Yield Forecasting

Blog/

What is trend yield, and how does it differ from a weather-adjusted yield estimate?

Low-angle view across a green corn field at sunset, with a tree line on the horizon.

Trend yield is the number you get by running a straight regression line through decades of yield data for a crop and region, then reading off the value for the current crop year. It's a baseline built on the long run of improving genetics and farming practice, extended forward by one more year, not a forecast of this season's weather.

USDA uses trend yield every February to set the first production number for corn and soybeans, months before anyone has planted a seed. At that point there's no crop to measure, so the only honest move is to extrapolate from history and assume a normal season.

A weather-adjusted yield estimate starts from that same trend line, then layers on what's happening in the field: rainfall against the normal, growing degree days, drought monitor categories, sometimes a satellite-derived vegetation index. The adjustment can push the number above trend in a good year or knock it down hard in a dry one.

Where the two numbers split

In a textbook season, trend yield and a weather-adjusted estimate land close together, because the weather itself is close to normal. The gap opens when conditions move away from average. A hot, dry stretch through pollination in the Corn Belt will drag a weather-adjusted estimate well below trend, while the trend line itself doesn't move at all. Trend yield answers what a normal year would produce. A weather-adjusted number answers what this season is producing.

That distinction is why a trend yield vs weather-adjusted yield comparison matters on a trading desk. The market has already priced trend yield into the forward curve by planting time. The money gets made or lost on how far reality drifts from that baseline, and how fast the drift shows up in something you can act on before the next USDA print.

Why a trend line alone isn't enough weeks before harvest

The problem with leaning on trend yield through the growing season is timing. It doesn't update when a region gets three inches of rain in August, or misses it. Weather-adjusted models try to close that gap, but most of the public ones run at a national or state level, which smooths over the regional spread that actually drives basis at a given elevator.

What a desk building a production call really wants is a yield trend line regional estimate: one that moves with the season instead of sitting fixed since February, and reads region by region instead of state by state. That means a current-season index curve, built off actual passes over the growing area, regressed against the same historical yield record that produced the original trend number.

That's the gap Yield Forecasting was built to close. It runs the index curve from wide-swath multispectral passes against the historical yield baseline for each region, then refreshes the production estimate and trend chart every week from planting through harvest, instead of leaving you with a trend number set once in spring and a single end-of-season satellite look that arrives too late to trade against.

For a desk pulling together a production call before harvest, neither trend yield nor a weather-adjusted number is granular or current enough for the region you're calling, most of the time.

If you're building that call now, see what a weekly regional print looks like for your region.

Start a project