Introduction

Gro’s Yield Forecast Models use a suite of machine-learning models to estimate in-season yields at the district, province, and/or national levels on a daily basis. Gro’s US Hard Red Wheat Yield Forecast Model updates daily to give national yield forecasts for hard red winter wheat, as well as winter wheat yield forecasts at the county level. The model offers in-season, near-real-time yield analysis, operating across three states—Kansas, Oklahoma, and Texas—which together account for about 60% of US total production of hard red winter wheat.

Customers Use the Model to:

  • Monitor crop production, the most significant part of the balance sheet, by tracking its most variable component, which is in-season yield
  • Gauge crop availability and crop prices
  • Forecast hard red winter wheat yield in areas that account for more than half of US production
  • Understand how weather impacts yields in microclimates
  • Inform other models focused on damage caused by pests and diseases

Why it Matters

The US is the third-biggest national exporter of wheat, after Russia and Canada. Despite a decline in US wheat acreage in recent years, the crop continues to occupy a significant portion of US farmland. Wheat grown across the country shows considerable variations in planting calendar, protein content, and market uses; specialized models that track different varieties are valuable for monitoring the downstream impact of changes in wheat yield. Hard red winter wheat (HRW), which is high in protein and suitable for baking, accounts for over a third of total US wheat production — nearly 18 million tonnes in 2020. About half of the crop is exported, which means that hard red winter wheat yields also have ramifications on global trade dynamics.

Methodology

Because the USDA doesn’t provide county-level yield specifically for HRW, we trained our model on overall winter wheat yield in regions that are almost exclusively HRW. We tested a number of geospatial inputs for our model, of which normalized difference vegetation index (NDVI) and evapotranspiration (ET) performed the best during the backtest. Besides these geospatial inputs, weekly winter wheat crop conditions from the USDA (at the state level) also helped improve the model’s performance.

Related Resources