Theses Doctoral

The mystery of observed and simulated precipitation trends in Southeastern South America since the early 20th century

Varuolo-Clarke, Arianna Marie

Southeastern South America (SESA), a region encompassing Paraguay, Southern Brazil, Uruguay, and northern Argentina, experienced a 23% increase in austral summer precipitation from 1902-2022, one of the largest precipitation trends observed globally. There is little consensus on the drivers of the precipitation trend, but Atlantic multidecadal variability, stratospheric ozone depletion, and greenhouse gas emissions stand out as key contributing factors.

The work presented in this dissertation addresses two main questions. First, what are the historical drivers of the SESA precipitation increase? To address this, I investigate simulations from the Coupled Model Intercomparison Project (CMIP) Phases 3, 5, and 6 and find that not only do fully-coupled climate models simulate positive SESA precipitation trends that are much weaker over the historical interval, but some models persistently simulate negative precipitation trends. The same is true of two atmospheric models forced with observed historical sea surface temperatures. While future 21st-century projections yield positive ensemble mean precipitation trends that grow with increasing greenhouse-gas emissions, the mean forced response never exceeds the observed historical trend. Finally, some pre-industrial control runs occasionally simulate centennial-scale trends that fall within the observational range, but most do not.

The second question I address is why climate models struggle to simulate the observed SESA precipitation trend. In an attempt to understand the model bias, I investigate one driver of SESA precipitation variability: the South American low-level jet. By developing a jet index from low-level moisture fluxes into SESA, I find that increased moisture flux through the jet accounts for 20-45% of the observed SESA precipitation trend from 1951-2020 in two reanalysis datasets. While results vary among reanalyses, both point to increased humidity as a fundamental driver of increased moisture flux and precipitation. Increased humidity within the jet is consistent with warming sea surface temperatures driven by anthropogenic forcing, although additional natural climate variations also may have played a role. The jet’s velocity also increased, further enhancing precipitation, but without a clear connection to anthropogenic forcing. These findings indicate that the SESA precipitation trend is partly attributable to jet intensification arising from both natural variability and anthropogenic forcing.

In my final research chapter, I explore whether CMIP6 models simulate a realistic relationship between SESA precipitation and the jet, as well as whether inaccuracies in the characterization of the jet could explain muted trends in simulated SESA precipitation. I find that the interannual variability in the simulated jet-precipitation relationship aligns well with results from observations from 1951-2014. Interannual precipitation variability across the models is primarily dominated by the jet’s velocity. The models simulate a forced increase in humidity within the jet, consistent with observations and theory, that contributes a positive trend to SESA precipitation. Given that the models generally simulate realistic jet-precipitation relationships, I conclude that model misrepresentation of the jet is not a likely explanation for the discrepancy between simulated and observed SESA precipitation trends.

Despite remaining uncertainties, my work sheds new light on our understanding of SESA precipitation variability and trends. Future work is needed to better understand the large-scale drivers of SESA precipitation outside of the jet and why climate models largely underestimate or fail to reproduce the observed precipitation trend. While Atlantic multidecadal variability is often cited as an important contributor to the SESA precipitation trend, I find austral summer forcing from the Atlantic to be ambiguous with regard to SESA precipitation and requires further analysis. Additionally, I highlight the Pacific South American mode as another contributing factor that warrants further exploration.

Geographic Areas


  • thumnail for VaruoloClarke_columbia_0054D_18005.pdf VaruoloClarke_columbia_0054D_18005.pdf application/pdf 5.1 MB Download File

More About This Work

Academic Units
Earth and Environmental Sciences
Thesis Advisors
Smerdon, Jason E.
Williams, A. Park
Ph.D., Columbia University
Published Here
July 26, 2023