Evaluating Statistical Methods for Detecting Trend of Precipitation (Case Study: Kermanshah Province)

Fluctuation in precipitation is one of the most interesting issues in climate change. Precipitation has an important rule in sustainable agriculture and water management. Trend series analysis is one of the methods used to detect changes in climate parameters. The aim of the recent study is to evaluate different methods including Mann Kendal, rating Mann Kendal and Sen’s estimator slope tests to detect the trend variability of precipitation in monthly, seasonal and annual time scales in Kermanshah Province locating in the west of Iran. In addition, TFPW (Trend-Free Pre-Whitening) method was used to remove autocorrelation on time series data. Trend analysis of precipitation carried out using 16 synoptic and rain gauge weather stations with 32 years data i.e., 1992-2013.The results showed that, precipitation trend analysis using Mann Kendal and rating Mann Kendal tests had the same findings in monthly, and annual scales. The results revealed that Mann Kendal TFPW was the best method to detect trend variability of precipitation in the region. Annual precipitation had a negative trend in 50 percent of weather stations which was significant at 5 percent level. The results also indicated that most of the weather stations had negative trend in annual and seasonal time scales. Winter and spring seasons showed the greatest changes which was significant at 5 percent level of significance. Although, the summer season indicated a positive trend in the whole synoptic stations. In addition, a significantly negative trend discovered in the months of Mehr, Azar and Esfand at 5 percent level. It could be concluded that land use changes associated with migration from rural areas to the urban ones were responsible for precipitation trend variability.