Development of the Weather Research and Forecast (WRF) mesoscale model is currently underway, and it is expected that this model will be used as the cloud model within the research community.  Development of the model dynamical formulations, physical parameterizations, and software architecture are underway, so now is the appropriate time to incorporate directly the needs of chemistry and cloud modeling into WRF.  The modeling issues that arise in compiling emissions inventories to be used in the new WRF depend on many factors, most of which are not generalized but may vary every time the model is run:

* the size of the domain used
* the grid size used
* time over which the model is run
* the details of the PBL parameterization, shallow and deep convection transport, chemical mechanisms, and other parameterizations used

The first two factors are most important in the level of detail at which the inventory must or can be compiled.  For example, compiling an inventory with 5 km resolution for the Eastern seaboard of the U.S. requires very large resources.  For short model runs detailed hourly (or better) emissions and the resolution of line sources may be desired; for other run times, weekday/weekend difference with by grid cell emissions may be all that is needed.  PBL parameterizations will dictate whether details are needed to estimate plume rise for major point sources.  Chemical mechanisms used will dictate the speciation needed in the emissions: SO2 alone vs. SO2 and sulfate, speciation detail of hydrocarbon emissions, etc.  In the U.S. the detailed inventory data that would be needed to derive inventories for WRF applications are usually collected by local and state air quality agencies; some of these detailed data are sent to the Environmental Protection Agency Office of Air Quality Planning and Standards in Research Triangle Park, NC.  These data are derived using several methodologies including direct measurement (example: CEMs), use of surrogates (ex: population, traffic counts), etc.  The state agencies can and have developed very detailed inventories (ex: 2 km resolution) for areas covering perhaps 100 km x 100 km for field campaign time periods; usually the state agency has been a participant in the campaign.  However, these detailed inventories are not developed routinely or even as special requests. These are some of the technical questions that participants need to address during the workshop.  Please send me your thoughts, comments, suggestions, etc. via e-mail at cmb@bnl.gov and I will modify this abstract accordingly so we can use it as a starting point in our discussions.