Soil modeling for soil loss tolerance estimations: exploring natural baselines and long-term variations

Ranathunga Arachchige, Keerthika Nirmani [UCL] Finke, Peter A. [Ghent University] Yin, Qiuzhen [UCL] Verdoodt, Ann [Ghent University] Yu, YanYan [Chinese Academy of Sciences]

Quantification of soil stocks and ecosystem services (ES) under changing climate need time series of climate and soil data. Here, we propose a new approach combining process-based soil and climate simulations to obtain such time series, and apply it onto a recent interglacial, Marine Isotope Stage (MIS) 5e. We combined the LOVECLIM climate model and the SoilGen2 soil evolution model to simulate soil development under two scenarios: dust addition plus erosion and erosion-only (both with five erosion rates corresponding to low, natural erosion rates: 0, 0.5, 1.0, 1.5 and 2.0 Mg ha−1y−1). We quantified five target variables: two soil stocks, Exchangeable Bases (EB) and Soil Organic Carbon (SOC), and three ES: Carbon Sequestration Capacity (CSC), Water Yield (WY) and the ratio of actual over potential evapotranspiration (Ω) at four sites on an aridity gradient on the Chinese Loess Plateau (CLP). We used the obtained time series of these target variables to estimate Soil Loss Tolerance (SLT) threshold values over the full extent of MIS 5e (22 ka). Under increased erosion, EB increases while SOC always declines in both scenarios. In both scenarios, the simulated CSC increases with erosion; contrastingly, the WY decreases with increasing erosion rates. Both CSC, WY and Ω gradually decrease towards the northwest CLP (semi-arid region). For EB and SOC, the determined SLT thresholds are relatively higher in the dust addition than in the erosion-only scenarios, and strictly follow the climate gradient in the CLP. Combined performance index for (1-CSC), WY and Ω, showed that soil ES performance is worse above 1.0 Mg ha−1y−1. This implies that benchmark levels must be chosen carefully. Our research highlights the potential of using SoilGen2 with LOVECLIM for quantifying soil-based ES and SLT.