GC53B-0517:
Environmental modeling in data-sparse regions: Mozambique demonstrator case
Friday, 19 December 2014
Guy Schumann1, Emily Niebuhr2, Kashif Rashid2, Vanessa M Escobar3, Konstantinos Andreadis4, Eni G Njoku4, Jeffrey C Neal5, Nathalie Voisin6, Florian Pappenberger7, Nuttavikhom Phanthuwongpakdee8, Paul D Bates5, Yi Chao9, Delwyn Moller9 and Paolo Paron10, (1)University of California Los Angeles, Los Angeles, CA, United States, (2)United Nations World Food Programme, Rome, Italy, (3)NASA Goddard Space Flight Center, Greenbelt, MD, United States, (4)NASA Jet Propulsion Laboratory, Pasadena, CA, United States, (5)University of Bristol, Bristol, United Kingdom, (6)PNNL, Seattle, WA, United States, (7)European Center for Medium-Range Weather Forecasts, Reading, United Kingdom, (8)King's College London, London, United Kingdom, (9)Remote Sensing Solutions, Inc., Sierra Madre, CA, United States, (10)UNESCO-IHE, Delft, Netherlands
Abstract:
Long time-series computations of seasonal and flood event inundation volumes from archived forecast rainfall events for the Lower Zambezi basin (Mozambique), using a coupled hydrology-hydrodynamic model, are correlated and regressed with satellite soil moisture observations and NWP rainfall forecasts as predictors for inundation volumes. This dynamic library of volume predictions can then be re-projected onto the topography to generate the corresponding floodplain and wetland inundation dynamics, including periods of flood and low flows. Especially for data-poor regions, the application potential of such a library of data is invaluable as the modeling chain is greatly simplified and readily available. The library is flexible, portable and transitional. Furthermore, deriving environmental indicators from this dynamic look-up catalogue would be relatively straightforward. Application fields are various and here we present conceptually a few that we plan to research in more detail and on some of which we already collaborate with other scientists and international institutions, though at the moment largely on an unfunded basis. The primary application is to implement an early warning system for flood inundation relief operations and flood inundation mitigation and resilience. Having this flood inundation warning system set up adequately would also allow looking into long-term predictions of crop productivity and consequently food security. Another potentially high-impact application is to relate flood inundation dynamics to disease modeling for public health monitoring and prediction, in particular focusing on Malaria. Last but not least, the dynamic inundation library we are building can be validated and complemented with advanced airborne radar imagery of flooding and inundated wetlands to study changes in wetland ecology and biodiversity with unprecedented detail in data-poor regions, in this case in particular the important wetlands of the Zambezi Delta.
