Sandra L. Winkler

A Study of the Association between Daily Mortality and Ambient Air Pollutant Concentrations in Pittsburgh, Pennsylvania

ABSTRACT We have studied the possible association of daily mortality with ambient pollutant concentrations (PM10, CO, O3, SO2, NO2, and fine [PM2 5] and coarse PM) and weather variables (temperature and dew point) in the Pittsburgh, PA, area for two age groups—less than 75, and 75 and over—for the 3-year period of 1989-1991. Correlation functions among pollutant concentrations show important seasonal dependence, and this fact necessitates the use of seasonal models to better identify the link between ambient pollutant concentrations and daily mortality. An analysis of the seasonal model results for the younger-age group reveals significant multicollinearity problems among the highly correlated concentrations of PM10, CO, and NO2 (and O3 in spring and summer), and calls into question the rather consistent results of the single- and multi-pollutant non-seasonal models that show a significant positive association between PM10 and daily mortality. For the older-age group, dew point consistently shows a significant association with daily mortality in all models. Collinearity problems appear in the multi-pollutant seasonal and non-seasonal models such that a significant, positive PM10 coefficient is accompanied by a significant, negative coefficient of another ambient pollutant, and the identity of this other pollutant changes with season. The PM25 data set is half that of PM10. Identical-model runs for both data sets reveal instability in the pollutant coefficients, especially for the younger age group. The concern for the instability of the pollutant coefficients due to a small signal-to-noise ratio makes it impossible to ascertain credibly the relative associations of the fine- and coarse-particle modes with daily mortality. In this connection, we call for caution in the interpretation of model results for causal inference when the models use fully or partially estimated PM values to fill large data gaps.

Comparison of stiff chemistry solvers for air quality modeling.

Four fast solvers and their variations are compared in terms of the accuracy of the solutions and computation time when they are used to solve a system of stiff ordinary differential equations describing the carbon bond IV mechanism in air quality modeling. The solvers are the Urban Airshed Model (UAM) solver, the quasi-steadystate assumption (QSSA), and the Hybrid solvers, each with an additional version employing the steady-state algorithm of the UAM solver, and the new implicit-explicit hybrid (IEH) solver under two different sets of error tolerance. The solvers were run for one 6-min time step under 256 different initial conditions for both daytime and nighttime. In terms of accuracy, the IEH solvers are the most accurate, while the QSSA solver is the least accurate. In terms of computation time, QSSA and QSSA with UAM steady state are the fastest, while UAM is the slowest for daytime integration and widely variable for nighttime integration. The more tolerant version of IEH (IEW26) and the Hybrid solver coupled with UAM steady state are fast, but the former is more accurate under all teoted conditions, assumes only one steady-state species (OlD), conserves the nitrogen mass, and is applicable to other stiff chemical systems. Accordingly, IEH26 should be an excellent candidate as a fast and accurate chemistry solver in air quality modeling, combustion, and other reactive flow systems. Hybrid schemes were recently carried out by Odman et al. (6) using a limited number of initial conditions and an integrating period of up to 3 days. This paper compares the performance of the above four solvers, each with one or two variations, in terms of their computation time and accuracy of the results. Rather than running the solvers for a long time period, only one time step will be used in a similar fashion as in 3D simulation. In addition, a wide range of initial conditions will be applied to assess the robustness of the solvers. Such a test allows one to explore, without the interference of extraneous factors, the performance of the solvers in greater detail under a wide range of conditions encountered in the full air quality model. Adoption of the IEH solver in the full UAM will be reported elsewhere. The solvers, test cases, and results are described in sections 2-4, respectively. Section 5 gives the conclusion.

A particle grid air quality modeling approach: 2. Coupling with chemistry

The particle grid method is applied to a system of 10 reacting chemical species in a two-dimensional rotating flow field with and without diffusion. Two types of chemistry grids are used to describe the chemical reactions: a fixed coarse grid and a moving (the advection case) or stationary (the advection plus diffusion case) fine grid. Two particle number densities are also used: 256 and 576 particles per fixed coarse grid cell. The species mass redistributed back to the particle after each reaction step is assumed to be proportional to the species mass in the particle before the reaction. The simulation results are very accurate, especially in the advection chemistry case. Accuracy improves with the use of a fine grid. A higher particle number density also reduces the concentration fluctuation in the cases involving diffusion. We also show by examples that chemistry can lead to significantly different results from numerical methods for the diffusion equation (let alone the advection equation) which otherwise yield almost identical solutions. The absence of this difficulty in the particle grid method further enhances its attractiveness.

