Characterizing Uncertainties in Hydrological Projections Under Climate Change in the Amazon Basin

Abstract:

Climatic Change projections are signed by irreducible uncertainties due to knowledge’s limitations, chaotic nature of climate system and human actions, which originate from multiple levels of the models chain. Such uncertainties affect the impact studies, mainly when associated to extreme events and difficult the decision-making process aimed at mitigation and adaptation. We carried out a study of the dispersion in hydrological projections using climate change projections from several climate models to feed the Distributed Hydrological Model of the National Institute for Spatial Research, MHD-INPE, applied in Amazonian sub-basins. We evaluate the dispersion in high, average and low discharges. Based on this approach we compare the effect of dispersion due to climate projections, bias correction approaches, non-stationary assumptions and available topographic data on the assessment of the estimation of number of affected people due to flood discharges. In addition to the uncertainties of climate change scenarios generated from different models, it is clear that the need of bias correction, the approach used to correct the bias and the absence of an accurate digital elevation model could introduce similar or bigger uncertainty than climate scenarios in the results.