Kenneth A. Sudduth

Accuracy issues in electromagnetic induction sensing of soil electrical conductivity for precision agriculture

Soil apparent electrical conductivity (ECa) has been used as a surrogate measure for such soil properties as salinity, moisture content, topsoil depth (TD), and clay content. Measurements of ECa can be accomplished with commercially available sensors and can be used to efficiently and inexpensively develop the dense datasets desirable for describing within-field spatial variability in precision agriculture. The objective of this research was to investigate accuracy issues in the collection of soil ECa data. A mobile data acquisition system for ECa was developed using the Geonics EM38 1 sensor. The sensor was mounted on a wooden cart pulled behind an all-terrain vehicle, which also carried a GPS receiver and data collection computer. Tests showed that drift of the EM38 could be a significant fraction of within-field ECa variation. Use of a calibration transect to document and adjust for this drift was recommended. A procedure was described and tested to evaluate positional offset of the mobile EM38 data. Positional offset was due to both the distance from the sensor to the GPS antenna and the data acquisition system time lags. Sensitivity of ECa to variations in sensor operating speed and height was relatively minor. Procedures were developed to estimate TD on claypan soils from ECa measurements. Linear equations of an inverse or power function transformation of ECa provided the best estimates of TD. Collection of individual calibration datasets within each surveyed field was necessary for best results. Multiple measurements of ECa on a field were similar if they were obtained at the same time of the year. Whole-field maps of ECa-determined TD from multiple surveys were similar but not identical. There was a significant effect of soil moisture and temperature differences across www.elsevier.com:locate:compag

Management Zone Analyst (MZA)

different management zones within a field? Two, how can information be processed into unique management Producers using site-specific crop management (SSCM) have a units (i.e., procedures for classification)? And three, how need for strategies to delineate areas within fields to which management can be tailored. These areas are often referred to as management many unique zones should a field be divided into? Quick, zones. Quick and automated procedures are desirable for creating efficient, and automated procedures are needed that admanagement zones and for testing the question of the number of zones dress these questions. to create. A software program called Management Zone Analyst A number of information sources have been used to (MZA) was developed using a fuzzy c-means unsupervised clustering delineate subfield management zones for SSCM. Tradialgorithm that assigns field information into like classes, or potential tional soil surveys often provide estimates of crop promanagement zones. An advantage of MZA over many other software ductivity for each soil map unit. In the USA, county programs is that it provides concurrent output for a range of cluster soil surveys report the average yield of major crops and numbers so that the user can evaluate how many management zones various soil properties by soil map unit; but the spatial should be used. Management Zone Analyst was developed using Microscale of county soil surveys has often been found inadesoft Visual Basic 6.0 and operates on any computer with Microsoft Windows (95 or newer). Concepts and theory behind MZA are prequate for use in SSCM (Mausbach et al., 1993). Digital sented as are the sequential steps of the program. Management Zone elevation data collected using global positioning systems Analyst calculates descriptive statistics, performs the