Unraveling the local and structured variation of soil nutrients using two-dimensional empirical model decomposition in Fen River Watershed, China

Abstract

ABSTRACT The spatial patterns of soil nutrients at the sampling scales are the interweaving of local and structured variances. In order to unravel the structured variation of soil nutrients from the original data, two-dimensional empirical mode decomposition (2DEMD) was used to decompose their spatial variability into different scales across a typical watershed. The results suggested that the 2DEMD could separate the local variation represented by IMF1 and structured variation represented by IMF2-4 and residue. Both variance percentages of each IMF and geostatistical parameters of structured variances (IMF2-4) indicated that majority of soil nutrient variations occurred within the local (< 3.2 km) and largest scales (60–90 km for SOM, SAP and SAK, and 219 km for STN km). Climatic factors controlled the soil nutrients under natural land use types, and the topographic factors mainly affected the soil nutrients under cultivated land use types at the largest structured variance. Therefore, at the largest scales, management practices considering slope or elevation are needed to improve the level of soil nutrients under cropland or orchard land in mountainous areas.