Document Type : Research Paper
Authors
1 urmia university
2 Associate Professor of Water Engineering, Urmia Lake Research Institute, Urmia University, Iran.
3 Associate Professor, Department of Water Resources Engineering, Faculty of Agriculture and Urmia Lake Research Institute,, Urmia University, Urmia, Iran
4 Associate Professor, Department of Watershed Management Engineering, Faculty of Natural Resources, Urmia University, Urmia, Iran
Abstract
Salmas plain represents one of the most critical areas in the country experiencing subsidence. Identifying the underlying factors contributing to subsidence becomes crucial in mitigating further land sinking. Consequently, this research seeks to investigate the intensity of subsidence in the Selmas plain using ArcGIS software and fuzzy logic to analyze the various factors influencing this phenomenon. Initially, comprehensive data on eight factors influencing subsidence, including groundwater level drop, well exploitation flow rate, aquifer storage coefficient, transmissibility coefficient, precipitation, Digital Elevation Model (DEM), soil texture, and bedrock depth, was collected. Subsequently, raster maps for each factor were extracted at the aquifer level. The standardization of these layers was conducted using fuzzy membership functions, considering the varying impact of each factor on land subsidence. These fuzzy operators (Gamma, OR, AND, SUM, and PRODUCT) were then applied and combined using overlay functions, resulting in a comprehensive unit map representing the subsidence intensity across the region. To determine the most optimal fuzzy-operation based maps, the outcomes were compared with field observation data, and the ROC curve performance index was applied for thorough control and validation. Throughout this process, due consideration was given to the observed subsidence patterns in the plain. The fuzzy OR operation exhibited the lowest overlap with the observed subsidence in the region, boasting an AUC of 0.693. Conversely, the Gamma models demonstrated the highest overlap with the subsidence observed in the plain, with an AUC above 70%. As a result, the Gamma 0.9 model was chosen for this research, presenting an AUC of 0.805. The findings underscore the critical nature of subsidence in the eastern part of the aquifer. Of the total area of Salmas plain, 25%, equivalent to 93 square kilometers, has subsidence with a very high susceptibility.
Keywords
)