Two primary automated observing systems are being deployed at airports across the United States, the Automated Surface Observing System (ASOS) and the Automated Weather Observation System (AWOS). Both systems are a collection of electronic sensors, connected to a computer, that measure, process, and create surface observations every minute. These systems provide 1-minute, 5- minute, Hourly, and Special observations 24 hours a day.
This discussion explains the need for automated observations and highlights the differences between automated and human observations. It will provide keys to understanding and applying the automated observing technology to daily operational decisions.
Modern technology can now provide a near real-time measurement of the atmosphere. The National Weather Service WSR-88D Doppler weather radar routinely updates its data sets every 5 to 6 minutes. Weather satellites provide new information every 15 to 30 minutes. Yet the routine surface observation taken by weather observers at airports across the nation remains a single hourly observation.
When the weather is quickly changing, the human observer can send only a few extra observations highlighting those changes. Only by automating the surface observation can we improve the timeliness of the surface weather information and provide data that can be integrated more effectively with the products of other systems.
Automation also removes the burden on individuals to create observations. Many small airports depend on the airport operators to provide observations. These operators often must make a special trip to the airport to take an observation for an incoming flight.
The National Weather Service also has personnel dedicated to observing the weather. These trained personnel can be better used to assist the forecast and warning program and provide more effective services to the public. Automated systems can provide observations 24 hours a day, freeing human observers for other tasks. There are many airports where hiring human observers is too costly or impractical. Automating many of these sites will increase the number of 24-hour airports available to pilots, increasing local business and adding safety.
Siting automated sensors is critical to providing an observation that is representative of the entire airport area. The favored location is near the touchdown zone of the main instrument runway. If that location does not consistently provide representative weather observations for most of the runway complex, an alternate site can be selected. At airports with great weather variations due to nearby rivers, lakes, oceans or terrain, additional sensors can be installed to provide a more accurate observation.
Human observers rely on their ability to see the entire atmosphere, horizon-to-horizon, to complete a weather observation. Any site that limits that viewing ability detracts from the accuracy of the observation. Often buildings or terrain will block the observer's view of the runway complex or limit the observer's ability to identify weather conditions away from the airport. At night, nearby lights frequently impede the observer's ability to see clearly, preventing an accurate assessment of certain weather elements. Lights, buildings or human perception do not affect automated systems. They are designed to create a representative observation from weather passing through the sensors 24 hours a day.
Automated systems measure only the weather that passes through the sensor array. A set of data is collected over time to provide a "representative" observation. The system applies mathematical logic, called algorithms, to the collected data to extrapolate the weather over a wider area. The table on this page provides a generalized summation of the area around an ASOS site where the data of the sensors are considered valid.
SAMPLE AREA BROADENING DUE TO ALGORITHMS PARAMETER PROCESSING RADIUS INTERVAL VALIDITY (MINUTES) (MILES) SKY CONDITIONS 30 3-5 VISIBILITY 10 2-3 PRECIPITATION 10 1-2 FREEZING RAIN 15 2-3 TEMP/DEW POINT 5 5 WIND 2 1-2 PRESSURE 1 5
Each of the various weather elements requires specific sample times for its algorithms. These times were selected to represent normal meteorological variations. For example, it was determined that 30 minutes of data provide an accurate description of sky conditions. This means that algorithms will process all the clouds passing over the sensor in the past 30 minutes to determine the height and amount of clouds transmitted in the observation. The last 10 minutes of data is "double weighted", counted twice, to enhance the response to the latest conditions. Each minute the data area is processed to provide up-to-date observations.
One major advantage of an automated observing system is consistency. Unlike human observers, all similarly configured automated systems measure the atmosphere the same way. ASOS is programmed to add remarks to the aviation observation to highlight significant changes or report important data. ASOS will create only a standard set of remarks concerning variable ceilings, visibility, precipitation, pressure changes and wind gusts. If the remarks are not included, sensors have not measured the condition.
By contrast, human observations are fraught with opportunities to provide inconsistent information. Observers at the same location will give different estimates of cloud heights and coverage. Human observers report mandatory remarks and also are allowed to provide "optional" information. Many of these "optional" remarks actually contribute to in-consistent observations.
One observer may ignore some off-airport weather that another observer deems important. Especially at night there are considerably fewer remarks concerning off-airport weather. A pilot cannot always assume that when a remark is not in the observation the condition is not occurring. Thus, human observers cannot provide site-to-site, day-to-night consistency.
A human observer may also be distracted by duties associated with critical weather elsewhere and not notice rapid changes in the airport weather. For example, if rain begins to fall when an observer is distracted, the observer can only estimate the actual beginning time. ASOS tirelessly and continuously measures the minute-by-minute weather and is never distracted by other duties.
Even though ASOS creates a completely new observation every minute, automated systems must have adequate sensor samples to develop an accurate observation. In rapidly changing conditions, the automated observations are known to lag slightly behind the actual weather. For instance, if skies are clear and a sudden overcast appears on the sensors, ASOS will take 2 minutes to report a scattered deck of clouds. Within 10 minutes, the system will indicate a broken layer.
