Technical Summary:

This proposes to develop methodology to delineate soil management zones that 1) best relate crop response to soil factors, and 2) minimize non-point source pollution attributable to field level nutrient management activities such as land application of animal production by-products and other fertilizers, and 3) develop an effective, responsible outreach and educational program based upon the research findings.  Soil management zones will be delineated using a variety of methods on a number of different land management areas.  Methods of delineation will include surrogate measures of soil differences (e.g. topography, hydrology, prior yield history (where available), apparent soil electrical conductivity) and direct soil sampling methods. Management zones will be constructed by clustering similar data into distinct areas and then comparing these areas on the basis of crop response or potential environmental impact using statistical analysis.  The statistical models will utilize state-of-the-art techniques for analyzing spatial data in order to separate systematic effects due to the characteristics of the management zones, as well as residual spatial correlation from random variability.  Results from this study will then be transferred to land managers in the from of extension publications, web sites, short courses, and roundtables.   From this research we expect to find a delineation method or combination of methods that delineate soil management zones which increase the efficiency of inputs while minimizing the environmental impact of those inputs.  These results will potentially increase the efficiency of land management decisions while protecting the environment and, hence, the general public, from any adverse effects of those decisions.  The principal investigators will coordinate with the Intellectual Property and Technology Licensing Office to protect any appropriate findings of this research. 

Objectives: 

1. Determine how to delineate soil management zones that best relate to crop response to soil factors.  Zones will attempt to manage soil inputs two ways.  Manageable inputs (lime, fertilizer, tillage, etc.) will be increased when limiting yield or decreased when a nonmanageable factor (texture, hydrology, etc.) limits yield.  Limiting inputs will be managed on a site-specific basis.
    
2. Determine how to delineate soil management zones that reduce the possibility of non-point source pollution from nutrient management activities (animal waste application, fertilization etc).
    
3. To transfer research findings and methodology to the client/user in the form of publications, short courses, web pages, etc, and to receive feedback from sources outside the university (farmers, industry, academia) to refine and help focus research efforts.

Procedures: 

Objective 1

In each of several different grain or fiber production areas, soil samples 0- to 5-cm and from 5- to 15-cm deep will be collected from management zones determined from a variety of different delineation techniques.  Delineation methods will include those based on surrogate, continuous datasets such as topography, hydrology, prior yield history (if available), and apparent soil electrical conductivity (ECa) as well as direct soil measurements such as fertility, pH, texture, etc.  Direct soil measurements will be conducted on a grid system with the grid spacing appropriate for the size of the field.  Samples will be air-dried, and sieved to < 2-mm particle size.  Basic soil nutrient characterization will include: 1) total carbon and nitrogen by combustion analysis; 2) 1:1 soil:water pH; 3) cation exchange capacity via sum of salt-exchangeable cations via inductively-couple plasma spectroscopy; 4) water- and dilute acid-extractable phosphorus via colorimetry; 5) salt-extractable potassium, calcium, magnesium, and sodium via inductively-couple plasma spectroscopy; 6) salt-extractable inorganic nitrogen (ammonium and nitrate) via colorimetry.  7) MS soil test extractable potassium, phosphorus, calcium, magnesium, and sodium via inductively-couple plasma spectroscopy.  In addition, several deep cores (0 to 1.5 m) from each management area will be collected for standard soil characterization: texture, mineralogy, pH, cation exchange capacity, organic matter content, etc.  Soil ECa will be determined by electromagnetic induction using a Geonics EM-38DD.  Elevation will be determined using RTK GPS.  Topography (slope and aspect) and hydrology will be determined using the continuous elevation measurements and GIS software.   The data from the direct soil measurements will be correlated to the data from the surrogate measurements to determine if the continuous measurements can be used to delineate differences in soil properties.  Both sets of data will be correlated with yield to determine their applicability to soil management zone delineation.  Correlations will be assessed using statistical models for spatial data.  The spatial analysis will begin with a linear regression model, or possibly a principal components regression model if the independent variables are extremely correlated.  After the initial model is fit, the residuals will be analyzed to detect any remaining spatial correlation.  If spatial correlation is present, an appropriate variogram structure will be selected to adjust for the spatial correlation, and then the full regression model will be re-estimated.  The process may be repeated until the best model is determined.  The advantage of this approach is that it removes systematic spatial correlation from the estimate of random variation, and therefore should provide better tests for the true effects of soil properties and management zones on yield, compared to an analysis which ignores spatial correlation.  The yield-limiting factors that are determined to be feasible to correct will then be managed on a site-specific basis.  Examples of these management practices may include, but are not limited to, VRT fertilizer application (available from commercial sources depending on vendor location) or tillage operations to be conducted on a limited spatial basis by the producer as examples.   

