Mobile Electromagnetic Induction Research Articles

Precision Geolocation of Active Electromagnetic Sensors Using Stationary Magnetic Sensors

Unexploded ordnance (UXO) detection and discrimination is currently performed using mobile electromagnetic induction (EMI) technologies, combined with sophisticated analysis algorithms. Accurate position information for the collected EMI data is essential for success in identifying UXO items. CSIRO Materials Science and Engineering and Sky Research have investigated the feasibility of using the magnetic field produced by an EMI-based UXO detection cart system as a means to accurately locate the position and optionally the tilt of the cart. Using a full tensor gradiometer constructed from fluxgate vector magnetometers; the gradient produced by a cart within 10 m of the instrument was estimated. Three techniques were implemented and compared for determining the position of the cart based on the gradient estimate. We were able to locate the position of the cart to within 0.20 m of its true location at distances of up to 9 m. Accurate tilt information was also demonstrated. We were limited by a noisy acquisition system, and believe improvements could be made to increase the accuracy of the positioning system.

Validation of the ANOCOVA model for regional‐scale ECa to ECe calibration

AbstractTwo approaches have emerged as the preferred means for assessing salinity at regional scale: (i) vegetative indices from satellite imagery (e.g., MODIS enhanced vegetative index, NDVI) and (ii) analysis of covariance (ANOCOVA) calibration of apparent soil electrical conductivity (ECa) to salinity. The later approach is most recent and least extensively validated. It is the objective of this study to provide extensive validation of the ANOCOVA approach. The validation comprised 77 fields in California's Coachella Valley, ranging from 1.25 to 30.0 ha in size with an average size of 12.8 ha. Mobile electromagnetic induction (EMI) equipment surveyed the fields obtaining geospatial measurements of ECa. Soil sample sites selected following ECa‐directed soil sampling protocols characterized the range and spatial variation in ECa across the field. From the data, a regional ANOCOVA model was developed. The regional ANOCOVA model successfully reduced cross‐validated, average log salinity prediction error (variance) estimate by more than 30% across the 77 fields and improved the depth‐averaged prediction accuracy in 58 of the 77 fields. The results show that the ANOCOVA modelling approach improves soil salinity predictions from EMI signal data in most of the surveys conducted, particularly fields where only a limited number of calibration sampling locations were available. The establishment of ANOCOVA models at each depth increment for a representative set of fields within a regional‐scale study area provides slope coefficients applicable to all future fields within the region, significantly reducing ground‐truth soil samples at future fields.

Apparent Soil Electrical Conductivity Used to Determine Soil Phosphorus Variability in Poultry Litter-Amended Pastures

The objectives of this research were to determine the relationship between soil apparent electrical conductivity (ECa) and soil P distribution, and to compare the effectiveness of noncontact mobile electromagnetic induction (EM) and direct contact methods for relating ECa to soil P. Studies were conducted at two locations in Southwest Missouri on a longterm forage fertility plot site and three 1 to 1.5 ha sites within beef cattle pasture fields, all having received long-term poultry litter applications. For the long-term plot site, both the direct contact ECa sensor deep reading and the EM-38 (Geonics) sensor in the shallow mode had significant positive correlations with soil test Bray-1 P at both the 0 to 5 and 5 to 15 cm sampling depths. Significant spatial variation in soluble, soil test Bray-1 and total P were observed by landscape position within pasture fields. In general, soil ECa was not significantly correlated with soluble, soil test Bray-1 and total P at each individual pasture site, but when data was combined over all three sites, significant relationships were observed between ECa measured by the EM-38 sensor and soil soluble P, soil test Bray-1 P and total P, especially when the vertical (deep) mode was used. The difference in performance of the two sensors between the two studies was attributed to the proportion of Research Article American Journal of Experimental Agriculture, 3(1): 124-141, 2013 125 coarse fragments contained in the soils and soil water content. These results suggest that soil ECa measurements may provide some useful information for evaluating spatial variation in soil P due to manure applications. However, further research is needed to assess the processes and factors affecting this relationship before it can be recommended for use for improved soil P management in individual farm fields with varying environmental conditions and management practices.

