Steps Toward A Universal Soil Classification

Abstract

Steps Toward A Universal Soil Classification Chapter one focusses on the segmented nature of many schemes around the world and the requirements for a unified system. The strengths, weaknesses and suitability of these world systems lead to the conclusion that there is no unified soil taxonomic system. Chapter two discusses the steps of creating a world taxonomic system. It suggests that many of the philosophical problems with Australian taxonomy can be solved. A new algorithm is formulated in which key extragrade points are nominated, turning them into the focal point of clusters. Data that is not extreme is free to resolve into typical fuzzy k means (FKM) clusters within the end points. Chapter three takes akromeson and applies it to data from the United States Department of Agriculture (USDA). This chapter discusses the implications of this new method of horizon partitioning. Chapter four takes the ideas of principal component analysis and taxonomic distances. Algorithms regarding next nearest neighbour analysis (NNA) at the Order level for both taxonomies enables a new test for taxonomic accuracy. Chapter five consists of work on the mathematical methods required for a global taxonomy. Data needs to be converted into the same taxonomic space. Two Orders almost identical in pedological description are selected for analysis. The mean taxonomic distances for each Order are calculated to gain an understanding on how each Order from each classification system compares to Orders of the opposing taxonomic system. NNA is then further developed to determine which tiers of which taxonomy are on a similar operational level. Chapter six suggests there are taxa within both ST and ASC that are so close in properties and attributes as to be considered duplicates. The tiers of ST are matched by NNA to the closest tier of the ASC. These levels are harmonized and an algorithm devised to remove duplicate soil taxa, then the taxonomies are combined and assessed. In Chapter seven, ST is assessed, compared and matched with the ASC. The combined data set is split into groups and the differentiating properties are found and a numerical taxonomic description scheme created. The role of a functional nomenclature is discussed. In chapter eight it is concluded that it is possible to harmonize soil information and to create a taxonomic scheme.