多维球面模型应用于草原监测数据分析的探讨

 提出分析归纳草原监测多元数据的新数学方法，多维球面模型。M个物种所组成的草原的状态可以用M元向量来表示，向量的长度用来表示草原的量值，向量的方向表示草原组成。从草原监测的角度，草原的相对组成是更重要的，成份变化的比率更能说明问题。在处理草原监测数据中使用向量等比系列和向量等差系列是可行的。等比系列比较能反映生物几何增长的本质。为了纠正中间过程被忽略的弱点，要坚持逐年监测，逐步分析，或使用滑动平均。二阶趋势值可尝试用于解决时间滞后问题，卡门滤波用于降低预报误差。因为草原监测需要使用分种数据，必须以人工监测记录为主，再辅以其它高科技手段，如遥感。

APPLICATION OF THE MULTI-DIMENSIONAL SPHERE MODEL IN RANGELAND MONITORING DATA ANALYSIS

Monitoring grassland is important and possible. Multi Dimensional Sphere Model (MDSM) is a new multivariate data analysis method to analyze the grassland monitoring data. Grassland can be expressed as a point in multivariate space, or a multi component vector. The magnitude of the vector expresses the biomass of the grassland, but the direction of the vector expresses the composition of the grassland. For temporal dynamic analysis, the composition of the grassland is more important than the production of the grassland, and the ratio of the changes can explain more than the differences do. Both the geometrical and arithmetic series can be adopted to analyze the grassland monitoring data. However, MDSM prefers the geometrical series. Moving average and second degree trend may be introduced to improve the model.