National Academies Press: OpenBook

Analysis of Aircraft Overruns and Undershoots for Runway Safety Areas (2008)

Chapter: Chapter 3 - Findings and Applications

« Previous: Chapter 2 - Research Approach
Page 20
Suggested Citation:"Chapter 3 - Findings and Applications." National Academies of Sciences, Engineering, and Medicine. 2008. Analysis of Aircraft Overruns and Undershoots for Runway Safety Areas. Washington, DC: The National Academies Press. doi: 10.17226/14137.
×
Page 20
Page 21
Suggested Citation:"Chapter 3 - Findings and Applications." National Academies of Sciences, Engineering, and Medicine. 2008. Analysis of Aircraft Overruns and Undershoots for Runway Safety Areas. Washington, DC: The National Academies Press. doi: 10.17226/14137.
×
Page 21
Page 22
Suggested Citation:"Chapter 3 - Findings and Applications." National Academies of Sciences, Engineering, and Medicine. 2008. Analysis of Aircraft Overruns and Undershoots for Runway Safety Areas. Washington, DC: The National Academies Press. doi: 10.17226/14137.
×
Page 22
Page 23
Suggested Citation:"Chapter 3 - Findings and Applications." National Academies of Sciences, Engineering, and Medicine. 2008. Analysis of Aircraft Overruns and Undershoots for Runway Safety Areas. Washington, DC: The National Academies Press. doi: 10.17226/14137.
×
Page 23
Page 24
Suggested Citation:"Chapter 3 - Findings and Applications." National Academies of Sciences, Engineering, and Medicine. 2008. Analysis of Aircraft Overruns and Undershoots for Runway Safety Areas. Washington, DC: The National Academies Press. doi: 10.17226/14137.
×
Page 24
Page 25
Suggested Citation:"Chapter 3 - Findings and Applications." National Academies of Sciences, Engineering, and Medicine. 2008. Analysis of Aircraft Overruns and Undershoots for Runway Safety Areas. Washington, DC: The National Academies Press. doi: 10.17226/14137.
×
Page 25
Page 26
Suggested Citation:"Chapter 3 - Findings and Applications." National Academies of Sciences, Engineering, and Medicine. 2008. Analysis of Aircraft Overruns and Undershoots for Runway Safety Areas. Washington, DC: The National Academies Press. doi: 10.17226/14137.
×
Page 26
Page 27
Suggested Citation:"Chapter 3 - Findings and Applications." National Academies of Sciences, Engineering, and Medicine. 2008. Analysis of Aircraft Overruns and Undershoots for Runway Safety Areas. Washington, DC: The National Academies Press. doi: 10.17226/14137.
×
Page 27
Page 28
Suggested Citation:"Chapter 3 - Findings and Applications." National Academies of Sciences, Engineering, and Medicine. 2008. Analysis of Aircraft Overruns and Undershoots for Runway Safety Areas. Washington, DC: The National Academies Press. doi: 10.17226/14137.
×
Page 28
Page 29
Suggested Citation:"Chapter 3 - Findings and Applications." National Academies of Sciences, Engineering, and Medicine. 2008. Analysis of Aircraft Overruns and Undershoots for Runway Safety Areas. Washington, DC: The National Academies Press. doi: 10.17226/14137.
×
Page 29
Page 30
Suggested Citation:"Chapter 3 - Findings and Applications." National Academies of Sciences, Engineering, and Medicine. 2008. Analysis of Aircraft Overruns and Undershoots for Runway Safety Areas. Washington, DC: The National Academies Press. doi: 10.17226/14137.
×
Page 30
Page 31
Suggested Citation:"Chapter 3 - Findings and Applications." National Academies of Sciences, Engineering, and Medicine. 2008. Analysis of Aircraft Overruns and Undershoots for Runway Safety Areas. Washington, DC: The National Academies Press. doi: 10.17226/14137.
×
Page 31
Page 32
Suggested Citation:"Chapter 3 - Findings and Applications." National Academies of Sciences, Engineering, and Medicine. 2008. Analysis of Aircraft Overruns and Undershoots for Runway Safety Areas. Washington, DC: The National Academies Press. doi: 10.17226/14137.
×
Page 32
Page 33
Suggested Citation:"Chapter 3 - Findings and Applications." National Academies of Sciences, Engineering, and Medicine. 2008. Analysis of Aircraft Overruns and Undershoots for Runway Safety Areas. Washington, DC: The National Academies Press. doi: 10.17226/14137.
×
Page 33
Page 34
Suggested Citation:"Chapter 3 - Findings and Applications." National Academies of Sciences, Engineering, and Medicine. 2008. Analysis of Aircraft Overruns and Undershoots for Runway Safety Areas. Washington, DC: The National Academies Press. doi: 10.17226/14137.
×
Page 34
Page 35
Suggested Citation:"Chapter 3 - Findings and Applications." National Academies of Sciences, Engineering, and Medicine. 2008. Analysis of Aircraft Overruns and Undershoots for Runway Safety Areas. Washington, DC: The National Academies Press. doi: 10.17226/14137.
×
Page 35
Page 36
Suggested Citation:"Chapter 3 - Findings and Applications." National Academies of Sciences, Engineering, and Medicine. 2008. Analysis of Aircraft Overruns and Undershoots for Runway Safety Areas. Washington, DC: The National Academies Press. doi: 10.17226/14137.
×
Page 36
Page 37
Suggested Citation:"Chapter 3 - Findings and Applications." National Academies of Sciences, Engineering, and Medicine. 2008. Analysis of Aircraft Overruns and Undershoots for Runway Safety Areas. Washington, DC: The National Academies Press. doi: 10.17226/14137.
×
Page 37
Page 38
Suggested Citation:"Chapter 3 - Findings and Applications." National Academies of Sciences, Engineering, and Medicine. 2008. Analysis of Aircraft Overruns and Undershoots for Runway Safety Areas. Washington, DC: The National Academies Press. doi: 10.17226/14137.
×
Page 38

Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

20 Database Summary Statistics This section provides a statistical description of the data gathered in this study and included in the ACRP 4-01 acci- dent and incident database. The final database incorporates 459 accidents and incidents. Figure 8 depicts the distribution of events by type of accident and incident (LDOR, LDUS, or TOOR). Proportionally, there have been more LDOR than LDUS and TOOR, as shown in Figure 9. The numbers of LDUS and TOOR are similar. Although some events prior to 1982 were included in the study, the majority of cases date from 1982 to 2006. The dis- tribution of incidents and accidents is variable along the period data was collected. However, in the average (from 1984 to 2004), the number of reported accidents and inci- dents was similar and averages 18 events per year (9 accidents and 9 incidents). Part of the reduction observed for 2005 to 2007 is due to the unavailability of the reports when the data for this study were collected. For many events during this period, either the reports had not been completed or they were not yet available in electronic format. Figure 10 sum- marizes the number of events per year. Summary of Anomalies Associated with Accidents and Incidents An FHA was conducted during the initial stages of this study to identify the relevant factors associated with aircraft overrun and undershoot events so that data on these param- eters could be gathered and included in the accident database and be used for developing risk models. The majority of investigation reports describe causal and contributing factors to accidents but in general these are not reported for incidents. In addition, certain factors not de- scribed as causal or contributing factors in the accident are relevant to the present study. For example, a runway overrun investigation report may describe high approach speed and long touchdown as causal factors, but the report described the runway surface as wet. Although the latter was not con- sidered a relevant factor in the investigation, the anomaly was present and is included in the summary statistics that follow. The anomalies were divided into six different categories to aid in understanding the factors leading to aircraft overrun and undershoot events: • Aircraft System Fault (SysF); • Wildlife Hazards (WH); • Weather Conditions (W); • Human Errors (H); • Runway Surface Conditions (R); and • Approach/Takeoff Procedures (AT). Several anomalies within each of these categories may be present during accidents and incidents. The majority of these anomalies were taken from the list of causal and contributing factors described in the investigation reports. In a few cases, even when not listed in the report, if an additional anomaly was identified, it was included in this analysis. For example, some investigation reports did not describe the wet runway as a causal or contributing factor to the accident, but rain during touch down on the runway was listed, and wet runway was included in the analysis. The complete list of anomalies within each of the above categories and used in this study is shown in Table 4. Figure 11 depicts the distribution by category for landing overruns. In this case, anomalies are mostly related to weather, human error, runway conditions, and approach procedures. Figure 12 shows the frequency of anomalies by the category for undershoots. Similarly to landing overruns, the anomalies are mostly related to weather, human error, runway condi- tions, and approach procedures. Except for runway conditions, the incidence of anomalies was higher for the accidents under the predominant categories. C H A P T E R 3 Findings and Applications

21 Accidents/Incidents by Type 0 20 40 60 80 100 120 140 160 180 Type of Event # of E ve nt s Accidents 121 48 55 Incidents 153 45 37 LDOR LDUS TOOR Figure 8. Distribution of events by type. Events by Type LDOR 60% LDUS 20% TOOR 20% Figure 9. Distribution by type of event. A summary of anomalies by category for takeoff overrun events is presented in Figure 13. For most of the events there were anomalies in the takeoff procedures. When there were anomalies related to weather conditions there were signifi- cantly more events in the accident category than incidents. The same conclusion is generally true for human errors, while system faults are mostly related to incidents. Anomalies reported or identified were included within each of these categories, as shown in Figure 14 for LDOR events. Only the anomalies having more than 10 percent incidence are reported here, but a comprehensive list of anomalies is avail- able in the accident/incident database. The highest incidence anomalies for landing overrun are contaminated and wet runways. Sometimes these anomalies occur in combination (e.g., a wet runway contaminated with rubber). For contaminated runways, ice was the most pre- dominant contaminant in the accidents and incidents evalu- ated. Three additional factors with high incidence for landing overruns are long touchdown, high speed during the ap- proach, and the presence of rain. According to the numbers presented in Figure 15 for LDUS, the most frequent anomaly was low visibility, followed by rain, particularly for the accidents. Gusting conditions had high incidence for these accidents. As expected, approaches below the glide path are an important anomaly for this type of event. Visual illusion was a significant factor only for land- ing undershoots. The presence of rain, gusting, crosswind, and low ceiling conditions were most predominant for these accidents when compared to incidents. Figure 16 depicts the most frequent anomalies for TOOR events. As expected, rejecting the takeoff operation at high speeds led to the majority of accidents and incidents. The sec- ond most important anomaly was incorrect planning, such as: aircraft overweight, short takeoff distance available, and incorrect load distribution in the aircraft. Basically, the fac- tors are equally frequent for accidents and incidents, except for the presence of rain, gusting, and crosswind conditions. These were more important for accidents when compared to incidents. A summary of the most frequent anomalies for all events by accident type is shown in Table 5. The “X” represents the anomaly was present in more than 10 percent of the cases for the specific event type: LDOR, LDUS, or TOOR. Unreported Events When using U.S. accidents and incidents as a sample, the number of reported incidents (53 percent) is close to the number of accidents (47 percent), when it was expected to see

22 Reported Events per Year 0 2 4 6 8 10 12 14 16 18 19 78 19 80 19 82 19 84 19 86 19 88 19 90 19 92 19 94 19 96 19 98 20 00 20 02 20 04 20 06 Year N um be r o f O cc ur re nc es ACC INC Figure 10. Events per year. a much higher number of incidents compared to accidents. One possible explanation for this phenomenon would be that some incidents are unreported. Therefore, an analysis of the number of unreported incidents was carried out. The methodology for evaluation of under reporting inci- dents is based on the assumption that there is a progressive decrease in the probability of travel to any given distance from the runway end with increasing distance. Such behavior is evident from the empirical accident and incident data set and is consistent with theoretical considerations of the nature of the event. The same basis considerations apply to LDOR, TOOR, and LDUS. Behavior of this type can be represented by a cumulative probability distribution function of the fol- lowing form: (3) where PROB{d > x} is the probability of traveling a distance d greater than x. Where there is full reporting of events it is expected that the available empirical data should fit a function consistent with this basic form. Where there is under-reporting, some dis- tortion in the apparent behavior can be expected. Failure to report is expected to be more likely for events where the dis- tance traveled off the runway is relatively low. The reported cumulative probability distribution (CPD) will be depressed at lower values of x but co-incident with the full distribution at higher values, as shown schematically in Figure 17. There may be other factors that distort the form of the CPD. A data set of incidents is expected to lack a dispro- portionate number of events at greater distances, since these would be expected to more likely result in more seri- ous consequences and be classified as accidents. On that basis, it is likely to be most appropriate to apply the analy- PROB d x e axn{ }> = − sis method proposed to a data set of incidents and accidents combined. Another factor that might distort the form of the curve is the obstacle environment beyond the runway end. An obsta- cle might cause the aircraft to come to a stop earlier than it might otherwise. Generally, there is an increased probability of an aircraft encountering an obstacle the farther it has trav- eled, and particularly when it has traveled farther than the RSA, this effect will lead to a reduced probability of aircraft traveling to greater distances than would otherwise be the case. The implications of this phenomenon required further consideration as part of this analysis. The analysis of unreported incidents is presented in Ap- pendix D, and the results are summarized in Table 6. Based on these numbers, different weights for statistical modeling were used to reflect the expected rate of incidents relative to accidents. Probability of Incident–Frequency Models The chance of an aircraft overrunning or undershooting a runway depends on the probability of accident per aircraft movement and the number of movements (landings and takeoffs) carried out per year. Logistic regression, discriminant analysis, and probit analysis were evaluated for modeling the probability of air- craft overrun and undershoot events. Discriminant analysis was not used because it involves numerous assumptions, including requirements of the independent variables to be normally distributed, linearly related, and to have equal vari- ance within each group (Tabachnick and Fidell, 1996). Logistic regression was chosen over probit analysis because

