عنوان مقاله [English]
Possibility of estimation of quantitative forest characteristics without field survey is a new idea for forest biometry. This investigation was conducted to evaluate possibility of spatial estimation of two main quantitative forest characteristics (trees number in hectare and mean diameter) using terrain analysis and linear regression models in the district 1 of Dr. Bahramnia (Shastkolate) forestry plan, Golestan Province. A digital elevation model (DEM) with 10×10 meter resolution was produced by contours interpolation of 1:25000 scale maps. The primary topographic attributes maps (slope, aspect, elevation from sea level, shaded relief, profile curvature, plan curvature and tangential curvature) and secondary attributes maps (wetness index and solar radiation) were derived from DEM using terrain analysis software. Field inventory data from second revision of Shastkolate forestry plan, district I (429 plots) were used as original data for forest quantitative characteristics. Information of topographic maps in plot location was derived. Then relationship between the forest characteristics and terrain attributes was analyzed and modeled using multiple linear regression models by stepwise approach. The developed models were validated using test data (85 plots). Adjusted R2 and root mean square error (RMSE) were determined to validate accuracy of prediction. Result of study showed that elevation from sea level, annual potential of solar radiation, slope and aspect were the most significant terrain attributes determined the spatial distribution of forest quantities characteristics. Results also showed that forest characteristics could be predicted about 40% to 47% of variation by these linear models. These models can be used for estimation of quantitative characteristics for adjacent forests with similar conditions (same management and morphology) using DEM without measurement or sampling. Finally, use of terrain analysis was emphasized for preliminary recognition of forest quantity characteristics.