Soil sampling: Grid, zones, composites or what?

Soil sampling: Grid, zones, composites or what?

Finding common sense to optimize N, P & K efficiency.

When talking nutrient management, it’s important to think about mobile nutrients (nitrate nitrogen) differently than immobile nutrients (phosphorus and potassium). They act different in the soil, and require different management and soil sampling schemes. 

The very fact that University Extension Services in the Corn Belt recommend a six or eight-inch core for P & K, and a 12 to 36-inch core for nitrate nitrogen establishes the difference in management needed. Zone soil sampling should not be confused with zone management, as they are two different subjects.

For nitrate nitrogen, zone sampling is best because nitrates move with the soil water, but for potassium and phosphorus grid sampling is superior, particularly when pH is of concern. Some agronomists tell their farmer customers that they do it differently (hopefully creating a sales advantage), when the focus should be on doing it better.

One distinct advantage for P & K grids is that they are standard across crops and soils, whereas zones are made of different data layers and they are not necessarily accurate or reflective of the P & K present or needed. There is no standard for zone maps, and that can be justified by numerous factors, but this also leaves any comparisons open to biases and judgment by an agronomist.  Therefore, making comparisons between different type zones and grids is problematical at best and irrelevant at worst.

Some researchers base their fertilizer conclusions on data from several commercial fields, which is not enough to make sweeping conclusions. The company I founded, Midwest Independent Soil Samplers (MISS), has sampled millions of acres, mostly grid, from the central, northern and western Corn Belt.

Nitrate and VR zones

USDA/SCS soil maps, mostly made decades ago without the benefit of GPS positioning or intense sampling, are often used in zone construction. The information in these maps about soil conditions has value, but may have changed over the years due to additional tile drainage and/or irrigation. Soil maps for nitrate management can easily be made from EC (electrical conductivity) or EM (electromagnetic) processes that accurately define changes in soil characteristics in the topsoil and subsoil strata. This data layer is crucial for nitrate management and for variable-rate planting or variable hybrid/variety changes in the field. EM or EC has one minor flaw regarding soils with high salt content – the readings will not be accurate since salts have a high electrical conductive capacity.  Soil moisture can slightly affect readings, but mostly of little significance.Potash application

Yield maps are good for zone management, but not for zone soil sampling, unless their accuracy can be assured. Field agronomists mostly agree that about 50% of yield maps are accurate enough for use, but they may not show where there is excess or deficient nitrogen without being ground truthed.

Dr. Newell Kitchen, USDA-ARS soil scientist in Columbia, Missouri, believes that nitrate soil sampling should be a component for nitrogen management along with whatever form of remote imagery is used. Remote imagery by drone, aircraft or satellite is very useful for rescue treatments of applied nitrogen when adverse weather causes high loss.  Otherwise NDVI has various flaws. It is only a snapshot of one day, and it cannot tell where there is excess nitrogen, nor tell if plant stress is caused by some other nutrient deficiency, weather factor, disease or insect damage.  And by the time stress can be detected, yield loss has already occurred. Nitrate management should be proactive rather than reactive.

 Nitrogen management is complex, but however done, it should be followed with stalk nitrate tests in the fall in the same GPS locations where the soil tests were taken.

Grids for P & K management

One rationalization for moving from grids to zones for P & K is reduced lab fees.  Surely the lab cost for three to 10 zone samples per 80 acres is less than 32 samples in a grid system. In general, all things being equal, more data is better than less; the assumption being that in both instances the data is precise. Not generally taken into consideration, other than the lab fees, are the associated costs for whatever sampling system is used.  The cost for the pickup to drive to the field is the same regardless of the system used in said field. The GPS and field computer equipment used is the same, the office work to accept the data, convert it to maps, send invoices, etc, is the same. None of these are dependent on the number of samples taken in that particular field.

After examining complete costs, just not lab costs, grids are not much more expensive than zones. Dr. Robert Miller, Colorado State and Coordinator of the Agricultural Laboratory Proficiency, adds that one of the big reasons for grid sampling is for pH levels that make a big impact on the bottom line. He says that over four years, any cost savings for zone sampling would be much less than a dollar per acre per year when soil tests impact fertilizer and lime decisions, which cost up to $70 per acre per year.

Craig Struve, General Manager of MISS, has assembled soil data from many different Midwest soil types.  The variance in nutrient and pH levels show very little to as high as nine to one for P in a Waldorf soil type.  pH ranges from none to 5.4-6.4 in a Fostoria soil type. MISS shows for nutrients and pH values, there is often more variation within a soil type as between soil types in the same field. As stated before, soil types are not ranked on their chemical composition, and farming practices over the last century have continued to alter what was there originally. Spatial algorithms are negated, where in the past animal agriculture existed and manure spreading was not done by soil type or zone.

Overlaying soil type maps over grid maps, to move the grid point when necessary, is advocated by some.  I am very cautious about that, since we are now letting the field person or somebody in the office make judgments about moving points. Thus we lose standardization. It is an advantage to take a sample on that eroded knoll, where the probable very high soil test exists, to reinforce just how much a waste of money can be given whole field or average fertilization.

