Priority 3: Conducting Research in Computer Science, Applied Mathematics, Statistics, and Numerical Methods

NCAR's research in computational science and math applied to geophysics enhances NCAR's computational resources and produces more efficient scientific simulations. This research is necessary to maintain an innovative computational facility and to lead the geophysics community in incorporating new numerical methods and models. NCAR's forecast verification research seeks to reduce forecast error and the potentially significant economic consequences of this uncertainty. This research spans several disciplines and addresses computational science at many levels such as network flow improvements and how to scale existing codes to different numerical methods for simulating geophysical phenomena. Integrated with computational science are areas of applied mathematics that include data analysis, models for multiscale processes, and techniques for assimilating data into numerical models. CISL's IMAGe group has a “Theme of the Year” to address these issues, and this year's theme was “Emerging Mathematical Strategies for Multi-Scale and Stochastic Modeling of the Atmosphere and Climate”. Because these different elements are coordinated through a single Laboratory at NCAR, technologies and ideas are more easily transferred from prototypes and theoretical results to implementation and workflow, and finally into incorporation as tools and models for the community. There is also a valuable reverse transfer whereby emerging computational capability and data storage spur particular research that takes advantage of these features.

In a separate research thrust, NCAR's forecast verification research seeks to reduce forecast error and the potentially significant economic consequences of this uncertainty. Standard forecast verification and evaluation activities are based on relatively simple measurement techniques that typically deal only with single aspects of overall performance. They generally do not provide information that can be used to improve the forecasts or that can be used for decision making by forecasters. Further, forecasts that are quite useful can have very poor scores when evaluated by using these standard metrics. To overcome these limitations, NCAR develops improved verification approaches and tools that provide more meaningful and relevant information about forecast performance. The focus of this effort is to develop diagnostic, statistically valid approaches that can provide more useful information about forecast performance for forecast developers and users.