Floods in the Pampas: Insights from over a Decade of Satellite Observations

Thursday, 18 December 2014
Sylvain Kuppel1,2, Javier Houspanossian1, Marcelo D Nosetto2 and Esteban G Jobbagy2, (1)CONICET, Buenos Aires, Argentina, (2)National University of San Luis, San Luis, Argentina
There is a general need to understand floods in very flat sedimentary regions, particularly in the context of current climate and land use changes. While most landscapes respond to extreme rainfall inputs by increasing their liquid water outputs, very flat and poorly drained ones have little capacity to do this and their most common responses include (i) increased water storage leading to rising water tables and floods, (ii) increased evaporative water losses and, only at very high levels of storage, (iii) increased liquid water losses. We explored the relative importance of these pathways in the extensive plains of the Argentine Pampas, where two significant flood episodes took place in 2000-2003 and 2012-2013. We combined a new remote sensing estimate of 8-day surface water cover (SWC, based on MODIS albedo product), monthly terrestrial water stock (TWS, based on GRACE gravity recovery data), daily precipitation (PP, based on TRMM data), evapotranspiration (ET, based on MODIS NDVI and surface temperature data). Focusing on two of the most flood-prone areas of the Argentine Pampas (60 000 km² each) we found that in 2000-2003, SWC reached 18 and 15% in each region, while the accumulated precipitation anomalies respectively approached 500 and 800 millimeters since the beginning of the flood cycle. In 2012-2013, SWC peaked at 8 and 7% after respective precipitation excesses of 420 and 250 mm. There was a direct, somewhat linear coupling between SWC and TWS in the western region during both flood events, while significant hysteresis features in the eastern region hint at a decoupling between the terrestrial water reservoir and its surface component. The relationships between yearly SWC, TWS, PP and ET point at a surface-water-balance-driven water cycle in the east, with below-ground water storage exerting a secondary and delayed control on the somewhat rapid flood dynamics. On the contrary, a groundwater-connected water cycle is suggested in the west, where groundwater would buffer water excesses and, when water tables rise, triggers longer-sustained flood events (path (i) defined above). These two hydrological pictures are discussed, notably taking into account the reduction of TWS by nearly 100 mm between the two flood events.