23 Category Anomaly Type Aircraft System Fault Tire Hydraulic Power Brake Other Wildlife Hazards Bird strike Other Weather Low Visibility Wind Shear Tailwind Crosswind Gusts Low Ceiling Strong Winds Turbulence Freezing Rain Rain Other Human Error Fatigue Communication/Coordination/ Planning Pressonitis Visual Illusion Other Runway Surface Wet Contamination / Low friction Standing Water Rubber Oil Slush Snow Ice Paint Construction Downslope Other Approach/Takeoff Procedures Unstabilized Approach Approach Below Flight Path Approach Above Flight Path High Speed Low Speed Long Touchdown Takeoff rejected at high speed Other Table 4. Anomalies during aircraft overrun and undershoot events. the latter does not give the equivalent of the odds ratio and changes in probability are harder to quantify (Pampel, 2000). Logistic regression is suited to models with a dichotomous outcome (incident and nonincident) with multiple predictor variables that include a mixture of continuous and categori- cal parameters. Logistic regression also is appropriate for case-control studies because it allows the use of samples with different sampling fractions depending on the outcome vari- able without giving biased results. In this study, it allowed the sampling fractions of accident flights and normal flights to be different. This property is not shared by most other types of regression analysis (Nagelkerke et al., 2005). Backward stepwise logistic regression was used to calibrate the three frequency models because of the predictive nature of the research. The selected technique is able to identify relationships missed by forward stepwise logistic regression (Hosmer and Lemeshow, 2000; Menard, 2001). Due to the more stringent data requirements of multivariate regression, cases with missing data were replaced by their respective series means. Every risk factor available in both Accident/Incident data- base and NOD were used to build each model. Table 7 shows the final parameters retained by the backward stepwise logis- tic regression as relevant independent variables for each of the frequency models. It should be noted that it was not possible to include some risk factors in the frequency models, for example, the ratio between the landing distance available and the landing distance required. Although a possible important factor to assess runway criticality, the lack of information for landing

24 Frequency Distribution of Anomalies for Landing Overruns 28% 0% 51% 53% 52% 58% 14% 0% 38% 27% 65% 39% 0% 10% 20% 30% 40% 50% 60% 70% Aircraft System Fault Wildlife Hazard Weather Condition Human Error Runway Conditions Approach/Takeoff Procedures % A CC /IN C w ith A no m al y ACC INC Figure 11. Frequency of anomalies by category (LDOR). Frequency Distribution of Anomalies for Landing Undershoots 12% 0% 59% 29% 67% 2% 2% 44% 38% 24% 31% 65% 0% 10% 20% 30% 40% 50% 60% 70% 80% Aircraft System Fault Wildlife Hazard Weather Condition Human Error Runway Conditions Approach/Takeoff Procedures % A CC /IN C w ith A no m al y ACC INC Figure 12. Frequency of anomalies by category (LDUS). distance required in the normal operations data precluded the use of such variables in the frequency models. However it is difficult to evaluate how much improvement such a fac- tor would bring to the model accuracy. Theoretically the runway length always should be compatible with the dis- tances required by the aircraft under certain conditions. In this sense the new factor may bring little benefit to the model but, on the other hand, a larger safety factor for distances required also should be expected for most flights operating in longer runways. The goal was to develop risk models based on actual ac- cidents/incidents and normal operation conditions so that the probability of occurrence for certain conditions may be estimated. The use of such models will help evaluate the

25 Frequency Distribution of Anomalies for Takeoff Overruns 33% 5% 56% 64% 24% 69% 51% 5% 14% 38% 11% 68% 0% 10% 20% 30% 40% 50% 60% 70% 80% % A CC /IN C w ith A n o m al y ACC INC Aircraft System Fault Wildlife Hazard Weather Condition Human Error Runway Conditions Approach/Takeoff Procedures Landing Overrun - Most Significant Anomalies 0% 10% 20% 30% 40% 50% 60% 70% Anomaly % o f T ot al fo r E ve nt T yp e All 12.8% 6.2% 31.5% 18.3% 17.9% 16.1% 11.4% 17.6% 20.1% 39.2% 53.5% 38.1% 25.3% Acc 14.2% 10.8% 35.0% 21.7% 21.7% 20.8% 17.5% 26.7% 30.8% 35.8% 42.5% 45.0% 33.3% Inc 11.8% 2.6% 28.8% 15.7% 15.0% 12.4% 6.5% 10.5% 11.8% 41.8% 62.1% 32.7% 19.0% SysF- Brake SysF- Other W- Rain W- Low Visib. W- Tailwind W- Xwind W- Gusts H-Inc Plan. H- Other R- Wet R- Contam /LF AT- Long TDown AT-High Spd App Figure 13. Frequency of anomalies by category (TOOR). Figure 14. Most frequent anomalies (LDOR). likelihood of incident occurrence for a runway that is sub- ject to certain environmental and traffic conditions over the year. The frequency model is in the following form: (4) where P Accident Occurrence e b b X b X b { _ } ( = + − + + + 1 1 0 1 1 2 2 3 3X +...) P{Accident_Occurrence} = s the probability (0-100%) of an accident type occurring given certain operational conditions; Xi = independent variables (e.g. ceil- ing, visibility, crosswind, tail- wind, aircraft weight, runway condition, etc.); and bi = regression coefficients.

