Grid Sampling of Soil

Grid soil sampling evolved after GPS became available to find your spot in the field over and over again. 2.5 acres quickly became the standard for a grid which is ten rectangles in a 50 acre field. Since most 50 acre fields aren't square, you have some larger than others, but you get the concept.

For us gardeners, most of our gardens are less than 2.5 acres so we are farming very tiny grids. Our 40 by 50 garden is only 2000 square feet, a tiny patch of ground compared 4,360 square feet in one acre. We can sample very accurately and get fairly reliable results and recommendations from our soil test.

I said "fairly reliable" because even a scientific soil test can seem to be a "a shot in the dark" as to what is really going on in a soil profile to cause plants to thrive or struggle. It's the best idea man has come up with in the last 100 years of farming so we start there.

Larger acreages are much more complex. Many farmers cover 2,000 acres in a year to make a living for a family of four. Some still make a living on a few hundred acres and it takes some many more thousands to make it. My subject today is why or why not grids?

I do not use grids. I still pull about one sample per acre and an 80 acre field that can be farmed by breaking it into 2 or 3 distinct soil types and yield levels can be treated as separate fields inside one boundary. Many farmers were sold or chose to grid sample their farms in an effort to increase yields and/or decrease costs. I have not seen it to do either very often, if at all.

Still the subject comes up often and my friend Joe Nester, owner of Nester Ag Consulting in Bryan, Ohio made a great post that explains one man's quest to prove grids right or wrong.

"I've been working with soil tests and trying to figure out the best way to be representative, so you can make good nutrient decisions for over 35 years. Grid sampling came out when GPS was made public, before yield monitors. Although we thought we were collecting lots of good data, turned out we were just collecting lots of data that was not very representative. It is actually point sampling, in most cases, 6 or 8 probes taken at the center of the grid. When the data is returned from the lab, it fills the boundary of the grid. Then most software divide each grid into 50' x 50' cells for application, and using mathematical equations, (several theories to chose from), the P & K is varied in each cell based on the theoretical value, that was determined by the theoretical values of the cells next to it. This system creates more variability than was normally present in the field to begin with. Over 97% of the data for 50' application cells is estimated in a 2.5 acre grid. You only sampled 1 of them. When it first came out, it was in conjunction with the invention of the Ag-Chem Soilection machine, and was used to sell the service of that equipment and differentiate from competitors that did not have it.

Even during that time, we were also using hand drawn maps, separating high ground from low ground, sands from clays, and high exchange soils from low exchange soils. We were having farmers spot treat for lime, even without VRT(Variable Rate Transfer) equipment. Then came Windows CE units, then iPac's, and handheld computers that we could hook GPS to. We started driving boundaries, and driving the lines that separated the soil type zones in the field, using the irregular shaped zones to sample and make recommendations, and spreading lime by those maps. Then came the yield monitor and digitized soil surveys, and these were fantastic additions to soils management.

Now- we can combine information from the crop, even normalize yields from different crops, and let the crop tell us where the soil changes. The digitized soil survey is decent information for how they were developed, but it is black and white- you're either in a soil type, or cross the line, and you're in a different soil type. It really doesn't happen that way in the field- there are transition zones where 1 soil type blends to another, and well calibrated yield maps show that.

So having had the opportunity to work with all the above, on lots of acres and lots of soil types, my #1 choice is zones developed from yield maps (well calibrated maps so they show accurate variation in yield), with input from the farmer, and input from the practicing agronomist. If you don't have a yield monitor, use digitized soil surveys, elevation data, and perhaps aerial imagery in crop, and veris data. (But think strongly about a yield monitor, because it can give you some very valuable data about your farm and your practices.)

Precision ag has had a different rate of adoption across the corn belt, with some hot spots having much more experience than others. Those early adopters have moved on from grid sampling to management zone development, and the NRCS and most land grant universities are even agreeing with them. You will seldom find an independent consultant that uses grid sampling anymore. Unfortunately, grid is totally driven by the computer, and does not include input from the farmer, the agronomist, or the crop being raised on the soils.

Yield zone management will be driven by the 2 most important factors in raising a crop: soil type and water holding capacity of those soils. A great blueprint to manage nutrients, lime, and plant populations by. I would also sample at least every other year- it's not exact and you need to keep building representative data. That was another pitfall of grid- so much was spent in one year that farmers tried to live 4 to 5 years off 1 testing cycle. With yield-zone management, and evaluating fields of progressive farmers, you will find that the lower yielding areas have the highest P & K. This is due to non-removal of blanket applications, and the real yield drop is more than likely due to drainage or lime needs. When those areas are found, corrections can be made if possible, and then you can harvest the P & K that is already present in those zones, and have dollars to spend elsewhere."

There you have it. I much agree with Joe's post on Crop Talk. What do you think?

Ed
Winkle