Objective 2

In each of several different nutrient management areas, i.e., pasture, forest, pasture-forest mixture, soil samples will be collected from 0- to 5-cm and from 5- to 15-cm depths. The sampling scheme will be determined by first determining spatial extent of each management area and then formulating a grid-spacing that will yield an appropriate number of sampling points per management area. This scheme will be developed using DGPS-GIS technology. In addition, sampling in this way will be conducted twice annually to determine changes in nutrient chemistry over time: September/October and March/April.  Samples will be air-dried, and sieved to < 2-mm particle size. Basic soil nutrient characterization will include: 1) total carbon and nitrogen by combustion analysis; 2) 1:1 soil:water pH; 3) cation exchange capacity via sum of salt-exchangeable cations via inductively-couple plasma spectroscopy; 4) water- and dilute acid-extractable phosphorus via colorimetry; 5) salt-extractable potassium, calcium, magnesium, and sodium via inductively-couple plasma spectroscopy; 5) salt-extractable inorganic nitrogen (ammonium and nitrate) via colorimetry.  In addition, several deep cores (0 to 1. 5 m) from each management area will be collected for standard soil characterization: texture, mineralogy, pH, cation exchange capacity, organic matter content, etc.  Continuous elevation and soil ECa datasets will be collected over each land management area.  These datasets and GIS software will be used to construct topography and hydrology maps for each land management area.   The measurements will be combined to formulate indices relating to the probability of contributing to non-point pollution.  Statistical modeling techniques similar to those described above (see Objective 1) will be used to assess the effects of soil characteristics on the potential for non-point pollution, and to test for differences among management areas. 

Objective 3

Outreach and education will be manifold with results promulgated through the scientific press, Extension curriculum development,  Experiment Station bulletins, websites, county, state, regional, and national presentations, classroom instruction, workforce experience development, dialogue development, and possibly other formats and venues.
 

The following output products are currently envisioned:

1.      Experiment Station technical bulletins,

2.      general audience Extension Information Sheets,

3.      producer information presentations, workshops, and short course,

4.      farm advisor (consultants and Extension personnel) directed training,

5.      publications in popular and peer-reviewed professional journals,

6.      presentations at local, regional, and national meetings,

7.      stakeholder roundtables,

8.      results and recommendations of research will be incorporated into various Extension Service electronic curricula such as websites, CD=s,  radio, and TV, and

9.      regular updates for state agency staff and leaders. 

This work will constantly evolve, thus the curriculum content will be updated often, and, hence, a continuously repeated short course will not meet the needs and demand of the defined publics. 

Specific examples of the extension component of this proposal include:

1.      Train the trainer InService education: Material derived from the project will be incorporated in the MSU-Extension Service's InService training for area agronomy agents as defined in the recent reorganization plan. This training will be offered as a background course in each track (however tracks are not yet defined within the MSU-ES Agriculture and Natural Resources program area) to be defined by each commodity Program Priority Group area.  Projected track title: Defining and Using Soil Management Zones.  The learning objectives of this course are to empower the area agents to provide local oriented programming using state staff originated curriculum materials in site specific soil management.

1. A public oriented short course with the same title, but with more general public oriented learning objectives for non-MSU natural resource management professionals using area agents trained in component 1 as resources and utilizing the state specialist developed curriculum materials.           
2. Relevant short courses and workshops for consultants and producers will be incorporated into existing models such as Mississippi Chemical Corporation Dealer Workshops, Mississippi Department of Agriculture and Commerce Fertilizer Short Courses, Delta Ag Expo, Agronomic Professional Continuing Education Workshop, Mississippi Agricultural Consultants Association annual meeting, and annual meetings of Mississippi focused agricultural science groups such as the MS Chapter of the American Society of Agronomy.  Learning objectives for these outreach activities will be oriented to the particular stakeholder group (consultants, producers, general public, etc.) involved in the activity.

