characterizing soil spatial variability with apparent soil ...€¦ · characterizing soil spatial...

Computers and Electronics in Agriculture 46 (2005) 103–133

Characterizing soil spatial variability with apparentsoil electrical conductivity

I. Survey protocols

D.L. Corwin∗, S.M. LeschUSDA-ARS, George E. Brown Jr. Salinity Laboratory, 450 West Big Springs Road,

Riverside, CA 92507-4617, USA

Abstract

Spatial characterization of the variability of soil physico-chemical properties is a fundamentalelement of (i) soil quality assessment, (ii) modeling non-point source pollutants in soil, and (iii) site-specific crop management. Apparent soil electrical conductivity (ECa) is a quick, reliable measurementthat is frequently used for the spatio-temporal characterization of edaphic (e.g., salinity, water content,texture, and bulk density) and anthropogenic (e.g., leaching fraction) properties. It is the objectiveof this paper to provide the protocols for conducting a field-scale ECa survey (Part I) and applythese protocols to a soil quality assessment in central California’s San Joaquin Valley (Part II). Theprotocols are comprised of eight general steps: (i) site description and ECa survey design; (ii) ECadata collection with mobile GPS-based equipment; (iii) soil sampling design; (iv) soil core sampling;(v) laboratory analysis; (vi) calibration of ECa to ECe; (vii) spatial statistical analysis; (viii) GISdatabase development and graphic display. For each outlined step, detailed discussion and guidelines

Abbreviations: CEC, cation exchange capacity; DPPC, dual pathway parallel conductance model; EC25◦C,electrical conductivity at 25◦ C; ECa, apparent soil electrical conductivity; ECe, electrical conductivity of thesaturation extract; ECw, electrical conductivity of the soil solution; EM, electromagnetic induction; EMh, elec-tromagnetic induction measurement in the horizontal coil-mode configuration; EMv, electromagnetic inductionmeasurement in the vertical coil-mode configuration; ER, electrical resistivity; ESAP, ECe sampling, assessmentand prediction software; ESP, exchangeable sodium percentage; GIS, geographic information system; GPS, globalpositioning system; IDW, inverse distance weighting;Ks, saturated hydraulic conductivity; LF, leaching fraction;NPS, non-point source; OM, organic matter; PW, percent water on a gravimetric basis; SAR, sodium adsorptionratio; SLRspatial linear regression; SP, saturation percentage; SSMU, site-specific management unit; TDR, timedomain reflectometry; USDA-ARS, United States Department of Agriculture-Agricultural Research Service

∗ Corresponding author. Tel.: +1 951 369 4819; fax: +1 951 342 4962.E-mail address:[email protected] (D.L. Corwin).

0168-1699/$ – see front matter. Published by Elsevier B.V.doi:10.1016/j.compag.2004.11.002

104 D.L. Corwin, S.M. Lesch / Computers and Electronics in Agriculture 46 (2005) 103–133

were presented. The developed protocols provide the guidelines to assure reliability, consistency, andcompatibility of ECa survey measurements and their interpretation.Published by Elsevier B.V.

Keywords:Precision agriculture; Salinity; ECa; Site-specific management units; Spatial heterogeneity; Soil quality

1. Introduction

The heterogeneity of soil physico-chemical properties has been known since the classicstudy of Nielsen et al. (1973), which characterized the spatial variability of soil-waterproperties for a 150 ha field at the University of California’s West Side Field Station in theSan Joaquin Valley. The characterization of soil spatial variability is fundamental to theunderstanding of landscape-scale processes of soils. Delineation of the spatial variation ofsoil properties is a crucial element of (i) non-point source (NPS) pollutant transport in thevadose zone, (ii) soil quality assessment, and (iii) site-specific crop management.

The spatial measurement of apparent soil electrical conductivity (ECa) is one meansof delineating spatial variation. Because of reliability, ease of measurement, and ability todetect a variety of soil properties, spatial ECa measurements have become a common toolused for field and landscape-scale studies related to edaphic properties. Spatial surveys ofECa have become widely used by a variety of scientists to spatially characterize soil salinityand nutrients (e.g., NO3−), texture-related properties, bulk density related properties suchas compaction, organic matter (OM) related properties, and a variety of other soil properties(seeTable 1; Corwin and Lesch, 2005a).

Geo-referenced ECa measurements have been correlated to associate yield-monitoringdata with mixed results (Jaynes et al., 1993; Sudduth et al., 1995; Kitchen et al., 1999;Johnson et al., 2001; Corwin et al., 2003b). These mixed results are due to confoundingfactors that complicate the relationship between ECa measurements and variations in cropyield. As pointed out byCorwin and Lesch (2003), crop yield inconsistently correlates withECa due to (i) the influence of soil properties (e.g., salinity, water content, texture, etc.) thatare measured by ECa, but may or may not influence yield within a particular field, (ii) atemporal component of yield variability that is poorly captured by a state variable such asECa, and (iii) confounding climatic factors.

Nevertheless, in instances where yield correlates with ECa, maps of ECa are useful fordevising soil sampling schemes to identify soil properties influencing yield within a field(Corwin et al., 2003b). When used as a means of directing soil sampling design, geo-referenced measurements of ECa have been shown by investigators to be a reliable, rapidmeans of establishing the spatial variability of soil physico-chemical properties associatedwith the leaching of NPS pollutants (Corwin et al., 1999), soil quality (Johnson et al., 2001;Corwin et al., 2003a), and variations in crop yield (Corwin et al., 2003b).

Because previous studies have varied in their approach of obtaining and interpretingspatial ECa measurements, a recent USDA-ARS workshop on precision agriculture (KansasCity, MO, 25–27 March 2003) concluded that protocols for conducting geo-referencedfield-scale ECa surveys and guidelines for interpreting the ECa measurements are neededto assure reliability, consistency, and compatibility of data. It is the objective of this paper

D.L. Corwin, S.M. Lesch / Computers and Electronics in Agriculture 46 (2005) 103–133 105

Table 1Compilation of literature measuring ECa with geophysical techniques (ER or EM) that have been categorizedaccording to soil-related properties that were either directly or indirectly measured by ECa

Soil property References

Directly measured soil propertiesSalinity (and nutrients, e.g. NO3−) Halvorson and Rhoades (1976); Rhoades et al. (1976); Rhoades

and Halvorson (1977); de Jong et al. (1979); Cameron et al.(1981); Rhoades and Corwin (1981, 1990); Corwin and Rhoades(1982, 1984); Williams and Baker (1982); Greenhouse and Slaine(1983); van der Lelij (1983); Wollenhaupt et al. (1986); Williamsand Hoey (1987); Corwin and Rhoades (1990); Rhoades etal. (1989, 1990, 1999a, 1999b); Slavich and Petterson (1990);Diaz and Herrero (1992); Hendrickx et al. (1992); Lesch et al.(1992, 1995a, 1995b,1998); Rhoades (1992, 1993); Cannon etal. (1994); Nettleton et al. (1994); Bennett and George (1995);Drommerhausen et al. (1995); Ranjan et al. (1995); Hansonand Kaita (1997); Johnston et al. (1997); Mankin et al. (1997);Eigenberg et al. (1998, 2002); Eigenberg and Nienaber (1998,1999, 2001); Mankin and Karthikeyan (2002); Herrero et al.(2003); Paine (2003); Kaffka et al. (2005)

Water content Fitterman and Stewart (1986); Kean et al. (1987); Kachanoski et al.(1988); Kachanoski (1990); Vaughan et al. (1995); Sheets andHendrickx (1995); Hanson and Kaita (1997); Khakural et al.(1998); Morgan et al. (2000); Freeland et al. (2001); Brevik andFenton (2002); Wilson et al. (2002); Kaffka et al. (2005)

Texture-related (e.g., sand, clay, depthto claypans or sand layers)

Williams and Hoey (1987); Brus et al. (1992); Jaynes et al. (1993);Stroh et al. (1993); Sudduth and Kitchen (1993); Doolittle et al.(1994, 2002); Kitchen et al. (1996); Banton et al. (1997); Boettingeret al. (1997); Rhoades et al. (1999b); Scanlon et al. (1999); Inmanet al. (2001); Triantafilis et al. (2001); Anderson-Cook et al. (2002);Brevik and Fenton (2002)

Bulk density related (e.g., compaction) Rhoades et al. (1999b); Gorucu et al. (2001)

Indirectly measured soil propertiesOrganic matter related (including soil

organic carbon, and organic chemicalplumes)

Greenhouse and Slaine (1983, 1986); Brune and Doolittle (1990);Nyquist and Blair (1991); Jaynes (1996); Benson et al. (1997);Bowling et al. (1997); Brune et al. (1999); Nobes et al. (2000)

Cation exchange capacity McBride et al., 1990; Triantafilis et al. (2002)Leaching Slavich and Yang (1990); Corwin et al. (1999); Rhoades et al.

(1999b)Groundwater recharge Cook and Kilty (1992), Cook et al. (1992); Salama et al. (1994)Herbicide partition coefficients Jaynes et al. (1995)Soil map unit boundaries Fenton and Lauterbach (1999); Stroh et al. (2001)Corn rootworm distributions Ellsbury et al. (1999)Soil drainage classes Kravchenko et al. (2002)

Taken fromCorwin and Lesch (2005a).

to (i) describe the GPS-based equipment used to conduct an ECa survey and (ii) outline adetailed set of protocols for conducting a field-scale ECa survey that is used to direct a soilsampling design to characterize soil spatial variability for use in soil quality assessment andsite-specific crop management.


2. Mobilized ECa-measurement equipment

A detailed description of the theory, operation, and construction of electrical resistiv-ity (ER) and electromagnetic induction (EM) instrumentation is provided byRhoades etal. (1999b)andHendrickx et al. (2002a). Mobilized ECa-measurement equipment, ER orEM instrumentation, has been in use for over a decade for the purpose of mapping andmonitoring field-scale spatial soil salinity patterns (Rhoades, 1992, 1993). More recentlythese mobilized systems have been used to spatially characterize soil condition by map-ping a variety of physico-chemical properties including salinity, water content, texture, bulkdensity (Johnson et al., 2001; Corwin et al., 2003a). The design of a mobilized ECa mea-surement system consists of four basic components: (i) ECa measurement sensor, (ii) globalpositioning system (GPS), (iii) hardware interfacing, and (iv) transport platform.

Three types of ECa measurement sensors are available: (i) invasive four-electrode ERsensors, (ii) noninvasive EM sensors, and (iii) time domain reflectometry (TDR) sensors.Invasive ER and noninvasive EM sensors are the most popular sensors because the com-mercial development of a TDR sensor for use on a mobile apparatus has not yet occurred.Invasive four-electrode sensors can take the form of either insertion probes or surface arrayswith the latter being the configuration used for mobilized ECa measurement systems. Ex-amples of invasive ER four-electrode sensors configured as fixed-surface arrays include theequipment developed byRhoades (1992, 1993)andCarter et al. (1993)and the commercialsensor technology used in the Veris 3100 system1 (Veris Technologies, Salina, KS). Com-mercial examples of EM sensors include the Geonics EM-31 and EM-38 soil conductivitymeters (see footnote 1) (Geonics Ltd., Mississauga, Ont., Canada), both of which can beeasily mobilized, but the EM-38 has been the primary instrument of choice for soil qual-ity and site-specific crop management applications because its depth of penetration mostclosely corresponds to the root zone (i.e., 0 to 1–1.5 m).

There are two basic GPS systems that can be used in mobile ECa-measurement equip-ment: (i) self-contained systems and (ii) stand-alone GPS receivers that require external datalogging. The difference between the two is not in the GPS receiver technology, but in theinterfacing. Self-contained GPS systems include data loggers and software programs thatallow the user to record, modify, and/or store GPS coordinate data independent of attachedsensors or hardware interfacing. Stand-alone GPS receivers typically must be connected toa microprocessor or electronic controller in order to store and/or process GPS coordinatedata. The Trimble Pathfinder Pro-XRS and Trimble Ag132 GPS systems are examples ofavailable commercial GPS systems.

