FSL in Review 2002 - 2003

Cover/Title Page

Organizational Chart

Office of the Director

Office of Administration
and Research

Information and
Technology Services

Forecast Research

Demonstration Division

Systems Development

Aviation Division

Modernization Division

International Division


Acronyms and Terms

Figures Listing

Contact the Editor
Nita Fullerton

Web Design:
Will von Dauster
John Osborn

Best Viewed With
Internet Explorer

FIR 2002 - 2003 DD MastHead
Margot H. Ackley, Chief
(Supervisory Physical Scientist)

Web Homepage: http://www.profiler.noaa.gov

Norman L. Abshire, Electrical Engineer III, 303-497-6179
Leon A. Benjamin, Programmer Analyst III, 303-497-6031
Michael M. Bowden, Engineering Technician II, 303-497-3260
James L. Budler, Engineering Technician II, 303-497-7258
James D. Bussard, Information Systems Specialist II, 303-497-6581
Michael A. Carrithers, Electronics Technician II, 303-497-4376
Robert L. Cinea, Data Technician III, 303-497-6200
Michael C. Foy, Programmer Analyst II, 303-497-4618
David J. Glaze, Electrical Engineer III, 303-497-6801
Daphne M. Grant, Meteorological Technician, 303-497-5627
Seth I. Gutman, Physical Scientist, 303-497-7031
Kirk L. Holub, Systems Analyst III, 303-497-6642
Bobby R. Kelley, Information Technology Specialist, 303-497-5635
Brian A. Koonsvitsky, Logistics Specialist, 303-497-3095
Michael J. Pando, Information Systems Specialist II, 303-497-6220
Brian R. Phillips, Field Electronics Technician, 303-497-6990
Alan E. Pihlak, Information Technology Specialist, 303-497-6022
Robert F. Prentice, Programmer Analyst, 303-497-6771
Susan R. Sahm, Information Technology Specialist, 303-497-6795
Michael K. Shanahan, Electrical Engineer, 303-497-6547
Jebb Q. Stewart, Programmer Analyst, 303-497-6832
Scott W. Stierle, Systems Analyst III, 303-497-6334
Richard G. Strauch, Senior Electrical Engineer, 303-497-6385
Douglas W. van de Kamp, Meteorologist, 303-497-6309
David W. Wheeler, Electronic Technician II, 303-497-6553

(The above roster, current when document is published, includes government, cooperative agreement, and commercial affiliate staff.)

Address: NOAA Forecast Systems Laboratory – Mail Code: FS3
David Skaggs Research Center
325 Broadway
Boulder, Colorado 80305-3328


The Demonstration Division evaluates promising new atmospheric observing technologies, such as the NOAA Profiler Network (NPN), developed by the NOAA Research Laboratories and other organizations and determines their value in the operational domain. Activities range from demonstrations of scientific and engineering innovations to the management of new systems and technologies. In support of NOAA's mission to serve society's need for weather and water information, the division uses new upper-air observing techniques to create and disseminate reliable assessments of weather, climate, space environment, and geodetic phenomena. The data and techniques developed and implemented by the division also support seasonal to interannual climate forecasts as well as the prediction and assessment of decadal to centennial climate change. Due to a largely unanticipated synergy between the requirements for atmospheric remote sensing and the more traditional applications of the Global Positioning System (GPS; i.e., positioning, navigation, and time transfer), the GPS-Met Observing Systems Branch within the Demonstration Division also promotes safe navigation by providing GPS and other observations to the National Geodetic Survey network of continuously operating reference stations (CORS), the U.S. Coast Guard (USCG), U.S. Department of Transportation (DOT), and other GPS users in the public and private sectors.

These activities are an investment in scientific research, the development of new technologies to improve current operations, and in NOAA's preparation for the future. The division has successfully demonstrated all major elements of three reliable, low-cost continuous upper-air observing systems — wind profilers, Radio Acoustic Sounding System (RASS) temperature profilers, and ground-based atmospheric water vapor sensing observing system (GPS-Met). These systems complement other operational and future ground- and space-based observing systems. New information network tools and techniques have been adapted to acquire and process Cooperative Agency Profilers (CAPs), GPS, and surface meteorological observations from NOAA and other public/private organizations and international partnerships. This capability allows rapid expansion of observing system coverage at extremely low cost. The division has been heavily involved in transferring environmental expertise/technologies to improve NOAA's ability to serve its customers and forge stronger ties with its partners, especially NOAA's National Weather Service (NWS), DOT, and Department of Defense (DOD).

During 12 years of operation, the NPN has been providing important upper-air data to a variety of customers. The NWS uses data from the NPN, CAP, and GPS networks routinely in computer-generated forecasts, and its field forecasters tailor model guidance to local conditions. The data are also available for interested users via the Global Telecommunication System and for the public via the Internet. Data are used by many other federal, state, and local organizations for support in weather forecasting, aviation, and monitoring climate and air quality. One of the NPN customers, the Lawrence Livermore National Laboratory, uses NPN data as critical input to its dispersion model. The model supports work under contract to DOD and the Department of Energy (DOE) for Homeland Security. At the request of NASA and the National Transportation Safety Board (NTSB), all NPN, CAP, and GPS-Met data taken during the break up of the space shuttle Columbia were provided to them for important forensic analysis in determining the root cause. In particular, the Palestine, Texas, and Winnfield, Louisiana, profiler data captured the time and horizontal and vertical positions of some of the falling fragments.

Currently the division is engaged in the following major projects:

  • Operation, maintenance, and enhancement of the 35-station NOAA Profiler Network (NPN), which includes three systems in Alaska and the CAP sites (Figures 44 and 45).

Figure 44 - NPN and CAP Sites

Figure 44. Location of all 35 NOAA Profiler Network sites and Cooperative Agency Profiler (CAP) sites providing data via the Profiler Website.

Figure 45 - NPN Sites

Figure 45. All NOAA Profiler Network sites, including Alaska (upper left),
with radars and surface instruments. Circles show location of NPN sites
without RASS; stars show location of NPN sites with RASS.
  • Collection, correction, and distribution of wind and temperature data from the CAP sites.
  • Planning and support activities for a Fiscal Year 2006 initiative for a national upper-air mesoscale observing system which will include profilers and GPS-Met systems.
  • Development, deployment, and evaluation of an all weather integrated precipitable water vapor (IPW) observing system using radio signals from the satellite GPS.
  • Evaluation of three GOES high data rate communication systems for network deployment.
  • Assessment of alternative network data communication technologies.
  • Upgrade of the surface meteorological sensor package at NPN sites to include winds and precipitation measurement capability.
The division comprises five branches organizationally; however, the branches work in a fully integrated team mode in supporting the overall objectives of the division.

