Climate Resilience and Benefit–Cost Analysis: A Handbook for Airports (2019)

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.

119 Four airports agreed to participate in illustrative case studies demonstrating the methodology described in this handbook to help airports evaluate the potential impacts of climate change. The project team presented example scenarios of specific climate risks faced by each airport using the most recent and localized climate data. This appendix provides further details on the interactions with each airport. It is important to emphasize that while the project mitigations are purely illustrative, the climate data shown in each case represent actual current estimates of potential future climate outcomes. It is also important to emphasize that, in all cases, the climate projections used were those from RCP8.5, which represents a high-emissions scenario for future climate change. New Orleans Historical experience and current climate projections indicate that MSY is at significant risk of flooding. After introducing the ACROS tool to the airport team, the summary results from that software were presented, as shown in Exhibit H-1. The results show that flood risk for the airport will occur every day of the year by 2030, regardless of the climate model employed. ACROS shows an increase of 2 ft in BFE, which would have implications for protecting existing infrastructure through the use of dikes or raising infrastructure to offset flood risks. Following the methodology suggested previously, ACROS screening indicated that further analysis was warranted. A slide from the MSY presentation, shown in Exhibit H-2, provides a general introduc- tion to historic EWL and projections of future RSL rise. NOAA is the source for both data series. The example in the slide is taken from data for Kings Point/Willets Point, New York, a tidal station near LaGuardia Airport. The top graph is the historic probability of water levels above mean high tide; for example, this area could expect a 1.5-m extreme water event every 10 years. Therefore, the annual probability of such an event is 1 in 10. The bottom table on the right shows local sea level rise predictions (expressed in centimeters) for the Kings Point station based on six different forecast GMSL rise scenarios (low to extreme). The probabilities of these outcomes are linked to three global emissions scenarios used by climate scientists—RCP2.4, 4.5, and 8.5, as shown in the bottom left table. For example, under RCP8.5, the probability of the King’s Point Intermediate scenario sea level rise of 41 cm is 17% in 2050. As described in Appendix D, after selecting one of the RCP scenarios, one can generate a random draw of future sea level rise by interpolating between the GMSL rise scenarios and A P P E N D I X H Case Study Details

120 Climate Resilience and Benefit–Cost Analysis: A Handbook for Airports Exhibit H-1. ACROS results for MSY. between the 10-year intervals for the localized projections. Combining that draw with a random draw from the historical EWL graph results in a probabilistic localized projection of the height of a future EWL event in any given year. Exhibit H-3 shows the results of 5,000 Monte Carlo simulations for the localized projection of EWL events for MSY.40 The top-most table shows both the historical probabilities and the projected mean annual outcomes for the future at 10-year intervals. The bottom part of the exhibit shows the assumptions for a generic mitigation project. The left side reports the assumed investment and operations and maintenance costs (including periodic restoration) and the right side reports the assumed costs to the airport of each extreme event with and without the mitiga- tion project. (These cost assumptions are completely generic and not based on any real data.) These values can be considered in relation to the conventional use of the 100-year storm as a metric for establishing building standards. Many localities and airports use the 100-year storm (based only on historical data) to set elevations for new building and for remediation of existing infrastructure, typically adding 1 to 3 ft of freeboard to the expected flood level to generically account for sea level rise. As shown in the Exhibit H-3, the 100-year event for MSY grows from about 6.4 ft based on historical data to well over 11 ft by 2095. These results show that using a conventional assumption of 3 ft of freeboard for infrastructure assets would provide a good chance of protection through at least 2065. Exhibit H-4 illustrates how the probability projections and cost data are combined to get results in the Monte Carlo model; each row is one of 5,000 simulations, with the columns being

Case Study Details 121 each year from 2020 to 2099. In this instance, the numbers shown in the table are the cost to the airport of extreme water events with the mitigation project. The implicit assumption in the sim- ulations is that the generic mitigation project does not fully eliminate the costs of all flood events. Notice that the costs increase over time as the probability of more extreme events increases. The main findings of the sample analysis are presented in Exhibit H-5, as expressed in a VaR graphic. The blue line is the range of probable outcomes without mitigation, and the red line is with mitigation. Each curve is composed of 5,000 possible outcomes for the airport (net costs to the airport expressed in NPVs). The box on the left side of the exhibit reports the average NPV and benefit–cost ratio for the project. In this generic example (using actual climate data), conventional decision making would suggest that the project is justified and should be pursued (absent capital constraints). The graphic is informative because it shows the distribution of potential outcomes based on the Monte Carlo simulations. It shows that 70% of the time, the project would pay off (benefit– cost ratio greater than 1), but if the airport does nothing, there is a 20% chance it would lose $40 million or more over the analysis period (expressed in today’s dollars). These findings could be relevant for both financial management and risk planning. Participant Feedback The airport participants appreciated the level of detail that went into the analysis and sug- gested that the approach could have some value as an adjunct to their consideration of future Exhibit H-2. Historical and future projections of extreme water events.

