THE AUSTRALIAN SOIL IDENTIFICATION SPREADSHEET (ASIS)

 

Introduction

     
    One of the two main conventional soil classification methods used by a majority of soil scientists in Australia is the Great Soil Groups (GSG) system (Stace et al., 1968). In this scheme, Australian soil has been classified into 43 great groups which have been arranged in seven categories. The Australian Great Soil Groups have been popular in Australia for many years and many people are familiar with them. Therefore, the Australian Soil Identification Spreadsheet (ASIS) was designed and developed to help soil users to allocate an unknown profile quickly as well as accurately to the GSG. Since some of the 43 original groups are multiple groupings, it has been expanded to 50.
     

    Fuzzy soil classification

    Soil classification, in the traditional sense, is based on the notion that soil forms discrete, internally homogeneous units, with sharp boundaries. In this sense each soil unit can be represented by a central idea known as a typical or representative profile. This model implies that the predicted value of a soil attribute at any unsampled area is either the value for the typical profile or the mean value for the soil unit.
    The problem of dealing with undefined classes and vague boundaries has led to fuzzy soil classification in which soil may belong totally, partially or not at all to soil classes. In practice, the extent of soil membership in any class is permitted from zero (no membership) to one (membership in one class only). The main feature of this system of classification is the grouping of soil individuals into classes where boundaries are not, should not or can not be, exactly defined. These diffuse classes are known as continuous classes and the system itself is called continuous classification (McBratney & De Gruijter, 1992). Allocating unknowns to fuzzy and continuous soil classes is dealt with in detail by McBratney (1994).
     

    Advantages of this system

    This system allows soil scientists to check the results of soil identification through the membership allocation; the greater the membership, the stronger the linkage to a great soil group. As this system is more flexible than conventional ones, it can transfer more and better information to users.
    Fuzzy soil classification is a continuous, quantitative or objective procedure in which the classes are defined in terms of membership and based on many properties  simultaneously (polythetic). Additionally, as for qualitative attributes such as soil structure, this system can accommodate these subjective assessments. No special expertise is required, except for a general knowledge of soil description.
     

    The ASIS program

    The ASIS is based on the fuzzy k-means with extragrades algorithm. Fuzzy k-means with extragrades has been applied to permit variation in the distance dependence for membership in the outlier group recognised by De Gruijter & McBratney (1988).

    The implementation on the Macintosh and PC makes the technique for generalised fuzzy k-means available for soil profile identification for the first time to everyday users. ASIS  requires an Apple Macintosh or Windows compatible PC with Microsoft Excel version 5.0 or higher installed.
    Download ASIS

     

    Using ASIS

    ASIS requires inputs to be coded and soil colour attributes transformed in a particular manner (Mazaheri et al., 1995a). Twenty attributes are needed: topsoil and subsoil texture for texture profile; topsoil salinity; subsoil salinity; topsoil organic matter %; subsoil pH (1:5 H2O); subsoil structure; subsoil mottling; subsoil clay %; bleaching of A2 (E) horizon (if present); concretions; crust; hardpan; topsoil calcium carbonate; subsoil (pedological) calcium carbonate; subsoil gypsum; pedological subsoil; alluvial layer; parent material and subsoil colour.

    For allocation of the profile to one of the texture profile classes of uniform, gradational and duplex (Northcote, 1984), the topsoil, midpoint (A2, A3, A/B or B1, if present) and subsoil texture should be determined. ASIS can allocate the profile automatically to the texture profile. Topsoil and subsoil salinity should be expressed in dS m–1 in a 5:1 water:soil extract. Acidity or pH should be measured in 1:5 soil:water ratio. As organic matter in soil is not usually measured, it can be calculated by approximate conversion that assumes soil organic matter contains generally 57 % carbon. Therefore,

    % O.M.= % O.C. ´ 100/57

    Moist soil colour is quantified and transformed automatically within the spreadsheet, provided the Munsell hue, chroma and value are entered. The final result is an allocation to one or more GSG. The membership values are displayed and may be printed out.

    Next:

    [Soil Physics Home Page]
     

    Designed by: Budiman
    (c) 1998