Analysis of the spatiotemporal distribution of soil organic carbon

Abstract

Feedbacks between atmospheric and terrestrial carbon stocks remain unclear. Soil carbon stores are affected by a complex interaction among several biophysical and hydroclimatic processes, including the dynamics of soil moisture, insolation, and temperature. Studying the spatiotemporal distribution of soil organic carbon (SOC) and its dynamic changes is necessary for building a soil carbon pool inventory, as well as predicting the potential for soil carbon sequestration (Dorji et al., 2014). Many environmental controls on soil organic carbon have been analysed at site-specific plot-scales, while long term temporal studies of SOC dynamics are less common. The continued uncertainty of carbon cycle feedbacks, and the complexity of interactions of controls on, and transport of, soil carbon at regional scales, justifies further investigation. This study investigates the spatiotemporal relationships between surface SOC and a number of environmental variables across a catchment. The findings of this research will contribute to overall understanding of SOC distribution and controls for large regional scales. The catchment for this study is located in the Upper Hunter Valley region of New South Wales, Australia. The study focuses on the Krui River catchment, having an area of 562 km 2 . Soil samples were obtained across the catchment using a 1 km grid scale. Cores with a depth of 220 mm were obtained from 59 grazing sites in 2006 and from 52 grazing sites in 2014, with 41 of the sites common to both 2006 and 2014. At each sampling location, aboveground biomass (AGB), soil moisture, and soil temperature was sampled within a 0.25m 2 quadrat. Land use at each sample site was classified as either cropping or grazing. For a comparison of the temporal variability in SOC concentrations, average SOC from the two sampling dates were compared using Student's t-test. To determine which variables were the most important, principle component analysis was performed for topographic (elevation, slope, aspect, plan curvature, profile curvature, TWI), soil (pH, Electroconductivity, clay %) and vegetation (sampled vegetation biomass, remotely sensed vegetation biomass) variables for the sample sites. SOC, elevation, Normalised Difference Vegetation Index (NDVI) and plan curvature were found to be the most important variables for the first 2 principle components. Linear regression and heteroscedasticity tests were applied to the strongest correlations between SOC and the other variables for both sampling periods. The results of this study show that soil carbon was spatially and temporally stable over medium time scales (8+ years), with the variables of SOC, elevation and NDVI having strong, positive correlations with each other for both sampling periods. Strong, positive Pearson's r correlations were observed between SOC and NDVI, a surrogate for Aboveground Net Primary Production (ANPP), for both sampling periods. Thus regions of higher net primary production corresponding with higher concentrations of SOC. Grazing intensity, represented in this study by sampled AGB, did not affect SOC. Topography strongly influences vegetation via its control on such climate variables as precipitation and temperature. Elevation was found to explain much of the variability in NDVI, and hence SOC, although slope and aspect also had weak to moderate correlations with NDVI. The relationship between SOC and aspect was weak. This study demonstrates that the variables of elevation and NDVI can be used to digitally map the spatiotemporal distribution of SOC across large (~500 km 2 ) catchments of elevations ranging from ~300 to ~1100 m. However, long term seasonal climate variability may affect the predictive ability of SOC using these variables. The spatial and long term temporal stability of catchment SOC demonstrated here has major implications for soil carbon sequestration. SOC across the catchment appears to be at equilibrium, with minimal variation observed after 8 years of continuous grazing. Carbon sequestration methods would therefore require major changes in grazing land-use to achieve observable increases in soil carbon.