Topographic metric predictions of soil organic carbon in Iowa fields

Authors: LI, XIA - University Of Maryland; McCarty, Gregory; Karlen, Douglas; Cambardella, Cynthia

Submitted to: Catena
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 9/25/2017
Publication Date: 1/15/2018
Citation: Li, X., McCarty, G.W., Karlen, D.L., Cambardella, C.A. 2018. Topographic metric predictions of soil organic carbon in Iowa fields. Catena. 160:222-232. https://doi.org/10.1016/j.catena.2017.09.026.
DOI: https://doi.org/10.1016/j.catena.2017.09.026

Interpretive Summary: As an important soil property and a key factor affecting soil quality, soil organic carbon (SOC) is lost from locations in a landscape by multiple processes, including soil aggregate disruption, carbon transport, and SOC mineralization with potential for deposition at other locations. SOC distribution is determined by a variety of factors in agroecosystems. Topography is an important factor regulating soil erosion since it affects surface runoff, soil texture, and vegetation. Landscape topography is a description of the shape and features of landforms, and topographic information can be derived from digital elevation models (DEMs). We adopted the radioactive fallout cesium -137 produced from atmospheric testing of atomic bombs in the 1960's to estimate soil redistribution rates and patterns of SOC distribution across 560 sampling locations at two field sites and at a larger scale for the Walnut Creek watershed in Iowa. Then, using ordinary least squares regression and principal component regression, topography-based models were developed to simulate spatial patterns of SOC content and soil redistribution. Results suggested that erosion and deposition of topsoil were controlled by topography with soil gain in lowland areas and soil loss in sloping areas. All topography-based models developed demonstrated good simulation performance, explaining more than 62% variability in SOC content and soil redistribution across two field sites with intensive samplings. However, the ordinary least squares regession models showed lower reliability than the principal component regression model for predicting SOC dynamics at a watershed scale. This finding indicates that models based on principal components can be effective tools for scaling of SOC content and soil redistribution at field to watershed scales

Technical Abstract: Topography is one of the key factors affecting soil organic carbon (SOC) redistribution (erosion or deposition) because it influences the gravity-driven movement of soil by water flow and tillage operations. In this study, we examined impacts of sixteen topographic metrics derived from Light Detection and Ranging (LiDAR) data on SOC distribution in agricultural fields. We adopted the fallout Cesium-137 technique to estimate soil redistribution rates and patterns of SOC distribution across 560 sampling locations at two field sites and at a larger scale for the Walnut Creek watershed in Iowa. Then, using stepwise ordinarily least squares regression (SOLSR) and stepwise principal component regression (SPCR), topography-based models were developed to simulate spatial patterns of SOC content and soil redistribution. Results suggested that erosion and deposition of topsoil were regulated by topography with soil gain in lowland areas and soil loss in sloping areas. Topographic wetness index (TWI) and relief were the most influential variables controlling SOC content and soil redistribution, respectively, and were of primary importance in SOLSR models. All topography-based models developed through SPCR and SOLSR demonstrated good simulation performance, explaining more than 62% variability in SOC content and soil redistribution across two field sites with intensive samplings. However, the SOLSR models showed lower reliability than the SPCR models in predicting SOC dynamics at a watershed scale. Results of this study highlighted the topography-based SPCR model as an effective and a promising tool allowing for scaling of in situ SOC content and soil redistribution at crop sites to a large-scale watershed, and provided valuable insight into the spatial patterns of SOC distribution.