HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Abstract Figures Only Full Text (PDF) Free Alert me when this article is cited Alert me if a correction is posted Services Similar articles in this journal Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via Google Scholar Google Scholar Articles by Nogués, J. Articles by Herrero, J. Search for Related Content PubMed Articles by Nogués, J. Articles by Herrero, J. GeoRef GeoRef Citation Agricola Articles by Nogués, J. Articles by Herrero, J. Related Collections Electromagnetic Induction, EMI Soil Classification and Mapping Irrigation

Pedology

Incorporating Electromagnetic Induction Methods into Regional Soil Salinity Survey of Irrigation Districts

^a Soils and Irrigation Dep., Agri-Research Center of Aragon, PO Box 727, 50080 Zaragoza, Spain
^b Dep. of Geophysics, Stanford Univ., 397 Panama Mall, Stanford CA 94305-2215

^* Corresponding author (jhi{at}aragon.es)

	ABSTRACT

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS AND DISCUSSION CONCLUSIONS REFERENCES

 
Many land use decisions are made without the aid of soil maps in countries without National soil surveys. The purpose of this research was to create a soil survey delineating salinity phases at a regional scale that could serve as the basis for improved decision making, or highlight areas for further investigation. Soil salinity survey at a regional scale remains a challenge. We demonstrate a survey methodology suited to small fields, often found in Europe and developing countries, and present a streamlined survey methodology making this typically labor-intensive process more efficient. A regional soil survey was conducted at a scale of 1:25 000, mapping consociations and complexes of soil series. The soil survey information was then used to guide an electromagnetic induction (EMI) survey of the soils with observed salinity, to delineate phases. A calibration method for estimating the electrical conductivity of the extract of the saturated paste (ECe) from the EMI measurements was devised. After testing several statistical methods, a single calibration of the EMI data was found to be suitable using a nonparametric regression. The use of the EMI method reduced operation costs, where time, personnel, analysis time, and budget are limitations.

Abbreviations: EMI, electromagnetic induction • ECe, electrical conductivity of the extract of the saturated paste

	INTRODUCTION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS AND DISCUSSION CONCLUSIONS REFERENCES

 
SOILS INFORMATION with enough regional scale cartographic detail is useful in decision making and planning. Unlike the USA and some other countries, Spain has no soil survey to aid in decision making. Soil salinity is an endemic problem in Spain, and severe environmental and economic damage may occur if it is not considered in land use planning. There is therefore a need for a soil survey that provides a comprehensive taxonomy of the soils, but that is flexible enough to incorporate regional scale properties of specific interest to planning. We chose the U.S. Soil Taxonomy (Soil Survey Staff, 1999) as the soil survey framework. Although this produces a map legend that indicates the presence of salinity, it provides little hard data on the level of that salinity. Soil survey procedures are flexible in allowing the assignment of salinity phases (Soil Survey Division Staff, 1993); however, no universally accepted sampling methods have been developed for assigning these salinity phases at the regional scale. In the USA, where fields are large (>20–40 ha), efforts have focused on salinity mapping at the field or management unit scale (Rhoades et al., 1999). Attempts have also been made to consider taxonomy of sodic soils as a function of landform position (Seelig et al., 1990). A method is needed that is rapid and that will provide an integrated salinity measurement that is representative of the pedon scale.

Collecting soil samples is time-consuming and expensive; as a consequence specialized statistical sampling procedures have been developed to minimize sample size. Software, such as ESAP (Lesch et al., 2000), is often used to calibrate georeferenced EMI measurements to ECe, water content, and texture; producing georeferenced salinity maps through directed sampling designs. This is a cost effective precision agriculture procedure in large fields, under a single management strategy, where 2- to 30-ha fields can be surveyed in 1 d and calibrated with as few as 12 samples. Mechanized salinity mapping equipment allows rapid, time saving survey, of large fields. However, this procedure becomes impractical for soil survey in the small fields of Europe and some developing countries, where field sizes are often <1 ha and managed differently. In many irrigated areas with these characteristics, there is a need for low cost regional information rather than detailed field specific information, so that the dominant processes and regional salinization patterns can be identified, and strategic regional policy decisions developed.