Each minute ASOS processes the most recent 10 minutes of visibility sensor data to obtain a representative value. Therefore, when visibility drops suddenly (in one minute) from 7 miles to 1 mile, ASOS needs about 4 minutes before the 10-minute mean values reach the 3 mile criteria. This criteria forces a Special observation to alert users to a significant change. A total of 9 minutes will pass before ASOS will report the 1 mile visibility.
When the visibility rapidly improves from 1 mile to 7 miles, ASOS generates a Special observation 4 minutes after reaching the 1.5 mile threshold. In 10-12 minutes, ASOS will report 7 miles. Why longer to improve? The algorithm is designed to raise visibility more slowly than to lower it. This design provides a margin of safety and buffers rapid changes when the visibility is widely fluctuating over a short period.
Hourly and Special observations, which denote a significant change in the weather, are the only ones created by human observers. In contrast, ASOS relentlessly measures the weather and can provide 1-minute, 5-minute, Hourly and Special observations. Because of ASOS' continuous weather measurements, the system will create more frequent Special observations than a human observer.
At towered sites, controllers, dispatchers, and large numbers of pilots rely on the weather observations to regulate flight operations. Unfortunately, too many or too frequent observations causes difficulty in making decisions (information overload). Most towered locations also have human observers to monitor and edit automated observations. The observer requires some time to evaluate the observations and, if necessary, edit the observations before transmission.
Thus, the system has a 5-minute time buffer to limit the number of observations that can be transmitted during an hour. The 5- minute time buffer, however, has created an unwarranted reputation of slow system response to fast weather changes. This slower response is seen mainly at towered airports where only the Hourly and Special observations are broadcast on the Airport Terminal Information System (ATIS) or on the national weather data circuits. These circuits also transmit only Hourly and Special observations. It is only at non-towered airports that pilots receive the 1- minute weather by calling the voice phone link or by the ground- to-air radio broadcasts. Users have not noted system responsiveness as a serious problem at non-towered locations.
Pilots must listen to a sequence of automated observations and not rely upon a single observation. A pilot cannot determine trends or changes in weather from a single observation. For example, if several 1-minute observations indicate unchanging weather conditions, the pilot can anticipate unchanging conditions in the vicinity of the airport.
If the indicated weather is deteriorating, a pilot should expect poorer conditions as they near the runway. When weather at the airport is rapidly changing, pilots should not expect an observation 5 to 10 minutes old to depict accurately the weather over the entire area. By monitoring the minute-by-minute observations, a pilot can determine a range of variations to be expected on the approach to a landing.
All observations, whether automated or taken by human observers, should be used with care. Users must be aware of how long ago the observation was taken, under what conditions, and whether there are Special observations. Even though automated systems are totally objective and maintain a certain uniformity among all sites, it does not mean they match perfectly the varied perception of users.
ASOS may occasionally report cloud decks lower than what is actually encountered. Sometimes precipitation, lower cloud fragments or fog triggers these lower values. Pilots have said that these "lower" reported values often indicated the height below which they had to fly before gaining enough forward visibility to see an airport and land. The key lesson here is to evaluate all reports closely before dismissing them as inaccurate.
Even though the visibility sensor is designed to objectively represent the visibility of the atmosphere over a wide range of weather conditions, day or night, it occasionally reports a visibility more optimistic than what a human perceives. During the day, the human eye can be overwhelmed by bright light reflected in clouds, light precipitation, fog or haze. Many individuals will even resort to wearing sunglasses to obtain some relief from the glare.
The ASOS visibility sensor is not as sensitive to this condition and sometimes reports a visibility approximately twice as high as what an individual may determine. Be alert for these bright conditions and expect a more optimistic value from the automated system. You can find more detail on evaluating visibility and sky conditions in other editions of Tool Box.
Automated systems will evaluate and report only weather that has passed through the sampling volume of the sensor array. Therefore, weather around the airport that has not encountered the sensors will not be measured. The system is not designed to report clouds above 12,000 feet, virga, tornadoes, funnel clouds, ice crystals, snow pellets, ice pellets, drizzle, freezing drizzle, blowing obstructions such as snow, dust, or sand, snow fall and snow depth. Many of these elements will be provided by other sources. New sensors will be added to measure some of these weather elements.
Automated observations are not identical to observations taken by humans observers. Automated observing systems are designed to provide consistent observations, day or night, representing the weather within 2 to 3 miles of the sensor site. They will not provide a horizon-to-horizon evaluation of the weather. Finally, automated observing systems are in their infancy. New developments in sensors and computer technology will improve their capabilities.
Bradley, J. T. and Imbembo, S.M., "1985: Automated Visibility Measurements for Airports," American Institute of Aeronautics and Astronautics: Preprints, Reno, NV
Bradley, J. T. and Lewis, Richard, "1993: Representativeness and Responsiveness of Automated Weather Systems," Fifth International Conference on Aviation Weather Systems, AMS, Vienna, VA, pp 163-167
Clark, P., "1995: Automated Surface Observations, New Challenges, New Tools," Sixth Conference on Aviation Weather Systems: Preprints, AMS, Dallas, TX
U.S. Dept. of Commerce, NOAA, 1992: ASOS Users Guide, Government Printing Office
To: ASOS HOME PAGE
Page last updated 10/99 by firstname.lastname@example.org