Does Planting Direction Matter?

Way back in 2000 I first got interested in the massive amount of data yield monitors were generating, and how we might process/analyze it.  There are numerous pieces of data captured every second that, along with a yield number, combine speed, swath width, and a direction of travel number (degree).  A producer that had been collecting yield data since 1999 was willing to share his data with me, and after doing quite a bit of data mining, I was intrigued by the planting direction results.

As you can see in the first table every single year from 1995 through 2000 we got a positive response to planting that was done in a North-South vs East-West direction.  Here's a brief description of our data processing technique; we farm in a very contoured environment so there are very few fields planted back and forth in a single direction.  We looked at each yield point and extract the direction at that time.  In a 360 degree circle, we extracted the 45 degrees around due North, 45 degrees around due South, etc.  This meant we basically pulled out half the data that was in, for example, a northeast or southwest direction and discarded it.  This is a very basic overview, if you are interested in the details on this process please contact me.

The first question we need to ask with any data analysis result (especially when we are not doing statistical data analysis, we are doing much more of what I have heard referred to as “observational data analysis”) is:  Why might the correlation I be seeing not be the underlying cause?  In this case - what other factors may be contributing to the data showing North-South rows being better.  The only one I can think of is that in this farmer's case his better/higher yielding soil was planted North-South more often.  

One way we evaluted this was to look at a single year of data from four different area producers.  We saw a very similar pattern.  Both of these data sets cover several thousand acres of corn in northeast Kansas.  While it is still possible that soil type or another unknown variable is causing this response, based on the available data I have a reasonable confidence that North South rows truly are yielding better.  Now, we are all aware that the corn seed doesn’t look at what direction the planter was running and say “Oh, I’m going to push out a bit more yield now” - that is ridiculous - so what is the underlying reason this might be occuring?  I am not an agronomist, and have no idea. In talking to several people over the years, the only realistic idea I have heard is that possibly there is a sunlight/plant canopy interaction, where the amount of sun hitting the plant on rows running North-South is higher than an East-West direction.  Valid?  No one can be sure, but if we believe the data to be valid, there has to be some underlying reason.  Do you have any thoughts as to why North-South planted rows could yield better than East-West planted ones?

Bottom line, it is difficult to use the knowledge gained from this data in our environment.  We do not feel that planting straight north-south, or east-west for that matter, is a wise decision in our environment, so there was only a few flatter fields/bottom ground that some producers did adjust their planting direction based on this data.  

Recently, we looked at the data on our own farm for the 2016 crop year.  The correlation/yield response using an entirely different analysis technique was much lower.  It is something I would like to dig deeper into, but at the present time I have not found a software package that makes this process easy, so there is a large amount of time and effort required to format and analyze the data this way.