Assessing the Multi-Temporal Changes of Soil Salinity and Waterlogging in Ismailia City, Egypt

Abstract

In this study, multi-temporal changes in soil salinity and water logging in Ismailia City, Egypt were analysed based on classification results acquired using the pixel-based and object-based approaches. Landsat images from 1986, 2000 and 2006 were used to carry out the image classification and ground truth data were collected from reconnaissance surveys, geological maps, Google Earth and personal knowledge. In pixel-based analysis, supervised classification was performed using maximum likelihood as the classifier with ERDAS Imagine. On the other hand, object-oriented image analysis was evaluated through eCognition software using the feature tool as the basis for classification. Results of the classified images show that the object-oriented approach gave an overall accuracy of 95.9%, 96.1% and 96.1% which was slightly higher than the pixel-based analysis with overall accuracy of 82.26%, 82.14% and 83.21% respectively. Change detection techniques including Image ratioing, image differencing and change vector analysis were deployed and analyzed to depict the history of changes as well as the spatial distribution of changes between the period of study. Results show that the study area experienced a decrease of 22.15KM2 and an increase of about 155.01 KM2 in soil salinity and waterlogging yearly between 1986 and 2006, it was also observed that waterlogged areas decreased towards the northeast and increased towards the west, southwest and southeast part of the study area. These findings suggest that remote sensing is a useful tool to detect saline soils and waterlogged areas.