Methodology

• Introduction
• Factors Influencing the Penetration Depth of GPR
• GPR Soil Indices
• Relative Suitability of Soils for GPR
• Ground-Penetrating Radar Soil Suitability Maps
• References
• Contact the Author
• Tables

Introduction

Ground-penetrating radar (GPR) is widely used by a diverse group of service providers that include agronomist, archaeologists, criminologists, engineers, environmental specialists, foresters, geologists, geophysicists, hydrologists, land use managers, and soil scientists. In recent years, GPR has gained recognition in the search for terrorism and military hazards. A common concern of GPR service providers is whether or not GPR will be able to achieve the desired depth of penetration in the soils of a project area. In many soils, high rates of signal attenuation severely restrict penetration depths and limit the suitability of GPR for a large number of applications. In saline soils, where penetration depths are often less than 10 inches (Daniels, 2004), GPR is unsuited to most applications. In wet clays, where penetration depths are typically less than 40 inches (Doolittle et al., 2002), GPR has very low potentials for most applications. However, GPR is highly suited to most applications in dry sands and gravels, where penetration depths can exceed 160 feet with low frequency antennas (Smith and Jol, 1995).

Most GPR service providers have limited knowledge of soils and are unable to foretell attenuation rates, penetration depths, and the general suitability of the soils within project areas to GPR. Knowledge of the probable penetration depth and the relative suitability of soils would help service providers assess the appropriateness of using GPR and the likelihood of achieving acceptable results. Soil attribute data contained in the State Soil Geographic (STATSGO) and the Soil Survey Geographic (SSURGO) data bases have been used to develop thematic maps showing, at different scales and levels of resolution, the relative suitability of soils for many GPR applications. Both STATSGO and SSURGO data bases consist of digital map data, attribute data, and Federal Geographic Data Committee compliant metadata. These data bases are linked to soil interpretation records, which contain data on the physical and chemical properties of approximately 22,000 different soils (USDA-Natural Resources Conservation Service, 1994).

The STATSGO data base was developed by the USDA-NRCS and published in 1994 (USDA-Natural Resources Conservation Service, 1994). State Soil Geographic data are available for the conterminous United States, Alaska, Hawaii, and Puerto Rico. Because of the small compilation scale (1:250,000) of STATSGO maps, soil map units and polygons that appear on soil survey maps had to be combined and generalized. This procedure resulted in fewer soil map units and larger soil polygons. The STATSGO data base contains 9,555 unique map units and 78,507 polygons. The minimum polygon size is about 1,544 acres. The composition of each map units was coordinated so that the names and relative extent of each soil component would remain the same between survey areas and across political boundaries. In areas where detailed soil maps are not available, existing data were assembled, reviewed, and the most probable classification and extent of soils determined (USDA-Natural Resources Conservation Service, 1994).

Larger scale, less generalized maps, which show in greater detail the spatial distribution of soil properties that influence the penetration depth and effectiveness of GPR, are prepared using the SSURGO data base. The SSURGO data base (USDA-Natural Resources Conservation Service, 1995) contains the most detailed level of soil geographic data developed by the USDA-NRCS. Soil maps in the SSURGO data base duplicate the original soil survey maps, which were prepared using national standards and field methods at scales ranging from 1:12,000 to 1:63,360 (with minimum delineation size of about 1.5 to 40 acres, respectively) (Soil Survey Staff, 1993). Base maps are USGS 7.5-minute topographic quadrangles and 1:12,000 or 1:24,000 orthophotoquads.

Tabular and spatial SSURGO data are available through the Soil Data Mart (http://soildatamart.nrcs.usda.gov/). The USDA-NRCS is presently compiling and digitizing data from additional soil survey areas. Completion of the SSURGO data digitizing is scheduled for 2011. A status map showing the digitized soil survey areas can be accessed at http://www.soils.usda.gov/survey/geography/ssurgo/.
 

