Automated Weather Forecasts

Numerical Weather Prediction

WDT has developed its own operational mesoscale NWP system based on based on the Advanced Research version of the Weather Research and Forecast (WRF) model.  WRF is the premier NWP system in the world and continues to be advanced via a large community of leading researchers.  WDT staff members have participated in the development of WRF and strive to incorporate the latest cutting edge science with each new release so that our customers can leverage these advantages much quicker than government-provided services offer. 

Unlike many users of WRF, WDT employs an advanced data assimilation technique that combines objective analysis with Four Dimensional Data Assimilation.  This system has the unique capability to assimilate geostationary weather satellite imagery and WDT’s quality-controlled three-dimensional radar mosaics in addition to virtually every type of operational meteorological observation systems.  This ensures optimum initialization and better forecasts than simply nesting WRF within a larger scale model.  WDT currently provides over 100 operational runs of WRF each day on a robust high-performance computing system located in its world class data center, allowing customers to rest assured that their data will be there when they need it.

In addition to WRF, WDT has partnered with NanoWeather to implement the award-winning Uncoupled Surface Layer (USL) microscale forecast model to provide the most precise and accurate forecasts of surface temperature, winds, and humidity possible.  USL can be configured to provide forecasts for precise individual locations or run as a grid of points at very high resolution (as low as 50 meters).  The combination of WRF and USL provide the foundation of WDT’s new WindPredictor system that provides the most accurate forecast of wind power down to the resolution of individual turbines.  When used as a corrector to any forecast model, USL provides accurate forecasts for any city in the world and is one of the key inputs to WDT’s Worldwide Point Forecasts available in iMap. 

In addition to the internally-created NWP content, WDT has an extensive data ingest capability that allows us to harvest a wide variety of numerical forecast products from many different providers, including domestic and international government providers as well as some specialized commercial providers.  These data provide additional input to WDT’s algorithms and decision tools and can also be customized for delivery to our clients.  These include marine forecast data from NOAA and ECMWF (waves, currents, sea ice, chlorophyll), global and regional atmospheric models from the National Weather Service, etc.

Advanced Algorithms

Raw NWP model output is not very useful to decision makers.  The NWP expertise and experience within WDT, combined with our operational focus, has resulted in the development of a proprietary post-processing capability to convert raw model fields into parameters that can be applied directly to the client’s particular problem.  For example, WDT’s NWP services support the aviation industry by applying the latest algorithms to diagnose turbulence, icing, and low-level wind shear hazards directly from the raw output and provide these derived parameters in standard formats directly to the end users’ processing systems. To support our energy clients, we employ algorithms that help predict the likelihood of wind ramp events (sudden changes in wind speed over a wind turbine farm) as well as the potential for lightning and wind that can lead to power outages. 

Worldwide Point Forecasts

WDT has developed a system that optimally blends a combination WDT’s various sources of NWP output (including the WRF and USL models) as well as the National Weather Service’s National Digital Forecast Database to produce site-specific forecasts for over 100,000 global locations.  The locations include every known airport and populated location in the world.  These forecasts are available with hourly temporal resolution out to 7 days and are updated four times per day.  This service provides the ClickCast capability available on iMap.  For customers requiring highly accurate forecasts for locations not already in the database, WDT can quickly and easily add new locations, to include the use of the USL model and its capability to “self-tune” based on observations and past performance.  This service also allows organizations with weather forecasting responsibilities to automated the basic elements of their forecasts (temperature, winds, humidity, cloud cover, preciptation, etc.), thereby allowing their meteorologists to focus their efforts on the more difficult and mission critical issues of the day.