Return to: December 1996 FSL Forum FSL Home Page

Introduction

Much progress has been made on the development of an operational surface-based integrated precipitable water vapor observing system for NOAA using the Global Positioning System (GPS). This system, now known as GPS-IPW, involves the joint efforts of two NOAA Environmental Research Laboratories in Boulder: FSL's Demonstration Division and the Environmental Technology Laboratory's (ETL) System Demonstration and Integration Division. Other organizations collaborating with FSL and ETL on this project are the Scripps Institution of Oceanography (SIO), the University of Hawaii (UH) at Manoa, and the University NAVSTAR Consortium (UNAVCO; affiliated with the University Corporation for Atmospheric Research (UCAR)).

The need for improved water vapor observations and the potential use of GPS to monitor the total (integrated) precipitable water vapor in the atmosphere were discussed in the September 1994 issue of the FSL Forum. A followup article in the December 1995 issue reported on the development of the GPS-IPW observing system and its installation at selected NOAA Profiler Network (NPN) sites, primarily for atmospheric remote sensing. This article also discussed the results of an ERL study which demonstrated that:

• Microwave water vapor radiometers are not required to calibrate the GPS observations, and

• It is possible to produce satellite orbits of sufficient accuracy so that the GPS data can be processed in less than 48 hours instead of two weeks, as previously required.

Significant Events in 1996

GPS Network - By last September, 10 ERL GPS-IPW systems had been installed: nine at NPN sites and one at the NWS National Data Buoy Center (NDBC) at the NASA Stennis Space Center in Mississippi (Figure 1). The GPS-IPW systems transmit data to the GPS Hub in Boulder with better than 97% reliability. These systems are part of the network of GPS Continuously Operating Reference Stations (CORS) managed by NOAA's National Geodetic Survey.

Three of the four significant GPS-IPW system failures in 1996 were associated with lightning strikes. We are working with UNAVCO to evaluate various in-line lightning surge suppressors at NOAA profiler sites. Once the most reliable model has been chosen, the units will be installed at the profiler sites.



Figure 1. Existing and future sites of an operational surface-based integrated precipitable water vapor observing system using the Global Positioning System (GPS-IPW).

GPS Hub - Last January, ERL installed the GAMIT geodetic processing software package, jointly developed by the Massachusetts Institute of Technology (MIT) and SIO, on a workstation in the Profiler Control Center in Boulder. ERL staff were trained on GAMIT fundamentals, and we began processing our own GPS data in February. We have been using this capability to evaluate many issues associated with real-time data processing and data quality. In the meantime, SIO and UNAVCO continue to handle daily processing of the water vapor data, although we expect to assume this responsibility next year. Our university collaborators will then be able to concentrate more on other critical research and development activities.

After handling the demands of GPS data processing for seven months, we are able to better define the near real-time data processing requirements for networks of GPS-IPW systems. To ensure our progress, we have ordered a new workstation to replace the original unit. A dedicated, more powerful computer will help us achieve significant improvements in performance in 1997.

Data and System Performance Monitoring - The results of data processing are routinely reviewed to evaluate system performance and to identify features of meteorological or climatological interest for further study. Some interesting examples of the data acquired during the past year are presented in Figures 2 and 3.



Figure 2 depicts data acquired by the radar wind profiler and GPS-IPW systems located at the Haskell, Oklahoma, NPN site during a spring storm in April 1996.

Also shown are radiosonde precipitable water vapor data acquired at the nearby Atmospheric Radiation Measurement (ARM) Cloud and Radiation Test (CART) Site Boundary Facility in Morris, Oklahoma. Of particular interest is the generally excellent correlation between high signal power from the vertical beam of the profiler, usually associated with [clouds and] rainfall, and peaks in precipitable water vapor measured by the GPS. With the notable exception of a small GPS-IPW peak at about 1200 UTC on day 112 (April 21), there is almost a one-to-one correlation between the radar signal power and the precipitable water vapor. Observations such as this are made possible as a result of the high temporal resolutions of the profiler and GPS, and the all-weather capability of the GPS-IPW system.