26 Landing Undershoot - Most Significant Anomalies 0% 5% 10% 15% 20% 25% 30% 35% 40% 45% 50% Anomaly % o f T ot al fo r E ve nt T yp e All 21.3% 43.6% 8.5% 9.6% 17.0% 7.4% 19.1% 16.0% 11.7% 8.5% 22.3% 11.7% 17.0% Acc 30.6% 46.9% 10.2% 18.4% 26.5% 12.2% 19.1% 16.0% 11.7% 4.1% 22.3% 14.3% 24.5% Inc 11.1% 40.0% 6.7% 0.0% 6.7% 2.2% 17.8% 15.6% 8.9% 13.3% 15.6% 8.9% 8.9% W- Rain W- Low Visib. W- Wind Shear W- Xwind W- Gusts W- Low Ceiling H-Inc Plan. H- Other H- Visual Illusion R- Constr. AT- Low App. AT-Low Spd App AT- Other Figure 15. Most frequent anomalies (LDUS). Takeoff Overrun - Most Significant Anomalies 0% 10% 20% 30% 40% 50% 60% 70% 80% Anomaly % o f T ot al fo r E ve nt T yp e All 9.8% 10.9% 20.7% 16.3% 12.0% 15.2% 12.0% 38.0% 12.0% 12.0% 67.4% Acc 7.3% 9.1% 16.4% 25.5% 14.5% 21.8% 18.2% 40.0% 20.0% 16.4% 69.1% Inc 13.5% 13.5% 27.0% 2.7% 8.1% 5.4% 2.7% 35.1% 0.0% 5.4% 64.9% SysF- Tire SysF- Power SysF- Other W- Rain W- Low Visib. W- Xwind W- Gusts H- Inc. Planning H- Other R- Wet AT- TO Rej. @ HS Figure 16. Most frequent anomalies (TOOR). Before logistic regressions were performed, it was ensured that all assumptions for the statistical procedure were met. Logistic regression is relatively free from assumptions, espe- cially compared to ordinary least squares regression. How- ever, a number of assumptions still apply. One of these is a linear relationship between the independents and the log odds (logit) of the dependent. The Box-Tidwell transformation test was used to check whether all continuous variables met this assumption (Garson, 1998). This involved adding to the model interac- tion terms that are the cross-product of each independent variable times its natural logarithm [(X)ln(X)]. The logit lin- earity assumption is violated if these terms are significant. In the current analysis, the continuous variables were found to have non-linear logits. As a solution, these variables were divided into different categories according to standard equal intervals using landing NOD and accident data. The vari- ables then were converted into categorical ones with these different levels, each being a separate logit independent variable.

27 LDOR LDUS TOORAnomaly ACC INC ACC INC ACC INC Brake system failure X X Power failure X Tire failure X Other aircraft system fault X X X Rain X X X X X Low Visibility X X X X X Low Ceiling X Tailwind X X Crosswind X X X X Wind shear X Gusts X X X Improper flight planning X X X X X X Visual illusion X Other human errors X X X X Wet runway X X X Contaminated runway X X Long touchdown X X High speed during approach X X Low speed during approach X Approach too low X X Other approach anomalies X Runway construction X Rejected takeoff at high speed X X Table 5. Summary of anomalies for aircraft overruns and undershoots. Form of Cumulative Probability Distribution 0 0.2 0.4 0.6 0.8 1 0 20 40 60 80 100 Distance Travelled Fr ac tio n of E ve nt s Normal CPD CPD with Under-reporting Figure 17. Schematic form of cumulative distribution functions. A test for multicollinearity is required for multivariate logistic regression. Collinearity among the predictor variables was assessed by conducting linear regression analyses to ob- tain the relevant tolerance and Variance Inflation Factor (VIF) values. None of the tolerance values were smaller than 1, and no VIF value was greater than 10, suggesting that collinearity among the variables is not serious (Myers, 1990; Menard, 2001). Kendall’s Tau also was used to assess potential correla- tions between predictor variables that are likely to be related. Two pairs of variables had Kendall’s Tau correlation coeffi- cient between 0.51 and 0.60, indicating moderate correlation: equipment class with airport hub size and icing conditions with frozen precipitation. Since none of the correlations were serious, all variables were kept in the multivariate model, and caution was applied in interpreting the results. This is preferred to the alternative solution of removing variables, which would lead to model misspecification. Although the R2 for the models ranged between 0.148 and 0.245, as shown in Table 8, relatively low values are the norm in logistic regression (Ash and Schwartz, 1999), and they should not be compared with the R2 of linear regressions (Hosmer and Lemeshow, 2000). The analysis of models using Receiver Operating Characteristic (ROC) curves to classify flights as “accident” or “normal” suggests good to excellent classification accuracy for such models (C-Statistic from 0.819 to 0.872).

28 Total # of Accidents and Incidents Total # of Incidents % Unreported Incidents ** Estimated # Unreported Incidents LDOR 240 121 28.8% 17 LDUS 81 38 9.6% 7 TOOR 75 28 28.8% * 1 Note: * value assumed based on comparisons with LDOR ** based on incidents occurring at small distances from threshold Table 6. Summary results for under-reported incidents. Variable LDOR LDUS TOOR Aircraft Weight/Size X X X Aircraft user class X X Ceiling X X X Visibility X X X Fog X X Crosswind X X Gusts Icing Conditions X X X Snow X X X Rain X Temperature X X X Electrical Storm X Turboprop/Jet X Foreign Origin/Destination X X Hub/Non-hub airport X Table 7. Independent variables used for frequency models. Model R2 C LDOR 0.245 0.872 LDUS 0.199 0.819 TOOR 0.148 0.861 Table 8. Summary statistics for frequency models. Type of Event Sampling Fraction (t1) Original Intercept (b*0) Adjusted Intercept (b0) LDOR 0.938274 -7.656 -15.45637 LDUS 0.943765 -7.158 -14.96421 TOOR 0.997447 -8.790 -16.65153 Table 9. Calculated model intercepts. Due to the case-control set-up of the study, the constant (intercept) term b0 of the final formula must be adjusted to account for the different sampling fractions between the cases and the controls. The following formula was used for this purpose (Hosmer and Lemeshow, 2000): b*0 = ln(t1/t0) + b0 (5) where b*0 = the original intercept, t1 = the sampling fraction of cases, t0 = the sampling fraction of controls, and b0 = s the adjusted intercept. Although parameter t1 is normally one when relevant in- formation is available for all events, it was necessary to adjust these values to reflect under-reporting of incidents. From the NOD sampling exercise, it was calculated that the total number of relevant normal operations from 2000 to 2005 inclusive is 191,902,290. That is 44.78 percent of the period’s total itinerant operations excluding military operations. From the TAF, the total number of itinerant operations from 1982 to 2002 inclusive (the accident sampling period) excluding military operations was computed to be 1,408,495,828 move- ments. Of the latter, 44.78 percent equates 630,792,133 move- ments. A detailed description on the calculation of relevant terminal area forecast traffic is presented in Appendix J. Since the total sampled normal operation population is 242,420 flights, t0 = 242420/630792133 = 3.843 × 10−4 With t1 and t0, the adjusted intercepts of each of the risk model formula can be calculated: b*0 = ln(t1/t0) + b0 = ln(t1/3.843 × 10−4) + b0 = 7.864 + b0 (6) Where b*0 is the original intercept, t1 is the sampling frac- tion of cases, t0 is the sampling fraction of controls, and b0 is the adjusted intercept. The calculated parameters for each model are shown on Table 9. Using the adjusted intercepts, the final frequency models are the following: Landing Overrun (7) b Acft CommuterAc= − + −15 456 0 551 2 113. . ( ) . (Heavy ft MediumAcft SmallAcft ) . ( ) . ( ) . − − + 1 064 0 876 0 445 0 857 1 832 ( ) . ( ) . ( TurbopropAcft ForeignOD C − + eilingHeight ft CeilingHeight < + 1000 1 639 10 ) . ( 01 2500 2 428 2 1 186 − + < + ft Visibility SM Vis ) . ( ) . ( ibility SM Visibility SM 2 4 1 741 4 6 0 322 − + − + ) . ( ) . ( ) . ( ) Visibility SM Crosswind knts 6 8 0 532 2 5 − − − +1 566 5 12 1 518 1 . ( ) . ( Crosswind knts Crosswind − + > 2 0 986 1 926 knts ElectStorm IcingCondit ) . ( ) . ( + + ions Snow Temp C Tem ) . ( ) . ( ) . ( + − < − 1 499 1 009 5 0 631 p C Temp C NonhubApt 5 15 0 265 25 1 006 0 − + > + + ) . ( ) . ( ) . ( )924 SignificantTerrain

