Office of the Director
Office of Administration
Acronyms and Terms
Contact the Editor
Will von Dauster
Best Viewed With
Carl S. Bullock, Chief
Web Homepage: http://www-md.fsl.noaa.gov
C. Vada Dreisbach, Senior Engineer, 303-497-7151
G. Joanne Edwards, Programmer, 303-497-6903
M. Thomas Filiaggi, Meteorologist, 303-497-6578
Kevin Fuell, Guest Worker, 303-497-4382
Tracy Hansen, Programmer, 303-497-6569
David Howard, Quality Assurance Specialist, 303-497-6577
Dr. Lynn E. Johnson, Hydrologist/Research Scientist, 303-497-6984
Youngsun Jung, Guest Worker, 303-497-5265
Thomas B. Kent, Programmer, 303-497-7004
Nam-Ouk Kim, Guest Worker, 303-497-5265
Patrice Kucera, Meteorologist, 303-497-6618
Thomas J. LeFebvre, Meteorologist, 303-497-6582
Mark A. Mathewson, Meteorologist/Chief, Enhanced Forecast Tools Branch, 303-497-6713
Deborah Miller, Systems Analyst III, 303-497-6770
Brian Motta, Guest Worker, 303-497-6561
Scott D. O'Donnell, Programmer/Hydrologist, 303-497-6562
Dale R. Perry, Meteorologist/Programmer, 303-497-6050
Robert F. Prentice, Systems Engineer/Leader, Advanced Development Group, 303-497-6771
William F. Roberts, Meteorologist, 303-497-6104
Michael Romberg, Programmer, 303-497-6544
MarySue Schultz, Systems Analyst, 303-497-6499
Michael Vrencur, Systems Administrator, 303-497-7526
Carol A. Werner, Secretary Office Automation, 303-497-6157
(The above roster, current when document is published, includes government,
cooperative agreement, and commercial affiliate staff.)
Address: NOAA Forecast Systems Laboratory Mail Code: FS6
David Skaggs Research Center
Boulder, Colorado 80305-3328
The Modernization Division produces functional designs or working prototypes of techniques, workstations, and systems that may be implemented into
National Weather Service (NWS), or other agency, operations up to a decade later. The process includes selecting, tailoring, and implementing advanced
techniques and devices produced by the research and development community, industry, or elsewhere. Developments are state of the art and continually
evolve commensurate with new technological advances, such as D2D (Display Two Dimensional) shown in Figure 61.
Figure 61. D2D screen showing satellite composite infrared image (large panel) over the Northern Hemisphere.
The modernization of NWS operations involves the development of a new radar system, an automated surface observing system, a new series of
geostationary satellites and products, and a communications and forecaster workstation system, the Advanced Weather Interactive Processing System
(AWIPS). FSL has been participating for years in risk reduction activities to help the NWS meet its goals in the development and deployment of AWIPS.
The Modernization Division comprises three branches:
Risk Reduction Branch AWIPS support and evaluation
Enhanced Forecaster Tools Branch AWIPS Forecast Preparation System
Advanced Development Branch AWIPS system evaluation and enhancement.
Risk Reduction Branch
Carl S. Bullock, Chief
Work in the Risk Reduction Branch is directed toward helping the NWS prepare for operations in the AWIPS era. The two focus areas include operation
and evaluation of risk reduction activities and the development of AWIPS. Since NWS announced in 1996 that the FSL-developed WFO-Advanced system
would form the core of the AWIPS software to run at all Weather Forecast Offices (WFOs) and River Forecast Centers (RFCs), the development of AWIPS
has been the primary activity. After AWIPS is completely deployed, the branch will return to its traditional tasks involving risk reduction.
AWIPS Software Builds
The branch provided significant portions of several AWIPS builds, including 5.1.1, 5.1.2, and 5.2.1. System performance improvements continue to be an
important area of development effort as AWIPS is used to perform time- critical operations such as the issuance of severe weather warnings. Some display
problems with a performance enhancement in 5.1.1 were carefully researched and analyzed and corrected. The performance of the function that notifies
the rest of AWIPS as new data arrive was improved. As the volume of data continues to increase with each new build, it becomes more vital that key
components operate efficiently. This key infrastructure improvement allows quicker updates of critical radar data, for example.
