Abstract:

Spatial autocorrelation can not be avoided in the forest landscape model, However, many studies often ignore this point, in order to solve the problem that combined with the spatial autocorrelation to build a model of forest landscape, Lushuihe Forest Bureau was taking as an case to analyze the spatial autocorrelation of the Red Forest landscape by using spatial autocorrelation analysis, semi-variance analysis means analysis. And that the Moran’s I and the semivariogram both showed that there were spatial autocorrelation of Korean pine in 1987, 1995 and 2003, but they showed a weakening trend of spatial autocorrelation, the distribution of red pine by non-random factors gradually became dominant random factors accounted for the dominant position, logging and other human factors had an important impact for the distribution of red pine. On this basis, the linear regression model and the spatial lag model were compared used to build landscape spatial model, the result showed that the fitness of the latter was better than that of the former.