Final Report for GNC14-192
Project Information
The advent of a soil health as a framework for management of row crops in the U.S. Midwest has major implications for agronomic practices. A soil health approach posits a new set of priorities for assessing soil compared to traditional forms of soil testing and resultant management recommendations. Our work examined soil health testing as a means to soil management using three distinct approaches. First, we worked with farmers across the state of Michigan to compare traditional field crop tests with results from soil health tests across a range of farmer fields, to compare how each test captured information related to soil management. Second, we evaluated an additional soil health assay for soil nitrification potential, to compare how this important biological process varied both along an experimental gradient of nitrogen application rates and across farmer fields. Finally, we held interviews with all growers to discuss how they understood the soil test results, and how this information fit with their own experience of managing individual fields. We found that soil health test results described a broad range of soil parameters that characterized differences in physical and biological properties on farmer fields. Compared to traditional soil tests, soil health tests were superior in revealing both positive parameters on good fields and constraints on poor fields, but also showed important limitations for some parameters such as soil compaction. A soil health assay for soil nitrification potential revealed clear increases in the magnitude of this process with increasing N fertilization, and concomitant increases in the microbial populations responsible for nitrification. Measuring nitrification potential on farmer fields suggested a possible new tool for assessing nitrogen management for improved agronomic and environmental outcomes. Finally, soil health test results elicited distinctly broader discussions with growers regarding soil management, compared to traditional soil test results. Growers compared and contrasted soil health results with their own management experience of individual fields and found them more informative to their understanding of specific fields than traditional soils tests. However, growers were less sure how to act on soil health test results. Nearly all growers appreciated the new information and wanted to continue monitoring soil health to help inform management decisions.
Introduction:
Row crop growers in Michigan and throughout the U.S. Midwest manage their fields for multiple outcomes from yield to profitability to environmental quality. To do this, they rely on many kinds of knowledge, including their own experience and information from outside sources such as crop advisers and soil testing agencies. Currently, soil testing is undergoing a major shift in emphasis, from managing for individual inputs/deficits of key nutrients, to overall soil management, a holistic approach often termed soil health. Soil health test parameters continue to evolve, but they differ from traditional soil tests in methodology and in the recommended management guidelines which follow from test results compared to traditional soil tests. Farmer decisions based on soil health test output may fit into a broader management framework, including tillage practices, residue management, or crop rotation. Thus, it is critical that soil health tests provide accurate, field-specific data which farmers can rely for these decisions. In contrast, traditional soil test have generated decades of fine scale data (i.e. recommendations for specific crops in specific counties). Soil health tests are still in the early stages of implementation and need to be vigorously tested. While soil health parameters have proven to be management-sensitive through research, this may not ensure that soil tests capture important differences on farmer fields, or in distinct regions such as those in Michigan. Finally, even if soil health tests accurately capture variability in soil physical, biological and chemical parameters on farmer fields, ultimately the manner in which farmers interpret and implement agronomic decisions based on these results is key to a soil health management framework, and whether it leads to positive agronomic performance for farmers.
Our work assesses multiple components of adapting soil health testing to Michigan. We worked with row crop farmers in three distinct regions of Michigan and asked them to identify four fields to sample for soil health testing, based on their judgment of ‘Best’ and “Worst” fields, a non-row crop (NRC) field, such as a woodlot or hay field, and an additional ‘Choice’ field which they wished to test. In so doing we wished to capture a range of field variability as identified by growers through their own management experience. We ran soil health tests alongside traditional field crop tests for each field. In addition we assessed nitrification potential as a possible additional soil health assay by measuring this process on these fields and on controlled experimental plots. Finally, we discussed test results with each grower to understand how well soil health results reflected their own management practices for each field type.
This work addressed three targets assessing soil health management on Michigan farms. First we analyzed the efficacy of soil health tests parameters across a range of Michigan farms. We also evaluated a test for nitrification potential as a new assay aimed at nitrogen management, on a controlled experimental gradient and on four farm fields. Finally we discussed soil test results with growers to assess how they understood test results and interpreted management options for their own fields. Specifically we sought to:
1) Evaluate how well soil health tests capture management variability on farmer fields.
