Hillslope redistribution of soil organic carbon in the depressional landscape in Minnesota

Abstract

Agricultural tillage has been estimated to cause a loss of 30-50% of the pre-settlement soil organic carbon (SOC) through enhanced decomposition and loss to the atmosphere or through erosion and subsequently loss to surface waters or burial in lower landscape positions. However, measures of whole landscape redistribution and fate of sediments and SOC are lacking. This research seeks to estimate change in SOC storage since agricultural settlement using soil-terrain modeling techniques in closed-depressional landscapes. The overall quantity of SOC in depressional landscapes may have not been lost to the atmosphere through enhanced decomposition but rather is redistributed downslope.I conducted field observations and soil sampling in hillslope transects in Lake Rebecca Park Reserve in East-Central Minnesota. The thickness of re-deposited sediments (termed post-settlement alluvium, or PSA) was identified by morphological indicators in the field. The spatial distribution of PSA presence and its thickness were modeled with local and regional terrain attributes using a two-stage regression approach. The current SOC inventory (1.119 Pg) in top 1-m soil at the Lake Rebecca site was estimated by spatial predictive models of SOC contents at four soil depths (0-10 cm, 10-30 cm, 30-60 cm, 60-100 cm). I estimated pre-settlement SOC inventory for erosional uplands with spatial predictive models for an uncultivated grassland in Morristown, Minnesota; for depositional lowlands, I calculated pre-settlement SOC inventory by applying models for soil profiles below the PSA depth at the study site. Erosional losses and depositional gains were determined by subtracting current SOC inventory from pre-settlement values.The results showed high SOC contents in surface soils at lower landscape positions, especially in wetlands near the surrounding marsh. Total SOC in the uppermost meter of this 6-ha study site was estimated as 1.528 Gg. The change in SOC density since European settlement was highly overestimated (36.7% increase). The prediction error is likely due to the lack of a mechanism to constrain the prediction of PSA under natural sedimentation patterns at the very bottom of the hillslope beyond the zone where PSA was observed. The model improvement is required to more accurately predict whole landscape SOC distribution and change over time.