ASOS Stations from FAA, NWS and Archived at NCDC

1. Progress Report February, 2002Evaluation of the Light Scattering Data from the ASOS NetworkSubmitted byRudolf B. HusarCenter for Air Pollution Impact and Trend AnalysisFebruary 18, 2002Submitted to James F. MeagherNOAA Aeronomy Laboratory R/ALBoulder Colorado

3. ASOS Stations from FAA, NWS and Archived at NCDC For this analysis data for 220 stations were available from the NCDC archive These ASOS sites are mostly NWS sites, uniformly distributed over the country (Imagine if we could get the entire set, including the DOD sites, not listed)

4. Data Quality Assessment • Co-located ASOS Sensors – absolute calibration • Lower detection limit (0.05 km-1, 50 km visual range) • Random sensor malfunctions • File format problems

5. Comparison of Sites with Duplicate ASOS Sensors • Co-located ASOS sensors are installed at different runways of the same airport. • Dual ASOS sensors (55) are distributed uniformly over the 800+ station network • Triple sensors are particularly useful for sensor calibration and consistency checking

6. Duplicate Sensors: Good Sites Dallas-FW, TX Erie, PA • At several duplicate sites the 2-sensor correlation is excellent and the absolute values also match. • This indicates that the scattering sensor per se has high precision and temporal stability. San Diego, CA Houston, TX

7. Tulsa, OK Atlanta, GA Duplicate Sensors:Mediocre Sites • Some sites (e.g. Tulsa, OK) show very high correlation between the sensors but they are off by a factor. • Other sites indicate poorer correlation and a significant offset.

8. Albuquerque, NM Duluth, MN Albuquerque, NM Duplicate Sensors:Poor Sites • Duplicate sensors at some sites show significant deviation in scale and offset. • The nature of these deviations indicate poor instrument calibration maintenance for the ASOS visibility sensors.

9. Three Sensor Comparison Cleveland, OH • At 7 NWS Sites, there are 3 ASOS visibility sensors which allow more detailed sensor evaluation. • Both at Cleveland, OH and Hartford, CT Bext1 and Bext3 show excellent correlation, R2=0.99 • On the other hand, the Bext1 and Bext2 correlation is poor. This indicates that the Bext2 sensor either • produces bad data or • it is located at a site with significantly different Bext Hartford, CT

10. Three Sensor Comparison New York JFK • At New York JKF airport, New York La Guardia

11. Washington, Dullas Philadelphia, PA

12. Comparison of Sites with Duplicate ASOS Sensors At the St. Louis Lambert airport, the Bext from the two ASOS sensors track well (right, top) However, the absolute magnitude of the Bext values differ by a factor When the ASOS sensor values are multiplied by a 1.66, the two signals are virtually identical (right, bottom) This indicates that at this site, the absolute calibration of the sensors differs by a factor of 1.66

13. Bext Comparison at other Duplicate Sites The signal pattern of adjacent sensors (at different runways) is consistent. However, the calibration differences between sensors may be up to a factor two. More rigorous calibration could reduce the calibration error.

14. ASOS Bext Threshold: 0.05 km(-1) • The Bext values below 0.05 km-1 are reported as 0.05. • Based on Koschmieder, the ASOS lower detection limit is ~ 50 mile visual range • In the pristine SW US, the ASOS threshold distorts the data • However over the East and Western US, the Bext signal is well over the detection limit

16. Evidence of ASOS Data Problems The ASOS data for Temperature and Dewpoint appear to be erratic for some stations The problems include constant values, spikes and rapid step changes.

17. Data Problems: Bad data records • The main data reading problems are due to bad records • Some records for some stations are not fixed length • Cause of the bad data records need to be identified

18. ASOS Data Pattern Analysis • Diurnal Cycle • RH Dependence of Bext • ASOS Bext – PM2.5 Relationship

19. Typical Diurnal Pattern of Bext, Temperature and Dewpoint • Typically, Bext shows a strong nighttime peak due to high relative humidity. • Most of the increase is due to water absorption by hygroscopic aerosols. At RH >90% , the aerosol is mostly water • At RH < 90%, the Bext is mostly influenced by the dry aerosol content; the RH effect can be corrected. Macon, GA, Jul 24, 2000

20. Diurnal Cycle of Relative Humidity and Bext Relative Humidity The diurnal RH cycle causes the high Bext values in the misty morning hours The shape of the RH-dependence is site (aerosol) dependent – needs work Bext

21. Adopted RH Correction Curve(To be validated for different locations/seasons) RH is calculated fromT – Temperature, deg C and D – Dewpoint, deg C RH = 100*((112-(0.1*T)+D)/(112+(0.9*T)))8 • The ASOS Bext value are filtered for high humidity • Values at RH >= 80% is not used • Later we will try to push the RH correction to 90%) • The Bext is also corrected for RH: RHCorrBext = Bext/RHFactor

