Fiscal Year 1995

FSL in Review

Demonstration Division


Division Home Page
Division Personnel Review Homepage

Russell B. Chadwick , Chief

Objectives

The Demonstration Division evaluates promising new atmospheric observing technologies developed by the Environmental 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. Currently the division is engaged in five major projects:

NOAA Profiler Network

The division manages, operates, and maintains the NPN consisting of 32 wind profilers located mostly in the central United States (Figure 19). The NPN wind profilers are upward looking, sensitive 404 MHz Doppler radars that measure the winds above the profiler site. They are specifically designed to measure vertical profiles of horizontal wind speed and direction. Typically, a wind profiler produces a vertical "stack" of winds once an hour from near the surface to above the tropopause. The wind profiling radars have an antenna with three fixed beams and no moving parts, and they can operate in most weather conditions including cloudy and precipitating weather. The radars are sensitive enough to detect fluctuations in refractive index caused by the turbulent mixing of volumes of air with slightly different temperature and moisture content. The resulting fluctuations are used as a tracer of the mean wind in the clear air.



Figure 19. NOAA Profiler Network site locations.

As part of NOAA's goals to modernize the National Weather Service, NPN data are distributed to all NWS forecast offices, government and university atmospheric researchers, private meteorologists, the National Centers for Environmental Prediction (NCEP), the Storm Prediction Center, and foreign agencies responsible for weather prediction. Division staff collaborate with other federal agencies and foreign governments in their efforts to develop and implement wind profilers, as listed in the examples below:

Forecasters and the research community routinely use profiler data to:

Radio Acoustic Sounding Systems (RASS) for Temperature Profiling

The original concept for an operational NPN envisioned the Doppler radar profiler as part of an integrated upper-air remote sensing system to measure wind speed and direction, temperature, and humidity. The division is now involved in a long-term effort to help meet the NPN productivity goals, which can be fulfilled, in part, through the addition of RASS to produce temperature profiles and the Global Positioning System (GPS) to measure atmospheric water vapor. (The photo of a profiler site in Figure 20 shows multiple observing systems: a profiler surface observing system, a profiler radar antenna, RASS, and a GPS water vapor antenna.)

RASS emits an acoustic signal whose propagation velocity is measured by the wind profiling radar. The speed of the sound signal, which is measured at each level, is converted to a virtual temperature (similar to the actual temperature except for a small correction due to water vapor in the air). Although the temperature measurements produced by the five newly installed RASS units maintain an accuracy better than 1oC, the altitude to which they can measure is limited. The stronger the horizontal wind, the greater the likelihood that the acoustic signal will be carried out of the radar beam, the reason for placing acoustic sources at all four corners of the wind profiling radar. RASS measurements with the 404-MHz profilers typically extend up to 3-5 km above ground.

Surface-Based GPS Water Vapor Monitoring

The measurement of atmospheric water vapor is essential for weather and climate research and operational weather forecasting. A high priority in modern weather prediction is to improve the accuracy of short-term cloud and precipitation forecasts, but the lack of timely water vapor data has severely limited this progress. The Department of Defense Global Positioning System, normally used for navigation, positioning, and time transfer, provides significant opportunities to measure elements of the atmosphere's thermodynamic structure. The division, in collaboration with the NOAA National Geodetic Survey and Geosciences Laboratory, Scripps Institution of Oceanography, the University of Hawaii, and the University NAVSTAR Consortium (affiliated with the University Corporation for Atmospheric Research, UCAR), has taken the lead in the use of the surface-based GPS technique for monitoring atmospheric water vapor. This technique involves the use of low-cost GPS receivers on the ground which continuously measure the propagation delays of the GPS signal due to variations in atmospheric water vapor. These delays are readily converted to estimates of total precipitable water vapor, the depth of water that would be realized if all the vapor in a vertical column were condensed. Measurements are made every half hour at seven NPN sites with better than 96% reliability.



Figure 20. White Sands Missile Range profiler site with multiple observing instruments.