Network Operations Branch – Monitors systems' health and data quality, and coordinates all field repair and maintenance activities.

Engineering and Field Support Branch – Provides high-level field repair, coordinates all network logistical support, designs and deploys engineering system upgrades, and redeploys GPS or profiler systems as needed.

Software Development and Web Services Branch – Provides software support of existing systems, develops new software and database systems as needed, provides Web support of the division's extensive Web activities, and designs software to support a national deployment of profilers.

GPS-Met Observing Systems Branch – Supports development and deployment of the GPS-IPWV Demonstration Network, and provides software development and scientific support.

Facilities Management and Systems Administration Branch – Manages all computers, data communications, network, and computer facilities used by the staff and projects of the division.

Network Operations Branch
Douglas W. van de Kamp, Chief


The Network Operations Branch is responsible for all aspects of NOAA Profiler Network (NPN) operations and monitoring, including the coordination of logistics associated with operating a network of 35 radars and surface instruments. The original concept for an operational profiler network envisioned the Doppler radar profiler as part of an integrated upper-air remote sensing system capable of measuring winds, temperature, and humidity. The Demonstration Division's progress toward these goals include the addition of the Radio Acoustic Sounding Systems (RASS) for temperature profiling in the lower troposphere at 11 NPN sites, and GPS integrated precipitable water vapor (GPS-IPWV) systems for moisture measurements at all NPN sites. In addition to the 35 NPN sites, another 200+ NOAA and other-agency sites are monitored for timely GPS positions and surface observations to produce real-time IPWV measurements. Additional wind and RASS data have been acquired from a growing number of independently operated profiler sites, now totalling about 80. These Cooperative Agency Profilers (CAPs) include many lower tropospheric boundary layer profiler sites plus a few higher power profilers similar to NPN sites. The data from these CAP sites are now available to the meteorological community in real time via the division's Webpage. Along with the four other branches within the division, this branch maintains and improves the NPN and CAP real-time data availability to the National Weather Service (NWS) and other worldwide users. The Network Operations Branch directly supports NOAA's mission of improving weather products and services by providing real-time comprehensive, high quality upper-air and surface observations to NWS forecasters and numerical weather prediction models.


The availability of hourly NPN winds to the NWS remained high through 2002, averaging about 95%. A summary of the overall performance of the network for the past 12 years is presented in Figure 46. This and other tracking mechanisms are used to assess the strengths and weaknesses of the NPN. During the past year, the NPN data never fell below 90%, compared to all previous years. It is interesting to note the general pattern of decreased availability of hourly winds each year during the spring and summer months, compared to slightly higher availability during the fall and winter months. This pattern has been analyzed and is attributed to increased lightning activity and severe weather during the convective season (cause of more commercial power failures and lightning induced profiler site hardware damage) and air conditioner failures during the summer. From this trend analysis, additional lightning suppression and communications equipment protection were added to the profiler sites.

Figure 46 - NPN Data Availability

Figure 46. NOAA Profiler Network 404-MHz profiler
data availability from January 1991 – January 2003.

The mean time between failure (MTBF) for individual NPN sites and the total number of failures resulting in a data loss of 24 hours or greater are presented in Figure 47. The more reliable sites are listed to the right, and the less reliable sites to the left. These MTBF statistics include any communication and commercial power outages (with duration of 24 or more hours) and any profiler site hardware failures. The profiler site hardware was designed for an MTBF of 6 months, and at least 6 NPN sites are exceeding this. Typical NWS commissioned systems such as NEXRAD, ASOS, and radiosondes have a data availability of 97% or better. Figure 47 shows that the NPN is comparable in data availability, but is currently less than the target of 97%. Also plotted here is the maximum time between failure for each NPN site. It is interesting to note that many sites have operated without a failure for 1 – 2 years, and 5 sites without a failure for over 2.5 years during the total 7-year period of this study.

Figure 47 - NPN Failure Analysis

Figure 47. Seven-year analysis (January 1996 – January 2003) of
mean time between failure (MTBF) and the total number of failures,
based on data outages for more than 24 hours.

A significant portion of personnel time involves the day-to-day operations and monitoring tasks related to the hardware, communications, and meteorological data quality aspects of the NPN. Constantly attending these tasks has resulted in high data availability rates for the past few years. Other tasks include initial diagnosis of equipment failures, coordination of all field repairs and maintenance activities, and maintenance of logs of all significant faults that cause an outage of profiler data. Figure 48 shows the total number of hours of profiler data lost by fault type (such as component failures, scheduled downtime for maintenance, and power and air conditioner failures) for the past seven fiscal years. The duration of each data outage is broken down into many different states, including how long it took to identify a failure, diagnose and evaluate the problem, wait for repair parts to be sent and received, restore commercial power or communications, and document when and how the fault was ultimately repaired. Figure 49 shows the distribution of these categories of downtime (normalized over the past 6 years). Analysis of all these states reveals important information regarding the operation of the network. In addition to the data monitoring tasks, there are the financial aspects related to the continued operation of the NPN, including tracking land leases, communications, and local commercial power and phone bills for all the profiler sites.

Figure 48 - NPN Data Loss By Fault Type

Figure 48. Hourly NOAA Profiler Network data lost by
fault type for the past four fiscal years, from 1999 – 2002.

Figure 49 - NPN Downtime By Categories

Figure 49. Distribution of NOAA Profiler Network downtime
by categories (normalized over six years) from 1996 – 2002.

Personnel in the Profiler Control Center (PCC) routinely monitor the NPN, currently noncommissioned by the NWS, only during normal working hours, 7:30 AM – 4:30 PM local time (27% of the total hours in a week). The remainder of the time, the profilers, dedicated communication lines, and Profiler Hub computer system operate unattended. The division has made significant improvements in its ability to remotely monitor activity within the NPN, Hub processing, and data communications via displays available on the Web and other tools. Activities that are now routinely monitored on the Web include information on profiler real-time status, data flow to the NWS Telecommunications Gateway (NWSTG), and ingest of profiler data into the Rapid Update Cycle (RUC) model at the NWS National Centers for Environmental Prediction (NCEP). Using these tools to remotely diagnose problems as they arise outside normal work hours has increased the availability of NPN data.