Build 5.1.1 gives users greater control over time matching in that they can now specify the time of the last frame and the interval between frames. Also, a
time offset can be used when multiple datasets are overlaid on the display, allowing data with different time frequencies to be more easily displayed together,
such as radar and satellite data.
A new approach was implemented in build 5.1.2 to deal with isentropic data on AWIPS. The isentropic coordinate system provides another way of viewing
meteorological data, complementing the constant pressure surface perspective. The software which calculates isentropic depictions was optimized so that
these computations can now be done on the fly, offering greater flexibility for the forecaster and eliminating a background processing load. A new function
was added that correlates temperature readings from infrared satellite images to soundings. A miniature sounding window pops up and shows the point on
the temperature profile that matches the reading from the satellite image. The user can specify observed soundings or forecast soundings from one of several
numerical models. A variety of other graphic improvements was made such as plots of buoy data that include wave height, steepness, swell, and user control
of the appearance of key radar graphics depicting hail and tornados. A limited monitoring capability was also added, and for certain text messages, forecasters
will be alerted to the possibility of hazardous weather based on its proximity to their area of responsibility.
An important new data acquisition capability was added in build 5.2.1. One of the standard formats used to transmit meteorological point data is Binary
Universal Form for Data Representation (BUFR). The branch developed a generic BUFR decoder for AWIPS that can be easily adapted to acquire a variety
of datasets. Initially this decoder will process forecast soundings from the Eta numerical weather prediction model. These soundings are created from the native
model output and have much higher vertical and temporal resolution than the current soundings extracted from AWIPS grids. The increased temporal resolution
of the soundings (every hour versus every 3 hours) necessitated the addition of a new display option that allows the user to exercise greater control over the time
interval between frames of a loop. Normally AWIPS will load at the highest temporal resolution available; however, there is a limit of 32 frames. Since the new
model forecast soundings from the Eta have hourly time steps out to 60 hours, all of the data cannot be loaded at once. The new time options allow the user to
view any part of that 60-hour period or to use a longer time interval, such as every other hour, to view the entire period at a lower time resolution.
The AWIPS interface to the WSR-88D begins to change in build 5.2.1. New radar products, including 8-bit (255 level) reflectivity and velocity products, will be
sent to AWIPS . The display code was modified to handle these new products. This should enable forecasters to see more detail in storms.
FSL and the National Weather Service decided that the Linux operating system would provide the most cost-effective path for evolving AWIPS, and FSL
ported its AWIPS workstation code to Linux. After conducting a demonstration of the Linux workstation at 14 NWS offices, with very positive results, the
NWS plans to install Linux workstations at all AWIPS sites during the next year.
Preparations were made to replace AWIPS hardware with Linux systems. First, the existing load on the AWIPS data server was reduced by transferring part
of the load on a Linux system. With the anticipated increase in numerical weather prediction model data, it is prudent to move that processing to a more
powerful (and less expensive) Linux system. In addition, the branch corrected a problem with the AWIPS network file serving perfomed by the data server
computer. By running the grid and satellite data ingests on a Linux system and storing the data there, the file serving load will be more distributed.
The migration to Linux has opened up new collaborations, including a group of developers from the Korean Meteorological Agency (KMA) who adapted the
WFO-Advanced workstation for use in Korea. Geographic scales appropriate for Korea were defined, and software modules were developed to transform
KMA data into the format needed by the workstation. For the severe weather season, radar data, satellite data, hurricane data, mesonet data, synoptic data,
upper air observations, and model data were all available on the workstation. FSL provided training to all KMA forecasters on the WFO-Advanced workstation
in advance of their summer severe weather season.
Online Evaluation Tool
User input and feedback have been an integral part of the AWIPS software development and training. Historically, collection of this information has been
a tedious and labor-intensive process. Over the past several years, the Modernization Division's E-Team has developed a password-protected Web-based
survey evaluation tool that provides an easily accessible and systematic method to gather user input for software tools, training, system performance, etc.
With this online survey tool, the E-Team has provided detailed and timely feedback on projects such as the AWIPS Operational Evaluation and Training,
the Interactive Forecasting Preparation and GFESuite/Rapid Prototype Processing programs, the Local Data Acquisition and Dissemination/Emergency
Management Dissemination System projects, and the NWS Linux demonstration.