Factors Influencing the Penetration Depth of GPR

The penetration depth of GPR is determined by antenna frequency and the electrical conductivity of the earthen materials being profiled (Daniels, 2004). Soils having high electrical conductivity rapidly attenuate radar energy, restrict penetration depths, and severely limit the effectiveness of GPR. The electrical conductivity of soils increases with increasing water, clay and soluble salt contents.

Electrical conductivity is directly related to the amount, distribution, chemical composition, and phase (liquid, solid, or gas) of the soil water (McNeill, 1980). At a given frequency, the attenuation of electromagnetic energy increases with increasing moisture contents (Daniels, 2004). The lack of adequate data on soil moisture and the high spatial and temporal variations in the degree of soil wetness within most soil map units precluded the use of moisture content in the preparation of GPR soil suitability maps. As a consequence, properties selected to prepare these maps principally reflect variations in the clay and soluble salt contents of soils. These properties include clay content and mineralogy, electrical conductivity, sodium absorption ratio, and calcium carbonate and calcium sulfate contents.

Clays have greater surface areas and can hold more water than the silt and sand fractions at moderate and higher water tensions. Because of their high adsorptive capacity for water and exchangeable cations, clays produce high attenuation losses (Daniels, 2004). As a consequence, the penetration depth of GPR is inversely related to clay content. While soils with more than 35 percent clay are restrictive, soils with less than 10 percent clay are generally favorable to deep penetration with GPR.

Soils contain various proportions of different clay minerals (e.g., members of kaolin, mica, chlorite, vermiculite, smectite groups). The size, surface area, cation-exchange capacity (CEC), and water holding capacity of clay minerals vary greatly. Variations in electrical conductivity are attributed to differences in the CEC associated with different clay minerals (Saarenketo, 1998). Electrical conductivity increases with increasing CEC (Saarenketo, 1998). Soils with clay fractions dominated by high cation exchange capacity clays (e.g., smectitic and vermiculitic soil mineralogy classes) are more attenuating to GPR than soils with an equivalent percentage of low cation exchange capacity clays (e.g., kaolinitic, gibbsitic, and halloysitic soil mineralogy classes). Soils classified as kaolinitic, gibbsitic, and halloysitic characteristically have low cation-exchange capacity and low base saturation. As a general rule, for soils with comparable clay and moisture contents, greater depths of penetration can be achieved in highly weathered soils of tropical and subtropical regions that have kandic or oxic horizons than in soils of temperate regions that have argillic horizons. Compared with argillic horizons, kandic and oxic horizons have greater concentrations of low activity clays (Soil Survey Staff, 1999).

Electrical conductivity is directly related to the concentration of dissolved salts in the soil solution, as well as the type of exchangeable cations and the degree of dissociation of the salts on soil particles (Soil Survey Staff, 1993). The concentration of salts in the soil solution is dependent upon the degree of water-filled porosity, the soil texture, and the minerals found in soils. In semi-arid and arid regions, soluble salts and exchangeable sodium accumulate in the upper part of some soil profiles. These salts produce high attenuation losses that restrict penetration depths (Doolittle and Collins, 1995). Because of their high electrical conductivity, saline (saturated extract electrical conductivity ≥ 4 mmhos cm -1) and sodic (sodium absorption ratio ≥ 13) soils are considered unsuited to GPR.

Calcareous and gypsiferous soils are characterized by layers with secondary accumulations of calcium carbonate and calcium sulfate, respectively. These soils mainly occur in base-rich, alkaline environments in semi-arid and arid regions. High concentrations of calcium carbonate and/or calcium sulfate imply less intense leaching, prevalence of other soluble salts, greater quantities of inherited minerals from parent rock, and accumulations of specific mineral products of weathering (Jackson, 1959). Grant and Schultz (1994) observed a reduction in the depth of GPR signal penetration in soils that have high concentrations of calcium carbonate.
 

