Developing a testbed at the ARM Climate Research Facility for Coupled Climate-Carbon Cycle model evaluation

Description

Climate simulations for the Fifth Assessment Report of IPCC (AR5) will utilize Coupled Climate-Carbon Cycle models (C4 models). Unfortunately, the terrestrial carbon components of these models are largely untested, particularly in coupled mode. Testing the coupled model performance requires both new data products and validation protocols. The Berkeley Lab/ARM Carbon is measuring terrestrial fluxes of CO2, water, and energy as well as concentrations of greenhouse gases and their isotopes from the surface to 15,000 feet above sea level, in the Southern Great Plains. We are also collaborating with NASA and JPL in estimating whole-column CO2 by upward-looking Fourier Transform Spectrometry. In this poster we describe our work to make the ARM Climate Research Facility a pre-eminent facility for evaluating C4 models, by integrating carbon cycle observations, forcing data, site information, and modeling analysis. By combining our efforts with other ASR and ARM efforts, we will create a testing platform for many carbon and non-carbon components of C4 models. Our focus will be on daily to multi-year time scales rather than dynamic vegetation and other long-term processes. The model-testing data sets may become a value-added product (VAP) for the ARM Climate Research Facility. In the first phase of this work we are focusing on testing the land-surface component of C4 models. We have collected AmeriFlux type data in grassland, wheat, and other dominant land-cover types, and have tower and aircraft data for concentrations, as well as observations of soil temperature and moisture, plant biomass, and 13C in soil and vegetation. These data will be used to test model predictions of plant physiological function (e.g., water-use efficiency, light-use efficiency), diurnal and seasonal cycles in land-surface C and energy exchanges, Bowen ratio, and land-surface responses to cloud type and cover. The models will be compared to regional-scale estimates we have made using a highly tuned land-surface model. In the second phase, we will provide data to test the fully coupled models. The coupled runs will be performed in both regional (e.g., WRF-CLM) and global (e.g., CCSM) climate models. Consistent boundary forcing will be specified for all the regional climate models (e.g., from reanalysis). Predictions of trace-gas concentrations, temperature, humidity, radiation, will be compared to ARM’s extensive atmospheric observations.