The concept of using induced electromagnetic fields to measure ground conductivity has been applied in the geosciences for more than 50 yr (Belluigi, 1948; Wait, 1954, 1955, 1982). Electromagnetic induction was applied to agriculture in the late 1970s (De Jong et al., 1979). Since then EMI has been used to map a variety of physical properties that correlated with EMI measurements (e.g., salinity, moisture, and clay contents). Applications of EMI sensors to salinity mapping include Cameron et al. (1981), Williams and Baker (1982), Rhoades et al. (1989), and Triantafilis et al. (2000, 2002). The development of EMI sensors and their integration with GPS have revolutionized salinity mapping (Rhoades et al., 1999).

Electromagnetic induction has proved particularly popular in estimating soil salinity because it is non-invasive, offers instantaneous readings, integrates measurements over a large volume, and can measure more than one depth depending on the orientation of the coils with respect to the ground. The reading obtained with an EMI meter increases with increasing clay, moisture, and soluble salt content in the soil. In areas with saline soils, salinity can account for 65% of the variations in the sensor readings (Williams and Baker, 1982).

An EM38 meter produced by Geonics Ltd. (Ontario, Canada) was used in this study. A transmitting coil in one end of the EMI meter creates a primary magnetic field. This field creates current loops in the ground. The current loops induce a secondary magnetic field. This induced magnetic field is superimposed on the primary magnetic field and measured in a receiving coil at the other end of the instrument (McNeill, 1980). The measured response is a function of ground conductivity, and is linear under conditions of low induction numbers (<10 dS m^–1). In the horizontal dipole orientation the EM38 meter measures to a depth of about 0.75 m with the greatest sensitivity at the soil surface. In the vertical dipole orientation measurements are to a depth of about 1.5 m, with the greatest sensitivity at a depth about 0.4 m. The difference between the two can be used to infer the source of the salts.

The objective of this study was to combine traditional soil survey methods with EMI salinity mapping according to the following procedure:1. Map the region using traditional survey methods at the scale of 1:25000. 2. Use this to select soils with evidence of, or susceptible to, salinization. 3. Map only those soils with the EM38 meter to delineate salinity phases.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS AND DISCUSSION CONCLUSIONS REFERENCES

 
Study Area
The study was performed in the municipalities of Barbués and Torres de Barbués (Aragón, Spain), with an extent of 34 km² (Fig. 1 ). These municipalities include 281 ha of rain-fed farmland spread among small enclaves, plus 208 ha of non-agricultural lands. There are 2877 ha of irrigated land, a few hundred of which have been irrigated since medieval times (Bolea, 1986) on both banks of the Flumen River. The majority of this land was put under irrigation gradually from 1940 until 1965. Elevation of the study area ranges from 320 to 390 m above sea level. The Flumen River drains the region, traversing it from North to South.

View larger version (89K): [in this window] [in a new window]   Fig. 1. Arrangement of the two municipalities, Barbués and Torres de Barbués, (dashed line) which constitute the study area, and disposition of the main irrigation canals and ditches. Below, sample area of the ortophotograph (scale bar in meters) with soil delineations superimposed. Note the small compact field size.

The average weather conditions can be characterized using the Grañén-Montesodeto weather station. The mean annual temperature is 14.3°C, and the mean annual precipitation is 525 mm. The mean annual potential evapotranspiration is 1304 mm (Faci and Martínez-Cob, 1991). The soil temperature regime is thermic (Soil Survey Staff, 1999), and the soil moisture regime is aridic or xeric, depending on the available water holding capacity of the soils (Jarauta, 1989).

Irrigation water is of good quality for crops (EC < 0.4 dS m^–1, SAR < 1). High evaporative demand and saline soil parent materials contribute to widespread soil salinity in bottoms and on slopes (Nogués et al., 2000). Most irrigated fields are quadrangular and <1 ha in size (Fig. 1). These fields are cropped with alfalfa, barley, maize, rice, sunflower or wheat (Casterad and Herrero, 1998). Rice is grown in depressional and salt-affected areas, where the paddies are continuously flooded by flowing water during the growing season. The other crops are irrigated by basin and border flooding. The irrigation scheme, designed in the 1940s, required extensive leveling and terracing to have plots adequate for this kind of irrigation; now most of these fields are laser leveled every 2 yr. An extensive network of pipes and ditches allows drainage to the Flumen River.