Hardware interfacing is needed to link the ECa measurement sensor data with associatedGPS coordinate data and control the timing of the data acquisition. The complexity of thehardware interface depends on the type and number of sensors and the extent of real-time dataprocessing; consequently, the hardware interface tends to be system specific and expensive.However, in a simple mobile ECa measurement system with a single electrical conductivitymeter (i.e., single EM-38) the hardware interface can be omitted by direct output of real-

1 Mention of trademark or proprietary products does not constitute an endorsement or guarantee/warranty of theproduct by the U.S. Department of Agriculture and does not imply its approval to the exclusion of other productsthat may also be available.


time sensor data through an RS-232 serial connection with the data capture capability of theGPS system. The Veris system comes complete with interfacing and recording hardware,only requiring the user to plug in a compatible GPS receiver.

The final component of the mobile ECa measurement system is the transport platform,which consists of either tow-able or self-mobilized platforms. Pickups, all-terrain vehicles(ATVs), and tractors have been used to tow ECa-measurement sensors. The fixed-arrayfour electrode (Rhoades, 1992, 1993) and Veris 3100 (Lund et al., 1999; Sudduth et al.,1999) are examples of ER sensor platforms that are towed. Simple non-metallic platformshave also been developed to tow EM instrumentation (Jaynes et al., 1993; Cannon et al.,1994; Kitchen et al., 1996; Freeland et al., 2002). An example of a self-mobilized platformincludes a modified hydraulic-driven spray tractor with insertion-type, four-electrode ECasensors located on the rig’s undercarriage that are driven into the ground with the hydraulicsystem, and a non-metallic cylinder located at the front end that houses an EM-38, which israised, lowered, and rotated 90E with the hydraulic system (Rhoades, 1992, 1993; Carter etal., 1993). In general, motorized platforms are more sophisticated, versatile, and expensiveto develop than tow-able platforms.

System integration can be dedicated or autonomous. In a dedicated system all ECameasurement sensors and GPS equipments are integrated directly into the transport platform.This system relies on extensive hardware interfacing and a central computer or controllerto manage the data acquisition and data storage. In an autonomous system the GPS receiverand each ECa measurement sensor can be easily removed from the platform and usedindependently. The Trimble Pathfinder Pro-XRS is an example of an autonomous system,whereas the Trimble Ag132 GPS is a dedicated system.

Surveys of ECa to characterize soil spatial variability are used as spatial informationto direct a soil sampling scheme that will provide the necessary ground-truth informationto establish the spatial distribution of those soil properties correlated with ECa within afield. For this reason, the inclusion of soil sampling equipment directly onto the platformis advantageous. The addition of soil sampling equipment allows the platform to serve asboth an ECa survey system and a soil sampling rig, which increases the versatility of thesystem.Figs. 1 and 2illustrate the components of a GPS-based mobile ECa measurementsystem developed at the George E. Brown Jr. Salinity Laboratory. The system consists ofa dual-dipole EM-38 that simultaneously measures vertical (EMv) and horizontal (EMh)electromagnetic induction ECa (seeFig. 1b), a Giddings soil core sampler (see footnote 1)mounted on the front of the rig (seeFig. 1c), and a Trimble Pro-XL GPS system (seeFig. 2).The Trimble Pro-XL GPS system consists of a MC-V datalogger, TANS receiver, batterypack, and dome antenna (seeFig. 2).

3. Protocols for conducting a field-scale ECa survey and soil sampling

Geo-referenced measurements of ECa are useful for establishing the spatial distributionof those soil properties influencing ECa within a field. In instances where ECa correlateswith a particular soil property, an ECa-directed soil sampling approach will establish thespatial distribution of that property with an optimum number of site locations to characterizethe variability and keep labor costs minimal (Corwin et al., 2003a). Also, if ECa is correlated


Fig. 1. Mobile EM equipment: (a) complete rig; (b) close-up of sled holding the Geonics dual-dipole EM-38 soilelectrical conductivity meter; (c) close-up of Giddings soil core sampler.

Fig. 2. Connection between (a) dual-dipole EM-38 meter and Trimble MC-V Pro-XL system consisting of (b)MC-V datalogger, (c) TANS receiver, (d) battery pack, and (e) dome antenna.


with crop yield, then an ECa-directed soil sampling approach can be used to identify thosesoil properties that are causing variability in crop yield (Corwin et al., 2003b). General ECasurvey guidelines can be derived fromCorwin and Lesch (2003)andCorwin et al. (2003a,2003b). Details for conducting a field-scale ECa survey for the purpose of characterizingthe spatial variability of soil properties influencing soil quality or crop yield variation areprovided herein.

The purpose of an ECa survey from a soil quality perspective is to establish the within-field variation of soil properties that influence the field’s intended use (e.g., agriculturalproductivity, environmental protection, waste recycling, etc.). The purpose of an ECa surveyfrom a site-specific crop management perspective is to establish the within-field variation ofsoil properties influencing the variation in crop yield. The basic elements of a field-scale ECasurvey applied to soil quality assessment and site-specific crop management include (i) sitedescription and ECa survey design, (ii) geo-referenced ECa data collection, (iii) soil sampledesign based on geo-referenced ECa data, (iv) soil sample collection, (v) physico-chemicalanalysis of pertinent soil properties, (vi) if soil salinity is a primary concern, developmentof a stochastic and/or deterministic calibration of ECa to soil salinity as determined by theelectrical conductivity of the saturation extract (ECe), (vii) spatial statistical analysis, and(viii) geographic information system (GIS) database development. The basic steps withineach component are outlined inTable 2.

3.1. ECa survey design and geo-referenced ECa data collection

An initial ECa survey with either mobile ER or EM equipment is conducted and used toestablish soil core sampling locations needed for calibration and/or characterization of thespatial distribution of soil properties correlated with ECa. Depending on the level of detaildesired, from 100 to several thousand spatial measurements of ECa are taken, generally inregularly spaced traverses across the field of interest. The use of mobile EM equipment hasthree advantages over the use of mobile ER equipment: (i) the ability to take measurementson dry and stony soils, (ii) the ability to traverse growing crops, and (iii) the ability to traversefields with beds and furrows. Under dry soil conditions the physical contact needed betweenER electrodes and soil for continuous electrical current flow is difficult to maintain. Stonysoils are damaging to the ER electrodes. Growing crops and bed–furrow systems posea problem for ER equipment because of contact problems with the invasive electrodes.The coulters or insertion probes of ER equipment are on a fixed-height or limited-heightadjustable platform that cannot clear most crops nor are they easily adjusted to conformto abrupt changes in microtopography as found in bed–furrow systems. In contrast, EMequipment is easily designed so the platform clears most crop canopies and the EM-38slides down furrows.

Prior to physically conducting the ECa survey, the survey’s objective(s) must be definedbased upon the project goals and available resources (e.g., manpower, funding, analyticalcapabilities, etc.). Pre-survey design tasks including site boundary definition, recording ofsite metadata, and selection of the GPS coordinate system and datum should be performed.The survey’s objectives and resources will determine the measurement intensity of thesurvey (i.e., spacing between ECa measurements). Measurement intensities generally varyfrom every 3 to 5 m for intense surveys used in detailed field-scale studies to 75–100 m


Table 2Outline of steps for an ECa field survey

1. Site description and ECa survey design(a) Record site metadata(b) Define the project’s/survey’s objective(c) Establish site boundaries(d) Select GPS coordinate system(e) Establish ECa measurement intensity

2. ECa data collection with mobile GPS-based equipment(a) Geo-reference site boundaries and significant physical geographic features with GPS(b) Measure geo-referenced ECa data at the pre-determined spatial intensity and record associated metadata

3. Soil sample design based on geo-referenced ECa data(a) Statistically analyze ECa data using an appropriate statistical sampling design to establish the soil samplesite locations(b) Establish site locations, depth of sampling, sample depth increments, and number of cores per site

4. Soil core sampling at specified sites designated by the sample design(a) Obtain measurements of soil temperature through the profile at selected sites(b) At randomly selected locations obtain duplicate soil cores within a 1 m distance of one another to establishlocal-scale variation of soil properties(c) Record soil core observations (e.g., mottling, horizonation, textural discontinuities, etc.)

5. Laboratory analysis of appropriate soil physico-chemical properties defined by project objectives

6. If needed, stochastic and/or deterministic calibration of ECa to ECe or to other soil properties (e.g., watercontent and texture)

7. Spatial statistical analysis to determine the soil properties influencing ECa and/or crop yield(a) Soil quality assessment

(1) Perform a basic statistical analysis of physico-chemical data by depth increment and by compositedepth over the depth of measurement of ECa

(2) Determine the correlation between ECa and physico-chemical soil properties by composite depth overthe depth of measurement of ECa

(b) Site-specific crop management (if ECa correlates with crop yield, then)(1) Perform a basic statistical analysis of physico-chemical data by depth increment and by composite

depths(2) Determine the correlation between ECa and physico-chemical soil properties by depth increment and

by composite depths(3) Determine the correlation between crop yield and physico-chemical soil properties by depth and by

composite depths to determine depth of concern (i.e., depth with consistently highest correlation, whetherpositive or negative, of soil properties to yield) and the significant soil properties influencing crop yield (orcrop quality)

(4) Conduct an exploratory graphical analysis to determine the relationship between the significant physico-chemical properties and crop yield (or crop quality)

(5) Formulate a spatial linear regression (SLR) model that relates soil properties (independent variables)to crop yield or crop quality (dependent variable)

(6) Adjust this model for spatial auto-correlation, if necessary, using restricted maximum likelihood orsome other technique

(7) Conduct a sensitivity analysis to establish dominant soil property influencing yield or quality(8) GIS database development and graphic display of spatial distribution of soil properties


for basin-scale studies of thousands of hectares. Typically, an 18 ha field can be surveyedwith mobile ECa equipment in one to two 8 h work days at a 5 m spacing, which results inroughly 7200 ECa measurements. This level of survey intensity provides a map of spatialvariation sufficiently detailed to meet nearly any intended purpose.

3.2. Soil sample design based on geo-referenced ECa data

Once a geo-referenced ECa survey is conducted, the data are used to establish the lo-cations of the soil core sample sites for (i) calibration of ECa to soil sample ECe and/or(ii) delineation of the spatial distribution of soil properties correlated to ECa within thefield surveyed. To establish the locations where soil cores are to be taken either design-based or model-based sampling schemes can be used. Design-based sampling schemeshave historically been the most commonly used and hence are more familiar to most re-search scientists. An excellent review of design-based methods can be found inThompson(1992). Design-based methods include simple random sampling, stratified random sam-pling, multistage sampling, cluster sampling, and network sampling schemes. The use ofunsupervised classification byFraisse et al. (2001)andJohnson et al. (2001)is an exampleof design-based sampling. Model-based sampling schemes are far less common, althoughsome statistical research has been performed in this area (Royall, 1988). Specific model-based sampling approaches that have direct application to agricultural and environmentalsurvey work are described byMcBratney and Webster (1983), Russo (1984)andLesch et al.(1995b).

The sampling approach introduced byLesch et al. (1995b)is specifically designed foruse with ground-based soil ECa data. This sampling approach attempts to optimize theestimation of a regression model (i.e., minimize the mean square prediction error producedby the calibration function), while simultaneously insuring that the independent regres-sion model residual error assumption remains approximately valid. This in turn allows anordinary regression model to be used to predict soil property levels at all remaining (i.e.,non-sampled) conductivity survey sites. The basis for this sampling approach stems directlyfrom traditional response-surface sampling methodology (Box and Draper, 1987).

There are two main advantages to the response-surface approach. First, a substantialreduction in the number of samples required for effectively estimating a calibration functioncan be achieved, in comparison to more traditional design-based sampling schemes. Second,this approach lends itself naturally to the analysis of remotely sensed ECadata. Indeed, manytypes of ground, airborne, and/or satellite-based remotely sensed data are often collectedspecifically because one expects this data to correlate strongly with some parameter ofinterest (e.g., crop stress, soil type, soil salinity, etc.), but the exact parameter estimates(associated with the calibration model) may still need to be determined via some type ofsite-specific sampling design. The response-surface approach explicitly optimizes this siteselection process.

A user-friendly software package (ESAP) developed byLesch et al. (2000), which usesa response-surface sampling design, has proven to be particularly effective in delineatingspatial distributions of soil properties from ECa survey data (Corwin and Lesch, 2003;Corwin et al., 2003a, 2003b). The ESAP software package identifies the optimal locationsfor soil sample sites from the ECa survey data. These sites are selected based on spatial


statistics to reflect the observed spatial variability in ECa survey measurements. Generally,6–20 sites are selected depending on the level of variability of the ECa measurements for asite. The optimal locations of a minimal subset of ECa survey sites are identified to obtainsoil samples.

