A coupled vegetation–mammal modeling framework for Earth system models

Terrestrial vegetation models are typically based on quasi-empirical or dynamical relationships that link climatic, soil, nutrient, and land-use conditions to the presence of different plant taxa, biomes, or plant functional types. The majority of these models neither include plant seed dispersal nor herbivore grazing effects, potentially leading to a misrepresentation of climate-biosphere feedbacks. Both of these factors, however, are known to play a critical role in the global distribution of plant types. 

In this presentation, we will introduce a new coupled global model with a 1x1 degree horizontal resolution that integrates vegetation dynamics with mammal herbivory. The vegetation model (ICCP global vegetation model, IGVM) simulates the fractional cover of grass, shrubs, trees, and desert using coupled reaction-diffusion equations. These quantities are translated into net primary productivity (NPP) through a non-parametric empirical method. The mammal model (ICCP global mammal model, IGMM) - also based on coupled reaction diffusion dynamics - realistically simulates the biomass of over 2,100 mammal species worldwide and has been extensively validated against observational datasets. The NPP from the vegetation model determines the biomass of individual herbivore species through their carrying capacities. In turn, herbivore grazing acts as a sink for vegetation carbon, and this effect is mapped back onto grass, shrub, and tree fractions in the vegetation model. Furthermore, the impact of mammal-mediated seed dispersal can be estimated.

By running the fully coupled model with and without mammal grazing, we determine the impact of mammal distributions on pre-Anthropocene global vegetation biogeography. This allows us to directly test the “Zimov Vegetation Hypothesis” and document the effects of trophic coupling on ecosystem functionality and stability at global-to-regional scales. We will further discuss how this new coupled modeling framework can be implemented into Earth System models.