Simulating floodplain extent and inland water storage in the Amazon basin from a multi LSM perspective

Several modeling attempts have been conducted trying to improve the simulation of floods at several temporal and spatial scales worldwide. These attempts consider different modeling approaches and forcings, which can have a non-negligible impact on streamflows, inland water storage and floodplain extent. In the framework of the Global Land Water Cycle Benchmarking (GLWCB), the capability of simulating water and energy budgets by state-of-the-art land surface models (LSMs) and their impacts on predicting floods are evaluated. Fifteen LSMs were run for the 1980-2008 period using Princeton's meteorological forcings on a 3-hourly time step and at a 1° resolution. Three experiments are performed using Princeton's precipitation dataset rescaled to match monthly global GPCC and GPCP datasets and a daily dataset (HOP) specially developed for the basin. Surface and sub-surface runoffs derived from LSMs are used to force the Hydrological Modeling and Analysis Platform (HyMAP) river routing scheme (RRS) and simulated discharges are compared against observations at 146 gauges in the Amazon basin. Simulated floodplain extent and surface water storage are also compared against satellite-based products. Results show that simulated streamflows vary significantly as a function of both the LSM and precipitation used. It is also noted that floodplain extent is highly dependent on RRS parameters and further calibration must be performed.