2) Determine how soil nitrification potential responds to experimentally controlled increases in
nitrogen addition, and if this metric is a meaningful metric for on-farm management.
3) Assess how farmers respond to soil health test results in comparison to traditional soil testing
from their own fields.
Cooperators
Research
1) Comparing soil tests parameters on-farm
We measured management-sensitive parameters of soil health to examine how well these tests characterized soils on Michigan farms, as compared to traditional field crop tests offered by Michigan State University (MSU). This objective informed our sampling approach. We asked growers to identify their Best and Worst fields, a field that was not under crop production, NRC, and finally a fourth ‘wild-card’ field that could be any field of their choosing. This comparison yielded a broad range of fields on which to ‘test’ soil health parameters. It also provided a participatory approach as a means to engage directly with farmers as decision-makers. For example, farmers had the freedom to choose their fields of interest even though it provided less researcher control of management history and other factors that would have allowed for more direct comparisons between fields and test parameters.
From May to June of 2014, we sampled 52 fields on thirteen farms across three counties, ranging from far northern to central to southwestern Michigan – all with contrasting soil types. Together with extension staff we identified growers in each region interested in testing their soils, and prior to sampling took a brief management history of each of the four fields on each farm. At each field we noted soil conditions and recorded crop management observations, and took five soil samples across a representative portion of each field. Using a shovel to make a small pit, we took a ‘slice of soil from the side of the pit, six inches in depth, and approximately 1.5 inches thick and 4 inches wide. All five samples were mixed thoroughly, sub-sampled in the field (2 Kg), and placed on ice until further processing. At each sampling point, two separate penetrometer readings were taken at 6-inch and 18-inch depths.
Samples were maintained at 4 degrees C in the lab and processed as soon as possible after return from farmer fields. At processing, samples were sieved to < 8mm to remove stones and a sub-sample was submitted to the Michigan State University Plant, Soil and Nutrient Laboratory (SPNL) for analysis of pH, total soil organic matter by loss on ignition (SOM), total Bray 1 phosphorus (P), potassium (K), Calcium (Ca), Magnesium (Mg) as well as ammonium (NH4-N) and nitrate (NO3-N). For all but SOM, we grouped these parameters into ‘chemical’ results.
In our lab we carried out soil health analysis following a similar approach as the Cornell Soil Health Laboratory, with the exception of measuring potential root health. For soil physical factors we measured compaction (in field, as above), soil texture, soil aggregate stability and available water capacity. For biological parameters (along with SOM from above) we took measures of permanganate oxidizable carbon (POXC), a colorimetric test for labile soil carbon, and potentially mineralizable nitrogen (PMN), an incubation assay that measures nitrogen derived from organic matter. In addition we added two more biological measures using incubation assays: short term carbon mineralization and soil nitrification potential. These tests were carried out in conjunction with a joint SARE recipient, Dr. Christine Sprunger (SARE GNC14-196) who will report separately on the results of short-term carbon respiration assay.
2) Soil Nitrification Potential
In addition to analyzing data from farms, we also sampled the Resource Gradient Experiment (RGE) at the Kellogg Biological Station (KBS) Long-Term Ecological Research site located in Michigan, USA. This experiment consists of a no-till, corn-soy-wheat rotation established in 2005 and spanning nine rates of nitrogen (N) fertilization in a randomized complete block design. Four blocks are under rain-fed conditions and a further four are under irrigation (dictated by weather and crop modeling predictions). We sampled this experiment prior to corn planting and application of 28% N addition (UAN, according to experimental rates), and after side-dress N application. We followed soil sampling protocols previously established for this experiment, taking multiple core at surface depth (0-10 cm). Soils from the first sampling passed through all of the soil health tests described above; we analyzed only nitrification potential for the second sampling. Soil nitrification potential was measured following Norton and Stark (2011). Briefly, to 8 grams of sieved (2mm), field moist soil were added 40 ml of 0.05 M NH4SO4 buffered at pH 7.2 by 0.2 M phosphate solution. The flask was covered with foil and placed on a rotary shaker at 200 rpm. Beginning within the first hour of shaking, 1.5 ml sub-samples of the slurry were taken at four time intervals over the next 24 hours. Slurry sub-samples were immediately spun down in microfuge tubes for two minutes at 8000g to remove soil, and supernatant was frozen for late measurement of ammonium and nitrite + nitrate concentration. Finally we extracted soil DNA and estimated populations of nitrifying microbes using quantitative PCR according to Dionisi et al. (2002) to estimate nitrification potential per abundance of nitrifier microbes.
