Identification of homogenous rainfall regions with trend analysis using fuzzy logic and clustering approach coupled with advanced trend analysis techniques in Mumbai city

Climatic conditions are significantly changing in the urban areas in the age of climate change. However, due the dynamic nature of urban areas and uneven distribution of meteorological stations, it is difficult to study rainfall pattern in the urban areas. Hence, in this study, trend and pattern of rainfall have been examined of Mumbai city using rainfall regionalization approach. For this, K-means and fuzzy C-means (FCM) clustering approaches have been used to examine the pattern of rainfall. The innovative trend analysis (ITA) and five tests of Mann-Kendall (MK) family have been utilized for anaylzing the rainfall trend. Result shows that both K-means and FCM techniques have identified two homogeneous rainfall regions in Mumbai exhibiting two distinctive rainfall patterns. The ITA curve shows that in the lower cluster, the data points are below 1:1 (45o) line, but in high and moderate cluster, they are above the line. This shows that there is a positive trend in the rainfall, but it is not monotonic. Further, the MK tests utilized in this study also shows positive trend in the rainfall in Mumbai. The result of this study may help in management of water resource and in urban flood mitigation in Mumbai city.