unsupervised (GPS) or total station surveys have been used for classifuzzy classification procedure for a range of cluster numbers, and profying a field into management zones (McCann et al., vides the user with two performance indices [fuzziness performance 1996; Lark, 1998; MacMillan et al., 1998; van Alphen and index (FPI) and normalized classification entropy (NCE)] to aid in Stoorvogel, 1998). Fleming et al. (2000) used aerial phodeciding how many clusters are most appropriate for creating managetographs of bare soil along with landscape position and ment zones. Example MZA output is provided for two Missouri claythe management experience of the producer to delinpan soil fields using soil electrical conductivity, slope, and elevation as eate within-field management zones. Because of the clustering variables. Management Zone Analyst performance indices relationship of bulk soil apparent electrical conductivity indicated that one field should be divided into either two (using NCE) or four (using FPI) management zones and the other field should be (ECa) to productivity on some soils (Kitchen et al., 1999, divided into four (using NCE or FPI) management zones. 2003), it has been used in the delineation of management units. Sudduth et al. (1996) and Fraisse et al. (2001a) used a combination of topographic attributes and soil ECa to delineate management zones. Long et al. (1994) S crop management promotes the concept investigated the accuracy and precision of field manageof identification and management of regions within ment maps created from several sources [e.g., soil survey the geographic area defined by field boundaries. Often map, aerial photograph, overlaying class values of kriged referred to as management zones, these subfield regions point data in a geographic information system (GIS)]. typically represent areas of the field that are similar based They concluded that aerial photographs of growing crops on some quantitative measure(s) (e.g., topography, yield, were the most accurate and precise for classifying a field and soil-test nutrients). Determination of subfield areas into management units to predict grain yield. Imagery is difficult due to the complex combination of soil, biotic, of a growing crop and yield data collected in the same and climate factors that may affect crop yield. These year would be highly correlated and thus an accurate factors dynamically interact, further complicating the representation of crop production potential for that spedecisions of how to manage by zones. Three questions typically arise when considering managing by zones. One, cific year (Boydell and McBratney, 1999). what information should be used as a basis for creating Delineating zones based on topographic attributes and/or soil physical properties most often captures yield J.J. Fridgen, ITD/Spectral Visions, 20407 South Neil Street Suite 2, variability due to differences in plant available water and Champaign, IL 61820; N.R. Kitchen, K.A. Sudduth, and S.T. Drumthus, crop production potential (McCann et al., 1996; mond, USDA–ARS, Cropping Syst. and Water Quality Res. Unit, van Alphen and Stoorvogel, 1998; Fraisse et al., 2001a). Columbia, MO 65211; W.J. Wiebold, Dep. of Agron., University of Missouri, Columbia, MO 65211; and C.W. Fraisse, Agric. and Biol. Eng. Dep., Univ. of Florida, Gainesville, FL 32611. Received 22 Aug. Abbreviations: ECa, apparent soil electrical conductivity; FPI, fuzzi2002. *Corresponding author (kitchenn@missouri.edu). ness performance index; GIS, geographic information systems; ISODATA, Iterative Self-Organizing Data Analysis Technique; MZA, Published in Agron. J. 96:100–108 (2004).  American Society of Agronomy Management Zone Analyst; NCE, normalized classification entropy; SSCM, site-specific crop management. 677 S. Segoe Rd., Madison, WI 53711 USA

