Runway visual range (RVR) systems measure or estimate three parameters [extinction coefficient (sigma), background luminance (B) and runway light intensity (I)] and calculate the distance a pilot can see down a runway using Koschmieder's Law (objects as targets) or Allard's Law (lights as targets). The errors in RVR resulting from errors in the three parameters were analyzed in a recent paper. Extinction coefficient errors translate roughly into equivalent percentage errors in RVR. RVR is less sensitive to errors in B and I; a factor of two (four) error translates into an RVR error of roughly 10 (20) percent.
RVR values are reported to pilots in relatively coarse increments (e.g. 100 feet or 25 meters). RVR accuracy requirements are typically derived from two considerations: (1) the achievable accuracy of the best extinction coefficient instruments (the two other measurements are ignored) or (2) the resolution of the reporting increments. This paper proposes that the instrumental accuracy requirements instead be based on the principle that instrumental errors be smaller than the other sources of error contributing to variance in the RVR estimate. This paper will show that this principle results in instrument accuracy requirements that are more relaxed than those derived from the two prior considerations. Requiring RVR systems to meet unrealistic accuracy criteria will result in more costly systems.
The reported RVR value is intended to represent how far a pilot can see down a runway. Errors in these values are generated by a large number of factors, such as:
For both Koschmieder's and Allard's Laws:
a) Variations in the pilot's eyesight,
b) Variations in aircraft cockpits,
c) Spatial variations in the obstruction to vision between the pilot's view and the location where the extinction coefficient is measured,
d) Measurement errors of the extinction coefficient (sigma) sensor,
For Koschmieder's Law:
e) Non ideal visibility targets,
For Allard's Law:
f) Angular and temporal variations in runway light intensity,
g) Errors in setting and/or measuring runway light intensity (I),
h) Differences in background luminance between the pilot's view and the direction where the background luminance is measured,
i) Errors in measuring background luminance (B), and
j) Errors in relating illumination threshold to background luminance.
Of all these errors, only (d), (g) and (i) pertain directly to the performance of an automated RVR system. In general, the design goal for an RVR system is to assure that the measurement errors are smaller than the other sources of error.
This paper will use measurements of spatial variations, error (c), to suggest practical accuracy requirements for extinction coefficient measurements. Instruments are typically installed 100 m from the runway centerline and are used to estimate the RVR for roughly ±1000 m along the runway. Spatial variations in extinction coefficient will be taken from three test sites using two methods: (1) direct comparison of two instruments at different locations and (2) comparisons of readings from the same instrument taken at different times, using the ambient wind to convert time displacements into distance displacements.
The 8th Conference on Aviation, Range, and Aerospace Meteorology