22. Seasonal Average Diurnal Bext Pattern • For each minute of the day, the data were averaged over June, July and August, 2000 • Average Bext was calculated for • Raw, as reported • For data with RH < 90% • RH < 90% and RH Corrected • Based on the three values, the role of water can be estimated for each location

23. Location of ASOS and Nearby Hourly PM2.5 Sites • There are no co-located ASOS and PM2.5 sites • The stations are not co-located but in the same city • Hourly PM2.5 data are compared to the filtered and RH-corrected one minute Bext

24. ASOS-Hourly PM2.5 Allentown, PA

25. ASOS-Hourly PM2.5 Des Moines, IO

26. ASOS-Hourly PM2.5 Grand Rapids, MI • In Grand Rapids, MI, July, the relationship is good. • Occasional spikes of Bext are probably weather events not adequately filtered

27. ASOS-Hourly PM2.5 Islip Long Island, NY

28. ASOS-Hourly PM2.5 Toledo, OH

29. ASOS-Hourly PM2.5 San Diego, CA

30. ASOS-Hourly PM2.5 Islip Long Island, NY

31. ASOS Bext – EPA PM2.5 Mass Comparison: Greensborough, NC • The PM2.5 station is in Winston-Salem while the ASOS Bext is at Greensborough. • The ASOS-PM2.5 relationship for Jun, Sep and Oct is good. • Occasional spikes of Bext are not yet explained ???

32. ASOS Bext – EPA PM2.5 Mass Comparison: Pensacola, FL • At Pensacola, FL, Aug, Sep, Oct, the relationship is more noisy. • There are more Bext spikes (to be filtered? Will see) • The Bext/PM2.5 is higher. ASOS calibration problem?

33. ASOS Bext – EPA PM2.5 Mass Comparison: Odessa, TX • At the arid site Odessa, TX, the relationship is poor: • more Bext ‘noise’ • systematically higher Bext/Pm2.5 • The higher Bext will be interesting to explore. • Is it coarse particle dust? • Will se

34. ASOS Bext – EPA PM2.5 Mass Comparison: Albuquerque, NM Lower detection limit: 0.05 • At the pristine Albuquerque, NM site, the PM2.5 values much lower than in the East • The lower detection limit of the ASOS Bext is about 0.05 km-1. • Clearly, over the pristine West, the ASOS will hit the lower detection limit

35. Summary – Tentative Conclusions • There are data for at least 220 Weather Service ASOS stations • Format problems with the data files forced us to discard 30-40 % of readings • To use the ASOS data as PM2.5 surrogate, RH filtering and correction can be applied. These procedures need more calibration. • Comparison of RH-filtered and corrected Bext with hourly PM2.5 at several cities is most encouraging. • ASOS will be a meaningful PM surrogate of PM2.5 concentration estimates with high time and spatial resolution over the Eastern US. For the pristine Southwest, the utility of ASOS is questionable. • The next steps will focus on further comparisons and calibrations.

36. AWOS Reference Guide • 1.2.2 VISIBILITY • Visibility Sensor: TYPE...Belfort 6200/6210 Visibility Meter. PERIPHERAL INTERFACE...VB with 6200, VC with 6210. PI ALGORITHM VERSION...VB1.4 AND VC 1.4. • 1.2.2.1 Sensor Functional Description • The sensor uses the forward scatter principle to measure the atmospheric extinction coefficient and extrapolates this into a visibility. A high intensity xenon strobe transmitter is used to illuminate the sensor's scatter volume. • 1.2.2.2 Accuracy • + / - 10 % of reading. • 1.2.2.3 Output • The unit takes measurements twice per second and reports a one minute average. The PI software then takes an average of ten minutes of extinction coefficient measurements every minute, and translates this into a visibility in statute miles. Reportable values are 0.1 mile steps up to 2.5 miles; 0.5 mile steps to four miles and one mile steps to nine miles. Visibilities over nine miles are reported as 9+. If the average visibility is less than three miles, and if three of the ten values considered in the averaging differ by 0.5 miles or more, then the visibility is reported as variable. These reportable values are later converted to eighths, quarters, halves and whole miles to conform with METAR and SPECI requirements. • 1.2.2.4 Special Criteria • Specials are issued if the visibility drops below, or rises to 0.3, 0.5, 0.8, 1.0, 1.5 or 3.0 miles. A five-minute time delay is built into the algorithm for the case of improving visibilities in an effort to reduce the number of specials in variable conditions. • 1.2.2.5 Limitations Of The AWOS Visibility Sensor • Overly sensitive to ice crystals (site specific); • This problem has proven to be very site specific, suggesting that the problem is related to the siting of the instrument. • Obstructions to vision (e.g. Fog) are not reported; • AWOS reports only the visibility at the sensor, not a prevailing visibility or off-site (sector) visibility.

