COMPUTER AIDED DAM-BREAK MODEL : CADAMBRK

Abstract

The text based interface dam-break program DAMBRK of the U.S. National Weather Service was modified to a graphic based interface program, called the Computer Aided Dam-Break (CADAMBRK) program. CADAMBRK consists of: (1) a graphical preprocessor; (2) a graphical postprocessor; and (3) a digital terrain model. CADAMBRK operates in a user friendly environment and its digital terrain model allows the user to derive terrain's cross-sectional geometries, required for flood routing, with little effort. CADAMBRK's graphic capability is much more sophisticated than that of DAMBRK: for example, flood surface may be superimposed on a 3-dimensional digitized terrain. The Manning flow roughness coefficients, one of the factors influencing the flood characteristics, were also studied. As the dam-break generated flood in Singapore is expected to be low in magnitude due to the relatively small dams, long breach time, and the generally flat topography, an effort has been made to study the roughness coefficients, for the local dam-break flood routing, using data from storm generated flood. Rainfall and runoff of 10 relatively small and highly urbanized catchments, of 8 different countries, were used. The analysis shows that when an effective overland flow roughness of 0.01 was used a probability of 85%, 82%, 57% and 95% may be achieved in predicting the peak flow rate, the time to peak flow, the averaged absolute flow rate deviation, and the time to centroid of the hydrograph, respectively, with an error of 30¾ or less. This roughness values was then tested on three heavy and long storms measured in the Upper Bukit Timah Catchme.1t in Singapore. Good results (with error well below 30¾) were obtained in the prediction of the peak flow rate and the time to peak flow for all the three storms studied. In flood flow prediction, the ability of a simulation model to predict, with high degree of certainly, the peak flow rate and its time of occurence is of utmost importance. With the successful verification runs, an effective roughness coefficient value of 0.01 was recommended for the dam-break flood prediction in Singapore.