Comparison of Three Methods of Estimating Surface Area and Biomass for a Forest of Young Eastern White Pine

This paper presents estimates of foliage, branch and stem surface area, and ovendry weight, with estimates of precision of these statistics, for a 10-yr-old stand of eastern white pine on a 16-ha watershed at the Coweeta Hydrologic Laboratory in the Southern Appalachians. Three different methods were used to estimate the forest surface area and biomass: (1) stratified two-phase sampling, (2) two-phase sampling with a regression estimator, and (3) two-phase sampling with a ratio-of-means estimator. Stratified two-phase sampling was the most precise and appropriate method; the population was estimated to contain 5.3 ha foliage, 0.76 ha branches, and 0.13 ha stems per hectare of land surface. The estimated ovendry weight of tree components was estimated to be 2.71, 6.83, and 7.01 metric tons per hectare, respectively, for foliage, branches, and stems. The standard error of estimate for surface area and biomass ranged from 5 to 10 percent, depending upon the tree components of interest. Forest Sci. 20:97-700. Additional key words. Pinus strobus, sampling, conifer, foliage, branches, stems. THE QUANTITY OF TREE SURFACE AREA and tree biomass per unit area of land are inventory data needed to understand the flow of energy, nutrients, and water through forest ecosystems. Surface area, in contrast to biomass, has received little attention. Although Whittaker and WoodweU (1967) described the surface area relations for several deciduous forests in the eastern United States, comparable studies for evergreen forests are lacking. A deficiency common to surface area and biomass estimates in early studies of forest communities was the lack of statistical statements about the reliability of estimated quantities. The quality of results can vary greatly because a series of sampling steps are involved; thus, the error term for a population estimate depends on the error associated with each sampling step. Attiwill and Ovington (1968) and Satoo (1967) pointed out the need for appraising different methods of determining forest biomass. Ovington et al. (1967) and Attiwill (1966) calculated the magnitude of errors which may occur when different methods are used to estimate the biomass on a forest plot. However, forest ecosystem investigations may involve populations considerably larger than a plot, and procedures for assessing the precision of population prediction from sampled values become more complicated. Our study of surface area and biomass was stimulated by the results of several watershed experiments at the Coweeta Hydrologic Laboratory in western North Carolina. These experiments demonstrated that streamflow reductions of 10 percent occurred 10 yr after watershed cover types were changed from mature hardwoods to The authors are, respectively, Principal Plant Ecologist, Southeastern Forest Experiment Station, USDA Forest Service, Coweeta Hydrologic Laboratory, Franklin, N.C., and Mathematical Statistician, SEFES, Forestry Sciences Laboratory, Research Triangle Park, N.C. Manuscript received Sept. 1, 1972. volume 20, number 1, 1974 / 91 white pine (Swank and Miner 1968). The results suggested that appropriate vegetative parameters should be measured which logically describe cause and effect relationships when water losses are compared within and between forest ecosystems. Therefore, the immediate objectives of this study were twofold: (1) to estimate the quantity of surface area and biomass for foliage, branches, and stems in a 10-yr-old stand of eastern white pine (Pinus strobus L.) at one point in time by several methods, and (2) to derive error terms for these