Determination of a Design Rainfall Pattern by Comparing with its Effect on Streamflow on Greater Colombo Watershed in Sri Lanka

Abstract

IDF curves provide the rainfall quantity corresponding to a particular critical duration and the design return period. In order to carryout infrastructure designs with the use of high resolution mathematical models it is necessary to select the most appropriate temporal distribution of design rain event. In cases of sufficient data availability, literature recommends the use of pattern based location specific design rainfalls for optimum designs. Present study aimed to develop design rainfall patterns based on rainfall observations, and compare with the Alternating Block, Uniform Intensity, and Greater Colombo Flood Design Patterns by evaluating the runoff response from S CS HEC HMS model developed for a sub watershed of Greater Colombo Region. A literature review was conducted to select the design rainfall pattern presently used for water infrastructure engineering.30 years of 15-minute resolution rainfall data of Colombo Meteorological station were used to separate events. Events separation were carried out by nominating Minimum Inter Event Time (MIT) of 6hrs. 220 observed events were separated into six groups considering event durations. Analysis were carried out by developing dimensionless mass curve and percentile curve for each category. Design patterns were developed from percentile curve for each event duration. Design hyetographs were developed for each duration corresponding to design rainfall depth calculated with IDF curves for Colombo and selected data for analysis. Average recurrence Interval (ARI). Runoff response for all patterns were evaluated using the nature of the outflow hydrographs with reference to flood peak and time to peak . It was observed that highest runoff response was given by Enveloped curve developed with observed data. A high runoff variation was observed between rainfall patterns. ABM base pattern can be used with reliable confidence where there is no data for analysis. Criticality Index was developed to account for the pattern of design event with regards to flood peak and time of occurrence. Enveloped curve and 10% probability distribution pattern showed the highest criticality and ABM showed the most consistent criticality for all event categories.