A Hierarchical Evaluation Framework for Assessing Climate Simulations Relevant to the Energy-Water-Land Nexus
Actionable climate science entails the provision of local-to-regional scale climate information to support decision-making for climate impacts, adaptation and mitigation. Despite advances in theories, observations, climate modeling techniques, and high performance computing, large uncertainties in model projections of climate change remain. These uncertainties challenge our understanding of the controlling processes and hamper the use and communication of climate information for societal benefits. We will develop a hierarchical model evaluation framework informed by different uses of climate models and their output by climate scientists and stakeholders for planning and managing resources. The framework features a wide variety of measures of model behavior (metrics) that include standard, existing metrics and new metrics developed for the framework. Metrics specifically tailored for the energy-water-land nexus will inform strategies for modeling human-Earth system dynamics at regional scales. The metrics will exploit of a wide range of observational data to evaluate model fidelity and explore the mechanisms and sources of model errors and uncertainty.
To demonstrate the usefulness of this evaluation framework, we select representative approaches for producing climate information for regions from a suite of different modeling methods and design structured, hierarchical experiments that feature baseline simulations across a range of spatial resolutions and modeling approaches. Some of these simulations will focus on the impacts of future land use and land cover changes associated with food and bioenergy crop production and urbanization, and expansion of wind turbine deployment, which highlight specific challenges for modeling the energy-water-land nexus. We will also apply our framework to output from selected coordinated model intercomparison experiments to demonstrate the broader functionality of our framework. Outreach through workshops will foster interactions with climate scientists and stakeholders, informing the development of our model evaluation framework.
The vast majority of our proposed research includes the conterminous U.S. However, some in-depth studies will focus on a specific data-rich region: the Southern Great Plains, which features growing competition for water, large-scale wind energy deployments, and remarkably consistent projected future changes in warm season precipitation. Key outcomes of the proposed research are: (1) A set of methodologies, algorithms, and software for individual and integrated metrics using both standard and high risk / high benefit approaches; (2) Determination of the relative value of the metrics, differentiation of model skill based on the metrics and hierarchical experiments, and understanding of the spatial scale dependence of the results; (3) Elucidation of local human impacts related to the energy-water-land nexus on climate and the requirements for simulating human-Earth systems at regional scales; (4) A rich dataset produced by the hierarchical modeling experiments archived and made available through the Earth System Grid Federation (ESGF) for use by the broader climate science community and stakeholders; and (5) A rigorously tested and community-vetted model evaluation framework and tools that form the basis for future development of a computationally enabled user-friendly model evaluation system for community use.