Monitoring annual precipitation changes in Dezful plain with statistical analysis and time series

 Predicting and studying the trend of climate variables in the future plays an important role in the optimal management of water resources. Different methods are used to determine the trend of change. One of the most common methods of trend change analysis is time series analysis. Time series is a set of observations about a variable that is measured at discrete points in time, usually at equal distances, and arranged in chronological order

  In the present study, the trend of precipitation changes in Dezful plain during 32 years was investigated and by selecting the appropriate time series model, a forecast was made for the next ten years. Man-Kendall’s non-parametric test was used to investigate the trend of precipitation changes.

 The result of this test showed that the annual precipitation of Dezful had a decreasing trend due to having a Man-Kendall statistic of -1.6. To select the appropriate time series model, data preparation (trend elimination and normalization) was performed first. Data stagnation was assessed with autocorrelation (ACF) and partial autocorrelation (PACF) charts. Using the differentiation method, the data became static (eliminating the mean trend) by applying one-time differentiation. By static data, random models were used to predict the average annual precipitation. Then, by fitting different Arima models and considering the criteria of T, P-VALUE less than 0.05 and Bayesian information criterion (BIC), the Arima model (3,1,1) was selected as the most appropriate model and to verify this the model was predicted for the period 2011 to 2018. The validation results showed that the prediction of this model is acceptable according to the actual values. Then, based on this model, a forecast was made for the next ten years from 2019 to 2028, which is predicted that the precipitation trend will decrease for the next period.