The Forecasting Models for Amount of Water in Nam Un Dam, Sakon Nakhon Province

Abstract

This study aimed to create forecasting models for the daily, weekly, and monthly water volumes of the Nam Un dam, Sakon Nakhon Province, using multiple regression analysis and the four-time series analysis techniques. The data consist of the water volume (in million cubic meters), water surface area, drainage volume, and related meteorological variables recorded by the Lam Num Un Irrigation Project from January 1, 2017, to November 30, 2021. To create the forecast models, the data were divided into two parts; the last 12 periods of data were used to compare the performance of the model, while the remaining data were used to create the model. The model’s efficiency was assessed based on the lowest MAPE criteria. The results showed that the Holt and Winter’s multiplicative method was the most accurate daily water volume forecasting model for the Nam Un dam, followed by the three-point simple moving average method. For weekly and monthly water volume forecasting models, the regression model had the lowest MAPE, followed by the three-point simple moving average method.