Figure 3. Very rapid change in total precipitable water vapor observed at the Lamont, Oklahoma, NPN site in July 1996.

Figure 3 captures what we believe to be a very unusual occurrence. On 22-23 July 1996 total precipitable water vapor observed by the GPS-IPW system at the Lamont, Oklahoma, NPN site rose more than 3.5 cm in about 6 hours, and then fell almost as much in the next 6 hours. This event is the result of thunderstorm outflow confirmed by surface measurements showing an increase in surface pressure of 5 mb in 3 hours, a temperature drop of about 15 degrees in the same period, winds gusting to 15 m s-1, and rainfall at the rate of 3 mm per hour. Radiosonde data acquired at the nearby ARM Central Facility (also shown in Figure 3) confirms the event, but not in the detail provided by GPS.

Data Distribution - As a contributor to the GPS CORS network, we send our GPS data and surface meteorological observations to the National Geodetic Survey daily. These data are available to the public through an Internet location:

ftp cors.ngs.noaa.gov
login: anonymous
password: your complete e-mail address.

In addition, processed GPS precipitable water vapor data are sent to several organizations every day for analysis and collaborative research and development, including the National Environmental Satellite, Data, and Information Service (NESDIS) Satellite Research Laboratory, FSL's Forecast Research Division, ETL and its System Demonstration and Integration Division, and the Department of Energy ARM Program. It is expected that the National Centers for Environmental Prediction (NCEP) and FSL's Facility Division will also start receiving these data early in 1997.

Near Real-time Data Processing Demonstrated - Last year UNAVCO demonstrated a significant breakthrough in GPS processing technology when they were the first to process GPS-IPW data acquired by ERL with only a 2-hour delay. UNAVCO near real-time precipitable water vapor estimates are available on the World Wide Web at: http://www.unavco. ucar.edu/~rocken realtime.html.

Assimilation of GPS Data into Numerical Weather Prediction Models - A necessary step in the development of any new observing system involves an assessment of the impact of these new data on weather forecast accuracy. FSL's Forecast Research Division is conducting a study on how surface-based GPS precipitable water vapor data affect mesoscale numerical forecasts.

Current plans are to assimilate the GPS data into FSL's Mesoscale Analysis and Prediction System (MAPS), the development version of the Rapid Update Cycle (RUC) numerical weather prediction model running at NCEP, using three-dimensional variational analysis. Identified options include direct analysis of the precipitable water vapor field and/or use of a forward model to estimate the zenith tropospheric signal delay (ZTD) from the forecast background, and use of the ZTD residual (difference between the forecast and observation) in variational analysis. Collocation studies, sensitivity tests, and verification and evaluation studies are planned for 1997.

Comparison of Precipitable Water Vapor from Satellite Sensors and Surface-Based GPS - Comparisons of GPS-IPW measurements with temporally and spatially sampled precipitable water vapor data from the Geostationary Operational Environmental Satellite (GOES), Polar Orbiting Environmental Satellite (POES)/Television and Infrared Operational Satellite (TIROS) Operational Vertical Sounder (TOVS) show the potential of using surface-based GPS data to correct wet biases in the satellite precipitable water vapor in an operational environment. A paper on this topic will be presented at the American Geophysical Union meeting this month, and plans are being made to extend this line of investigation to different seasons and climate regimes next year.

GPS Surface Observing System (GSOS) - As reported last December, FSL is working with the National Data Buoy Center to develop a small surface meteorological sensor package called GSOS, which can be deployed at GPS sites with power and communications, such as the U.S. Coast Guard (USCG) and the U.S. Army Corps of Engineers (USACE) differential GPS (DGPS) sites along the East and West Coasts and major river systems in the United States. The first GSOS was installed at the USCG Electronics Engineering Center at Wildwood, New Jersey, last February, and a second unit has been operating continuously at ERL in Boulder since August.

