Uncertainty assessment transport policy modeling on the conversion of vehicles from petroleum to compressed natural gas

Abstract

In order to control the vehicular emission-induced air-pollution and consequent health hazards in Dhaka, recently one major policy initiative was taken by Bangladesh government to switch to a better alternative fuel - Compressed Natural Gas (CNG), from the conventional diesel and/or gasoline fuels. CNG is an attractive alternate automobile fuel primarily due to its less particulate emissions performance. However, CNG conversion can have implications on climate changes through emissions of well-identified green-house gases (carbon-di-oxide, methane) and aerosols (black carbon, organic carbon and sulfur-di-oxide). Therefore, the evaluation of the true impacts of such a wide-scale transport policy requires a comprehensive model. Uncertainty assessment is an integral part of such comprehensive policy-impact assessing models to support the decision-making processes. It is a study of communicating the model results with the complex combination of uncertainty and sensitivity analyses. This research proposes an overall precise framework for evaluation of the stated transport policy impacts by including uncertainty assessment as an important analyzing tool. The policy is being analyzed for two major impacts – urban-air quality and climate impacts. Following impact-pathway approach, a model is developed in programming language C++ to determine health benefits in monetary terms from reduced PM2.5 emissions resulting from the policy. Grid-based vehicular emission of PM2.5 for Dhaka city is estimated over Dhaka City Corporation (DCC) and greater Dhaka (GD) region. The corresponding concentrations are estimated using grid based source-receptor matrix (SRM) recently developed for Dhaka. Climate impacts are quantified by climate model through estimating the changes in emissions of the relevant species which affect the overall climate balance by contributing to global warming and/or cooling processes. To communicate the policy model results with uncertainty studies, an approach of seven-step methodology has been formulated. Uncertainties in model factors are represented with sampling-based probabilistic approaches. Uncertainty analysis is conducted by Monte-Carlo simulation method that involves random sampling from the distribution of inputs and successive model runs until a statistically significant distribution of outputs is obtained. 5000 random numbers are generated corresponding to the continuous probability distributions assigned to each uncertain input factor. Without the consideration of uncertainty, urban-air quality model gives total health benefits of USD 937 million over DCC and USD 1134 million over GD grids (13.45 and 16.28 million BDT respectively, 2010 prices) accrued from the policy. The climate model estimates total increase in emissions of about 941,000 tons/year and a climate cost of about USD 42 million (about 6,03,000 BDT) due to policy. With the inclusion of uncertainty analysis, the mean health benefits is obtained as about USD 1227 million with 95% confidence interval of USD (1213-1241) million (17.41-17.82 million BDT) over DCC. The corresponding values for GD are about USD 1490 million, USD (1473-1506) million respectively or 21.4, (21.15- 21.62) million BDT. The mean climate cost accrued from the policy is about USD 26 million (3,73,295 BDT) resulting from a mean change (increase) in global emissions of about 592,000 tons/year. Sensitivity studies ascertain most-priority transport-specific factor as PM2.5 emission factor from gasoline cars for air-quality model. For the climate model input factors, the resource allocation priority order is obtained as emission factors of methane followed by the annual vehicle activity, black carbon and carbon-di-oxide emission factors from specific vehicle-fuel combinations.