Author(s): Astria Nugrahany
Linked Author(s): Astria Nugrahany
Keywords: Flash flood guidance (FFG); Bayesian FFG; Konto; Flash flood; Forecasting; Probabilistic rainfall threshold
Abstract: Flash flood is a short and sudden local flood with great volume. Characteristic of flash flood which is the short time of the basin to response compare to riverine flood, it raised consequences of short lead time to detect, forecast and issued a warning. The Flash Flood Guidance (FFG) is defined as a numerical estimate of the average rainfall over a specified area and time duration required to initiate flooding in small streams. Flash flood guidance in the United States is expressed in units of inches for 1-, 3-, and 6-hour durations. As an alternative to the existing approaches, a methodology to determine Bayesian probabilistic rainfall thresholds in forecasting occurrences of flash flood was explored in this paper. The computation was combined between rainfall range and antecedent moisture condition (AMC) of the catchment. The developed methodology was applied to the Konto River Basin in Indonesia. The Bayesian model was built using measured data from 1994–1999. The results of the Bayesian model show the advantage of following the Bayesian approach. The prior probability computed based on the total rainfall amount is updated with additional information such as rainfall ranges and antecedent moisture condition. The posterior probabilities are more realistic estimates of flooding. There are anomalies condition as well such as for AMC I for 3-hourly rainfall events and AMC II for 6-hourly rainfall events, the reason is only limited of data was used in this study. A larger dataset is required to eliminate this anomaly.
Year: 2018