Towards mechanistic understanding and prediction of water, sediment and carbon budgets of large tropical floodplains

Abstract:

The ability to quantify water movement between the main river channel and floodplain is fundamental for the examination of ecohydrological processes and assessment of sediment and nutrient budgets of floodplain reaches. Precise hydraulic modelling is difficult at large scales because of the lack of appropriately scaled information to represent floodplain topography and surface roughness, both of which vary in space and time. Determining river-floodplain exchanges through field measurements is also difficult because currently available technologies still present limitations to measure flow over long stretches of vegetated levees and innumerous floodplain channels. Such limitations have led to over simplification of floodplain hydrological processes in large scale models and lack of data for calibration and validation. However, inaccuracies in representing variations in flows in these pathways propagate into calculations and interpretations of the functioning of physical and biogeochemical processes on floodplains. We computed river-floodplain exchanges for a large (~2400 km²) flood basin encompassing the Lago Grande de Curuai in the lower reach of the Amazon River using two-dimensional inundation modeling. We analyzed seasonal variation of suspended sediment and organic carbon fluxes by combining a daily water budget with field measurements of sediment and carbon concentrations and pools. Along with the river-floodplain exchanges, the quantification of aquatic carbon fluxes included carbon fixation by aquatic plants, biogenic gas evasion, and fluxes in other hydrological pathways. Total input of organic carbon was estimated as 1,250 Gg C y^-1 and total output as 1,850 Gg C y^-1. Organic carbon loads from the Amazon River represented 10% of all input sources to the floodplain. Floodplain discharge to the river represented 11% of the outputs. The main input and output were net primary production of macrophytes (48%) and CO₂ outgassing (85%), respectively. Based on our results, we explore concepts related to mechanistic assessment of the water balance of large tropical river system and discuss critical processes that require finer representation in large-scale models to dynamically predict flooding and recession of floodplains and the lateral redistribution of water and materials.