A photochemical extent parameter to aid ozone air quality management

Abstract Interest in semiempirical approaches for relating precursor emissions (VOC and NOx to ozone air quality has been renewed recently. Of particular interest is a semi-empirical approach [called Smog Production Model (SPM) or SP Algorithm] initiated by Johnson and coworkers. SPM defines a photochemical extent parameter, which provides the directional guidance on the relative effectiveness of NOx and VOC controls in reducing ozone levels. In the present paper, a modified version of SPM is introduced, and the parameters involved in SPM are evaluated using urban airshed model (UAM) simulation results. UAM results are also used to examine relationships of extent parameters to ozone reductions resulting from VOC or NOx emission reductions. A modified version of SPM is applied to ambient air quality data. With improved quantified model parameter values and improved measurements of NOy, SPM may provide important information on the relative effectiveness of precursor controls. An attractive feature of SPM is that the required data are simultaneously measured NOx NOy and ozone.

A comparison of advection algorithms coupled with chemistry

Abstract Three-dimensional air quality models face the problem of solving the advection or advection-dominated transport equation. The performance of six advection algorithms coupled with chemistry is assessed here. The six algorithms are the forward-Euler Taylor-Galerkin (FETG) method, the implicit Chapeau-function method, a streamline upwind Petrov-Galerkin (SUPG) method, Smolarkiewiczs method, a semi-Lagrangian (SemiLag6) method, and the accurate space-derivative (ASD) method coupled with periodicity recovery. These methods are coupled with a 10-step chemistry involving 10 species simulating the essence of atmospheric photochemical reactions. The flow field is a two-dimensional, non-divergent rotating velocity field. The domain has 33 × 33 grid cells. Two test cases defined by the initial conditions are considered: (1) non-zero background concentrations for all species except four that have an additional conic profile; (2) same as (1) for a region that covers 11 × 11 grid cells with the conic profile at its center, while the concentrations of all species outside the 11 × 11 concentration platform are set equal to zero. After one full rotation in 24 h, for test case (1), the ASD gives the most accurate results, followed by FETG and the chapeau-function method. The Smolarkiewicz method gives the least-accurate result. For test case (2), no method performs well for NO and OH. For most species, ASD and FETG still yield accurate results. To assure highly accurate advection results for all concentration profiles, however, one may have to resort to a particle-trajectory method such as the particle-grid modelling approach.

Light-duty vehicle CO2 targets consistent with 450 ppm CO2 stabilization.

We present a global analysis of CO2 emission reductions from the light-duty vehicle (LDV) fleet consistent with stabilization of atmospheric CO2 concentration at 450 ppm. The CO2 emission reductions are described by g CO2/km emission targets for average new light-duty vehicles on a tank-to-wheel basis between 2010 and 2050 that we call CO2 glide paths. The analysis accounts for growth of the vehicle fleet, changing patterns in driving distance, regional availability of biofuels, and the changing composition of fossil fuels. New light-duty vehicle fuel economy and CO2 regulations in the U.S. through 2025 and in the EU through 2020 are broadly consistent with the CO2 glide paths. The glide path is at the upper end of the discussed 2025 EU range of 68-78 g CO2/km. The proposed China regulation for 2020 is more stringent than the glide path, while the 2017 Brazil regulation is less stringent. Existing regulations through 2025 are broadly consistent with the light-duty vehicle sector contributing to stabilizing CO2 at approximately 450 ppm. The glide paths provide long-term guidance for LDV powertrain/fuel development.

Trajectory-grid: An accurate sign-preserving advection-diffusion approach for air quality modeling

Abstract We propose a new method, called the trajectory-grid (T-G) approach, which contains a fully Lagrangian advection scheme coupled with an Eulerian diffusion scheme, to remove the numerical problems associated with the advection equation in air quality modeling. This method is sign preserving, mass conserving, and very accurate. The method assigns the spatial locations of points on a given concentration profile to a set of concentration pulses; then tracks the pulse positions as they move downwind undergoing diffusion and reactions. Multigrid nesting can be conveniently accounted for by increasing the number density of concentration pulses in given areas in the modeling domain. The method can also serve as a tool to test the validity of the observation-based models. While the execution time may be comparable to that of the Accurate Space Derivative scheme, the advection step of the method is highly amenable to multiply-parallel processing.

