基于克里格插值与序贯高斯协同模拟的森林碳密度空间估计

Comparison of Forest Carbon Spatial Distribution Based on Kriging Interpolation and Sequential Gaussian Co-Simulation

以浙江省仙居县为例,利用普通克里格插值法和序贯高斯协同模拟法对森林碳储量(地上部分)的空间分布进行估计,利用交叉检验方法对其结果进行对比分析。结果表明,仙居县的森林碳密度分布差异较大,大部分地区森林碳密度较低;普通克里格法未能体现森林碳密度的空间差异,具有明显的"平滑"效应,序贯高斯协同模拟在减少平滑影响方面优于前者,序贯高斯协同模拟法的预测结果较克里格插值法高。

The spatial distribution of aboveground forest carbon storage in Xianju County,Zhejiang Province was estimated and simulated by means of applying Ordinary Kriging interpolation method and sequential Gaussian co-simulation method,and the outcomes were comparatively analyzed.The results showed that the aboveground for-est carbon storage of Xianju County was relatively low and it was unevenly distributed.The Ordinary Kriging inter-polation method greatly changed the spatial distribution of aboveground forest carbon storage,which showed a dis-tinctly smooth effect.The sequential Gaussian co-simulation method was better than Ordinary Kriging interpolation method in reducing the smooth affect.For the spatial distribution,Ordinary Kriging interpolation method had to re-cover the non-forest land prior to representing its distribution,but the sequential Gaussian co-simulation method could represent distribution of the non-forest-land without having recovered it.And for the prediction accuracy,the predicted result from sequential Gaussian co-simulation method was relatively higher than that of the Ordinary Krig-ing interpolation method.Therefore,the sequential Gaussian co-simulation method should be a better and more effective method for studying the forest carbon spatial distribution.