x

Preliminary Workshop Announcement: Development of Predictive Carbon Cycle Science

Location
Washington, D.C.

Workshop Date: March 7-March 9, 2016 (2.5 days)

Workshop Venue: NOAA NCWCP Conference Center, College Park, Maryland

The objective

This invitation-only workshop will gather collective knowledge on predictive carbon cycle science from the community. We will generate a roadmap on how to further develop and enhance our capacity to predict future carbon dynamics in North America. The possibility to establish a new project “Development of Predictive Carbon Cycle Science” for the second decade of NACP will be explored. The workshop outcomes will also inform the 2nd State of the Carbon Cycle Report (SOCCR-2).

Background

The first decade of NACP has greatly advanced our observation and monitoring systems of carbon cycle research.  For example, the long-term AmeriFlux network, primarily under DOE support, has been successfully established for nearly two decades to quantify exchange of CO2 between the atmosphere and land surface. NASA has launched OCO-2 to retrieve a global geographic distribution of CO2 sources and sinks. NOAA’s Global Greenhouse Gas Reference Network measures the atmospheric distribution and trends of CO2, CH4, and N2O at 4 observatories and 8 tall towers, from air samples at more than 50 sites, and by small aircraft. Those observational and monitoring programs are essential to improve understanding of the carbon cycle through diagnosis and attribution. Their values would be much augmented if their results can be used to constrain future projections of carbon cycle dynamics in response to climate change and human activities.

Traditionally, model has been a primary tool for projecting the future states of carbon cycle dynamics. For example, biogeochemical models have been incorporated into ESMs to predict responses and feedbacks of the carbon cycle to climate change. However, the current generation of Earth system models (ESMs) makes widely different predictions across models, either in direction or magnitude, and fit observations poorly. This model-to-model variation has been shown in any model intercomparison projects (MIPs), including Coupled Model Intercomparison Project Phase 5 (CMIP5) and Multi-scale Synthesis and Terrestrial Model Intercomparison Project (MsTMIP). Even more daunting is that current ongoing research has not effective to improve projections but rather widen the uncertainty. The uncertainty cannot be easily explained in any of those ongoing MIPs.

If the uncertainty issue is not adequately addressed, terrestrial carbon sink potentials cannot be fully understood, carbon-climate feedback cannot be well quantified, and therefore future climate change cannot be predicted with confidence. As a consequence, policies to stabilize CO2 concentrations will fall short in meeting targets of climate mitigation or incur unnecessary costs.

Developing predictive carbon cycle science is an important step toward carbon and climate management, which are overarching challenges facing the global community.  Decisions concerning carbon management are most effective if informed by the best possible scientific understanding.  The latest U.S. Carbon Cycle Science Plan (2011) calls throughout the document for a focus on the application of carbon cycle science to the issues of carbon and climate management.  Integral to this topic are the need to 1) incorporate state of the art findings in carbon cycle sciences into integrated assessment models, 2) improve our ability to quantify uncertainties in carbon cycle predictions, 3) improve the prognostic skill of carbon cycle models, 4) engage with existing international efforts in this area and 5) expand the research agenda to those areas of social sciences, humanities and engineering integral to and engaged in the study of carbon and climate management. 

Approaches

To develop the predictive science of the carbon cycle, we need (1) theoretical understanding of fundamental properties of the carbon cycle system that determine its future trajectories, (2) assessment of intrinsic predictability of the carbon cycle system in response to exogenous forcing variables, (3) a new generation of carbon cycle models with their realized predictive abilities in accordance with system’s predictability, that is, the models have well-constrained structures and parameters of their components for which we have solid theoretical and empirical understanding while alternative hypotheses are explored for less-understood components, (4) synthesis of all relevant data to constrain model structures and parameters and meanwhile to continuously improve our understanding of those less-understood carbon cycle components, (5) technology that can assimilate highly disparate and heterogonous datasets into ESMs, and more importantly (6) a new integrative research and education culture that not only promote multi-disciplinary collaboration but also facilitate the fusion of theory, data, experiments, observation, and models.
 

Key questions to be discussed during the workshop

  1. What are the predictive skills of current regional and global carbon models?
  2. What are the reasons that the current generation of carbon cycle models has low predictive skills?
  3. How much are model predictive skills related to the following issues?
    1. The knowledge on intrinsic predictability – the fundamental properties of the system that determine the future trajectory of that system in response to changes in external variables
    2. Measure of the difference between system’s intrinsic predictability and model’s achieved predictive ability
    3. Technical and scientific limits to realize the intrinsic predictability
    4. Demonstrable projects for predicting carbon cycle dynamics
    5. Research culture that may not be conducive to predictive science  
    6. How can we synthesize results from experimental, observational, and modeling studies to improve model predictive skills
    7. How closely are predictive skills of models related to effectiveness of supporting policy-making?
    8. What would be your five recommendations that can be effective to improve predictive carbon cycle science?

Expected outcomes

We plan to write one manuscript targeted as Science Policy forum on predictive carbon cycle science vs. effectiveness of supporting policy-making.

We plan to publish a meeting report on EOS to summarize key discussion points and consensus of the workshop.

The above and other workshop outcomes will also inform related topics in the 2nd State of the Carbon Cycle Report (SOCCR-2), a special USGCRP Sustained Assessment Report being led by the Carbon Cycle Interagency Working Group/U.S. Carbon Cycle Science Program.

 

Organizing committee (tentative)

one “monitoring & attribution (M&A)” and one “predictive modeling (Future)” expert for each of the four domains or disciplines.

  • Wei-Jun Cai, U of Delaware (M&A, Ocean/coastal)
  • Scott Denning, Colorado State University (Co-chair) (Future, Atmosphere)
  • Kevin Gurney, Ariz State (M&A, Emissions)
  • Gretchen Keppel-Aleks, U Michigan (M&A, Atmosphere)
  • Libby Larson, NASA Goddard, NACP
  • Yiqi Luo, University of Oklahoma (Co-chair) (Future, Land)
  • Yude Pan, USDA FS (M&A, Land)
  • Adam Schlosser, MIT (Future, Atmosphere)
  • Gyami Shrestha, U.S. Carbon Cycle Science Program Office/CCIWG, USGCRP/UCAR (Co-chair)

 

For additional details and questions, please contact workshop lead Dr. Yiqi Luo, University of Oklahoma.