The end-to-end survey development process starts with discussions with developers and management to determine appropriate questions that address the
overall goals of the project, software, or training, along with questions on specific elements important for future system development. The survey software
can support several types of questions, whether they are "yes/no," rating scales, or open-ended commentary questions. Using HTML, the survey is drafted
and the various widgets are implemented in the survey text file. The survey scripts (perl and cgi) are easily configurable to each project, as is the user
access/password setup. The respondents' completed surveys, whether just a few or many dozens, are saved as individual files and easily imported into
spreadsheets, and the text files are ported into a text editor. The E-Team then carefully reviews all of the feedback, and then stratifies and compiles a
catalog of the survey results, either as a hard copy or online, including (if desired) graphics of the numerical results. The catalogs do not include the
individual names, so that every respondent's anonymity is protected. The catalogs are then delivered to the development teams and management and
results are reviewed with the E-Team. Participation in this streamlined process has been exceptional. Nearly all (>90%) of the participating organizations
and individuals routinely and comprehensively fill out these surveys, generating a reliable and very useful database which helps direct future system development.
The final AWIPS software build of the "5" series, build 5.2.2, will be completed and installed during 2002. This will complete the additional capabilities
needed to make AWIPS fully functional. Soundings and wind reports from satellite and aircraft will be added. Model ensemble data will be introduced.
Build 5.2.2 will start addressing special needs (such as surface synoptic reports) of offices outside the continental United States. Although portions of the
Interactive Forecast Preparation System have been available in previous builds, a significant milestone will be met when the complete system becomes
available in 2002. This will change the way forecast information is generated by the National Weather Service, allowing forecasters to prepare a set of
grids that capture forecast weather conditions. A wide variety of products will be generated from these grids. In addition, FSL will prepare software that
aids in the coordination of grids between neighboring offices to resolve inconsistencies. The grids will also be collected on a national basis to provide a
nationwide view of forecast conditions, something never before possible. (Figure 62 shows a model-derived (Eta) cross section of a 130-knot jet core over
the Northern Rockies.)
Figure 62. D2D model-derived cross section of a 130-knot jet core over the Northern Rockies.
The next step in the Linux evolution will be a demonstration of a preprocessor system, which handles the processing of satellite and gridded data to free up
resources on the data server. A data broadcast function will be developed that will enable storage of time-critical datasets, such as satellite and radar, locally
on the workstation. This should greatly speed up loading of these datasets, which are vital for diagnosing severe weather. Plans are to investigate a complete
AWIPS system based on Linux to identify remaining issues with the Linux evolution.
Return to Top of Modernization Division Section
Enhanced Forecaster Tools Branch
Mark A. Mathewson, Chief
The focus of the Enhanced Forecaster Tools Branch is the Interactive Forecast Preparation System (IFPS). In consultation with a working group of NWS
weather forecasters, staff are designing and building the graphical forecast support system for AWIPS. A basic NWS concept driving the design of the IFPS
is that forecast duties in the WFO will no longer be divided along service boundaries. Instead, a nowcaster/forecaster division of duties will be in place, with
each forecaster responsible for all forecast elements for a specific block of time. Rather than spending time typing information from forecast products,
forecasters will maintain a graphical database of forecast weather elements. A goal of the IFPS designers is to minimize forecast preparation time and
to maximize forecasters' ability to interact with the data, thus allowing more time to focus on the art and science of forecasting. Starting with the current
forecast database and/or gridded fields initialized from numerical models or central guidance, and using a set of highly interactive graphical tools, forecasters
will visualize and modify surface sensible weather elements. Product generation utilities will then format the digital forecast data into a wide variety of
graphical, gridded, and text-based products with little additional effort from forecasters.
Interactive Forecast Preparation System
The branch concentrated on four development tasks of the advanced IFPS: improving the Graphical Forecast Editor (GFE), developing GFE products,
developing other CFESuite capabilities, and supporting the Rapid Prototype Project (RPP).
Improving the Graphical Forecast Editor The GFE provides tools for the forecaster to view and edit grid fields that capture the essential
information needed to generate a variety of forecast products. Productive interactions between field forecasters and developers prompted numerous
enhancements of the GFE graphical user interface, GFE scripting capability, and other capabilities. Smart tools continue to be enhanced, and forecasters
are using the smart tool technique to adjust algorithms and write their own editing tools. Three versions of the GFE were delivered to the NWS for
inclusion in AWIPS. A screen from the GFE is shown in Figure 63.
Figure 63. A screen showing the Graphical Forecast Editor, part of the AWIPS Interactive Forecast Preparation System.
