Application of improved sequential indicator simulation to spatial distribution of forest types

摘要

Spatial distribution of forest types is the base and precondition of conducting forest landscape pattern. In current practical forest production, the attainment ways of forest distributing map have very strong subjectivity or are restricted by environmental factors. The purpose of spatial statistics is to describe the distribution characteristics of the things in the space. Conditional simulation algorithm is one kind of effective means of conducting spatial interpolation in the spatial statistics. The forest type is a regionalized categorical variable, and perhaps there are some regions with patch fragmentation of forest types in research regions. For these reasons, this paper used the sequential indicator simulation (SIS) algorithm of categoricalvariables to simulate the spatial distribution of forest types. The SIS had very great chanciness when the adscription of forest type of simulated point was confirmed, and hence the precision of the SIS was relatively low. This paper presented the improved method for the SIS. Using the sample plots of the China's Northeast Wangqing Forestry Bureau and the Guangxi Zhuang autonomous region as the tested materials, the simulation of forest distributing map was successfully implemented with the improved sequential indicator simulation (ISIS) algorithm, and thus obtaining a simulation precision of about 80%. Results of precision analysis indicated that the ISIS enhanced the simulation precision. Therefore, the ISIS was feasible and effective. Using the sample plots as the tested materials, the ISIS might be used as an effective way of obtaining the forest distributing map.