Towards sustainable agriculture: behaviors, spatial dynamics and policy in an evolutionary agent-based model

Economic and population growth increasingly pressure the Earth system. Fertile soils are essential to ensure global food security, requiring high-yielding agro-technological regimes to cope with rising soil degradation and macro-nutrients deficiencies, which may be further exacerbated by climate change. In this work, we extend the AgriLOVE land-use agent-based model (Coronese et al., 2023) to investigate trade-offs in the transition between conventional and sustainable farming regimes in a smallholder economy exposed to explicit environmental boundaries. We investigate the ability of the system to favor a sustainable transition when prolonged conventional farming leads to soil depletion. First, we showcase the emergence of three endogenous scenarios of transition and lock-in. Then, we analyze transition dynamics under several behavioral, environmental and policy scenarios. Our results highlights a strong path-dependence of the agricultural sector, with scarce capacity to foster successful transitions to a sustainable regime in absence of external interventions. The role of behavioral changes is limited and we find evidence of negative tipping points induced by mismanagement of grassland and forests. These findings call for policies strongly supporting sustainable agriculture. We test regulatory measures aimed at protecting common environmental goods and public incentives to encourage the search for novel production techniques targeted at closing the sustainable-conventional yield gap. We find that their effectiveness is highly time-dependent, with rapidly closing windows of opportunity.