Heat at Flowering Will Overtake Drought as the Primary Threat to Wheat Yields by 2050

Wheat feeds roughly 2.5 billion people. When its yields fall, food prices rise, import-dependent nations face shortages, and farming communities absorb losses that take years to recover from. Climate projections have long pointed to drought as the dominant future threat to wheat production. A study from Rothamsted Research challenges that hierarchy - and replaces drought at the top of the threat ranking with something more specific and more difficult to manage.

The critical variable is not total rainfall or annual temperature averages. It is what happens to wheat during flowering - a window that lasts only a few days but determines how many grains a plant will produce, and therefore how much it will yield at harvest.

Flowering: A Few Days That Determine the Season

Wheat flowering is the stage when pollen must be released and fertilization must occur. The process is exquisitely sensitive to temperature. Heat above roughly 30-35 degrees Celsius during anthesis - the period when pollen is shed - damages pollen viability, reduces fertilization rates, and cuts grain set. A brief heat event of three to five days at the wrong moment can reduce yield by 20 percent or more, with relatively little the plant can do to compensate afterward.

Dr. Mikhail Semenov, Mathematical Modeller and Emeritus Fellow at Rothamsted Research, described the sensitivity directly: "Flowering is one of the most sensitive stages in wheat development. It's when the plant sets grain, which ultimately determines yield. Even a few days of very high temperatures or severe water stress at this stage can reduce grain numbers and significantly cut final harvests."

What the Sirius Model Found

The research team used the Sirius wheat growth model - a well-validated mechanistic model that simulates wheat development under specific weather conditions - combined with advanced climate projections to estimate how the frequency and intensity of extreme heat and drought events during flowering will change across major wheat-growing regions.

The results show a clear divergence. Drought during flowering currently causes more yield loss globally than heat. But drought's projected impact is expected to remain roughly stable or decline slightly in some regions. Heat stress during flowering is on a steep upward trajectory. By 2050, global yield losses attributable to extreme heat at flowering are projected to increase by approximately one-third above current levels. By 2090, those losses could exceed current levels by more than three-quarters. The reversal from drought-dominant to heat-dominant risk is projected to occur in the 2040s to 2060s depending on the climate scenario and region.

What Plant Breeders and Farmers Face

Drought tolerance has been a breeding priority in wheat for decades. The genetics of drought response are relatively well characterized. Heat tolerance during flowering is a different and less thoroughly understood challenge. Varieties that can maintain pollen viability and fertilization success at high temperatures exist but have not been bred at scale for the major commercial wheat varieties that dominate global production. Developing them requires identifying the genetic loci that confer heat tolerance during anthesis, backcrossing them into elite germplasm, and conducting multi-environment field trials - a process that typically takes 10 to 15 years from gene identification to commercial release.

Professor Malcolm Hawkesford, leader of the "Delivering Sustainable Wheat" Institute Strategic Programme at Rothamsted, emphasized the urgency: "This kind of modelling study provides critical information on, and pointers to, the traits we should be breeding for now, ready for predicted future climate conditions."

For farmers with shorter time horizons, the practical implications include adjusting sowing dates to shift flowering away from the peak heat period and selecting varieties with earlier or later flowering times that reduce exposure to extreme heat events. Neither approach fully solves the problem; they redistribute risk rather than eliminating it.

Limits of the Analysis

The study uses modeled climate projections, which carry inherent uncertainty at regional scales. The Sirius model captures many aspects of wheat physiology accurately but cannot simulate every interaction between plant stress responses, soil conditions, and extreme weather. Real-world impacts will depend on how effectively farmers and breeding programs adapt - and how rapidly warming actually proceeds relative to the scenarios modeled. Regional variation is also large: the adaptation implications differ substantially between a temperate zone in northern Europe and a semi-arid growing region in northern India.