基于logistic模型的人类开发活动影响下河北省滨海湿地演变驱动分析

DRIVING FORCES ANALYSIS OF COASTAL WETLAND EVOLUTION IN HEBEI COASTAL ZONE BASED ON LOGISTIC REGRESSIONG MODEL

关键词:  滨海湿地Logistic回归农业活动开发区域建设活动开发区域河北省海岸带

Key words:  coastal wetland  Logistic regression  constructive development region  agricultural development region  Hebei coastal zone

摘要:

［目的］为定量分析河北省海岸带地区滨海湿地向农业开发区域和建设用地开发区域转变的驱动机制，［方法］文章基于遥感数据，采用面向对象与目视解译相结合方法，提取研究区2005年和2013年两个时期土地利用空间格局信息，分析滨海湿地向建设开发区域和农业开发区域演变发生的空间位置信息，并作为空间拟合分析的因变量样本。同时，通过ARCGIS软件基于DEM和专题数据，构建高程、坡度、气温、降水、距城镇距离、距农村居民点距离、距海岸线距离和距道路距离8个自变量样本，使之与湿地空间演变样本进行Logistic回归拟合，计算各自变量的贡献率和演变发生概率在空间上的变化特征。［结果］在湿地向农业开发区域演变的过程中，距农村居民点距离和距海岸线距离两个变量贡献较大，距农村居民点距离减小1个单位，其转变为农业活动开发区域的概率将增大155倍。在湿地向建设开发区域演变的过程中，距海岸线距离、距城镇距离和距农村居民点距离贡献较大，其发生概率随距海岸线距离减小而增大，减小1个单位，概率将增大172倍。［结论］基于概率诊断与RS、GIS技术结合的综合研究方法，可以为定量识别滨海湿地演变的驱动因子的贡献率，为海岸带地区的湿地保护提供科学依据。

Abstract:

The study area lies in Hebei coastal zone. The spatial distribution of coastal wetlands in 2005 and 2013 was extracted by means of a combined method of object orientation and visual interpretation based TM images. The overlaying analysis of these data was conducted, the types and spatial location of change were discovered and served as the sample of dependent variable．At the same time, variables including altitude, slope, temperature, precipitation, distance to cities , distance to rural residence, distance to coastline, distance to road were extracted served as sample independent variable, supposed by ARCGIS based on DEM, and then Logistic Regression were conducted based these variables. The contribution of each independent variable as well as the spatial property of the occurrence probability of change was calculated. The result showed that, during the change from coastal wetlands to agricultural development regions, the contribution of distance to rural residence and distance to coastline were greater than other variables, when the distance to residence decreased by one unit, the probability for its conversion to agricultural development regions would be increased 1.55times. During the change from coastal wetlands to constructive development regions, the contribution of distance to coastline, distance to cities, distance to rural residence were greater than other variables, and when the distance to coastline decreased by one unit, the probability for its conversion to constructive development regions would be increased 1.72 times. That comprehensive research method based on probabilistic diagnosis RS and GIS can identify the contribution rate of driving factors for coastal wetland evolution quantitatively, it is beneficial to the protection of coastal wetlands.