Crop Residue Bias and Effect on Correlation to Yield
Different crop residues have very different reflectance and can have a large influence on crop reflectance and index calculation if the canopy is not closed enough. (Graphics are from Hollinger [2011].)
Imagery (OGRIP, 2011) of a field in spring of 2006 showing winter wheat on northern area.
Landsat band 3 (red) in the spring of 2007; the darker area corresponds to 2006 harvested wheat.
Below left, is 2007 soybean yield average per pixel (green is higher yield). Below right, is NDVI on 6/11/07 for soybean field, values are higher on northern side of field due to lower red (band 3) values of wheat residue; the lower red band values contribute to the higher NDVI values. (Near infrared values [not shown] are more similar throughout the field than the red band values.)
Landsat NDVI on 7/13/07, canopy has closed enough to reduce the visibility of different residues enough whereby the overall correlation to soybean yield is correct
References
Hollinger, D. 2011. Crop Condition and yield prediction at the field scale with geospatial and artificial neural network applications. Dissertation. Kent State University.
OGRIP. 2012. Ohio Geographically Referenced Information Program. Cited at: http://ogrip.oit.ohio.gov/. Last updated: 2014.