Plans for 1997

Goals for the next year are to:
• Demonstrate the ability to acquire and process in real time GPS-IPW data from a network of sites consist of NOAA and other-agency GPS receivers.

• Provide these data to modelers and forecasters every 30-60 minutes with no more than a 15-minute delay.

• Assist in demonstrating that GPS-IPW data provide a positive impact on mesoscale forecast accuracy.

To accomplish these activities, we must have both long baseline geodetic control and highly accurate satellite orbits in real time. Currently, these data are available with about a one-day delay.

Real-time Geodetic Control - ERL will install four "Long Baseline Sites" (locations shown in Figure 1) in 1997. These sites will furnish the geodetic control that is traditionally provided by the widely spaced International GPS Service (IGS) for Geodynamics tracking stations, but will do so in real time. Data from these sites (identified in the Table below) will enable us to form the baselines for calculating precipitable water vapor in real time using the "absolute" technique first demonstrated by SIO and UH in 1995.

Table
Long Baseline Sites and Approximate Baseline Lengths



Real-time Satellite Orbits - Most civilian orbit facilities, seven at this writing, generate improved satellite orbits for the previous day in less than 24 hours. Real-time precipitable water vapor calculations, however, require orbits of sufficient accuracy to be available at all times during the current day. Right now, the most promising way to accomplish this appears to be by predicting the orbits of the GPS satellites at least one day in advance. Some orbit facilities, including SIO, Jet Pro-pulsion Laboratory, and the University of Bern, Switzerland, are already producing predicted orbits on an experimental basis. We plan to perform an assessment of the impact of these orbits on precipitable water vapor, as we did in 1995, in evaluating the impact of rapid orbits on precipitable water vapor accuracy.

Additional GPS-IPW Systems - In 1997, we plan to install GPS-IPW systems at two additional NPN sites in the Central United States: Granada, Colorado, and DeQueen, Arkansas. Sometime next spring, we also plan to install GPS and GPS Surface Observing Systems at the three new NPN sites in Alaska: Central, Glennallen, and Talkeetna.

Expanding the GPS-IPW Network - As previously discussed, a very cost-effective way of expanding the number of GPS-IPW sites is to utilize the GPS data acquired by other federal agencies such as the U.S. Coast Guard. We hope to be able to amend an existing Memorandum of Agreement between the NOAA National Geodetic Survey and the U.S. Coast Guard which will allow us to install GPS Surface Observing Systems at U.S. Coast Guard and U.S. Army Corps of Engineers differential GPS sites, and to be able to transmit the surface meteorological data acquired by these units to the National Geodetic Survey along with the GPS data already acquired there. In cooperation with the U.S. Coast Guard, when the amended Agreement is ratified, we plan to install 10 U.S. Coast Guard differential GPS units at the sites identified in Figure 1.

Assessments of GPS Effect on Forecast Accuracy - For their study, scientists in FSL's Forecast Research Division selected the U.S. Coast Guard differential GPS sites because they are expected to provide the most definitive information needed to assess the impact of GPS data on mesoscale forecast accuracy. During 1997, we believe that these activities will lead to demonstrated improvements in operational moisture, cloud, and precipitation forecasts in the MAPS/RUC. In addition, we hope to demonstrate the ability to predict GPS tropospheric delays using the information provided by GPS-IPW observations and surface meteorological data. This capability is expected to have a significant impact on GPS surveying and positioning accuracy.

(Seth Gutman and Anthony Simon are systems analysts in the Forecast Systems Laboratory Demonstration Division, headed by Russell B. Chadwick. Daniel Wolfe is a meteorologist in the Environmental Technology Laboratory System Demonstration & Integration Division, headed by M.J. Post. More information on this topic can be found at the FSL Demonstration Division Website Homepage: http://www-dd.fsl.noaa.gov, and at the ETL GPS Water Vapor Homepage: http://www4. etl.noaa.gov/gps.html.)