Figure 63. User Interface of the FX-Net client workstation.
The available FX-Net products are categorized into four groups: satellite data, model graphics and observations, radar
imagery, and model imagery. Wavelet transform is used to compress model and satellite imagery. The application of this
relatively new compression technique is critical to the success of delivering very large-size imagery via the Internet in
a reasonable amount of time. The small loss of fidelity in the imagery is acceptable in exchange for very high compression
ratios. Processing time can be further minimized by pregenerating and compressing all satellite data on the FX-Net server
side. In contrast to the satellite imagery, the radar imagery is encoded in a standard lossless image compression format
(GIF) and the small-sized model graphics are represented in a standard vector graphics format.
Accomplishments
Work continued toward meeting the FX-Net project's goal of providing access to the interactive functionality of the
WFO-Advanced workstation via the Internet. The architecture of the system was constrained mainly by bandwidth
limitations and the need to leverage existing workstation code. These issues were addressed by focusing on data
compression techniques along with multithreaded client-side processing and communication. Other important tasks
include mirroring of the user interface, making the client Java code portable, and modifying the existing WFO-Advanced
software.
Software Development – A milestone was met with the completion of an operational version 2.0 of FX-Net
released in September 2000. Risk reduction studies and the design of a robust operational server and client version
were part of this development. Some of the many new AWIPS products and features included in FX-Net 2.0 are:
- Color Editor – This tool emulates the AWIPS color editor dialog and allows users to modify the color of
an image represented by a given product. For example, users can choose to "black out" portions of an image represented
by a specific range of color spectrum and highlight a more interesting image related to a current weather event.
- Product Families – FX-Net now emulates the AWIPS product family capability, enabling users to request,
via the "Families" section of the Volume menu, entire product sets based on a given model family.
- Procedures – FX-Net will emulate the AWIPS procedures capability, allowing users to save and load
custom product sets.
- Auto Update – AWIPS always automatically updates products loaded in the primary viewport. FX-Net,
however, cannot purely emulate this capability, but will allow users to activate a similar feature. Once enabled
(via a toggle button on the control panel or from the View menu), the system will initiate a background thread that
continually checks for new data based on the product set loaded in the primary viewport, and will refresh that product
set by appending the new frames or overlays to it when it becomes available.
- Hot Keys – Certain "short-cut" keyboard commands will be available that closely emulate those found in
the AWIPS product.
- Product Status – A text window is now accessible that contains the history of those messages displayed
on the status bar. This window is accessed by pressing the button next to the status indicator on the status bar.
- Splash Screen – The FX-Net splash screen was added.
- Product Load Indicator – This application now contains a small animated globe in the upper-left comer
of the control bar which spins when a product is in the process of loading.
- Logging – A more detailed logging facility now exists in the products which will help users and developers
track down problems. This feature allows for six distinct operational logs (based on the software package), and includes
configurable properties to control, by severity level, which messages get sent to any given log.
FX-Net Applications – The operational use of FX-Net began two years ago at Plymouth State College (PSC) in
New Hampshire. PSC installed a special meteorological computer laboratory, which included 20 FX-Net workstations and
a data projector. FX-Net is used extensively in an undergraduate class that focuses on high-tech forecasting tools.
Close cooperation between FSL's FX-Net team and PSC resulted in important improvements on the FX-Net client side.
The FX-Net upgrade (version 2.0) was received very well by PSC.
Last year, FX-Net was selected as the official forecaster workstation to support weather forecasting at the 2002
Winter Olympics outdoor venues in Salt Lake City. In this scenario, the FX-Net server is located at the NWS Western
Region headquarters in Salt Lake City. Forecasting offices at each of the five Winter Olympics outdoor venues will
be equipped with FX-Net PC clients. With strong support from the NWS staff, an FX-Net server was installed at the
Western Region headquarters. High-speed communication lines were installed to connect the server with the outdoor
venues' forecasting offices. FX-Net was successful in its debut at several pre-Olympics trials held in November 2000.
Projections
The next major software development goal for the FX-Net project is the adaptation to the latest AWIPS software,
builds 5.0 and 5.1. Future client-side enhancements will include access to case study data and support of national
composite high-resolution radar datasets.
