Ensheng Weng
Terrestrial vegetation, as a key component of the Earth system, defines the boundary conditions of land surface for the exchange of energy, momentum, and water vapor between land and atmosphere, regulates long-term atmospheric CO2 concentration, and thus deeply shapes Earth’s climate dynamics. Dynamic global vegetation models (DGVMs) are normally used in Earth system models (ESMs) to simulate plant physiological activities, vegetation dynamics, ecosystem biogeochemical cycles, and land surface characteristics for atmospheric components. DGVMs bin vegetation into a small number of plant functional types (PFTs) and predict the geographic distribution of PFTs by bioclimatic and physiological rules. These models are able to track ecosystem carbon and/or nitrogen cycles as pools and fluxes, and predict their feedbacks on climate systems at large spatial scales. However, these models are unable to predict transient vegetation compositional and transient changes because of underrepresentation of functional diversity and lack of detailed demographic processes. Vegetation demographic models (VDMs) are thus developed to explicitly simulate demographic processes and individual-based competition for light and soil resources. In VDMs, the stochastic birth, growth, and mortality processes replace the deterministic carbon processes that are in current DGVMs, potentially altering the representation of vegetation dynamics and carbon cycle. The inclusion of individual-based competition effectively implements adaptive dynamics – a method used in evolutionary game theoretic analysis to determine the best fit strategy in a given context – into ESMs and thus significantly increases the functional diversity of PFTs. In this presentation, I will summarize our studies on the modeling of stochastic disturbance effects on ecosystem carbon dynamics, vegetation demographic processes, and competitively dominant plant traits to show how the fundamental plant physiological and ecological processes are represented in vegetation models and thus significantly increase model predictive skills. With the case studies, I will show how the predictions of vegetation dynamics and carbon cycle are improved by exploring plant competitively dominant strategy in response to elevated CO2, variations of soil nitrogen and precipitation regimes. I will also discuss the possible approaches of bridging the gaps between vegetation modeling and conventional ecological studies for improving Earth system modeling.
