Penentuan Batas Ambang Curah Hujan Penyebab Banjir (Studi Kasus DAS Ciliwung Hulu)

Abstract

One of the flood occurrences that always causes loss to the state almost every year is the Flood of Jakarta. The loss might be minimized if the flooding discharge is able to be early predicted. The objective of this study determined rainfall generating flooding. We applied multiple regression method to develop relation between rainfall and discharge in Ciliwung Hulu Sub-Watershed. Rainfall data from three stations (Gunung Mas, Citeko, and Katulampa) and flood discharge at Bendung Katulampa were used in this study. This study consist of three categories model, the first model was relation of every station against discharge Alert IV until Alert I, the second model was relation of stations against discharge in 1996 – 2008, and the last model was relation of stations against disharge which was not use zero data of rainfall or disharge with the same period. Model for estimating rainfall threshold using Multiple Linear Regression has been incompatible to be developed for flood early warning. Because small value of determination coefficien which was only less than 50%. Based on the first model the rainfall causes of flood was undeterminable. The second model gave rainfall threshold for Gunung Mas, Citeko and Katulampa as value as 183 mm, 176 mm, and 149 mm. Variance inflation factor for Gunung Mas, Citeko and Katulampa as value as 1.5, 1.4, and 1.1, p-value was 0.000, and R2 was 33.6%. The third model gave threshold rainfall as value as 253 mm, 253 mm and 125 mm, with variance inflation factor for Gunung Mas, Citeko and Katulampa as value as 1.4, 1.4, and 1.0, p-value was 0.000, with R2 was 24.6%. Based on the historical data, flood is usually happened on January and February. To minimize the loss, in that month the official functionary in upstream should be ready in reporting the discharge for the downstream official functionary.