Crisp and Fuzzy Signal Detection Theory and Pilot Weather Judgment: Implications for VFR Flights Into IMC

Abstract

Weather represents one of the greatest hazards to general aviation (GA), accounting for 15% of the GA accident fatalities. Of the fatal weather accidents 90% are attributed to visual flight rules (VFR) flight into instrument meteorological conditions (IMC). The situation assessment hypothesis suggests that pilots may inadvertently enter IMC because they lack the sensitivity needed to distinguish between visual meteorological conditions (VMC) and IMC. An alternative hypothesis is that pilots recognize conditions have deteriorated but are motivated by some other factor, such as pressure from passengers. The present study uses Jensen's Pilot Judgment Model and Signal Detection Theory to explain pilot judgment. The impact of Graphical Weather Information Systems (GWIS), particularly graphical METARs on pilot judgment was also assessed. Twenty-four general aviation pilots were shown simulated video images of different weather conditions. Several of these trials contained GWIS surface data of varying accuracy. Results indicated that pilots had trouble in distinguishing between VFR and IFR conditions, especially determining ceiling. Overall, pilots had a low sensitivity in determining whether the ceiling was VMC or IMC and tended to overestimate ceilings. This problem was amplified by an interaction with visibility. Pilots' estimates of IMC ceilings actually increased as the visibility increased. A similar effect of ceiling influencing visibility judgments was also found. Pilots' judgments were additionally influenced by the inaccurate METAR information presented in the GWIS. GWIS data that suggested conditions were worse than those seen out-the-window caused a liberal shift in response bias, and conditions that were better than those out-the-window had a corresponding conservative shift in response bias. Overall the experiment found evidence to suggest both situation assessment and motivation could contribute to a decision to continue into IMC. The interaction of ceiling and visibility also suggests a new potential factor in inadvertent VFR flight into IMC. The improper evaluation of one weather dimension based upon a bias from the other weather dimension needs to be further examined for its role in pilots' decision to continue into deteriorating weather conditions.