Comparison of electromagnetic induction and direct sensing of soil electrical conductivity

et al. (1989) modeled ECa as a function of soil water content (both the mobile and immobile fractions), the Apparent profile soil electrical conductivity (ECa) can be an indielectrical conductivity (EC) of the soil water, soil bulk rect indicator of a number of soil physical and chemical properties. Commercially available ECa sensors can be used to efficiently and density, and the EC of the soil solid phase. inexpensively develop the spatially dense data sets desirable for deMeasurements of ECa can be used to provide indirect scribing within-field spatial soil variability in precision agriculture. measures of the soil properties listed above if the contriThe objective of this research was to compare ECa measurements butions of the other soil properties affecting the ECa from a noncontact, electromagnetic induction–based sensor (Geonics measurement are known or can be estimated. If the ECa EM38)1 to those obtained with a coulter-based sensor (Veris 3100) changes due to one soil property are much larger than and to relate ECa data to soil physical properties. Data were collected those attributable to other factors, then ECa can be on two fields in Illinois (Argiudoll and Endoaquoll soils) and two in calibrated as a direct measurement of that dominant Missouri (Aqualfs). At 12 to 21 sampling sites in each field, 120-cmfactor. Lesch et al. (1995a, 1995b) used this direct-calideep soil cores were obtained for soil property determination. Depth bration approach to quantify variations in soil salinity response curves for each ECa sensor were derived or obtained from the literature. Within a single field and measurement date, EM38 data within a field where water content, bulk density, and and Veris deep (0–100 cm depth) data were most highly correlated other soil properties were “reasonably homogeneous.” (r 0.74–0.88). Differences between ECa sensors were more proResearch in Missouri has established direct, within-field nounced on the more layered Missouri soils due to differences in calibrations between ECa and the depth of topsoil above depth-weighted response curves. Correlations of ECa with response a subsoil claypan horizon (Doolittle et al., 1994; Sudduth curve–weighted clay content and cation exchange capacity were generet al., 1995, 2001; Kitchen et al., 1999). ally highest and most persistent across all fields and ECa data types. Mapped ECa measurements have been found to be Significant correlations were also seen with organic C on the Missouri related to a number of soil properties of interest in fields and with silt content. Significant correlations of ECa with soil precision agriculture, including soil water content (Sheets moisture, sand content, or paste EC were observed only about 10% and Hendrickx, 1995), clay content (Williams and Hoey, of the time. Data obtained with both types of ECa sensors were similar and exhibited similar relationships to soil physical and chemical prop1987), CEC, and exchangeable Ca and Mg (McBride et erties. al., 1990). Because ECa integrates texture and moisture availability, two soil characteristics that affect productivity, it can help to interpret spatial grain yield variations, at least in certain soils (e.g., Sudduth et al., 1995; Jaynes E and accurate methods of measuring et al., 1993; Kitchen et al., 1999). Other uses of ECa in within-field variations in soil properties are imporprecision agriculture have included refining the boundtant for precision agriculture (Bullock and Bullock, aries of soil map units (Fenton and Lauterbach, 1999), 2000). Apparent profile soil electrical conductivity is interpreting within-field corn rootworm (Diabrotica one sensor-based measurement that can provide an indibarberi Smith and Lawrence) distributions (Ellsbury et rect indicator of important soil physical and chemical al., 1999), and creating subfield management zones properties. Soil salinity, clay content, cation exchange (Fraisse et al., 2001). capacity (CEC), clay mineralogy, soil pore size and disTwo types of portable, within-field ECa sensors have tribution, soil moisture content, and temperature all been used in agriculture—an electrode-based sensor reaffect ECa (McNeill, 1992; Rhoades et al., 1999). In quiring direct contact with the soil and a noncontact saline soils, most of the variation in ECa can be related electromagnetic induction (EM) sensor. The earliest to salt concentration (Williams and Baker, 1982). In sensors were of the contact type and included four elecnonsaline soils, conductivity variations are primarily a trodes inserted into the soil, coupled with an electric function of soil texture, moisture content, and CEC current source and resistance meter. Hand-carried four(Rhoades et al., 1976; Kachanoski et al., 1988). Rhoades electrode sensors were initially used in salinity surveys (Rhoades, 1993), and later versions were tractor1Mention of trade names or commercial products is solely for the mounted for mobile, georeferenced measurement of purpose of providing specific information and does not imply recommendation or endorsement by the USDA or the Univ. of Illinois. ECa. The electrode-based sensing concept formed the K.A. Sudduth and N.R. Kitchen, USDA-ARS, Cropping Syst. and Abbreviations: CV, coefficient of variation; DGPS, differential global Water Qual. Res. Unit, 269 Agric. Eng. Bldg., Univ. of Missouri, positioning system; EC, electrical conductivity; ECa, apparent soil Columbia, MO 65211; G.A. Bollero and D.G. Bullock, Dep. of Crop electrical conductivity; ECa-sh, shallow (0–30 cm) apparent soil electriSci., Univ. of Illinois, 1102 S. Goodwin Ave., Urbana, IL 61801; and cal conductivity measured by Veris 3100; ECa-dp, deep (0–100 cm) W.J. Wiebold, Dep. of Agron., Univ. of Missouri, 214 Waters Hall, apparent soil electrical conductivity measured by Veris 3100; ECa-em, Columbia, MO 65211. Received 10 July 2001. *Corresponding author vertical-mode apparent soil electrical conductivity measured by Geo(SudduthK@missouri. edu). nics EM38; EM, electromagnetic induction; GPS, global positioning system; TD, topsoil depth. Published in Agron. J. 95:472–482 (2003).

