植被监测及趋势分析——植被数量生态学中几个理论问题的探讨

 系统监测可以对危机发出预警, 是防治灾害的重要手段。生态监测的基础是植被监测。多物种 多样本 多年的植被定位监测数据隐含 着植被变化的信息。该文探索描述植被的数学工具, 提出植被监测数据的趋势分析方法。植被是资源竞争系统, 可以用多维空间的向量来表示 。在向量空间(射影空间), 不是“距离” , 而是“方向”决定区别; 在植被科学, 不是“产量”, 而是“组成”决定区别。新方法用多维空间 的位置向量来表示植被: 向量的方向表示植被的组成、两向量夹角余弦值表示相似、向量长度表示植被总体。在简缩数据时, 用“中心化”滤 去样本噪音、“标准化”滤去系统噪音, 得到状态向量。在趋势分析时, 定义后、前状态向量的比值为变化趋势; 用当年的状态和趋势的乘积 来预报次年的状态。到次年, 再用实测数据修正、更新来自去年的预报, 是为“卡尔曼滤波”。卡尔曼滤波能降低监测成本, 有效地使用历史 数据, 提高分析精度。

VEGETATION MONITORING AND TREND ANALYSIS: DISCUSSIONS ON QUANTITATIVE VEGETATION ECOLOGY

Vegetation monitoring is important. This paper introduces a trend analysis method for vegetation sciences. A unit of homogeneous vegetation can be treated as a point, so it can have dynamic analysis. However, to carry enough information, this point has to be put into multi-variable space. Vegetation can be expressed as a position vector in multidimensional space. Vegetation is a resource competing system. All plant species compete for limited resources, and demands of all species can not exceed available resources. This can be expressed as the sum of the squares of all species equals one. Therefore, all the complementary plant species can be treated as mutually orthogonal. When using position vectors to represent vegetation, the magnitudes of the vectors carry information of the total biomass, while the directions of the vectors carry information of composition of vegetation; thus, the position vectors have to be standardized. Vegetation growth based on cell duplication is expressed as exponential growth. Changing trend is defined as present state over the past. The trend can be used to monitor the vegetation changes and to predict future states. Kalman filter is used to increase accuracy and lower monitoring cost.