GPR Soil Indices

Soil properties selected to prepare GPR soil suitability maps are summarized in Table 1. Attribute index values (AIV) were assigned to each of the selected soil properties based on field experiences. Lower AIVs are associated with lower rates of signal attenuation, greater penetration depths, and soil properties that are characteristically more suited to GPR. For clay content, AIVs range from 1 to 5. Mineral soils with clay contents less than 10 percent in all horizons within a depth of 6.6 feet (2.0 m) have a very high potential for most GPR applications and are assigned an AIV of 1. Sandy soils with one or more finer textured layers between depths of 3.3 and 6.6 feet (1.0 to 2.0 m) have a lower potential for GPR and are assigned an AIV of 2. Mineral soils with clay contents of 18 to 35 percent, 35 to 60 percent, or greater than 60 percent in one or more horizons within a depth of 3.3 feet are assigned index values of 3, 4, and 5, respectively. A mineralogy override is used for highly weathered soils that are dominated by low activity clay minerals. Based on taxonomic criteria, a textural adjustment (-1) is applied to soils that have kandic or oxic horizons with between 10 and 60 percent clay. Fabric adjustments and separate indices are used for organic soils. Organic soils with more acidic reactions (dysic; AIV of 1) are typically more nutrient deficient and less limiting to GPR than organic soils with more neutral or alkaline soil reactions (euic; AIV of 2). Distinctions are also made for organic soils with mineral layers that are more than 12 inches (30 cm) thick and occur within depths of 4.1 feet (1.25 m; terric taxonomic subgroup). Because of their unsuitability to GPR, soils that are recognized as being saline or alkaline (sodic) are assigned an AIV of 6. Typically, these soils have adverse concentrations of soluble salts within the upper 20 inches (50 cm) of the soil profile. Soils having adverse concentrations of soluble salts below depths of 20 inches are assigned an added AIV of +1. Based principally on taxonomic criteria, mineral soils with less than 60 percent clay that are characterized by high gypsum and/or calcium carbonate contents within the upper 40 inches (100 cm) are assigned an added AIV of +1. In addition, soils with both high salts and carbonates within depths of 20 to 40 inches are assigned an additive adjustment of +1. Very fine textured soils (>60 percent clay) have very low potential for GPR and are not assigned these added AIVs.

For each component of a soil map unit, the most limiting condition (highest AIV) for each of the selected properties (clay content, electrical conductivity or SAR, and calcium carbonate or calcium sulfate content) at the soil horizon level is selected. These index values are summed to a depth of 6.6 feet for mineral soils and 4.1 feet for organic soils. The summation of the most limiting conditions represents the component index value (CIV). A CIV is computed for each soil in the map unit. Table 2 shows the relative composition (percentage) of different soil components and their CIVs for a hypothetical soil map unit.

A relative suitability index (SI) is computed for each map unit by summing the percentages of soils with the same CIV. Table 3 shows the results of summing the soil component percentages by the CIVs shown in Table 2. The dominant CIV is 4 for the map unit shown in Table 3. Soils with this index value make up 38 percent of the map unit. However, this map unit is also composed of soils that have more limiting (21%) and more favorable (41%) CIVs. The SI for a map unit represents the most dominant attribute properties, but does not identify or weighs the proportion of other soils that have different CIVs and occur within the map unit.

The final product is a lookup table consisting of the map unit identifiers and dominant GPR suitability indices (SI) shown in Table 4. The dominant SI for the soil components is joined to the map unit identifiers and displayed in graduated colors on digital maps.
 

Relative Suitability of Soils for GPR

Soil attribute index values and relative soil suitability indices are based on observed responses from antennas with center frequencies between 100 and 200 MHz. For mineral soils, the inferred SI is based on unsaturated conditions and the absence of contrasting materials within depths of 6.6 feet. Penetration depths and the relative suitability of mineral soils will be less under saturated conditions. Contrasting physical and chemical properties will also affect attenuation rates and penetration depths.