Examination of several years of data showed that a significant number of lost hours of data were attributed to the local main power breaker (200 amps) being tripped to the off position, usually caused by lightning related power surges. Simply resetting the breaker would restore operation, but still required a site visit, typically by an NWS technician of the local landowner. From this analysis, the Engineering and Field Support Branch designed and installed a device to remotely reset the main breaker via a phone call to the site. The Network Operations Branch routinely uses this method to restore profiler operations, as well as "power cycling" a site in an attempt to clear software "hangs" and other problems. Last year the breaker reset capability was attempted 226 times outside of normal work hours to restore operations. It was successful 180 times (80%), resulting in an additional 4,700+ hours of profiler and GPS-IPWV data availability to our customers. These resets performed outside normal work hours alone increased data availability by 1.5%. This is quite impressive when our data availability is already routinely >90%.

The Network Operations Branch provided support for all aspects of planning, installation, activation, and evaluation of two relocated NPN profilers. The original 404-MHz profiler located near Platteville, Colorado, was disassembled, transported, and reassembled at a site near Ledbetter, Texas, in cooperation with the NWS Southern Region and the Lower Colorado River Authority in Texas. The 449-MHz profiler originally located at the Vandenberg Air Force Base in California was disassembled, transported, and reassembled at the Platteville site to replace the original 404-MHz profiler. Compared with the 404-MHz system, the 449-MHz system is much more versatile. It is a hybrid system that uses transmitter and antenna hardware similar to that originally installed in the NPN, but with different receiver, data processing, and beam steering capabilities.

Eleven NPN sites have RASS capabilities, typically providing measurements up to 2.5 – 4 km above the ground. In general, the velocity of the lower tropospheric wind limits the maximum height coverage of RASS by advecting the acoustic signal outside the radar beam. Each RASS-equipped site has four acoustic sources that are located inside the antenna field fence near the corners of the wind profiler antenna. Ongoing experiments are being conducted at Platteville, Colorado, and Purcell, Oklahoma, to investigate the impact of acoustic sources placed 35–140 m upwind of the profiler sites. Typical improvements of 500 – 1000 m in the RASS height coverage are observed when the 70 – 140 m upwind acoustic sources are activated, and the low altitude winds are from that direction.

Low-power profilers that measure winds and temperature in the boundary layer to the lower troposphere (60 m to ~3 km above ground) have begun operating in greater numbers around the Northern Hemisphere in recent years. They primarily support air quality measurements and meteorological forecasting and research programs, and typically operate independently or in small groups. Approximately 80 CAP sites are currently operating and providing data to the Profiler Control Center in Boulder. The division is working in cooperation with other agencies to acquire CAP wind and temperature data that are processed into hourly and subhourly quality-controlled products, and are ultimately distributed along with products from the NPN. CAP data are primarily used for air quality monitoring and forecasting, but have applications to homeland security, and numerical weather prediction and subjective weather forecasting in support of NOAA's mission.

To gain a better understanding of how and when NPN data are used by the NWS field offices, this branch started monitoring their Area Forecast Discussions (AFDs). Each day NWS offices typically write two AFDs, which describe the current forecasting issues, both in the short-term and longer-term forecast period. These AFDs are generally technical in detail and more of a "thought process" to share among the forecasters, both within a forecast office between shifts and in adjacent NWS forecast offices. All AFDs are searched for the term "profiler." During a 45-day period (1 December 2002 – 14 January 2003), 51 NWS offices (out of a possible 114) mentioned the use of profiler data in at least one of their AFDs. Figure 50 displays the geographical location of these offices. Note that the spatial distribution is very similar to that of all the profiler sites shown in Figure 44. The NWS offices located in the central U.S. are of course primarily using NPN data, while those offices near the East and West Coasts are all using CAP data. Of the 51 offices indicating the use of profiler data, a total of 220 AFDs (~5 per day) mentioned the use of profiler data in their decisionmaking process.

Figure 50 - NWS WFOs Using NPN Data

Figure 50. Location of National Weather Service offices (in red) identifying
the use of profiler data in their Area Forecast Discussions (AFDs) during
a 45-day period, from 1 December 2002 – 14 January 2003.


The Bird Contamination Check algorithm will be examined for possible additional improvements. The original algorithm analyzed only the hourly averaged north and east beams to detect the broader spectral widths caused by migrating birds. Recently the spectral width from the vertical beam was incorporated into the algorithm. The next significant improvement is likely to involve more sophisticated processing of the 6-minute moment data. Since the current Profiler Hub cannot incorporate any more processing, additional QC development work is limited at this time.

The division will continue to operate and maintain the 11 RASS-equipped profiler sites. Experiments will continue at Platteville and Purcell to investigate the optimum acoustic source locations (distance upwind) and acoustic output power. Improvements are expected in the quality control of RASS data, primarily during periods of internal interference, and in the presentation (i.e., contouring specific temperatures) of RASS data on the Profiler Webpage.

The operations of the CAP Hub will continue and will be used to acquire additional tropospheric profiler data from targets of opportunity, provide quality control for these data, and distribute that data to users via the Web and the NWSTG. Additional automated monitoring procedures will be investigated to handle the increasing number of CAP sites available and monitored by the PCC.

The capabilities of the new 449-MHz profiler at Platteville will be investigated. This will include data quality and height coverage of two different data processing methods (standard consensus versus a multiple peak tracking algorithm), three beams compared to five beams in terms of data quality and cost/complexity issues, higher temporal resolution data, and reduced height of the first sample height. These are all issues related to the design and implementation of a national profiler network.

The capability to remotely reset the main breaker via a phone call to the site has proven so successful that the procedure will be automated. Data availability is manually checked each evening during weekdays, and (typically) mornings and evenings on weekends. Sites that are "hung" due to software failure or missing data for other reasons are reset at that time. The average time between a site shutting down and being reset is currently 5.6 hours. Plans are underway to automatically initiate a breaker reset after two hours of missing data.

