Old News
Roberts and Wolfram Schlenker's 2009 PNAS publication established that U.S. corn and soybean yields demonstrate significant vulnerability to temperature extremes. Their analysis employed a novel methodology by accounting for the complete temperature distribution rather than simple averages, yielding consistent results across different geographic regions and identification approaches — cross-sectional, time-series, and panel regression methods all converged on similar findings.
The study's strength derived from weather's quasi-random annual variation and demonstrated out-of-sample predictive accuracy, suggesting genuine causal relationships with limited farmer adaptation capacity.
Any Impacts Yet?
Despite nearly two decades of global warming since publication (approximately 0.4°C increase), U.S. yield growth has remained linear without apparent deceleration. The 2012 growing season provided the notable exception — an exceptionally hot year coinciding with La Niña conditions.
However, global agricultural productivity growth has slowed measurably, with climate change appearing as a contributing factor, though American agricultural output has largely escaped comparable slowdowns.
Why No Apparent Warming Impacts?
A critical geographic anomaly explains this paradox: the U.S. Corn Belt has experienced a "warming hole" during summer months while global temperatures surge elsewhere. This region — comprising nearly 40% of worldwide corn and soybean production — has actually cooled during critical growing seasons over recent decades.
The cooling reflects multiple factors:
- The 1930s Dust Bowl represented exceptionally elevated temperatures, creating a baseline effect
- Summer temperatures specifically have declined in recent decades across the Corn Belt
- This regional cooling has functioned as a beneficial tailwind for yield expansion
Theories on the Warming Hole
Several hypotheses attempt to explain this phenomenon:
Evapotranspiration Hypothesis
Dense corn and soybean cultivation might generate cooling through massive moisture release during peak summer growth, though modeling evidence remains inconclusive.
Land Use Effects
The 1930s Dust Bowl likely involved significant land degradation amplifying heat and drought. Subsequent reforestation and improved land management may have partially reversed these conditions.
Climate Oscillations
Extreme La Niña patterns — blamed as primary factors in the Dust Bowl and recurring in 2012 — may drive cyclical temperature variations independent of long-term warming trends.
Roberts remains skeptical that crops themselves significantly cool the climate, though undergraduate researcher Eleanor Yuan's honors thesis suggested the relationship involves "much larger and more complex interactions between crops and weather."
Recent Years' Weather Patterns
Global temperatures have risen sharply in 2023–2024, driven by El Niño, climate change, and various cyclical factors. Remarkably, the Corn Belt has remained cool during these globally warm years, with 2023 and 2024 showing exceptionally low accumulation of extreme heat relative to historical averages since 1971.
AE4cast — Roberts's startup with former graduate student Sisi Zhang — forecasts corn yields in real-time using sophisticated weather modeling. Their 2023 projection proved accurate, though certain states exhibited notable errors. The 2024 crop is projected to set records.
The underlying model employs greater complexity than published research, accounting for nonlinear temperature relationships and inherent weather forecast uncertainty. The team plans international expansion and continued refinements.
Future Prospects
Multiple scenarios appear plausible:
Decades resembling 1930s conditions or worse could devastate American agricultural production and trigger global food crises, despite current record crops and declining prices obscuring these risks.
Climate impacts might prove manageable, though other productivity-limiting factors could emerge.
Continued agricultural luck combined with precision agriculture innovations — AI-driven equipment, drone monitoring, enhanced soil management — could accelerate productivity growth while reducing land requirements and climate sensitivity.
Roberts emphasizes that "uncertainty is not our friend," advocating rapid decarbonization as warranted given significant downside risks.
Update: Afforestation Effects
A new working paper by Grosset-Touba, Papp, and Taylor examines the Great Shelterbelt program's post-Dust Bowl impacts, demonstrating that large-scale forestation significantly increased rainfall and reduced temperatures, substantially affecting corn production. This research suggests afforestation may represent "at least half the story" explaining the Corn Belt's cooling.
Roberts remains concerned about whether favorable weather patterns will persist, particularly during future extreme La Niña events.