Beaver (Castor canadensis Kuhl) and willow (Salix spp,) are important components of riparian restoration on degraded western rangelands, Land managers need quantitative information to evaluate carrying capacity and potential habitat quality for beavers in riparian-willow systems, Our objectives were to determine the best model to predict biomass components of coyote willow (S, exigua Nuttall) from basal stem diameters and compare model predictions to diameter class averages, The study was conducted in a shrub-steppe ecosystem of northwestern Colorado, We estimated oven-dried weights of annual and total beaver food and total live biomass by diameter class from a sample of 160 willow stems. Several variants of a logistic function were fit with nonlinear least squares regression to select a model that best predicted mean biomass by stem diameter, A four-parameter logistic model provided the best fit for all 3 stem components. Predicted biomass estimates of beaver food and total live biomass had smaller standard errors than sample means for all 10 stem diameter class midpoints, Percentage of stem weight that was beaver food varied from 93.6% for the smallest stems to 12.2% for the largest, We concluded that the logistic model provided reliable estimates of beaver food biomass and could be used with food consumption rates and stem density data to evaluate carrying capacity for beaver or test assumptions in the beaver habitat suitability index model.