Continued collaboration within the division to support the operation and maintenance of the NPN will help maintain consistently high data availability statistics. This ultimately supports NOAA's mission of improving weather products and services, resulting in reduced loss of life and property damage from weather related events.

Return to Top of Demonstration Division Section

Engineering and Field Support Branch
Michael K. Shanahan, Chief


The primary focus of the Engineering and Field Support Branch is to carry out the operation, maintenance, and improvement of the NOAA Profiler Network (NPN). Through collaboration with the FSL Profiler Control Center (PCC), the 35-site NPN network is monitored to assure data quality and reliability. Constant network upgrades, identification of network problems (using remote diagnostics analysis), and prompt corrective actions result in increased data availability.

Most of the preventive and remedial maintenance is performed by electronics technicians from the National Weather Service (NWS) in accordance with network maintenance agreements. The PCC uses the remote diagnostics capabilities to recognize failed components, order line replaceable units, and coordinate with the NWS electronics technicians regarding field repairs. More complex problems are handled by a team of specialized engineer/technicians, called rangers, who are experienced in the design and operation of the profiler systems. Based in Boulder, the rangers can be mobilized to the field on short notice to repair the profilers.


In collaboration with the the Lower Colorado River Authority (LCRA) and NWS, this branch installed a 404-MHz profiler &$151 complete with the Radio Acoustic Sounding System (RASS), GPS Surface Observing System (GSOS), and GPS instruments — at Ledbetter, Texas (Figure 51). This profiler, located at the Natural Science Laboratory (75 miles east of Austin at Cooper Farm), will support the NWS Southern Region and the growing Texas Mesonet.

Figure 51 - 404-MHz Profiler in Ledbetter, TX

Figure 51. The 404-MHz profiler at the Lower Colorado River
Authority's Natural Science Laboratory in Ledbetter, Texas.
(Photo courtesy of Brian Phillips, SRG, Inc.)

The 449-MHz profiler at Vandenberg Air Force Base was relocated to Platteville, Colorado, and will be used as an operational testbed for the conversion of the 404-MHz systems to 449-MHz. The Platteville profiler (Figure 52) is a hybrid system consisting of Vaisala and Lockheed Martin components.

Figure 52 - 449-MHz Platteville, CO Testbed

Figure 52. The 449-MHz profiler at Platteville, Colorado, planned
as a testbed site for a national 449-MHz profiler network.
(Photo courtesy of Brian Koonsvitsky, SRG Inc.)

The grounding systems at 30 profiler sites were tested and upgraded to meet current electrical specifications. To ensure that each site is more reliable during severe weather events, lightning and current surge protection devices were installed to protect the electronics and communications equipment.

Eleven data processors were acquired and installed at profiler sites to replace obsolete ones that can no longer be purchased or repaired.


An all-digital surface meteorological sensor package, the Profiler Surface Observing System (PSOS-II), will be installed to replace the GSOS and PSOS units operating at some profiler sites. A 10-meter mast with an anemometer and rain gauge will be added to sites currently without surface wind measuring capability. This will bring equipment availability at the 35 profiler sites into uniformity and provide additional meteorological data.

New signal processing techniques will be tested at the Platteville profiler to determine the best method for acquiring quality data. The new techniques will help alleviate the problems associated with ground and sea clutter and bird contamination.

The Engineering and Field Support Branch will provide operations and maintenance support to the 10 quarter-scale profilers to be installed for the Air Force Tethered Aerostat Radar System (TARS).

Return to Top of Demonstration Division Section

Software Development and Web Services Branch
Alan E. Pihlak, Chief


The responsibilities of the Software Development and Web Services Branch are to provide software support for existing systems, develop new software and database systems as needed, provide Web support for the division's extensive Web activities, and design software to support a national deployment of profilers. To help improve short-term warning and forecast services, up-to-the-minute profiler data are provided on the NOAA Profiler Network (NPN) Website, http://profiler.noaa.gov — the first place to go for wind profiler data. The Profiler Website provides historical archives of wind, temperature, and other profiler information beneficial to researchers for forecasting and modeling both long-term and short-term climate change. A constant goal is to improve the timeliness of profiler data delivery and distribution through work on existing software systems and development of new software.

Branch resources are used in the operation of the Cooperative Agency Profiler (CAP) network, a non-NPN network of profiler sites. FSL, in cooperation with other agencies, ingests profiler data in near real time from different sources ranging from the Environmental Protection Agency to the Japanese Meteorological Agency. Unlike the radars used in the NPN, the CAP sites are operated in many different ways, are owned by about 30 different agencies, and are optimized for different applications. FSL acquires these data, applies its own quality control algorithms to the data, and makes the value-added data available on the Web and to the National Weather Service (NWS). The data from these profilers, distributed primarily via the Profiler Website, contribute significantly to NWS forecasts in areas where the NPN does not operate tropospheric profilers.


In 2002, this branch began delivery of single-station, single-time-period messages to the National Weather Service Telecommunications Gateway (NWSTG). The message format is a simplified subset of the COST-76 standard BUFR format for profiler data. These changes are steps toward future use of alternate communications solutions. FSL's Modernization Division received a grant for modifying the AWIPS workstation to accept and process data in these new formats, which allow data to be processed closer to their time of arrival. These data now arrive at the NWSTG an average of 6 minutes faster every hour. Before changing the formats, data would occasionally arrive too late to be used in the NCEP (National Centers for Environmental Prediction) numerical weather models. Also, GPS precipitable water vapor measurements are now delivered to the NWSTG twice each hour.

The number of profilers in the CAP project more than doubled during the last two years, from approximately 50 in 2001 to over 100 currently. Division resources were shifted to accommodate this explosive growth. Data from CAP are routinely used by NWS forecasters from California to Virginia.

Work on other subsystems required to eliminate the aging NPN Hub is estimated to be 50% complete. Progress on this goal was affected by the growth of the CAP network.

The Wind Profiler Processing Platform (WPPP) is progressing, with decisions made regarding the standard hardware and software configuration.

Low-level, reusable software components based on a station-instrument model have been completed. In this model, station objects own instrument objects capable of producing different types of meteorological data in various composites. Now every station, whether CAP or NPN, can be treated in a similar fashion without writing new software every time a new station is added, removed, or relocated. These components can reside on any Java-enabled processor on the system, as long as it is connected to the Internet, regardless of operating system. Many of these components are currently in use in the production environment.

