Properties and classification of residual soils derived from Cambrian and Ordovician limestones and dolomites in southwestern Virginia

Abstract

Profiles of residual soils developed in materials weathered from limestones and dolomites in Washington, Smyth, Wythe, Roanoke, Botetourt, and Rockbridge Counties were characterized in order (i) to evaluate differences in distributions of soils related to differences in parent materials; (ii) to evaluate differences in distributions of soil as they relate to identification of soil profiles as members of Frederick, Groseclose, and Lodi series by soil surveyors in the field; (iii) to evaluate relationships of taxa limits defined by Soil Taxonomy and distributions of soils produced by grouping soil profiles by numerical taxonomic procedures; and (iv) to produce a regional correlation of these one-half million acres of agriculturally important soils. Soil profiles selected randomly from areas of residual soils developed in materials weathered from limestones and dolomites were characterized to evaluate differences in soils related to differences in parent materials. Specifically, soils studied were derived ( i) from the Beekmantown and Chepultepec formations, limestones and dolomites that contain argillaceous materials and chert, parent materials for the Frederick soils; (ii) from the Conococheague and Copper Ridge formations, limestones and dolomites that contain argillaceous materials and sandstones, parent materials for the Lodi soils; and (iii) from the Elbrook and Honaker formations, dolomites that contain argillaceous materials, shales, and siltstones, parent materials for the Groseclose soils. Distributions of properties of soils developed from these three types of parent materials were similar. Therefore, differencs in soils assumed to result from differences in the lithologic characters of these geologic formations were not supported by this study. Soil profiles were identified as members of the Frederick, Groseclose, or Lodi series by field-observable soil properties. The series representing the highest percentage of agreement among the soil surveyors from the six counties was used to name each soil profile. Distributions of properties for soils identified as members of these series were similar. Therefore, we conclude that the concepts of these three series represent names for similar soils. Principal component and cluster analyses were used to group the soil profiles. Distributions of properties of soils in the four clusters were similar, i.e., did not warrant the use of more than one series for naming map units of these soils. Coincidence of taxa limits and median values for the distributions of properties used as differentiae by Soil Taxonomy was the primary reason for soil profiles not being classified dominantly into one family. The number of taxonomic placements for soils developed from each of the three geologic formations and for soils identified as members of each of the soil series precludes the use of one soil series defined by current criteria in Soil Taxonomy for objectively naming and interpreting map units of these soils. Seven taxonomic placements were possible using current criteria in Soil Taxonomy. The proposed kandic great groups and subgroups will increase the number of placements to nine. Because the distributions of properties of soils developed from the three parent materials, of soils identified as members of the three series, and of soils identified as members of the four cluster were similar, we propose the use of one soil series for naming map units of these soils. Soils in all the other taxa should be considered as taxonomic noise, i.e. similar soils. Since Frederick is a benchmark, Hall-of-Fame series and has precedence through longer use, we recommend that the Frederick series be used to name map units of soils currently identified as members of the Frederick, Lodi, and Groseclose series in the approved field mapping legends of Washington, Smyth, Wythe, Roanoke, Botetourt, and Rockbridge Counties. We propose that the Groseclose series be redefined to include soils with 20- to 40-inch depths to BCt or C horizons that contain weathered shales and siltstones. The redefined Groseclose series would then be limited to soils primarily developed in materials weathered from the Rome-Waynesboro formation. The Groseclose series should take precedence over the Lodi series since the soils are similar and contain small amounts of sand. Data presented in this study were used to revise the Frederick and Groseclose series descriptions given in the section "Series Concepts - Soil Taxonomy."