Two questions motivated this study: 1) Will meteorological droughts become more frequent and severe during the twenty-first century? 2) Given the projected global temperature rise, to what extent does the inclusion of temperature (in addition to precipitation) in drought indicators play a role in future meteorological droughts? To answer, we analyzed the changes in drought frequency, severity, and historically undocumented extreme droughts over 1981–2100, using the standardized precipitation index (SPI; including precipitation only) and standardized precipitation-evapotranspiration index (SPEI; indirectly including temperature), and under two representative concentration pathways (RCP4.5 and RCP8.5). As input data, we employed 103 high-resolution (0.44°) simulations from the Coordinated Regional Climate Downscaling Experiment (CORDEX), based on a combination of 16 global circulation models (GCMs) and 20 regional circulation models (RCMs). This is the first study on global drought projections including RCMs based on such a large ensemble of RCMs. Based on precipitation only, ~15% of the global land is likely to experience more frequent and severe droughts during 2071–2100 versus 1981–2010 for both scenarios. This increase is larger (~47% under RCP4.5, ~49% under RCP8.5) when precipitation and temperature are used. Both SPI and SPEI project more frequent and severe droughts, especially under RCP8.5, over southern South America, the Mediterranean region, southern Africa, southeastern China, Japan, and southern Australia. A decrease in drought is projected for high latitudes in Northern Hemisphere and Southeast Asia. If temperature is included, drought characteristics are projected to increase over North America, Amazonia, central Europe and Asia, the Horn of Africa, India, and central Australia; if only precipitation is considered, they are found to decrease over those areas.

One of the most important anthropogenic influences on climate is land use change (LUC). In particular, the Amazon (AMZ) basin is a highly vulnerable area to climate change due to substantial modifications of the hydroclimatology of the region expected as a result of LUC. However, both the magnitude of these changes and the physical process underlying this scenario are still uncertain. This work aims to analyze the simulated Amazon deforestation and its impacts on local mean climate. We used the Common Land Model (CLM) version 4.5 coupled with the Regional Climate Model (RegCM4) over the Coordinated Regional Climate Downscaling Experiment (CORDEX) South America domain. We performed one simulation with the RegCM4 default land cover map (CTRL) and one simulation under a scenario of deforestation (LUC), i.e., replacing broadleaf evergreen trees with C₃ grass over the Amazon basin. Both simulations were driven by ERA Interim reanalysis from 1979 to 2009. The climate change signal due to AMZ deforestation was evaluated by comparing the climatology of CTRL with LUC. Concerning the temperature, the deforested areas are about 2 °C warmer compared to the CTRL experiment, which contributes to decrease the surface pressure. Higher air temperature is associated with a decrease of the latent heat flux and an increase of the sensible heat flux over the deforested areas. AMZ deforestation induces a dipole pattern response in the precipitation over the region: a reduction over the west (about 7.9%) and an increase over the east (about 8.3%). Analyzing the water balance in the atmospheric column over the AMZ basin, the results show that under the deforestation scenario the land surface processes play an important role and drive the precipitation in the western AMZ; on the other hand, on the east side, the large scale circulation drives the precipitation change signal. Dipole patterns over scenarios of deforestation in the Amazon was also found by other authors, but the precipitation decrease on the west side was never fully explained. Using budget equations, this work highlights the physical processes that control the climate in the Amazon basin under a deforestation scenario.

We analyze the local and remote impacts of climate change on the hydroclimate of the Amazon and La Plata basins of South America (SA) in an ensemble of four 21st century projections (1970–2100, RCP8.5 scenario) with the regional climate model RegCM4 driven by the HadGEM, GFDL and MPI global climate models (GCMs) over the SA CORDEX domain. Two RegCM4 configurations are used, one employing the CLM land surface and the Emanuel convective schemes, and one using the BATS land surface and Grell (over land) convection schemes. First, we find considerable sensitivity of the precipitation change signal to both the driving GCM and the RegCM4 physics schemes (with the latter even greater than the first), highlighting the pronounced uncertainty of regional projections over the region. However, some improvements in the simulation of the annual cycle of precipitation over the Amazon and La Plata basins is found when using RegCM4, and some consistent change signals across the experiments are found. One is a tendency towards an extension of the dry season over central SA deriving from a late onset and an early retreat of the SA monsoon. The second is a dipolar response consisting of reduced precipitation over the broad Amazon and Central Brazil region and increased precipitation over the La Plata basin and central Argentina. An analysis of the relative influence on the change signal of local soil-moisture feedbacks and remote effects of Sea Surface Temperature (SST) over the Niño 3.4 region indicates that the former is prevalent over the Amazon basin while the latter dominates over the La Plata Basin. Also, the soil moisture feedback has a larger role in RegCM4 than in the GCMs.