New, reusable graphics software for displaying winds, temperatures, and moments from the Web and local or remote datasets was completed and is being used on the Demonstration Division Website.


As we enter the third generation of wireless technology, organizations have been sponsored by the National Science Foundation to create, demonstrate, and evaluate a noncommercial prototype high-performance wide-area wireless network for research and education. Wireless technology is in daily use by millions of consumers, and an evolution in observing systems communications must progress in the same direction. It is expected that at least four NPN sites will be converted to use satellite Internet in place of direct phone line connections in 2003. Staff will explore the results of last year's COMET project grant for JINI/TINI technology using packet radio communications, as well as other technologies such as cellular and point-to-multipoint IP.

In 2003, completion of the Phase 1 transition will include retirement of the obsolete NPN Hub and supporting 1980s era peripherals, and completion of a general-purpose serial data client-server component, the basis for several subsystems such as a new Web-based PMT control system.

Improvements and updates will be made to the Demonstration Division Website.

Final work will be completed related to the National Climatic Data Center's (NCDC) profiler data archive. These data, covering the last 11 years, have been reprocessed into a modern storage format and will be sent back to NCDC within the year.

Archived NPN and CAP data will be placed online (http://profiler.noaa.gov) and available for downloading or viewing. At least one year of data will be available online, and requests for older historical data will be processed as received.

Return to Top of Demonstration Division Section

GPS-Met Observing Systems Branch
Seth I. Gutman, Chief


The GPS-Met Observing Systems Branch was formed in response to the need for improved moisture observations to support weather forecasting, climate monitoring, and research within NOAA. The activities of the branch primarily support NOAA's environmental assessment and prediction mission. It creates and disseminates reliable assessments of weather, climate, the space environment and geodetic phenomena using a new upper-air observing system technique in support of advance short-term warning and forecast services. The data and techniques developed and implemented by the branch also support seasonal to interannual climate forecasts as well as the prediction and assessment of decadal to centennial climate change. Due to a largely unanticipated synergy between the requirements for atmospheric remote sensing and the more traditional applications of the Global Positioning System (i.e., positioning, navigation, and time transfer), the branch's activities also promote safe navigation by providing GPS and other observations to the National Geodetic Survey (NGS) Continuously Operating Reference Station (CORS) network, the U.S. Coast Guard (USCG), U.S. Department of Transportation (DOT), and other GPS users in the public and private sectors.

The primary objectives of the GPS-Met Observing Systems Branch are to define and demonstrate the major aspects of an operational ground-based Global Positioning System (GPS) integrated precipitable water vapor (IPWV) monitoring system, facilitate assessments of the impact of GPS meteorological data on weather forecasts, assist in the transition of GPS-Met data acquisition and data processing techniques to operational use, and encourage the use of GPS in atmospheric research and other applications. The work utilizes the resources and infrastructure established to operate and maintain the NOAA Profiler Network (NPN) in achieving these objectives at low cost and risk. The branch collaborates with other FSL divisions (especially the Forecast Research Division, Aviation Division, International Division, Systems Development Division, and the Director's Office which includes the Information and Technology Services group) to achieve objectives of mutual interest and benefit the laboratory, its customers and partners.

This work represents an investment in scientific research, the development of new technologies to improve current operations, and assistance in helping NOAA prepare for the future. The branch has successfully demonstrated all major elements of a reliable, low-cost continuous upper-air observing system that complements other operational and future ground- and space-based observing systems. Newly adapted information network tools and techniques acquire and process GPS and surface meteorological data from NOAA and other public, private, and international partnerships. This capability has permitted rapid expansion of GPS-Met coverage at extremely low cost. The branch has been heavily involved in developing and implementing environmental expertise and technologies to improve NOAA's ability to serve its customers and forge stronger ties with its partners, especially the National Weather Service (NWS), DOT, and Department of Defense (DOD).


The focus of the GPS-Met project during 2002 was on expanding the demonstration network to facilitate assessment of these observations on weather forecast accuracy, investigating other uses for GPS data and meteorological models assimilating GPS observations, and making NOAA and other agencies more aware of these research and development activities. In this regard, the branch completed its second technical review (available at http://gpsmet.fsl.noaa.gov/jsp/review2002.jsp), published four peer-reviewed papers, presented two papers at the annual meeting of the American Meteorological Society, and presented briefings to numerous organizations, including the ASOS Program Management Committee, CORS User Conference, the NOAA Space Environment Center, U.S. Air Force Space Command, and Purdue University. Support, information, and materials were also provided to schools and other institutions, federal agencies, foreign countries, and private companies.

Impact of GPS Water Vapor Data on Weather Forecast Accuracy

The GPS-Met Observing Systems Branch assisted the Regional Analysis and Prediction Branch in performing the fifth consecutive assessment of the impact of GPS integrated precipitable water vapor (IPW) retrievals on weather forecast accuracy. The annual assessments are actually data denial experiments using the 60-km Mesoscale Analysis and Prediction System (MAPS), a research version of the operational Rapid Update Cycle numerical weather prediction model (RUC2) currently running at the National Centers for Environmental Prediction (NCEP). The 60-km MAPS model was again run in a 3-hour data assimilation/forecast cycle over the central U.S. (Figure 53). Each forecast cycle used the same boundary conditions and observations (including rawinsondes, surface, aircraft, wind profiler, and GOES precipitable water); the only difference was the addition of GPS IPW (integrated precipitable water) observations in a second "parallel" run. The 3-hour relative humidity forecasts (with and without GPS) were compared with twice daily rawinsonde observations at 17 NWS upper-air sites to assess the improvement in the relative humidity (RH) forecast accuracy at 5 pressure levels (850 hPa, 700 hPa, 500 hPa, 400 hPa, and 300 hPa). The results are summarized in Table 1, which compares GPS-Met impact assessments over the past 5 years. Table 2 presents the results of Table 1 in terms of percent improvement in 3-hour relative humidity (RH) forecast skill for the lowest two levels evaluated, 850 hPa and 700 hPa. From a weather forecast perspective, these are the two most important levels, since most of the moisture in the atmosphere resides at or below these levels. Figure 54 shows the reduction (increase) in forecast error as a function of the number of GPS stations used in the data denial experiments: 18 in 1998 and 1999, 55 in 2000, 74 in 2001, and 100 in 2002. Figure 55 is the improvement in 3-hour RH forecast skill in 2002 as a function of month of the year. Although RH forecast skill was again greatest during the cold months, improvements during the warm months appear to be larger than previously observed. This is probably related to the increasing size of the GPS network as discussed below.

