GIS Ag Maps.org

Landsat Correlation to Soybean Yield

(field maps are developed based on pixels as is shown near bottom of page)

Images transition every 4 seconds (or click arrows or numbers); statistics appear below graphic.

(Left: Landsat NIR for mid R-stage; Right: clean soybean yield averaged to each pixel; darker green is higher yield)

Landsat NIR reflectance of soybean fields during mid R-stages has been shown to highly correlate to spatial patterns of yield (Hollinger, 2011; see Landsat Correlation to Crop Yield folder on left). Below, Landsat NIR is correlated (linear regression) with the average of clean yield points within the pixel extent (based on 4-meter spacing; approximately 56 yield points per pixel). For soybeans, visible bands should not be used for yield prediction during R-stages because they saturate; (indices such as NDVI or GNDVI should not be used as they include visible band data which lowers the correlation with yield as is shown below). 

Correlation (R²) for above pixel groups between NIR and other reflectance-based indices and soybean yield; correlation level is largely a function of the standard deviation of the index value for a field (based on data from Hollinger [2011]):

* Stan. Dev. is NIR reflectance standard deviation. Correlation level is largely a function of index standard deviation as shown here; R² = 0.6692 between NIR and standard deviation (polynomial 2nd order); R² = 0.6477 for linear correlation.

NDVI = (NIR – red) / (NIR + red); GNDVI = (NIR – green) / (NIR + green); NNIR = NIR / (NIR + red + green)

Saturation of visible bands is a main reason why NDVI and other indices above have lower correlation to yield than solely NIR.

Pixels are only used for correlations if they do not average in areas from outside the field (this eliminates many near boundary) or major non-crop surfaces within the field, and if they correspond to valid clean yield points (for example, there can be a void of points adjacent to headland areas while points near obstacles such as electrical installations are typically erroneous). For both maps, darker green is higher value.  Pixels are 30 x 30 meters. 

Knowing that Landsat can provide high yield correlations as shown above, valid pixels from throughout a field (not just those that correspond to valid clean yield points, which is limited throughout a field but necessary for the above correlations) can be applied as the basis for a yield prediction map. The map only includes pixels that represent surface from within the boundary of a field (and excludes non-crop areas from within the boundary if necessary) and has the data extended to the boundary. The map can be calibrated to yield amounts based on an equation or any specified yield range and/or average value; the data is then essentially a generalized yield map (see below). An example of how Landsat can be used to produce a field yield prediction map of continuous data can be viewed though the link that follows:

Soybean yield prediction mapping to field extent 

Steps that show how Landsat can be used to produce a yield prediction map of continuous data or zones are included below (same process for any crop; soybean field shown below) and can be seen in different locations of the website. The map can be produced to a field extent of different shapes and sizes and can be calibrated with yield values based on an equation or a specified yield range and/or average value; the data is then essentially a generalized yield map.    

The progression from raw Landsat imagery to maps to a field extent (or any other extent) is the same for any crop. From left to right the steps are: 1) produce Landsat with pixels that represent correct value to predict yield well enough; 2) use only pixels that represent the crop, not pixels that average in surface outside the field extent or major non-crop surface within the field perimeter; 3) data is interpolated or resampled to a finer resolution for a more coherent map (below, the pixels have been modified from the 30 x 30 meter native resolution to a one-meter resolution); 4) zones can be developed (the appropriate classification method, amount, and minimum size of zone need to be determined).  

Calibration of Landsat Maps to Yield

Landsat yield prediction maps can be calibrated to yield based on solely a field average or a specified yield range and average can be applied. In either case, the calibrated map will keep the same proportions to the Landsat value map (in other words, the map symbology will look the same before and after calibrated to yield).  

If just a field average is known, equations from Hollinger (2011) can be applied that predict yield for the entire field. NIR reflectance values are necessary to apply the equation. The equation predicts normalized (to the mean) yield amounts based on normalized (to the mean) NIR reflectance. Once normalized yield values are calculated they are multiplied by a field average to acquire a yield prediction map. To produce the equation for NIR, pixel groups from many fields over four seasons from mid R-stage image dates were normalized to the mean and combined into one plot where they were correlated with corresponding clean yield monitor data (R² = 0.63; n = 3,807 pixels). The yield variability prediction model is based on fairly typical seasons. 

Maps can also be calibrated based on a specified yield range and field average. An advantage of this is that DN values can be used (atmospheric correction is not necessary) which makes the process less time-consuming.

Reference

Hollinger, D. 2011. Crop Condition and yield prediction at the field scale with geospatial and artificial neural network applications. Dissertation. Kent State University.