LAUSR

1. Healthy soils are crucial for sustainable food production, but tillage limits the biological regulation of essential ecosystem services. Better understanding of the mechanisms driving management effects on soil ecosystem engineers is needed to support sustainable management under environmental change. 2. This paper presents the EEEworm (Energy–Environment–Earthworm) model, a mechanistic individual-based model (IBM) of Lumbricus terrestris populations. L. terrestris is a dominant earthworm species in undisturbed habitats and is closely associated with numerous ecosystem services such as water flow regulation, soil structure and crop production. In reduced tillage agriculture a decline in mechanical disturbance allows for L. terrestris proliferation, whilst the activities of L. terrestris can replace many of the soil functions provided by tillage. 3. Extensive EEEworm validation with eight published studies (average R2 = 0.84) demonstrates a mechanistic approach which can extrapolate between diverse soil, management and weather conditions. EEEworm simulation experiments elucidate that a combination of direct and indirect tillage effects lead to population declines in tilled fields, with litter removal from the soil surface being the main driver. 4. We investigate the effects of different tillage intensities under historical and projected soil warming conditions, and find that future warmer and drier soils in our simulation exacerbate the effects of deep ploughing on L. terrestris population declines. These effects result from warmer and drier soil conditions increasing individual metabolic rates and tillage reducing food availability to meet energy demands. 5. Synthesis and applications. Pre-emptive strategies to mitigate climate change impacts on soil health in agroecosystems should focus on decreasing tillage intensity and retention of crop residues following tillage. EEEworm has the potential to benefit land managers, policy makers, risk assessors and regulators by providing a tool to forecast how soil systems respond to combinations of land management and environmental change. To allow better cost-benefit analysis of contrasting land management systems a future aim of mechanistic models like EEEworm is to incorporate the links between earthworm populations, soil functions and ecosystem services.