Soil Mapping
Soils were mapped at a scale of 1:25 000 (Nogués, 2002). A total of 27 tentative soil series were established and classified to the family level using Soil Taxonomy (Soil Survey Staff, 1999), with phases (Soil Survey Division Staff, 1993) based on texture differences in surface layer, slope, and salinity. Soil profiles were described and horizons sampled for laboratory analysis in 40 pits distributed over the four geomorphic units. The pits were dug with a backhoe up to 2 m deep or to the lithic or paralithic contact.

During soil mapping holes were drilled using an Edelman auger (Eijelkamp Agrisearch Equipment, Giesbeek, the Netherlands) to a depth 1.5 m or to a lithic or paralithic contact at 217 points distributed among all the geomorphic units. Samples were not collected for laboratory analyses but were studied for their color, texture (estimated in the field by feel), effervescence of soil to HCl, accumulations (mainly gypsum or calcium carbonate), stains by redox phenomena, water table level, and root-limiting features. The soil survey data were combined with the geomorphic units so as to draw provisional map units on a 1:10 000 scale enlargement of the aerial photographs used in soil mapping. The soil map units consist of consociations and complexes and were named according to the guidelines of Van Wambeke and Forbes (1989). All maps obtained in this work are in vector format, overlaid on orthophotographs and linked with associated databases.

EMI Sensor and Soil Salinity Survey
EMI has been used extensively in the Flumen irrigation district to map soil salinity (Díaz and Herrero, 1992; López-Bruna and Herrero, 1996; Lesch et al., 1998; Herrero et al., 2003). In these studies, simple regressions of EM38 measurements with ECe of soil samples at a reduced number of points were developed rather than multiple calibrations for various parcels of land. To minimize the effects of moisture content variations, EMI surveys were completed on relatively dry soils before irrigation, constraining the time in which the survey could be conducted. The ECe salinity values in the soils were expected to be as much as 20 dS m^–1, over riding any textural effects that might affect calibration.

Surveys attempt to obtain a balanced distribution of sample points throughout the EMI measurement range and for the region studied. It is often necessary to decide at which measurement points to extract samples based on the sensor readings. This requires one to complete a set of readings and then determine the sample points, using, for example the ESAP software (Lesch et al., 2000), and then return to those points to take the samples. This procedure is impractical for an area the size of this study, with small and numerous individually managed parcels, and with differing irrigation timing.

The sample strategy of García and Virgili (1996) was adopted to establish phases of soil salinity (Nogués, 2002) by integrating EMI methods with traditional soil survey methods. In this survey, soil taxonomic and geomorphic units were used to direct the EMI surveying. This strategy is suited to cases where a wide variety of soils are to be mapped by a single soil surveyor, over an extended period and under different management conditions.

The EMI survey was conducted in field plots that were considered to be representative of provisional soil delineation. Five points were then selected within this plot, one point situated at the center of the plot, and the others at the four corners, 2 m from the borders. At each point electromagnetic sensor readings were made in both the horizontal and vertical dipole orientations. A correction factor was applied to the readings according to the soil temperature at a depth of 50 cm, so as to reference them to 25°C. After dividing them by 100 to facilitate the calibration equations, the two corrected readings were termed EMh and EMv, respectively.

From 10 April through 1 May 1998, 48 field plots were selected, and 240 points were measured with the EM38 meter. At 44 of the measurement points, auger samples were extracted from the soil to correlate the ECe with the EM38 response. At each sample point located in the bottoms two soil samples were taken, one from 0- to 60-cm depth and the other from the 60- to 120-cm depth. On the slopes only one sample up to 60 cm was made at each sample point, due to the shallow soil constraint.

In consideration of the lack of information about the spatial dependence of soil salinity, the 44 measurement points were randomly chosen (Halvorson et al., 1997). During the course of sampling, adjustments were made to cover all the different visual salinity degrees, and to sample the whole range of EMI measurements. This procedure resulted in 68 soil samples taken for determination of the saturated paste extract electrical conductivity (ECe in dS m^–1 at 25°C) in the laboratory with standard methods (United States Salinity Laboratory Staff, 1954). This effort resulted in 65 EMh and EMv measurements within the 7 soils delineated as slopes, with 20 soil samples collected; 135 EMI measurements in the 25 soils delineated as bottoms without gypsum with 28 soil samples from 14 points; and 40 measurements made in the 7 soils delineated as bottoms with gypsum that included 20 soil samples from 10 sampling points.

