During the past years, the growth of population, development of industries and agriculture have brought an increase in the extraction of groundwater resources. The increase in extraction has been done in most of the study areas of the country without taking into account the capacity of the aquifers, while logically, the exploitation of an aquifer should be done at the limit of its dynamic storage. The purpose of this research is to model and predict the groundwater level of Najaf Abad plain under the conditions of climate change. After collecting the data in the Najaf Abad plain, due to the different climatic conditions of the region, it was divided into five regions for more accurate modeling, and according to the effective parameters in the groundwater level, from four different patterns of input parameters to the models. These five areas include parts of the Nekoabad irrigation network, Zayndeh-Roud river boundary and Khmiran irrigation network, and the western and northern elevations of the Najaf Abad plain. The input parameters include water level data of wells, meteorological data including temperature and precipitation, hydrometric data including river flow, water withdrawal from wells in the area and data related to irrigation for the years 2011 to 2021. After checking outlier data and estimating missing data and data preparation, seventy percent of the data related to each area is used for modeling the training through machine learning algorithms, including support vector machine, random forest, and enhanced gradient modeling. Thirty percent was used to eva‎luate the models. The enhanced gradient algorithm was identified as the superior algorithm in the most regions. The mean square errors for the validation data of the enhanced gradient model for five regions are 0.85, 0.87, 0.04 and 0.23, respectively. Meteorological data, including precipitation and temperature, were extracted from the sixth global climate change reports and downscaled for Najafabad region. Then these data were entered into the enhanced gradient model and the groundwater level in Najaf Abad Plain was predicted for the years 2021 to 2032, according to the climate data of the 6th Global Climate Change Report. The results of groundwater level forecasting in the conditions of climate change and no change in water extraction from wells and dryness of ZayandehRud with the enhanced gradient model, the maximum drawdown is 8 meters in the right irrigation network of Nekoabad and 5.5 meters for near the river.

سعيد اسلاميان

حميدرضا صفوي , سعيد سلطاني كوپائي

منوچهر حيدرپوراسفرجاني , سعيد سلطاني كوپائي