Importance of modelling microbial dynamics in soil: evaluation of conventional and microbial soil models informed by observations across various plant functional types

Soil is the largest terrestrial pool of carbon (C), storing 1395-2293 Pg C. Under changing climate a large portion of soil C could potentially transfer back to the atmosphere as CO₂, pushing the earth system into a positive feedback loop between increasing soil CO₂ emissions and rising temperatures. We rely on models to estimate soil responses to climate change; however recent global carbon cycle model intercomparisons have shown poor model performance in capturing C cycle processes in the soil. To gain more confidence in the range of potential soil C emissions over a long-term period of climate change it is necessary to match the model outputs with the observations. Studies have shown that data-model fusion techniques can improve conventional models' spatial representation of observed soil carbon, and that adding an explicit microbial component further improves model performance. The soil carbon datasets used to inform conventional models often do not contribute much information to model parameters associated with processes happening on a short time scale, such as turnover rates of fast organic matter, rates of enzyme production and loss, activation energy of microbial uptake of dissolved organic carbon, etc. In this study we (1) inform conventional and microbial soil models with observed time series of soil respiration and estimate the optimum parameter sets for various plant functional types, (2) explore the drivers of variability in the optimum parameters across plant functional types, and (3) investigate whether including explicit microbial dynamics improves model performance compared to that of a conventional model.