»WeatherFlow RAMS Cluster

Since 2004, WeatherFlow has operationally run its own numerical model, RAMS, developed by meteorologists at Colorado State University. RAMS is run at higher resolutions (down to 2km) than the traditional numeric models (12km+) used by the National Weather Service. WeatherFlow is continually expanding RAMS grid coverage across the United States.

»WRAMS Advantage

The advantage of WeatherFlow’s modeling is that we take our model to the next level…WRAMS. By consulting with the model’s developer, Weatherflow’s Meteorologists and Real-time Mesonet are able to provide accurate feedback to improve the accuracy of the RAMS model. This feedback mechanism ensures that the WRAMS model is always improving and becoming the most accurate short range model available.

A well-established fact in the world of atmospheric modeling is that the quality of forecasts is directly related the successful input of as much high quality atmospheric data as possible. The exploitation of this relationship is even more pronounced in the world of mesoscale modeling where resolving weather variability on the order of kilometers is critical. WeatherFlow’s own mesonet database is constantly growing with our ”home-grown” quality reporting sites, plus a growing set of other carefully selected data sources are assimilated into each model run to yield a superior initial state at model start-up. The WeatherFlow unique mesonet/model tandom equates to a superior solution that no other institution can equal.

Mesoscale models are initialized by utilizing input from larger-scale synoptic scale model input fields. An inherent limitation to mesoscale model forecast performance is the accuracy of the “parent” synoptic scale model. This limiting factor can be minimized by having the freedom of choosing among more than one synoptic model to initialize a mesoscale model run.

Quite often, model performance is compromised during the first few hours of output because input data contains valuable information on the moisture distribution in the troposhere, but does *not* initially know where moisture is sufficient to spawn cloud development. That is, if clouds, especially of a convective nature, are actually present early in the model run, they are often un-diagnosed until a few hours into the run. Furthermore, winds associated with precipitation will be absent. In an effort to eliminate this model difficiency, WeatherFlow is releasing a “warm start” version of its model that, in essence, starts model run data input and moisture computations in a staggered sequence to allow clouds to be forecast by hour one.

»Sample Applications

Chesapeake Bay Sea Breeze Experiment
Radar Propagation - Navy Dahlgren