Areas dominated by mineral soil materials with less than 10 percent clay or very deep organic soils with pH values < 4.5 in all layers have very high potential (SI of 1) for GPR applications. Areas with very high potential afford the greatest possibility for deep, high resolution profiling with GPR. However, depending on the ionic concentration of the soil solution and the amounts and types of clay minerals in the soil matrix, signal attenuation and penetration depths will vary. With a 200 MHz antenna, in soils with very high potential for GPR, the effective penetration depth has averaged about 16.5 feet. However, because of variations in textural layering, mineralogy, soil water content, and the ionic concentration of the soil water, the depth of penetration can range from 3.3 to greater than 50 feet.

Areas dominated by mineral soils with 18 to 35 percent clay or with 35 to 60 percent clay that are mostly low-activity clay minerals have moderate potential (SI of 3) for GPR. Low activity clays are principally associated with older, more intensely weathered soils. In soils with moderate potential for GPR, the effective penetration depth with a 200 MHz antenna has averaged about 7 feet with a range of about 1.6 to 16 feet. Though penetration depths are restricted, soil polygons with moderate potential are suited to many GPR applications.

Mineral soils with 35 to 60 percent clay, or calcareous and/or gypsiferous soils with 18 to 35 percent clay have low potential (SI of 4) for GPR. Areas with low potential are very depth restrictive to GPR. In soils with low potential for GPR, the depth of penetration with a 200 MHz antenna has averaged about 1.6 feet with a range of about 0.8 to 6.5 feet.

Areas that are unsuited (SI >5) to GPR consist of saline and sodic soils. These soil map units are principally restricted to arid and semiarid regions and coastal areas of the United States.
 

Ground-Penetrating Radar Soil Suitability Maps

Ground-Penetrating Radar soil suitability maps have been prepared for the conterminous United States, most states and Puerto Rico. The Ground-Penetrating Radar Soil Suitability Map of the Conterminous United States is based on attribute data contained in the STATSGO data base and offers service providers an indication of the relative suitability of soils to GPR within broadly defined soil and physiographic areas. Within any broadly defined area, the actual performance of GPR will depend on the local soil properties, the type of application, and the characteristics of the subsurface target. State GPR soil suitability maps are based on attribute data contained in the SSURGO data base and provide a more detailed overview of the spatial distribution of soil properties that influence the depth of penetration and effectiveness of GPR. The spatial information contained on state GPR soil suitability maps can aid investigators who are unfamiliar with soils assess the likely penetration depth and relative effectiveness of GPR within project areas. However, as soil delineations are not homogenous and contain dissimilar inclusions, on-site investigations are needed to confirm the suitability of each soil polygon for different GPR applications.

Several reference layers are included on the GPR soil suitability maps to facilitate location of GPR project areas. Major political boundaries are displayed using segmented black lines outlined in gray.

The Ground-Penetrating Radar Soil Suitability Map of the Conterminous United States was compiled at a scale of 1:250,000 and incorporates the state and national boundaries as reference layers (U.S. Census Bureau, 2008B). Six different shades of color were chosen to represent the GPR suitability indices.

County or parish boundaries and names are also included on state GPR soil suitability maps (U.S. Census Bureau, 2008a). Major highways are depicted with red lines (U.S. Geological Survey, 1999). Selected urban areas are labeled and shown with gray hatch-filled patterns on state maps (U.S. Census Bureau, 2008c). Bodies of water, if mapped, digitized and archived, are shown in light blue. Bordering areas are in shaded relief with major political boundaries shown (U.S. Geological Survey, 2008).

The state GPR soil suitability maps use different publication scales depending upon the size of the projected areas. Display scales vary from 1:330,000 (Connecticut, Massachusetts, and Rhode Island) to 1:1,550,000 (Texas). The dominant GPR suitability indices are displayed as a graduated color map. Ground-penetrating radar soil suitability maps can be printed and select areas may be “zoomed-in” for better clarity of soil polygons. However, image quality will not support enlargements greater than 200 percent.

