4. Data Acquisition, Response Time, Calibration

The real-time communication system used by the airline community is described in Fleming (1996). WVSS data is acquired at a rate of 4 Hertz on the aircraft and the four samples are averaged to one-second values. This water vapor information is mixing ratio (r), which is converted from the measured RH values by Eq. (10) and continually broadcast to the digital flight data acquisition unit (DFDAU) on the aircraft along with time, latitude, longitude, winds, and temperature. This meteorological data stream is routinely sent in real-time to the user community via the Aircraft Communication Addressing and Reporting System (ACARS) hardware on the aircraft. The format and frequency are defined in the Aeronautical Radio, Inc. (ARINC)-620 document and vary according to the stage of flight: ascent, descent or en route. Details of these formats can be found in Fleming (1996), but the typical rates used by the UPS aircraft are:
ascent:        observations every 6 seconds (approximately 100-meter-vertical
                  resolution) until 90 seconds into the flight, then every 20 records
                  until an altitude of 20,000 feet
descent      observations every 30 seconds from approximately 20,000 feet to
                  the surface
en route     observations every 3 minutes
These rates can be adjusted up or down via a real-time ACARS message.

The original FAA requirement focused on ascent/descent information only and the response time specifications were modestly set to entice greater creativity in the approach to measurements of water vapor on commercial aircraft and to encourage several companies to bid in the competitive procurement process. These specifications were a response time of <6 s to 90% full-scale response for heights less than 5000 feet and <20 s for heights less than 20 000 feet. Results from early testing (Hills and Fleming, 1994) suggested mean values of 4 s and 12 s for corresponding altitudes for thin-film sensors. Typically radiosonde response times are about 1 s at 0°C to 10 s at -20°C to 100 s at -43°C (Salasmaa and Kostamo, 1975). In the upper troposphere, where temperatures are often as cold as -60°C, these response times must be much longer.

At flight level, a major advantage of the thin-film technology on the aircraft, as opposed to its use on the radiosonde, is the dynamically heated environment in which the measurement is made. As an example, using Eq. (2) and values of Mach number = 0.8, Ts = -60°C (213.15K) the temperature in the probe is 240.4K - greater than 27 degrees of warming. Thus, one can estimate response times to be in the range of 40 - 60 s and thus perhaps 4 - 6 times faster than for the radiosonde at this same cold static (ambient) temperature.

Two other characteristics of measurement systems that are useful to know are the sensitivity of the sensor (e.g., the lowest value of RH that can be measured) and the precision. The LMC WVSS sensor system is quite sensitive at the low RH range and is considered valid down to RH = 0.4%. The lowest RH value recorded in the probe (in an early sample of 12 000 reports) was 1.02% and the lowest value of RH static recorded (over all temperatures in the same sample) was 1.75%. [Note that this WVSS has a precision of 0.01% and this is why the RH values in this report are indicated to two digits beyond the decimal point - accuracy alone would not warrant this.]

The accuracy of the WVSS will depend on several factors. One of these is initial calibration and how well that calibration is maintained over time.

Initial calibration of the WVSS is performed by BFG. This is performed by standard laboratory procedures at RH values near 0% (using dry nitrogen) and 70% (using a chilled mirror reference traceable to NIST). The HMM30D provides a voltage from 0 to 5 volts that represents RH from 0 to 100%. The HMM30D signal conditioner is affected by ambient temperature. Therefore, BFG added circuitry to the electronics to compensate for output changes due to temperature. This typically reduced the error by a factor of five so the errors are less than +/- 0.5% RH over the temperature range likely to be encountered in the probe. Tests by LMC of the WVSS at BFG and at the National Center for Atmospheric Research (NCAR) indicated that the WVSS could be calibrated to within 1% for the range tested.