Alaska Profiler Network

The concept for a profiler network in Alaska dates back to December 1989 when Mt. Redoubt, located southwest of Anchorage, erupted and sent volcanic ash 38,000 feet high into the atmosphere. The dangers that could be caused by the ash became obvious when the passengers and crew of a KLM 747 were placed in peril as the airliner flew through the ash cloud. The ash caused extensive damage to the aircraft, but good piloting prevented a major catastrophe. The NWS Alaska Region management immediately began investigating ways to forecast aviation hazards resulting from airborne volcanic ash. A likely tool for predicting ash trajectories was the profiler, which can measure wind profiles up to 53,000 feet and give updates every 6 minutes. In March 1995 NOAA management asked the Demonstration Division to carry out a plan to deploy three new profilers in Alaska. The data from these profilers will serve two purposes: to warn of hazards to commercial aviation in the event of a volcano eruption, and to support other National Weather Service activities in Alaska. (Figure 21 shows the currently operating profiler site at Homer, Alaska, and planned sites at Glennallen, Talkeetna, and Central.)

Division staff collaborate with DOD and the ERL Environmental Technology Laboratory (ETL) in the implementation of the Alaska Profiler Network. At the time planning for the Alaska project began, the U.S. Air Force was using an NPN profiler in support of launch operations at Vandenberg AFB, California. This 404 MHz profiler needed to be upgraded to the new frequency of 449 MHz to accommodate operational considerations. NOAA and the Air Force agreed to jointly develop 449 MHz profilers combining their similar schedules and requirements. The Vandenberg AFB profiler upgrades will include a profiler data processor supplied by Radian International and an antenna and power amplifier supplied by Loral. The Alaska profilers will use existing data processing systems and the Loral antennas and power amplifiers. Also, the specifications for the antenna and power amplifier meet both Air Force and NOAA requirements, allowing joint use of certain spare parts to reduce logistics costs. Staff support to ETL continues in the ongoing upgrades to the Vandenberg AFB profiler.



Figure 21. The Alaska Profiler Network. The Homer profiler is operational, and the Talkeetna, Glennallen, and Central sites are in the planning stage.

Boundary Layer Profilers

One of the main functions of the Demonstration Division is to provide a development testbed which can be used to validate new sensors that may become part of NOAA's future upper-air observation system. In order to improve forecasts, NOAA needs more high-quality observations, but does not have the necessary resources to produce all of the data that could be of benefit.

In recent years, different agencies in several parts of the country have begun operating radars that measure winds and temperature in the boundary layer (normally less than 3-4 km). Approximately 50 boundary layer profilers (BLP) are either currently in operation or are scheduled to be installed soon. If the data from these radars can be gathered together into a real-time network, it is likely that the resultant dataset will have utility in numerical weather prediction, subjective forecasting, and research applications.

The Demonstration Division initiated a project designed to acquire BLP data with little cost to NOAA. Appropriate real-time quality control algorithms are being developed, and the availability and general quality of the BLP network are being evaluated. Once BLP data have been acquired in sufficient volume and of sufficient quality, plans will be made to output these data to the operational community along with data from the NPN.


Accomplishments

NOAA Profiler Network
Profiler Power Amplifier and Antenna Reliability Enhancements
Quality Control of Hourly Winds
RASS Temperature Profilers
GPS Water Vapor Monitoring
Alaska Profiler Network
Boundary Layer Profilers

Projections


NOAA Profiler Network

Staff continued to operate and maintain the NPN and supply upper-air and surface data to a wide range of users. All 32 profiler sites are operating and routinely sending data to the Profiler Control Center (PCC) in Boulder. The data from each site are transmitted to the Profiler Hub computer system, within the PCC, for processing and quality control. The hourly-averaged profiles are then sent to the NWS Telecommunications Gateway (NWSTG) for distribution to the regional NWS AFOS circuits and NCEP. The datasets continue to be available to about 130 universities, the private sector subscribers, the government research community, and the World Meteorological Organization (WMO) community. All six-minute and hourly-averaged profiler data are archived by the National Climatic Data Center.

...return to index

Profiler Power Amplifier and Antenna Reliability Enhancements

Within about a year after acceptance, nearly all of the profilers had exhibited numerous failures of components in the power amplifier and antenna subsystems. Working with the prime contractors and their subcontractors, division staff initiated a program to identify, test, and implement design changes required to increase the reliability of these subsystems. The most extensive changes involved modifications to the power amplifiers. The amplifier contractor redesigned the circuitry in the 1 kilowatt solid-state power amplifier modules, changed the way that the power is reduced for the profiler low-mode operation, and increased the cooling airflow in the power amplifier cabinet. These modifications, combined with careful rebuilding of the amplifiers and lower operating power, have reduced the number of component failures to levels consistent with the overall profiler mean time between failure requirements. The amplifiers are being modified at a rate of about one per month.

