SoLIM home
study area
results and evaluations




---- Utilize components of the “SoLIM” approach to help assist in the development and evaluation of a computer generated-landscape model that captures existing soils and environmental knowledge, creates data layers for input into a GIS, and produces raster and vector maps suitable for used in soil survey, as well as providing the potential for the use of “fuzzy soil logic” in the development of “soilscape-based” soil interpretations.

---- Evaluate these procedures and products as to their applications with the National Cooperative Soil Survey, as well as within the National Park Service and the Natural Resources Conservation Service. In particular, assess the adoption of these techniques for mapping soil resources over areas that are not easily accessible for field observation.

SoLIM stands for “Soil Land Inference Model”. It is a fuzzy inference scheme for estimating and representing the spatial distribution of soil types in a landscape. There are four key components included in this framework:
(1) Soil-Landscape model: The knowledge on the relationships between the soil series and their surrounding environmental variables can be obtained from soil experts using a set of knowledge acquisition techniques.
(2) GIS Database: The soil formation environmental database of an area can be characterized and represented using GIS and RS techniques.
(3) Inference Engine: A set of computer programs under fuzzy logic used to estimate and predict the spatial distribution of soil types in a landscape based on Soil-Landscape model and GIS database.
(4) Fuzzy Representation: A Similarity Vector Model to represent soil as continuum.

Our study area is three quadrangles in the Great Smoky Mountains National Park, which straddles the boundary between North Carolina State and Tennessee State. In phase I, we worked on Mt. Guyot quadrangle and in phase II, we worked on Mt. LeConte quadrangle and Clingmans Dome quadrangle.

Three soil information products were provided: fuzzy membership maps of individual soil types (series), raster soil series maps, and conventional soil polygon maps.
---- Fuzzy Membership Maps
Fuzzy membership maps showing the spatial gradation of soils and preserving the intermediate nature (between types nature) of soils. This preservation would assist in soil interpretation. The images show the membership variation of Anakeesta (Right Upper) and Sylco (Right Lower) respectively. Anakeesta occupies most ridge top areas and slopes areas above frigid line while Sylco occupies most ridge top areas and slopes areas below frigid line. The membership maps show a realistic gradation of the two soil types when one travels from high elevation to low elevation position.

---- Raster Soil Maps
Conventional soil type maps can be produced through harden process from the fuzzy membership maps derived from the SoLIM approach.


---- Polygon Soil Maps (with the minimum mapping size of 1 hectare)
Conventional soil polygon like maps can also be created from the fuzzy representation. We know that it is inevitable for a soil polygon to include some small soil bodies which are different from what the soil polygon is labeled to be. These inclusions can be reported per soil polygon basis which is a great improvement over the conventional approach to reporting inclusions (often lumped into a mapping unit).

---- Accuracy for model develop area – Mt. Guyot: 30 meter resolution - 96.43% (54 out of 56); 10 meter resolution – 98.21% (55 out of 56)
---- Accuracy for extrapolated areas – Mt. Le Conte: 30 meter resolution - 92.68% (38 out of 41); 10 meter resolution– 92.68% (38 out of 41)
---- Accuracy for extrapolated areas – Clingmans Dome: 30 meter resolution - 97.67% (42 out of 43)

University of Wisconsin - Madison
A-Xing Zhu and James Burt, principal investigators, provide leadership in applying and evaluating the SoLIM approach in the study area;
Rongxun Wang, project assistant, do soil inference

Natural Resources Conservation Service
Darwin Newton and Roy Vick, provide administrative leadership respectively for the states of Tenne4ssee and North Carolina;
Anthony Khiel and Doug Thomas, serve as the local NRCS project coordinator;
Berman Hudson and Sheryl Kunickis, provide guidance and support from
the national NCSS perspective and oversee partial funding of the proejct from the NRCS persepetive

National Park Service

Pete Biggam, serve as project coordinator from the NPS;
Mike Jenkins, serve as the local contact representing GRSM

last updated: March 8, 2004
go top of the page