Weather Research and Forecasting (WRF) Model

WRF NEWS !

(Octobert 08, 2009)

We are now running WRF v3 three to four runs per day. The 3.1 upgrade has been delayed, and is expected in late October. Our runs are initialized at 00,06,12 & 18UTC with the 12km NCEP NAM. There may be a partial 18UTC runs as a result of testing.

There have been numerous challenges which after failed testing, we will await the 3.1 upgrade. There is no BUFKIT data and the Maine web graphics have been permanently turned off. There have been occasional missing, and premature termination of runs due to issues related to OS upgrades at UAlbany. We’ve solved this by running it on a cluster using the previous OS for the time being, but the runs take longer. These issues have been less frequent since we’ve installed a patch from the developer.

The new version of the WRF has numerous physics and configuration options. We will run with the baseline until the model run is stabilized. In the late Fall we plan to add a 1km nest in support of a research project covering the Hudson River Estuary. Convective Precipitation Graphics are blank for the HiRes/CWSU nest since they use explicit convection.

Warren Snyder – Science & Operations Officer, WFO Albany, NY

The WEATHER RESEARCH and FORECASTING (WRF) models displayed here are experimental local high resolution mesoscale models, run in real time in collaboration with UAlbany. The models are run at the Research IT (RIT) high performance computing center at UAlbany, and transferred to WFO Albany, New York as part of the CSTAR projects.  They are used experimentally and to support model studies at the Albany & Upton, New York, Grey & Caribou, Maine Weather Forecast Offices, and the Center Weather Service Units at Nashua, New Hampshire and Oberlin, Ohio. There are three domains (model areas) in this package. 

The Northeastern US WRF domain covers much of the United States from the Mississippi Valley east, as well as adjacent areas of Canada. This domain is 15km resolution and is designated D01. It is meant as a big picture overview, and provides hourly output to the inner domains. At this scale Kain-Fristch Convective parameterization is used, and Convective Precipitation graphics are produced. There are two smaller domains nested inside the large regional domain and convection is explicit, therefore the Convective Precipitation Graphics are blank. This version of the WRF is an ARW core, or MM6 based core.  It is a non-hydrostatic model.

The other two domains are D02 and D03 and are at 5km resolution. D02 covers Maine and surrounding areas. D03 is the HiRes WRF and covers both the CWSU Oberlin and Nashua service areas as well as much of the area from the Great Lakes to New England. The HiRes domain (D03) is part of an ongoing CSTAR project with the CWSUs at Oberlin and Nshua, New Hampshire, on developing ways to use mesoscale models to improve aviation forecasts, and reduce ground stops and ground delays (the events that keep you stranded in the airport).  This data is included by CWSU meteorologists as part of the forecast process for the air traffic controllers at the Cleveland and Boston Centers.

The WRF uses a variety of state of the art physics and dynamics packages. It is particularly adept at forecasting local effects, due to mountains, coastlines and lakes, winds, as well as localized heavy precipitation. It also is useful in identifying areas where thunderstorms are likely to form first. The D03 domain produces BUFKIT soundings for dozens of locations in the area it covers, and low level wind forecasts are developed from this data.

Currently the WRF is run twice per day with a 00 UTC and 12 UTC initialization time. The data on this page is updated as the model runs, so be aware that when you loop the data you may get data from two different runs.  Typically data from the runs begins to reach the web by 615am/pm, and finishes by 2pm/2am EST.

The following issues remain to be resolved with the graphics:

This project is managed by Science and Operations Officer Warren R. Snyder and Information Technology Officer Vasil T. Koleci. Address any modeling questions to warren.snyder@noaa.gov .  Problems with the webpage should be addressed to vasil.koleci@noaa.gov .