The profiler antennas exhibited failures in mechanical switches in the beam steering unit and water intrusion in the antenna radomes. The switches were redesigned and new beam steering units are being installed in all profilers. The source of the water intrusion was identified and antennas are being tested and repaired at a rate of about one per month.

...return to index

Quality Control of Hourly Winds

Although the implementation of the Weber/Wuertz continuity method in October 1993 greatly improved the quality of the hourly winds, difficulties with detecting interference contamination and bird contamination remain. Interference is particularly difficult to detect because it generates clusters of coherent, erroneous winds. Since these winds mutually agree, single-station, profiler-only quality control methods are incapable of establishing them as bad. However, pre-liminary investigations indicate that model "background" or "consistency" checks may have the skill to locate these erroneous winds. The background checks utilize forecasts generated by the NWS operational Rapid Update Cycle model. Case studies have shown encouraging results: in cases where the profiler data were bad, the differences between the RUC and profiler winds were large, and in cases where the profiler data were good, the differences were small.

Bird contamination is another difficult quality control problem. Until recently, not enough was known about the degradation of hourly profiler winds in the presence of bird migration to accurately evaluate quality control algorithms. Rawinsonde data from the four profiler sites at the DOE Southern Great Plains CART have provided revealing information on the effects of bird contamination on the quality of the NPN hourly winds and have benefitted the evaluation of "bird detection" quality control algorithms. Three of these CART sites are located in Oklahoma, at Lamont Vici, and Haskell, and the fourth at Hillsboro, Kansas. At the Lamont site, radiosondes are launched 5 times each day under normal conditions and 8 times each day during Intensive Operating Periods (IOPs). At the remaining sites, launches are performed once per day, except during IOPs when 8 radiosondes are launched. Data from all four sites were obtained for 4 three-week intervals during the period 1 June 1994 - 30 March 1995. Root-mean-square (RMS) differences comparing profiler data with the rawinsonde data in bird migration and nonbird migration seasons were surprisingly close. Perhaps the bird contamination problem is not as severe as previously thought. Previous estimates of RMS differences in bird and nonbird situations were given at approximately 10-15 m s-1. Preliminary statistics from the current study show differences closer to 1-2 m s-1.

...return to index

RASS Temperature Profilers

The quality of the hourly temperature data was investigated by comparing virtual temperatures measured by the RASS at Platteville, Colorado, and virtual temperatures measured by the NWS rawinsonde at Denver. Temperatures were compared for the periods 8 June 1993 - 7 September 1993 and 1 December 1993 - 28 February 1994. Results showed that, overall, the comparison was quite favorable: the RMS difference between the two instruments was 1.58oC and 1.27oC for the summer and winter datasets, respectively, which is in good agreement with other RASS-to-rawinsonde comparisons and with rawinsonde-to-rawinsonde comparisons. Approximately 2% of the comparison points, however, were outliers caused by errors in the RASS measurements. Most of the outliers occurred at higher altitudes where the acoustic signal is expected to be weak. In fact, in nearly 40% of the outlier cases, the RASS "temperatures" were not temperatures at all, but interference passing consensus in the probable absence of acoustic signal.

Quality control procedures to flag these outliers as bad were designed, tested, subjectively inspected, and applied to the datasets described above. The procedures are based on signal strength thresholds and the Weber/Wuertz continuity method that is also used to quality control hourly winds. They were evaluated by repeating the RASS-to-rawinsonde study described above. The results were very good; 80% of the RASS errors identified in the dataset were located and flagged as bad by the quality control procedurs. All of the remaining (undetected) errors were the result of interference. As described above, for hourly winds, interference-generated errors are particularly difficult to detect with single-station, profiler-only quality control methods. Model consistency or horizontal consistency checks will need to be applied to detect interference. In addition to the high probability of detection indicated above, the algorithm also displayed a very low false alarm rate: only 2 points not labeled as RASS errors were labeled as bad by the algorithm. The algorithm is now applied to the operational RASS data stream.

...return to index

GPS Water Vapor Monitoring

In April 1995, the Demonstration Division installed four additional GPS systems at NPN sites: three in Kansas and Oklahoma, and one at White Sands, New Mexico. Two of the GPS systems in Kansas and Oklahoma were funded by the DOE ARM program, and the White Sands GPS receiver was funded by the U.S. Army. The computer hardware and software needed to continuously monitor the status and to download the data from other GPS profiler sites was installed during the year.