Figure 53 - Verification Area Using MAPS

Figure 53. Verification area for GPS-Met impact
assessments using the 60-km MAPS model.

Table 1.
Comparison of five years of GPS impact on RH forecasts in the 60-km RUC. Numbers in the columns labeled "Control 2002" and "with GPS" are (forecast–RAOB in % RH) without and with GPS IPW, respectively.
Level 1998-1999 2000 2001 2002 Control 2002 With GPS
850 .15 .38 .39 .50 14.26 13.76
700 .11 .41 .63 .65 16.43 15.78
500 .07 .21 .20 .24 18.07 17.83
400 .03 .01 -.04 -.05 18.54 18.59
300 .01 .01 -.12 -.25 17.84 18.09

Table 2.
Results for 850 hPa and 700 hPa, binned by whether
they were better, worse, or the same with GPS.
850 mb 1998 1999 2000 2001 2002 (No. of Cases in 2002)
Better 25% 28% 37% 38% 45% (266)
Worse 21% 19% 23% 24% 26% (152)
Same 54% 53% 40% 38% 29% (174)
          (74% Same/Better)
700 mb 1998 1999 2000 2001 2002 (No. of Cases in 2002)
Better 23% 27% 39% 49% 52% (310)
Worse 21% 21% 21% 18% 20% (119)
Same 56% 52% 40% 33% 28% (163)
          (80% Same/Better)

Figure 54 - RH Forecast Improvement

Figure 54. 3-hour RH forecast improvement as a function of the
number of GPS stations used in the data denial experiments:
18 in 1988 and 1999, 55 in 2000, 74 in 2001, and 100 in 2002.

Figure 55 - RH Forecast Skill Improvement Vs. Month

Figure 55. Improvement in 3-hour RH forecast skill in 2002 as a function
of month of the year. Percent improvement in RH is defined as 1-*100
(3-hour forecast error with GPS and 3-hour forecast error without GPS).

The magnitude of the improvement shown in Table 1 is relatively small in absolute terms, (0.5% at 850 hPa in 2002), primarily because the evaluation includes days when the addition of GPS has little or no impact on the forecast, or the addition of GPS makes the forecast slightly worse. In fact, the former constitutes the majority of the cases. This occurs when the moisture field is not changing rapidly and is well described by the model using the current suite of operational observing systems. However, a major goal of modern weather prediction is to improve forecasts of severe weather, and this is precisely when GPS appears to be making its greatest contribution. With this in mind, we present the following points.

  • The NWP model is assimilating an integrated quantity (IPW), comparing the difference between the model-predicted values and observed moisture in the vertical column, and distributing the difference to the model as a percent correction. Obviously this is a simplistic approach to distributing the errors, since it has absolutely no physical basis. Nonetheless, we see more or less continuous improvement below 500 hPa, and the improvement diminishes with altitude becoming (on average) slightly negative above 500 hPa. From a pragmatic standpoint, it is better to have an improvement in moisture forecast accuracy where the moisture is, rather then where it is not (i.e., above 500 hPa).
  • A major problem in estimating an initial state for a numerical forecast comes from spatial and temporal aliasing when interpolating discrete observations into an "analysis increment" field. This is especially true for a vertically integrated quantity, because the forecast background error at discrete vertical levels must be estimated from the difference between observed and forecast integrated quantities.
  • The assessment is being carried out in one of the best observed regions on Earth. Considering the number of observations going into modern mesoscale models such as the RUC2, it is surprising that we see any positive impact at all from the assimilation of GPS-IPW. For example, the number GOES PW estimates under cloud free conditions is nearly an order of magnitude greater than the number of GPS sites.
  • Other areas in the western United States, the Caribbean, Mexico, and Canada are not as well observed, however. As a consequence, the potential contribution of GPS-Met to improved forecasts downstream of these regions is probably much greater than it is from the central United States.
  • After 5 years, it is clear that the impact on 3-hour forecasts steadily grows as the network expands, and it seems that there is no reason to doubt that this trend will continue. This is not so much an expression of having a large number of sites per se, but of having a sufficient number of observations when and where they are needed to better define the initial conditions for the model.
  • The relationship between an improvement in RH forecast skill and an improvement in precipitation forecast skill is not straight forward since a large number of other factors besides the amount of moisture in the atmosphere are involved in determining if, when, where, and how much precipitation will occur.

Expansion of the GPS-Met Network

GPS Surface Observing System (GSOS) packages were installed at 5 Nationwide Differential GPS sites and at 25 U.S. Coast Guard and U.S. Army Corps of Engineers Maritime Differential GPS sites. This brings the number of "backbone" sites in the network to 110, with a goal of 200 sites nationwide by 2005.

The GPS-Met team collaborated with NGS in a Center for Operational Oceanographic Products and Services (CO-OPS) program to use GPS to monitor the levels of the Great Lakes and its tributaries. Other participants include the Ohio State University, NOS Office of Ocean and Coastal Resource Management, NOS Office of the Coast Survey, OAR Great Lakes Environmental Research Laboratory, Canadian Hydrographic Service, and the National Resources Canada Geodetic Survey Division. Ultimately eight new sites were added to the network.

To facilitate the expansion of the GPS-Met network, the branch worked on techniques to ingest data from a growing number of GPS continuously operating reference stations (CORS) in the United States. These CORS sites are established by state and local government agencies to improve local high accuracy positioning and navigation services (including surveying, 911 response, and intelligent transportation system applications), and by universities for teaching and research. Most state and local government CORS do not have collocated surface meteorological sensors that permit the GPS data to be used directly for GPS meteorology. However, since these agencies have graciously provided NOAA with access to their observations at no cost, it is worthwhile to see if they could still be used for operational weather forecasting. The branch investigated how water vapor retrieval accuracy degraded as the distance between the GPS antenna and the pressure and temperature sensors increased. Results showed that if an automated surface observing system such as ASOS was within a reasonably short distance (less than 50 km horizontally and 100 meters vertically), then the water vapor retrieval error could be kept to less than 1 mm IPW through a process called "bias fixing." This was done using products generated by the Mesoscale Analysis and Prediction System/Rapid Update Cycle (RUC) Surface Assimilation Systems (MSAS/RSAS), at http://msas.noaa.gov/msas.html, developed by the FSL Systems Development Division. While this technique is not deemed suitable for climate monitoring applications, the level of accuracy is more than sufficient for current mesoscale weather forecasting applications. As seen in Figure 56, the implementation of this technique in 2002 more than doubled the number of GPS-Met stations available in the conterminous United States at virtually no direct cost to the government. Figure 57 shows the configuration of the GPS-Met network, including "backbone sites" operated by United States federal agencies (identified by triangles) and "infill sites" operated by other government agencies, universities, and the private sector (identified by circles).

