Testing the Lake States variant of FVS (Forest Vegetation Simulator) for the main forest types of northern Ontario

Abstract

The Lake States variant of the FVS (Forest Vegetation Simulator) model (LS-FVS), also known as the LS-TWIGS variant of FVS, was validated for black spruce (Picea mariana (Mill.) BSP), white spruce (Picea glauca (Moench) Voss), jack pine (Pinus banksiana Lamb.) and trembling aspen (Populus tremuloides Michx.) forests in northern Ontario. Individual-tree data from 537 remeasured sample plots were used. This dataset included different combinations of site index, stand density and age. It was possible to compare observations and predictions for different projection length periods. The validation exercise included a biological consistency analysis, the computation of mean percent difference (MPD) for stand density, stand basal area, top height and quadratic mean diameter (QMD) and the comparison of observed and predicted individual-tree dbh. The biological consistency analysis indicated that LS-FVS logically predicted the effect of site index on top height, stand basal area and QMD for black spruce and jack pine. However, the decrease in stand basal area at young ages was inconsistent with the normal development pattern of the forest stands under study and was attributed to deficiencies in the prediction of mortality. LS-FVS was found to underpredict stand density, stand basal area and top height and to over-predict QMD. Even though there were large errors in the prediction of change in stand density, LS-FVS was nevertheless consistent in the prediction of the shape of the dbh size distribution. Key words: FVS, Forest Vegetation Simulator, validation, biological consistency analysis