Soil moisture and organic matter prediction of surface and subsurface soils using an NIR soil sensor

Sensors are needed to document the spatial variability of soil parameters for successful implementation of Site-Specific Management (SSM). This paper reports research conducted to document the ability of a previously developed near infrared (NIR) reflectance sensor to predict soil organic matter and soil moisture contents of surface and subsurface soils. Three soil cores (5.56 cm dia. ×1.5 m long) were collected at each of 16 sites across a 144 000 km 2 area of the US Cornbelt. Cores were subsampled at eight depth increments, and wetted to six soil moisture levels ranging from air-dry to saturated. Spectral reflectance data (1603–2598 nm) were obtained in the laboratory on undisturbed soil samples. Data were collected on a 6.6 nm spacing with each reflectance value having a 45 nm bandpass. The data were normalized, transformed to optical density [OD, defined as log10 (1/normalized reflectance)], and analyzed using stepwise multiple linear regression. Standard errors of prediction for organic matter and soil moisture were 0.62 and 5.31%, respectively. NIR soil moisture prediction can be more easily commercialized than can soil organic matter prediction, since a reduced number of wavelength bands are required (four versus nine, respectively).

Soil Electrical Conductivity and Topography Related to Yield for Three Contrasting Soil–Crop Systems

Along with yield mapping, producers have expressed increased interest in characterizing soil and topographic Many producers who map yield want to know how soil and landvariability (Wiebold et al., 1998). Numerous properties scape information can be used to help account for yield variability influence the suitability of soil as a medium for crop and provide insight into improving production. This study was conducted to investigate the relationship of profile apparent soil electrical root growth and yield. These include soil water-holding conductivity (ECa) and topographic measures to grain yield for three capacity, water infiltration rate, texture, structure, bulk contrasting soil–crop systems. Yield data were collected with combine density, organic matter, pH, fertility, soil depth, topograyield-monitoring systems on three fields [Colorado (Ustic Haplarphy features (i.e., slope, aspect, etc.), the presence of gids), Kansas (Cumuic Haplustoll), and Missouri (Aeric Vertic Epiarestrictive soil layers, and the quantity and distribution qualfs)] during 1997–1999. Crops included four site-years of corn (Zea of crop residues. These properties are complex and vary mays L.), three site-years of soybean (Glycine max L.), and one sitespatially (and with some, temporally) within fields. No year each of grain sorghum [Sorghum bicolor (L.) Moench] and winter single measurement adequately describes the influence wheat (Triticum aestivum L.). Apparent soil electrical conductivity of the soil environment on rooting and crop growth and was obtained using a Veris model 3100 sensor cart system. Elevation, obtained by either conventional surveying techniques or real-time yield. Georeferenced soil sampling for fertility status, kinematic global positioning system, was used to determine slope, typically from the surface layer from 0 to 20 cm, is often curvature, and aspect. Four analysis procedures were employed to used by producers in developing recommendation maps investigate the relationship of these variables to yield: correlation, for variable-rate fertilizer application. Information obforward stepwise regression, nonlinear neural networks (NNs), and tained from these samples [including fertility, organic boundary-line analysis. Correlation results, while often statistically matter, cation exchange capacity (CEC), and texture] significant, were generally not very useful in explaining yield. Using has also been used