Division staff conducted a study to compare GPS-derived water vapor estimates with 338 rawinsonde measurements and 4416 microwave water vapor radiometer measurements, and to evaluate how more frequent GPS processing (the "rapid" orbit) affects precipitable water vapor calculations. Until a breakthrough by the Scripps Institution of Oceanography (SIO), the improved orbits needed to calculate water vapor from GPS signal delays were not available for 2-3 weeks after the GPS data were acquired. Since the SIO breakthrough, these orbits are available within one day, which makes it possible to process the data and provide water vapor data to the numerical weather prediction models in the Forecast Research Division for evaluation and sensitivity testing. The Demonstration Division compared different data processing techniques, one of which requires a water vapor radiometer to be collocated with a GPS station for calibration purposes. The comparison demonstrated that 1) water vapor calculated with the rapid orbit is just as accurate as the estimates made with a conventional "precise" orbit; moreover, the time needed to calculate water vapor with GPS was reduced from three weeks to one day, and 2) data processing techniques not requiring a radiometer for calibration were just as accurate as those that did; this eliminated the need for a radiometer in the operational GPS water vapor monitoring system thus reducing cost and improving reliability without sacrificing data accuracy. (Figure 22 shows a typical comparison of GPS, radiometer, and rawinsonde-derived water vapor data at Lamont, Oklahoma, for 30 days during April and May 1995.) Figure 22. Water vapor comparison at the Lamont, Oklahoma, site during April and May 1995. The division identified opportunities to increase the number of GPS observing systems suitable for monitoring atmospheric water vapor at very low cost to NOAA. The U.S. Coast Guard and U.S. Army Corps of Engineers are deploying a network of 50 GPS receivers along the coasts and major river systems of the United States to facilitate ship navigation and enhance safety. Data from these systems will be continuously transmitted to a sister NOAA organization, the National Geodetic Survey (NGS), where they will be available for hourly water vapor calculations. Staff initiated discussions with NGS and the U.S. Coast Guard for permission to deploy small surface meteorological packages at Coast Guard and Corps of Engineers sites, and to receive these data along with the GPS data in Boulder for water vapor monitoring. Staff developed the requirements specification for the surface meteorological package (called the GPS Surface Observing System - GSOS) in collaboration with ETL and the NWS National Data Buoy Center (NDBC).

...return to index

Alaska Profiler Network

Since approval of the Alaska Profiler Network project in March 1995, progress has been rapid in the areas of overall planning, siting of the new profilers, and system acquisition. To complement these plans, division staff also developed project management tools for tracking schedules. An interagency management committee was organized involving NOAA staff from FSL, ETL, NDBC, the Systems Acquisition Office, and the NWS Office of Systems Development, Office of Meteorology, and the Alaska Region. Discussions began with the NWS Office of Meteorology and the Alaska Region to determine general locations of the Alaska profilers, which must comply with specific technical and logistics requirements. A "Right of Access" for each site was obtained by staff at the Government Real Property office to perform electromagnetic measurements. Reports were prepared covering an electromagnetic spectrum survey for each identified site. The technical requirements for a new generation of 449 MHz power amplifiers and 449 MHz antennas were specified, using input from members of the interagency management committee. Industry proposals were solicited and ranked, and contracts were awarded in September to Loral Defense Systems (formerly Unisys Corporation). Contract awards included a sole source contract for receivers, exciters, signal processors, and shelters. An indoor staging space in Boulder, the Boulder Assembly Facility, was acquired to support the assembly and testing of all system components except the antenna.

...return to index

Boundary Layer Profilers

Most of the work to date on this new project has involved planning and coordination. Access to 80% of the BLPs can be accomplished by collaborating with four agencies, and the main emphasis has been on working with these groups. For development purposes, the division has been ingesting data from BLPs operated by ETL. These data are being used to get most of the processsing in place which can then be used for any of the BLPs that are subsequently acquired from other agencies.

...return to index

Projections

NOAA Profiler Network

Radio Acoustic Sounding Systems

GPS Water Vapor Monitoring

Alaska Profiler Network

Boundary Layer Profilers

...return to index

Maintained by: Wilfred von Dauster