Carbon saturation drives spatial patterns of soil organic matter losses under long-term bare fallow

Spatial controls of soil organic carbon (SOC) turnover are not well understood. We hypothesized that spatial patterns of SOC turnover are related to carbon (C) saturation rather than to the size of measurable SOC-pools such as particulate organic matter (POM), determined as SOC in particle-size fractions. Therefore, we repeatedly grid-sampled a field after one, three, seven, and eleven years under bare fallow management, which revealed a spatial gradient from high to low degrees of C saturation. We measured the contents of SOC and the contents of SOC in coarse sand-size (2000-250 pm, POM1), fine sand-size (250-53 mu m, POM2), silt-size (53-20 mu m, POM3), and fine silt to clay-size fractions (nonPOM, < 20 mu m), calculated the degree of C saturation from textural properties and nonPOM contents, and related these parameters to SOC losses. In the first year of bare fallow, the soil contained on average 12.1 g SOC kg(-1), of which 0.6 g kg(-1), 1.7 g kg(-1), and 2.1 g kg(-1) were lost after three, seven, and eleven years of bare fallow, respectively. The SOC losses within eleven years were spatially variable and varied between 1% and 46% relative to the initial SOC content. In support of our hypothesis, SOC losses were largest at subsites with largest degrees of C saturation (R-2 = 0.83). Although the POM fractions declined most drastically, they only comprised 4 to 9% of bulk SOC, and they did neither correlate with nor explain spatial patterns of SOC losses. We conclude that the concept of C saturation is superior to conventional physical fractionation approaches for predicting spatio-temporal patterns of SOC turnover at sites with a high degree of C saturation.