An environmental gradient model predicts the spatial distribution of potential habitat for Hypogymnia duplicata in the Cascade Mountains of northwestern Washington

A spatially explicit environmental gradient model was developed and field validated to determine the ecology and distribution of an epiphytic lichen of conservation concern in the Pacific Northwest. Hypogymnia duplicata was a Survey and Manage species in the Northwest Forest Plan and required special management on federal lands. The model output was a potential habitat map for Hypogymnia duplicata in four likelihood classes at a 90-meter square pixel resolution for the Cascade Range of northwestern Washington. A basic assumption of this habitat model was the frequency of occurrence of a species along an environmental gradient resembles a Gaussian distribution. When this study was initiated, H. duplicata was known from three locations in the study area; 76 new sites were discovered while field calibrating preliminary models. Known locations of H. duplicata and mathematical associations with environmental variables were used to predict and map sites where similar environments occurred. Frequency distributions of H. duplicata were compared to frequency distributions of the study area for 15 continuous environmental variables representing direct gradients, plus a qualitative variable for potential vegetation. Analysis of actual gradients in a spatial context provided insight into which variables were predictive in describing H. duplicata habitat. The final model had five variables: Precipitation at Sea Level, Temperature Lapse Rate, Fog Effect, Cold Air Drainage Effect, and Plant Association Group. Variables representing atmospheric qualities were most important for this epiphytic lichen. The final habitat model was field validated with an independent random sample of 143 plots stratified by four habitat classes and two forest age classes. Hypogymnia duplicata occurred on 53 plots (37%) of the validation sample; 98% of these occurrence plots were mapped as high and moderate likelihood habitat. Hypogymnia duplicata showed strong association with older forests, occurring in older forests 96% of the time. High likelihood habitat was 10% (110,163 ha) of the study area (1,096,265 ha). This modeling approach successfully described and mapped potential habitat, and provided information on the species' ecology and distribution. The results showed H. duplicata was not at risk on federal lands in the study area under current climate conditions and reserved land status.