Effects excluding the effect of autocorrelation in analysis trend of hydro-climatic variables (Fars Province)

Abstract

Trend analysis and understanding of the historical volatility in the climate system, is one of the most important requirements in research on climate change. In order to study the trend, there are different methods covering parametric and non-parametric approaches. The non-parametric Mann - Kendall methods are widely used to study the hydrological series. But the autocorrelation in the hydrological series increases probability occurrence of type I error. In this study in order to evaluate the effects of autocorrelation, the time series trend without excluding the effect of autocorrelation and then by applying TFPW method and excluding the effect of autocorrelation were analyzed. The results showed that by removing the effect of autocorrelation, number of stations with significant negative trend was decreased. Therefore, in annual and monthly scales, none of rainfall data series in 22 rain gauge stations were significant at the 95 percentage confidence level. Therefore, it revealed that in evaluating the hydro-climatological data, the effects of autocorrelation must be removed from time series to provide accurate and reliable results.

Keywords

• trend analysis
• mann- kendall test
• autocorrelation
• tfpw method
• fars province