Relative efficacies of GPR and EM-31 compared to traditional soil investigation techniques on a loess : Tertiary deposits in Southwest Tennessee

Remote sensing technology such as ground penetrating radar (GPR) and electromagnetic induction (EMI) have great promise as soil investigation tools. For certain applications, using tools such as these can potentially provide a more complete picture of the sub-surface than traditional soil investigation techniques. In order to evaluate the efficacy of such tools, GPR and EMI data were compared to traditional soil investigation data on loess / alluvium / Tertiary deposits in southwest Tennessee. The objectives of this study were to i) conduct a complete soil morphological, chemical, and physical characterization of a landscape at the Ames Plantation; ii) classify the soils to the family level of soil taxonomy; iii) conduct non-intrusive soil investigation using both GPR and EMI; and iv) compare remote sensing data to the traditional soils data. A total of nine pedons were sampled on field #75 at the Ames Plantation. Sites were sampled on a 30.5-m X 30.5-m grid. Each site was spaced 15.25 m apart. Sampling depths ranged from 2.5 m to 3.0 m. Standard soil survey laboratory methods were performed on all sites for particle size, cation exchange capacity, exchangeable bases, free iron oxides, easily reducible manganese, KCl extractable aluminum, pH, total elemental analysis, total carbon, organic carbon, and BaCl2-TEA extractable acidity. Bulk density was determined using the Varsol method. Saturated hydraulic conductivity was measured in situ using an Amoozemeter. Soil physical data were analyzed and grouped using Ward\u27s Minimum Variance Clustering Analysis. Non-intrusive mapping was completed using a Trimble Ag. GPS™ 132, a Geonics EM-31™, and an SIR system 10-A™ GPR unit. Maps of bulk soil conductivity to the 3-m and 6-m depths as well as GPR images were produced. Sites were clustered based upon similarities seen in the remote sensing data. Final groupings from both techniques were analyzed for agreement using SAS. All sites were located on an upland position with loess over alluvium over Tertiary sand parent material sequence. Loess thickness ranged from 90 cm to 144 cm. Alluvium thickness ranged from 82 cm to 151 cm. Pedons sampled at sites four and seven did not reach the Tertiary sand. Carbon distribution was somewhat irregular throughout the alluvium at all sites except site six. A fining upward sequence was seen throughout the alluvium at all sites. Data were insufficient to distinguish different loess depositions. All soils were classified as Ultic Hapludalfs. Statistical analysis of soil morphological and physical data yielded the following groups: 1) 1, 3, 4, 7; 2) 5, 6, 8; 3) 2, 7; site 9 was not grouped. GPR and EM-31 data produced the following groups: 1) 1, 3 ,4, 7; 2) 6, 8, 9; 3) 2, 5. Agreement between the groupings was strong, K = .8 ± 0.34. The GPR and EM-31 were effective for detecting slight variations in physical properties and grouping soils based on these differences in soils formed in loess/alluvium/Tertiary sands