Methods for processing information of radar topographic survey for solving the tasks for determining heights and wood stocks for forest stands

The paper studies the possibilities of using global elevation models SRTM (Shuttle radar topographic mission) to assess the  characteristics of forest stands – average heights and wood stocks. It is known that in process of radar shooting  vegetation is considered as a barrier to correctly determining the height of the earth’s surface. The surface, fixed by the sensors above the forest covered territory (scattering phase center height), is located in the upper part of the forest canopy. The review of publications confirms the relevance of this area of investigation in the world. A brief theoretical basis of the SRTM survey, factors determining the values of the heights and the errors associated with them are presented based on literature reviews. The possibility of determining the height of forest canopy based on evaluation models is shown.
The part of Uchebno-Opytnoe Forest District of the Leningrad region was chosen as the model territory. The geographic information databases, data of radar survey SRTM and topographic maps were the origin data for the study. The model territory is divided by a regular network into cells with  1 km step. Most of the territory is covered with forest vegetation. Relief modeling was performed on the basis of digitized topographic maps of 1:25000 scale by interpolation using  TIN method. A visual and statistical comparison of both evaluation models – a surface model (based on SRTM radar survey data) and a relief model (based on topographic maps) was done. With help of the profiles construction an assessment of the nature of  changes in the heights of the relief and surface models was performed. It is noted that for most of the model territory, the differences in height between the surface and relief models are 15-20 m. The convergence of graphs for heights is observed in areas, which are not covered with forest vegetation.
The set of linear regression dependencies between the scattering phase center heights (independent variable) and average heights or wood stocks (dependent variable) within the borders of forest compartments, grouped by the predominated tree species, was obtained. The influence on the closeness of the relationship and the value of the regression coefficient of such factors as the value of basal area and the share of predominant tree species in composition was found.
The established regularities can be used to improve the theory and practice of forest inventory, as well as to solve other problems related to the assessment of natural resources at the regional and global level (national forest inventory, carbon stock determination, assessment of biomass).