A simple exercise that will change your views on variable rate fertilizing

Sunday, March 8, 2015

Using yields from your farm field, you can create Yield Probability Indexes for each zone in the field and use them to guide fertilizer applications or seeding densities, or calculate the profit in those areas

by MIKE R. DUNCAN

For a moment, forget the technology, forget the computers and forget the myriad opinions of what management zones are or should be. I want to introduce a very simple exercise that will likely change your views on variable rate and give you, the farmer, a tool of your own that can easily be used to evaluate what anyone is telling you.

The calculation. Imagine a small area of your farm field which you have been working for 10 years. Going back in your memory year by year, try to remember whether the yield within that cell was higher or lower than the average for that year. Take a pencil and paper and write down the years. Then, if the yield was higher than average for that area or cell, write a one next to the year; otherwise write a zero (0). Add up the ones and you'll get a number between zero and 10. I call this the Yield Probability Index, or YPI for that cell.

If you can do this exercise for every area in your field in your head, then do so. What you will rapidly discover is that the YPI numbers tend to cluster. This means that 10s often have 10s next to them and, if not, then it is most likely a nine (very rarely will it be less than nine). If you make a map of the YPI values, you will see that they, by themselves, delineate zones in your field (YPZones). The 10s are where crops exceed the average every year and the zero areas are where the field under-performs every year.

Of course, the easiest way to do this is with the years of yield data collected using the GPS-driven yield monitor and a small program (see Figure 1). I first developed YPI using a spreadsheet (Figure 2). The calculations can be done with as few as three years of data. We have found that many years of data only add detail to the patterns, which are pretty much fixed with three years.

I tend to group the YPZones into three zones (Figure 3). In this case, I would group 0,1,2,3, into "low," 4, 5, 6 and 7 into "medium" and 8, 9 and 10 into "high." The three zones are the field's low (< 35 per cent), medium (~35-70 per cent) and high (>~80 per cent) potential yield areas, but feel free to group them any way you see fit.

These zones can easily be used to guide variable rate fertilizer applications or seeding densities, or to calculate the profit of the field in these areas. These are the field's "natural" management zones. I say natural since no thresholds or decisions of any kind were required on the part of the farmer – the zones are a result of the year-to-year probability of growth in the field.

Yield Probability Zones are a direct result of yield pattern stability in farm fields (Figure 4). In our work at Niagara College we are finding this in every field we work with.

The idea of yield pattern stability was first introduced to Ontario farmers two decades ago by Doug Aspinall of the Ontario agriculture ministry. In his work, Doug has found that the patterns most closely match landforms within the field. From our perspective, the stability means that we can exploit the patterns to show the YPZones and introduce variable rate techniques.

Interpretations. From an economic perspective, the YPZones with eight to 10 are your highest-producing areas, while YPZones with zero to three are your lowest-producing zones. Zones with high YPI can be pushed to produce more, while areas with really low YPI are usually hopeless in getting them to average or above average productivity. These are areas that are typically so far below average yields, year after year, that economic viability is only achieved by putting nothing there. YPZones with a value of one at least show some indication that they might be recoverable.

What also might jump out at you is the fact that the patterns differ in small ways from year to year, but the overall patterns are stable.

This stability is what Doug Aspinall has been pointing out to farmers for the last two decades.  Stability of yield patterns means that you can exploit the field for better profit. You can exploit the stability using the management zones that result from this simple calculation. Overlaying the grid sampling pattern that you use for soil sampling might also be illuminating, because it is quite possible that the grid entirely misses both the high-producing and low-producing areas.

To me, what a management zone should be doing is separating the field into areas of different statistical behaviour. In the case of this simple YPI calculation, the statistics of the zeros differ greatly from the statistics of the 10s. In fact, the distributions don't even overlap – they are statistically distinct. So, whatever is driving the growth in the zero areas is very different from what is driving growth in the 10 areas, and YPI can separate it all out.

YPI and YPZones are interesting to us because there is no a priori reason that YPI should result in zones. A YPI of five (from a 10-year sample) means that, in five out of the 10 years, this cell over-performed; in the other five years it didn't. An interesting characteristic of this technique is that year-to-year comparisons will show that while the fives tend to cluster, each cell typically over-performed in different years from its neighbour.

For those that can't do this from memory but have all the data, then use one of the popular farm data manipulation tools, take the 10 years and clean and grid each year's yield data. Export the 10 years and the numbered cells to Excel (or something similar) and then use an "if  (cell > average, 1, 0)" statement for each cell for all years and then add up the results for each cell. You'll immediately notice the clustering of values.

You then export the results cell (0-10) back into your software and view the map. The zones are Yield Probability Zones (YPZ or YPZones). A YPZone with a value of six represents a 54 per cent chance of above average yield for that entire zone.

These simply defined YPZones require no decisions or thresholds defined. They are defined solely by the data and represent the potential yield at every point in your field. Fertilize accordingly. BF

Dr. Mike R. Duncan, Ph.D., is Natural Sciences and Engineering Research Council of Canada  Industrial Research Chair for Colleges in Precision Agriculture and Environmental Technologies, Niagara College Research and Innovation.