基于投影寻踪的洞庭湖生态系统健康评价

ECOSYSTEM HEALTH EVALUATION OF DONGTING LAKE BASED ON PROJECTION PURSUIT MODEL

以历史文献、图件及1986～2010年水沙、水质等实测资料为依据，选取环境要素、生物指标、水动力条件和外部指标等4类共13个表征洞庭湖生态系统健康状况的指标，采用加速遗传算法的遗传寻踪模型，基于matlab7．0平台对典型年洞庭湖生态系统健康状况进行评价。结果表明：（1）1986-2010年，洞庭湖生态系统健康样本投影值呈下降趋势。1991年左右出现较高点（投影值为1．7264），2006年出现了较低点（投影值为1．6258），而2010年投影值有所回升（投影值为1．679O）；（2）从空间分布上看，南洞庭湖健康状况相对较好，其次是西洞庭湖，最差的是东洞庭湖。这说明受湖泊围垦、退田还湖和三峡水库运用等影响，洞庭湖的水文水动力条件、湖泊生物多样性以及水质状况等发生了较为明显的变化，且各湖泊存在一定的差异。

According to historical literatures, maps, water and sediment data, water quality monitoring data from 1986 to 2010,13 indicators including external and environmental indicators,hydrodynamic conditions and biological indicators were selected to evaluate the healthy condition of the lake ecosystem during the typical years by Projection Pursuit method based on Genetic Algorithm in the Dongting Lake. The results show as follows. From 1986 to 2010,the projection value of ecosystem health in the Dongting Lake showed a tendency of ＂wavy descend＂. The highest value was in 1991（1.7264）, the lowest healthy membership grade was in 2006（1. 625 8）,and in 2010 the projection value began to rise again（1.679 0）. The South Dongting Lake had a relatively good health status, then followed by West Dongting Lake, and the East Dongting Lake was the worst one. It illustrated that water dynamic conditions, biodiversity and water quality have been changed a lot under the effect of the reclamation of lakes,returning farmland to lake and the Three Gorges Reservoir, but it exists some differences in different lakes.