Mid-Holocene climate simulations over Brazil using the Eta regional paleoclimate model

Abstract

Paleoclimate simulations are generally performed using coarse global climate models. However, validation of these simulations using a local dataset may be penalized by the coarse grid of the global models. Using a regional climate model, which benefits from the use of higher resolution in a limited area and specific period of time, may help improve the validation of the simulations and help to understand the climate of the past. The objective of this study is to evaluate the Mid-Holocene (MH) simulations for Brazil at 20 km spatial resolution using the regional Eta Model. The simulations are produced by nesting the Eta Model to the National Institute for Space Research (INPE) Global Spectral Atmospheric Model at T062 resolution and 28 vertical levels. Both global and regional models used the same orbital parameters to produce the Milankovic Cycles based on the Berger parameterization. The models adopted the typical carbon dioxide values for the present time and the MH. Downscaling simulations were performed with the Eta Model, resulting in simulations of 2 periods of time, the Eta 0k (present) and the Eta 6k (MH), each 10 years long. The evaluation compares these simulations against proxy data and other paleoclimate model simulations for the region. The difference between the two simulations, 6k and 0k, provides the changes between the two climatic periods. The Eta simulations indicate that the climate during MH was more humid over Northeast Brazil; this agrees partially with paleoclimate data in eastern Northeast Brazil. The Amazon region simulations were mostly drier during MH, in agreement with the paleoclimate data. A weaker convergence of winds in the MH, with winds blowing from the Amazon toward Southeast Brazil, affected the formation and the positioning of the South Atlantic Convergence Zone. Consequently, the associated moisture transport toward the Southeast was smaller; this reduced precipitation in the Brazilian southeastern, central-west, and southern regions, which agrees with the proxy data. The Eta Model simulated a cooler climate for the MH over the Northeast, Central, Southeast, and South Brazil. However, this simulated temperature showed less agreement against proxy data. The Amazon region was slightly warmer, and other transition regions showed no climate change. Overall, the results show that this modified version of the Eta Model is suitable for paleoclimate studies and provides added value over the driver model.