.

Forest carnivores and other wide-ranging species such as grizzly bear (Ursus arctos), wolf (Canis lupus), wolverine (Gulo gulo), lynx (Lynx canadensis), cougar (Felis concolor), and woodland or mountain caribou (Rangifer tarandus caribou) all need large landscapes to maintain viable populations. Among the shortcomings of most current conservation planning efforts is the fact that they do not prioritize sufficient habitat necessary to maintain viable populations and metapopulations, and they do not address or identify adequate habitat for wildlife movement; or connectivity.

We began by accessing the best available data and applying computer habitat suitability models as a first step to identify core and connectivity habitats. The overall study area extends from Prince George in British Columbia, Canada, to the Clearwater River in northern Idaho, United States (Figure1). It encompasses the lowlands comprised of Interior Cedar-Hemlock forest as described by the Province of British Columbia Ministry of Forests (DeMarchi 1996). At higher elevations this area is comprised primarily of Engelmann spruce-subalpine fir forest and alpine tundra. The boundaries of this area include pockets of sub-boreal spruce forest, ponderosa pine forest, and montane spruce.

Study Area

For the initial phase of this project we restricted the analysis to the Canadian portion of Interior Cedar-Hemlock forest region that is within the Canadian Rocky Mountains (CRM) Ecoregional Plan boundary because equivalent datasets were not available for areas outside that analysis area. For the second phase of this project we conducted a similar analysis of the US portion. Our study area remained within the CRM Ecoregional boundary although significant cedar-hemlock is found further south, particularly along the Selway River.

For the third phase of this project we used coarser resolution data and a slightly different habitat modeling approach over the entire region.

Figure 1. The study area


We emphasized focal species analysis using habitat suitability models of grizzly bear, wolf, wolverine, lynx, cougar, and woodland caribou. There were adequate data and scientific understanding to develop habitat suitability models for these species (Carroll and others 2001, 2003; Lambeck 1997; Roberge and Angelstam 2004). Models identified habitat concentration areas (cores). Core area results were combined for all focal species to produce a Composite Core model. Subsequently we addressed habitat suitability for aquatic and avian species.

During the first phase of the project, landscape and habitat suitability characteristics were evaluated for the Interior Cedar-Hemlock forest region in terms of Land Cover Class, Human Population Density, Road Density, Slope and Elevation for each species. We then used the least-cost-path connectivity methodology of Singleton and Lemkuhl (1999, 2000) and Singleton and others (2002, 2003) to identify probable movement corridors between core areas. We thus prioritized core protected areas and zones of interconnection that should be sustainably managed: critical ecological foundations that have been inadequately addressed in previous planning efforts.

To address aquatic focal species we utilized available BC government data on salmon (Salmo spp.) escapement and distribution. Native salmon species found in the study area are Coho (Onchorhynchus kisutch), Chinook (Onchorhynchus tshawytscha), Steelhead (Onchorhynchus mykiss), Sockeye (Onchorhynchus nerka), and Pink Salmon (Onchorhynchus gorbuscha).


We adapted the TNC/NCC representation analysis (coarse filter) which resulted from the Canadian Rockies Ecoregional (CRM) Assessment (Rumsey and others 2003) for an initial prioritization of conservation areas (Figure 2).

Figure 2. An initial prioritization of conservation areas

For the second phase of the project we followed the same approach but applied them to the Interior Cedar-Hemlock forest of the US in Idaho, Montana, and Washington states using different landcover datasets. Core area results were combined for all focal species to produce a Composite Core model (Figure 3).

Figure 3. A composite core model

Subsequently a linear programming model was developed (Williams et. al. 2004, Williams 2008) to minimize the total cost of simultaneously setting aside a required threshold of percentage of total area for each focal species (Figure 4).

Figure 4. A spatial optimization model

For the third phase of the project we addressed differences in landcover classifications between BC data and US data using a coarser-resolution data set: Shining Mountains, which covered the entire study area at 1 km2 resolution (DeMarchi 1996). Habitat suitability models were run for each of the focal species. Core area results were combined for all focal species to produce a Composite Core model (Figure 5).

Figure 5. The final composite core model

Next, an optimal solution of core areas was developed using a range of thresholds: 80% of Grizzly Core Habitat, 70% of Wolverine Core Habitat, 70% of Wolf Core Habitat, 50% of Cougar Core Habitat, 50% of Lynx Core Habitat, 100% of Caribou Core (Canada), and 80% of Caribou Core (U.S.) (Figure 6).

