ClusterOS

Crop development

Low expected yield: a season-long forecast from satellite images

Low expected yield is a forecast that a field will harvest less than is normal for its crop and area. Yield grows with the green biomass the crop builds up over the season, so a lag in biomass signals a shortfall, and contract losses, well before harvest. The risk is a summary: it adds up the effect of drought, salinity, heat, nutrient shortage and missed field work, but does not name the cause itself.

Crops: Cotton, Wheat, Rice

When it is dangerous

The forecast is refined throughout the season. A useful estimate is available from mid-season, and the most reliable one about two months before harvest. Accuracy improves fastest in the reproductive period, from flowering to grain or boll filling.

Causes

  • Drought and shortage of irrigation water
  • Soil salinity
  • Heat
  • Nutrient shortage
  • Missed field operations

Signs on the field

  • The crop has built up less greenness this season than usual for the field or area
  • A low peak of green biomass at mid-season
  • Little accumulated biomass by the filling stage

How ClusterOS notices it

ClusterOS sums the greenness indices NDVI and EVI, measures of green biomass from Sentinel-2 optical images, over the whole season and compares them with the history of the same field since 2019 and with fields in the area. The earliest signal is a lag behind the normal level for the field or area. The forecast does not name the cause: that comes from the other risks on the field, such as drought, salinity, heat or nutrition. In dense crops the index saturates and separates fields less well, and early forecasts can be unstable from year to year.

What to do

  1. Check which other risks are flagged on the field, as they point to the cause
  2. Inspect the field with an agronomist and compare with the irrigation and fertilizer log
  3. If irrigation or fertilizer can fix the cause, act before the season ends
  4. Factor the expected shortfall into delivery plans and contracts

Questions

Can satellite images predict cotton yield?

Yes, with caveats. Green biomass accumulated over the season is linked to yield, and such models have been tested for cotton in Xinjiang, a region with a similar climate. The coefficients still need local calibration, and a reliable forecast is available about two months before harvest.

Why does the yield forecast not say what the cause is?

A low yield is the sum of every problem in the season. To find the cause, the forecast is compared with the other risks on the field and with a field inspection.

Learn about this risk on your fields before it becomes a loss

We will show ClusterOS on your own fields: history since 2019, crop stage and all 24 risks.

Other risks in «Crop development»

Scientific basis

  • Lang P., Zhang L., Huang C., Chen J., Kang X., Zhang Z., Tong Q. (2023) Frontiers in Plant Science 13:1048479, DOI: 10.3389/fpls.2022.1048479
  • Bokusheva R. et al. (2016) Agricultural and Forest Meteorology 220:200–206, DOI: 10.1016/j.agrformet.2015.12.066
  • Monteith J.L. (1977), DOI: 10.1098/rstb.1977.0140
  • Steduto P. et al. (2012) Crop Yield Response to Water, FAO Irrigation and Drainage Paper 66

Request a demo

We will call or message you in Telegram and show ClusterOS on your fields.

Who you are