RMS Error and Correlation for Hardinge Bridge on Ganges River 1.60 1.000 RMSE, m 0.900 0.80 0.850 Danger Level 14.25 m Correlation 0.950 1.20 RMSE_Jason-2 0.40 RMSE_FFWC 0.800 Correlation_Jason-2 Correlation_FFWC 0.00 0.750 1 2 3 4 5 6 7 8 Lead Time, Day (a) RMS Error and Correlation for Bahadurabad on Jamuna River 1.00 1.000 0.950 0.900 0.850 0.800 0.750 0.50 0.25 Danger Level 19.50 m 0.700 RMSE_Jason-2 0.650 RMSE_FFWC 0.600 Correlation_Jason-2 Correlation RMSE, m 0.75 0.550 Correlation_FFWC 0.500 0.00 1 2 3 4 5 6 7 8 Lead Time, Day (b) Figure 6. Assessment of JASON-2 forecasting skill for warning stations Hardinge Bridge (a) and Bahadurabad (b). For location of the sta- tions, refer to Figure 1. The lines without dark circles refer to the Govt's official 3 day forecast. The 3rd day was often times not reported in www.ffwc.gov.bd. as the current FFWC system combines persistence with extensive human intervention and applies complex hydrologic and atmospheric models for forecasting boundary flow conditions. On the other hand, the JASON-2 system is entirely hydrodynamic model-based (non-persistence based) and automated (with no routine human intervention for quality check). 5. ConClusion To the best of our knowledge, the rigorous assessment of flood forecast skill up to 8 days of lead time for a real-time 34 and independently managed (by a stakeholder agency) satellite altimeter-based system is perhaps the first of its kind reported in literature. Such an assessment provides a consistent and verifiably affirmative response to the question we posed earlier-Can a Radar Altimetry Satellite Deliver on the Promise of an Operational and Real-Time Transboundary Flood Forecasting System for Floodprone Bangladesh? When a satellite-based system consistently generates high SNR (and low RMSE below 1 m) even at the 8 days lead time at most flood warning locations, there remains little doubt about the tremendous potential that satellite ieee Geoscience and remote sensing magazine september 2014http://www.ffwc.gov.bd