The next generation of global climate models are soon to perform numerous simulations with a horizontal resolution on the order of a few tens of kilometers (10 to 50 km), minimizing or potentially eliminating the need for downscaled projections. In addition, higher resolution (on the order of a few km) datasets for impact analysis are available from dynamical downscaling or statistical methods. However, high-resolution simulations/datasets are not guaranteed to be synonymous with high fidelity, usability, or relevance to impact research and consideration to each of these areas is required to maximize their benefit to science and society.
For example, while higher resolution tends to improve the representation of tropical cyclones and topographically forced features, results with a current set of climate models have shown that it can also distort extreme precipitation and cloud representations. Meanwhile, it is rare that detailed information is available to prospective users regarding which impact-relevant features of the climate are represented well in high resolution data products. Thus, for advances in resolution to aid decision-making and address the broader climate information usability gap, the process of model evaluation and development would benefit from greater input from users of climate information, as well as input from social science research.
As global models and datasets approach the spatial scales relevant for impacts on human systems, there is a need to understand many cross-disciplinary questions, including:
Which are the phenomena that are most important for climate impact and adaptation analyses? How well do climate models represent those? Can we define impact-based metrics to evaluate climate models?
Which features of the climate system improve with increasing resolution? Which do not improve, and why?
Can high-resolution simulations be paired with lower resolution simulations to explore multiple scenarios and more fully characterize uncertainty in the changing frequency of extreme events?
What are additional barriers to the usability of high resolution datasets for impacts research, planning, and decision-making that should be addressed during model development and analysis stages?
The answers to (or paths towards answering) such questions would provide critical information to the impacts and adaptation research community to support uncertainty quantification and appropriate use of simulation data, while providing a missing feedback mechanism to the climate modeling community regarding the decision-relevance of its outputs, ultimately improving the usability of future scenario projections and decadal predictions.
This Aspen Global Change Institute session will gather together approximately 30 researchers and stakeholders across diverse disciplines and applied contexts. Researchers in physical climate modeling, impacts, vulnerability, and adaptation research, and social science research on the usability of science will join with senior decision-makers and stakeholders representing four sectors of focus:
Coastal planning & management
Water resources management
Public health & safety
It is anticipated that the session will include a number of international participants, as well as several postdoctoral and early career scientists.
Proposed outcomes of the meeting include a preliminary set of impacts-relevant model evaluation metrics with associated tolerance levels, as well as a community strategy for refining these diagnostics over time and incorporating them into model intercomparison exercises such as CMIP6. We also expect that the meeting will foster new collaborative relationships and ongoing dialogue between the physical climate and impacts and adaptation research communities.
Workshop Topic (s):
- Climate Variability and Change (including Climate Modeling)