A relative weight model for soil productivity assessment

Abstract

Soil productivity is a function of inherent factors such as topography, parent material, physical and chemical properties of the soil, and the infrastructure for irrigation and drainage. As multi-criteria evaluation methods in soil productivity assessment, the least-factor and weight methods, while popular, have limitations. The least-factor method is not accurate enough in the absence of a vital constraint factor, and the weight method leads to an inaccurate, and even incorrect result when there is a vital constraint factor. In order to overcome these limitations a new concept, relative weight, was introduced and a prototype model developed. In this prototype model, every factor has different relative weights in different soil units, thus allowing it to overcome shortcomings of the Weight method where the weights of a given factor are assumed to be equal in all soil units. This prototype model was then applied in a case study on the Loess Plateau in Northwest China. Results from the case study indicated this prototype model was more precise than either the least-factor or weight methods, and was able to avoid the invalid results of the Weight method. Key words: Soil productivity index, relative weight, multi-criteria evaluation, Loess Plateau