Methods

Fig. 1 - Western United States
and Southwestern Canada

Study Area

The area of study consisted of the region supported by the climate surfaces of Rehfeldt (2006). This area encompasses the Western United States and Southwestern Canada, latitudes 31° to 51° N and longitudes 102° to 125° W (figure at right).

Climate Estimates

Spline climate model (Rehfeldt 2006) was used to estimate climate at point locations (latitude, longitude, and altitude). The model is based on monthly minimum, maximum, and average temperatures and precipitation normalized for the period of 1961-90. This data was obtained from more than 3000 weather stations dispersed throughout the conterminous Western United Sates and Southwestern Canada. Hutchinson’s (1991, 2000) thin plate splines were used to develop geographic surfaces for monthly data. The result produced 48 surfaces representing minimum, maximum, and average temperature and precipitation for each month of the year. For our analyses, these surfaces were used to estimate 33 climate variables, which represent a range in climate variables from simple sums to complex interactions of temperature and precipitation (Table: Climate Variables).

An updated version of the spline climate model was used to predict the effects of global warming for decades beginning in 2030, 2060, and 2090. The model used a regional summary of the IS92a scenario (1 percent per year increase in greenhouse gases after 1990) of the International Panel on Climate Change (Houghton and others 2001) from General Circulation Models (GCMs) produced by the Hadley Centre (HadCM3GGa1, Gordon and others 2000) and the Canadian Centre for Climate Modeling and Analysis (CGCM2_ghga, Flato and Boer 2001). These two models are generally well respected, and averages of their predictions should provide an illustration of the potential effects of global warming.

Vegetation Data

Species were selected for this analysis based on their small geographic range and the availability of point location data to the authors. Species presence data were acquired from Barnes (1996) for Macfarlane’s four-o’clock, Arno (1970), Khasa and others (2006), and Forest Inventory and Analysis, USDA Forest Service (FIA) Data Base for subalpine larch, and unpublished field data from Forest Service RMRS, Moscow Forestry Sciences Laboratory, for smooth Arizona cypress and Piute cypress (see Rehfeldt 1999). FIA plot data were used to identify locations where the species did not occur. Their database was developed from a systematic sampling of woody vegetation permanent plots on forested and nonforested lands (Alerich and others 2004, Bechtold and Patterson 2005) within the United States. In accordance with FIA proprietary restrictions, plot locations are not available for publication. However, we were given access to their database to generate climate estimates for each plot directly from uncompromised field data. The resulting datasets include a tabulation of the presence-absence of species at approximately 117,000 locations.