Proof of Concept of an Altimeter-Based River Forecasting System for Transboundary Flow Inside Bangladesh

Recent work by Biancamaria et al. (Geophysical Research Letters, 2011) has demonstrated the potential of satellite altimetry to forecast incoming transboundary flow for downstream nations by detecting river levels at locations in upstream nations. Using the Ganges-Brahmaputra (GB) basin as an example, we assessed the operational feasibility of using JASON-2 satellite altimetry for forecasting such transboundary flow at locations further inside the downstream nation of Bangladesh by propagating forecasts derived from upstream (Indian) locations through a hydrodynamic river model. The 5-day forecast of river levels at upstream boundary points inside Bangladesh were used to initialize daily simulation of the hydrodynamic river model and yield the 5-day forecast river level further downstream inside Bangladesh. The forecast river levels were then compared with the 5-day-later nowcast simulation by the river model based on in-situ river level at the upstream boundary points in Bangladesh. Results show that JASON-2 retains good fidelity at 5-day lead forecast with an average RMSE (relative to nowcast) ranging from 0.5 m to 1.5 m and a mean bias (underestimation) of 0.25 m to 1.25 m in river water level estimation. Based on the proof-of-concept feasibility, a 4 month-long capacity building of the Bangladesh flood forecasting agency was undertaken. This facilitated a 20-day JASON-2 based forecasting of flooding during Aug 1, 2012 to Aug 20, 2012 up to a 5 day lead time in a real-time operational environment. Comparison against observed water levels at select river stations revealed an average error of forecast ranging from -0.4 m to 0.4 m and an RMSE ranging from 0.2 m to 0.7 m. In general, this study shows that satellite altimeter such as JASON-2 can indeed be an efficient and practical tool for building a robust forecasting system for transboundary flow.