U.S. winter wheat yield loss attributed to compound hot-dry-windy events

Wheat yields have stagnated or collapsed in some regions, raising concerns about sustainable production amidst climate change. Previous research focused on single extreme events like drought, heat waves, or cold spells, or on multiple events without considering their compound effects. The Intergovernmental Panel on Climate Change (IPCC) defines compound events as combinations of extreme or non-extreme events leading to amplified impacts. Compound hot-dry-windy events (HDW), combining high temperature, low humidity, and high wind, pose a particular risk, especially in the U.S. Great Plains, historically vulnerable to such conditions (as seen in the 1930s Dust Bowl). While HDW events have increased since the 1950s, their impact on U.S. hard winter wheat production hasn't been comprehensively assessed. This study uses phenological stages (planting to jointing, jointing to heading, heading to maturity) to analyze climate-crop relationships, avoiding the limitations of fixed calendar-based analyses. This approach is crucial because phenological timing, crop sensitivity, management practices, and regional climates vary spatially.

Existing literature highlights the impacts of single extreme climatic events (drought, heat waves, cold events) on agricultural yields. Some studies examined the effects of multiple extreme events, but not as compound events. The IPCC's definition of compound events has evolved, recognizing the combination of multiple drivers contributing to risk. Research on compound events has been limited, focusing primarily on changes in crop water supply and temperature. Studies on the effects of atmospheric dryness are limited, despite its importance in rapid water loss from plants. The focus on inadequate soil water supply has overshadowed the importance of rapid water demand due to atmospheric dryness which can impact plants even with adequate soil moisture.

The study focused on hourly HDW events during the heading-to-maturity (HD-MT) stage of winter wheat growth (1982–2020) in the U.S. winter wheat belt (South Dakota, Nebraska, Colorado, Kansas, Oklahoma, and Texas). Hourly HDW events were defined as the co-occurrence of temperature ≥32°C, relative humidity ≤30%, and wind speed ≥7 m s⁻¹ at 10 m above ground. County-level hard winter wheat yield data were obtained from the USDA-NASS. Hourly climate data (temperature, relative humidity, wind speed) were from the ERA5-Land dataset. Daily precipitation data were from the PRISM dataset. Phenological data (planting, jointing, heading, maturity dates) were digitized from hard copies of Winter Wheat Performance Tests Reports. A linear mixed-effects model was used to assess the impacts of climate indices (HDW hours, freezing days, extreme degree days, precipitation) on wheat yields. The model included county-specific and year-fixed effects. Standardized climate indices (z-scores) were used to compare yield sensitivities to different climate indices. Yield sensitivity to HDW was analyzed during early, middle, and last sub-stages of the HD-MT stage. The impacts of individual, bivariate, and trivariate compound events (hot, dry, windy, and their combinations) were compared. Yield shock years (years where yield change percentage fell below the 25th percentile) were analyzed to determine spatial patterns of yield loss and associate them with climatic indices. Robustness checks included using different HDW thresholds and alternative temperature modeling approaches (quadratic temperature, temperature bins).

The study found that HDW events significantly increased in the southwestern Kansas and panhandle areas of Oklahoma and Texas, mirroring the Dust Bowl areas of the 1930s. Upward HDW trends during HD-MT were statistically significant, with rates up to 8 h per decade in the most affected areas. High-temperature events were a major control variable for HDW occurrence. HDW during HD-MT (HDW_HD-MT) was the most significant driver of yield variability, accounting for a 3.5% yield loss per standard unit or 4% yield loss per 10 h. EDD_JT-HD (extreme degree days during the jointing-to-heading stage) also significantly impacted yields. Compound hot-dry events also significantly reduced yields. HDW impacts were greatest during the middle sub-stage of the HD-MT stage. HDW_HD-MT are associated with yield losses up to 0.09 t ha⁻¹ per decade in severely affected areas. The Pacific Decadal Oscillation (PDO) appears to be an atmospheric bridge, influencing decadal variations in HDW. Yield shock years were characterized by anomalous EDD_JT-HD in the central winter wheat belt and EDD_HD-MT in western areas, as well as excessive precipitation in eastern areas and drought in western and northern areas. The most HDW-affected areas during yield shocks were spatially consistent with 1930s Dust Bowl areas.

The findings demonstrate that compound HDW events, often overlooked, are major drivers of U.S. winter wheat yield loss. The significant impact of HDW surpasses the effects of individual extreme climate events. The spatial patterns of HDW impacts align with the historical Dust Bowl region, highlighting the vulnerability of this area to these compound events. The link between HDW and the PDO suggests potential for decadal variations in HDW frequency and intensity. These results underscore the need for adaptation strategies, including improved crop management (water and chemical management) and the development of wheat varieties with enhanced tolerance to heat, drought, and wind stress. The interaction of various stresses under HDW conditions complicates crop vulnerability, highlighting the importance of considering multiple factors.

This study reveals the significant and previously underappreciated impact of compound HDW events on U.S. winter wheat yields. The spatial overlap with the Dust Bowl area highlights the persistent vulnerability of this region to such climatic extremes. Future research should focus on developing more robust adaptation strategies to mitigate the effects of HDW and further explore the relationship between compound extreme events and crop yield, considering the interactions of frequency, intensity, and various stressors at multiple scales.

The study relies on reanalysis data, which may contain uncertainties, potentially affecting the precision of HDW event estimations. The model assumes a linear relationship between climate indices and yields, and non-linear effects may exist. The study mainly focuses on the impact of HDW during HD-MT, neglecting potential impacts during other growth stages. The spatial resolution of climate data might not perfectly capture the variability within each county.