Author(s): N. Nithila Devi; Soumendra Nath Kuiry

Linked Author(s):

Keywords: Probabilistic inundation forecasting; Sub-grid modeling; Computationally efficient forecasts; Ensemble weather forecasts

Abstract: The flood forecasting framework links to weather, hydrologic and hydraulic models. The weather forecast errors that are due to errors associated with observation, initial conditions, and model uncertainties can be addressed by deploying ensemble weather forecasts and post-processing techniques. On a global level, in order to increase the prediction of high-impact weather, at a lead time of 1–14 days, under the World Meteorological Organization, TIGGE facilitates access to medium-range forecasts from various centers across the globe. However, the observed data for extreme rainfall remains limited due to the low frequency of occurrence, which makes the post-processing of the ensemble forecasts difficult. Therefore, in this study, the predictive performance of raw ensemble forecast in the short term is examined for the flood-prone Adyar River basin that comprises the highly urbanized Chennai Metropolitan Area (CMA) in the state of Tamil Nadu, India. An existing hydrologic model is set up and calibrated for the entire basin for generating runoff response for a predicted rainfall, while a hydraulic model is used for the CMA for simulating the resulting flood. In this regard, a state-of-the-art local-inertial hydraulic model that runs on a course numerical grid, while still capturing the small urban features using sub-grids is employed to simulate the complex flood pattern in the highly urbanized CMA. The computationally efficient and reliable hydraulic model successfully simulates flood scenarios for all the ensemble members and eventually, flood hazard probability maps are generated for short-term forecasts. The obtained flood hazard maps are then compared with those the observed rainfall to realize the performance of raw ensemble forecasts in flood hazard mapping. The study shows that TIGGE ensemble forecasts can be utilized for predicting flood inundation in Chennai City with a reasonable lead time for issuing early warnings.

DOI: https://doi.org/10.1007/978-981-97-6009-1_17

Year: 2022