The EM38 was calibrated with ECe laboratory determinations split into four distinct soil sample groupings, reflecting four facets of soil salinity: that is, salinity of the surface layer, of the subsurface layer, of both layers, and agronomically limiting salinity. These calibrations correspond to the ECe of the samples for: (i) 0- to 60-cm depth; (ii) 60- to 120-cm depth; (iii) 0- to 120-cm depth; and (iv) ECemax, which is the larger ECe if two extracted samples were taken at a location. This last calibration reflects the soil salinity impact on agricultural practice, in that the ECemax is the key limiting factor for crop development at a certain time in the crop cycle.

Regression and Data Analysis Procedures
Several linear regression procedures, including parametric and nonparametric, were tested. After an exploratory analysis of the data and a study of the normality using the Kolmogorov–Smirnov test, the straight-line regression parameters were determined of the forms ECe = a + b EMh, and ECe = a + b EMv. The hypotheses of the simple linear regression model are checked by studying the residuals.

Since the hypotheses of the simple linear regression model were not satisfied by the data, Box-Cox transformations were applied (Myers, 1990). These transformations try to obtain normality when the larger values have greater variance than the smaller ones, and when the distribution is unknown. The best transformation is searched for by using a power or a logarithm of the dependent variable, with:

[1]

where w is the transformed variable, y is the original dependent variable, and is the optimum constant for the transformation. The value of is selected automatically using the Minitab (Minitab Inc., State College PA, USA) software package.

The Box-Cox transformations did not produce satisfactory results. As a consequence, a nonparametric simple regression was tested, using the Theil Method (Daniel, 1990); this estimates the slope coefficient of a regression line. Here, all the data pairs are taken in order from small to large for the dependent variable, from which one obtains the slopes (B_ij) for each pair, that is:

[2]

For each regression the values of B_ij are calculated and the median of the B_ij values is taken as the slope of the straight-line regression. The different calibration modes for the EM38 readings, that is, by geomorphic calibration zones, by soil sample groupings, or by the parametric or nonparametric method, were used to assign the soil delineations to salinity phases. ECmax was selected as best representing the agronomically limiting salinity. Summing the surface area of the delineations of each phase, one obtains salinity estimations which vary somewhat according the calibration mode. The distributions of the EM38 readings in the three geomorphic calibration zones were compared using boxplots, with the conventions of Chambers et al. (1983, p. 21). The confidence interval for the median is calculated using nonlinear interpolation (Hettmansperger and Sheather, 1986).

	RESULTS AND DISCUSSION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS AND DISCUSSION CONCLUSIONS REFERENCES

 
Soil Survey
The taxonomic classification of the 27 soil series established by Nogués (2002) are listed in Table 1, with their extent. These soil series are considered provisional, as they have not been officially correlated. The study area is on the border between the aridic soil moisture regime of the central Ebro valley with the xeric regime of the fringe transition to the Pyrenees. Following the model of Jarauta (1989), soils with AWHC 50 mm are considered to have an aridic soil moisture regime (Rodríguez et al., 1989). Other soils have xeric soil moisture regime. Soils had ochric epipedons and calcic, petrocalcic or gypsic subsurface diagnostic horizons. Subsurface horizons of Xerofluvents and Xerorthents contained some reduction or stagnation features, often associated with the low river terraces or to paddies, but did not meet the requirements for the Aquic subgroup. These were classified into the Oxyaquic great group. A shallow petrocalcic horizon occurs on a platform topped by gravels with the soil classified as a Petrocalcic Calcixerept. The occurrence of a calcic horizon leads to Calcixerepts, which cannot be split at the Subgroup level.

Soil salinity, as a relevant genetic, agricultural, and environmental feature, is depicted in the map by using salinity phases, whose delineation was improved by the EMI survey. Salt-affected soils included the Callén, Cordel, Francal, Osplanos, and Valfonda soils.

