In this study, we investigate whether the climate over South America has changed as a result of human activity since the beginning of the industrial revolution and to what extent the recently observed changes in daily minimum (Tmin) and daily maximum (Tmax) temperature are consistent with climate change AR5 projections. To this end, we assess whether the observed changes are likely to have been due to natural (internal) variability alone, and if not, whether they are consistent with what models simulate as response to anthropogenic forcing. We use several data sources: 1) projected response to greenhouse gas (GHG) forcing derived from 30 global climate simulations from CMIP5 archive, 2) projected response to GHG forcing from 1 regional climate model based on RCP4.5 scenario from CORDEX, 3) historical well-mixed greenhouse gas forcing only simulations 4) historical aerosol only forcing simulations with and without the “second indirect effect” of aerosol (“cloud lifetime effect”), 5) output from 20,000 years of pre-industrial control simulations derived of CMIP5 archive.
Results indicate that, over the past decades, observed warming trends in Tmin and Tmax in dry seasons (JJA, SON) over northern South America cannot be explained by natural (internal) variability alone and that externally forced changes have a significant (at 2.5% level) influence in the observed warming trends. Significance is estimated using the output from long pre-industrial control simulations. We further assess the influence on daily temperature over SA of several key climate modes, such as the El Niño Southern Oscillation (ENSO), the Atlantic Multi-Decadal Oscillation (AMO), the Pacific Decadal Oscillation (PDO), and the Southern Annual Mode (SAM). We assess the robustness of detection results against subtracting from the observations that part of the temperature variability that can be attributed to those four natural modes of climate. The detection of externally forced changes in Tmin and Tmax in dry seasons are robust against the removal of the fingerprint of the four natural modes. However, in wet season (DJF, MAM) the natural modes of variability explain a substantial portion of Tmax and Tmin variability.
The spatial correlation and regression patterns derived from observations clearly illustrate the emergence of GHG signal (at 2% level) in the 21st century. However, none of the global and regional climate change projections we used captures the observed warming of up to 0.6 K/Decade in Tmax in 1983-2102 over Amazon Basin in austral spring (SON). Thus, besides the regional manifestation of GHG forcing, other external drivers have an imprint. Using aerosols-only forcing simulations, our results provide evidence that anthropogenic aerosols also have a detectable influence in observed warming in SON and that the indirect effect of aerosols on cloud’s lifetime is compatible (at 2.5% level) with the observed daily maximum temperature increase over northern South America. We further show that there is a positive trend in the incoming solar radiation over the northern SA in SON that cannot be explained by natural (internal) variability (at 2.5% level) and that is distinct from the expected response to GHG forcing. We conclude here that the effect of aerosols-radiation-cloud interactions due to biomass burning aerosols from the large number of fires over Amazon Basin is a missing driver in climate change projections.