Mechanistic Microbial Death Pathways Improve Regional Soil Carbon Simulations across Diverse Cropland Ecosystems

Abstract

Microbial death pathways (MDPs) are increasingly recognized as key drivers of terrestrial carbon cycling, primarily through their regulation of microbial necromass carbon (MNC), a critical pool in global carbon dynamics. Yet explicit representation of MDPs in soil organic carbon (SOC) models remains limited. Here, we developed and evaluated three SOC models that differ in their structure of MNC pool: the multiple-pathway necromass (MPN) model, which partitions microbial necromass carbon (MNC) into four MDP-derived subpools; the dual necromass (DUN) model, which differentiates two necromass pools with distinct decay rates; and the single necromass (SIN) model, which aggregates necromass into a single pool. Using a unified data assimilation framework and SOC observations from six major agricultural regions in China, we found that the MPN model consistently outperformed DUN and SIN models across most regions, producing necromass subpool dynamics, scenario responses, and parameter sensitivities that closely reflect the mechanistic understanding of MDPs. In cold or nutrient-limited regions, however, the simpler DUN model performed similarly while requiring fewer parameters, emphasizing the importance of balancing model complexity with regional ecological constraints. Our results demonstrate that explicitly incorporating MDPs enhances the robustness and mechanistic realism of SOC simulations and provides a robust foundation for more explicit representations of MDPs to assess the soil carbon sequestration potential and guide sustainable land management.