|
Arkansas | Dave,
I agree with what Farmpond said above.
My additional thoughts: I wasn’t sure if you were using the point grid or the grid cell sampling strategy.
Point Grid Strategy. if you took samples from a single location on the grid (rather than averaging the soil cores from the grid and getting an analysis for the average of the grid), then you may want to decide at what distance from the soil sample point will the soil test values be reasonable. If for instance you’re using 2.5 acre grids, there are 330 feet between the closest two sample sites; and in many fields even half this distance is enough for soil conditions and analysis to change. If you opt to go this route you’ll have to use your judgment in determining the pertinent distance from the sample locations. You won’t be using all the yield data available; but if you don’t have the soil test sites then it probably will not correlate well with what you are trying to find.
Grid Cell Strategy: If you took samples from across the grid and averaged together for a composite, then the decision process is simpler with respect to yield data.
We usually do not see a statistical response of crop yield to soil test analyses unless one area of the field has very low soil test levels. In many cases, the fertility levels will be in the adequate (or sufficient) range and therefore not expected to influence yield. It is also normal to see high soil test levels where yields are typically low as Farmpond suggested; where yield was limited by something other than fertility and fertility levels increased by repeated fertilizer applications.
The link below to a document on the Purdue SSMC wesbsite may be useful for some of the technical (operational) computer software steps to perform the above suggestion using GIS.
http://www.agriculture.purdue.edu/ssmc/publications/YieldDataAnalys...
| |
|
Integrating grid sampling with yield data?
