科研之友
微信
新浪微博
Facebook
分享链接
摘要
Pedodiversity studies greatly depend on the discretization of the soil continuum via various soil classification standards. Soil taxonomic systems are constructed considering soil properties and their organization in the soil profile. Therefore, it seems logical that the pedodiversity of an area could be calculated by considering the simultaneous variation of multiple soil properties within that area. With the aim of creating a bridge between these two lines of thinking, this study presents the development of two new indices of pedodiversity (HULLdiv and EIGENdiv) which consider continuous variables as input. The soil input information tested was a) Visible and Near Infrared (Vis-NIR) reflectance values and b) values of five soil properties predicted from Vis-NIR spectra. Both indices were employed to measure the pedodiversity at different extents in two perpendicular transects containing 27 (North to South) and 22 (East to West) locations respectively in New South Wales (NSW), Australia. Each location considered natural and intervened land Use. The indices successfully represented the pedodiversity of the area of study, however they were different depending on the input soil information i.e., raw Vis-NIR and predicted properties. HULLdiv was affected by extreme soil observations and as a result its discrimination power between areas with different soil diversity was inferior. On the other hand, EIGENdiv represented well the pedodiversity of an area, despite the extreme observations. The results showed good agreement with conventional methods for assessing pedodiversity i.e., Shannon's and Simpson's indices. However, the new indices were more discriminating by being able to better represent the landuse effect. This study represents the first attempt to measure pedodiversity in a continuous way using Vis-NIR information.
- 出版日期2017-5