29 Equipment Class Ref: C Large jet of MTOW 41k-255k lb (B737, A320 etc.) HeavyAcft AB Heavy jets of MTOW 255k lb+ CommuterAcft D Large commuter of MTOW 41k-255k lb (small RJs, ATR42 etc.) MediumAcft E Medium aircraft of MTOW 12.5k-41k lb (biz jets, Embraer 120 Learjet 35 etc.) SmallAcft F Small aircraft of MTOW 12.5k or less (small, single or twin engine Beech90, Cessna Caravan etc.) User Class Ref: C = Commercial UserClass1 F = Cargo UserClass2 G = GA ForeignOD Foreign origin/destination (yes/no) - Ref: domestic CeilingHeight Ref: >2500ft CeilingHeight<1000ft <1000 CeilingHeight1001- 2500ft 1001-2500 Visibility Ref: 8-10 statute miles (SM) Visibility<2SM < 2 SM Visibility2-4SM 2-4 SM Visibility4-6SM 4-6 SM Visibility6-8SM 6-8 SM Crosswind Ref:< 2 knots Xwind2-5knts 2-5 knots Xwind5-12knts 5-12 knots Xwind>12knts >12 ElectStorm Electrical storm (yes/no) – Ref: no IcingConditions Icing conditions (yes/no) – Ref: no Snow Snow (yes/no) – Ref: no Air Temperature Ref: 15 – 25 deg.C Temp<5C < 5 deg.C Temp5-15C 5 – 15 deg.C Temp>25C > 25 deg.C NonhubApt Non-hub airport (yes/no) – Ref: hub airport SignificantTerrain Significant terrain (yes/no) – Ref: no Notes: Ref: indicates the reference category against which the odds ratios should be interpreted. Non-hub airport: airport having less than 0.05% of annual passenger boardings Significant terrain: terrain within the plan view of airport exceeds 4,000 feet above the airport elevation, or if the terrain within a 6.0 nautical mile radius of the Airport Reference Point rises to at least 2,000 feet above the airport elevation. Landing Undershoot (8) b Acft CommuterA= − + −14 9642 0 036 1 699. . ( ) . (Heavy cft MediumAcft SmallAcft ) . ( ) . ( ) . − + + 0 427 1 760 0 288 1 0 908 2 1 042( ) . ( ) . (UserClass UserClass Fo+ − reignOD CeilingHeight ft ) . ( ) . + < + 0 199 1000 1 463( ) . ( CeilingHeight ft Visibilit 1001 2500 2 074 − + y SM Visibility SM Visibi < + − − 2 0 069 2 4 0 185 ) . ( ) . ( lity SM Visibility SM Fo 4 6 0 295 6 8 1 830 − − − + ) . ( ) . ( g Rain Temp C Temp ) . ( ) . ( ) . ( − − < − − 1 705 0 505 5 0 874 5 15 0 446 25 2 815 2 412 C Temp C Icing Sn ) . ( ) . ( ) . ( − > + + ow) Takeoff Overrun (9) Where: b Acft CommuterA= − + −16 6515 0 721 0 619. . ( ) . (Heavy cft MediumAcft SmallAcft ) . ( ) . ( ) .− + +0 009 1 669 1 336 1 1 052 2 1 225 ( ) . ( ) . ( UserClass UserClass Ce+ + ilingHeight ft CeilingHeight < + 1000 1 497 100 ) . ( 1 2500 0 201 2 1 941 − + < − ft Visibility SM Visi ) . ( ) . ( bility SM Visibility SM 2 4 0 366 4 6 0 317 − − − + ) . ( ) . (Visibility SM Fog Xwind 6 8 1 660 0 292 2 5 − + − − ) . ( ) . ( knts Xwind knts Xwind ) . ( ) . ( )+ − + >1 598 5 12 1 781 12 − < − − + 0 536 5 0 507 5 15 0 502 . ( ) . ( ) . ( Temp C Temp C Temp C Icing Snow> + +25 1 805 2 567) . ( ) . ( )

30 Raw Distances Model for Landing Overruns 0% 20% 40% 60% 80% 100% 0 500 1000 1500 2000 2500 Distance x from Threshold Pr ob {d ist an ce > x ) % Actual Raw Distances Raw Distances Model P{d>x}=exp(-0.003871*x^0.955175) R2=99.8%; n=257 Figure 18. LDOR location model using raw (nonnormalized) distances. Appendix M provides the results for multivariate logistic regression analysis used to obtain the model coefficients described earlier. Accident Location Models Based on the accident/incident data for wreckage loca- tions, three sets of complementary cumulative probability distribution (CCPD) models were developed in this study. With CCPDs, the fraction of accidents involving locations exceeding a given distance from the runway end or thresh- old can be estimated. When the CCPD is multiplied by the frequency of accident occurrence, a complementary cumu- lative frequency distribution (CCFD) is obtained. The latter quantifies the overall frequency of accidents involving loca- tions exceeding a given distance from the runway end or threshold. The CCPD model structure selected was used by Eddowes et al. (2001) and is in the following form: For the longitudinal distribution, the basic model is: (10) where P{Location > x} = the probability the overrun/undershoot distance along the runway centerline beyond the threshold is greater than x; x = a given location or distance beyond the threshold; and a, n = regression coefficients. For the transverse distribution, the same model structure was selected. However, given the accidents transverse loca- tion is not reported, in general, if the wreckage location is within the extended runway lateral limits, it was necessary to P Location x e axn{ }> = − use weight factors to reduce model bias, particularly for mod- eling the tail of the probability distribution. Therefore the model can be represented by the following equation: (11) where P{Location>y} = the probability the overrun/undershoot distance from the runway centerline is greater than y (P{Location<=0} = c); y = a given location or distance beyond the threshold; and b, m = regression coefficients. The correlations between the overrun and undershoot dis- tances to the lateral distance relative to the runway axis also were evaluated for assessing the correlation between x and y locations. A high correlation would suggest the best geome- try for RSAs is not a rectangle. When plotting the percent of accidents beyond a certain distance from the threshold, shown in Figure 18, it can be noted that an RSA with 1000 ft in length will encompass close to 95 percent of all landing overruns. It should be noted that the raw data includes reported accidents and incidents, but incidents were weighted to account for unreported cases. Figure 19 depicts the distribution of raw lateral distances from the extended runway centerline. For many events the distance was very close to the runway centerline and the actual distance was not reported. For such cases when possi- ble, the y-distance was assumed to be 0.0. As mentioned earlier, weighting factors were used to obtain unbiased esti- mates at the tails of the distribution. In this case, weighting was applied to the events having y-distances above 400 ft. The LDOR CCPD for normalized distances is shown in Figures 20 and 21. Using transformed distances, a 1000 ft-long P Location y e bym{ }> = −