122 Climate Resilience and Benefit–Cost Analysis: A Handbook for Airports flood elevation maps. They stated that it could also be useful in master planning. While the airport representatives were focused on stormwater management at that time, they commented that the methodology could be useful in the analysis of any projects with long life spans such as runways, parking garages, terminals, and levees. They were also aware that the specific localized sea level projections used in the sample analysis might not be accurate for MSY due to their remote geographic location. Boston Similar to MSY, Logan Airport in Boston may be threatened by flooding from sea level rise or storm surge. The ACROS screen for BOS shows no expected days of flooding from sea level rise. However, ACROS does show an increase in base flood elevation, which could threaten some BOS infrastructure or access routes. Exhibit H-6 shows the localized projection of the probability of extreme water events for the airport, using the same data sources and methodology described for MSY.41 The exhibit also presents the same $5 million generic project to mitigate the effects of climate change. Exhibit H-3. Simulation summary and cost assumptions for MSY.

Case Study Details 123 Exhibit H-4. Monte Carlo simulations for MSY. It is interesting to compare the climate simulation results with more traditional analysis based on 100-year events. The 100-year event for BOS represents an EWL of about 4.6 ft based on historical data, and it grows to nearly 9 ft by 2095. These results show that using a conventional assumption of 3 ft of freeboard for infrastructure assets would provide a good chance of protec- tion through at least 2075. Exhibit H-7 shows the results of the VaR analysis for the generic mitigation project. The mean project shows marginal net benefits with a benefit–cost ratio of 1.02. However, the project would pay off only 35% of the time. There is a 10% chance of the airport losing at least $40 million (NPV). Participant Feedback BOS was already doing its own modeling of flood risks using the Boston Harbor Flood Model. It used that model to provide estimates for the Massport Flood Proofing Design Guide, which uses a 500-year event (probability 0.2%) and the “intermediate-high” GMSL rise scenario as baselines. Given this background, airport representatives had no trouble understanding the methodology and the added benefit of sampling across different uncertain future outcomes. They appreciated the fact that the Excel model allows the user to select from among different emissions scenarios. They also suggested that in addition to analyzing new mitigation projects, the methodology could also be used to summarize risk in terms of the probability of occurrence over the remaining life of existing assets.

124 Climate Resilience and Benefit–Cost Analysis: A Handbook for Airports Although the presentation did not directly discuss the potential use of inundation maps, the Boston participants suggested that linking such maps to flooding occurrence probabilities could be useful for their own internal illustration purposes. They also thought that a high-level hand- book would be useful for upper management. Phoenix One of the important climate issues faced by PHX is increased frequency of very high ambient temperature days. In the summer of 2017, there were about 50 regional jet operations that were cancelled when temperatures breached 118°F. Standard narrow-body jets would face similar cancellation issues at about 126°F (Wang 2017). The case study presented to the airport examined the number of days the airport would face extreme temperatures in the future (118°F for regional jets and 126°F for standard jets). For the purposes of the case study, in the base case it was assumed that flights would be delayed by an average of 3 hours during the middle part of the day. The mitigation project analyzed was a runway extension that would completely eliminate these delays. It is important to note that this is a simplified analysis where flights are delayed due to very high temperatures. A more realistic analysis would assess the impact of weight restrictions (which could begin to occur at much lower temperatures), where flights are not actually delayed Exhibit H-5. Sample VaR results for MSY.

