Processes of streamflow generation in a small rainforest catchment in Central Amazonia
Martin
George
Hodnett, Free University of Amsterdam, mhodnett@lineone.net
(Presenting)
Javier
Tomasella, CPTEC/INPE, javier@cptec.inpe.br
Luz
Adriana
Cuartas, CPTEC/INPE, adriana@cptec.inpe.br
Antonio
Donato
Nobre, INPA, anobre@ltid.inpe.br
Maarten
Waterloo, Free University of Amsterdam, watm@geo.vu.nl
Sylvia
Mota de
Oliveira, INPA, smoli@inpa.gov.br
Camilo
Renno, CPTEC/INPE, camilo@cptec.inpe.br
In 2002 and 2003, runoff from the Asu catchment (6.37 km2) was 1455 mm and 832 mm, of which 45% and 31% was stormflow respectively. Earlier studies in the Barro Branco catchment (1.3 km2) in Reserva Ducke (with similar geology and soils) showed that stormflow was 9% of total runoff in 1981, 1982 and 1983. Other studies in the same catchment suggested that the storm runoff was mainly generated from the valley floor, where the water table is close to the soil surface, leading to saturation excess overland flow (SOF). The Asu data confirm that the valley floor is the source of the storm runoff. The water table in the valley floor is controlled by two factors, (i) in the short-term, local vertical recharge direct to the water table and (ii) in the longer term, the discharge of groundwater from beneath the plateau and slope areas, fed by deep drainage (recharge) from these areas. There is also some evidence that interflow and return flow contribute to storm runoff in very large events.
The main reason for the large difference in storm runoff between the catchments is that in the much larger Asu catchment, the proportion of the catchment occupied by valley floor (32%), with a shallow water table, is much greater than in the Barro Branco. In addition, the rainfall was well above average in 2002. These results show that it may not be appropriate to scale up from process studies in very small catchments, because of scale-related changes in processes.