Variations in the concentration, composition, and sources of WEOM in erosion and deposition landscapes over an erosion intensity gradient on the Loess Plateau of China

Abstract

Water-extractable organic matter (WEOM) is the most active and mobile fraction of soil organic matter and responds sensitively to soil erosion. However, previous studies on the chemical nature of WEOM in the erosion and deposition landscapes were controversial at site-level studies and mainly focused on surface soils, limiting our assessment of WEOM’s environmental fate at larger and deeper scales. Here, we collected eroded soils and deposited sediments from a depth of 0–200 cm in five eroding catchments with an erosion intensity gradient across the Loess Plateau. We examined the concentration, composition, and potential sources of WEOM using UV–visible absorption and fluorescence excitation-emission matrix-parallel factor analysis. Soil WEOM quantity and quality varied significantly in the erosion and deposition landscapes, with greater variations in soils below 20 cm, where WEOM in the deposition zone was characterized by higher concentration, aromaticity, hydrophobicity, and UVA and UVC humic-like components derived from terrestrial and soil organic matter. WEOM variations in deep soils increased with increasing erosion modulus among sites, and were mainly driven by changes in soil organic carbon concentration, aggregate stability, and carbon to nitrogen ratio. However, the pattern of WEOM parameters between erosion and deposition zones at 0–20 cm was opposite to that in soils below 20 cm. Vegetation inputs compensated the loss of organic matter in sites with suitable climatic status and weak erosion, and led to higher concentrations and exogenous features of WEOM in the surface erosion area. Concludingly, the WEOM concentration, components, and sources differed significantly between erosion and deposition zones, and these differences, as well as their response to erosion intensity, revealed an opposing pattern in surface and deep soils.