The objectives of this study were to determine the loading of dead and down forest fuels in 6 forest cover types on the upper Coastal Plain of Virginia, specifically in the Richmond National Battlefield Park, and to determine the relationships between the fuels and the vegetation. The forest fuels and vegetation were sampled in stands of the following cover types: pine, oak-pine, mixed hardwood-pine, oak, mixed hardwood, and oak-mixed hardwood. The planar intersect method was utilized to quantify the woody fuels. and the forest floor was sampled in fixed area plots. The weight. volume, depth. and species of fuel particles was determined for each stand sampled in the Park, and this information was combined for an estimate of fuel loading by cover type. The overstory trees were sampled within prism plots, and the understory was sampled in fixed area plots. Standard forestry information was collected for each sampled stand, and estimates of species dominance, density, frequency and importance were determined.

In general, the fuel loading was not different between the six cover types as determined with ANOVA procedures. due to the large variation of loads within each. The loading of certain sizes of fuel particles were different between a few cover types, due to both the influence of the overstory species and the site they had occupied. For example, the 1-hr. time-lag branch fuel loading was significantly greater in the oak cover type than in the pine or mixed hardwood cover types. Another significant difference was in the forest floor loads: the mixed hardwood cover type had a lighter forest floor than the pine or oak-pine cover types. This was most likely due to the higher quality of sites upon which the mixed hardwood stand were found, and the higher palatability of litter produced by these species, which would favor the decomposition of litter more than conditions found in the pine or oak-pine cover types.

To quantify relationships between the fuel loads and forest vegetation and site characteristics, Spearman's rank correlation coefficients were calculated between all the variables. Many significant relationships were found, though all but one correlation coefficient was under 0.51: total forest floor load and forest floor depth had a correlation coefficient of 0.75. Forest and site variables significantly related to fuel loadings were tested for their predictive value with multiple regression statistics. The resulting set of regressions had low coefficients of determination and varying levels of precision; their usefulness would depend on the level of sampling intensity and precision one would want to expend to get an estimate of fuel loading in a stand. If a rough estimate requiring little time in the field is desirable, then these regressions would be useful.