Death comes to us all. Life's but a walking shadow | Curious as to whether the data would show a real effect of current crop conditions on yield, I put together a data set for Illinois using the USDA crop conditions reports from 1986 to 2016. Yield trend and June & July precipitation accounts for about 79% of the observed variation in yield but the addition of initial crop conditions contributes another 10% to the model for a r-value of 89%. There is however, still substantial variation in yield not explained by the model parameters. A graph showing the relationship between the estimated model yield and the actual yield is shown below.
Given that the analysis uses multiple regression it's a little hard to easily display the results. One way to illustrate the effect of the initial crop conditions on yield is by example. For instance, the model predicts that 2016 yields of 199 bpa as opposed to the 197 bpa actually observed. If you use the 2017 crop conditions along with the 2016 precipitation observations, the model predicts a yield of 183 bpa, a difference of 16 bpa. If you use the average precipitation values for the past 31 years of about 4.5 & 4.0 inches for June & July respectively, the model predicts a yield of 179 bpa using the 2017 initial crop conditions.
Either way, the model lends some support to the idea that this year's initial crop conditions will have an effect on yield.
For what it's worth.
 (ill cond s model-page-001.jpg)
Attachments
----------------
 ill cond s model-page-001.jpg (55KB - 119 downloads)
|
Zen, crop conditions model results