Once the number and location of the sample sites have been established, the depth ofsoil core sampling, sample depth increments, and number of sites where duplicate or repli-cate core samples should be taken are established. The depth of sampling should be thesame at each sample site and should extend over the depth of penetration by the ECa-measurement equipment used. For instance, the Geonics EM-38 measures to a depth ofroughly 0.75–1.0 m in the horizontal coil configuration (EMh) and 1.2–1.5 m in the verti-cal coil configuration (EMv). Sample depth increments are flexible and depend to a greatextent on the study objectives. A depth increment of 0.3 m has been commonly used atthe USDA-ARS Salinity Laboratory because it provides sufficient soil profile informationover the root zone (i.e., 0–1.2 to 1.5 m) for statistical analysis without an overly burden-some number of samples to conduct physico-chemical analyses. Depth increments shouldbe the same from one sample site to the next. The number of duplicates or replicatestaken at each sample site are determined by the desired accuracy for characterizing soilproperties and the need for establishing the level of local-scale variability at the site. Du-plicates or replicates are not necessarily needed at every sample site to establish local-scalevariability.

3.3. Soil core sampling

General soil core sampling protocols (Peterson and Calvin, 1996) and associated qualitycontrol and quality assurance procedures (Klestra and Bartz, 1996) should be followed. Soilcores are acquired to the same depth and over the same depth increments at all the selectedsites. Soil cores are taken directly over the location of the ECa measurement. As the coresare taken, soil temperature through the profile can be measured at selected sites, if needed.During sampling, attention should be taken to avoid including any dry, loose soil that maybe present at the surface. Dry, loose topsoil is not reflected in the ECa measurement becauseit has only residual moisture content making it non-conductive.

The extent of the spatio-temporal variation of soil temperature determines the overallsignificance of soil temperature as a factor influencing ECa measurements. Customarily,electrical conductivity is expressed at a reference temperature of 25◦C. Electrolytic con-ductivity increases at a rate of approximately 1.9% per◦C increase in temperature. Forcomparison purposes, if the spatio-temporal variations in soil temperature are greater than10◦C, then temperature data is needed to adjust the electrical conductivity to the referencetemperature of 25◦C. The following equation expresses electrical conductivity at a standardreference temperature of 25◦C:

EC25◦C = fT ECT (1)

where EC25◦C is the electrical conductivity at the reference temperature of 25◦C, fT thetemperature conversion factor, and ECT the electrical conductivity at temperatureT (◦C). Anapproximation equation for the temperature conversion factor has been derived bySheets


and Hendrickx (1995):

fT = 0.4470+ 1.4034 exp

(− T

26.815

)(2)

Duplicate or replicate samples are collected at a minimum of 4–6 of the sample sites within astudy area. The primary and replicate soil core samples are taken within a 0.5–1.0 m radius.If the field consists of beds and furrows, all primary and replicate samples are acquired fromthe same respective locations of the bed–furrow.

An appropriate identification system for labeling the soil samples should be established.Any metadata associated with each sample site and the samples such as observations aboutthe depth to water table, abrupt changes in soil texture, horizonation, mottling, color change,surface crusting, etc. should be recorded. All soil samples should be placed in sealed,air-tight containers to minimize the loss of moisture, which would affect water contentmeasurements. The soil samples should be immediately placed in an insulated storagecontainer (refrigerated, if possible) for protection and keep the samples cool.

3.4. Laboratory analysis of soil physico-chemical properties

General quality control and quality assurance protocols for laboratory analysis should befollowed (Klestra and Bartz, 1996). Soil samples should be analyzed as soon as possible aftertheir collection. All soil sample preparation and analyses should be conducted followingscientifically accepted procedures such as those outlined in the Soil Science Society ofAmerica’s Methods of Soil Analysis (Sparks, 1996; Dane and Topp, 2002) and the UnitedStates Department of Agriculture’s Handbook 60 (U.S. Salinity Laboratory Staff, 1954).

The appropriate analyses will depend upon the objective of the study. There is generalagreement by soil scientists upon recommended minimum data sets of soil parameters thatshould be used to quantify soil quality including biological (microbial biomass, potentiallymineralizable N, and soil respiration), chemical (pH, ECe, OM, N, P, and K) and physical(texture, bulk density, depth of rooting, infiltration, and water holding capacity) parameters(Bouma, 1989; Larson and Pierce, 1991, 1994; Arshad and Coen, 1992; Doran and Parkin,1994, 1996). However, soil quality is a function of the intended use of the soil; conse-quently, the appropriate soil properties to assess quality will vary accordingly.Corwin et al.(2003a)found the following properties appropriate for assessing quality of a salt-affected,sodic soil used for agricultural production: ECe; pH; anions (HCO3−, Cl−, NO3

−, SO42−)

and cations (Na+, K+, Ca2+, Mg2+) in the saturation extract; trace elements (B, Se, As,Mo) in the saturation extract; lime (CaCO3); gypsum (CaSO4); cation exchange capacity(CEC); exchangeable Na+, K+, Mg2+, and Ca2+; ESP (exchangeable sodium percentage);SAR; saturated hydraulic conductivity (Ks); and leaching fraction (LF). Site-specific cropmanagement applications generally require a knowledge of those soil physico-chemicalproperties influencing the yield of a specific crop and impacting the environment. For in-stance, for an irrigated, arid zone soil on the San Joaquin Valley’s west side,Corwin et al.(2003b)found pH, B, NO3 and N, Cl−, ECe, LF, gravimetric water content,ρb, % clay,and saturation percentage (SP) to be the most significant soil properties when consideringedaphic influences on within-field variations in cotton production.


3.5. Stochastic and/or deterministic calibration of ECa to ECe or to other soilproperties

Apparent soil electrical conductivity can be calibrated to any soil property that signifi-cantly influences the ECa measurement such as salinity, water content, clay content, SP, bulkdensity (ρb), and OM. As previously mentioned, there are numerous studies that documentthe relationships between soil electrical conductivity and various soil physical and chemicalproperties, including soil salinity (Rhoades, 1992, 1996; Rhoades et al., 1989; Lesch et al.,1995a; Williams and Baker, 1982), clay content (Williams and Hoey, 1987), depth to claylayers (Doolittle et al., 1994), nutrient status (Sudduth et al., 1995), and moisture content(Kachanoski et al., 1988), just to list a few. Additionally, there are articles documentingthe use of conductivity survey information to determine salt loading and field irrigationefficiency (Rhoades et al., 1997; Corwin et al., 1999) and for estimating deep drainage(Triantafilis et al., 2003). All the data analysis and interpretation presented in these paperscan be classified into two data modeling categories: deterministic and stochastic.

In general, stochastic models are based on some form of objective sampling methodologyused in conjunction with various statistical calibration techniques. The most common typesof calibration equations are geostatistical models (generalized universal kriging models andcokriging models) and spatially referenced regression models.

Traditionally, universal kriging models have been viewed as an extension of the ordinarykriging technique and used primarily to account for large-scale (non-stationary) trends inspatial data. However, this modeling technique can be easily generalized to model ancillarysurvey data (such as EM-38 data) when this data correlates well with some spatially vary-ing soil property of interest (e.g., soil salinity). This generalization is commonly referredto as a “spatial linear model” or “spatial random field model” in the statistical literature(Christensen et al., 1992). This modeling approach requires the estimation of a regressionequation with a spatially correlated error structure. This type of model probably representsthe most versatile and accurate statistical calibration approach, provided enough calibrationsample sites are collected (n≥ 50) to ensure a good estimate of the correlated error structure.

Regardless of their versatility, spatial linear models are typically used in regional situa-tions. Such an approach is rarely used for field-scale survey work, due to the large numberof required calibration soil samples, which makes this approach economically impractical.Instead, most calibration equations of soil properties are spatially referenced regressionmodels. A spatially referenced regression model is just an ordinary regression equationthat includes the soil property being calibrated with ECa and trend surface parameters. Themodel assumes an independent error structure that can usually be achieved through care-fully designed sampling plans, such as the response-surface sampling design. In practice,these are the only models that can be reasonably estimated with a limited number of soilsamples (n< 15).

Deterministic conductivity data modeling and interpretation can be carried out eitherfrom a geophysical or a soil science approach. In the geophysical approach, mathemati-cally sophisticated inversion algorithms are generally employed. These approaches, whichrely heavily on geophysical theory, have met with limited success for the interpretation ofnear-surface ECa data. Part of the reason for the lack of success is that most geophysicalinversion approaches assume that (i) there are multiple conductivity signal readings avail-


able for each survey point and (ii) that distinct, physical strata differences exist within thenear-surface soil horizon. Neither of these conditions are typically satisfied in most ECasurveys.

A more common interpretation technique used, particularly in salinity inventorying work,is to employ some form of deterministic ECa-to-salinity model (i.e., an equation whichconverts ECa to ECe based on knowledge of other soil properties). One model of this typethat has been shown to be useful is the DPPC (dual pathway parallel conductance) modeldeveloped byRhoades et al. (1989, 1990)and extended byLesch and Corwin (2003).This model is based on the idea that electrical conductivity of soil can be modeled as amulti-pathway parallel electrical conductance equation. This model has been shown to beapplicable to a wide range of typical agricultural situations (Corwin and Lesch, 2003). TheDPPC model demonstrates that soil electrical conductivity can be reduced to a nonlinearfunction of five soil physico-chemical properties: ECe, SP, volumetric soil water content,�b, and soil temperature. InRhoades et al. (1990)the DPPC model was used to estimatefield soil salinity levels based on ECa survey data and measured or inferred informationabout the remaining soil physical properties.Corwin and Lesch (2003)andLesch et al.(2000)showed that this model can also be used to assess the degree of influence that eachof these soil properties has on the acquired ECa-survey data.

Soil salinity, as conventionally expressed in terms of the electrical conductivity of thesaturated-paste extract, ECe, can be determined from ECa in two ways: a deterministic anda stochastic approach (Rhoades et al., 1999b). The preferred approach will vary with thesize of the area to be assessed, availability of equipment, and the specific objectives. Inthe deterministic approach, either theoretically or empirically determined models convertECa into ECe. Deterministic models are “static” (i.e., all model parameters are consideredknown and no ECe data needs to be determined). For example, Eq.(3) from the DPPCmodel ofRhoades et al. (1989)is a deterministic approach:

ECa =(

(θss+ θws)2 · ECws · ECss

(θss · ECws) + (θws · ECs)

)+ (θw − θws) · ECwc (3)

whereθw = θws + θwc = total volumetric water content (cm3 cm−3); θws andθwc are the volu-metric soil water content in the soil-water pathway (cm3 cm−3) and in the continuous-liquidpathway (cm3 cm−3), respectively;θss the volumetric content of the surface-conductance(cm3 cm−3); ECws and ECwc the specific electrical conductivities of the soil-water pathway(dS m−1) and continuous-liquid pathway (dS m−1); and ECss the electrical conductivity ofthe surface-conductance (dS m−1). Soil ECa is converted into estimated soil salinity (i.e.,ECe) using Eqs.(3)–(8)originally developed byRhoades et al. (1989):

θw = (PW · ρb)

100(4)

θws = 0.639θw + 0.011 (5)

θss = ρb

2.65(6)


ECss = 0.019(SP)− 0.434 (7)

ECw =[

ECe · ρb · SP

100· θb

](8)

where PW is the percent water on a gravimetric basis,ρb the bulk density (Mg m−3), SP thesaturation percentage, ECw the average electrical conductivity of the soil water assumingequilibrium (i.e., ECw = ECsw = ECwc), and ECe the electrical conductivity of the saturationextract (dS m−1).