3) Farmer Interviews
After compiling all soil testing data, we held interviews with all participant growers. These sessions were divided into three sections. First, we discussed the management history of each field, including crop rotation, tillage, how the farmer had chosen specific management decisions for each field, and why the farmer had assigned each field to each category (i.e., ‘Best’ and ‘Worst’). Second, we discussed specific test results for all fields, grouping results into physical, biological and chemical categories. Third, we facilitated an unstructured discussion of soil test results and the factors that influenced management decisions for specific fields. All interviews were recorded, and notes were written up within 24 hours of each interview. Recordings were transcribed and analyzed for common themes related to each soil test parameter, different approaches to soil management based on field type, and the influence of soil testing on management practices. Finally w combined soil test results and themes that emerged from farmer interviews to design and facilitate a workshop with participating farmers. We recorded and took notes during this group discussion on soil testing and agronomic management.
1) Comparing soil tests parameters on-farm
Soil testing is an important component of agronomic management for row crop farmers. Traditional soil tests have centered on measurement of, and recommendations for, key nutrients (N,P, K) as well as soil pH. In contrast, soil health testing incorporates soil physical and biological measures meant to both reflect management-sensitive parameters and point toward alternative management options for improved soil health. We compared output from traditional soil field crop tests and soil health tests across a broad range of farm fields.
We identified numerous patterns in individual soil health assays across three counties in Michigan. In terms of physical parameters, both aggregate stability (t-test=2.26; p=0.05) and available water capacity (t-test=2.60; p=0.02) significantly differed between pared ‘best’ and ‘worst’ fields among growers (Table 1). Conversely, measures of surface and sub-surface hardness did not differ between these fields and showed no pattern across regions, in spite of a large range of values. The sampled regions (and Michigan in general) have more coarsely textured soils, so it might be expected that aggregate stability and available water capacity were sensitive to changes in organic matter content or soil management that would influence these factors. Compaction, however fluctuated more with the status of the field on the day of sampling. Different fields had been tilled to differing degrees, or had variable amounts of residue or cover which did not necessarily vary by the grower’s ‘best’ and ‘worst’ designations. For biological parameters, neither PMN nor POXC showed significant differences between best and worst fields. Percent SOM was significant in describing differences between pared ‘Best’ and Worst’ fields (t-test=3.51; p=0.005) as did labile carbon measured as respiration (see Sprunger SARE GNC14-196). None of the chemical data measured in the traditional field crops test - pH, P, K, Ca, Mg, ammonium or nitrate - showed any significant difference between Best and Worst fields.