Use expanded center point grid

It has been well established that an expanded center point grid is most efficient in time and the most representative, according to research by Dr. Robert Miller, soil scientist, Colorado State University.  Dr. Miller is Director of Ag Laboratory Proficiency Program, which tests over a hundred soil laboratories for accuracy.  His research shows that 11-12 cores are needed for best representation in a ten foot radius around the center point, particularly in no till or manure applied fields. Taking cores all in a row is totally inadequate because of row-to-row yield variation that affects nutrient removal.  Also inadequate is pulling 5 or 6 cores right around the ATV. Subsequent research by Dr. Fabian Fernandez, University of Minnesota soil scientist, shows that taking cores some distance from the row (not in the row or in the middle of the rows) is best depending on tillage and fertilization soil-sampling

At the advent of variable-rate fertilization in the late 80s and/or 90s, there was a question of how the soil cores should be taken in the grid: center point, random unaligned or composite (as in a transect across the grid as recommended by the University of Minnesota).

Three fields were sampled near Fort Dodge, Iowa, and the soils sent to the University of Minnesota for analysis and recommendation. The analysis showed the center point and composite were about equal and the random unaligned was the least representative. Since the center point system was the most efficient in time and labor, this system became the standard and the U of M discontinued recommending the composite grid because of the extra time and effort required.

Even if random unaligned sampling was superior (it was not), the system has two flaws. First nobody uses the unaligned system commercially so any research comparing it to any other system is suspect, because it has no commercial agricultural application. Also, Dr. Antonio Mallarino, Iowa State soil scientist, has said that the best time of the year for soil sampling is after planting, because by then much of the nutrients in the previous year’s crop residue have been deposited back into the soil.  Farmers will not allow any soil sampler to drive randomly over the field because of crop damage in a growing crop, thus taking away the best time to soil sample.

What then can be gained by using the random unaligned system, no matter its merits or detriments, in research when it is not germane to commercial agriculture. Dr. Jeffrey Strock, Director of the Southwest Experiment Station, University of Minnesota, says he researches what farmers need and want. In my opinion research on unaligned grids is unnecessary or even wasteful.

Fall sampling bad for K

Which brings up the recommendation the Extension services give about what time of the year to sample, and their answer has always been to sample at the same time of the year.  That answer today is only half right. Fall sampling, whether right behind the combine or later in the fall after tillage and rainfall, will give widely divergent answers for potassium particularly because of the large amounts of K held in the corn crop residue that has no chance to convert back into the soil.

Even in southern corn growing areas like Missouri, Kansas and southern Illinois, microbial activity in the soil slows down considerably over late fall and winter, which slows the breakdown of crop residue and subsequent release of nutrients.  So, given fall sampling, consideration must be made for what nutrients will be held in the residue or converted back into the soil before planting next year’s crop.

How often to grid sample is another question.  Because of the vagaries of weather and other factors, the best answer should be, “It depends.”  There are often anomalies in soil sampling that requires two things: retesting the anomaly or tightening the schedule for resampling of the whole field to create a trend line. Trend lines in agronomy are everything. Dr. Mallarino recommends soil testing every two years.  He is correct. Sampling every four years would take nine years to create a trend line, but only five years to create a trend line for a two-year rotation. The four-year rotation was started thirty years ago because that would encompass two crops of soybeans and two of corn for a cost of about $2.50 per acre per year to the farmer. That sounded reasonable to farmers, even when corn was less than two dollars and soybeans under five.

Even as fertilizer prices have risen over the years, there just hasn’t been the impetus to follow Dr. Mallarino’s advice as there should be.  Antonio’s direct quote: “Economic and practical considerations suggest that a new soil test should be planned every two years for most cropping systems. Sampling every three or four years may be acceptable when fields are near optimum levels and maintenance fertilization is being used. For P and K applications between the soil testing years, adjustments should be made to the recommended annual application.”

Soil sampling accuracy

According to Iowa State University, the Sampling Efficacy Index calculates that 0.3 to 0.5 acre grids are 100% accurate, 2.5 acre grids are 50% accurate, zone sampling is 39% accurate and soil type sampling is 22% accurate in nutrient data. I take issue with grid sampling being 50% Dr. Paul Fixen, DIrector of Research for International Plant Nutrition Institute (IPNI), has stated that every field should be PROPERLY grid sampled at least once. However, after the initial sampling is completed, if there is little variance, 2.5-acre grid sampling would be much higher than 50% in accuracy.But if there is considerable variability, a smaller grid size would be needed in subsequent samplings to increase accuracy. That, in the commercial agricultural world, will depend on the price and yield of the crop(s) grown. Nevertheless, the grid system for fixed nutrients is superior according to the Sampling Efficacy Index. Dr. Mallarino concludes speaking about grids, “It describes P, K and pH variation better.  It can be used with yield maps to describe P & K removal.  It adapts well to varable rate applications.”yields-soil sampling

All of the above systems of soil sampling depend on the correct number of cores, properly spaced at a consistent depth.  As one who has sampled soils in New York, Tennessee, Texas, Colorado, the Dakotas, Minnesota and all points in between, that is easier said than done. Hard compacted clay soils are vastly different from loose peat or dry sandy soil.  Inaccurate sampling, leads to inaccurate lab results and subsequent inadequate advice to the farmer. The industry needs to put much more emphasize on accurate core collection. Precise data comes before big data in any sampling scheme.

TAGS: Crops
Hide comments


  • Allowed HTML tags: <em> <strong> <blockquote> <br> <p>

Plain text

  • No HTML tags allowed.
  • Web page addresses and e-mail addresses turn into links automatically.
  • Lines and paragraphs break automatically.