Spatial and temporal contribution of flood flows in the upstream area of Bang Pakong river basin using rainfall runoff inundation model

Abstract

Flooding in the Bang Pakong River Basin is strongly influenced by spatial rainfall variability, catchment characteristics, and reservoir operations. This study investigated the spatio-temporal separation of flood hydrographs in the upstream Bang Pakong River Basin, Thailand, during the 2021 wet season using an integrated framework combining the distributed Rainfall–Runoff–Inundation (RRI) model and the Time–Space Accounting Scheme (T–SAS). The RRI model was calibrated and validated during 2020–2021 and 2022–2023, respectively, with reservoir releases from Khlong Phra Sathueng, Pra Pong, and Naruebodindrachinta reservoirs incorporated as boundary conditions. Simulations were conducted for the period 5 June–30 November 2021 using daily rainfall data from the Thai Meteorological Department (TMD) and the Hydro-Informatics Institute (HII), which encompassed six storm events. Model results demonstrated satisfactory performance, particularly at Kgt.3 station. Comparison of rainfall inputs revealed that rain gauge density significantly affected simulated peak discharge and hydrograph temporal patterns. The HII rainfall dataset better captured localized rainfall variability, while the combined TMD and HII dataset provided the most balanced representation of downstream runoff dynamics. T–SAS analysis at Kgt.3 station indicated that flood response was governed not only by rainfall magnitude but also by antecedent catchment conditions. The third and fifth storm events generated disproportionately large flow contributions due to elevated soil moisture and reduced storage capacity. Spatial analysis further showed that Phra Prong subbasin was the dominant contributor to downstream discharge, whereas Khlong Phra Sathueng subbasin, despite having the largest drainage area, highlighting the influence of rainfall distribution, land use, and basin storage characteristics. Reservoir operations moderated flood peaks and altered flow contribution patterns by storing excess inflows and gradually releasing water during subsequent periods. Furthermore, comparisons between multi-storm and individual storm simulations confirmed the critical role of antecedent rainfall in enhancing catchment wetness and amplifying downstream flood response. The integrated RRI and T–SAS framework provides a robust approach for tracing the temporal and spatial sources of floodwater. The findings demonstrate the critical role of antecedent conditions and reservoir operation in flood response and offer valuable tools for flood forecasting, reservoir operation, and flood mitigation.