A method to represent ozone response to large changes in precursor emissions using high-order sensitivity analysis in photochemical models

摘要

Photochemical grid models (PGMs) are used to simulate tropospheric ozone andquantify its response to emission changes. PGMs are often applied for annualsimulations to provide both maximum concentrations for assessing compliancewith air quality standards and frequency distributions for assessing humanexposure. Efficient methods for computing ozone at different emission levelscan improve the quality of ozone air quality management efforts. This studydemonstrates the feasibility of using the decoupled direct method (DDM) tocalculate first- and second-order sensitivity of ozone to anthropogenic NO_xand VOC emissions in annual PGM simulations at continental scale. Algebraicmodels are developed that use Taylor series to produce complete annualfrequency distributions of hourly ozone at any location and anyanthropogenic emission level between zero and 100%, adjustedindependently for NO_x and VOC. We recommend computing the sensitivitycoefficients at the midpoint of the emissions range over which they areintended to be applied, in this case with 50% anthropogenic emissions.The algebraic model predictions can be improved by combining sensitivitycoefficients computed at 10 and 50% anthropogenic emissions. Comparedto brute force simulations, algebraic model predictions tend to be moreaccurate in summer than winter, at rural than urban locations, and with100% than zero anthropogenic emissions. Equations developed to combinesensitivity coefficients computed with 10 and 50% anthropogenicemissions are able to reproduce brute force simulation results with zero and100% anthropogenic emissions with a mean bias of less than 2 ppb and meanerror of less than 3 ppb averaged over 22 US cities.