U. S. Department of Transportation
Download 2.66 Mb.
|
RadarThese variables are derived from the ATQA OE dataset. Correspondingly, they can only be analyzed for OE incidents. They describe the radar systems available to the controller at the airport where the incident took place. In some instances the radar variables have been combined to cover multiple similar versions of a system. Where this occurs, the specific systems included will be noted. Brief definitions of the different radar systems examined follow:49
STARS(ATQA OE) Table 105 and Table 106 present the observed and expected distributions of STARS by severity. Fisher’s exact test indicates that there is a relationship between severity and the availability of the STARS radar system. Categories A, B and D appear underrepresented while Category C is over represented. Table – Observed Distribution of STARS by Severity
Table – Expected Distribution of STARS by Severity
To better understand how this variable impacts severity, category D incursions are excluded from the following tables (Table 107 and Table 108). This eliminates the conflict versus non-conflict dynamic that appears in Table 105. As with the entire range of severity, categories A and B are underrepresented while category C is overrepresented. However, the relationship between STARS and severity is weaker. This indicates that some of the relationship seen in Table 105 can be attributed to discriminating between conflict and non-conflict. This may be a product of where STARS is deployed; that is, STARS may be deployed where the baseline rate for conflict events is higher regardless of its impact on severity. Nevertheless there appears to be weak evidence suggesting that the presence of STARS is associated with lower severity incidents. Table – Observed Distribution of STARS by Severity, Conflict Only
Table – Expected Distribution of STARS by Severity, Conflict Only
ASDE(ATQA OE) It is important to acknowledge that this ASDE variable does not discriminate between different versions of the ASDE system. That is, this variable indicates the presence of ASDE-3 or ASDE-X. This is due to how the information was ended in the Runway Incursion Database. Regardless, Table 109 and Table 110 present the distribution of this variable. Interestingly, there appears to be a strong relationship between severity and the presence of ASDE. Categories A, B, and D are underrepresented while category C is overrepresented. This is likely a product of how the ASDE systems were deployed. ASDE is deployed at major airports, where a non-conflict event (category D) is less likely. Therefore, it is instructive to look at the conflict only distribution as presented in Table 111 and Table 112. Table – Observed Distribution of ASDE by Severity
Table – Expected Distribution of ASDE by Severity
The conflict only distribution indicates a similar pattern to the overall distribution. There is some evidence that ASDE is associated with lower severity events (in this case, category C incursions). This indicates that the lower than expected number of D incursions seen in Table 109 is likely a product of the distribution of ASDE systems with respect to airports. Table – Observed Distribution of ASDE by Severity, Conflict Only
Table – Expected Distribution of ASDE by Severity, Conflict Only
Given that both ASDE and STARS appear to reduce the severity of runway incursions, it would be interesting to investigate whether or not there is any synergy between STARS and ASDE. The logit results presented in Table 113 indicate that STARS and ASDE are both associated with lower severity incidents, but there is no synergy between the systems. That is, the effect of STARS and ASDE is exactly the sum of its parts. Note, however, that the odds ratios for STARS and ASDE in isolation are not precisely estimated; this is likely a product of including the interaction term in the estimation. Though the evidence for the isolated impact of ASDE or STARS is weaker in this logit model, combining these results with those from the Fisher’s Exact test indicates that there is evidence that these radar systems reduce severity. Table – Logit Estimates of Impact on Severity, ASDE and STARS
ARTS II(ATQA OE) As mentioned previously, ARTS II represents a lower cost version of the ARTS III system. This variable indicates if any version of ARTS II was available to the controller at the time of the incident. Table 114 and Table 115 present the observed and expected distribution. There is no indication of any relationship between the presents of ARTS II and severity. As the ARTS systems are focused on airborne traffic, this is not an unexpected result. Table – Observed Distribution of ARTS II by Severity
Table – Expected Distribution of ARTS II by Severity
ARTS III(ATQA OE) The ARTS III system is the more feature rich and expensive version of the ARTS systems under consideration. This variable indicates whether ARTS III was available to the controller at the time of the incident. Table 116 and Table 117 present the observed and expected distribution of this variable. Table – Observed Distribution of ARTS III by Severity
Table – Expected Distribution of ARTS III by Severity
Categories A and C appear over represented while categories B and D appear under represented. It is important to reiterate that the Fisher’s Exact test indicates that there is some relationship between the two variables (i.e., there is systematic relationship between rows and columns in the table). It does not test for any particular direction or even if that relationship is consistent. For a better understanding of how ARTS III may impact severity, category D incursions can be excluded, removing the conflict versus non-conflict dynamic. Table 118 and Table 119 examine ARTS III in terms of conflict events only. The relationship seen in Table 116 is no longer present. As with ASDE it is possible that this relationship is due to how ARTS III is deployed – busier airports received the expensive ARTS III system. Table – Observed Distribution of ARTS III by Severity, Conflict Only
Table – Expected Distribution of ARTS III by Severity, Conflict Only
|