State GPR soil suitability maps use nine different colors to represent the GPR suitability indices, and areas that are not digitized, have not been rated, or bodies of water. The dominant GPR suitability indices are displayed in the same color scheme used for the Ground-Penetrating Radar Soil Suitability Map of the Conterminous United States.

Areas that are “Not Digitized” are shown in white on state GPR soil suitability maps. These areas are either not yet inventoried or mapped (e.g., non project soil survey areas), or the soils have been mapped but the spatial data have not been digitized, archived, and available for public distribution through the Soil Data Mart. Also appearing as “Not Digitized” are areas where the soils were not inventoried, but polygons were assigned a descriptive map unit name (e.g., Access denied, Arlington National Cemetery, Not mapped).

Polygons shown as “Not Rated” contain missing soil property data in more than 50 percent of their area. These polygons are shown in gray on state GPR soil suitability maps. Most miscellaneous areas and some areas mapped at higher levels of soil classification (e.g., Ustorthents, Udifluvents) lack pertinent soil property data and are shown as “Not Rated.” Miscellaneous areas contain little or no soil and support little or no vegetation. Examples include areas of exposed parent rock (e.g., Cinder land, Lava flows, Quarries, Rock outcrop, Rubble land), recently exposed or deposited materials (e.g., Badlands, Beaches, Rough broken land), and culturally modified materials (e.g., Borrow pits, Dumps, Earthen dams, Made land, Paved areas, Urban land). Some miscellaneous areas, because of intrinsic properties (Salt flats, Playas, Sand dune, Dunes, and Dune land), have been assigned index values. This was accomplished using a companion lookup table that contains keys, which overrides the programmed codes and adjusts the ratings.

Some multi-component map units consist of soils and/or miscellaneous areas that have exceedingly different GPR index values. In these map units, the rating for the most extensive component is used. If the component percentages are identical, the most limiting rating is selected. Examples are the Isolde-Appian complex and the Dune land-Playas complex from Nevada. Isolde and Appian are recognized soil series and are members of the mixed, mesic Typic Torripsamments and the fine-loamy over sandy or sandy skeletal, mixed, superactive, mesic Typic Natrargids families, respectively. Dune land and Playas are two highly contrasting miscellaneous areas that have soil property data. In each of these soil map units, the named components occupy about 40 percent of the polygon and have CIVs of either 1 (Isolde and Dune land) or 6 (Appian and Playas). These units are assigned an SI of 6, based on the most limiting rating.

An “urban rule” was introduced to provide some information for areas of “Urban land” that were mapped with at least one named soil component. For urban land map units, if named soil components make up 25 percent or more of the soil polygons, the CIV for the most extensive soil is used. If the extent of two or more soils is equal, the most limiting CIV is used (“tie-rule”).

Because of changes in mapping concepts, the recognition of new soils, additional laboratory data and revised soil interpretations, changes in GPR soil suitability indices are evident along some county and soil survey borders. These discrepancies in GPR indices are artificial and represent the patchwork collection of soil data over time. Modern soil surveys conform to natural soil and physiographic features rather than political boundaries. Under modern soil survey concepts, common standards and quality control will be applied and soil polygons and interpretations will join across political boundaries. As soil surveys are updated, these mapping artifacts (soil boundaries that follow political rather than soil or physiographic boundaries) will be eliminated.
 

References

Daniels, D. J., 2004. Ground Penetrating Radar, 2nd Edition. The Institute of Electrical Engineers, London, United Kingdom.

Doolittle, J. A. and M. E. Collins, 1995. Use of soil information to determine application of ground-penetrating radar. Journal of Applied Geophysics, 33:101-108.

Doolittle, J. A., F. E. Minzenmayer, S. W. Waltman, and E. C. Benham, 2002. Ground penetrating radar soil suitability map of the conterminous United States. 7-12 pp. In: Koppenjan, S. K., and L. Hua (Eds). Ninth International Conference on Ground Penetrating Radar. Proceedings of SPIE Volume 4158. 30 April to 2 May 2002. Santa Barbara, CA.