in some research to evaluate yield either regression or NN analysis, ECa alone explained yield variability variation (Kravchenko and Bullock, 2000; Nolin et al., (averaged over sites and years R2 0.21) better than topographic 2001; Ward and Cox, 2001), but usually little or no variables (averaged over sites and years R2 0.17). In six of the nine site-years, the model R2 was better with ECa than with topography. significance has been found. Combining ECa and topography measures together usually improved Inexpensive and accurate methods for measuring model R2 values (averaged over sites and years R2 0.32). Boundary within-field soil variation would have the potential to lines generally showed yield decreasing with increasing ECa for Kansas greatly improve site-specific crop management. Sensors and Missouri fields. Results of this study can benefit farmers and are ideal for mapping soil properties because they can consultants by helping them understand the degree to which sensorprovide data without the need to collect and analyze based soil and topography information can be related to yield variation samples and can be linked to global positioning systems for planning site-specific management. (GPS) and computers for on-the-go spatial data collection. Sensors that measure soil properties could play an important role in helping to characterize yield variation. Y monitoring and mapping have given producOne sensor-based measurement that has shown ers a direct method for measuring spatial variability promise is ECa, which is a measure of the ability to in crop yield (Lark and Stafford, 1996; Pierce and Noconduct electrical current through the soil profile. Sevwak, 1999). Yield maps have shown high-yielding areas eral authors have reported on relating ECa to variation to be as much as 150% higher than low-yielding areas in crop production caused by soil differences (Jaynes et (Kitchen et al., 1999) and have revolutionized the way al., 1995; Kitchen et al., 1999; Luchiari et al., 2001; Zhang producers view yield as they seek to learn how they and Taylor, 2001). Rapid spatial measurement of ECa might improve production. However, yield maps are can be accomplished using noncontact electromagnetic confounded by many potential causes of yield variability induction sensors (McNeil, 1992; Jaynes et al., 1993; (Pierce et al., 1997) as well as potential error sources Sudduth et al., 2001) or with direct-contact sensors such from combine yield sensors (Lamb et al., 1995; Blackas rolling coulters that measure electrical resistance dimore and Marshall, 1996). When other georeferenced rectly (Lund et al., 1999; Sudduth et al., 1999). In geninformation is available, producers naturally want to eral, ECa can be affected by a number of different soil know if and how these various layers of data can be properties, including clay content, soil water content analyzed to help explain yield variability and provide (Kachanoski et al., 1990; Morgan et al., 2001), varying insight into improving production practices. depths of conductive soil layers, temperature, salinity, N.R. Kitchen, S.T. Drummond, and K.A. Sudduth, USDA-ARS, Abbreviations: CEC, cation exchange capacity; DEM, digital elevaCropping Syst. and Water Qual. Res. Unit, Columbia, MO 65211; tion model; ECa, apparent soil electrical conductivity; ECa-dp, deep E.D Lund, Veris Technol., 601 N. Broadway, Salina, KS 67401; and (100 cm) apparent soil electrical conductivity; ECa-sh, shallow (30 cm) G.W. Buchleiter, USDA-ARS, Water Manage. Unit, Ft. Collins, apparent soil electrical conductivity; GPS, global positioning system; CO 80523. Received 1 June 2001. *Corresponding author (kitchenn@ MLR, multiple linear regression; MQR, multiple quadratic regression; missouri.edu). MQR Int, multiple quadratic regression including two-way linear interactions; NN, neural network. Published in Agron. J. 95:483–495 (2003).