The International Division will work with the Salt Lake City Organizing Committee to provide an FX-Net workstation
system for onsite forecasting at each outdoor site of the 2002 Winter Olympic Games. The system will be used for
forecasting during the pre-Olympics events from December 2000 to March 200 1, and for the Weather Support Systems
tests planned for January 2001. FX-Net will be installed at a number of competition venues for use by an onsite
forecaster. A test server at the NWS Western Region headquarters will be further evaluated to determine FX-Net
requirements. Modifications and enhancements will be made to the system according to forecaster feedback regarding
these requirements.
FX-Net will also continue to support the PSC undergraduate meteorology laboratory. The University of Northem Iowa
(UNI) will be one of the new customers using FX-Net by summer 2001. The UNI application will be comparable to the
one at PSC, but will also include special summer weather forecasting workshops for teachers, emergency response
personnel, and the general public.
In support of the many other possible applications of FX-Net, development of more advanced features will continue,
such as the ability of a single FX-Net server to provide data for many AWIPS localizations, and optimization of data
transmission in batch mode for fire weather use. Whenever FX-Net servers are deployed outside FSL, enhancements
will be made to allow easy operational support of these systems.
With the advent of wireless Internet connectivity and powerful laptop computers, FX-Net can be accessed even in remote
areas. This makes FX-Net especially interesting for forecast regions with numerous remotely located forecasting offices,
for example, the Pacific and Alaska regions. Cooperation with both of these NWS regional offices should begin in 2001,
as well as serious planning efforts for the use of FX-Net as a fire weather forecasting tool in the field.
Additional information on the above FX-Net activities is available on the International Division homepage,
http://www-id.fsl.noaa.gov.
The Wavelet Compression Initiative
Renate Brümmer, Project Manager
Objectives
After successfully applying the wavelet data compression technique to satellite imagery, the Wavelet Data Compression
initiative was established to further investigate the possibility of using the technology for other meteorological
datasets.
Accomplishments
To meet the requirements from other projects to transmit model products via low bandwidth communication channels, effort
was focused on the application of wavelet compression to gridded model data. Compared to imagery datasets, model data
usually have higher numbers of dimensions, but each dimension has a smaller size. Therefore, special treatments are
needed to exploit the correlation among all dimensions.
A multidimensional data arrangement and transform scheme has been developed to accommodate the special features of the
model dataset. An experimental encoder and decoder package has been implemented to test various datasets with different
standard waves and different post transform compression algorithms. The preliminary test of the model grids with
prototype software has shown promising results.
Researchers completed investigations of different methods to control and minimize errors that were made in various
error metrics.
Projections
Studies will continue on one of the critical portions of the data compression package, quantification. A zero-tree
like algorithm will be designed and implemented to further improve the data compression performances, including
compression ratio, error distribution, and encoding time. Work will continue on the preprocessing of datasets.
These preprocessing steps, along with the new quantification method, will help to take advantage of the correlation
between data points, thus improving the performance of the compression algorithms.
The WorldWide Weather Workstation Initiative
Wayne Fischer, Project Manager
Objectives
The WorldWide Weather Workstation (W4) initiative is one of FSL's latest developments
utilizing WFO-Advanced workstation technology. The primary objective is to design a system that will meet the
forecasting needs of developing nations worldwide. Developing countries, of course, experience the same types
of natural disasters as occur elsewhere, but they lack the infrastructure available in modernized forecast offices,
such as data communications, local capability for operating numerical forecast models, and access to observations
and data. The advanced workstation technology and varied datasets offered by the W4 system
will help forecasters
and emergency response managers in these countries better deal with hazardous weather events, ultimately saving
lives. The W4 system would first be made available to forecasters in various countries
of Central and South America.
In studying the forecast situations and needs of developing countries, the design of the W4
system was based on the following considerations:
- Low capital and operating costs
- Ease of installation and maintenance
- Portability (installation in remote areas for short-term operation)
- Flexibility (modification of forecast products and addition of local data products at the forecast site).