Case Study Details 125 but are forced to offload passengers in order to take off. The Excel template developed for high temperatures directly addresses these impacts rather than assuming flight delays. The ACROS model shows an appreciable increase in hot days (90°F or more) and very hot days (100°F or more). Using ACROS as a screening tool suggests that the airport needs to prepare for increased frequency of such days, but the information is not precise enough to examine the problems that aircraft with current technology may face when temperatures exceed 118°F. The more detailed VaR evaluation for PHX uses the high-emissions RCP8.5 climate scenario. For each of the 5,000 simulations run for PHX, and for each year between 2020 and 2089, the model randomly selects from one of 31 different climate models, each of which makes daily projections of high temperatures throughout the selected period. For each year, one can simply count up how many days the selected model forecasts in excess of 118°F and 126°F. Exhibit H-8 summarizes the range of outcomes for 118°F days by decade. The dots in the graph show the median number of days each year. The lines represent the range of average number of annual days (by decade) produced by the Monte Carlo simulations. Notice that the uncertainty increases significantly the further into the future one looks. 0-1 Exhibit H-6. Simulation summary and cost assumptions for BOS.

126 Climate Resilience and Benefit–Cost Analysis: A Handbook for Airports It was assumed that the airport would be developing a benefit–cost study to support an LOI application for an AIP grant to extend its longest runway. The main purpose of the extension would be to prevent the cancellation/rescheduling of flights on days with very high tempera- tures.42 The life-cycle costs (discounted present value) of the extension were assumed to be $30 million. In the base case, each time temperatures reached 118°F, regional jet flights during the middle of the day incurred 3-hour delays. Standard jets experienced the same delays at 126°F. The FAA’s airport benefit–cost guidance methodology (FAA 1999b) and its Economic Values for FAA Investment and Regulatory Decisions (FAA 2016b) were used to value passenger delays, operator crew costs, and depreciation (the latter due to the aircraft not being productively used for 3 hours). Exhibit H-9 summarizes the assumptions used in the analysis example. The results of the VaR analysis are shown in Exhibit H-10. In this instance, the NPV of the project is negative (–$5.1 million), with only a 15% chance of being positive. There is a 3% chance that passengers and operators would lose $35 million (NPV) over the analysis period. Participant Feedback PHX personnel appreciated the level of detail that went into the methodology and thought the Monte Carlo analysis made sense and was well organized. However, they suggested that material presented to senior management would have to focus much less on the modeling and more on the primary takeaways from the summary results. There also appeared to be some sense that due Exhibit H-7. Sample VaR results for BOS.

Case Study Details 127 to the highly uncertain nature of future climate change, one would have to be careful not to put too much stock in the results from any one analysis. As for the specific analysis, airport representatives correctly noted that there will be payload/range restrictions that would become relevant at much lower temperatures than the very high ones used in the example. They also noted that high temperatures are also detrimen- tal to asphalt taxiways and ramps and to personnel working outside (which PHX is already addressing). Little Rock LIT has an 8,200-ft runway. Its longest commercial flight at the time of writing was to Los Angeles, but it had aspirations for longer flights. The airport indicated that it would be concerned if airlines faced frequent payload penalties for service to these destinations due to increasing high temperatures. The case study presented to the airport examined the potential impact of payload restrictions when daily high temperatures exceeded 100°F. The ACROS model shows a substantial increase in very hot days (more than 100°F) by 2060 for LIT. Thus, a more precise analysis for a possible runway extension could be warranted. Exhibit H-8. High-temperature projections at PHX.