For comparison, another optimization solution solved for 30% of each species’ core habitat (Figure 7). This threshold, 30%, is in the general range of conservation plans such as the CIT ecological science assessment. Another solution solved for 30% of each species’ core habitat but included 100% of caribou habitat in Canada and 80% of caribou habitat in the US (Figure 8). Compared to Figure 6, these 30% results provide much less area for the persistence of large carnivores, and the total area is fragmented into many smaller patches of contiguous habitat.

Figure 6. Optimized core areas

Figure 7. Optimized 30% of all species cores

Figure 8. Optimized 30% of all species cores plus 100% caribou in BC

We then used least-cost-path connectivity methodology to identify probable movement corridors between core areas. We conducted a general analysis of the amount of unvegetated areas within the entire ITR and also within optimized core areas. 17,847 km2 or 17.44% of the cores was non-habitat. Thus, the actual vegetated land area of habitat cores comprised 42.44% of the ITR. We then developed map layers of the highest quality avian habitat (50% of the habitat above the mean bird richness within the ITR) from datasets of bird species richness (Jones 2002). All of the highest quality avian habitats were included within our terrestrial habitat optimal solutions.

Finally we developed a map layer of the highest quality aquatic habitat based upon Aquatic Integrity Area (AIA) scores that have been derived for most of the U.S. portion of the ITR study area (Hitt and Frissell 2000, and Oechsli, L. and C. Frissell. 2003). We used percent roadless areas within watershed basins to estimate aquatic integrity across the entire ITR. Roads are good predictors of aquatic habitat quality (USDA Forest Service 1997). A roadless score was calculated for each watershed basin in the ITR following Hitt and Frissell (2000).

A small percentage of the highest quality aquatic habitat (best 50%) was outside of the terrestrial cores. These drainages added an additional 4,584.04 km2, or 2.35%, to the final CAD solution (Figure 9).. With this inclusion the actual vegetated land area of the optimized habitat cores plus adjoining high quality watersheds, comprised 44.79 % of the ITR.

Figure 9. Watersheds within cores plus additional high quality watersheds

A final CAD map will consist of Figure 6 plus the additional watersheds in Figure 9. Inclusion of special elements will involve adding the Tier 1 and Tier 2 solutions from the CRM Ecoregional Plan (Rumsey and others 2003). In summary, the CAD is just a broad blueprint. Similar mapping projects at a finer scale are now needed to make conservation decisions, accompanied by analysis on-the-ground in the real landscape. The broad-scale CAD type of analysis should help to put local conservation values in perspective and add support for local efforts by showing that a given area is part of an important core or corridor. The results of this CAD should constitute a defensible scientific basis for implementation of conservation planning and for campaigns to facilitate such implementation.

Literature Cited

Carroll, C.; Noss, R. F.; Paquet, P. C. 2001. Carnivores as focal species for conservation planning in the Rocky Mountain region. Ecological Applications. 11: 961–980.

Carroll, C.; Noss, R. F.; Paquet, P. C.; Schumaker, N. H. 2003. Use of population viability analysis and reserve selection algorithms in regional conservation plans. Ecological Applications. 13(6): 1773–1789.

DeMarchi, D. 1996. An introduction to the ecoregions of British Columbia. Wildlife Branch, Ministry of Environment, Lands and Parks, Victoria, BC. 46 p. plus appendices.

Lambeck, R. J. 1997. Focal species define landscape requirements for nature conservation. Conservation Biology. 11: 849–856.

Hitt, N.P. and C.A. Frissell. 2000. An Evaluation of Wilderness and Aquatic Biointegrity in Western Montana. USDA Forest Service Proceedings RMRS-P-15-VOL-2.

Jones, K. 2002. Mapping Bird Abundance and Community Diversity from Satellite Imagery: Validation of AVHRR and MODIS Models. Unpublished MSc Thesis: Montana State University.

Noss, R. F.; O’Connell, M. A.; Murphy, D. D. 1997. The science of conservation Planning—habitat conservation under the Endangered Species Act. Washington DC: Island Press. 246 p.

Noss, R. F.; Wuerthner, G.; Vance-Borland, K.; Carroll, C. 2001. A biological conservation assessment for the Greater Yellowstone Ecosystem: draft report to the Greater Yellowstone Coalition. Conservation Science, Inc. Corvallis, OR.

