Coupled Atmospheric and Land-use Shifts Amplify Drought-driven Crop Losses Across the U.S.
Lili Yao1*, Hongxiang Yan1, Ning Sun1*, Eva Sinha1, Kanishka B. Narayan1, Travis B. Thurber1, and Jennie Rice1
1 Pacific Northwest National Laboratory, Richland, WA, USA
* Correspondence: Lili Yao, lili.yao@pnnl.gov; Ning Sun, ning.sun@pnnl.gov
Agricultural drought (AD) poses a major threat to food security, yet its future risk remains uncertain under co-evolving atmospheric conditions and land-use trajectories. Using an integrated, multi-sector modeling framework, we projected AD risks for major crops across the contiguous United States (CONUS) through 2055 under a range of plausible futures that capture thermodynamic changes and land-use and land-cover change (LULCC) trajectories. We benchmarked simulated drought hazard, crop production losses, and financial impacts against multiple independent datasets, demonstrating robust performance across metrics and regions. With this validated framework, model projections reveal that drought-driven crop production losses increase sharply by nearly 60% for corn, 250% for wheat, and 135% for soybean relative to historical levels. Wheat exhibits the largest projected loss increases, a result that remains robust across scenarios. The primary drivers of these increases, which include atmospheric shifts and LULCC, vary by region and crop type. LULCC acts as an important risk amplifier in regions experiencing cropland expansion into drought-prone regions, such as the Great Plains and northwestern U.S. These findings highlight that interactions between atmospheric conditions and land-use trajectories shape future agricultural drought risk and should be jointly considered to support effective adaptation and food-system planning.
Coupled Atmospheric and Land-use Shifts Amplify Drought-driven Crop Losses Across the U.S. Submitted to Earth's Future – January 2026.
| Dataset | URL | DOI |
|---|---|---|
| TGW-WRF | https://tgw-data.msdlive.org/ | https://doi.org/10.1038/s41597-023-02485-5, https://doi.org/10.57931/1885756 |
| GCAM-SELECT-Demeter | https://data.msdlive.org/records/vy529-6eg15 | https://doi.org/10.57931/2502083 |
| Dataset | URL | DOI |
|---|---|---|
| CLM5 soil moisture and crop yield simulations | https://data.msdlive.org/records/gmcgt-pvx90 | https://doi.org/10.57931/3012125 |
| Model | Version | URL | DOI |
|---|---|---|---|
| CLM5 | ctsm5.1.dev118 | https://github.com/IMMM-SFA/im3-clm | https://zenodo.org/records/6653705 |
| IM3 Components | 0cf45e8 | https://github.com/IMMM-SFA/im3components/tree/main/im3components/wrf\_to\_clm |
Clone the CLM5 repository to set up the CLM5 model. You will need to download the TGW forcing data and convert them into CLM input format using these scripts. You will also need to replace the default CLM surface and landuse timeseries files using data from the GCAM-SELECT-Demeter. In addition, hydrological parameter values in the default parameter file and the user name list file should be updated based the behavioral parameter values. The output data repository already contains the soil moisutre and crop yield output from the CLM5 model so you can skip rerunning the CLM5 model if you want to save time.
Use the scripts found in the figures directory to reproduce the figures used in this publication.
| Figure Numbers | Script Name | Description | Figure |
|---|---|---|---|
| 1 | Figure_1.py | Validation of CLM’s performance |  |
| 2 | Figure_2.m | Rainfed crop planting areas changes |  |
| 3 | Figure_3.py | Drought exposure and intensity |  |
| 4 | Figure_4.py | Mean projected changes in production loss |  |
| 5 | Figure_5.py | Scenario and ESM variant uncertainty in projection |  |
| 6 | Figure_6.py | Relative differences in financial loss between scenarios |  |
| S1 | Figure_S1.py | Ratio of crop prices between scenarios |  |
| S2 | Figure_S2.m | Projected changes for nine land use and land cover types |  |
| S3 | Figure_S3.m | Projected changes in annual and seasonal air temperature |  |
| S4 | Figure_S4.m | Projected changes in annual and seasonal precipitation |  |
| S5 | Figure_S5.m | Projected changes in annual and seasonal potential evapotranspiration |  |
| S6 | Figure_S6.m | Projected changes in annual and seasonal aridity index |  |
| S7 | Figure_S7.py | Agricultural drought duration under historical and future scenarios |  |
| S8 | Figure_S8.py | Relative differences in financial loss for cooler and hotter variants |  |