Using Earth system models and long-term observations to generate ecological hypotheses

Abstract:

In the geoscience community, Earth system models are typically used to investigate carbon cycle – climate feedbacks at global scales, but increasingly this class of models applies ecological understanding to ask biologically relevant questions. Fewer studies, however, explore the utility and feasibility of using Earth system models to generate and evaluate ecological hypotheses. Here, we assess simulations from the Community Land Model (CLM, the terrestrial component of the Community Earth System Model) configured to represent an extreme tundra environment at the Niwot Ridge Long Term Ecological Research (LTER) site in Colorado. The Niwot Ridge LTER has the highest continuously operating weather station in North America (‘D-1l’, 3700 m asl), with climate data extending from 1953 to present. The site also has what are believed to be the highest operating eddy-covariance towers in North America (‘T-Van’, 3500 m asl). Using these observations as well as ancillary data from the Niwot LTER, we simulated four distinct vegetation communities that characterize this alpine tundra site. We demonstrate how CLM reproduces carbon, water, and energy fluxes that broadly match observations and generates testable hypotheses about the shifting nature of energy, water, and nutrient limitations across space and through time in this heterogeneous landscape. Our results highlight the value in using diverse data streams related to hydrology, eddy covariance measurements, and biogeochemistry to critically evaluate and improve process-level representations in Earth system models at multiple scales of interest. These findings also illustrate how models like CLM can be used to generate hypotheses that ultimately guide further experimental work and model development.