128 Climate Resilience and Benefit–Cost Analysis: A Handbook for Airports Based on estimates from Airbus A320 and Boeing 737-800 aircraft performance charts, Exhibit H-11 shows estimated payload penalties in terms of passengers who would not be able to fly when temperatures breach 100°F or 110°F. It is important to note that these estimates of passenger reductions do not take into account the possibility that an airline might be able to offload cargo, which would reduce the number of passengers affected. In addition, these are manual, informal projections based on visual approx- imations taken from the aircraft performance charts, and they assume 100% load factors.43 The presentation assumed that LIT was developing a benefit–cost study to support an LOI application for an AIP grant to extend its runway. It was assumed that passengers would be delayed by an average of 6 hours if not able to fly due to weight restrictions.44 Unit costs to pas- sengers are the same as those assumed for PHX. The mitigation project being analyzed was a runway extension costing $30 million. The sampling of the 31 climate models used in the PHX analysis (assuming emission scenario RCP8.5) was repeated for LIT. The temperature projections are illustrated in Exhibit H-12, which shows the range of forecasts (by decade) for annual days at LIT exceeding 110°F. Again, the uncertainty increases the further into the future one looks. Assumptions Source Construction cost for 2,500-ft runway extension $21,875,000 Assumes $350/sq yard for 150-ft wide runway + 75-ft wide taxiway 20-year rehabilitation cost % construction cost 50% Assumed value Annual O&M expense % construction cost 3% Assumed value Affected Flights: 320 321 738 739 CR7 CR9 E75 Source Threshold Temperature (°F) Avg daily flights 1300-1759 in 2017 20.2 24.9 42.0 6.3 22.5 24.7 5.9 Avg block hrs per flight 2.6 2.7 3.0 2.7 1.4 1.4 2.1 Avg seatsize 170 187 158 175 63 84 75 Passenger Impacts: Avg load factor 83.0% 83.2% 81.0% 85.2% 86.4% 75.4% 84.8% FAA T-100 Domestic Segment report -- PHX load factors by eqpt type for May-Sep 2016 Avg daily pax per flt 141.1 155.6 128.0 149.1 54.4 63.3 63.6 = Avg seatsize * Avg load factor Avg hrs of delay per passenger 3.0 3.0 3.0 3.0 3.0 3.0 3.0 Assumed value Passenger delay cost per hr $44.30 $44.30 $44.30 $44.30 $44.30 $44.30 $44.30 FAA Economic Values, Table 1-1, All Purpose Intercity Air and High Speed Rail Total passenger delay cost at threshold in 2017 $378,794 $514,860 $714,359 $124,837 $162,765 $207,909 $49,869 = Avg daily flights in 2017 * pax per flt * hrs of delay per pax * delay cost per hr Total Passenger Impacts at threshold in 2017 Airline Impacts: Crew cost per block hr $777 $777 $724 $777 $349 $349 $349 Aircraft depreciation per block hr $352 $352 $221 $352 $144 $144 $144 PHX delay propagation multiplier 1.49 1.49 1.49 1.49 1.49 1.49 1.49 FAA Economic Values, Table 10-1 Total airline cost at threshold in 2017 $88,350 $113,095 $177,414 $28,614 $23,139 $25,401 $9,101 = Avg daily flights in 2017 * block hrs per flt * (crew costs + depreciation per block hr) * delay propagation multiplier Total Airline Impacts at threshold in 2017 Total Daily Impacts at threshold in 2017 PHX annual departure growth rate, 2017-2045 2.1% FAA TAF Forecast 2016, ITN_AC ops avg annual growth rate at PHX, 2017-2045 PHX annual departure growth rate, 2045-2079 1.0% Assumed value Official Airline Guide (OAG) -- based on May- Sep 2017, 1300-1759 hrs FAA Economic Values, Table 4-6 126 118 $631,636$1,986,872 $1,608,013 $545,380 $378,859 $86,256 Exhibit H-9. BCA assumptions for PHX.

Case Study Details 129 Exhibit H-10. Sample VaR results for PHX. Exhibit H-13 shows the results of the VaR analysis. The benefit–cost study showed nega- tive mean results (–$1.8 million NPV and benefit–cost ratio of 0.91). There is only about a 7% chance of the project paying off if it were built today. Large losses for passengers appear to be unlikely, even in the worst cases, if nothing is done (base case). Participant Feedback Although the participants from LIT were not familiar with ACROS, they found the summary projections informative and potentially indicative of a future issue regarding payload restric- tions. They specifically viewed payload penalties as an issue for air service development efforts involving service to other airports. Based on their engineering backgrounds, the participants found the logic of the Monte Carlo VaR analysis easy to follow and agreed that it might be useful to apply this type of analysis to justify AIP funding. They did noted that the wide variance in outcomes is similar in nature to the uncertainties of enplanement scenarios used in master planning. As a general matter, the airport personnel offered that they typically would depend on the airlines to identify needed infrastructure improvements to support long-haul flights.

130 Climate Resilience and Benefit–Cost Analysis: A Handbook for Airports Passengers Removed per Flight 100°F day 110°F day Airport Served Today Distance (nm) A320 737-8 A320 737-8 LAS Yes 1122 10 16 LAX Yes 1298 10 16 LGA No 943 3 BOS No 1095 9 14 SFO No 1467 24 44 SEA No 1552 28 51 Note: LAS = McCarran International Airport, LAX = Los Angeles International Airport; LGA = LaGuardia Airport; SFO = San Francisco International Airport; SEA = Seattle–Tacoma International Airport. Exhibit H-11. Estimated passenger payload restrictions at LIT. Exhibit H-12. High-temperature projections at LIT.

Case Study Details 131 Exhibit H-13. Sample VaR results for LIT.