Both the server and user subsystems will utilize the PC Linux platform, substantially decreasing system capital costs
while providing high-level system performance. Since the Linux platform can receive a highly compressed stream of
satellite broadcast data over modest bandwidths, the W4 system will deliver forecast products using a broadcast from
a geostationary satellite. Very Small Aperture Terminal (VSAT) technology will provide the broadcast of weather data.
FSL expertise in data compression techniques, particularly wavelet transform compression, will allow the transmission
of a full set of forecast data over a 128-kbps satellite link. These techniques will be applied to satellite imagery,
observational data, and gridded numerical weather prediction model datasets. Significant improvement in efficiency of
transmission will decrease the system's operating costs. W4 users will have access to a wide range of data, including
full resolution geostationary satellite data from all channels, full resolution numerical weather prediction (NWP)
data from various forecast models (initially, from the Medium-Range Forecast (MRF) and the Aviation (AVN) models), all
internationally available observation data (such as surface observations, ship and buoy reports, upper-air soundings,
and aircraft reports), and local observations (available at individual forecast offices).
The W4 system will use a series of scales, as is done in AWIPS, that will allow a display
compatible with the size and type of weather feature that is being studied. Important workstation features will be
available, such as roam and zoom, predefined procedures (sets of forecast products designed for a specific type of
storm), family graphics, and toggle capabilities for displayed products.
Accomplishments
Development of the W4 system began during Fiscal Year 2000, with progress accomplished in
the areas of system configuration, datasets and forecast products, and data compression.
System Configuration – The W4 system has four major components: the
W4 server, the satellite broadcast system, the user satellite ground station, and
the forecaster workstation. A significant investment is required in establishing the W4
capability and implementing the server and satellite broadcast system to disseminate the forecast information.
The system is economically viable with hundreds of users. Design studies indicate that the commonly used VSAT
technology provides the required quality and reliable performance at comparatively low cost.
The target workstation platform is a PC with Dual 1 GHz Pentium III processors with 512 MB of memory, a Linux
operating system, and a 2l-inch color monitor. Such a PC configuration is readily available anywhere in the
world at the current moderate cost of under $5,000 from several manufacturers. The Linux version of the AWIPS
workstation allows the extension of weather products to other geographic areas. FSL performed a test to determine
potential problems when forecasting in Bogota, Colombia. Forecast products generated by the local workstation itself
were displayed without major problems, but many local considerations are necessary to create a fully functional system.
Datasets and Forecast Products – A fundamental system decision to be addressed is whether a remotely located
forecast workstation is to be provided with full datasets or with specific forecast products. With a predefined set of
forecast products, it is only necessary to send a small volume of data to simplify the forecaster workstation. Providing
full sets of data, both observation and numerical model data, offers the potential for far greater flexibility in the
use of the forecast workstation. An analysis of costs associated with satellite broadcast showed that 128-kbps is a
reasonable cost/performance capability. Using this data rate, the full set of products required for synoptic and
meso-alpha scale forecasting can be efficiently broadcast to forecast offices.
Data Compression – The International Division's earlier work with the FX-Net system produced important
advances in data compression of NWP data files, particularly in satellite imagery. These techniques were applied
to NWP model output with good results; this convinced the development team that full NWP model datasets could be
transmitted over 128-kbps channels. However, the size of operational forecast model runs increases at a faster rate
than communication costs decrease along with improved capacities. The distribution of model output is a growing
problem of broader importance to the meteorology community, beyond the operation of W4 and
the development team will continue to investigate a solution.
Projections
During Fiscal Year 2001, emphasis will be on developing a fully operational beta-test version of the
W4 system, conducting an operational test and evaluation of the beta-test version,
and improving the data compression technology.
The International Division will develop partnerships with countries in Central and South America, and will seek a
weather service organization that will work with FSL in defining forecast products and procedures that fit the
forecast needs of the Latin American countries. This partner organization would serve as the initial beta test site
and as an eventual operational site for the W4 system. Collaboration will entail an
exchange of visits for orientation and training purposes. FSL meteorologists and system developers are skilled in
planning, designing, testing, and evaluating complex meteorological computer/communication systems such as
W4, but the installation, operation, and maintenance of such a system deployed at various
locations would need to be handled by a private sector corporate partner. Therefore, a corporate partner will be
sought whose skills and experience complement FSL's.
