渭干—库车河三角洲绿洲土壤盐渍化空间分布特征遥感监测研究

渭干-库车河三角洲绿洲是位于塔里木盆地北缘的具有典型干旱特征的绿洲。干旱地区生态环境脆弱,土壤盐渍化严重破坏了国民经济的生产和生态环境的平衡,所以对干旱区盐渍化问题进行系统的研究是非常必要的。本文采用研究区的2001年8月6日遥感图像进行K-L和K-T变换,经过一系列波段结合．我们得到TM3、KL3、KT2波段结合是盐碱地信息提取的最佳波段结合;然后把研究区的等高线矢量图转换成DEM立体图,以及分类结果和DEM之间进行对比,结果表明,立体图的凹区和分类图的盐碱地基本重合在一起,从而我们得到导致该地区的盐渍化问题的自然和人文因素中,地形因素是占在主导地位的。     

基于贝叶斯网络分类的土壤盐渍化遥感监测

贝叶斯网络是一种将贝叶斯概率方法和有向无环图的网络拓扑结构有机结合的概率模型。采用贝叶斯网络分类对具有典型干旱特征的库车县土壤盐渍化情况进行监测，首先应用条件独立性测试原理建立贝叶斯网络结构，把研究区遥感数据进行离散化，然后应用贝叶斯定理作为分类原则，将每个像元分为像元最大概率的类别。研究结果表明该方法分类6种地类的整体分类精度达到96％，为该区盐渍地面积、空间分布等特征监测提供了较好的依据。     

Parametrization of groundwater quality of the Quaternary aquifer in N’Djamena (Chad), Lake Chad Basin: application of numerical and multivariate analyses

The urban groundwater of the Quaternary aquifer of the Lake Chad basin in N’Djamena has been subject to many hydrochemical studies. However, the results are often not presented in a way that enables water quality managers to make an appropriate decisions, which restrict development and poverty reduction efforts. The objective of the present study was to contribute the improved management of the local groundwater resources. A total of 85 groundwater samples were interpreted using hydrochemical techniques associated with integrated numerical indices and multivariate statistical analysis. The hydrochemical results coupled with the relative residence time of water have shown that the chemical composition of these waters is linked to geogenic and anthropogenic factors and to their proximity to the Chari-Logone rivers. These investigations showed that the groundwater quality in N’Djamena is characterized by a high spatial variability. This study also assessed the suitability of groundwater for user needs and identified areas which are more/less favorable for a specific use. The evaluation of water quality and its suitability for human consumption is also a problem of optimizing data acquisition strategy, and this study used the correlation between water quality index (WQI) and electrical conductivity (EC) to orientate future data acquisition strategies. This parametrization can assist the decision makers and water management professionals in evaluating groundwater availability and setting up a robust water quality management plan in areas with similar hydrogeological and climatic conditions.

     

基于贝叶斯网络分类的土壤盐渍化遥感监测

贝叶斯网络是一种将贝叶斯概率方法和有向无环图的网络拓扑结构有机结合的概率模型.采用贝叶斯网络分类对具有典型干旱特征的库车县土壤盐渍化情况进行监测,首先应用条件独立性测试原理建立贝叶斯网络结构,把研究区遥感数据进行离散化,然后应用贝叶斯定理作为分类原则,将每个像元分为像元最大概率的类别.研究结果表明该方法来分类6种地类的整体分类精度达到96%,并为该区盐渍地面积,空间分布等特征监测提供较好的依据.