The deterministic approach is the preferred approach when significant, localized varia-tions in soil type exist in the field. However, this approach typically requires knowledge ofadditional soil properties (e.g., soil water content, SP,ρb, temperature, etc.). In instanceswhere extensive soil property information is lacking, the stochastic approach is more appro-priate. In the stochastic approach, statistical modeling techniques such as spatial regressionor co-kriging are used to directly predict the soil salinity from ECa survey data. In this ap-proach, the models are “dynamic” (i.e., the model parameters are estimated using soil sampledata collected during the survey). The calibration is developed by acquiring soil salinitydata (or other soil property data such as SP, texture,ρb, etc.) from a small percentage ofthe ECa measurement sites and estimating an appropriate stochastic-prediction model foreach depth increment using the paired soil sample and ECa data. Using the remaining ECadata in conjunction with the established model, the soil salinity levels (or other calibratedproperties) are predicted at all of the remaining non-sampled, measurement locations. Thestochastic- and deterministic calibration approaches are described in detail byLesch et al.(1995a, 1995b, 2000)and incorporated into the ESAP software (Lesch et al., 2000).

3.6. Spatial statistical analysis to determine the soil properties influencing ECa and/orcrop yield

In the past, the fact that ECa is a function of several soil properties (i.e., soil salinity,texture, and water content) has sometimes been overlooked in the application of ECa mea-surements to site-specific crop management. In areas of saline soils, salinity dominates theECa measurements and interpretations are often straightforward. However, in areas otherthan arid zone soils, texture and water content or even OM may be the dominant propertiesmeasured by ECa. To use spatial measurements of ECa in a soil quality or site-specificcrop management context, it is necessary to understand what factors are most significantlyinfluencing the ECa measurements within the field of study. There are two commonly usedapproaches for determining the predominant factors influencing ECa measurement: (1)wavelet analysis and (2) simple statistical correlation.

An explanation of the use of wavelet analysis for determining the soil properties influ-encing ECa measurements is provided byLark et al. (2003). Even though wavelet analysisis a powerful tool for determining the dominant complex interrelated factors influencingECa measurement, it requires soil sample data collected on a regular grid or equal-spacedtransect. Grid or equal-spaced transect sampling schemes are not as practical for deter-mining spatial distributions of soil salinity (or some other correlated soil property) from


ECa measurements as the statistical and graphical approach first developed byLesch et al.(1995a, 1995b, 2000).

The most practical means of interpreting and understanding the tremendous volume ofspatial data from an ECa survey is through statistical analysis and graphic display. De-tails describing the statistical and graphical approach for determining the predominant soilproperty influencing ECa measurements are found inCorwin and Lesch (2003). For a soilquality assessment a basic statistical analysis of all physico-chemical data by depth in-crement provides an understanding of the vertical profile distribution. A basic statisticalanalysis consists of the determination of the mean, minimum, maximum, range, standarddeviation, standard error, coefficient of variation, and skewness for each depth increment(e.g., 0–0.3, 0.3–0.6, 0.6–0.9, and 0.9–1.2 m) and by composite depth (e.g., 0–1.2 m) overthe depth of measurement of ECa. In the case of ECa measured with ER (e.g., Veris orfixed-array four electrode equipment), the composite depth over the depth of measurementof ECa is based on the spacing between the electrodes, while in the case of EM-38 mea-surements of ECa the composite depth for the EMh measurement is about 0–0.75 to 1.0 mand 0–1.2 to 1.5 m for EMv. The calculation of the correlation coefficient between ECaand mean value of each physico-chemical soil property by depth increment and compositedepth over multiple sample sites determines those soil properties that correlate best withECa and those soil properties that are spatially represented by the ECa-directed samplingdesign. Those properties that are not correlated with ECa are not spatially characterizedwith the ECa-directed sampling design indicating that a design-based sampling schemesuch as stratified random sampling is probably needed to better spatially characterize thesesoil properties.

Crop yield monitoring data in conjunction with ECa survey data can be used from a site-specific crop management perspective to (i) identify those soil properties influencing yieldand (ii) delineate site-specific management units (SSMU). For site-specific crop manage-ment, an understanding of the influence of spatial variation in soil properties on within-fieldcrop-yield (or crop quality) variation is desired. To accomplish this using ECa, crop yield(or crop quality)mustcorrelate with ECa within a field. If crop yield (or crop quality) andECa are correlated, then basic statistical analyses by depth increment (e.g., 0–0.3, 0.3–0.6,0.6–0.9, and 0.9–1.2 m) and by composite depths (e.g., 0–0.3, 0–0.6, 0–0.9, and 0–1.2 m)are performed. As before, the correlation between ECa and mean values of each physico-chemical soil property for each depth increment and each composite depth establishes thosesoil properties that are spatially characterized with the ECa-directed sampling design. Thecorrelations between crop yield (or crop quality) and physico-chemical soil properties willalso establish the depth of concern (i.e., the root zone of the crop), which will be the com-posite depth that consistently has the highest correlation of each soil property (i.e., each soilproperty determined to be significant to influencing yield) with crop yield (or crop quality).Exploratory graphical analyses (i.e., scatter plots of crop yield or crop quality and each soilproperty) are then conducted for the depth of concern to determine the linear or curvi-linearrelationship between the significant physico-chemical properties and crop yield (or cropquality). A spatial linear regression (SLR) is formulated that relates the significant soilproperties as the independent variables to crop yield (or crop quality) as the dependent vari-able. The functional form of the model is developed from the exploratory graphic analysis.The model is adjusted for spatial auto-correlation, if necessary, using restricted maximum


likelihood or some other technique. This entire spatial statistical analysis process is clearlydemonstrated byCorwin et al. (2003b)andCorwin and Lesch (2005b).

To use spatial measurements of ECa in a site-specific crop management context, it is notonly necessary to understand what factors most significantly influence ECa measurementswithin the field of study, but also know those factors that most significantly influence within-field variation in crop yield (or crop quality).Corwin et al. (2003b)used sensitivity analysissimulations to arrive at the dominant edaphic and anthropogenic factors influencing within-field cotton yield variations. Sensitivity analysis involves increasing a single independentvariable (i.e., edaphic factors) and observing the resultant effect on the dependent variable(i.e., crop yield or crop quality). This is done for each independent variable. The relativeeffect of each independent variable on the dependent variable determines the independentvariable that most significantly influences the dependent variable.

3.7. GIS development and graphic display of spatial distribution of soil properties

The organization, manipulation, and graphic display of spatial soil and ECa data is bestaccomplished with a geographic information system (GIS). Spatial soil property data areentered into any of the several GIS software packages such as ArcView or ArcGIS (seefootnote 1). Once the spatial data is entered, maps of the soil physico-chemical propertiescan be easily prepared. A variety of interpolation techniques such as inverse-distance-weighting (IDW) interpolation and various kriging approaches can be used. Previous studiescomparing interpolation methods for mapping soil properties have found mixed results. Insome instances kriging has been found to perform the best (Laslett et al., 1987; Warricket al., 1988; Leenaers et al., 1990; Kravchenko and Bullock, 1999) and in others IDW hasbeen found superior (Weber and Englund, 1992; Wollenhaupt et al., 1994; Gotway et al.,1996). A common means of determining which method is the best to use for a particularspatial data set is to use the statistical approach of jackknifing to establish the interpolationmethod that minimizes the prediction error (Isaaks and Srivastava, 1989).

4. Additional considerations

There are a number of additional issues to consider in an ECa survey that may havesubtle effects on the reliability and accuracy of the ECa measurements. These effects relateto factors that are easily overlooked, but may collectively mean the difference betweenuseful and unreliable data.

4.1. Accuracy issues concerning EM measurements

The issue of accuracy in EM measurements of ECa for precision agriculture is addressedbySudduth et al. (2001). The authors point out that over time the EM-38 sensor is subject todrift, which can contribute a significant fraction of the within-field ECa variation. A studyby Robinson et al. (2004)indicated that the drift observed in the EM-38 is likely due totemperature effects on the EM-38 sensor and that a simple reflective shade over the sensorcould reduce drift effects considerably. However, an added precaution would be to conduct


regular “drift runs” where repeated data is periodically acquired along a transect to adjust forthe drift in the post-processing of ECa data. Drift runs are conducted in the morning, noon,and late afternoon to provide a range of diurnal temperature effects on the EM instrument.The variations in drift provide the basis for adjusting the ECa measurements.

Positional offset can also be a problem due to both the distance from the sensor to theGPS antenna and the data acquisition system time lags.Sudduth et al. (2001)found thatthe sensitivity of ECa to variations in sensor operating speed and height was relativelyminor. Nevertheless, mobile ECa equipment should not be operated at speeds higher than10 km h−1 to minimize positional offset effects.

4.2. Factors influencing within-field ECa variations and ECa-survey and soil-samplingstrategies to account for their occurrence

Variation of ECa within a field is due to spatial variation in soil properties influencingECa. The spatial heterogeneity of these soil properties is the consequence of the interactionof (i) soil formation processes, (ii) meteorologic processes, and (iii) anthropogenic influ-ences. Soil formation processes are the result of complex interactions between biological,physical, and chemical mechanisms acting on a parent material over time and influencedby topography. Meteorologic processes directly and indirectly influence soil formation pro-cesses. Anthropogenic influences are typically related to management practices includingleaching fraction and irrigation water quality. To implement an efficient ECa survey andassociated soil sampling plan to reliably characterize spatial variability, an awareness andunderstanding of the factors influencing within-field ECa variations are crucial. Suggestionsfor ECa surveys and sampling strategies are provided to account for the occurrence of thedeterministic and stochastic mechanisms that create local and within-field ECa variations.

4.2.1. ECa variation with salinityIn semi-arid regions where saline seeps occur due to shallow water tables and in arid agri-

cultural areas, salinity is generally the soil property that dominates the ECa measurement.Salinity accumulation occurs where evaporation or evapotranspiration exceed irrigationand/or precipitation. The predominant mechanism causing the accumulation of salt in irri-gated agricultural soils is loss of water through evapotranspiration, leaving ever increasingconcentrations of salts in the remaining water. Unlike texture or bulk density, which arestatic properties of soils, salinity is a dynamic property. It varies temporally and spatiallywith depth and across the landscape and exhibits high variation across a field and moderateto high local-scale variability (Corwin et al., 2003a). Local-scale variation determined fromnear-surface furrow samples acquired 0.5 m apart can vary anywhere from 10 to 100% dueto micro-scale soil composition characteristics and/or fluctuations in preferential water flow.

Electromagnetic induction instrument readings tend to average out this local-scale vari-ation. The degree to which this averaging occurs depends directly upon the instrument’s“foot print” (i.e., the volume of soil incorporated into the signal response). For example,an EM-38 signal will be influenced by any electrically conductive material within about1–2 m of the instrument (both laterally and vertically). Hence, it is typically assumed tohave about a 1.5 m foot print.


Because the volume of soil measured by the EM-38 is so much larger than the volumeobtained by conventional soil sampling techniques, an estimate of the degree of local-scalesalinity variation needs to be acquired for calibration model purposes and for soil qualityassessments of spatial variability. Such an estimate can be acquired by obtaining duplicateor replicate sample cores within a 1 m radius at some of the calibration sites during the soilsampling process. The replicate cores can then be used to estimate the local-scale salinityvariation (referred to as the “nugget variation” in geostatistical models and as the “pureerror estimate” in spatial regression models). The measured salinity data from these corescan be used to construct a residual autocorrelation test, known as a “lack-of-fit” test, forassessing the spatial residual independence assumption (Lesch et al., 1995a). Techniquesfor estimating the local-scale salinity variation and performing residual lack-of-fit tests aredescribed inLesch et al. (1995a).

In a typical 12–16 site calibration sampling design, replicate sample cores are commonlytaken at 4–6 of the calibration sites. These 4–6 sites can either be chosen at random (fromamongst the 16 sites) or selected throughout the survey area. Additionally, the core separa-tion spacing should be the same at all sites and both the primary and replicate cores shouldalways come from the same location with respect to the bed–furrow environment (i.e., bothfrom the bed, or both from the furrow).