Salinity Mapping
The EMI survey area was partitioned into geomorphic units. Geomorphic units have been used previously in the USA (Seelig et al., 1990) in an attempt to delineate salt-affected soils. Of the four basic geomorphic units recognized by Artieda (2002) for this area, two are non-saline (platforms, and fluvial terraces), and two are either saline or salinity-prone (bottoms, and slopes). Therefore, soil salinity was only surveyed in the valley bottoms and on slopes, but not on the platforms or the fluvial terraces; the geomorphic units lacking visual evidence of salinity or a risk of salinization. Considering that the soil depth and the presence of gypsum could affect the response of the electromagnetic sensor, three calibration zones were delineated, based on the geomorphic units. These geomorphic calibration zones were: (i) slopes, 161 ha, where the most represented soils are Escalerón and Salagones; (ii) valley bottoms with gypsum, 289 ha, mainly Valfonda soil; and (iii) valley bottoms without gypsum, 577 ha, where the main soils are Callén, Cordel, Francal, and Osplanos. These geomorphic calibration zones are presented in Fig. 2 . Where present, gypsum in these soils comes from outcrops of gyprock strata in the upper portions of some valleys, outside the study region. Providing that the source of gypsum is outside the study area, we qualify the whole geomorphic delineation as bottom with gypsum, when gypsum was found in the field in some part of the unit.

View larger version (82K): [in this window] [in a new window]   Fig. 2. Calibration zones for the EM38. White areas are nonsaline, and were not surveyed with EM38.

View larger version (8K): [in this window] [in a new window]   Fig. 3. Frequency histograms for the 240 electromagnetic sensor readings for both orientations (EMh and EMv), and for the 44 ECemax determined in laboratory.

View larger version (12K): [in this window] [in a new window]   Fig. 4. Boxplots of EMh and EMv in the 240 reading points and in the 44 reading points with drilling (left hand plots), and boxplots of four different samples' groupings: 0- to 60-cm, 60- to 120-cm, 0- to 120-cm, and the sample of highest ECe in each drilling, that is, ECemax (right hand plots).

[3]

and

[4]

Using the ECemax estimated with the EM38 meter, we assigned salinity phases to each map unit in the EMI surveyed area. To do this, readings were converted to ECe selecting between the regressions of ECemax on EMh or on EMv, whichever gave the greater coefficient of determination, R². When the salinity phases were established, the ECe values determined directly from soil samples were not taken into account, provided that the soil salinity of the soil delineation is attributed to the sampled field, with five EM38-measured points, but not to any individual drilled point (which could represent extreme values within the field). To group the estimated ECemax and thereby establish the salinity phase, the ECe class intervals from Soil Survey Division Staff (1993) were used with modifications in the class names (Table 2). Occasionally, this generated new map units when the initial units were split by phase of soil salinity.

View larger version (9K): [in this window] [in a new window]   Fig. 5. Plot of ECemax (maximum ECe determined in each drilling) versus the EMh readings (circle) and EMv readings (cross).

In the same way for each calibration zone as for the three together, the slopes of the simple linear regressions (Table 3) varied in a coherent way when separate adjustments were made for the 60-cm surface depth, the 60- to 120-cm depth layer, the 0- to 120-cm depth, and the ECemax. This permits us to estimate salinity for different soil depths, in the present article we approach salinity as an agronomic limiter with a precautionary focus, and then we establish the salinity phases based on ECemax.

Estimate of the Surface Areas of Soils of Distinct Salinity
In this study the calibration equations were used to assign salinity phases, defined (Table 2) by unequal ECe intervals (Soil Survey Division Staff, 1993), to soil delineations whose sizes are heterogeneous, giving surface area estimations that are numerically discontinuous and at irregular intervals.

Figure 6 presents the estimated surface area of the salinity phases for the previously calculated regressions. The parametric cases (white and gray bars) show the surface areas obtained with regressions separated by calibration zone and with a single regression for the three zones together; the regression used is the one with the highest R². For the nonparametric case (Theil's method, black bars), estimated surface areas are based on the regression obtained for the three calibration zones together using EMv, since the slopes (b_T) for EMv are, on the average, closer to those of the parametric regressions than those of EMh (Table 3).

View larger version (29K): [in this window] [in a new window]   Fig. 6. Surface area of the soil salinity phases in Barbués and Torres de Barbués using the five intervals of the Soil Survey Division Staff (1993). The graphs correspond to: each calibration zone (slopes, bottoms without gypsum, bottoms with gypsum), the sum of the three calibration zones, and the whole study area, that is, including also the platforms and fluvial terraces, non surveyed with EMI. Bars show the areas obtained from sensor calibrations that use separate regressions for each of the three calibration zones under consideration (white bars), and a single calibration for the three zones with either parametric method (gray bars) or nonparametric method (black bars).

