Extended Abstract (168K)4.3
A Nowcast/Forecast Model of the St. Johns River and Its Performance Assessment
Aijun Zhang, NOAA/National Ocean Service, Silver Spring, MD; and E. P. Myers and F. Aikman III
As part of the Coastal Storms Initiative (CSI), an interdisciplinary NOAA project aimed at providing local communities with enhanced tools and resources to help mitigate the hazards of coastal storms, an experimental circulation model-based nowcast/forecast system of the St. Johns River, Florida has been implemented. The hydrodynamic model of the river and estuary makes nowcasts and forecasts of water levels, currents, salinity, and temperature and provides a means for the community to view the results. Two circulation model applications have been implemented and evaluated in the development stage of the nowcast/forecast system. The first of these model applications uses the Environmental Fluid Dynamics Code (EFDC) to perform operational hourly nowcasts and 36-hour forecasts. This model application was originally developed by the St. Johns River Water Management District (SJRWMD) and calibrated for 1995-1998. The SJRWMD provided the calibrated model to NOAA’s Coast Survey Development Laboratory (CSDL) for it’s implementation as a real-time nowcast-forecast system.
The other model being examined for use in the St. Johns River is ELCIRC, a finite volume/finite difference baroclinic model for unstructured grids. As the EFDC model is converted into an operational mode, the ELCIRC application is likewise being calibrated to see if the same level of accuracy, or better, can be attained. ELCIRC offers some benefits that include: 1) an unstructured grid that can easily be modified to add refinement where needed; 2) an advection algorithm that permits large time steps and thus affords more grid refinement if necessary; and 3) a natural treatment of wetting and drying. The latter is an important consideration for this CSI pilot study, as flooding from coastal storms is one of the critical hazards that local communities need to address in their mitigation strategies.
In order to quality assess the performance of model applications such as these, software tools were developed to compute a standard suite of NOS skill assessment statistics. As part of the Coastal Ocean Modeling Framework that NOS is implementing, all models are first adjusted to output in standardized netCDF formats. The skill assessment software is then designed to read such output from any model and compute statistics using modeled and observed water levels, currents, salinities and temperatures. The statistics address the performance of the models under several types of simulation scenarios in both a hindcast and a forecast framework. The St. Johns River model applications are evaluated using this software, and these statistical results will help guide the best approach for transitioning the models to an operational environment.
Session 4, Development and operation of coastal forecast systems and data assimilation
Wednesday, 12 January 2005, 8:30 AM-5:30 PM
Previous paper Next paper
Recorded presentation