Remote- and Ground-Based Sensor Techniques to Map Soil Properties

Farm managers are becoming increasingly aware of the spatial variability in crop production with the growing availability of yield monitors. Often this variability can be related to differences in soil properties (e.g., texture, organic matter, salinity levels, and nutrient status) within the field. To develop management approaches to address this variability, high spatial resolution soil property maps are often needed. Some soil properties have been related directly to a soil spectral response, or inferred based on remotely sensed measurements of crop canopies, including soil texture, nitrogen level, organic matter content, and salinity status. While many studies have obtained promising results, several interfering factors can limit approaches solely based on spectral response, including tillage conditions and crop residue. A number of different ground-based sensors have been used to rapidly assess soil properties “on the go” (e.g., sensor mounted on a tractor and data mapped with coincident position information) and the data from these sensors compliment image-based data. On-the-go sensors have been developed to rapidly map soil organic matter content, electrical conductivity, nitrate content, and compaction. Model and statistical methods show promise to integrate these groundand image-based data sources to maximize the information from each source for soil property mapping.

Proximal Soil Sensing: An Effective Approach for Soil Measurements in Space and Time

Abstract This chapter reviews proximal soil sensing (PSS). Our intent is for it to be a source of up-to-date information on PSS, the technologies that are currently available and their use for measuring soil properties. We first define PSS and discuss the sampling dilemma. Using the range of frequencies in the electromagnetic spectrum as a framework, we describe a large range of technologies that can be used for PSS, including electrochemical and mechanical sensors, telemetry, geographic positioning and elevation, multisensor platforms, and core measuring and down-borehole sensors. Because soil properties can be measured with different proximal soil sensors, we provide examples of the alternative techniques that are available for measuring soil properties. We also indicate the developmental stage of technologies for PSS and the current approximate cost of commercial sensors. Our discussion focuses on the development of PSS over the past 30 years and on its current state. Finally, we provide a short list of general considerations for future work and suggest that we need research and development to: (i) improve soil sampling designs for PSS, (ii) define the most suitable technique or combination of techniques for measuring key soil properties, (iii) better understand the interactions between soil and sensor signals, (iv) derive theoretical sensor calibrations, (v) understand the basis for local versus global sensor calibrations, (vi) improve signal processing, analysis, and reconstruction techniques, (vii) derive and improve methods for sensor data fusion, and (viii) explore the many and varied soil, agricultural, and environmental applications where proximal soil sensors could be used. PSS provides soil scientists with an effective approach to learn more about soils. Proximal soil sensors allow rapid and inexpensive collection of precise, quantitative, high-resolution data, which can be used to better understand soil spatial and temporal variability. We hope that this review raises awareness about PSS to further its research and development and to encourage the use of proximal soil sensors in different applications. PSS can help provide sustainable solutions to the global issues that we face: food, water, and energy security and climate change.

Statistical and Neural Methods for Site-Specific Yield Prediction

Understanding the relationships between yield and soil properties and topographic characteristics is of critical importance in precision agriculture. A necessary first step is to identify techniques to reliably quantify the relationships between soil and topographic characteristics and crop yield. Stepwise multiple linear regression (SMLR), projection pursuit regression (PPR), and several types of supervised feed-forward neural networks were investigated in an attempt to identify methods able to relate soil properties and grain yields on a point-by-point basis within ten individual site-years. To avoid overfitting, evaluations were based on predictive ability using a 5-fold cross-validation technique. The neural techniques consistently outperformed both SMLR and PPR and provided minimal prediction errors in every site-year. However, in site-years with relatively fewer observations and in site-years where a single, overriding factor was not apparent, the improvements achieved by neural networks over both SMLR and PPR were small. A second phase of the experiment involved estimation of crop yield across multiple site-years by including climatological data. The ten site-years of data were appended with climatological variables, and prediction errors were computed. The results showed that significant overfitting had occurred and indicated that a much larger number of climatologically unique site-years would be required in this type of analysis.

Comparison of sensors and techniques for crop yield mapping

Abstract The implementation of site-specific crop management is dependent on the variations in yield and yield potential within a field. Crop yield maps are important for both the implementation and evaluation of site-specific crop management strategies. Management decisions and evaluations based on yield maps must take into consideration the accuracy and resolution of the maps. An impact-based yield monitor and a volumetric yield monitor were compared. The effect of different dynamic models of combine grain flow on the calculated instantaneous yields were investigated. Both simple time delay models and first order models could be used to model the grain flow. In general, a simple time delay model with minimal smoothing provided the best yield maps. Yield maps developed using different methods of Kriging and other mapping techniques were compared. The maps showed the same general trends. However, localized yield features were represented differently due to the methods used for developing the maps and the degree of smoothing.

Geographic Operating Range Evaluation of a NIR Soil Sensor

A previously developed prototype soil sensor which used near infrared (NIR) reflectance techniques was tested in the laboratory. The test set of soils included 30 Illinois soils and another 33 soils collected from across the continental U.S., all prepared at soil moisture tensions ranging from 1.5 MPa (wilting point) to 0.033 MPa (field capacity). As the geographic range represented by the included samples was increased, the predictions of organic carbon, cation exchange capacity, and soil moisture became less accurate. Calibrations obtained with soils from several states of the lower Corn Belt (Illinois, Missouri, Indiana, and Ohio) were only slightly less predictive of soil organic carbon than calibrations developed for the Illinois soils alone. Extension of the geographic range westward or northward from this area reduced the organic carbon prediction accuracy to an unacceptable level.