93 



implications? 



The Federal Aviation Administration, its forerunners, and the National Weather Service, 

 and its forerunners employ professionally trained staff whose duty it is to report visibility at 

 civil airports. The purpose is obviously to maintain aviation safety. 



There are four airports that surround the Shenandoah Park (Figure 1): Elkins, West 

 Virginia, Dulles International, Lynchburg Regional (also known as Preston Glen Field) and 

 Richmond Byrd International. Together, these airports have been served by well over 100 

 uained visibility observers in the last three decades. 



Airport data often differ in mean visibility because of different characteristics surrounding 

 each site. An airport located on a flat plain, for example, contains few horizon markers for 

 an observer to base readings upon, but a mountainous one does. The result is that different 

 airports show different average visibility, but the fluctuations from month to month or from 

 year to year are in unison. 



One measure of the degree to which the airports behave simultaneously is the mathematical 

 correlation between them. A certain calculation reveals whether the data are independent of 

 each other or not. In the case of the data for these four airports, the correlations are highly 

 significant in the statistical sense. This means that in a month when visibility is reduced at, 

 say, Dulles Airport, it also is reduced at the others. 



Because these airports surround the Park, and because the data are all significantly 

 correlated between airports, it is clear that they must represent also a location that is 

 mutually between all of them, which is the Shenandoah National Park. 



We analyzed over 950,000 separate observations of visibility since 1959, creating mean 

 monthly values over thirty years. The results, supplied Figure 2, show mean noon 

 visibility trends for each individual month. 



There are no overall trends whatsoever in any of the monthly visibility records, either for 

 an individual airport or in aggregate. However, there is a downward trend in July and 

 August visibility ending in 1969 (July) and 1973 (August). Since 1969 there are no 

 statistically significant downward trends in any of the data. Five of the airport-month 

 combinations show significant increases in visibility since then. 



Perhaps most interesting is the fact that, in those two cases in which there is a trend, all 

 airports show it simultaneously. Thus the subsequent data, with no declining trend in any 

 month, also appear reliable. 