31 Raw Lateral Distances Model for LDOR 0% 20% 40% 60% 80% 100% 0 200 400 600 800 1000 1200 Distance Y from Extended Runway Axis (ft) Pr ob ai lit y{d ist an ce > y} Actual Data Model P{d>x}=exp(-0.20174*C2^0.489009) R2=94.7%; n=141 Figure 19. LDOR lateral location model using raw (nonnormalized) distances. Normalized Distances Model for Landing Overruns 0% 20% 40% 60% 80% 100% 0 1000 2000 3000 4000 5000 6000 Distance x from Threshold (ft) Pr ob {d ist an ce > x} Actual Data Norm Model Raw Model P{d>x}=exp(-0.004692*B2^0.824513) R2=99.5%; n=232 Figure 20. LDOR location model using normalized distances. Normalized Lateral Distances Model for LDOR 0% 20% 40% 60% 80% 100% 0 500 1000 1500 2000 2500 Distance Y from Extended Runway Axis (ft) Pr ob ai lit y{ dis tan ce > y} Actual Data Model P{d>x}=exp(-0.243692*Y^0.388726) R2=93.4%; n=138 Figure 21. LDOR lateral location model using normalized distances.

32 Raw Distances Model for Landing Undershoots 0% 20% 40% 60% 80% 100% 0 500 1000 1500 2000 2500 Distance x to Threshold Pr ob {d ist an ce > x} Actual Data Alt Model P{d>x}=exp(-0.024445*C4^0.643232) R2=98.5%; n=82 Figure 22. LDUS location model using raw (nonnormalized) distances. RSA will encompass approximately 80 percent of all landing overruns. The probability that the point of first impact is beyond a certain distance for landing undershoots is depicted in Figures 22 and 23, for raw distances, and in Figures 24 and 25, for normalized distances. For nearly 13 percent of landing undershoots, the aircraft point of first impact will occur at distances greater than 1000 ft from the runway threshold. The raw location probability trend for takeoff overruns is depicted in Figures 26 and 27. From the raw, unweighted accident and incident data, close to 20 percent of takeoff overruns will occur beyond a 1000 ft distance from the threshold. The normalized distance models for takeoff over- runs is presented in Figures 28 and 29. For each set of location models, one model was developed with the raw distance locations and one model used normal- ized distances relative to terrain type, runway elevation, and the air temperature during the accident/incident. Tables 10 and 11 show the location models developed in this study. The sample sizes available to develop the models shown in Table 11 were smaller than those used for the models shown in Table 10. A number of investigation reports provide only the distance from the threshold, but not the lateral distance. Sample sizes for normalized models also are smaller than those developed with raw data. For a few cases in each acci- dent group there was no information on the terrain type used to normalize the distance. Analysis of RSA Geometry The correlation between the overrun and undershoot dis- tances to the lateral distance relative to the runway axis was evaluated to define the geometry of the safety areas. RSA are normally rectangular-shaped areas, but it was possible that a strong correlation between longitudinal and lateral could exist. In other words, a statistical analysis was necessary to evaluate if greater longitudinal distances for wreckage loca- tion can lead to greater transverse distances. The correlation between the longitudinal and lateral dis- tance for each type of event is shown in Table 12. Although the correlation between x and y locations is not zero for LDOR and LDUS (P < 0.05), the level is relatively low; it was assumed that the correlation is not important. This leads to the assumption that the transverse location dis- tribution of accidents is fairly constant along the longitudinal locations from the threshold. Consequences As described earlier, accident costs were used to integrate consequences related to injuries and property loss into a single parameter. The initial intent was to relate the consequences, represented by the accident cost, with the wreckage distance for the accident. The relationship could be used to estimate the consequences of accidents based on the wreckage location, providing a link between the location and consequences mod- els. Unfortunately these relationships were found to be quite poor, as consequences depend not only on the speed when the aircraft departs the runway, but also the nature and location of existing obstacles, as well as the type and size of aircraft. Additional analysis attempted to relate accident location with aircraft damage. Four categories of damage—none, minor, substantial, and hull loss—were correlated to accident location. The use of raw distances proved to hold very low correlations between wreckage path distance and the aircraft damage. However, there was an improvement when normal- ized distances relative to terrain, elevation, and temperature were used. The correlations are quite reasonable, as shown in Table 13.

33 Raw Lateral Distances Model for LDUS 0% 20% 40% 60% 80% 100% 0 800600400200 Distance Y from Extended Runway Axis (ft) Pr ob ai lit y{ dis tan ce > y} Actual Data Model P{d>x}=exp(-0.409268*Y^0.351851) R2=92.0%; n=48 Figure 23. LDUS lateral location model using raw (non-normalized) distances. Normalized Distances Model for Landing Undershoots 0% 20% 40% 60% 80% 100% 0 2000 4000 80006000 10000 12000 Distance x to Threshold (ft) Pr ob {d ist an ce > x } Actual Norm Data Norm Model Raw Model P{d>x}=exp(-0.022078*X^0.585959) R2=99.1%; n=69 Figure 24. LDUS location model using normalized distances. Normalized Lateral Distances Model for LDUS 0% 20% 40% 60% 80% 100% 0 400200 600 800 1000 1200 1400 Distance Y from Extended Runway Axis (ft) Pr ob ai lit y{ dis tan ce > y} Actual Data Model P{d>x}=exp(-0.19539*Y^0.433399) R2=90.3%; n=41 Figure 25. LDUS lateral location model using normalized distances.