A trajectory-grid approach for solving the condensation and evaporation equations of aerosols

Abstract The trajectory-grid method (Chock et al., Atmospheric Environment 30 (1996) 857–868) for solving the transport equation in a grid model can be easily adapted to solve the condensation/evaporation equations of internally mixed multicomponent aerosols based on the sectional approach. Compared to the Bott scheme (Monthly Weather Review 117 (1998) 1006–1015, 2633–2636) often used for solving these equations, the method is significantly more accurate, of comparable speed, but gaining in speed as the number of aerosol components increases. Most notably, there is no mismatch in size change of the different components of the same aerosol. The method is quite flexible, can handle an abrupt change in aerosol size due to evaporation of fog particles, and is ideal for multiply-parallel processors.

URBAN OZONE AIR QUALITY IMPACT OF EMISSIONS FROM VEHICLES USING REFORMULATED GASOLINES AND M85

The urban ozone air quality impact of exhaust emissions from vehicles using reformulated gasolines and flexible/variable-fuel vehicles using M85 has been studied using emissions data from the Auto/Oil Air Quality Improvement Research Program and a single-cell trajectory air quality model with two different chemical mechanisms (the updated version of Carbon-Bond-IV (CB4) and the LCC mechanisms). Peak ozone concentrations are predicted for each fuel for all combinations of the following ambient conditions: low and high atmospheric dilution or mixing height, four NMOG/NOx ratios, two each of the initial NMOG concentration, the vehicular contribution to the ambient air, and the NMOG composition of the initial ambient mixture. The ozone impact of a fuel depends strongly on the atmospheric dilution and NMOG/NOx ratio of an area. The differences in ozone impact among fuels are limited under the condition of high atmospheric dilution and a high NMOG/NOx ratio. The ozone-forming potentials (OFPs) for the exhaust emissions based on the maximum incremental reactivities (MIRs) for various fuels are generally well correlated with model-calculated peak ozone levels at a low NMOG/NOx ratio. These OFPs can serve to separate out fuels with rather different reactivities, but not fuels with comparable reactivities. Model-calculated ozone levels for various fuels based on CB4 and LCC mechanisms are relatively well correlated at low NMOG/NOx ratios, but much less so at higher ratios. Fuels with a high aromatic content, including high-toluene fuels, tend to be ranked more favorably by CB4 than by LCC. On the other hand, M85 is ranked more favorably by LCC than by CB4. Fuels with a low 90% boiling point and a low content on aromatics and olefins are generally less reactive. M85 would be an attractive fuel if the formaldehyde emissions could be curtailed significantly. (A)

Effect of grid resolution and subgrid assumptions on the model prediction of a reactive bouyant plume under convective conditions

Abstract We have introduced a new and elaborate approach to understand the impact of grid resolution and subgrid chemistry assumption on the grid-model prediction of species concentrations for a system with highly non-homogeneous chemistry—a reactive buoyant plume immediately downwind of the stack in a convective boundary layer. The Parcel-Grid approach was used to describe both the air parcel turbulent transport and chemistry. This approach allows an identical transport process for all simulations. It also allows a description of subgrid chemistry. The ambient and plume parcel transport follows the description of Luhar and Britter (Atmos. Environ. 23 (1989) 1911, 26A (1992) 1283). The chemistry follows that of the Carbon-Bond mechanism. Three different grid sizes were considered: fine, medium and coarse, together with three different subgrid chemistry assumptions: micro-scale or individual parcel, tagged-parcel (plume and ambient parcels treated separately), and untagged-parcel (plume and ambient parcels treated indiscriminately). Reducing the subgrid information is not necessarily similar to increasing the model grid size. In our example, increasing the grid size leads to a reduction in the suppression of ozone in the presence of a high-NOx stack plume, and a reduction in the effectiveness of the NOx-inhibition effect. On the other hand, reducing the subgrid information (by using the untagged-parcel assumption) leads to an increase in ozone reduction and an enhancement of the NOx-inhibition effect insofar as the ozone extremum is concerned.