2.      Specific publications: it is difficult to predict titles of extension publications based on research not yet done. However, a series will be developed with the theme and common title: "Managing Mississippi Soils". Individual information sheets will be hyphenated from the common title to establish continuity. 

5. Soil management zones will be featured on the MSUcares website using the same general theme-hyphenated specific subject described for the print publication.

Justification: 

Mississippi has a broad spectrum of soils making up its landscape. It is one of the few states that contains as many as eight of the twelve soil orders. This contributes to the high amount of variability seen throughout its land management areas. Variability in soil chemical and physical properties within individual management areas can potentially affect soil management. In practice, soil management zone delineation is often based solely on fertility measurements with little or no thought given to other yield-limiting factors or the potential environmental impact of land management operations. To be effective however, these management zones must address both soil variability and those soil properties limiting yield while reducing the potential for runoff or erosion. Soil variability can arise from an assortment of different factors. Interactions among parent materials, topography, vegetation, tillage, fertilization, cropping history, etc. can influence the variability of the physical and chemical properties of fields. 

Land managers in Mississippi are increasingly forced to account for these soil changes as they develop nutrient management plans for either the application of fertilizer to crops for increased yield or the application of by-product material as a disposal method. The general public can also be considered to be confronted by this problem in terms of water quality. Land mismanagement can lead to an impaired water quality which affects the general public in terms of aesthetic uses as well as increased costs of remediation or purification. 

This problem is of growing importance in a number of ways. Current commodity prices demand that producers maximize the efficiency of applied fertilizers and lime in order to return a net profit. Determining appropriate soil management zones can lead to an increased profit by either increasing yield in areas of fields that are being underutilized or decreasing fertilization in areas of fields where maximum economic yield has already been attained. In the areas where soils are used for by-product disposal, appropriate soil management zone delineation can reduce the potential for non point source pollution arising from misapplication of by-products.  This proposal attempts to define methodology that is adequate for management zone delineation and is unique in that it uses geospatial technologies to encompass both the production and the environmental aspects of agriculture 

Literature Review:

Variability in soil physical and chemical properties and their relation to crop yield has been well documented in the literature (Pierce et al., 1994; Stein et al., 1997; Mallarino, 1996; Han et al., 1996). As fields may contain a variety of soil series and topographical features, variability in soil chemical and physical properties is much more common than homogenous conditions across a field. Agbu and Olsen (1990) found a wide range of variability in selected soil physical properties. In their study, CV's for slope and aspect were 93.5% and 83.6%, respectively, while the CV for pH in the B horizon was only 9.5%. Brubaker et al., (1993) found that soil physical and chemical properties have the potential to vary significantly with landscape position. Sand, silt, pH, exchangeable Ca, and base saturation generally increased with decreasing elevation, and clay, organic matter, cation-exchange capacity, and available potassium generally decreased with decreasing elevation. In another study, the coefficient of variation (CV) for pH, P, K, Ca, and Mg ranged from 6% (pH) to 81% (Mg) for three fields sampled at 30.5-m grid spacings (Pierce et al., 1994). Corn (Zea mays) yield was also highly variable and little correlation was found between yield and soil fertility, indicating that other factors are influencing yield more than soil fertility. A study by Nolin et al. (1996), investigated soil texture, pH, organic matter, and fertility and showed a positive relation between corn yield and pH. It was shown that pH had a larger influence on corn yield than any other measured soil variable. However, Mallorino et al. (1996) considered several soil physical and chemical parameters in a field study and found that the lack of consistent correlation between variables is not uncommon. They found when relating crop yields to soil variables not all yield variability can be explained by the variability of the measured soil factors and that high variations in soil factors can not always be correlated to high variations in yield. Another important finding by Mallorino et al. (1996) was the importance of which soil factors to consider. The soil factor that may affect yield in one area within a field is subject to change throughout the field hence, it's influence on yield can change. A 1997 study had similar findings (Stein et al., 1997). Of eight soil fertility parameters tested, none were highly correlated with millet (Pennisetum glaucum L.) yield. However, they did find a negative correlation existed between yield and CEC, and attributed this correlation to higher clay, lower fertility soil exposed due to erosion.  Research conducted as part of prior ASTA programs has found a great deal of variability in soil chemical and physical properties in Mississippi land management areas.  Variability in soil parameters, as evidenced by coefficient of variation (CV), differed with pH having the lowest CV and P having the highest CV.  Principal component analysis was used to group highly correlated variables into independent variables which could then be used in regression analysis.  Using this approach, yield variability tended to be explained by principal components representing topography, clay content, and general fertility although these influences were not always straightforward.  These results indicate that, although commonly used, soil fertility parameters are not dependable when delineating soil management zones.  Several attempts have been made to delineate management zones using measurements other than soil fertility.  Myers et al. (2000) used apparent soil electrical conductivity to estimate the physical properties of soils to produce a productivity index in Missouri claypan soils.  This productivity index was based essentially on soil management zones. They found that variations in ECa measurements were significantly related to the soil productivity index, bulk density, plant available water-holding capacity, and soil texture indicating that this measurement may be useful in explaining variation found in yield.  Other methods employed to delineate soil management zones include aerial photos, yield maps, soil survey maps, topography, and farmers management experience (Gerwig et al.,2000; Fleming et al., 2000).  Each of these methods had mixed results and indicated that none of the methods individually could be used to delineate soil management zones.   

