Estimating Browse Biomass Using Multiple Regression and Plotless Density Estimates

Abstract

We estimated current annual growth (CAG) of woody material on stems of bog birch (Betula gladulosa) and willow (Salix spp.) growing on 4 sites in central British Columbia using 4 single-stem approaches. Three approaches used regression equations of CAG on stem dimensions; 2 of these involved single equations representing all sites, and the other applied separate equations to each site. The fourth approach employed a clip-and-weigh technique. Estimates of similar precision were obtained from the 3 regression approaches, although the single equation based on the common data set gave biased results. The clip-andweigh method provided the most accurate estimates for a given sample size, but may not be the most costefficient at achieving a specific precision level. A regression equation designed to encompass several sites can likely be developed from a smaller sample than several site-specific equations. It is important that sample data for such an equation be selected equitably across all the sites. J. WILDL. MANAGE. 54(1):180-186 Current annual growth of twigs and leaves of woody plants, commonly termed browse (Basile and Hutchings 1966), is an important forage source for cattle and wildlife in British Columbia (McLean 1979). Effective allocation of browse between domestic stock and wild ungulates requires reliable, quantitative inventories. Browse inventories pose complex problems because of highly variable plant physiognomy, life history, and spatial distribution among browse species. Available browse sampling methods range from simple observation of plant presence to accurate delineation of biomass (Telfer 1981), each with inherent strengths and weaknesses relative to specific objectives (Rutherford 1979, Habgood 1985:4-28, Pitt and Schwab 1988). We present single-stem regression equations relating CAG to measurements of stem and canopy dimensions and quality for bog birch and willow on 4 wetland sites. Bias and precision of estimates from equations derived to represent all the sites, using data collected in 2 different ways, are compared to estimates from site-specific equations. Predictions of CAG production (g/m2), derived by combining estimates of mean stem CAG based on the different regression equations with an estimate of stem density based on the corrected-point-distance (CPD) method (Batcheler 1973, 1975), are compared to estimates obtained using a clip-and-weigh approach. Financial support was provided by the British Columbia Ministry of Environment (Fish and Wildlife Branch). L. Larson, Oregon State University, initiated the project and was instrumental in arranging financing.