露水河林业局森林景观空间自相关分析与拟合

空间自相关是建立森林景观模型中不可回避的问题，然而目前众多研究中常忽略了这一点，为了探解决考虑空间自相关性，建立森林景观模型的问题，以露水河林业局为例，采用空间自相关分析、半方差分析等手段分析红松森林景观等景观类型的空间自相关性。研究结果发现，Moran’s I指数以及半方差图显示红松在空间分布上具有较强的空间自相关性，但这种空间自相关性从1987-2003年呈现减弱趋势，红松的分布由非随机因素占主导地位逐渐变为随机因素占主导地位，采伐等人为因素对于红松的分布产生了重要影响。在此基础上分别利用经典线性回归模型和空间滞后模型建立了红松森林景观分布与环境驱动因子的模拟模型，经过比较，认为空间滞后模型的拟合度优于经典线性回归模型。

Spatial Autocorrelation Analysis of Forest Landscape in Lushuihe Forestry Bureau

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 Forestry 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 to build landscape spatial model,the result showed that the fitness of the latter was better than that of the former.