For the slopes calibration zone, the equation derived from EMh, with R² = 75.9, was used in the linear regression. The white bars of Fig. 6 show that 51% of the surface area of the slopes is in non-saline delineations, 34% is slightly saline, and 15% is moderately saline. In each of the plots in Fig. 6, gray bars represent the estimated surface areas using a single linear regression for all the EM38-surveyed area. With this method, the estimate of non-saline surface area for the slopes is 4% lower than that using a separate calibration, for slightly saline it is 18% lower, for moderately saline it is 5% higher, and for strongly saline it is 16% higher. The non-parametric regression using a single calibration for all the demarcation is shown in the black bars and is identical to the results obtained through the joint parametric calibration.

For the bottoms without gypsum, using parametric methods and calibrating separately for each calibration zone, the ECemax is estimated by a regression on EMv, with an R² = 86.9 (Table 3). Figure 6 displays the resulting surface areas: 19% non-saline, 14% slightly saline, 39% moderately saline and 28% strongly saline. The joint calibration produces (Fig. 6) an approximate increase of 4% in the non-saline surface area, an increase of slightly more than 2% for the slightly saline, and slightly more than 1% in the moderately saline, whereas a reduction of 7% of the estimate of strongly saline surface area. In contrast with these figures, the Theil Method, using a joint calibration and shown in the black bars reduces the estimated non-saline surface area by 9%, raises the slightly saline by 7% and the moderately saline surface area by 2%. The strongly saline surface area is unchanged against the separate calibration, but increases 7% if compared with the joint parametric calibration.

For bottoms with gypsum, the separated linear regressions produce (Table 3) the highest coefficient of determination for the estimation of ECemax using EMh, with an R² = 80.1. This equation, moving back to the white bars of Fig. 6, estimates the percentage of non-saline surface area to be 29%, the slightly saline as 16%, the moderately saline 3%, the strongly saline 17%, and the very strongly saline 35%. The joint calibration (gray bars of Fig. 6) reduces the first estimates for slightly saline surface area by 9% and the estimate of the moderately saline by 26% increments, reducing the estimate (0 ha) of the strongly saline surface by 17%. In contrast, Theil's Method using joint calibration (black bars of Fig. 6), reduces the non-saline surface estimate (0 ha) by 29%, increases that of slightly saline by 15%, and increases the moderately saline surface area by 14%. This estimate of the surface area of the very strongly saline soil did not change, as was the case previously.

The slopes achieved the smallest standard error (1.1 dS m^–1) for the linear regressions selected to estimate ECemax (Table 3). For the three zones jointly, an intermediate value is obtained (2.7 dS m^–1), where the repercussion when assigning salinity phases to individual soil demarcations should be studied case-by-case. The medians of the salinity estimates in each of the three calibration zones do not differ significantly. This becomes clear if in Fig. 7 one compares two-by-two the distributions in the calibration zones, for EMh and for EMv, given that these magnitudes are linearly related to soil salinity. If we add to this the lack of unidirectional bias (Fig. 6) when using one or several areas for calibrating the sensor, the result (Fig. 8 ) achieved with a single calibration for the three calibration zones can be accepted, with the added benefit of lower survey costs.

View larger version (8K): [in this window] [in a new window]   Fig. 7. Boxplots of the EM38 readings for horizontal and vertical dipole orientation (EMh and EMv) for each of the calibration zones.

View larger version (79K): [in this window] [in a new window]   Fig. 8. Salinity phases for soil mapping, adapted from Soil Survey Division Staff (1993), established with EM38 estimates of ECe for the most saline horizon of the profile (ECemax).

	CONCLUSIONS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS AND DISCUSSION CONCLUSIONS REFERENCES

 
The single-calibration EMI procedure is particularly suited to rapid soil survey. Our strategy for rapid soil diagnosis was based on the discrete units inherent in soil mapping. Using the EMI method with the traditional soil survey method to assign salinity phases is better suited to our study area than interpolation and kriging used by many geostatistical programs. Traditional soil survey ‘judgment sampling’ combined with modern tools such as the EM38 meter allow cost efficient collection of salinity data for improved decision making. This work demonstrates that there is still a strong need for the development of regional scale survey methods in areas with saline soils.