Variations in salinity through the soil profile also influence the ECa measurement. Thisinfluence is particularly complex when ECa is measured with EM because the depth-weighted response function of the instrument (e.g., EM-31 or EM-38) is non-linear. Thedepth-weighted nonlinearity is shown inFig. 3, which illustrates the cumulative relativecontributions to ECa [i.e., R(z)] for a homogeneously conductive material below a normal-ized depth ofzbased on Eqs.(9) and (10)from McNeill (1980)for vertical and horizontal

Fig. 3. Cumulative relative contribution of all soil electrical conductivity,R(z), below various depths for the EM-38 apparent soil electrical conductivity reading when the device is held in a horizontal (parallel) and vertical(perpendicular) position. Taken fromMcNeill (1980).


dipoles, respectively:

Rv(z) = 1

(4z2 + 1)1/2(9)

Rh(z) = (4z2 + 1)1/2 − 2z (10)

When considering 0.3 m depth increments through the soil profile, the greatest responsefor EMh occurs over the 0–0.3 m depth increment and over the 0.3–0.6 m depth incrementfor EMv. Nevertheless, it is possible to determine the general shape of the salinity profileby the relative magnitudes of the EMh and EMv readings at a point in the field. For salinity-driven ECa surveys, if EMh > EMv, then the salinity profile is inverted and decreases withdepth; if EMh < EMv; then salinity increases; if salinity is uniform, then EMh ≈ EMv.

4.2.2. ECa variation with water contentSoil water content variations affect ECa measurements. Like salinity, soil water content

is a dynamic soil property that varies with depth and across the landscape, generally withmoderate to high local-scale variability. In areas under uniform irrigation management prac-tices, the degree of spatial water content variability is typically minimal provided significantsoil texture variation is not present. However, some fields demonstrate gradual trends in wa-ter content across the extent of the field, which may be due to gradual changes in shallowwater table levels close to the depth of penetration of measurement or to abrupt textural dis-continuities, or due to non-uniformity of water application (e.g., flood irrigation has a trendof high to low from the head water to tail water ends of a field, respectively). In instanceswhere gradual changes in the soil water content level occur, trend surface parameters in theregression model can be used.

Like salinity, variations in the water content through the soil profile influence the ECameasurement and this influence is particularly complex when using EM. As with salinity,it is possible to determine the general profile shape of water content. For fields where watercontent is the dominant soil property influencing the ECa measurement, if water contentdecreases with depth, then EMh > EMv; if water content increases, then EMh < EMv; ifwater content uniform, then EMh ≈ EMv.

It is important to remember that if the water content of the soil drops too low (e.g.,<0.10 cm3 cm−3), then the EM signal readings can become seriously dampened. In mostpractical applications, reliable EM signal data will be obtained when the soil is at or nearfield capacity. Surveying dry areas should be avoided. This is especially true for surveysthat use ER, which requires close contact between the electrodes and soil that can only beattained when the soil is moist.

4.2.3. ECa variation induced by changes in soil textureSoil texture can cause extremely complex spatial patterns of ECa. Under a uniform

irrigation distribution, water content will generally coincide with texture. Soils higher insand have lower water contents than soils higher in clay. Minimal complexities in spatialpatterns of ECa occur (i) when there is minimal soil texture variability, (ii) when the texture


changes are smooth and gradual across the field, or (iii) when the texture, water content,and salinity variations are strongly correlated.

4.2.4. ECa variation induced by other soil propertiesOther physical soil properties also affect the ECa survey readings to various degrees.

These properties include OM, magnetic susceptibility, and temperature. Significant variationin these properties needs to be present before any meaningful influence upon the ECa signalreading occurs. With respect to temperature, a 1◦C change in temperature throughout theentire soil profile typically causes no more than a 2% change in the EM-38 signal readings.Since soil temperature fluctuations below 0.3 m in the soil profile occur rather slowly, theentire survey process can usually be completed before a significant change in the bulk-average soil profile temperature occurs. Magnetic susceptibility is seldom a factor exceptfor soils high in free iron oxides. Apparent soil electrical conductivity variation related toOM has been investigated byJaynes (1996).

4.2.5. ECa variation with depthThe variation of ECa with depth is primarily due to gradations in salinity, texture,

and water content through the soil profile. Soil salinity levels can change quite rapidlywith depth; it is not unusual in some arid zone areas to see relative salinity profilelevels fluctuate by an order of magnitude within the top meter of soil. Water contenttends to increase with depth and will vary according to textural distribution. Textu-ral distribution is mainly a consequence of soil formation processes. However, anthro-pogenic effects can result in increased uniformity of texture within the plow layer orcan produce other localized variations. Salinity will vary in the soil profile primarilyfrom the process of leaching with plant, chemical, and topographic effects contributing tovariations.

Devising a sampling scheme that accounts for temporal and depth variations while main-taining accurate and consistent sampling depths throughout a survey area is critical. Withoutprior knowledge of the distribution of salinity, water content, and texture within a profile,it can be difficult to infer the appropriate sample depth design. For this reason, soil coresshould be acquired to a depth of at least 1.2–1.5 m at each sample site. If resources permit,each core can be sliced into subsamples, thereby facilitating the estimation of predictionfunctions (regression models) for multiple sample depths.

Electrical resistivity and EM techniques are both well suited for field-scale applicationsbecause their volumes of measurement are large, which reduces the influence of local-scalevariability. However, ER has a flexibility that has proven advantageous for field application,i.e., the depth and volume of measurement can be easily changed by altering the spacingbetween the electrodes. This allows the ECa for a discrete depth interval of soil to be easilycalculated with a fixed-array four electrode by measuring the ECa of successive layers forincreasing inter-electrode spacings and using the following equation (Barnes, 1952; Telfordet al., 1976):

ECx = ECai − ECai−1 =(

ECaiai − ECai−1ai−1

ai − ai−1

)(11)


whereai is the inter-electrode spacing, which equals the depth of sampling,ai−1 the previousinter-electrode spacing, which equals the depth of previous sampling, and ECx the apparentsoil electrical conductivity for a specific depth interval.

Electromagnetic induction can also measure ECa at variable depths determined by theheight of the EM instrument above the soil surface. Unlike ER, depth profiling of ECa withEM is mathematically complex (Borchers et al., 1997; McBratney et al., 2000; Hendrickxet al., 2002b). Measurements of ECa at variable depths with EM are usually achievedby positioning the EM instrument at various heights above the soil surface in either thevertical (EMv) or horizontal (EMh) dipole mode (Rhoades and Corwin, 1981; Corwin andRhoades, 1982). Though not required, the measurement of ECa near the soil surface (i.e.,top 0.25–0.5 m) along with spatially associated larger soil volume ECa measurements withEM equipment can be used to increase the accuracy of the fitted prediction functions thatdefine ECa profile variation (Lesch et al., 1992). Insertion four-electrode probes and small,hand-held fixed-arrays are both very useful for measuring the soil ECa within the first0.25–0.5 m of topsoil.

One sampling strategy commonly used in association with an ECa survey is to acquiresoil samples at each sample site in 0.3 m increments, typically down to a depth of either1.2 or 1.5 m. When sampling by hand (i.e., using a hand auger), each soil sample can beremoved individually. If a drill rig is available, then the entire core is usually bored at onetime and then split into subsamples after being brought to the surface. To minimize temporalchanges that may occur in dynamic soil properties such as water content and salinity, ECasurveys and associated soil sampling are conducted when the soil is at or near field capacity(i.e., water content of soil after free drainage has occurred, generally 3 and 4 days followingan irrigation).

4.2.6. ECa variation induced by surface topographySurface topography plays a significant role in influencing spatial ECavariation. Slope and

aspect will determine the level and location of runoff and infiltration, which will influencethe variation in water content and salinity at local scales and larger. Areas where the slopeis steep tend to have lower water content than areas where a depression occurs. All otherfactors being equal, flat areas tend to be more spatially uniform in areal variation of watercontent. The influence of surface topography on salinity distribution coincides with theinfluence of surface topography on water flow gradients, which result in salt transport.

The bed–furrow environment is an example of surface topography that can have subtlelocal-scale variation effects. The bed–furrow topography can be a source of considerablewater and salinity variation, particularly over the cropping season. A percentage of theirrigation water applied to a furrow will move laterally and upwards into adjacent beds dueto capillary flow. This water movement in turn will carry near-surface soluble salts up intothe bed. In flood irrigated fields the relative difference between the near-surface furrow andbed salinity levels can become pronounced over time.

Fig. 4 displays the geometrical distribution of soil salinity throughout the near-surfacebed–furrow environment within a fixed-bed, flood-irrigated cotton field in the CoachellaValley, California (sampled in 1992). The high mean salinity level was 23.2 dS m−1 through-out the bed–furrow; the ratio of bed to furrow near-surface salinity levels was 4:1. At the lowlevel (5.8 dS m−1) this ratio actually increased to 8:1.Fig. 4indicates that the overall mean


Fig. 4. Two-dimensional salinity distribution within the bed–furrow.

salinity level would have been very poorly estimated by samples acquired either only in thefurrows or in the bed. In this particular survey, it took 14 soil samples at each sample site toadequately describe the two-dimensional pattern of salinity present within the bed–furrowenvironment. The main point conveyed byFig. 4 is that more extensive sampling schemesmust be considered for bed–furrow systems, if knowledge of the two-dimensional salin-ity and/or water content distributions is an essential objective of the survey. However, if adetailed knowledge of the two-dimensional distributions of a bed–furrow system is unnec-essary, then there is a critical need for consistency with respect to soil core locations. Inthese instances, all soil cores should be sampled from the same place within the bed–furrow.Furthermore, the ECa survey data, particularly from EM equipment, and soil sample coresshould be acquired from exactly the same location within the bed–furrow. For example, ifthe ECa data are acquired over the furrows, then the soil samples should also be acquiredfrom the furrows.

Composite or bulk sampling strategies for averaging variations in a bed–furrow sys-tem are generally not recommended. In composite sampling, soil samples would be ac-quired from both bed and furrow locations and then mixed together in an effort to obtain amore “representative” sample. Composite samples are not recommended for two reasons.First, it doubles the field work without providing any knowledge of the two-dimensional,bed–furrow salinity distribution and second, it often introduces more variability into thesample data (through poor mixing processes) than it removes through averaging.

Irrigated, agricultural land is typically laser-leveled to improve irrigation efficiency.Various leveling designs are used, depending on the method of irrigation and the agricultural


crop under production. The three most common designs include dead leveling, single-slopeleveling, and dual slope leveling. All these three designs create a theoretical plane that canbe written mathematically as:

Surface elevation= α0 + α1x + α2y (12)

wherexandy represent the physical (x, y) coordinates, and theα0, α1 andα2 coefficients theprimary and secondary slopes of the field. Eq.(12)is referred to as a first-order trend surfaceequation. The regression relationship between the geo-referenced ECa measurements andsoil properties (i.e., salinity, water content, texture) influencing the ECa measurement canbe influenced by gradual changes in salinity, water content, and/or texture across the surveyarea due to changes in elevation. These gradual changes in salinity, water content, and/ortexture due to elevation can be taken into account through the inclusion of a first-order orsecond-order trend surface equation in the regression analysis.

If an ECa survey is conducted on non-graded farmland, which exhibits significant localvariation in surface elevation, then it will usually be necessary to conduct a surface elevationsurvey along with the ECa survey. This elevation data can then be directly incorporated intoany regression analyses relating ECa to soil properties or into the sample design strategy.

4.2.7. ECa variation induced by traffic patternsAnother source of potential ECa variation arises from soil compaction caused by repet-

itive traffic patterns of heavy agricultural equipment. In many fields, heavy equipment isconsistently driven down the same set of furrows when performing tillage and cultivationoperations over the growing season. This leads to a systematic pattern of compaction in asubset of furrows throughout the field.

Fig. 5displays the EM-38 horizontal (EMh) readings acquired in 1991 along 30 adjacentfurrows in a buried drip-irrigated cotton field (Westlands Water District, California) subjectto repetitive traffic influences. In this case the traffic pattern induced a clearly cyclic pattern inthe EMh readings; the highest conductivity readings consistently occurred in the compactedfurrows whereρb was characteristically higher near the soil surface.

The data shown inFig. 5 is atypical. In general, compaction induced cyclic patternsdo not have such pronounced effects on ECa. Nonetheless, caution should be taken tosystematically avoid taking ECa measurements down compacted furrows, particularly withEM equipment. Excessive soil compaction will nearly always have at least some effecton theρb, soil salinity, and water content levels. Random surveying (and sampling) ofboth compacted and non-compacted furrows in the same field will introduce variabilitythat should be avoided whenever possible. A simple, but effective, means of determiningfurrows that have been compacted by repeated traffic from heavy equipment is by pushing along screwdriver or rod into the furrow to determine if the resistance is greater than adjacentfurrows.

