Why are we having this meeting?

            
The scheduling of this meeting was strongly motivated by NWS plans to present forecasts on a fine-resolution grid. The NWS's experimental National Digital Forecast Database (NDFD) has a nominal grid-spacing of 5 km across the United States and represents a blend of objective forecast guidance and forecaster edits. Early evaluation efforts of the new gridded forecasts have been hampered by the lack of any matching, gridded analysis of the various forecast parameters, which include, among others, temperature, dew-point temperature, wind, precipitation, clouds, and weather. Thus, the National Weather Service has an immediate and critical need to develop a process to produce real-time, NDFD matching resolution, analyses, which has begun to be referred to as an "Analysis Of Record (AOR)".

Yet, this is not solely a National Weather Service need. In fact, demands for similar, high-resolution objective analyses are growing rapidly across our community. These demands rise from many facets, including: a multitude of mesoscale modeling efforts for both operational weather forecasting and fundamental scientific investigation; dispersion modeling for real-time prediction of hazardous materials, air pollution, and homeland defense; environmental issues from the coastal zone to national forests, including fire management; and, a number of other applications requiring environmental information on high temporal and spatial scales. Additionally, accurate high-resolutions analyses form the basic building blocks of a climate database for use in statistical applications that help to assess the impacts of climate change.

This meeting is intended to bring together various groups who are already working on mesoscale objective analysis methodologies, and to discuss advantages and disadvantages of these methods. Ideally, partnerships can be formed and ideas scoped out that will organize our community's attempt to meet these growing demands for high-resolution analyses. To provide some focus for this meeting, a strawman proposal was requested from Geoff DiMego of the Environmental Modeling Center, National Center for Environmental Prediction. A draft presentation of this proposal is available here. Other groups may wish to provide draft proposals prior to the meeting so that attendees come prepared to discuss the pros and cons of all of the approaches. Contact one of the cochairs if interested to do so.

What are the operational requirements for the AOR?

For the National Weather Service, the multi-parameter analyses need to be provided in real time at NDFD-matching resolution (5 km grid spacing, possibly 2.5 km), and on an hourly frequency. Latency of no more than 30 minutes is also required from nominal time of observation to delivery. Parameters are many and diverse, including both basic parameters like temperature and wind, to weather type and intensity, sky cover, and freezing level.

What are the science issues that may limit implementation of the AOR at the outset? What science questions need to be addressed over the next several years in order to lead to improved AORs in the future?

Some fundamental science questions regarding how to relate specific physical processes at local scales to gridded fields at 2.5-5 km need to be addressed for this project to succeed. For example, to what extent can analyses resolve the temporal and spatial extent of severe weather events in order to develop adequate statistics for their occurrence? Presuming that the analyses are likely to fail to develop adequate statistics for many types of severe weather, how can that limitation be assessed quantitatively? How do parameterizations of the boundary layer, convection, soil moisture, and other physical processes that may be used in a model to provide the background fields affect the AOR as a function of time of day, season, region?

What are the research & development issues that need to be resolved at the outset of the AOR project? What R&D issues need to be addressed over the next several years?

Clearly the AOR will be sensitive to the choice(s) for the data assimilation methods employed and the types and quality of the data used. Which assimilation systems show the greatest promise? Can statistical downscaling approaches, such as PRISM, be utilized? What can be learned from the global and regional reanalysis efforts? What, if any, data collection efforts are required? What are the operational/resource limitations for completing AOR?