Climate modeling has evolved from physical atmosphere models with specified atmospheric constituents, to today’s earth system models coupling physical, chemical and biological land, atmosphere and ocean components. As newer components are added, a major problem is assuring that the underlying processes and coupled dynamics are accurately portraying the earth system. Most Earth System models now include active, predictive carbon cycling, a huge step forward. However, carbon simulations have not been compared against observations with the same rigor as done for simulated physical climate over the past 30 years (e.g. past IPCC reports). For climate projections over the next decades and beyond, carbon cycle feedbacks represent approximately 40% of the uncertainties associated with the physical climate system. Therefore a grand challenge is to develop models that can accurately simulate the carbon cycle, not just of the land or ocean system in isolation, but for the full coupled climate-carbon system.

Currently, less than 50% of the emitted anthropogenic carbon dioxide remains in the atmosphere, with the remainder taken up by the land and ocean. Whether this will continue in the future and whether land and ocean carbon reservoirs will remain stable is not well known. Studies from the previous IPCC assessments indicate large differences in model predictions of future atmospheric carbon dioxide concentrations, even with similar human emissions. The wide range reflects model-to-model differences in the sensitivity of land and ocean carbon uptake to atmospheric CO2 and climate as well as the sensitivity of the physical climate system to increasing levels of atmospheric CO2. The substantial variations in carbon model formulations produce large differences in carbon-climate feedbacks. We argue that this large uncertainty could be reduced by comprehensive data-based evaluation and improved model parameterizations.

Carbon cycle models also provide critical information about climate change impacts on humans and ecosystems including changes in net primary productivity, standing biomass, biogeochemistry and ocean chemistry. Thus, improvements in the carbon cycle models would allow for better projections of climate, climate impacts, requirements for adaptation, and vulnerability assessments. Finally, improved carbon cycle models would allow scientists to better inform policy makers of whether climate change mitigation policies are likely to be successful and whether policies that are put in place are working.

With NSF support from this grant, the research team from the Woods Hole Oceanographic Institution (WHOI) collaborated on evaluating the simulated carbon cycle within the Community Earth System Model, a publically available model that is widely used to study the mechanisms connecting physical climate to the biogeochemistry and ecology of the atmosphere, ocean and land biosphere.

As part of the international RECCAP (REgional Carbon Cycle Assessment and Processes) program, we quantified the pattern and magnitude of air-sea CO2 fluxes across the globe and for specific basins: Arctic and Atlantic Oceans, Indian Ocean, Southern Ocean, and Pacific Ocean. The analysis identified specific regions where models such as CESM have particular problems simulating the real ocean, providing a starting point for future model development and improvement. In a similar fashion analyzed simulated and observed patterns of ocean particulate organic carbon sinking flux and relationship to plankton community structure and ballast material, resulting in more systematic indication of model biogeochemistry biases.

We documented data analysis and modeling methodology used in international Global Carbon Project to quantify trends and inter-annual variability for human CO2 emissions as well as ocean, atmosphere, and land CO2 sources/sinks. The Global Carbon Project an...

Please report errors in award information by writing to: awardsearch@nsf.gov.