System level assessment of uncertainty in aviation environmental policy impact analysis

Abstract

This thesis demonstrates the assessment of uncertainty of a simulation model at the system level, which takes into account the interaction between the modules that comprise the system. Results from this system level assessment process aid policy-makers by identifying the key drivers of uncertainty in model outputs, among the input factors of the various modules that comprise the system. This knowledge can help direct resource allocation for research to reduce the uncertainty in policy outputs. The assessment results can also identify input factors that, when treated as deterministic variables, will not significantly affect the output variability. The system level assessment process is demonstrated on a model that estimates the air quality impacts of aviation. The model comprises two modules: the Aviation Environmental Design Tool (AEDT), which simulates aircraft operations to estimate performance and emissions inventories, and the Aviation environmental Portfolio Management Tool (APMT)- Impacts Air Quality module, which estimates the health and welfare impacts associated with aviation emissions. Global sensitivity analysis is employed to quantify the contribution of uncertainty in each input factor to the variability of system outputs, which here are adult mortality rates and total health cost. The assessment results show that none of the input factors of AEDT contribute significantly to the variability of system outputs. Therefore, if uncertainty reduction in the estimation of adult mortality and total health cost is desired, future research efforts should be directed towards gaining more knowledge on the input factors of the APMT-Impacts Air Quality module. This thesis also demonstrates the application of system level assessment in policy impact analysis, where policy impact is defined as the incremental change between baseline and policy outputs. In such an analysis, it is important to ensure that the uncertainty in policy impacts only accounts for the uncertainty corresponding to the difference between baseline and policy scenarios. Some input factors have a common source of uncertainty between scenarios, in which case the same representation of uncertainty must be used. Other input factors, on the other hand, are assumed to have independent variability between the different scenarios, and therefore need to have independent representation of uncertainty. This thesis demonstrates uncertainty assessment of a technology infusion policy analysis.