Using High Resolution Weather Model Output in Support of the HYSPLIT Dispersion Model
The NWS in Melbourne (MLB) Florida worked with NASA's Applied Meteorology Unit (AMU) in 2009 to install the HYSPLIT model on NWS-MLB servers for operation within the local forecast office. The HYSPLIT setup allows for the production of routine trajectory and concentration information for pre-positioned locations of concern, as well as the ability to add additional locations as the need arises. The NWS-MLB HYSPLIT configuration also has the ability to generate emergency on-demand concentration plumes for locations that forecasters request through their Advanced Weather Interactive Processing System (AWIPS) workstations. These capabilities give the MLB meteorologists ready access to HYSPLIT dispersion output based not only on high resolution models from environmental modeling centers, but also based on high resolution output from locally configured Weather Research and Forecasting (WRF) models. Several NWS-MLB WRF model configurations are at higher spatial and temporal resolutions than that afforded by national scale guidance. Wind fields which are often governed by transient mesoscale features such as sea breezes and convective outflow boundaries (which are common in Florida) can be more accurately depicted. In turn, this improves the quality of the dependent HYSPLIT plume forecast. It has been suggested by Bowman et. al (BAMS, 2013) that advances in the quality of output from Lagrangian type dispersion models, such as HYSPLIT, can occur with higher spatial and temporal input model data such as that provided by a optimally configured WRF.
This paper will discuss the HYSPLIT configuration at NWS-MLB and describe how local forecasters interact with the routine, as well as emergency on-demand, dispersion run capabilities. The benefits of using the high spatial and temporal resolution WRF model to provide input to HYSPLIT for decision support services will be discussed, along with the associated leverages toward a weather-ready nation.