The correlation between available deer browse, forest cover type, and forest site
This study is one phase of a ten-year project designed to evaluate forest-wildlife relationships. The project was initiated in 1958 on the Jefferson National Forest, Broad Run Wildlife Management Area, Craig County, Virginia.
The objective of the study was to determine if correlations existed between quantities of available deer browse in the understory of a particular forest cover type and several site quality measurements. The purpose of the study was to determine if weights of browse per acre could be estimated reliably by utilizing easily obtained site quality measurements instead of clipping and weighing browse. Eight site quality measurements (variables) were tested. These were: site index, depth of the A₁ soil horizon, position on slope, basal area per acre, aspect (exposure), percent of slope, elevation, and the number of clipped stems per sampling unit.
Fieldwork was conducted in two major forest cover types; the oak, hickory, poplar, white pine type (cove hardwoods type), and the mixed oak-pine type. Sampling units were located randomly, in pairs, within each of these two forest cover types. Each sampling unit consisted of a circular 1/4 acre plot and a square 1/100 acre plot located at the center of the circular plot. A system of double sampling was used to obtain browse weight data and site quality data for comparisons. Data on eight variables recorded at each 1/4 acre sampling unit were compared with the quantity of browse clipped from the 1/100 acre sampling unit located at the center of that particular 1/4 acre plot.
A multiple regression analysis was used to determine the degree of correlation between quantities of browse (available browse) clipped from sampling units and all measurements of the eight independent variables (site quality measurements) recorded on sampling units.
The final analysis of the oak, hickory, poplar, white pine cover type data indicated that the variables significantly related to browsing weights per acre were the number of stems clipped per sampling unit and the depth of the A₁ soil horizon. These two significant variables explained 48,63 percent of the total variation in browse weights occurring between sampling units. Using only the two significant site quality measurements (independent variables), the final estimating equation was: Y (pounds of browse per acre) = -0.14 + 0.06 (number of stems clipped per sampling unit) + 0.61 (depth of the A₁ horizon, inches). The final estimating equation should not be used for reliable estimates of browse production in the oak, hickory, poplar, white pine forest cover type. A total of 51.37 percent of the variation in browse weights occurring between sampling units is unaccounted for in this equation.
The final analysis of the mixed oak-pine cover type data indicated that the only variable significantly related to browsing weights per acre was the number of stems clipped per sampling unit. However, this significant variable explained only 17.97 percent of the total variation in browse weights occurring between sampling units. Using only the one significant site quality measurement (independent variable), the final estimating equation was: Y (pounds of browse per acre) = 2.24 + 0.07 (number of stems clipped per sampling unit). The final estimating equation should not be used for reliable estimates of available browse production in the mixed oak-pine forest cover type. A total of 82.03 percent of the variation in browse weights occurring between sampling units was unexplained in this equation.
More research is necessary to determine other easily measured environmental factors (variables) which might bear a significant relationship to quantities of available deer browse produced in the two forest cover types sampled. When several more of these significant variables are discovered, the addition of these variables to the estimating equations for the two cover types might account for a large enough percent of the explained variation to enable the game biologist to use the equations for reliable estimates of browse production.