Blog

Field notes: the week every wheat field in the batch looked like barley

19 October 2026

Ripe golden wheat field under a blue sky with large white clouds.

Second week of April, a batch of declared winter cereal parcels came through looking wrong. Not wrong in the sense of bad data. Wrong in the sense that a third of the wheat polygons were sitting on a reflectance curve that, three weeks earlier, had belonged firmly to barley.

It's worth saying what normally separates the two on a seasonal signature, because the confusion only makes sense once you've seen the normal case. Winter barley moves first. It tillers earlier, it heads earlier, and by late April it's usually already showing the NDVI plateau and the early senescence tip that marks a crop heading into grain fill. Winter wheat lags behind it by ten days to two weeks in most years, which is plenty of separation for a parcel's seasonal curve to tell the two apart without anyone needing to walk the field.

That gap is also the thing a declared-crop check leans on. A single overpass catches wheat and barley close enough in canopy color that a human eye, or a one-date classifier, can mix them up. The seasonal trajectory doesn't have that problem, because it isn't asking "what does this field look like today," it's asking "what shape did this field's season take." Barley's curve peaks and bends down early. Wheat's peaks later. Different shapes, same sensor.

This spring the gap closed. A cold snap sat over the region through most of March, then a wet stretch in early April kept fields too saturated to push growth. Wheat development that's usually staggered behind barley got compressed against it instead, both crops sitting in roughly the same growth stage at the same time. For about ten days, if you'd pulled a single pass and asked "wheat or barley," you'd have gotten the wrong answer on parcels that were declared, planted, and verified as wheat going back to the sowing record.

The parcels that stayed correctly called were the ones where enough of the season had already accumulated before the cold snap hit. Early-planted wheat had already put down enough of its own curve shape in February and early March that the April compression didn't erase it. The parcels that gave trouble were the later-drilled ones, where most of the available signature was sitting right inside that ten-day window where everything looked the same.

None of this shows up if you're working off a bulletin that tells you "cereal conditions normal for the region." It only shows up when you're looking at the actual shape of the curve on a parcel-by-parcel basis, and it's a good reminder of why a crop-type call that leans on one week of imagery is a coin flip at exactly the weeks it matters most. Weather doesn't delay crops evenly. It bunches them up, and whichever species gets bunched into whichever other species is going to depend on the year, the region, and the two or three weeks that happened to get skipped.

By the first week of May the gap had reopened, barley pushing into head emergence while the wheat sat a clear step behind it, and the batch sorted itself back out without anyone touching a boundary line. That's the part worth remembering before the next cold spring: the confusion is real for a window, but it's a window, not a verdict, which is the whole argument for building a crop-type call off the full season's reflectance signature rather than whatever single pass happened to land during the compressed week. That's the approach behind the parcel-level layer we put together at Crop Type Mapping, and it's worth a look before your next declared-crop check runs into the same wheat-barley week.