Sedimentation in stormwater management and road tunnel holding pond

The Berembang holding pond of the Stormwater Management and Road Tunnel (SMART), located in the state Selangor, Malaysia, is designed to store floodwater diverted from the confluence of the Klang and Ampang rivers during major storm events. The sub-catchments of the Klang and Ampang Rivers are among two sub-catchments that drain into the main Klang River. The excess floodwater diverted from the confluence of the Klang and Ampang Rivers were detained in the holding pond before it is released through a bypass tunnel at the Desa Park detention pond. The diverted floodwater from the river confluence carries a high amount of suspended sediment and organic debris that will continuously settle, and, consequently, reduce the capacity of the holding pond storage. The information on sediment concentration discharge relationship and the factors affecting the supply of sediment is imperative in providing qualitative insights and quantitative data for the development and evaluation of annual suspended sediment load estimation models. Deposited bed materials comprising a high composition of cohesive or finer-grained sediment generally constitutes a relatively higher contaminant level than a coarser-grain-sized bed sediment. Therefore, the spatial distribution data of the deposited sediment properties can be used implicitly to distinguish contaminated bed surface areas. The aim of this research is to determine the relationships between the rainfall, discharge and suspended sediment transport and the variations thereof during the dry and wet periods in a tropical urban catchment, to statistically examine the spatial variation of in situ wet sediment bulk density profiles of deposited sediment, to determine the sand mass and organic matter distribution based on in situ measured wet sediment bulk density profile data, simulated water velocity data of diverted flood water and deposition thickness results of a single grain-sized particle in the Berembang stormwater holding pond. Factors that have a major influence on the suspended sediment yields during both the dry and wet periods were examined. The clockwise and counter-clockwise hysteresis loops occurred during the dry period can be described as events with moderate total rainfall, rainfall intensity, average discharge and suspended sediment load. The extra sediment load during the dry period is a consequence of the removal of the sediment produced during the interstorm period by the first flush of water and the different time period between events. The complex loop events in the wet period occur with moderate moisture condition generated the highest suspended sediment load, which can be associated with the erosion caused by the high discharge flow at random sediment contributing areas. Two-way analysis of variance (ANOVA) was applied to investigate the significant effects of spatial sampling locations and depth variation on in situ wet sediment bulk density profiles of deposited sediment in the holding pond. Fifty-three sampling locations were hierarchically clustered into two groups, based on wet sediment bulk density profile data measured at six depth levels with 5 cm increments. The wet sediment bulk densities are significantly affected by depth variations at both sampling location groups. The sampling location groups have a significant main effect on the wet sediment bulk densities. There is a significant interaction effect between the sampling location groups and the depth levels on the mean wet sediment bulk densities. The effect of depth variations on the wet sediment bulk density measured at sampling location group one and two are significantly different. The measured wet sediment bulk densities indicate a stronger relationship with depth variation compared to both sampling location groups. The consolidation process rate from 5 to 10 cm depth level in group one sampling location is relatively higher than that of the group two sampling areas. Discriminant analysis (DA) was applied to spatially distinguish areas of relatively low and high composition of sand mass and organic matter content based on in situ measured wet sediment bulk density profile data, the simulated depth average water velocity variations and deposition thickness results of a single grain-sized particle (d50 = 0.375 mm). The spatial distribution sand and organic matter composition of surface sediment were predicted using Fisher’s linear discriminant functions. Based on in situ measured wet sediment bulk density profile data, the model correctly predicts 88.9 and 71.4% of the sampling locations consisting of relatively low and high sand weight percentages, respectively. For organic matter distribution, the model correctly predicts 70.0 and 86.7% of sampling locations consisting of relatively low and high organic matter percentage, respectively. Wet sediment bulk density profile measured at more than 15 cm depth levels indicates better predictor for the distribution of sand mass and organic matter composition area. Based on the simulated depth average water velocity variations, the model correctly predicts 88.9% and 100.0% of sampling locations consisting of relatively high and low sand mass percentage, respectively, with the cross-validated classification showing that, overall, 82.8% are correctly classified. Based on simulated deposition thickness results of a single grain-sized particle, the discriminant function significantly differentiated the two sampling location group areas composed of a relatively high and low sand mass content with the classification results showing that the model correctly predicts 66.7% and 100.0% of the surface sediment sampling locations consisting of relatively high and low sand weight percentage, respectively. As a conclusion, the suspended sediment loads generated during the dry period are highly influenced by the rainfall intensity. During the wet period, the suspended sediment loads are related to the total rainfall and storm event duration. Generally, the spatial variation in the wet sediment bulk density in the sampling location groups is predominantly influenced by the sediment grain size, consolidation rate and flow velocity variations in the deposited sediment. Fisher’s linear discriminant functions can be used to spatial classified sand mass and organic matter composition based on in situ measured wet sediment bulk density profile data, the simulated depth average water velocity variations and deposition thickness results of a single grain-sized particle.

Preview Text FPAS 2016 8 IR.pdf Download (1MB) | Preview

Actions (login required)

View Item