Prediction of Precipitation-Temperature Data and Drought Assessment of Turkey with Stochastic Time Series Models

Abstract

Throughout the geological history of Earth, there have been many changes in the climate system due to natural and external factors. In the past, it can be said that changes in climate were caused by natural causes, while today they are largely caused by human activities. Turkey is among the countries that will be affected by climate change. Therefore, In this study, a stochastic time series model was constructed to forecast the precipitation and temperature data of Turkey between 2020 and 2050. Seasonal Autoregressive Integrated Moving Average models were used to take into account the relationship between the data and seasonality factors. In addition, the most appropriate model for each station was established separately. The accuracy of the predicted data was tested by correlation test (r) and root mean square error (RMSE) test. As a result of the study, the average r value for temperature data was 99% and RMSE value was calculated as 1.46. For precipitation data, the average r value was calculated as 66% and RMSE value as 34.6. In addition, in this study, drought models for Turkey until 2050 were established and spatial and temporal evaluation of these models were made. These models were obtained by analyzing the data of uniformly distributed stations over Turkey between 1990 and 2050 with standard precipitation evapotranspiration index (SPEI). Different time scales (SPEI₃, SPEI₆, SPEI₉ and SPEI₁₂) were used in drought analysis. As a result of this study, drought return interval maps of Turkey and drought maps between 1990 and 2050 were created.