We found that the use of traditional soil survey methods combined with the identification of geomorphic units could effectively direct soil salinity surveying. The use of the EM38 meter increased the number of observations by five-fold over the survey area. Within the study area, correlation of EMI measurements with ECe using separate linear regressions for the three geomorphic calibration zones (slopes, bottoms with gypsum, and bottoms without gypsum) compared with a single lumped calibration, did not produce any considerable surface area estimation difference for any of the salinity phases. This relieves us of making separate sensor calibrations for the region studied, a significant cost saving. According to these estimates, 79% of the surface area under study is nonsaline or slightly saline, 10% is moderately saline, and 11% is strongly or very strongly saline. A nonparametric regression—the Theil method—for the whole demarcation gives negligible differences in the estimated surface area of each soil salinity phase. Since the distributions of the data do not meet the statistical requirements for a linear regression, we recommend the use of nonparametric regression methods.

	ACKNOWLEDGMENTS

 
The article is a result of the research project RTA2005–00230, funded by INIA (Spanish Government). The authors thank the anonymous referees for their helpful comments and suggestions.

Received for publication December 14, 2005.

	REFERENCES

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS AND DISCUSSION CONCLUSIONS REFERENCES

 

• Artieda, O. 2002. Marco geológico y geomorfología. p. 20–23 in Nogués, J. Mapa de suelos (E 1/25.000) de Barbués y Torres de Barbués (Huesca). Aplicaciones para modernización de regadíos. Consejo de Protección de la Naturaleza de Aragón. Zaragoza, Spain.
• Belluigi, A. 1948. Coupling of a vertical coil with a homogeneous earth. Geophysics 14:501–507.
• Bolea, J.A. 1986. Los riegos de Aragón, 2nd ed. Huesca, Spain.
• Cameron, D.R., E. De Jong, D.L.W. Read, and M. Oosterveld. 1981. Mapping salinity using resistivity and electromagnetic inductive techniques. Can. J. Soil Sci. 61:67–78.
• Casterad, M.A., and J. Herrero. 1998. Irrivol: A method to estimate the yearly and monthly water applied in an irrigation district. Water Resour. Res. 34:3045–3049.[CrossRef]
• Chambers, J.M., W.S. Cleveland, B. Kleiner, and P.A. Tukey. 1983. Graphical methods for data analysis. Chapman and Hall. New York.
• Daniel, W.W. 1990. Applied nonparametric statistics, 2nd ed. Pws-Kent Publishing Company. Boston, MA.
• De Jong, E., A.K. Ballantyne, D.R. Cameron, and D.W.L. Read. 1979. Measurement of apparent electrical conductivity of soils by an electromagnetic induction probe to aid salinity surveys. Soil Sci. Soc. Am. J. 43:810–812.[Web of Science]
• Díaz, L., and J. Herrero. 1992. Salinity estimates in irrigated soils using electromagnetic induction. Soil Sci. 154:151–157.
• Faci, J.M., and A. Martínez-Cob. 1991. Cálculo de la evapotranspiración de referencia en Aragón. Departamento de Agricultura. Gobierno de Aragón. Zaragoza, Spain.
• García, J.F., and J.M. Virgili. 1996. Utilización del sensor electromagnético en la cartografía detallada de suelos: Determinación de las fases salinas. Actas IV Congreso de la Sociedad Española de la Ciencia del Suelo. 231–236. Lleida, Spain.
• Halvorson, J.L., J.L. Smith, and R.I. Papendick. 1997. Issues of scale for evaluating soil quality. J. Soil Water Conserv. 52:26–30.
• Herrero, J., and O. Pérez-Coveta. 2005. Soil salinity changes over 24 years in a Mediterranean irrigated district. Geoderma 125:287–308.[CrossRef][Web of Science]
• Herrero, J., A.A. Ba, and R. Aragüés. 2003. Soil salinity and its distribution determined by soil sampling and electromagnetic techniques. Soil Use Manage. 19:119–126.
• Hettmansperger, T.P., and S.J. Sheather. 1986. Confidence intervals based on interpolated order statistics. Stat. Probab. Lett. 4:75–79.
• Jarauta, E. 1989. Modelos matemáticos de régimen de humedad del suelo. PhD Thesis, Universidad Politécnica de Barcelona, Spain.