Figure 56 - GPS-Met Network Growth

Figure 56. Growth of the GPS-Met network and major milestones.

Figure 57 - GPS-Met Network 2002

Figure 57. GPS-Met Network at the end of 2002. Triangles identify
"backbone sites" owned and operated primarily by U.S. federal
agencies. These provide the highest quality GPS IPW retrievals
and are maintained as operational systems. Circles identify
"infill sites" used for network densification. The quality of
retrievals from infill sites vary from site-to site, but all are
suitable for use in weather forecasting. Since infill sites are
not operated or maintained as operational systems, they
have (in general) lower reliability than backbone sites.

Using Meteorological Models to Improve GPS Positioning Accuracy

During average (quiet) geomagnetic and tropospheric conditions, the gradients in total electron content in the ionosphere over the continental United States, and temperature, pressure, and moisture in the lower atmosphere are usually small. Under these circumstances, existing methods to correct for excess signal delays caused by the ionosphere and troposphere over short to moderately long baselines work reasonably well for many but not all (e.g., rapid static and real-time kinematic, or RTK) positioning and navigation requirements. During significant space and tropospheric weather events, however, the constituents of both the ionosphere and troposphere can vary greatly in time and space, leading to rapid changes and large errors in GPS accuracy. During geomagnetic storms, for example, gradients in the electron plasma density of the ionosphere increase considerably, and transients are expected to propagate from high latitudes. At these disturbed times the steep gradients associated with the equatorial ionization anomaly begin to penetrate into midlatitude regions. In similar fashion, strong gradients in pressure, temperature, and moisture commonly associated with severe weather in the lower atmosphere also affect GPS accuracy. While the magnitude of the impact of the lower atmosphere on positioning accuracy is less than the ionosphere, the need for higher accuracy dynamic positioning and navigation requires significant improvements in both areas.

GPS observations can be used to improve weather forecast accuracy, especially during active weather. The observations improve the NWP model's initial description of the moisture field, which, under most circumstances, leads to improved short-term forecasts of atmospheric moisture and precipitation. The parameter being assimilated, integrated precipitable water vapor, is retrieved from the GPS tropospheric signal delay. The tropospheric signal delay is estimated as a free parameter in the solution of the double-difference equation that is used to measure changes in the position of the GPS antenna over time. The NWP model assimilates GPS IPW and other thermodynamic quantities (or proxies thereof) and provides a three-dimensional analysis of the mass and momentum fields. From this, a prediction of the future state of the atmosphere comes from a finite difference or spectral representation to the equations describing geophysical fluid dynamic flow on an unevenly heated rotating sphere. This is an initial value problem (and also a lateral boundary problem for a non-global, limited-area model of a closed set of non-linear partial differential equations) that describes the physical laws governing change of atmospheric parameters including temperature, moisture, wind, and pressure. These parameters can be inverted to provide an estimate of the tropospheric signal delays over the model domain. In similar fashion, improved space weather models will soon be able to do this for the ionosphere using an analogous procedure applied to different physical parameters. These delays can be used to constrain two of the five elements of the GPS error budget: namely the ionospheric and tropospheric signal delay to each satellite in view, thus improving the GPS estimate of position.

The U.S. DOT Federal Highway Administration (FHWA) received funding in 2002 from the Interagency GPS Executive Board (http://www.igeb.gov/org/execsec.shtml) to determine the feasibility of developing more robust atmospheric (ionospheric and tropospheric) corrections using space and tropospheric weather models. In this project, FSL is collaborating with the NOAA Space Environment Center (SEC) and the NOAA National Geodetic Survey (NGS) to investigate how the improved ionosphere and troposphere models could be used to improve GPS data processing, especially the integer-fixing problem for mobile GPS observations and rapid-static GPS surveying.

FSL has already demonstrated that it is possible to use the 20-km RUC model assimilating GPS IPW data to characterize and make short-term predictions (nowcasts) of the signal delays caused by the lower atmosphere with an accuracy about 5 times better then the best available techniques currently used in the FAA Wide Area Augmentation System. Techniques to characterize and predict the total electron content of the ionosphere are under evaluation, with a goal of improving these delays by a factor of about 3.

Atmospheric Infrared Sounder (AIRS) Calibration and Validation

NASA selected the FSL GPS-Met group to collaborate as co-principal investigator with the NOAA NESDIS Office of Research and Applications on a project to evaluate the AIRS (Atmospheric Infrared Sounder) radiometer and other moisture sensing systems aboard the Aqua spacecraft using ground-based GPS water vapor observations. GPS-IPW is now recognized internationally as a base level, climate-quality observation, in part because of the collaborative work conducted by FSL between 1994 and 2002.

The Aqua/AIRS mission provides observations that contain information about the state of the Earth's surface and atmosphere. These observations should be sufficient in quality, quantity, and timeliness of delivery to advance the state of the art of numerical weather prediction and the characterization and understanding of atmospheric and climate processes. To help meet these broad objectives, the Aqua/AIRS project will provide more accurate and higher resolution measurements of atmospheric water vapor profiles than has been available from satellite sensors previously. This investigation is intended to provide accurate measurements that are closely matched in time and space of the column integrated precipitable water vapor (IPWV) product from AIRS. In addition, it is being used to identify the best radiosonde data to use as the basis for AIRS water vapor profile retrievals. The AIRS validation process has been designed to establish in sequence the reliability of Level 0 data – EDRs (eddy dissipation rates); Level 1 data – radiances; Level 2 data – temperature retrievals; and Level 2 data – trace gases, such as H2O, CO, and CO2.