Reconstructing palaeochannel morphology with a mobile multicoil electromagnetic induction sensor

Field methods to map and reconstruct the morphology of buried river systems are highly dependent on spatial interpolation. Conventional methods, such as standard borehole survey, allow a detailed vertical reconstruction of the shallow subsurface but leave lateral connections between sample locations open to interpretation. Geophysical survey techniques have recently introduced more detail. Mobile electromagnetic induction (EMI) survey combines high density sampling with full lateral coverage but fails to produce detailed information about vertical facies changes. Recently, multicoil EMI survey added vertical discrimination potential to this lateral continuity. In this study, we present an integrated approach for reconstructing the morphology of a known palaeochannel segment by modelling the depth to the sandy substrate. In addition, a calibration method based on a limited number of auger data is proposed. In a first phase, the modelling procedure was evaluated along two transects on a test site, showing palaeochannel depths ranging from 1 to >4m beneath the surface. In a second phase, the morphology of the entire site was reconstructed. These three resulting depth models were then compared with auger observations and electrical resistivity tomography (ERT) data. The high correlation coefficients (>0.9) between observed and modelled depths showed that even in complex pedological environments, palaeochannel morphology could be predicted precisely using multicoil EMI data. Therefore, we concluded that a multicoil EMI survey proves to be an efficient and reliable solution for mapping and reconstructing the morphology of the shallow subsurface.

Quantitative evaluation of soil salinity and its spatial distribution using electromagnetic induction method

In the Lower Yellow River Delta, soil salinity is a problem due to the presence of a shallow, saline water table and marine sediments. Spatial information on soil salinity at the field level is increasingly needed, particularly for better soil management and crop allocation in this area. In this paper, a mobile electromagnetic induction (EMI) system including EM38 and EM31 is employed to perform field electromagnetic (EM) survey, and fast determination and quantitative evaluation of the spatial pattern of soil salinity is discussed using the field EM survey data. Optimal operation modes of EM38 and EM31 are determined to establish multiple linear regression models for estimating salinity from apparent soil electrical conductivity (ECa). Spatial trend and semivariogram are illustrated and spatial distribution of field salinity status is further visualized and quantitatified. The results suggest that ECa (EM38 and EM31) data is highly correlated with salinity, and that the interpretation precision of soil salinity at various layers can be improved using EM38h and EM31h (where h represents the horizontal mode of EM measurement). Both EM38h and EM31h exhibit significant geographic trend. Nested spherical models fit the semivariance of EM38h and EM31h better than single spherical models. Spatial autocorrelation of EM31h is stronger than that of EM38h, and short-range variation is the chief constitute of spatial heterogeneity for both EM38h and EM31h. Quantitative classification shows that soil salinity exhibits the trend of accumulation in the root zone. In 0–1.0 m solum, heavy salinized and saline soils are the predominant soil types, accounting for 54% and 41% of total survey area, respectively. The area of light and moderate salinized soils is comparatively small, which accounts for only 0.4% and 4.6%, respectively.

Within-field spatial distribution of earthworm populations related to species interactions and soil apparent electrical conductivity

Within-field spatial distribution of ecosystem engineers such as earthworms determine the spatial patterns of important ecosystem processes at the field scale. But the driving factors that shape the within-field spatial variability of earthworm populations remain largely unclear. The aim of this study was to describe the earthworm distribution patterns in a tilled arable field and to explain earthworm spatial variability as a function of biotic interactions within populations and of abiotic soil heterogeneity measured as the soil apparent electrical conductivity (EC a). Earthworms were sampled at 100 locations within an area of (105 × 75) m 2 in a harvested wheat field on a loess soil in central Belgium. The soil EC a was measured using a mobile electromagnetic induction (EMI) sensor as a proxy for soil textural composition. Maps of earthworm density and soil EC a were produced by variogram modeling and ordinary kriging. Two approaches were followed in the data analysis: (i) a pixel-by-pixel comparison of the spatial patterns based on categorical maps derived from a fuzzy k-means clustering, and (ii) causal modeling based on point-by-point Mantel tests. The endogeic species Aporrectodea caliginosa and A. rosea inhabited similarly sized and overlapping patches, which were neither related to the spatial occurrence of the deep-burrowing species Lumbricus terrestris and A. longa, nor to the measured soil EC a variability. Endogeic adults and juveniles lived closely associated in the same spatial clusters. The segregated field distributions of both deep-burrowing species were largely determined by the subsoil textural properties (as measured by EC a) and not by competition. A. longa individuals lived in field areas with high ECav values (related to relatively higher clay content) while L. terrestris juveniles occupied regions with low ECav values. Anecic juveniles were found in larger and spatially differing clusters than adults, suggesting the dispersal of juveniles from parental clusters into neighbouring areas. L. terrestris adults were spatially organized in distinct patches of ∼15 m diameter and it is hypothesized that the particular mating behaviour of this species requires such intimate distributions. The rapid, easy and non-invasive geo-referenced soil characterization by means of EMI-based measurements proved to be a useful tool for determining and understanding the within-field spatial distributions of earthworms but requires further testing in a variety of (agro-)ecosystems.

