Development of an integrated spatial distributed travel time method using GIS to model rainfall runoff in Bentong catchment, Malaysia

Rainfall runoff (RR) modelling over a catchment wide basis is a challenging task. The emerging of GIS technology has make the usage of distributed GIS more convenient nowadays. There many types of distributed model but in this study, it is performed using combination of kinematic wave approximation (KWA) and manning’s equation along with National Resources Conservation Services Curve Number (NRCS CN) to provide the estimation of effective rainfall. This study proposed an improved method by introducing the usage of new way to route the overland flow with enhancing the applicability of Digital Elevation Model (DEM) characteristics. One of parameter used in KWA calculation is distance a flow path or ridges, x (m). Previously, most models used assumption or estimated input namely overland flow, L and length of slope, ls to be applied for the whole grids. This approach deems not fully suit with the term distributed model, which every single grid supposedly has their own unique value. Therefore, an improvement proposed by multiplying the rainfall intensity with the longest perpendicular distance in a grid. Then, to predict discharge at the outlet, time area (TA) approach is applied. This model named as Spatial Distributed Travel Time (SDTT) in this study. Besides SDTT, spatial lumped model (SLM) is also applied to compare the result of both model. From DEM resolution and sensitivity analysis (SA), it is determined that using 30 m DEM size with input of x equal to 42.2 m gave the best predicted hydrograph. Using coarser DEM resolution increase the peak discharge and shorten the time to peak. SDTT model performed better than SLM when compared to the observed discharge. For calibration, SDTT gave result of NSE = 0.86; PBIAS = -1.95; r = 0.87 (p<0.0001) and SLM produce NSE = 0.78; PBIAS = -21.58; r = 0.85 (p>0.0001). In validation, SDTT result NSE = 0.81; PBIAS = 0.17; r = 0.88 (p<0.0001) and SLM produce NSE = 0.62; PBIAS = 15.87; r = 0.85 (p<0.0001). Furthermore, SDTT also perform better in predicting peak discharge (PD and time to peak (TP) compare to SLM. Land use land cover change (LULC) effect on flood hydrology of Bentong catchment also indicate there is significant of LULC changes from period of 2000 to 2016 found out through Chi- square goodness of fit test with result (χ² = 7.403, p –value = 0.0006 (P <0.05)). There is also significant increase of peak discharge and decrease in time to peak and travel time.

Preview Text FPAS 2018 32 IR.pdf Download (3MB) | Preview

Actions (login required)

View Item