Modeling of land subsidence o Salmas plain by using adaptive neuro-fuzzy inference system (ANFIS)

Document Type : Research Article




In this study, with the aim of investigating ground subsidence, a fuzzy neural inference system model has been used to predict subsidence prone areas. To achieve this goal, seven important factors in subsidence of the region including land slope, digital elevation model, vegetation, water table or groundwater depth, distance from the road, distance from the river and distance from piezometric wells have been used. The information of the above factors was collected in ArcGis software and transferred to MATLAB software to implement the model.Using C-flodCV method, the data were randomly divided into three groups: 70% as training data, 20% as test data and 10% as validation data, and the data were introduced to MATLAB for training, testing and validation. Data were accurately trained and validated 〖10 ^ ^ (- 8). The data were modeled in several different membership functions including trapezoidal, triangular, Gaussian, two-sided Gaussian and bell function functions.The results showed that the trapezoidal membership function with a regression correlation coefficient of 0.86, and the Gaussian membership function with a regression correlation coefficient of 0.81 had the best performance in identifying areas prone to subsidence. According to the results of studies, it can be said that the eastern areas and the outlet of the plain have faced a large drop in groundwater and subsidence.Statistical studies also showed that groundwater has a total drop of more than 18 meters, which is reported to be 1.23 meters per year in the past few years. In the last two decades, the crisis has been exacerbated by over-extraction of groundwater in the agricultural sector, and a storage or feeding ratio of 0.03%, and climatic conditions.


Main Subjects


Articles in Press, Accepted Manuscript
Available Online from 15 June 2024
  • Receive Date: 27 March 2024
  • Revise Date: 11 June 2024
  • Accept Date: 15 June 2024