4.2.8. ECa variation induced by irrigation management practicesIrrigation management has a pronounced effect on determining areal and profile ECa

distributions within a field. The amount and frequency of irrigation will directly influence


Fig. 5. Variations of ECa due to compaction from repeated heavy equipment operation on a drip-irrigated cottonfield.

the movement of water and soluble salts through the profile and across the field. The watercontent of the soil during the ECa survey will be at least partially determined by the elapsedtime from the last irrigation event, the presence or absence of a crop, and if present, thematurity of the crop.

Since a change in irrigation management (e.g., drip, sprinkler, or flood irrigation) canseriously affect the three-dimensional salinity and water content distributions within a field,it is important to avoid conducting an ECa survey (and/or soil sampling) for an area undermore than one irrigation management strategy. This means that any survey area must berestricted to farmland under similar irrigation management practices. Each survey must beconducted entirely within a single, homogeneous irrigation management area. The failureto restrict EM readings to an area under a single water management strategy can result inserious regression model and sample design bias with inflated errors that corrupt the entiresurveying process.

5. Conclusions

All landscape-scale environmental and agricultural studies involving soil require a spa-tial and often times a temporal knowledge of soil physico-chemical properties. Surveys ofECa provide one of the most reliable and comprehensive means for obtaining this spatio-temporal information. The development of mobile ECa equipment has made it possibleto characterize spatial variability of a variety of physico-chemical properties both rapidlyand cost effectively. Nevertheless, previous surveys have met with inconsistent results attimes due to the inexperience of the researcher and/or the lack of a standardized set of ECasurvey protocols. Concomitantly, the ability to use ECa survey information in a compar-


ative sense from one site to another and between different researchers or to collectivelybuild a database comprised of ECa measurements from multiple locations by multiple re-searchers requires a framework that standardizes the collection of ECa survey and associatedspatial edaphic information. The outlined protocols for conducting ECa surveys provide ameans of characterizing the spatial variability of soil physico-chemical properties that canbe used in site-specific crop management, landscape-scale modeling of non-point sourcepollutants in the vadose zone, and soil quality assessment. Guidelines are also presented forunderstanding and interpreting ECa survey measurements. It is the intent of the developedprotocols to improve the reliability, consistency, and compatibility of ECa survey measure-ments and their interpretation for characterizing spatial variability of soil physico-chemicalproperties.

Acknowledgements

The authors wish to acknowledge the collaborative work during the 1980s and 1990swith Dr. James Rhoades, which ultimately led to the development of the ECa survey pro-tocols. Dr. Rhoades’ formative work on the measurement of soil salinity with ECa and thedevelopment of mobile ECa-measurement equipment laid the foundation for subsequentspatial variability characterization with ECa. The authors also thank the three anonymousreviewers for their constructive comments, which helped to improve the quality of this paper,and the suggestions of Dr. Daniel Schmoldt, Editor-in-Chief ofComputers and Electronicsin Agriculture.

References

Anderson-Cook, C.M., Alley, M.M., Roygard, J.K.F., Khosia, R., Noble, R.B., Doolittle, J.A., 2002. Differentiatingsoil types using electromagnetic conductivity and crop yield maps. Soil Sci. Soc. Am. J. 66, 1562–1570.

Arshad, M.A., Coen, G.M., 1992. Characterization of soil quality: physical and chemical criteria. Am. J. Alt.Agric. 7, 25–31.

Banton, O., Seguin, M.K., Cimon, M.A., 1997. Mapping field-scale physical properties of soil with electricalresistivity. Soil Sci. Soc. Am. J. 61 (4), 1010–1017.

Barnes, H.E., 1952. Soil investigation employing a new method of layer-value determination for earth resistivityinterpretation. Highway Res. Board Bull. 65, 26–36.

Bennett, D.L., George, R.J., 1995. Using the EM38 to measure the effect of soil salinity onEucalyptus globulusin south-western Australia. Agric. Water Manage. 27, 69–86.

Benson, A.K., Payne, K.L., Stubben, M.A., 1997. Mapping groundwater contamination using DC resistivity andVLF geophysical methods—a case study. Geophysics 62 (1), 80–86.

Boettinger, J.L., Doolittle, J.A., West, N.E., Bork, E.W., Schupp, E.W., 1997. Nondestructive assessment ofrangeland soil depth to petrocalcic horizon using electromagnetic induction. Arid Soil Res. Rehab. 11 (4),372–390.

Borchers, B., Uram, T., Hendrickx, J.M.H., 1997. Tikhonov regularization of electrical conductivity depth profilesin field soils. Soil Sci. Soc. Am. J. 61, 1004–1009.

Bouma, J., 1989. Using soil survey data for quantitative land evaluation. Adv. Soil Sci. 9, 177–213.Bowling, S.D., Schulte, D.D., Woldt, W.E., 1997. A geophysical and geostatistical methodology for evaluating

potential subsurface contamination from feedlot runoff retention ponds. ASAE Paper No. 972087. ASAE, St.Joseph, MI, USA.


Box, G.E.P., Draper, N.R., 1987. Empirical Model-building and Response Surfaces. Wiley, New York, NY, USA,669 pp.

Brevik, E.C., Fenton, T.E., 2002. The relative influence of soil water, clay, temperature, and carbonate minerals onsoil electrical conductivity readings taken with an EM-38 along a Mollisol catena in central Iowa. Soil Surv.Horizons 43, 9–13.

Brune, D.E., Doolittle, J., 1990. Locating lagoon seepage with radar and electromagnetic survey. Environ. Geol.Water Sci. 16, 195–207.

Brune, D.E., Drapcho, C.M., Radcliff, D.E., Harter, T., Zhang, R., 1999. Electromagnetic survey to rapidly assesswater quality in agricultural watersheds. ASAE Paper No. 992176. ASAE, St. Joseph, MI, USA.

Brus, D.J., Knotters, M., van Dooremolen, W.A., van Kernebeek, P., van Seeters, R.J.M., 1992. The use of electro-magnetic measurements of apparent soil electrical conductivity to predict the boulder clay depth. Geoderma55 (1–2), 79–93.

Cameron, D.R., de Jong, E., Read, D.W.L., Oosterveld, M., 1981. Mapping salinity using resistivity and electro-magnetic inductive techniques. Can. J. Soil Sci. 61, 67–78.

Cannon, M.E., McKenzie, R.C., Lachapelle, G., 1994. Soil-salinity mapping with electromagnetic induction andsatellite-based navigation methods. Can. J. Soil Sci. 74 (3), 335–343.

Carter, L.M., Rhoades, J.D., Chesson, J.H., 1993. Mechanization of soil salinity assessment for mapping. ASAEPaper No. 931557. ASAE, St. Joseph, MI, USA.

Christensen, R., Johnson, W., Pearson, L.M., 1992. Prediction diagnostics for spatial linear models. Biometrika79 (3), 583–591.

Cook, P.G., Kilty, S., 1992. A helicopter-borne electromagnetic survey to delineate groundwater recharge rates.Water Resour. Res. 28 (11), 2953–2961.

Cook, P.G., Walker, G.R., Buselli, G., Potts, I., Dodds, A.R., 1992. The application of electromagnetic techniquesto groundwater recharge investigations. J. Hydrol. 130, 201–229.

Corwin, D.L., Carrillo, M.L.K., Vaughan, P.J., Rhoades, J.D., Cone, D.G., 1999. Evaluation of GIS-linked modelof salt loading to groundwater. J. Environ. Qual. 28, 471–480.

Corwin, D.L., Kaffka, S.R., Hopmans, J.W., Mori, Y., Lesch, S.M., Oster, J.D., 2003a. Assessment and field-scalemapping of soil quality properties of a saline-sodic soil. Geoderma 114 (3–4), 231–259.

Corwin, D.L., Lesch, S.M., 2003. Application of soil electrical conductivity to precision agriculture: theory,principles, and guidelines. Agron. J. 95 (3), 455–471.

Corwin, D.L., Lesch, S.M., 2005a. Apparent soil electrical conductivity measurements in agriculture. Comput.Electron. Agric. 46, 11–43.

Corwin, D.L., Lesch, S.M., 2005b. Characterizing soil spatial variability with apparent soil electrical conductivityPart II. Case study. Comput. Electron. Agric. 46, 135–152.

Corwin, D.L., Lesch, S.M., Shouse, P.J., Soppe, R., Ayars, J.E., 2003b. Identifying soil properties that influencecotton yield using soil sampling directed by apparent soil electrical conductivity. Agron. J. 95 (2), 352–364.

Corwin, D.L., Rhoades, J.D., 1982. An improved technique for determining soil electrical conductivity–depthrelations from above-ground electromagnetic measurements. Soil Sci. Soc. Am. J. 46, 517–520.

Corwin, D.L., Rhoades, J.D., 1984. Measurement of inverted electrical conductivity profiles using electromagneticinduction. Soil Sci. Soc. Am. J. 48, 288–291.

Corwin, D.L., Rhoades, J.D., 1990. Establishing soil electrical conductivity–depth relations from electromagneticinduction measurements. Commun. Soil Sci. Plant Anal. 21 (11–12), 861–901.

Dane, J.H., Topp, G.C. (Eds.), 2002. Methods of Soil Analysis, Part 4—Physical Methods. Soil Sci. Soc. Am.Book Series 5. Soil Science Society of America, Madison, WI, USA.

de Jong, E., Ballantyne, A.K., Caneron, D.R., Read, D.W., 1979. Measurement of apparent electrical conductivityof soils by an electromagnetic induction probe to aid salinity surveys. Soil Sci. Soc. Am. J. 43, 810–812.

Diaz, L., Herrero, J., 1992. Salinity estimates in irrigated soils using electromagnetic induction. Soil Sci. 154,151–157.

Doolittle, J.A., Indorante, S.J., Potter, D.K., Hefner, S.G., McCauley, W.M., 2002. Comparing three geophysicaltools for locating sand blows in alluvial soils of southeast Missouri. J. Soil Water Conserv. 57 (3), 175–182.

Doolittle, J.A., Sudduth, K.A., Kitchen, N.R., Indorante, S.J., 1994. Estimating depths to claypans using electro-magnetic induction methods. J. Soil Water Conserv. 49 (6), 572–575.


Doran, J.W., Parkin, T.B., 1994. Defining and assessing soil quality. In: Doran, J.W., Coleman, D.C., Bedzicek,D.F., Stewart, B.A. (Eds.), Defining Soil Quality for a Sustainable Environment. SSSA Special PublicationNo. 35. Soil Science Society of America, Madison, WI, USA, pp. 3–22.

Doran, J.W., Parkin, T.B., 1996. Quantitative indicators of soil quality: a minimum data set. In: Doran, J.W., Jones,A.J. (Eds.), Methods for Assessing Soil Quality. SSSA Special Publication No. 49. SSSA, Madison, WI, USA,pp. 25–38.

Drommerhausen, D.J., Radcliffe, D.E., Brune, D.E., Gunter, H.D., 1995. Electromagnetic conductivity surveys ofdairies for groundwater nitrate. J. Environ. Qual. 24, 1083–1091.

Eigenberg, R.A., Doran, J.W., Nienaber, J.A., Ferguson, R.B., Woodbury, B.L., 2002. Electrical conductivitymonitoring of soil condition and available N with animal manure and a cover crop. Agric. Ecosyst. Environ.88, 183–193.

Eigenberg, R.A., Korthals, R.L., Neinaber, J.A., 1998. Geophysical electromagnetic survey methods applied toagricultural waste sites. J. Environ. Qual. 27, 215–219.

Eigenberg, R.A., Nienaber, J.A., 1998. Electromagnetic survey of cornfield with repeated manure applications. J.Environ. Qual. 27, 1511–1515.

Eigenberg, R.A., Nienaber, J.A., 1999. Soil conductivity map differences for monitoring temporal changes in anagronomic field. ASAE Paper No. 992176.ASAE, St. Joseph, MI, USA.

Eigenberg, R.A., Nienaber, J.A., 2001. Identification of nutrient distribution at abandoned livestock manure han-dling site using electromagnetic induction. ASAE Paper No. 012193. ASAE, St. Joseph, MI, USA.