|
Paired t-test comparisons : t (p-value) |
||
|
Soil health parameter |
Best and Worst |
Best and NRC |
|
% Aggregate Stability |
2.26 (0.05) |
0.11 (0.91) |
|
Available Water Capacity |
2.60 (0.02) |
1.54 (0.15) |
|
Surface Hardness (6-inch) |
-1.42 (0.18) |
2.66 (0.02) |
|
Subsurface hardness (18-inch) |
0.04 (0.97) |
1.49 (0.17) |
|
Total organic matter |
3.51 (0.005) |
1.90 (0.09) |
|
Permanganate oxidizable carbon (POXC) |
-0.35 (0.73) |
2.66 (0.02) |
|
Potentially mineralizable nitrogen (PMN) |
0.35 (0.73) |
2.09 (0.06) |
|
Nitrification potential |
0.70 (0.50) |
2.67 (0.02) |
|
pH |
0.26 (0.80) |
0.00 (1.00) |
|
Bray phosphorus (P) |
-0.07 (0.94) |
0.26 (0.80) |
|
Potassium (K) |
0.44 (0.67) |
0.42 (0.68) |
|
Calcium (Ca) |
1.15 (0.28) |
0.54 (0.60) |
|
Magnesium (Mg) |
-0.03 (0.98) |
0.21 (0.84) |
|
Ammonium |
-0.89 (0.39) |
1.30 (0.22) |
|
Nitrate |
0.68 (0.51) |
0.45 (0.66) |
NRC fields ranged greatly in their management/land-use history, depending on the individual farm, from woodlot to hay field to buffer strip. However, all had been under some kind of perennial cover for at least five years. Comparing Best and NRC fields showed a differing pattern in significant soil health parameters (Table1). The total organic matter tended to be higher in these fields compared to best fields, but percent aggregate stability and available water capacity did not significantly differ from Best fields. Surface hardness was much higher than any other treatment type, almost certainly due to the high degree of cover and rooting. Biological parameters on NRC fields revealed much greater differences compared to Best fields. Labile, POXC carbon was significantly higher (t-test=2.66 , p-value=0.02) and PMN was considerably higher (t-test=2.09 , p-value=0.06). Nitrification potential was also significantly lower on NRC treatments (t-test=2.67 , p-value=0.02). Once again, none of the field crop tests for chemical parameters showed any significant trends, with pH, K, Ca, Mg and ammonium tending to be higher on NRC fields, while P and nitrate were generally higher in Best fields.
Of the 52 fields measured for P, K, Mg, or Ca, (204 separate results) only ten results rated below ‘Optimum’ concentration, as designated by the Michigan State University Soil Testing Laboratory. The majority of these were for K, and included all four field types selected by participant growers. By contrast, soil health tests showed some strong patterns in parameter sensitivity across the selected field designations. Soil physical parameters and biological (organic matter) parameters corresponded to Best and Worst field designations, in spite of the limits to exerting control over sampling approach. Further, the stronger contrast of biological parameter measures under NRC fields compared with Best fields revealed the importance of management for soil ecological functions as key to soil health.
2) Nitrification potential as a management-sensitive indicator for soil health
Nitrification is a critical biological process, mediated by microbes, that converts ammonium to nitrate. We propose it as an important soil health parameter because i) nitrification is the process that links two types of plant available N, ii) it is also a potential source of a potent greenhouse gas, nitrous oxide, and iii) its end product, nitrate, is highly mobile and can readily leave the field in surface or ground waters and contribute to downstream eutrophication.
Rates of nitrification differ by land management as well as by other edaphic factors such as soil moisture, temperature, or texture, some of which may be alterable via management (Booth et al 2005). A management-sensitive measure of nitrification potential – or the rate of nitrate generation by a soil under controlled lab conditions (Hart et al. 1994) - could inform growers’ nutrient management plans.
In conjunction with testing nitrification potential on farmer fields, we tested nitrification at the KBS RGE. This was an ideal experiment for testing nitrification potential because of the gradient of N addition as a key experimental variable. Indeed, in spite of the experiment being in place since 2005, soil health test results did not vary across the gradient. These included measures of labile C that might arise from higher-yielding treatments at the higher end of the gradient, and no difference in PMN across the gradient, which might have suggested different-sized pools of SOM-derived ammonia apart from those manipulated in the experiment. The similarity among treatments in other characteristics (aside from N input) increased our confidence that the RGE was an ideal system for measuring nitrification potential responses to increasing rates of N addition. We measured nitrification potential at N rates of 0, 120 and 260 lb/a N) on both irrigated (IR) and rain fed (RF) treatments.
At both at pre-plant and side dress N applications, nitrification potentials showed a strong correlation with treatments based on N rate. IR and RF treatments showed increasing nitrification potential with N rate, but nitrification potentials on IR plots were much higher – ranging from 1.48 mmol N kg-1 day-1 at zero N addition to 2.12 mmol N kg-1 day-1 at the high end of the N gradient. By contrast, the RF treatments had lower nitrification potentials, but a larger magnitude of increase along the N rate gradient, from 0.41 mmol N kg-1 day-1 at zero N addition to 1.55 mmol N kg-1 day-1 at the highest N rate. Finally, we used quantitative PCR to estimate populations of nitrifying bacteria carrying out this this process and found that populations of ammonia oxidizing bacteria and nitrite oxidizing bacteria increased in proportion to nitrification potential rates. As fertilizer N increased along the RGE, the population of nitrifying bacteria increased, in effect priming treatments receiving greater N addition to process, and potentially lose more N at a faster rate.