34 Raw Distances Model for Takeoff Overruns 0% 20% 40% 60% 80% 100% 0 500 1000 1500 2000 2500 Distance x from Threshold (ft) Pr ob {d ist an ce > x } Actual Data Model P{d>x}=exp(-0.001033*x^1.065025) R2=99.0%; n=76 Figure 26. TOOR location model using raw (nonnormalized) distances. Raw Lateral Distances Model for TOOR 0% 20% 40% 60% 80% 100% 0 200 400 600 800 1000 1200 1400 Distance Y from Extended Runway Axis (ft) Pr ob ai lit y{ dis tan ce > y} Actual Data Model P{d>x}=exp(-0.182098*Y^0.448346) R2=95.6%; n=44 Figure 27. TOOR lateral location model using raw (nonnormalized) distances. Normalized Distances Model for Takeoff Overruns 0% 20% 40% 60% 80% 100% 0 20001000 3000 4000 5000 6000 Distance x from Threshold (ft) Pr ob {d ist an ce > x } Actual Data Norm Model Raw Model P{d>x}=exp(-0.003364*x^0.807138) R2=98.5%; n=72 Figure 28. TOOR location model using normalized distances.

35 Both Spearman R and Kendal Tau correlation coefficients provide an indicator of the degree of co-variation in the vari- ables. Both tests require that variables are represented at least in ordinal scale (rank), which is the case for aircraft damage. While Spearman R has an approach similar to the regular Pearson product-moment correlation coefficient, Kendall Tau rather represents a probability. Despite these reasonable correlations, a more rational approach to model consequences was preferred to assess the effect of different obstacles at various locations in the vicinity of the RSA. Examples of such obstacles include fences, drops and elevations in the terrain, existing facilities, culverts, ALS, and ILS structures, trees, etc. Modeling Approach The main purpose for modeling consequences of aircraft accidents is to quantify the risk based on the probability of occurrence and the results in term of injuries and property loss. It was not possible to develop one model for each type of accident, as previously done to model frequency and location. However, a rational probabilistic approach is suggested to evaluate the probability of accidents or serious accidents. The basic idea is to use the location model to estimate the incident occurrences when the aircraft will have high energy resulting in serious consequences. Figure 30 can be used to illustrate and help understand this approach. The x-axis represents the longitudinal location of the wreckage relative to the threshold. The y-axis is the probabil- ity that the wreckage location exceeds a given distance “x.” The location distance can be normalized or not, according to the criteria selected. In this example, an obstacle is located at a distance “D” from the threshold and the example scenario being analyzed is an aircraft landing overrun incident. The figure shows an exponential location model developed for the specific acci- dent scenario, in this case, landing overrun. There are three distinct regions in this plot. The first region (medium shaded area) represents those occurrences that the aircraft departed the runway, but the exit speed was relatively Normalized Lateral Distances Model for TOOR 0% 20% 40% 60% 80% 100% 0 500 1000 1500 2000 2500 3000 Distance Y from Extended Runway Axis (ft) Pr ob ai lit y{ dis tan ce > y} Actual Data Model P{d>x}=exp(-0.181046*Y^0.406544) R2=97.1%; n=42 Figure 29. TOOR lateral location model using normalized distances. Type of Accident Type of Data Model Eq.# R2 # of Points Raw 955175.0003871.0}{ xexdP (12) 99.8% 257 LDOR Normalized 824513.0004692.0}{ xexdP (13) 99.5% 232 Raw 643232.0024445.0}{ xexdP (14) 98.52% 82 LDUS Normalized 585959.0022078.0}{ xexdP (15) 99.1% 69 Raw 065025.1001033.0}{ xexdP (16) 99.0% 76 TOOR Normalized 807138.0003364.0}{ xexdP (17) 98.5 72 where P{d > x} is the probability the wreckage location exceeds distance x from the threshold, and x is the longitudinal distance from the threshold. Table 10. Summary of X-location models.

36 Type of Accident Type of Data Model Eq.# R2 # of Points Raw 489009.020174.0}{ yeydP (18) 94.7% 141 LDOR Normalized 388726.0243692.0}{ yeydP (19) 93.4% 138 Raw 351851.0409268.0}{ yeydP (20) 92.0% 48 LDUS Normalized 433399.019539.0}{ yeydP (21) 90.3% 41 Raw 448346.0182098.0}{ yeydP (22) 95.6% 44 TOOR Normalized 406544.0181046.0}{ yeydP (23) 97.1% 42 Table 11. Summary of Y-location models. low, and the aircraft came to a stop before reaching the exist- ing obstacle. The consequences for such incidents are expected to be none to minor as the aircraft may hit only frangible ob- jects (e.g., threshold lights) within these small distances. The rest of the curve represents events that the aircraft exited the runway at speeds high enough for the wreckage path to extend beyond an existing obstacle. However, a por- tion of these accidents will have relatively higher energy and should result in more severe consequences, while for some cases the aircraft will be slow when hitting the obstacle so that catastrophic consequences are less likely to happen. Using this approach, it is possible to assign three scenarios: the probability that the aircraft will not hit the obstacle (re- sulting in none or minor consequences); the probability that the aircraft will hit the obstacle with low speed and energy (with substantial damage to aircraft but minor injuries); and the probability that the aircraft will hit the obstacle with high energy (with substantial damage and injuries). For events with low energy when impacting the obstacle, it is possible to assume that if no obstacle was present the aircraft would stop within a distance Δ from the location of the obsta- cle. The problem is to evaluate the rate of these accidents hav- ing low speeds at the obstacle location and this is possible based on the same location model. This probability can be estimated by excluding the cases when the speed is high and the final wreckage location is significantly beyond the obstacle location. A similar approach was developed to combine the longitu- dinal and transverse location distribution with the presence, type, and dimensions of existing obstacles. The basic approach is represented in Figure 31 for a single and simple obstacle. A few simplifying assumptions were necessary when devel- oping this approach. One simplification is to assume the lateral distribution is random and does not depend on the presence of obstacles. This is a conservative assumption be- cause there are events when the pilot will avoid some obsta- cles if he has some control of the aircraft. The database contains a number of cases when the pilot avoided ILS and Approach Lighting System (ALS) structures in the RSA. A second assumption is that the aircraft follows a path near parallel to the extended runway axis. Again, this assumption will lead to calculations of higher than actual risk and is con- servative. The aircraft may hit or avoid obstacles in paths that are nonparallel to the runway axis. The shaded area in Figure 31 represents the area of analy- sis. Accident data was considered relevant when wreckage location challenged an area of 2000 × 2000 ft beyond the threshold. The example shown in the figure depicts an over- run example. Obstacle 1 is located at a distance xo, yo from the threshold and has dimensions W1 × L1. When evaluating the possibility of severe consequences it is possible to assume this will be the case if the aircraft fuselage or a section of the wing close to the fuselage hits the obstacle. Thus, it is possible to assume the accident will have severe consequences if the y location is between Yc and Yf, as shown in the figure. Based on Equa- tion 11 for transverse distance, the probability the aircraft axis is within this range can be calculated as follows: (24) where Psc = the probability of high consequences; b, m = regression coefficients for y-location model; Yc = the critical aircraft location, relative to the obstacle, closest to the extended runway axis; and P e e sc byc m by f m = − − − 2 Type of Event Sample Size Spearman R Kendall Tau LDOR 224 0.62 0.49 LDUS 68 0.30 0.23 TOOR 67 0.55 0.44 All 359 0.56 0.44 Table 13. Correlation between normalized wreckage location and aircraft damage. Type of Event R R2 CI 95% p n LDOR 0.320 10.2% 0.20 - 0.43 < 0.0001 235 LDUS 0.316 10.0% 0.11 - 0.50 0.0040 81 TOOR 0.113 1.3% -0.12 a 0.33 0.3430 73 Table 12. Correlation between lateral and longitudinal overrun/undershoot distances.

