Study of Mann-Kendall test performance in detecting the series of autocorrelation

In all stations the least amount of P was related to seasonal Mann-Kendall method that series data has autocorrelation. The stations have first-order correlation MK-PW method have highest amount of P. TFPW method has shown different behavior because of calculating the autocorrelation after removing trend. Comparisons of methods showed MK-VC, MK-BB and HERISH methods better than others and able to reduce the probability of first error. To further trend study and relationship between climate and hydrological parameters, rainfall and river flow in Nodehkhandooz station was studied using MASH approach in different months. The results showed that river flow has decreasing trend in the months when withdrawal of river is high. Most of climatic and hydrological data has autocorrelation, which makes Kendall method detect trend the series data without trend. In this study the effect of different removal aoutocorrelation methods were studied in Kendall trend detection of rainfall and runoff in 4 hydrometric stations in Golestan province. The methods used to remove the autocorrelation was MK, MK-PW, TFPW, MK-VC, MK-BB, SEAS and HERISH. The results showed that rainfall in Nodehkhormaloo station and river flow in Ghazaghly and Gholitapeh stations have significant increasing and decreasing trend respectively.