Once applied to a field, ammonia-based fertilizers begin to undergo nitrification. A soil health test for nitrification could inform a grower how vulnerable a field is to rapid and costly N losses. Management options may be readily deployable, such as timing and placement of N, but a broader list of management tools need to be carefully developed and tested to maximize farmer control of reducing N loss.
3) Farmer interviews on soil health test results
All the farmers we worked with were familiar with results and output from traditional soil tests, and used similar services on a regular basis in evaluating fertilizer recommendations and to adjust soil pH. Indeed many farmers already had a good idea of what these results would look like for each field because they manage according to this information. What emerged from our discussions of soil health tests was the extent to which soil health test results reflected the experience of growers for individual fields. This occurred particularly for soil physical parameters.
One grower described how his Best field acts when it rains, “this one, the water hits it and it absorbs it,” and his Worst field, “but when water hits off this one it beats it, and then hits a table and then just has to evaporate off…then the sun bakes it and you got clay.” This observation was corroborated with strongly contrasting aggregate stability measures for these two fields, with the Best field have highly stable soil aggregates. Similarly, another farmer had great concerns about compaction, “you got to try to be conscientious to what you do in a field. Compaction is huge. That adversely affects yield instantly.” Their farm managed this in terms of limiting field traffic to the extent possible, although they felt they needed to till, even though they were reluctant to do so. Reflecting on his poor soil aggregate stability result he expressed his experience of what was happening in his fields in a new way, “But when those come apart [the aggregates], they'll either become kind of compacted like that and will start to lay flat and get compaction, or will just slake off and you'll just get - I don't know if that crusts at all but it will crust once those aggregates come apart.”
Six out of thirteen farmers were familiar with nitrification as a process and one clearly articulated what this process entailed. Three farmers had experimented with commercial nitrification inhibitors, and two had given up on them as a management tool. Some growers remarked on how nitrification, PMN and soil organic matter were higher in NRC soils and how this must have been what the soil “used to be like,” supporting their usefulness as indicators of biological quality.
Farmers were also interested in the time frame for improving soil health. Even though some knew that increasing soil organic matter was important, it did not impact their management decision horizon, “I am not going to live long enough to do that…”. By contrast, the results we shared regarding more temporally sensitive measures of soil health, as well as the contrasting findings among their fields motivated all participating farmers to either want to test more of their fields for soil health parameters, or track change over time in a given field after several more growing seasons. “I’m interested in it, I think that there’s money to be made in soil health.” “And if we can follow up some of these, …through extension, that's kind of our long-term goal.”
For group discussions, we followed a three-tiered scoring approach to soil health test results, similar to those employed by the traditional field crops test which use ‘below optimum’, ‘optimum’ and ‘above optimum’ to rate soil nutrient levels. However, we normalized for sand content ranked-ordered each parameter (for 52 fields), and then roughly divided into bottom, middle and top thirds. Farmers were able to easily point out how the bulk of top physical and biological soil health measures, across all farms, were disproportionately from Best fields while the Worst fields fell into the boom third of soil health results. A notable point emerging from ensuing discussion was the farmer’s noting that they would want extension staff to know about and be able interpret soil health results to make them a more useful county-level tool.
Educational & Outreach Activities
Participation Summary:
Sprunger, C.D. and B.E. O’Neill. Soil health testing on Michigan farms. Kellogg Biological Station: Brown Bag Seminar Series. Oral. January 2015.
O’Neill, B.E. and C.D. Sprunger. Results of soil health tests on Michigan farms. Farming for the Future Conference. Paw Paw, MI. Extension Talk: Oral. March 2015.
