Assessment of a Double-Sampling Technique for Estimating Browse Biomass

Abstract

We assessed the potential for determining current annual growth (CAG) of browse with visual estimates of shrub canopy volume, combined with double sampling for clip-and-weigh measurements of biomass. This sampling approach permits direct assessment of CAG per unit area without estimates of shrub density. Ratio estimates based on double sampling provided suitable CAG estimates, assuming 80% confidence, for total production when browse species were combined. This double-sampling approach for individual browse species is likely possible only when shrub taxa are abundant or uniformly distributed, or when the sample site has been stratified according to the distribution of CAG values of specific browse species. J. WILDL. MANAGE. 54(2):342-348 Inventory programs to estimate browse production available for use by wildlife and livestock pose complex problems because of highly variable plant physiognomy, life history, and spatial distribution of browse species (Telfer 1981). Sampling methods range from visual estimates to detailed clip-and-weigh determinations of CAG or standing crop; each method has inherent strengths and weaknesses (Rutherford 1979, Pitt and Schwab 1988). Visual estimates offer an inexpensive and relatively easy method to assess shrub CAG, but are hindered by observer and statistical bias (Pechanec and Pickford 1937, Jolly 1954) and often fail to provide management agencies with verifiable levels of statistical confidence and precision (Pitt and Schwab 1988). Shrub CAG may be determined directly by clipping within randomly or systematically selected plots. Plant material is then oven-dried to obtain biomass, CAG, or availability per unit area (Mannetje 1978). Clip-and-weigh techniques are typically expensive and tedious and are more appropriate for intensive research than for extensive browse inventory programs (Pitt and Schwab 1988). Regression and ratio estimates reduce the arduous, expensive aspects of direct measurements by predicting CAG from more easily measured shrub components. Biomass of leaves, twigs, or entire shrubs have been associated with stems per unit area (Wolff 1978); twig diameter (Telfer 1969, Provenza and Urness 1981); stem diameter (Felker et al. 1982); shrub height (Ohmann et al. 1976); crown volume (Guy 1981); crown diameter, depth, and density (Dean et al. 1981); crown width (Rittenhouse and Sneva 1977); crown area, or cover (Anderson and Kothmann 1982); height and stem diameter products (Stone and Crawford 1981); twig lengths (Schuster 1965); twigs per plant (Peek 1970); and shrub ring widths (Davis et al. 1972). Despite the abundant literature describing methodology, extensive shrub inventories remain prohibitively costly, perhaps because most regression techniques provide CAG per stem or shrub. Such CAG estimates must be combined with stem or shrub density estimates to yield CAG per unit area. Appropriate techniques for estimating density vary according to physiognomy and spatial distribution of shrub species. Many density estimates are laborious or contain statistical bias (Pitt and Schwab 1988). The most appropriate shrub inventory method, therefore, is a regression or ratio estimate of CAG predicted from shrub components measured on a unit-area basis, such as canopy volume. These predictions estimate CAG per unit area directly, without density estimates. Shrub canopy volume also correlates well with biomass and CAG and has generally lower coefficients of variation compared to other predictor variables (Bryant and Kothmann 1979, Murray and Jacobson 1982). Direct predictions of CAG per unit area are feasible for browse inventory when double-sampling programs are used (Cochran 1977:327, Nichols 1979). We assessed a double-sampling approach for predicting shrub CAG using visual volume estimates recorded during first-phase sampling and actual CAG clipped and weighed during second-phase sampling. Feasibility for browse inventory programs was also evaluated relative to field sampling costs and statistical precision. M. L. Beets, C. A. Easthope, R. C. Gordon,