Analysis of change point of rainfall and its trend in Doimukh (Itanagar), Arunachal Pradesh

Abstract

In this study change point and the trend was detected in monthly, seasonal and annual rainfall series for Doimukh (Itanagar), Arunachal Pradesh for the years 1988 to 2016. The Pettitt’s test was used to detect change point in monthly, seasonal and annual rainfall series. For detecting the trend in the same series 05 tests were used i.e. Least squares linear regression (LR) test, Mann–Kendall test, Spearman rank correlation (SRC) test, Sen’s slope and innovative trend analysis technique. A change point in the months of February (2007), May (1998), June (2008), July (2004,2008), September (2004), October (1994,1999), annual (2008), monsoon (2008), pre-monsoon (2008) and post-monsoon (2006) rainfall series. Mann- Kendall and Spearman’s rho tests statistics indicated a significant falling trend in the months of January, February, May to October. A significant falling trend was also observed in the seasonal and annual rainfall series using both the methods. The results of innovative trend analysis indicated that there was falling trend in the months January, February, and May to December. Falling trend also existed in the post monsoon, pre-monsoon, monsoon and annual rainfall series.The results of Sen’s slope method indicated that there was falling trend in all the selected rainfall series.LR test indicated, a negative value of slope i.e. falling trend in monthly, seasonal and annual rainfall series. However, estimated value of the slope was significant in the months of January, March, April and November and December. By comparing the results all five methods of trend detection tests, it was observed that the results of the Mann-Kendall, Spearman’s Rho test and innovative trend analysis technique were exactly matching in seasonal, annual rainfall series. Mann-Kendall and Spearman’s Rho and innovative trend analysis were also exactly matching in rainfall series of the months January, February, and May to October only. Whereas, varying negative values of the slope were observed in all monthly, seasonal and annual rainfall series in both Sen’s slope and LR test methods. However, a statistical test for slope in LR method gave different trend inference compared to all other selected methods.Therefore, Mann-Kendall, Spearman’s rho test and innovative trend analysis technique can be successfully applied for detecting the trend in rainfall time series of the study area.Moreover, Sen’s slope is a better choice for obtaining the value of slope in the trend line.