Simulation of Temperature and Rainfall Using the Seasonal Auto-Regressive Integrated Moving Average (SARIMA) Model and Trend Analysis in Arid Regions

This paper analyzes the temperature and rainfall data series collected by Dezful stations in a 31-year period (1986 to 2017) in order to evaluate the magnitude of these changes statistically and to forecast their behavior for the 2018-2020 period using SARIMA models. The Mann-Kendall test was used to analyze climate change in the past and future. The results show that rainfall has a decreasing trend and minimum and maximum temperatures have increasing trends. The results of the SARIMA model show that the coefficient of correlation (r) between the observed and forecasted values was 0.95, 0.9 and 0.58 for rainfall, minimum temperature, and maximum temperature and the mean absolute error (MAE) was 1.24, 1.45 and 20.24 for them, respectively. The results of trend analysis reveal that Mann-Kendall s statistics (Z-value) for the data on minimum temperature, maximum temperature and rainfall are 3.81, 1.78 and -2.71, respectively implying a descending trend for temperature and an ascending trend for rainfall. Minimum and maximum temperatures have been rising at the rates of 0.07 and 0.04°C per year, but they are forecasted to have increased by 0.084 and 0.06°C by 2020, respectively. The rate of rainfall variation will decrease from 4.4 mm to 4.85 mm per year. Improved understanding of recent climate change helps to elucidate the impacts and vulnerability of the local population in order to implement the most appropriate practices to cope with climate change and manage the changing situation in a better way.

Simulation of Temperature and Rainfall Using the Seasonal Auto-Regressive Integrated Moving Average (SARIMA) Model and Trend Analysis in Arid Regions Keywords:

Climatic Factors , Trend Analysis , Multiplicative Time Series Models , Forecasting