Ellsbury, M.M., Woodson, W.D., Malo, D.D., Clay, D.E., Carlson, C.G., Clay, S.A., 1999. Spatial variability incorn rootworm distribution in relation to spatially variable soil factors and crop condition. In: Robert, P.C.,Rust, R.H., Larson, W.E. (Eds.), Proceedings of the Fourth International Conference on Precision Agriculture.St. Paul, MN, July 19 and 22, 1998. ASA-CSSA-SSSA, Madison, WI, USA, pp. 523–533.

Fenton, T.E., Lauterbach, M.A., 1999. Soil map unit composition and scale of mapping related to interpretationsfor precision soil and crop management in Iowa. In: Robert, P.C., Rust, R.H., Larson, W.E. (Eds.), Proceedingsof the Fourth International Conference on Precision Agriculture. St. Paul, MN, July 19 and 22, 1998. ASA-CSSA-SSSA, Madison, WI, USA, pp. 239–251.

Fitterman, D.V., Stewart, M.T., 1986. Transient electromagnetic sounding for groundwater. Geophysics 51,995–1005.

Fraisse, C.W., Sudduth, K.A., Kitchen, N.R., 2001. Delineation of site-specific management zones by unsupervisedclassification of topographic attributes and soil electrical conductivity. Trans. ASAE 44 (1), 155–166.

Freeland, R.S., Branson, J.L., Ammons, J.T., Leonard, L.L., 2001. Surveying perched water on anthropogenicsoils using non-intrusive imagery. Trans. ASAE 44, 1955–1963.

Freeland, R.S., Yoder, R.E., Ammons, J.T., Leonard, L.L., 2002. Mobilized surveying of soil conductivity usingelectromagnetic induction. Appl. Eng. Agric. 18 (1), 121–126.

Gorucu, S., Khalilian, A., Han, Y.J., Dodd, R.B., Wolak, F.J., Keskin, M., 2001. Variable depth tillage basedon geo-referenced soil compaction data in coastal plain region of South Carolina. ASAE Paper No. 011016.ASAE, St. Joseph, MI, USA.

Gotway, C.A., Ferguson, R.B., Hergert, G.W., Peterson, T.A., 1996. Comparison of kriging and inverse-distancemethods for mapping soil parameters. Soil Sci. Soc. Am. J. 60, 1237–1247.

Greenhouse, J.P., Slaine, D.D., 1983. The use of reconnaissance electromagnetic methods to map contaminantmigration. Ground Water Monit. Rev. 3 (2), 47–59.

Greenhouse, J.P., Slaine, D.D., 1986. Geophysical modelling and mapping of contaminated groundwater aroundthree waste disposal sites in southern Ontario. Can. Geotechnol. J. 23, 372–384.

Halvorson, A.D., Rhoades, J.D., 1976. Field mapping soil conductivity to delineate dryland seeps with four-electrode techniques. Soil Sci. Soc. Am. J. 44, 571–575.

Hanson, B.R., Kaita, K., 1997. Response of electromagnetic conductivity meter to soil salinity and soil-watercontent. J. Irrig. Drain. Eng. 123, 141–143.

Hendrickx, J.M.H., Baerends, B., Raza, Z.I., Sadig, M., Chaudhry, M.A., 1992. Soil salinity assessment by elec-tromagnetic induction of irrigated land. Soil Sci. Soc. Am. J. 56, 1933–1941.

Hendrickx, J.M.H., Borchers, B., Corwin, D.L., Lesch, S.M., Hilgendorf, A.C., Schlue, J., 2002b. Inversion ofsoil conductivity profiles from electromagnetic induction measurements: theory and experimental verification.Soil Sci. Soc. Am. J. 66, 673–685.


Hendrickx, J.M.H., Das, B., Corwin, D.L., Wraith, J.M., Kachanoski, R.G., 2002a. Indirect measurement ofsolute concentration. In: Dane, J.H., Topp, G.C. (Eds.), Methods of Soil Analysis, Part 4—Physical Meth-ods. Soil Sci. Soc. Am. Book Series 5. Soil Science Society of America, Madison, WI, USA, pp. 1274–1306.

Herrero, J., Ba, A.A., Aragues, R., 2003. Soil salinity and its distribution determined by soil sampling and elec-tromagnetic techniques. Soil Use Manage. 19 (2), 119–126.

Inman, D.J., Freeland, R.S., Yoder, R.E., Ammons, J.T., Leonard, L.L., 2001. Evaluating GPR and EMI formorphological studies of loessial soils. Soil Sci. 166 (9), 622–630.

Isaaks, E.H., Srivastava, R.M., 1989. Applied Geostatistics. Oxford University Press, New York, NY.Jaynes, D.B., 1996. Mapping the areal distribution of soil parameters with geophysical techniques. In: Corwin,

D.L., Loague, K. (Eds.), Applications of GIS to the Modeling of Non-point Source Pollutants in the VadoseZone. SSSA Special Publication No. 48. SSSA, Madison, WI, USA, pp. 205–216.

Jaynes, D.B., Colvin, T.S., Ambuel, J., 1993. Soil type and crop yield determinations from ground conductivitysurveys. ASAE Paper No. 933552. ASAE, St. Joseph, MI, USA.

Jaynes, D.B., Novak, J.M., Moorman, T.B., Cambardella, C.A., 1995. Estimating herbicide partition coefficientsfrom electromagnetic induction measurements. J. Environ. Qual. 24, 36–41.

Johnson, C.K., Doran, J.W., Duke, H.R., Weinhold, B.J., Eskridge, K.M., Shanahan, J.F., 2001. Field-scale elec-trical conductivity mapping for delineating soil condition. Soil Sci. Soc. Am. J. 65, 1829–1837.

Johnston, M.A., Savage, M.J., Moolman, J.H., du Pleiss, H.M., 1997. Evaluation of calibration methods forinterpreting soil salinity from electromagnetic induction measurements. Soil Sci. Soc. Am. J. 61, 1627–1633.

Kachanoski, R.G, de Jong, E., Van-Wesenbeeck, I.J., 1990. Field scale patterns of soil water storage from non-contacting measurements of bulk electrical conductivity. Can. J. Soil Sci. 70, 537–541.

Kachanoski, R.G., Gregorich, E.G., Van-Wesenbeeck, I.J., 1988. Estimating spatial variations of soil water contentusing noncontacting electromagnetic inductive methods. Can. J. Soil Sci. 68, 715–722.

Kaffka, S.R., Lesch, S.M., Bali, K.M., Corwin, D.L., 2005. Site-specific management in salt-affected sugar beetfields using electromagnetic induction. Comput. Electron. Agric. 46, 329–350.

Kean, W.F., Jennings Walker, M., Layson, H.R., 1987. Monitoring moisture migration in the vadose zone withresistivity. Ground Water 25, 562–571.

Khakural, B.R., Robert, P.C., Hugins, D.R., 1998. Use of non-contacting electromagnetic inductive method forestimating soil moisture across a landscape. Commun. Soil Sci. Plant Anal. 29, 2055–2065.

Kitchen, N.R., Sudduth, K.A., Drummond, S.T., 1996. Mapping of sand deposition from 1993 midwest floodswith electromagnetic induction measurements. J. Soil Water Conserv. 51 (4), 336–340.

Kitchen, N.R., Sudduth, K.A., Drummond, S.T., 1999. Soil electrical conductivity as a crop productivity measurefor claypan soils. J. Prod. Agric. 12, 607–617.

Klestra, E.J., Bartz, J.K., 1996. Quality assurance and quality control. In: Sparks, D.L. (Ed.), Methods of SoilAnalysis, Part 3—Chemical Methods. Soil Sci. Soc. Am. Book Series No. 5. Soil Science Society of America,Madison, WI, USA, pp. 19–48.

Kravchenko, A.N., Bollero, G.A., Omonode, R.A., Bullock, D.G., 2002. Quantitative mapping of soil drainageclasses using topographical data and soil electrical conductivity. Soil Sci. Soc. Am. J. 66, 235–243.

Kravchenko, A.N., Bullock, D.G., 1999. A comparative study of interpolation methods for mapping soil properties.Agron. J. 91, 393–400.

Lark, R.M., Kaffka, S.R., Corwin, D.L., 2003. Multiresolution analysis of data of electrical conductivity of soilusing wavelets. J. Hydrol. 272, 276–290.

Larson, W.E., Pierce, F.J., 1991. Conservation and enhancement of soil quality. In: Dumanski, J. (Ed.), Proceedingsof the International Workshop on Evaluation for Sustainable Land Management in the Developing World, vol.2, Technical Papers, Chaning Rai, Thailand, September 15 and 21, 1990. Int. Board Soil Res. Manage, Bangkok,Thailand, pp. 175–203.

Larson, W.E., Pierce, F.J., 1994. The dynamics of soil quality as a measure of sustainable management. In:Doran, J.W., Coleman, D.C., Bedzicek, D.F., Stewart, B.A. (Eds.), Defining Soil Quality for a SustainableEnvironment. SSSA Special Publication No. 35. Soil Science Society of America, Madison, WI, USA, pp.37–51.

Laslett, G.M., McBratney, A.B., Pahl, P.J., Hutchinson, M.F., 1987. Comparison of several spatial predictionmethods for soil pH. J. Soil Sci. 38, 325–341.


Leenaers, H., Okx, J.P., Burrough, P.A., 1990. Comparison of spatial prediction methods for mapping floodplainsoil pollution. Catena 17, 535–550.

Lesch, S.M., Corwin, D.L., 2003. Predicting EM/soil property correlation estimates via the dual pathway parallelconductance model. Agron. J. 95 (2), 365–379.

Lesch, S.M., Herrero, J., Rhoades, J.D., 1998. Monitoring for temporal changes in soil salinity using electromag-netic induction techniques. Soil Sci. Soc. Am. J. 62, 232–242.

Lesch, S.M., Rhoades, J.D., Lund, L.J., Corwin, D.L., 1992. Mapping soil salinity using calibrated electromagneticmeasurements. Soil Sci. Soc. Am. J. 56, 540–548.

Lesch, S.M., Rhoades, J.D., Corwin, D.L., 2000. ESAP-95 Version 2.10R: user manual and tutorial guide. ResearchReport 146. USDA-ARS George E. Brown Jr. Salinity Laboratory, Riverside, CA, USA.

Lesch, S.M., Strauss, D.J., Rhoades, J.D., 1995a. Spatial prediction of soil salinity using electromagnetic inductiontechniques. 1. Statistical prediction models: a comparison of multiple linear regression and cokriging. WaterResour. Res. 31, 373–386.

Lesch, S.M., Strauss, D.J., Rhoades, J.D., 1995b. Spatial prediction of soil salinity using electromagnetic inductiontechniques. 2. An efficient spatial sampling algorithm suitable for multiple linear regression model identificationand estimation. Water Resour. Res. 31, 387–398.

Lund, E.D., Chrissty, C.D., Drummond, P.E., 1999. Applying soil electrical conductivity to precision agriculture.In: Robert, P.C., Rust, R.H., Larson, W.E. (Eds.), Proceedings of the Fourth International Conference onPrecision Agriculture. St. Paul, MN, July 19–22, 1998. ASA-CSSA-SSSA, Madison, WI, USA, pp. 1089–1100.

Mankin, K.R., Ewing, K.L., Schrock, M.D., Kluitenberg, G.J., 1997. Field measurement and mapping of soilsalinity in saline seeps. ASAE Paper No. 973145. ASAE, St. Joseph, MI, USA.

Mankin, K.R., Karthikeyan, R., 2002. Field assessment of saline seep remediation using electromagnetic induction.Trans. ASAE 45 (1), 99–107.

McBratney, A.B., Bishop, T.F.A., Teliatnikov, I.S., 2000. Two soil profile reconstruction techniques. Geoderma97, 209–221.

McBratney, A.B., Webster, R., 1983. Optimal interpolation and isarithmic mapping of soil properties V. Co-regionalization and multiple sampling strategy. J. Soil Sci. 34, 137–162.

McBride, R.A., Gordon, A.M., Shrive, S.C., 1990. Estimating forest soil quality from terrain measurements ofapparent electrical conductivity. Soil Sci. Soc. Am. J. 54, 290–293.

McNeill, J.D., 1980. Electromagnetic Terrain Conductivity Measurement at Low Induction Numbers. Tech. NoteTN-6. Geonics Limited, Ontario, Canada.

