Akademik Veri Yönetim Sistemi
Arabian Journal of Geosciences, cilt.14, sa.2, 2021 (SCI-Expanded)
© 2021, Saudi Society for Geosciences.Precipitation regimes that change with global warming and climate changes affect the countries in environmental, economic, and social dimensions. The Marmara region is an important region located in the northwest of Turkey. The impact of economic, environmental, and social dimensions in the region is high. For this reason, the Marmara region is in a situation that can be affected more by climate change and drought. Precipitation forecasting is the first step for the management of agricultural planning, flood controls, and use of drinking water resources. Time series analysis is an important statistics tool that allows forecasting the amount of future precipitation based on the historical data analysis. Autoregressive Moving Average (ARMA), Autoregressive Integrated Moving Average (ARIMA), and Seasonal Autoregressive Integrated Moving Average (SARIMA) models are the most common statistical methods used to estimate precipitation based on time series. The ARMA, ARIMA, and SARIMA models are based on the assumption that past conditions will remain the same in the future. In this study, precipitation for the 9 cities in Turkey’s Marmara region is examined based on the 51-year (1969–2019) historical data and the ARMA, ARIMA, and SARIMA models are used to predict the precipitation in the next 60 months (up to 2024). While determining the model, the lowest AIC (Akaike information criterion) and AICc (corrected Akaike information criterion) are preferred and, generally, the AICc value is used to select the prediction model. After, the forecast measure errors of the models are checked with mean absolute error (MAE), root mean squared error (RMSE), and mean absolute scaled error (MASE) indicators. Finally, the ARIMA model is chosen as the most suitable model with the lowest estimation error.