Developing GFE Products The GFESuite system comprises the GFE and other programs to generate products. Product generators have been
written to provide netCDF output and PNG graphical imagery. The output capability is highly tailorable through the use of Python configuration files.
The infrastructure has been developed to allow field sites to create tabular and narrative format text products. All of the products are derived directly
from the gridded database.
Developing other GFESuite Capabilities The GFESuite of software is maturing with the introduction of additional capabilities. Intersite
coordination of grids was developed to permit the exchange of forecast grids among sites during the creation of the forecast. Forecasters can now determine
if their forecast deviates from their neighbor's in a significant way. Another capability developed recently is smart initialization, which is a very flexible,
field tailorable system that is used to derive first-guess foercast grids directly from the models.
Supporting the Rapid Prototype Project The RPP was created to accelerate the GFE development process and to deliver a workable version to
NWS field offices in the shortest amount of time. The 16 field RPP sites work closely with FSL developers to enhance the IFP software so that it better facilities
the operational forecast process.
Many suggestions for improvements in the GFE comes from the RPP sites, and are implemented and tested a year before the same improvements appear
in the AWIPS IFPS. Seven software releases, consisting of over 300 enhancements and 100 bug fixes, were issued to the RPP sites during the last year.
FSL hosted GFE workshops which provided advanced training for smart tools and the GFE, as well as a Web forum for feedback to improve the GFE and
A major task during 2002 is to continue support of the RPP process at NWS forecast offices by meeting with forecasters, listening to their comments and
improvement suggestions, and integrating feedback into the IFPS development process. Tasks include further development of intersite coordination, improved
performance of smart tools and other field-tailorable algorithms, and introduction of a point- and grid-based daily forecast critique. Support will continue for
the IFPS in AWIPS by responding to user feedback and assisting the field in the implementation of AWIPS builds 5.21 and 5.2.2. Improvement of GFE will
continue toward future goals related to initial operating capabilities for the IFPS.
Return to Top of Modernization Division Section
Advanced Development Branch
Carl S. Bullock, Acting Chief
The Advanced Development Branch endeavors to keep the National Weather Service abreast of advanced technology, particularly with regard to how it
can be effectively utilized in forecasting systems related to the operational AWIPS.
An attempt was made to broaden the scope of the FX-Advanced systems by working with the Environmental Protection Agency (EPA), specifically to adapt
our Web dissemination system to an application within the EMPACT (Environmental Monitoring for Public Access and Community Tracking) program.
EMPACT disseminates environmental information to communities via the Internet. Although there was an interest in the dissemination system, funding
was not available. The experience, however, increased our awareness of a growing need to incorporate more GIS (geographical information system)
capabilities into the FX-Advanced systems. It was also useful in adapting FSL systems for agencies other than the National Weather Service. The evolving
OpenGIS standard is inviting and will be followed as staff pursue methods to interface FSL systems with external GIS systems.
Significant effort has been expended to validate certain configuration management techniques applicable to AWIPS. These techniques mostly involved ways
in which the software build process can be modified to be more efficient and reliable, and freed from constraints imposed by use of a proprietary build product
(PCMSmake) that is currently used by the AWIPS contractor. AWIPS program management decided to defer application of these techniques until after the
build 5 series of releases has been completed.
Our current successful strategy for fostering collaboration and improving productivity has been to surround the NWS change management system with our
own localdev environment, a Perl-based collection of development scripts. A major upgrade to localdev was released last year that fully supports the migration
of software between projects with automated merge support. This enables project managers to periodically upgrade their software with software from other
projects in an efficient and reliable fashion. This capability has been especially useful in the Range Standardization and Automation (RSA) project for NASA,
a project for the Korean Meteorological Administration (KMA), and the FX-Net project.
During 2002, the branch will investigate changing the software development process used in developing AWIPS code. This process has been based on
providing major software releases twice a year. There is interest in changing to a system that releases software more frequently with a focus on smaller
segments of the overall system. This may require changing the software tools that are used for configuration management. FSL will make every effort
to move toward an efficient, cost-effective system that offers levels of reliability, automation, and ease-of-use that are much higher than any experienced
in the AWIPS program so far.
A primary goal is to start investigation into a next-generation forecasting system. The concentration on fielding the existing AWIPS system and porting it
to Linux has consumed all resources for the last several years, thus curtailing any investigation into a longer term direction.
Return to Top of Modernization Division Section