Simulation of groundwater level and groundwater salinity parameters of Ramhormoz plain using artificial neural network model and optimized artificial neural network model

Because of their high effectiveness and fewer expenses than other methods, groundwater models have been developed and used by hydrogeologists as water resource management tools. In this regard, many models have been developed, which propose better management to protect water resources. Most of these models require input parameters that are hardly available or their measurements are time-consuming and expensive. Among them, Artificial Neural Network (ANN) models inspired by the human brain are a better choice.

The present study simulated the groundwater level and salinity in Ramhormoz plain using ANN and ANN+PSO models and compared their results with the measured data. The data collected as inputs of the two models included minimum temperature, maximum temperature, average temperature, wind speed at 2 m altitude, minimum relative humidity, maximum relative humidity, average relative humidity, and sunshine hours gathered from 2011 to 2017.

The results indicated that the highest prediction accuracy of groundwater level and salinity was achieved by the ANN-PSO model with the logarithm sigmoid activation function. Thus, the MAE and RMSE statistics had the minimum and R^2 had the maximum value for the model.

Considering the high efficiency of artificial neural network models with Particle Swarm Optimization algorithm training, it can be used to make managerial decisions, ensure the results of monitoring, and reduce costs.