摘要
文章以辽河口湿地景观格局为例,考虑湿地景观格局变化过程的空间变量,建立了不同湿地景观格局变化过程的二元Logistic回归模型。模拟结果表明:模型中转为建筑用地的主要解释变量是城镇人口变化率;转为芦苇地的主要解释变量是农村人口变化率和过境水资源变化率;转为水稻田的主要解释变量是农村人口变化率;转为养殖池的主要解释变量是第一产值变化率和农村人口变化率。空间异质性和湿地景观格局变化过程的时间变量共同影响着使用二元Logistic回归模型来解释湿地景观格局变化驱动力的能力;通过对各个模型的二元Logistic回归模型进行检验,得出空间统计模型能较好地揭示不同湿地景观格局变化过程的主要驱动力及其作用机理。
Taking the case study of wetland of Liaohe, the paper analyzed and simulated the data using ARCGIS and Logistic regression model with the data of landscape map, DEM, population and GDP. Taking into account the spatial variability of the landscape changes process, the descriptive spatial models developed suggested some important driving factors in the landscape changes. Results showed that the binary logistic regression model of landscape patterns has better goodness-of- fit. The most important explanatory variable of the spatial model of building land suggested a model, where conversion to building land was controlled by the urban population rate, while reed conversion was controlled by rural population rate and the rate of change of cross-border water resource. Rice land was controlled by the first value rate, while expansion of pond was controlled by urban population rate and the first value rate. The spatial models can identify main driving forces of different forces of different landscape changes through validation of logistic regression model of wetland of Liaohe.
出处
《环境科学与技术》
CAS
CSCD
北大核心
2012年第6期193-198,共6页
Environmental Science & Technology
基金
国家水体污染控制与治理科技重大专项资助(2008ZX07526-001-02)
辽河河口区大型湿地生态修复关键技术与示范研究(2008ZX07208-009)
本科生研究发展计划项目(OUC-SRDP)
关键词
二元LOGISTIC回归
HL检验
驱动力分析
binary Logistic regression analysis
Hosmer-Lemeshow test
driving forces analysis
