Diagnosing and evaluating the dynamics of terrestrial carbon cycle models
Abstract
Terrestrial Biosphere Models (TBMs) simulate the dynamics of the land carbon cycle as a cascade of C fluxes between pools, ultimately driven by the assimilation of C from atmospheric CO2 by photosynthesis (gross primary production, GPP). As a simplification, their dynamics can be described by linear dynamics, leading to a source-driven behavior of the terrestrial C cycle. Source-driven models have been criticized for their inability to predict widely observed phenomena, including nutrient limitation of biomass growth, the tree growth rate-longevity trade-off, and soil C dynamics. However, whether models actually behave in a source-driven fashion is not clearly understood nor quantified or systematically evaluated against empirical evidence. Here, we introduce a framework to diagnose linearity of TBM dynamics by quantifying the ratio of relative enhancements in simulated pools and fluxes, assessed from simulations where CO2 is increasing. Our analysis of a comprehensive set of TBMs reveals a tightly linear relationship between GPP and NPP for almost all models. Exceptions are apparent in models with a coupled N cycle, where the relative increase in NPP is greater than the relative increase in GPP, likely due to the feedback between intensified N limitation, decreasing tissue N content, and resulting reductions in respiration rates. No general linear relationship is simulated for the link between cVeg and NPP, and between cSoil and NPP. This analysis demonstrates that current TBMs clearly deviate from a purely source-driven behavior and reveals links between model structure and emergent dynamics in terrestrial carbon cycle simulations.
- Publication:
-
AGU Fall Meeting Abstracts
- Pub Date:
- December 2021
- Bibcode:
- 2021AGUFM.B53A..07S