37. ASOS STATUS REPORT • ASOS All 569 FAA-sponsored and 313 NWS Automated Surface Observing System sites are commissioned. For a state-by-state listing of commissioned sites, see "Map of Automated Weather Observing Sites and Current Weather." For your convenience, we have included the ground-to-air frequencies and telephone numbers of these systems. The ASOS systems at each site operate to comply with the "Weather Observation Service Standards." A Next Generation Runway Visual Range (NGRVR) interface to ASOS was developed and has completed the Operational Test & Evaluation testing. There are currently 120 RVR long line sites planned with 72 sites operational. • ASOS "TEST" MESSAGE The following information is being promulgated to clear up some possible misunderstanding regarding the ASOS "Test" message: After an ASOS system is installed at a site, it undergoes a testing and evaluation period to ensure that the system is operating properly before it is commissioned as an operational system. During this pre-commissioning period, the ASOS generated weather message includes the word "TEST". Although the weather data in the test message may be accurate, it does not undergo the same level of quality assurance checks as a commissioned site does, and therefore should not be used operationally. Once the site is commissioned, the test message is removed and the ASOS becomes fully operational.

39. AWOS Block Diagram

40. Belfort Visibility Sensor • Visibility Sensor: Model 6230A • COMPACT, MODULAR DESIGNEASY INSTALLATION & MAINTENANCEVISIBLE LIGHT FOR GREATER ACCURACYWIDE DYNAMIC RANGEDIGITAL AND ANALOG OUTPUTSEXTENSIVE SELF-DIAGNOSTICSNATIONAL WEATHER SERVICE APPROVED • Specifications | Standard Features | Options | PDF • The Belfort Model 6230A Visibility Sensor is a highly reliable, compact and rugged atmospheric visibility sensor. It combines new digital signal processing with the proven reliability of the previous analog model. • The sensor can be used to replace existing transmissometers, while providing better stability, accuracy, and maintainability, at a much lower price. Applications include visibility measurements at airports, lighthouses, along highways, and on ships, as well as for meteorological observations. The sensor can also be used in the detection of hazards to transportation caused by fog, snow, smoke, blowing sand, etc. • The Model 6230A uses the Forward Scatter Principle to measure the atmospheric extinction coefficient and visibility. A high intensity Xenon strobe transmitter is used to illuminate the sensor's scatter volume. This results in a high signal-to-noise ratio and reduces the effects of natural background light variations. • Accurate visibility measurements are possible over a standard range of 0.01 to more than 30 miles, the widest dynamic range available. The use of light within the visible spectrum allows the sensor to most accurately simulate human perception of visibility. • Field calibration is simple and can even be accomplished during periods of impaired visibility. • Digital and optional analog outputs simplify integration of the sensor into networks or systems. • The Belfort Visibility Sensor has been extensively tested and widely accepted by the World Meteorological Organization, the British Meteorological Office, the U.S. National Weather Service and the Canadian Atmospheric Environment Service. The sensor is being installed as part of the National Weather Service's Automated Surface Observing System (ASOS) Program. • Specifications • Range: 17 ft to 30 miles (5m to 50km)Accuracy: ±10% of readingScatter Angle: 20° to 50°Illumination Source: Xenon flash lampSample Volume: .75 Ft(.021m)Time Constant: 1 sec. (selectable)Flash Rate: 2HzOutput: RS232C @2400 BaudEnvironmental:With Low Temp. Opt: -67° to +140° F -55° to +60° C)Without Low Temp: -4° to +140° F (-20° to +60° C)MTBF: 10 years (MIL-217E)EMI: MIL-STD-461Power Requirements: 20 watts at 24 Vdc. With Low Temperature OptionPerformance Package: 150 watts at 24Vdc.Dimensions:Sensor: 19"Hx64"Wx25"D (48.3x162.6x63.5cm)Support: 10.0 Ft. (3 m)Weight:Sensor: 30 lbs(13.6kg)Support: 34 lbs(15.4kg) • Standard Features • Uses visible light - proven the most reliable source for accurately simulating human perception of visibilityDigital output, or combination digital and analog output, will meet most interface requirementsSingle pedestal mounting for easy, inexpensive installationModular construction for rapid, easy field replacement of partsSelf-diagnostics assure operator of correct performance at all timesLarge scattering volume exceeding 0.75 ft (0.21m)24 VDC or 12 VDC Operating Voltage for efficient operation and safety protection to personnel during servicing