Can atmospheric reanalysis datasets be used to reproduce flood characteristics?

Wednesday, 17 December 2014: 3:25 PM
Konstantinos Andreadis1, Guy Schumann2 and Dimitrios Stampoulis1, (1)NASA Jet Propulsion Laboratory, Pasadena, CA, United States, (2)University of California Los Angeles, Los Angeles, CA, United States
Floods are one of the costliest natural disasters and the ability to understand their characteristics and their interactions with population, land cover and climate changes is of paramount importance. In order to accurately reproduce flood characteristics such as water inundation and heights both in the river channels and floodplains, hydrodynamic models are required. Most of these models operate at very high resolutions and are computationally very expensive, making their application over large areas very difficult. However, a need exists for such models to be applied at regional to global scales so that the effects of climate change with regards to flood risk can be examined. We use the LISFLOOD-FP hydrodynamic model to simulate a 40-year history of flood characteristics at the continental scale, particularly over Australia. LISFLOOD-FP is a 2-D hydrodynamic model that solves the approximate Saint-Venant equations at large scales (on the order of 1 km) using a sub-grid representation of the river channel. This implementation is part of an effort towards a global 1-km flood modeling framework that will allow the reconstruction of a long-term flood climatology. The components of this framework include a hydrologic model (the widely-used Variable Infiltration Capacity model) and a meteorological dataset that forces it. In order to extend the simulated flood climatology to 50-100 years in a consistent manner, reanalysis datasets have to be used. The objective of this study is the evaluation of multiple atmospheric reanalysis datasets (ERA, NCEP, MERRA, JRA) as inputs to the VIC/LISFLOOD-FP model. Comparisons of the simulated flood characteristics are made with both satellite observations of inundation and a benchmark simulation of LISFLOOD-FP being forced by observed flows. Finally, the implications of the availability of a global flood modeling framework for producing flood hazard maps and disseminating disaster information are discussed.