Grant, J. A. and P. H. Schultz, 1994. Erosion of ejecta at Meteor Crater: Constraints from ground penetrating radar. 789-803 pp. In: Proceedings Fifth International Conference on Ground-Penetrating Radar. Waterloo Centre for Groundwater Research and the Canadian Geotechnical Society. June 12–14, 1994, Kitchner, Ontario, Canada.

Jackson, M. L., 1959. Frequency distribution of clay minerals in major great soil groups as related to the factors of soil formation. Clays and Clay Minerals 6: 133-143.

McNeill, J. D., 1980. Electrical conductivity of soils and rock. Technical Note TN-5. Geonics Limited, Mississauga, Ontario.

Saarenketo, T., 1998. Electrical properties of water in clay and silty soils. Journal of Applied Geophysics 40: 73-88.

Smith, D. G. and H. M. Jol, 1995. Ground-penetrating radar: antenna frequencies and maximum probable depths of penetration in Quaternary sediments. Journal of Applied Geophysics 33: 93-1.

Soil Survey Staff, 1993. Soil Survey Manual. US Department of Agriculture - Soil Conservation Service, Handbook No. 18, US Government Printing Office. Washington, DC.

Soil Survey Staff, 1999. Soil Taxonomy, A Basic System of Soil Classification for Making and Interpreting Soil Surveys 2nd Edition. US Department of Agriculture - Natural Resources Conservation Service, Agriculture Handbook No. 436, US Government Printing Office, Washington, DC.

U. S. Census Bureau, 2008a. Current County and Equivalent, TIGER/Line 2008 (cartographic boundary file, tl_2008_us_county.zip). Available FTP: ftp://ftp2.census.gov/geo/tiger/TIGER2008/. [Accessed on February 20, 2009]

U. S. Census Bureau, 2008b. Current State and Equivalent, TIGER/Line 2008 (cartographic boundary file, tl_2008_us_state.zip). Available FTP: ftp://ftp2.census.gov/geo/tiger/TIGER2008/. [Accessed on February 20, 2009]

U. S. Census Bureau, 2008c. Urban Areas (generalized cartographic boundary file, ua99_d00_shp.zip). Available FTP: http://www.census.gov/geo/cob/bdy/ua/ua00shp/. [Accessed on February 20, 2009]

USDA - Natural Resources Conservation Service, 1994. State Soil Geographic (STATSGO) Database - Data Use Information. Misc. Publication No. 1492. National Soil Survey Center, Lincoln, NE.

USDA - Natural Resources Conservation Service, 1995. Soil Survey Geographic (SSURGO) Database - Data Use Information. Misc. Publication No. 1527. National Soil Survey Center, Lincoln, NE.

U.S. Geological Survey, 1999. Major Roads of the United States: U.S. Geological Survey, Reston, Virginia. Available FTP: http://nationalatlas.gov/atlasftp.html

U.S. Geological Survey, 2008. Analytical Hillshade computed from 1 kilometer National Elevation Dataset (NEDS) using the following parameters: 315 degrees altitude, 45 degrees azimuth, and z factor 1x. Prepared by USDA-NRCS-NSSC, Lincoln, NE.
 

Contact the Author

Jim Doolittle
E-mail:  jim.doolittle@lin.usda.gov

Table 1. Soil properties and attribute index values (AIV) used to calculate soil component index values.
CIV = (A + B + C).

A. Clay
A1.1. Mineral Soils

A1.2. Mineralogy override for low activity clays

A2. Fabric override for Organic Soils

B. Electrical Conductivity (mmho/cm) and Sodium Absorption Ratio

C. Added AIV for Calcium Carbonate and Calcium Sulfate

Table 2. Relative composition (%) and component index values (CIV) for a hypothetical map unit.

Table 3. Relative proportion (%) of soil component with the same component index value (CIV) for a hypothetical soil map unit.

Table 4. GPR potential ratings based on grouped suitability indices (SI).

< Back to Ground-Penetrating Radar Soil Suitability Maps