Message from the DirectorA. E. "Sandy" MacDonald, Director
This website details the activities of the Forecast Systems Laboratory during the last year, and discusses our plans for the future. In this introduction, I would like to highlight a few exciting new developments within the laboratory.
A very important accomplishment of the year was the installation of a Massively Parallel Processor (MPP). This system, which was built by HPTi Corporation and named Jet, was installed in December, and announced in a press conference on 26 April 2000. The computer, as installed initially, consists of 270 fast processors based on Compaq's Alpha architecture. When fully configured, in about two years, it will consist of over 1,500 processors with a peak speed of over 5 trillion arithmetic calculations per second. It brings FSL into the ranks of NOAA supercomputing, and will allow rapid advances in a number of areas. These include the development of very high-resolution mesoscale models. FSL is a participant, along with a number of organizations (including the NWS National Centers for Environmental Prediction and the National Center for Atmospheric Research) in the development of the Weather Research and Forecast (WRF) model. This model will be fully nonhydrostatic, and is being designed to serve both the research and operational communities. Since it will be running at resolutions as small as a kilometer, the availability of the HPTi supercomputer will be crucial to FSL's role in the development and testing of WRF. A second very important use for Jet will be in testing future observing systems. The North American Atmospheric Observing System is a program to study the composite observing system of the future. By running observing system tests, the best mix of systems can be identified. Observing system simulations require extraordinary amounts of processing, since they are accomplished by running models repeatedly with different inputs. Finally, FSL will be participating, along with NASA, NOAA NCEP, and NOAA NESDIS, in an effort to develop and improve techniques for better use of satellite data in numerical weather prediction. The MPP will allow rapid progress in these three important efforts.
FSL has continued an important role with the AWIPS system of the National Weather Service (NWS). FSL developers have improved AWIPS significantly in the software for "build 5," which will be installed nationally during the coming year. Similarly, FSL has played a major role in the expansion of the GLOBE program to more countries and more schools. FSL's FX-Net and World Wide Web Workstation are examples of new systems being developed. In its efforts to develop paths for NWS to keep AWIPS current in technology, FSL has been investigating the use of the Linux operating system, high speed networks, and PC technology.
Finally, FSL has made significant progress in observing technology. During the last year, a small network of profilers in Alaska were brought to operations and turned over to NWS. Our demonstration network in the central United States reached new highs in reliability and usage. FSL investigated new techniques for recovery of water vapor using GPS, and made significant progress in the design of observing systems from balloon and automated aircraft.
In summary, we believe that you will find this review both interesting and surprising. Our job is to anticipate the science and technology that will be needed in the nation's operational services in the next 5 to 10 years. More than ever, the rapid pace of technological change requires that this type of effort be made. FSL's record at anticipating change is excellent, so this review can be seen as a look into the future of operational weather prediction.