Morgan, C.L.S., Norman, J.M., Wolkowski, R.P., Lowery, B., Morgan, G.D., Schuler, R., 2000. Two approaches tomapping plant available water: EM-38 measurements and inverse yield modeling. In: Roberts, P.C., Rust, R.H.,Larson, W.E. (Eds.), Proceedings of the Fifth International Conference on Precision Agriculture (CD-ROM).Minneapolis, MN, July 16–19, 2000. ASA-CSSA-SSSA, Madison, WI, USA, p. 14.

Nettleton, W.D., Bushue, L., Doolittle, J.A., Wndres, T.J., Indorante, S.J., 1994. Sodium affected soil identificationin south-central Illinois by electromagnetic induction. Soil Sci. Soc. Am. J. 58, 1190–1193.

Nielsen, D.R., Biggar, J.W., Erh, K.T., 1973. Spatial variability of field-measured soil-water properties. Hilgardia42 (7), 215–259.

Nobes, D.C., Armstrong, M.J., Close, M.E., 2000. Delineation of a landfill leachate plume and flow channels incoastal sands near Christchurch, New Zealand, using a shallow electromagnetic survey method. Hydrogeol. J.8 (3), 328–336.

Nyquist, J.E., Blair, M.S., 1991. Geophysical tracking and data logging system: description and case history.Geophysics 56 (7), 1114–1121.

Paine, J.G., 2003. Determining salinization extent, identifying salinity sources, and estimating chloride massusing surface, borehole, an airborne electromagnetic induction methods. Water Resour. Res. 39 (3),1059.

Peterson, R.G., Calvin, L.D., 1996. Sampling. In: Sparks, D.L. (Ed.), Methods of Soil Analysis, Part 3—ChemicalMethods. Soil Sci. Soc. Am. Book Series No. 5. Soil Science Society of America, Madison, WI, USA, pp.1–17.

Ranjan, R.S., Karthigesu, T., Bulley, N.R., 1995. Evaluation of an electromagnetic method for detecting lateralseepage around manure storage lagoons. ASAE Paper No. 952440. ASAE, St. Joseph, MI, USA.


Rhoades, J.D., 1992. Instrumental field methods of salinity appraisal. In: Topp, G.C., Reynolds, W.D.,Green, R.E. (Eds.), Advances in Measurement of Soil Physical Properties: Bring Theory into Prac-tice. SSSA Special Publication No. 30. Soil Science Society of America, Madison, WI, USA, pp. 231–248.

Rhoades, J.D., 1993. Electrical conductivity methods for measuring and mapping soil salinity. In: Sparks,D.L. (Ed.), Advances in Agronomy, vol. 49. Academic Press, San Diego, CA, USA, pp. 201–251.

Rhoades, J.D., 1996. Salinity: electrical conductivity and total dissolved salts. In: Sparks, D.L. (Ed.), Methodsof Soil Analysis, Part 3—Chemical Methods. Soil Sci. Soc. Am. Book Series No. 5. Soil Science Society ofAmerica, Madison, WI, USA, pp. 417–435.

Rhoades, J.D., Chanduvi, F., Lesch, S., 1999b. Soil salinity assessment: methods and interpretation of electricalconductivity measurements. FAO Irrigation and Drainage Paper No. 57. Food and Agriculture Organizationof the United Nations, Rome, Italy, 150 pp.

Rhoades, J.D., Corwin, D.L., 1981. Determining soil electrical conductivity–depth relations using an inductiveelectromagnetic soil conductivity meter. Soil Sci. Soc. Am. J. 45, 255–260.

Rhoades, J.D., Corwin, D.L., 1990. Soil electrical conductivity: effects of soil properties and application to soilsalinity appraisal. Commun. Soil Sci. Plant Anal. 21, 837–860.

Rhoades, J.D., Corwin, D.L., Lesch, S.M., 1999a. Geospatial measurements of soil electrical conductivity toassess soil salinity and diffuse salt loading from irrigation. In: Corwin, D.L., Loague, K., Ellsworth, T.R.(Eds.), Assessment of Non-point Source Pollution in the Vadose Zone. Geophysical Monograph 108. AmericanGeophysical Union, Washington, DC, USA, pp. 197–215.

Rhoades, J.D., Halvorson, A.D., 1977. Electrical conductivity methods for detecting and delineating saline seepsand measuring salinity in Northern Great Plains soils, ARS W-42. USDA-ARS Western Region, Berkeley,CA, USA, pp. 1–45.

Rhoades, J.D., Lesch, S.M., LeMert, R.D., Alves, W.J., 1997. Assessing irrigation/drainage/salinity managementusing spatially referenced salinity measurements. Agric. Water Manage. 35, 147–165.

Rhoades, J.D., Manteghi, N.A., Shouse, P.J., Alves, W.J., 1989. Soil electrical conductivity and soil salinity: newformulations and calibrations. Soil Sci. Soc. Am. J. 53, 433–439.

Rhoades, J.D., Raats, P.A.C., Prather, R.J., 1976. Effects of liquid-phase electrical conductivity, water content andsurface conductivity on bulk soil electrical conductivity. Soil Sci. Soc. Am. J. 40, 651–655.

Rhoades, J.D., Shouse, P.J., Alves, W.J., Manteghi, N.M., Lesch, S.M., 1990. Determining soil salinity from soilelectrical conductivity using different models and estimates. Soil Sci. Soc. Am. J. 54, 46–54.

Robinson, D.A., Lebron, I., Lesch, S.M., Shouse, P., 2004. Minimizing drift in electrical conductivity measurementsin high temperature environments using the EM-38. Soil Sci. Soc. Am. J. 68, 339–345.

Royall, R.M., 1988. The prediction approach to sampling theory. In: Krishnaiah, P.R., Rao, C.R. (Eds.), Handbookof Statistics, vol. 6. Elsevier, Amsterdam, The Netherlands, pp. 399–413 (sampling).

Russo, D., 1984. Design of a optimal sampling network for estimating the variogram. Soil Sci. Soc. Am. J. 48,708–716.

Salama, R.B., Bartle, G., Farrington, P., Wilson, V., 1994. Basin geomorphological controls on the mechanism ofrecharge and discharge and its effect on salt storage and mobilization: comparative study using geophysicalsurveys. J. Hydrol. 155 (1/2), 1–26.

Scanlon, B.R., Paine, J.G., Goldsmith, R.S., 1999. Evaluation of electromagnetic induction as a reconnaissancetechnique to characterize unsaturated flow in an arid setting. Ground Water 37 (2), 296–304.

Sheets, K.R., Hendrickx, J.M.H., 1995. Non-invasive soil water content measurement using electromagneticinduction. Water Resour. Res. 31, 2401–2409.

Slavich, P.G., Petterson, G.H., 1990. Estimating average rootzone salinity from electromagnetic induction (EM-38)measurements. Aust. J. Soil Res. 28, 453–463.

Slavich, P.G., Yang, J., 1990. Estimation of field-scale leaching rates from chloride mass balance and electromag-netic induction measurements. Irrig. Sci. 11, 7–14.

Sparks, D.L. (Ed.), 1996. Methods of Soil Analysis, Part 3—Chemical Methods. Soil Sci. Soc. Am. Book Series5. Soil Science Society of America, Madison, WI, USA.

Stroh, J.C., Archer, S.R., Doolittle, J.A., Wilding, L.P., 2001. Detection of edaphic discontinuities with ground-penetrating radar and electromagnetic induction. Landsc. Ecol. 16 (5), 377–390.


Stroh, J.C., Archer, S.R., Wilding, L.P., Doolittle, J.A., 1993. Assessing the influence of subsoil heterogeneityon vegetation in the Rio Grande Plains of south Texas using electromagnetic induction and geographicalinformation system, College Station, TX, USA. The Station, 39–42.

Sudduth, K.A., Drummond, S.T., Kitchen, N.R., 2001. Accuracy issues in electromagnetic induction sensing ofsoil electrical conductivity for precision agriculture. Comput. Electron. Agric. 31, 239–264.

Sudduth, K.A., Kitchen, N.R., 1993. Electromagnetic induction sensing of claypan depth. ASAE Paper No. 931531.ASAE, St, Joseph, MI, USA.

Sudduth, K.A., Kitchen, N.R., Drummond, S.T., 1999. Soil conductivity sensing on claypan soils: comparison ofelectromagnetic induction and direct methods. In: Robert, P.C., Rust, R.H., Larson, W.E. (Eds.), Proceedingsof the Fourth International Conference on Precision Agriculture. St. Paul, MN, July 19–22, 1998. ASA-CSSA-SSSA, Madison, WI, USA, pp. 979–990.

Sudduth, K.A., Kitchen, N.R., Hughes, D.F., Drummond, S.T., 1995. Electromagnetic induction sensing as anindicator of productivity on claypan soils. In: Robert, P.C., Rust, R.H., Larson, W.E. (Eds.), Proceedings of theSecond International Conference on Site-specific Management for Agricultural Systems. ASA-CSSA-SSSA,Madison, WI, USA, pp. 671–681.

Telford, W.M., Geldart, L.P., Sheriff, R.E., Keys, D.A., 1976. Applied Geophysics. Cambridge University Press,Cambridge, UK.

Thompson, S.K., 1992. Sampling. Wiley, New York, NY, USA, 343 pp.Triantafilis, J., Ahmed, M.F., Odeh, I.O.A., 2002. Application of a mobile electromagnetic sensing system (MESS)

to assess cause and management of soil salinization in an irrigated cotton-growing field. Soil Use Manage. 18(4), 330–339.

Triantafilis, J., Huckel, A.I., Odeh, I.O.A., 2001. Comparison of statistical prediction methods for estimatingfield-scale clay content using different combinations of ancillary variables. Soil Sci. 166 (6), 415–427.

Triantafilis, J., Huckel, A.I., Odeh, I.O.A., 2003. Field-scale assessment of deep drainage risk. Irrig. Sci. 21 (4),183–192.

U.S. Salinity Laboratory Staff, 1954. Diagnosis and improvement of saline and alkali soils. USDA Handbook 60.U.S. Government Printing Office, Washington, DC, USA, pp. 1–160.

van der Lelij, A., 1983. Use of an electromagnetic induction instrument (type EM38) for mapping of soil salinity.Internal Report Research Branch, Water Resources Commission, NSW, Australia.

Vaughan, P.J., Lesch, S.M., Corwin, D.L., Cone, D.G., 1995. Water content effect on soil salinity prediction: ageostatistical study using cokriging. Soil Sci. Soc. Am. J. 59, 1146–1156.

Warrick, A.W., Zhang, R., El-Harris, M.K., Myers, D.E., 1988. Direct comparisons between kriging and otherinterpolators. In: Wierenga, P.J., Bachelet, D. (Eds.), Proceedings of the Conference on Validation of Flow andTransport Models for the Unsaturated Zone. New Mexico State University, Las Cruces, NM, USA.

Weber, D.D., Englund, E.J., 1992. Evaluation and comparison of spatial interpolators. Math. Geol. 24, 381–391.Williams, B.G., Baker, G.C., 1982. An electromagnetic induction technique for reconnaissance surveys of soil

salinity hazards. Aust. J. Soil Res. 20, 107–118.Williams, B.G., Hoey, D., 1987. The use of electromagnetic induction to detect the spatial variability of the salt

and clay contents of soils. Aust. J. Soil Res. 25, 21–27.Wilson, R.C., Freeland, R.S., Wilkerson, J.B., Yoder, R.E., 2002. Imaging the lateral migration of subsurface

moisture using electromagnetic induction. ASAE Paper No. 023070. ASAE, St. Joseph, MI, USA.Wollenhaupt, N.C., Richardson, J.L., Foss, J.E., Doll, E.C., 1986. A rapid method for estimating weighted soil

salinity from apparent soil electrical conductivity measured with an aboveground electromagnetic inductionmeter. Can. J. Soil Sci. 66, 315–321.

Wollenhaupt, N.C., Wolkowski, R.P., Clayton, M.K., 1994. Mapping soil test phosphorus and potassium forvariable-rate fertilizer application. J. Prod. Agric. 7, 441–448.

characterizing soil spatial variability with apparent soil ...€¦ · characterizing soil spatial...

Documents