37 a, n = regression coefficients for the x-model; and Δi = the location parameter for obstacle i. The value of Δ may be estimated based on Kirkland’s model for aircraft deceleration over different types of terrain (Kirkland et al., 2004) and crashworthiness speed criteria for aircraft. It should be noted that Δ depends on the type of terrain, type and size of aircraft, and type of obstacle. Frangible objects in the RSA are less prone to causing severe consequences. Lighter aircraft may stop faster and the landing gear configuration also may have an effect on the aircraft deceleration in soft terrain, but these factors are not accounted for in Kirkland’s model. The probability and location models should provide a quantitative assessment based on operating conditions for a specific airplane landing or takeoff at a specific runway. The consequences model should provide a qualitative assessment of the severity of an accident, based on the location model and the existing runway characteristics, to include dimen- sions of existing RSA, airplane weight, type, location and size of obstacles, and the topography of the surrounding terrain. The procedure will allow modeling overrun and under- shoot risks for the conditions of the airport being evaluated. The probability of the accident occurring, as well as stopping location distances, will be compared to existing geometry of safety areas and existing obstacles to assess the possible con- sequences of the accident, at least qualitatively. Cost of Accidents As described in the previous chapter, the direct costs of ac- cidents and incidents were estimated for each event having sufficient information for the computation. This section presents a summary of these costs. Yf = the critical aircraft location, relative to the obstacle, farther from the extended runway axis. The same example is depicted in Figure 32 showing the probability of severe consequences can be represented by the lightly shaded area in the probability distribution. Combining this approach with the longitudinal distribu- tion approach and the possibility of multiple obstacles, the risk for accidents with severe consequences can be estimated using the following model: (25) where N = the number of existing obstacles; P e e esc byci m by fi m a xi i n i N = −( )− − − +( ) = ∑ 21 Δ D0 is distance to Obstacle, d is distance the aircraft came to stop Area (lightly shaded) between D0 and D0 +Δ represent % occurrences at low speed (energy) when hitting obstacle (low consequences) Figure 30. Approach to model consequences of overrun/undershoot accidents. x y 2000ft 20 00 ft RSA y1 y2 x (dist. to obstacle) Figure 31. Modeling consequences.

As mentioned earlier, the intent was to use the data and find relationships between certain parameters of the accident (e.g., wreckage path distance) and the number and level of injuries, as well as damage to aircraft. The total consequences were estimated in terms of total direct costs for injuries, air- craft damage, and accident investigation. Figure 33 depicts the average cost by type of accident and by severity. Most of the cost for LDORs is attributed to loss of property or aircraft damage. On the other hand, loss of dollars due to injuries is significantly higher for LDUSs, most likely due to the high speed and energy during these accidents. The average loss of property among the three types of ac- cidents was fairly similar. As expected, the cost of incidents was significantly lower for all three types of events. The cost of investigation is not represented in the figure, but rather is shown in Table 14. The costs for injuries, aircraft damage, and accident/incident investigation are available in the accident database for each event included in this study. Appendix L provides more details on the calculation of accident costs. 38 Lateral Location Probability y Obstacle y1 y2 Psc Figure 32. Modeling likelihood of striking an obstacle. Accident and Incident Average Direct Costs (2007 dollars) 5.60 0.11 4.19 0.06 6.72 0.05 1.31 0.00 22.11 0.00 15.43 0.21 0.00 5.00 10.00 15.00 20.00 25.00 Type of Event A ve ra ge C os t ( $ M illi on ) Aircraft 5.60 0.11 4.19 0.06 6.72 0.05 Injuries 1.31 0.00 22.11 0.00 15.43 0.21 LDOR ACC LDOR INC LDUS ACC TOOR ACC TOOR INCLDUS INC Figure 33. Direct cost of accidents and incidents. Type of Event Total Cost Investigation Cost LDOR ACC 7.08 0.17 LDOR INC 0.11 0.00 LDUS ACC 26.75 0.45 LDUS INC 0.06 0.00 TOOR ACC 22.60 0.46 TOOR INC 0.26 0.00 Table 14. Total and investigation costs (2007 dollars, millions).

Next: Chapter 4 - Practical Application of Models »
Analysis of Aircraft Overruns and Undershoots for Runway Safety Areas Get This Book
×
MyNAP members save 10% online.
Login or Register to save!
Download Free PDF

TRB's Airport Cooperative Research Program (ACRP) Report 3: Analysis of Aircraft Overruns and Undershoots for Runway Safety Areas explores overrun and undershoot accident and incident data conditions relating to these occurrences. The report also includes an assessment of risk in relation to the runway safety area and highlights a set of alternatives to the traditional runway safety area. The appendices to ACRP Report 3 are available online.

ACRP Report 50: Improved Models for Risk Assessment of Runway Safety Areas, which was released in July 2011, expands on the researc presented in ACRP Report 3. ACRP Report 50 analyses aircraft veer-offs, the use of declared distances, the implementation of the Engineered Material Arresting System, and the incorporation of a risk approach for consideration of obstacles in or in the vicinity of the runway safety area.

View the Impact on Practice related to this report.

  1. ×

    Welcome to OpenBook!

    You're looking at OpenBook, NAP.edu's online reading room since 1999. Based on feedback from you, our users, we've made some improvements that make it easier than ever to read thousands of publications on our website.

    Do you want to take a quick tour of the OpenBook's features?

    No Thanks Take a Tour »
  2. ×

    Show this book's table of contents, where you can jump to any chapter by name.

    « Back Next »
  3. ×

    ...or use these buttons to go back to the previous chapter or skip to the next one.

    « Back Next »
  4. ×

    Jump up to the previous page or down to the next one. Also, you can type in a page number and press Enter to go directly to that page in the book.

    « Back Next »
  5. ×

    To search the entire text of this book, type in your search term here and press Enter.

    « Back Next »
  6. ×

    Share a link to this book page on your preferred social network or via email.

    « Back Next »
  7. ×

    View our suggested citation for this chapter.

    « Back Next »
  8. ×

    Ready to take your reading offline? Click here to buy this book in print or download it as a free PDF, if available.

    « Back Next »
Stay Connected!