O’Neill, B.E. and C.D. Sprunger. Comparing soil health tests with traditional soil. Extension workshop. Mt. Pleasant, MI. February 2015.
O’Neill, B.E. and C.D. Sprunger. Testing for soil health in Michigan. Extension Bulletin. In progress.
Project Outcomes
1) Soil health testing captures soil factors important for grower management
For farmers across three regions of Michigan, soil health tests demonstrated clear patterns that more completely captured agronomic differences across farmer fields, compared to traditional field crop tests.
For some parameters, such as aggregate stability and available water capacity, soil health measures revealed important differences in farmer fields that clearly related to management, even given the limited scope of this survey. Some parameters often used in soil health surveys, such as POXC and PMN, did not show as much of a contrast on farmer fields of interest to growers, which may have to do with the nature and timing of the sampling as much as the lack of clear discrimination by field designation.
Perhaps more important than how well soil health testing accurately described the soil management practices of each farmer, or what each individual parameter revealed, was the capacity to compare multiple facets of soil management. For example, total organic matter was strongly correlated with PMN on Best soils (slope = 0.22, R2 = 0.70), but not at all on Worst soils for which the relationship was slightly negative. Higher SOM on poor soils did not appear to be related to improved capacity to supply organic N in these soils. Another example of insights from the soil health test comes from examining soil aggregate stability. In our survey, coarser soils tended to have both less SOM and fewer soil aggregates, but those aggregates might be highly stable. For a grower on soils with little capacity to gain SOM, careful monitoring of soil structure and aggregation may be a more meaningful metric for soil management.
Finally soil health metrics provided a powerful means to engage with growers on critical management questions. Farmers had extremely detailed knowledge and observations of most of the fields they designated for testing, including how well the soil ‘worked’, how the soils responded to different crops, how they drained or retained moisture, how well stands established, and how management of each field fit into overall considerations of their farming operation. Information from soil health tests were a powerful tool sparking broader discussion of farmers’ experiences of individual fields including their approach to organic matter management, crop rotation, tillage and other critical soil management factors. By contrast, many growers remarked that the nutrient values from traditional soil test from Best and Worst field were similar. Farmers carefully followed recommendations from traditional tests, but they knew they did not have the specific information they would need to improve problematic (i.e.‘Worst’) fields.
2) Nitrification assay as a potential tool to improve nutrient management
Along experimental treatments of increasing N fertilization rates, on both rain-fed and irrigated treatments, nitrification potential increased significantly with increased N. Along this same gradient, rates of nitrous oxide emissions and nitrate leaching have previously been show to increase – both of these downstream effects are exacerbated by high nitrification rates. This results not only from the effect of increased N addition, but also from growth of the microbial populations that carry out nitrification – a potential positive feedback which can exacerbate N losses from fields. While the limited scope of our farm surveys did not reveal differences in nitrification on cropped fields, comparisons with adjacent NRC fields indicated how nitrification potential changes with management.
Economic Analysis
No economic analysis was conducted for this project.
Farmer Adoption
We did not analyze farmer adoption in this study. However, in interviews, farmers overwhelmingly requested additional soil health test either from MSU, or asked for recommendations for other places to receive these tests. Furthermore, farmer interviews revealed that discussion of soil health test results shifted farmers’ consideration of management options much more than traditional field crop tests.
Areas needing additional study
Soil health tests on Michigan farms captured soil physical and biological variability across a large range field types. Much more testing on farmer fields is required to build stronger links between soil health parameters and specific soil conditions and agronomic management in Michigan. Similarly, recommendations resulting from test results need to be more tightly linked to results from farmer fields and field trials in Michigan.
Many growers wished to pursue soil health testing, but had concerns about the cost. An economic assessment that includes careful agronomic considerations about the frequency and intensity with which soil health tests should be carried out is necessary.
Nitrification potential holds promise as a management-sensitive parameter for nutrient management. Two areas need more careful study for the future i) how well nitrification potential acts as management-sensitive parameter for soil health and ii) to what extent nitrification can be managed-for through reasonable changes in famer practice. Both of these targets are tractable through study on experimental plots and on farmer fields.