Eighty-eight million hectares of extra cropland came into production between 1992 and 2020, across 110 countries, to compensate for yields lost to heat stress. The figure comes from Extreme Heat and Agriculture, the FAO–WMO joint report published in April 2026. The land was cleared and planted across 28 years not to expand global output, but to keep it where it was.
The 88 million hectares is the visible part. On top of that, the report attributes 21.8 billion tonnes of CO₂ equivalent in emissions to the expansion itself, or 18.9% of total land-use emissions for the countries involved. Less visible, and harder to recover from, is a 21% decline in agricultural total factor productivity since 1961 attributable to heat stress. Roughly seven years of global productivity gains, gone.
Per-degree yield losses
The per-crop numbers are worth stating in one place. Across the global meta-analyses the report draws on, every 1°C of growing-season warming costs maize about 7.5% of yield, soybean 6.8%, wheat 6.0%, and rice 1.2%. These are observed losses from tens of thousands of crop-year records, not model output. Forward projections add another 4 to 10% for maize, soybean, and wheat per additional degree, with the spread depending on which model is run.
The damage threshold for most major crops sits around 30°C. Potatoes and barley take damage earlier. How much that matters on a given farm depends on how close the growing-season mean already sits to that threshold, and how often it crosses. Tropical and Mediterranean systems are the ones most exposed. Northern systems still have headroom, for now.
Heat alone, heat with drought
One of the more useful numbers in the report concerns the relative weight of heat versus drought as drivers of yield anomalies. Heat alone, on average, accounts for about a 9% yield loss in affected regions. When heat coincides with dry conditions, compound losses jump to roughly 24.89%. For most southern European producers, that compound hot-dry signature is now the default summer event, not the exception.
The mechanism is unsurprising. Heat shortens the grain-filling window by speeding up maturation, raises atmospheric vapour pressure deficit, and pulls moisture from soil and tissue. When soil moisture is also low, the plant closes its stomata to conserve water, and photosynthesis stops with them. The two stresses combine multiplicatively rather than additively. Cotton, cereal, and oilseed flowering windows are particularly exposed, because pollen viability drops sharply once tissue temperatures climb above the species threshold for more than a few hours.
Why the global signal hasn't moved
A 2024 paper by Marshall Burke and colleagues examined fifty years of grain yield data for evidence that crops have become less sensitive to extreme heat. They didn't find it. Across 21 outcomes the team examined, six showed declining sensitivity, five showed rising sensitivity, and the rest showed nothing clear. For crops, several major producers have grown more sensitive to heat over time, not less.
This isn't a contradiction of the adaptation growers already do. Switching to climate-resilient varieties, shifting planting dates, leaving residue on the surface, irrigating more efficiently, all of it reduces damage on the farm where it is applied. Burke's point is that the global aggregate hasn't moved, which means warming is outpacing those gains roughly as fast as they accrue. The FAO–WMO 21% productivity gap is the same finding from the other direction.
What that looks like in southern Europe
Last summer was a clear preview. Temperatures hit 46°C in southern Spain, and severe stress was reported across cotton, olives, tomatoes, and wine grapes from Andalusia to Thrace. Cotton flowering windows were the worst hit, and yield losses above 50% in the affected Mediterranean basins are documented in our analysis of the July 2025 heatwave.
The FAO–WMO modelling projects that compound drought-and-heat events across Europe, the Mediterranean basin, the Middle East, and Western and Central Asia could rise 200 to 300% by the end of the century relative to a 1986–2005 baseline. Northern European cereals and oilseeds will hold their yields longer than southern systems will. The Mediterranean rainfed cereal belt is already losing the assumption that a standard variety can carry an average season.
Where the gains are coming from
The report's adaptation chapter sorts interventions into three buckets, each with concrete examples.
Genetics is the most flexible. Where the local climate has shifted enough to make existing varieties marginal, there is usually a shorter- or longer-cycle alternative that fits the new window. Where it has shifted further, the answer is a different crop or breed. International programmes like CIMMYT's heat-tolerant wheat lines and Drought Tolerant Maize for Africa have shown 20-50% better performance under stress than conventional varieties, depending on trial and location.
Soil environment is the second, and the cheapest. Conservation tillage and surface mulch insulate the root zone. The report cites trials where ploughed soil under maize hit 41.4°C while neighbouring zero-till plots stayed about 10°C cooler at the same hour, under the same sky. The Brazilian sugarcane sector is the working example at scale. No-till on sugarcane straw left soybean fields measurably less damaged through the 2023 and 2024 seasons because the residue blanket cut both surface temperature and evaporative loss. The same physics carries to vegetable beds and orchards.
Management is the third. Adjusted planting dates, agrometeorological advisories, and anticipatory action protocols are often the highest-return interventions and the most underused. A heatwave is one of the few extreme events that genuinely can be forecast days in advance. The constraint, particularly in southern and eastern European smallholder systems, is access to the forecast and to a workable response plan, not the science behind it.
A note on sourcing
For grain and processed-food buyers, the spatial pattern of heat-driven yield losses is changing how supply contracts are written. The southern Mediterranean, the Black Sea basin, and the North Sea farms don't fail in the same weather. Procurement portfolios concentrated in two or three suppliers in the same climate zone are the most exposed when a regional shock arrives, as the 2022 Black Sea wheat market and the 2025 southern European cotton, olive oil, and processing tomato seasons both made plain.
The ceiling
Adaptation reduces damage. It doesn't set the ceiling. Under high-emission pathways, total agricultural productivity continues to decline in the FAO–WMO modelling even with full uptake of every known on-farm intervention. The 88 million hectares is the early phase of the adjustment, and the carbon cost of those hectares is already in the atmosphere working against the next round.
What holds up across most climate trajectories is short and unglamorous. Residue cover, reduced tillage, varieties chosen for the climate the farm actually has rather than the one in the planting handbook, and a forecast that arrives in time to act on. None of it new. Producers and buyers already running all of it are the ones least exposed when the next compound event lands.
References
- FAO and WMO. (2026). Extreme heat and agriculture – FAO–WMO joint report. Rome and Geneva.
- Burke, M., Zahid, M., Martins, M., Callahan, C., Lee, R., Avirmed, T., Heft-Neal, S., Kiang, M. V., Hsiang, S., & Lobell, D. (2024). Are We Adapting to Climate Change? NBER Working Paper No. w32985.