• Lesch, S.M., J. Herrero, and J.D. Rhoades. 1998. Testing for changes in field salinity levels over time using electromagnetic induction techniques. Soil Sci. Soc. Am. J. 62:232–242.[Abstract/Free Full Text]
• Lesch, S.M., J.D. Rhoades, and D.L. Corwin. 2000. ESAP-95 version 2.01R. User manual and tutorial guide. Research report No. 146. USDA-ARS, G.E. Brown Salinity Laboratory. Riverside, CA.
• López-Bruna, D., and J. Herrero. 1996. El comportamiento del sensor electromagnético y su calibración frente a la salinidad edáfica. Agronomie 16:95–105.
• McNeill, J.D. 1980. Survey interpretation techniques EM38. TN-6 Geonics Ltd., Ontario, Canada.
• Myers, R.H. 1990. Classical and modern regression with applications. 2nd ed. Wadsworth Publishing Co., Belmont, CA.
• Nogués, J. 2002. Mapa de suelos (E 1/25.000) de Barbués y Torres de Barbués (Huesca). Aplicaciones para modernización de regadíos. Consejo de Protección de la Naturaleza de Aragón. Zaragoza, Spain.
• Nogués, J., J. Herrero, R. Rodríguez-Ochoa, and J. Boixadera. 2000. Land evaluation in a salt-affected irrigated district using an index of productive potential. Environ. Manage. 25:143–152.
• Rhoades, J.D., F. Chanduvi, and S.M. Lesch. 1999. Soil salinity assessment. Methods and interpretation of electrical conductivity measurements. FAO Irrigation and Drainage Paper 57. FAO, Rome.
• Rhoades, J.D., S.M. Lesch, P.J. Shouse, and W.J. Alves. 1989. New calibrations for determining soil electrical conductivity-depth relations from electromagnetic measurements. Soil Sci. Soc. Am. J. 53:74–79.[Web of Science]
• Rodríguez, R., J. Herrero, and J. Porta. 1989. Suelos de regadío con drenaje enterrado. XVI Reunión de la Sociedad Española de la Ciencia del Suelo. Lérida, Spain.
• Seelig, B.D., J.L. Richardson, and W.T. Barker. 1990. Characteristics of sodic soils as a function of landform position. Soil Sci. Soc. Am. J. 54:1690–1697.[Web of Science]
• Soil Survey Division Staff. 1993. Soil survey manual. Natural Resources Conservation Service, Handb. 18. USDA. Washington, DC.
• Soil Survey Staff. 1999. Soil Taxonomy. 2nd ed. Natural Resources Conservation Service, Agriculture Handbook 436. USDA. Washington, DC.
• Triantafilis, J., M.F. Ahmed, and I.O.A. Odeh. 2002. Application of a mobile electromagnetic sensing system (MESS) to assess cause and management of soil salinisation in an irrigated cotton-growing field. Soil Use Manage. 18:330–339.[CrossRef]
• Triantafilis, J., G.M. Laslett, and A.B. McBratney. 2000. Calibrating an electromagnetic induction instrument to measure salinity in soil under irrigated cotton. Soil Sci. Soc. Am. J. 64:1009–1017.[Abstract/Free Full Text]
• United States Salinity Laboratory Staff. 1954. Diagnosis and improvement of saline and alkali soils. Agriculture Handb. 60. USDA. Washington, DC.
• Van Wambeke, A., and T.R. Forbes. 1989. Guidelines for using Soil Taxonomy in the names of soil map units. Tech. Monogr. No. 10. USDA, Soil Management Support Services. Washington, DC.
• Wait, J.R. 1954. Mutual coupling of loops lying on the ground. Geophysics 19:290–296.[CrossRef]
• Wait, J.R. 1955. Mutual electromagnetic coupling of loops over a homogeneous ground. Geophysics 20:630–637.[CrossRef]
• Wait, J.R. 1982. Geo-electromagnetism. Academic Press, New York.
• Williams, B.G., and G.C. Baker. 1982. An electromagnetic induction technique for reconnaissance surveys of soil salinity hazards. Aust. J. Soil Res. 20:107–118.[CrossRef]

This Article Abstract Figures Only Full Text (PDF) Free Alert me when this article is cited Alert me if a correction is posted Services Similar articles in this journal Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via Google Scholar Google Scholar Articles by Nogués, J. Articles by Herrero, J. Search for Related Content PubMed Articles by Nogués, J. Articles by Herrero, J. GeoRef GeoRef Citation Agricola Articles by Nogués, J. Articles by Herrero, J. Related Collections Electromagnetic Induction, EMI Soil Classification and Mapping Irrigation