The validation of AIRS radiances is a necessary prerequisite to the validation of derived products (Level 2), since high quality, well-characterized radiance measurements are needed as input to the Level 2 retrieval algorithms. Nine months after the launch of Aqua in May 2002, the high radiometric quality of AIRS observations has been established by other members of the validation team, and derived temperature profiles have been assigned a high degree of confidence. The AIRS science and validation teams have begun producing initial AIRS water vapor profiles and IPWV data. Preliminary comparisons of those data with the NOAA GPS IPWV are now underway. Definitive GPS-to-AIRS IPWV comparisons await the optimization of AIRS water vapor retrieval algorithms. Careful has been taken for data exchange and rapid analysis once the water vapor product stream is ready for full-scale validation.

IPW measurements at 30-minute temporal resolution from the GPS-Met network continue to be archived on an operational basis, and are available as images or data via the Web. Moreover, sample GPS IPWV datasets have been transmitted to NASA/JPL and to individual AIRS validation team members upon request in designated formats, typically netCDF or ASCII. Finally, the fixed locations of the GPS receivers (latitude, longitude, and altitude) have been provided to the AIRS science and validation team so that relevant members can select portions of their datasets (both AIRS and alternate validation measurements) that match the GPS data in time and space.


An outreach effort was funded to inform federal, state, and local government agencies about the NOAA Profiler Network, GPS Meteorology, and the advantages of joint use and sharing of GPS and surface meteorological data in near real time. The Aviation Division assisted in providing booth displays and presentations at major meetings, including the American Meteorological Society, World Space Congress, and the Intelligent Transporation Society of America.


During 2003, GSOS packages will be installed at approximately 20 new backbone sites: 6 new NDGPS sites; 10 Maritime DGPS sites in the conterminous U.S., Puerto Rico, and Hawaii; and about 6 other backbone sites to be selected in collaboration with other NOAA organizations. Tools to incorporate and display GPS-Met data on operational AWIPS workstations will be prototyped and evaluated. The GPS-Met network will continue to expand through the incorporation of infill sites operated by State Departments of Transportation, universities via SuomiNet, and other organizations. Assessment of GPS-Met on weather forecast accuracy will be facilitated. The branch will work with other organizations to assess the utility of meteorological models assimilating GPS observations to improve real-time GPS positioning accuracy. It will also become more involved in multiagency efforts to utilize ground-based GPS observations for satellite calibration and validation.

Return to Top of Demonstration Division Section

Facilities and Systems Administration Branch
Bobby R. Kelley, Chief


The objectives of the Facilities and Systems Administration Branch are to manage and support the Demonstration Division communications and computer requirements. Duties include performing systems operations, systems maintenance, systems administration, network administration, NOAA Profiler Network (NPN) telecommunications administration, and support of the Global Positioning System Integrated Precipitable Water Vapor (GPS-IPWV) demonstration project.


While NPN processing is in the process of being converted to a network of off-the-shelf PCs running Linux, the original two clusters of 13 micro-VAXs are still operating as primary and backup data processing and distribution systems. Components of the new modernized NPN processing system are currently running in parallel with the original NPN processing and distribution system. Work is ongoing to replace the NPN VAXs and NPN Hub software with robust production, backup, and development environments, using low-cost off-the-shelf equipment. Backup communications for NPN data acquisition are being converted to a newer PC running Linux to acquire data as necessary from a DOMSAT receiver. A PC-based server running the Linux operating system has replaced Sun Microsystems equipment for data processing and Webpage hosting. Two remaining Sun systems that are still in modest use will be decommissioned by April 2003. Support for the GPS-IPWV demonstration project has grown from 18 PCs running Linux to a total of 34.

Day-to-day work includes new component installations and system configuration on the division network, network problem isolation and maintenance, system configuration modifications to meet division requirements, system problem isolation and maintenance, in-house telecommunications maintenance or coordination of contracted maintenance, peripheral installation and configuration, computer and network security, preventive maintenance, information technology purchasing, and routine file system backups. A primary focus of the Facilities Management and Systems Administration Branch is computer and network security responsibilities: ensuring system and data integrity and maintaining dependable NPN and GPS-IPWV data acquisition, processing, and distribution. Full-time (24/7) operations coverage is provided during normal workdays through the Profiler Control Center in Boulder and via pager during nights, weekends and federal holidays.

Data telecommunications responsibilities cover 38 NPN data circuits within the lower 48 states and in Alaska. When Unisys elected to discontinue providing telecommunications services in 2002, it became necessary to obtain an alternate provider. AT&T provided telecommunications services directly on an interim basis at approximately the same cost as the original services through Unisys. With the establishment of a Memorandum of Understanding (pursuant to the Economy Act) between FSL and the Department of the Interior's Minerals Management Service (DOI/MMS), an interagency fund transfer was accomplished to obtain telecommunications services at reasonable cost and without requiring replacement of $130,000 of existing NOAA equipment. The DOI/MMS contract can provide services through Fiscal Year 2005. Meanwhile, investigation is ongoing to determine the viability and cost effectiveness of alternatives such as wireless technologies and satellite-based communications to provide future communications services, meeting potentially expanding requirements for additional circuits and greater bandwidth while reducing communications costs.

A stand-alone computer room air conditioning unit was installed in the NPN computer facility. This has served to further stabilize the computer room temperature and provides redundancy for the building air conditioning system which was originally the only source of cooling. The installation was successfully accomplished without interrupting NPN operations. The investment serves to further improve NPN processing reliability.


The branch will maintain current operations and ensure continuous and dependable NPN acquisition, processing, and distribution of NPN and GPS-IPWV data (Figure 58) to all customers. Further development, testing, and implementation of the modernized NPN processing system will continue. To expand earlier successes, a low-cost approach of using off-the-shelf PCs running the Linux operating system will continue. Computer network equipment in the NPN computer facility will be replaced to maintain pace with building network upgrades, enable easier maintenance of the NPN network, and support expansion of the NPN network. Investigation of alternative communications options will be undertaken to provide increased bandwidth for NPN data acquisition and remote system control, and reduce future communications costs.

Figure 58 - Real-Time Water Vapor Interface

Figure 58. GPS-Met Network Webpage screen
showing the Real-Time Water Vapor Interface.

Return to Top of Demonstration Division Section

FSL Staff FSL in Review (Other Years) FSL Forum Publications