Characterizing soil spatial variability with apparent soil electrical conductivity: Part II. Case study

Geospatial measurements of apparent soil electrical conductivity (EC a) are recognized as a means of characterizing soil spatial variability at field and landscape scales. However, inconsistencies in the measurement and interpretation of field- and landscape-scale geospatial EC a measurements have resulted in data sets that are unreliable and/or incompatible. These inconsistencies are, in part, a consequence of the lack of EC a-survey protocols that provide standardized guidelines to assure reliability, consistency, and compatibility. It is the objective of this paper to apply EC a-survey protocols to a soil quality assessment to demonstrate their utility in characterizing spatial variability. The soil quality assessment was conducted on a 32.4-ha field on the westside of central California's San Joaquin Valley where a mobile electromagnetic induction (EM) survey was performed following outlined protocols. The EM survey consisted of EC a measurements taken at 22,177 locations in April 2002. A response-surface sampling design was used to identify 40 sites where soil-core samples were taken at 0.3-m increments to a depth of 1.2 m. Duplicate samples were taken at eight sites to evaluate the local-scale variability. Soil samples were analyzed for a variety of physico-chemical properties associated with soil quality for an arid zone soil. Analysis characterized the soil as montmorillonitic, saline, and sodic with EC e (electrical conductivity of the saturation extract) varying from 4.83 to 45.3 dS m −1, SAR (sodium adsorption ratio) from 5.62 to 103.12, and clay content from 2.5 to 48.3%. Spatial trends showed high areas of salinity and SAR in the center of the southern half of the study area. Strong correlation was obtained between EC a and the soil properties of the saturation extract (EC e; Cl −, HCO 3 −, SO 4 2−, Na +, K +, and Mg 2+), exchangeable Na +, and SAR. Other properties were poorly correlated, including: volumetric water content ( θ v), bulk density ( D b), percent clay (% clay), saturation percentage (SP), exchangeable sodium percentage (ESP), Mo, CaCO 3, gypsum, total N, Ca 2+ in the saturation extract, and exchangeable cations (K +, Ca 2+, and Mg 2+). The spatial distribution of the poorly correlated properties is not as well represented with a response-surface sampling design suggesting the need for a complementary stratified random sample design.

Application of the theory of optimal experiments to adaptive electromagnetic-induction sensing of buried targets

A mobile electromagnetic-induction (EMI) sensor is considered for detection and characterization of buried conducting and/or ferrous targets. The sensor may be placed on a robot and, here, we consider design of an optimal adaptive-search strategy. A frequency-dependent magnetic-dipole model is used to characterize the target at EMI frequencies. The goal of the search is accurate characterization of the dipole-model parameters, denoted bythe vector theta; the target position and orientation are a subset of theta. The sensor position and operating frequency are denoted by the parameter vector p and a measurement is represented by the pair (p, O), where O denotes the observed data. The parametersp are fixed for a given measurement, but, in the context of a sequence of measurements p may be changed adaptively. In a locally optimal sequence of measurements, we desire the optimal sensor parameters, P(N+1) for estimation of theta, based on the previous measurements (p(n), On)n=1,N. The search strategy is based on the theory of optimal experiments, as discussed in detail and demonstrated via several numerical examples.