Identification of Extreme Events in Climate Data from Multiple Sites

Extreme weather events have introduced significant losses and it still becomes an important issue in natural disaster management field. This paper proposes a procedure to identify extreme rainfall events occurred in a region, which is an important information for water resource management. The extreme identification can be easily conducted if it involves only climate data recorded in a single site. However, specific treatment has to be defined when the region has multiple sites. It particularly becomes a problem when one needs to study the large scale meteorological pattern dealing with the event. This paper combines the Extreme Value Theory (EVT) with some field considerations for identifying the dates of extreme rainfall events in Indramayu, Indonesia. The results show that extreme rainfall event is defined as rainfall observed in at least 5 stations with extreme amount in each station.