Oechsli, L. and C. Frissell. 2003. Aquatic Integrity Areas: Upper Columbia River Basin. American Wildlands, Bozeman, MT. 28 pp.

Roberge, J-M.; Angelstam, P. 2004. Usefulness of the umbrella species concept as a conservation tool. Conservation Biology. 18 (1): 76–85.

Rumsey, C.; Wood, M.; Butterfield, B.; Comer, P.; Hillary, D.; Bryer, M.; Carroll, C.; Kittel, G.; Torgerson, K. J.; Jean, C.; Mullen, R.; Iachetti, P.; Lewis, J. 2003. Canadian Rocky Mountains Ecoregional Assessment, Volume One: Report. Prepared for The Nature Conservancy and the Nature Conservancy of Canada.

Sanjayan, M. A.; Jeo, R.; Sizemore, D. 2000. A Conservation Area Design for the central coast of British Columbia. Wild Earth. 10(1): 68–77.

Singleton, P. H.; Lehmkuhl, J. F. 1999. Assessing wildlife habitat connectivity in the Interstate-90 Snoqualmie Pass corridor, Washington. In: Evink, G. L.; Garrett, P.; Zeigler, D., eds. Proceedings of the third international conference on wildlife ecology and transportation; 1999 September 13–16; Missoula, MT. FL-ER-73-99. Tallahassee, FL: Florida Department of Transportation: 75–83.

Singleton, P. H.; Lehmkuhl, J. 2000. I-90 Snoqualmie Pass wildlife habitat linkage assessment: final report. Report No. WA: RD489.1. Olympia, WA: Washington State Department of Transportation. 97 p.

Singleton, Peter H.; Gaines, William L.; Lehmkuhl, John F. 2002. Landscape permeability for large carnivores in Washington: a geographic information system weighted-distance and least-cost corridor assessment. Res. Pap. PNW-RP-549. Portland, OR: U.S. Department of Agriculture, Forest Service, Pacific Northwest Research Station. 89 p.

Singleton, Peter H.; Gaines, William L.; Lehmkuhl, John F. 2004. Landscape permeability for grizzly bear movements in Washington and Southwestern British Columbia. Proceedings of the workshop on border bears: small populations of grizzly bear in the US-Canada transborder region. Ursus. 15(1) Workshop Supplement: 90–103. Available: www.huntingandfishingjournal.org/archives/ issues/GB-BBW-ALL-FINAL.pdf. [April 14, 2006].

USDA Forest Service. 1997. Water/Road Interaction Technology Series. San Dimas Technology and Development Program, San Dimas, California.

Williams, Justin C., Optimal reserve site selection with distance requirements, Computers & Operations Research.Volume 35, Issue 2, February 2008, Pages 488-498.

Williams, Justin C., Charles S. ReVelle, and Simon A Levin, 2004, Using mathematical optimization models to design nature reserves, Frontiers in Ecology and the Environment 2(2), pp. 98-105.

Figure 1. The study area.
Figure 2. An initial prioritization of conservation areas
Figure 3. A composite core model
Figure 4. A spatial optimization model
Figure 5. The final composite core model
Figure 6. Optimized core areas
Figure 7. Optimized 30% of all species cores
Figure 8. Optimized 30% of all species cores plus 100% caribou in BC.
Figure 9. Watersheds within cores plus additional high quality watersheds.

  1. Lance Craighead, Craighead Environmental Research Institute 201 S Wallace Ave, Ste B2D, Bozeman MT 59715, 406-585-8705
  2. Thomas J. Olenicki, Trail Creek Wildlife Planning,467 Storrs Rd., Bozeman, MT 59715, (406) 587 - 5793, tcwildlife@gmail.com
  3. Brent L. Brock, Landscape Ecologist, Craighead Environmental Research Institute, 201 S Wallace Ave, Ste B2D, Bozeman MT 59715, 406-585-8705 bbrock@craigheadresearch.org
  4. Justin C. Williams, Associate Research Professor, Johns Hopkins University, Dept. Geography & Environmental Engineering - Ames Hall, 3400 N. Charles Street Baltimore, MD 21218, 410 516 5079, jcwjr@jhu.edu
  5. Baden Cross, Applied Conservation GIS, 826 Woodcreek Dr. North Saanich, B.C. V8L 5K4 Canada. Email: badency@telusmail.net

Return to Conference Main Page


"Treebeard" (above) is portrayed courtesy of McBride artist Sheilagh Foster.

The University of Northern British Columbia
Indigo Ink Graphic Design