Literature Cited: 

Agbu, P.A. and K.R. Olsen, Spatial variability of soil properties in selected Illinois mollisols, Soil Sci. vol 150, pp. 777‑786.

Bemdtsson, R. and A. Bahri. 1995. Field variability of element concentrations in wheat and soil. Soil Sci. 159:311‑320. 

Brubaker, S.C., A.J. Jones, D.T. Lewis, and K. Frank. 1993. Soil properties associated with landscape positions. Soil Sci. Soc. Am. J. 57:235‑239. 

Fleming, K.L., D.G. Westfall, and W.C. Bausch, 2000, Evaluating management zone technology and grid soil sampling for variable rate nitrogen application, in Proceedings, Fifth Int. Conf. On Precision Agriculture, P.C. Robert et al., (Eds.), Minneapolis, MN, ASA, SSSA, CSSA, July 16-19, 2000. 

Gerwig, B.K., E.J. Sadler, and D.E. Evans, 2000, Evaluating techniques for defining management zones in the SE coastal plain, in Proceedings, Fifth Int. Conf. On Precision Agriculture, P.C. Robert et al., (Eds.), Minneapolis, MN, ASA, SSSA, CSSA, July 16-19, 2000. 

Mallarino, A.P. Spatial variability patterns of phosphorus and potassium in no‑tilled soils for two sampling scales, Soil Sci. Soc. Am J. vol 60, pp. 1473‑1481, 1996. 

Myers, D.B. N.R. Kitchen, K.A. Sudduth, and R.J. Miles, 2000, Estimation of a soil productivity index on claypan soils using soil electrical conductivity. in Proceedings, Fifth Int. Conf. On Precision Agriculture, P.C. Robert et al., (Eds.), Minneapolis, MN, ASA, SSSA, CSSA, July 16-19, 2000

Nolin, M.C., S.P. Guertin, C. Wang, 1996. Within‑field variability of soil nutrients and corn yield in a Montreal lowlands clay soil. p. 257‑270. In P.C. Robert et al. (Eds.) Precision Agriculture, Proceedings of the 3 rd International Conference. ASA/CSSA/SSSA, Madison WI. 

Pierce, F.S., D.D. Wamcke and M.W. Everett, Yield and nutrient variability in glacial soils of Michigan, in Proceedings, 2"d Int. Conf. On Site‑Specific Management for Agricultural Systems, P.C. Robert et al., (Eds.), Minneapolis, MN, ASA, SSSA, CSSA, Madison, WI, March 27 ‑ 30 1994, pp. 133‑150. 

Stein, A., J. Brouwer, and J. Bouma, Methods for comparing spatial variability patterns of millet yield and soil data, Soil Sci. Soc. Am. J. vol. 61, pp. 861‑870, 1997. 

Current Research:

Current research is underway attempting to delineate soil management zones in rice and soybean production.  Initial findings indicate that no single soil property serves as a parameter for management zone delineation.  In both soybean and rice production, topography appears to influence yield directly (soybean) and indirectly (rice).  Topography appears to influence plant available water content and hence, directly influence soybean yield while in rice production, leveling high spots and filling low spots in the field appears to influence yield. 

Back to Top

Back to Menu