Forecasting of Water Quantities in the Iraqi Marshes using MSARIMAX Models

Abstract

The marshes of Iraq are particularly significant, as they are a historical legacy of the cities of the Mesopotamian civilization. The importance of the Iraqi marshes lies in the fact that they represent the richness of both history and nature, so they were placed among the treasures of the world that must be preserved and perpetuated. The Iraqi marshes suffered from neglect and experienced varying degrees of dryness, desiccation, and overfishing for different periods of time. Studying and analyzing the quantities of water in the marshes and the factors affecting them contributes to an attempt to address the problem of drought and develop effective solutions for it. In this research, the monthly rates of water quantities in the southern marshes of Iraq were estimated using the (SARIMA) model. The (SARIMAX) model was also employed, with the effect of three factors separately, which are the monthly rates of Iraqi river imports, the monthly rates of evaporation in dams and reservoirs, and the monthly rates of average temperatures in southern Iraq as external factors, as well as employing the (MSARIMAX) model to estimate the monthly averages of water quantities in the marshes with the influence of the three factors combined. In order to determine the best model for predicting the monthly rates of water quantities in the marshes, a comparison was made between the three models using the criteria (AIC), (Log-L), (RMSE), and (MAPE). The results of the comparison demonstrated the superiority of the MSARIMAX(1,1,2)(0,0,2)4 model over the other two models. The results of the estimation of the cross-correlation coefficients of the filtered time series indicated that there was a time lag with a positive effect of one month until the effect of water imports appeared, as well as a time lag with a negative effect of six months until the effect of monthly rates of evaporation appeared, and a time lag with a positive effect of five months until the effect of temperature. The values of the monthly averages of the quantities of water in the marshes of Iraq were predicted using the (MSARIMAX) model. The results of the prediction proved that there are noticeable differences in the predicted amounts of water compared to what it was in the previous months, as the monthly averages of the quantities of marsh waters will decrease over the next three years, as the amount of water may reach 114.53 million cubic meter in the marshes, and this decrease behaves seasonally in 2022 and 2023, and the seasonal effect decreases in 2024 and 2025.