Amazon deforestation effects on mean and extreme riverflows: insights from ecohydrological scaling

Abstract:

Land cover changes associated with both human activities and global change processes (such as drought, fire and invasive species) can lead to changes in byophisical processes that support hydrological processes in watersheds, leading to potential changes in longterm streamflow regimes. Traditional hydrological approaches have successfully associated stremaflow regimes to watershed physical attributes such as area via hydrological scaling. Yet, such approaches preclude the incorporation of biophysical processes, such as land cover changes, into the statistics of streamflow and scaling approaches. In this work, we test the hypothesis that mean and extreme riverflows in a watershed should scale not only with surface area, but also with leaf area. To test this hypothesis, we use a dataset consisiting of daily stremflow observations for 87 sub-basins of the Amazon basin, covering a range of watershed sizes from 1,617 km² to 4,680,000 km². We developed scaling relations of the type E[Q^r]= aA^b or E[Q^r]= c(LA)^d between the order r statistical moments of the probability distribution functions of mean and extreme riverflows (E[Q^r]) and both the watershed area (A) and the leaf area (LA) - which we define as ecohydrological scaling. Our results show that our proposed ecohydrological scaling represents the statistics of streamflow as well as traditional hydrological scaling relationships. However, ecohydrological scaling allows the assessment of the effects of land use-land cover changes on hydrological regimes. We used previously developed scenarios of Amazon deforestation to assess these effects and found maximum reductions of the order of 35% in the longterm mean and extreme riverflows as a result of deforestation. The approach that we propose uses the simplicity of hydrological scaling and improves it by incorporating ecohydrological processes into physical scaling. We illustrate its potential by applying it to future vegetation cover scenarios but highlight how, when combined with more sophisticated ecosystem and atmospheric models, ecohydrological scaling can become a powerful predictive tool. In summary, we present a new, ecohydrological scaling approach that combines biological and hydrological scaling approaches to assess the regional effects of global-change type processes.