Discrepancies between observation and ERA5 reanalysis in the Amazon deforestation context: A case study

Abstract

Deforestation in tropical areas is broadly reported to change the climate both locally and regionally. Warmer and drier conditions, as well as changes in precipitation patterns, are linked to deforestation in the Amazon. In this study, we identified two areas of distinct land use in Amazon: a preserved native forest and an increasingly deforested region southward. Due to the scarcity of available ground data, we propose assessing the impacts of deforestation on surface climate using two different datasets: a station-based reference product and the ERA5 reanalysis. However, as ERA5 does not include a recent and dynamic land use map in its development, an additional goal is to evaluate the potential discrepancies in the reanalysis for not accounting for these changes. Despite some consistent and similar patterns in relative humidity and low intensity (9th decile) precipitation, our results show, indeed, different trends among the datasets, with reference trends always more accentuated than in the reanalysis. Despite being broadly used in numerous studies, reanalysis data under intensive land use change and ungauged areas need to be used with caution to avoid inconclusive or misleading findings.