Investigation of probable maximum precipitation for disaster risk reduction in Sri Lanka

Abstract

The torrential rain during January - February 2011 caused many major reservoirs of Sri Lanka to spill and damaged over 400 medium and minor irrigation systems. In this extreme event, during a 30 day period commencing from 15th Dec 2010, Batticaloa experienced its average annual rainfall. The consequences were devastating, with nearly two million people being affected, the loss of 62 lives and severe damage caused to infrastructure. Furthermore, during the last one hundred years, many parts of the country experienced major floods due to extreme rainfall. Given this scenario and serious concerns about climate change, it is vital and prudent to review research methodologies in order to minimize the impact of natural disasters. The failure of a spillway of a large dam could be catastrophic. The design of dams, considering public safety is a challenging task in hydrology. The estimation of Probable Maximum Precipitation (PMP), which is indicative of probable extreme rainfall, is required to assess the adequacy of spillways under the criterion of Probable Maximum Flood (PMF). The objective of this paper is to provide a dependable and realistic method to estimate PMP in the context of Sri Lanka, considering its monsoonal weather pattern and the occurrence of cyclones. While several methods are available to estimate PMP, the hydro-meteorological method and the statistical technique developed by Hershfield are widely used procedures which are also recommended by the World Meteorological Organization. These two methods are characteristic of the deterministic and probabilistic approaches respectively. The physically based hydro-meteorological procedure has been adopted in this study and applied to the annual maximum daily rainfall series from seven meteorological stations. The research shows that maximizing moisture and using the corresponding wind run or maximizing moisture alone, are two realistic scenarios for making reliable estimates of PMP. The findings of research can be effectively utilized for developing PMP maps for Sri Lanka.