PARENT SESSION
OOS 27: From Idea to Reality: Applied Science and Ecological Problem Solving in the National Forest System.
Organized by: H Safford, R Holthausen, and T DeMeo
Thursday, August 5, 8:00 AM to 11:30 AM, Meeting Room E 143.

Use of potential natural vegetation modeling to identify rare species habitat in the Pacific Northwest.

Henderson, Jan¹, Lesher, Robin^*,1, DeMeo, Thomas², ¹ USDA Forest Service, Mountlake Terrace, WA, USA² USDA Forest Service, Portland, OR, USA

ABSTRACT- A potential natural vegetation (PNV) modeling technique was used to identify and map rare plant and animal habitats in support of the Northwest Forest Plan Survey and Manage Program, an effort to manage species of concern not covered by coarse-scale conservation strategies. Climatic grids of temperature and moisture variables are developed, and then used to generate potential vegetation maps at the vegetation zone and plant association group level. The habitat model is developed using known locations of the species and associated modeled environmental variables to predict where on the landscape similar environmental conditions occur. We compared the known site distribution with the distribution of that variable across the landscape to determine if the distribution of plots is different than random. The model solves a series of Lorentzian functions and produces a value for each 30 m pixel. A map is produced that displays categories of "likely habitat". This preliminary model is field calibrated and refined to improve the accuracy of mapping potential habitat. The final stage is a field validation sample, where a random set of pixels is surveyed to determine species presence, and is used to calculate an accuracy assessment. This PNV modeling approach has been effective in predicting the habitat areas of many species (plant and animal), because their ranges often correlate well with temperature and moisture gradients. For example, the middle of the range of red tree vole correlates well with a mean annual temperature of 9.5°C. Models have been developed for twelve species representing various taxonomic groups. This approach has been successful in finding new sites of the target species, and in understanding and quantifying the ecological distribution of the species. Results have saved funding and personnel time by eliminating project areas from rare species concern, and by directing field searches towards areas of greater probability of occurrence.

Key words: species habitat models, conservation biology, potential natural vegetation, Pacific Northwest