The 1930s Dust Bowl witnessed record-breaking summer heatwaves across the central US, exceeding even those simulated by historically forced coupled climate models. These extreme temperatures, along with the severe drought, caused significant socio-economic and ecological damage. The extent to which these heatwaves were driven by sea surface temperatures (SSTs) and exacerbated by human-induced land cover changes remains a key research question. This study aims to disentangle the contributions of ocean and land forcing to the exceptional heatwaves of the Dust Bowl. Understanding this is crucial, given the projected increase in future global heatwave activity and the need to assess the potential for amplification of natural extreme events through human impacts on the environment. The research will utilize atmospheric-only models and simulations to investigate the relationships between SST anomalies (particularly in the North Atlantic and Pacific), land cover changes (specifically devegetation and its impact on evapotranspiration), and the frequency and intensity of the Dust Bowl heatwaves. The findings will provide insights into the complex interactions between natural climate variability and human-induced environmental changes in shaping extreme weather events.

Previous research has highlighted the role of both SSTs and land-cover changes in the Dust Bowl drought. Studies have indicated a connection between cooler-than-average North Pacific SSTs and warmer North Atlantic SSTs during the 1930s and the drought's development. However, atmospheric-only general circulation models (AGCMs) often underestimate the drought's severity when forced with observed SST anomalies. Other studies have shown that incorporating realistic historical land-cover changes and dust aerosol forcing improves the simulation of precipitation and temperature during this period. The observed extreme heat during the Dust Bowl has been linked to persistent upper-level atmospheric ridging and land-atmosphere interactions, and previous research has also emphasized the importance of springtime preconditioning through dry conditions. While the influence of Pacific SST anomalies on central US droughts has been noted, there is less consensus on the relative contribution of Atlantic SST anomalies compared to Pacific SST anomalies. Existing research lays the foundation for this study, which will further investigate the relative roles of SSTs and land-cover changes using advanced modeling techniques.

This study employed a multi-faceted approach combining observational data, coupled climate model simulations, and atmospheric-only general circulation model (AGCM) experiments. Heatwaves were identified using a consistent definition based on consecutive hot days and nights exceeding their respective 90th percentile thresholds. Observational data included daily temperature data from the Global Historical Climatology Network-Daily (GHCN-D) archive, gridded data from HadEX2, and the Berkeley Earth Surface Temperature (BEST) dataset. Sea surface temperature (SST) data were derived from HadISST2.1. Atmospheric circulation data (mean sea-level pressure (MSLP) and 500 hPa geopotential heights (Z500)) came from the Twentieth Century Reanalysis (20CR) version 2. The HadGEM3-GA6 AGCM was used to conduct a series of experiments, including historical simulations forced with observed SSTs (HIST) and idealized simulations with either Atlantic (ATLHIST) or Pacific (PACHIST) SST anomalies held constant while the other ocean basins were represented by climatological values. Sensitivity experiments using HadGEM3 explored the impact of varying percentages (30%, 50%, and 80%) of bare soil in the central US in 1930 to simulate the land-cover changes of the Dust Bowl. CMIP5 pre-industrial control (piControl) simulations from 22 models were analyzed to assess the frequency of Dust Bowl-like heatwaves in unforced simulations. Statistical tests (Mann-Whitney U test, bootstrapping) were used to assess the significance of differences between experimental results. The study carefully considered model biases, particularly in HadGEM3's tendency to overestimate land-atmosphere coupling strength, and primarily focused on heatwave frequency (HWF) to mitigate this bias, as this metric is less sensitive to absolute temperature biases.

The study's key findings reveal a complex interplay of ocean and land processes in shaping the Dust Bowl heatwaves. Analysis of observational data confirmed the exceptional nature of the 1930s heatwave activity, exceeding the variability observed in CMIP5 historical simulations. Using the HadGEM3 AGCM, the Atlantic SST anomalies were found to have a stronger influence on summer heatwave activity than Pacific SST anomalies. This was primarily due to the creation of drier spring conditions caused by a weakening of moisture transport from the Gulf of Mexico during the spring months prior to summer. This drier spring then led to a preconditioning of the land surface, promoting more extreme summer heatwaves. Bare-soil sensitivity experiments, designed to represent the extensive devegetation during the Dust Bowl, demonstrated a substantial increase in heatwave frequency and intensity with increasing bare soil coverage. This increase was attributed to enhanced sensible heat fluxes and a rapid drying of exposed soil, leading to an earlier drought onset. Analysis of CMIP5 pre-industrial control simulations showed that while Dust Bowl-like heatwave events can occur naturally, the combined intensity and spatial extent observed during the 1930s was extremely rare. The results highlighted the interplay between natural climate variability (SST patterns linked to the AMO and PDO) and human-induced land-cover changes (devegetation). The results showed that although HadGEM3 had significant biases, the Atlantic SSTs were still important in inducing drier springs, and that the increase in bare soil from devegetation was important in creating more frequent and intense heatwaves.

The findings of this study demonstrate the crucial roles of both ocean and land processes in the extreme heatwaves of the 1930s Dust Bowl. The influence of Atlantic SST anomalies in creating drier springs and preconditioning the land surface for extreme summer heat highlights the importance of considering remote climate drivers. The strong sensitivity of heatwave activity to land cover changes, as illustrated by the bare-soil experiments, underscores the significant impact of human activities on amplifying the intensity of naturally occurring drought conditions. This underscores that changes to land cover can have significant impacts on extreme weather events. The limitations in CMIP5 models in capturing the extreme nature of the Dust Bowl, likely due to underrepresentation of land-cover changes, suggest the need for improved model parameterizations of land surface processes. The interaction of Atlantic SST anomalies and vegetation feedbacks illustrates the potential for future amplification of extreme heatwaves in a warmer world with altered vegetation patterns caused by climate change. The study provides insights into how different factors interact to create extreme climate events and emphasizes the need for further research into how climate change might exacerbate these interactions in the future.

This research reveals a complex interaction between naturally occurring climate variability in the North Atlantic and human-induced devegetation in driving the extreme heatwaves of the 1930s Dust Bowl. Atlantic SST anomalies were a significant driver, inducing drier springs which led to more extreme heatwaves during summer. Human-induced land cover changes, specifically widespread devegetation, dramatically amplified these heatwaves. The combined effects of these factors, along with shortcomings of current climate models to account for them, emphasizes the potential for greater extreme heatwaves in the future due to similar patterns of climate variability. Further research into the interactions between climate variability and land cover feedback, with improved model simulations, is crucial for understanding and predicting future extreme weather events.

The study acknowledges limitations arising from model biases, particularly the overestimation of land-atmosphere coupling strength in HadGEM3. While the focus on HWF helped mitigate the impacts of the temperature bias, uncertainties still exist regarding the exact amount of devegetation during the Dust Bowl and potential effects of dust aerosols on precipitation. The bare-soil experiments, while demonstrating the importance of land cover changes, were simplified representations of the complex processes involved. This simplified model approach may underrepresent the actual level of complexity in the feedback systems involved.