The atmosphere carries a heavy burden of air pollution, with large contributions to the problem coming from the combustion of coal and petroleum-derived fuels. As a society, we need to evolve towards a more sustainable, environmentally benign approach to meeting growing demands for energy. My research group uses mathematical models and data from field experiments to help understand air pollution problems and related issues in atmospheric chemistry, climate change, and emission source characterization and control.
Some air pollutants are formed in situ from other precursor emissions by photochemical reactions in the atmosphere. Air pollution problems of this type, including tropospheric ozone and some components of airborne particulate matter, have complex relationships to precursor emissions. We use mathematical models to synthesize understanding of relevant processes that take place in the real atmosphere.
I am interested in development and use of diagnostic tools to assess source contributions to air pollutant concentrations, as there are typically multiple source types and regions that contribute to the problem. We quantify model sensitivity and uncertainty with respect to underlying processes and model input data (see Publication List, refs 14, 23, 41, 50-51, 53, 58, 65, 70). We use models to illuminate the reasons for observed atmospheric responses to changes in emissions that occur on various time scales ranging from diurnal to decadal (refs 15, 40, 51, 60).
An example of research on this topic is the paper by Martien and Harley (2006), Adjoint Sensitivity Analysis for a Three-Dimensional Photochemical Model: Application to Southern California. Environmental Science & Technology 40, 4200-4210.
Time Series Analysis
Analysis of measured pollutant concentrations provides a complementary perspective to model-based studies. Unfortunately, the signals that we seek to detect are often hard to separate from natural variability in the system that occurs from day to day and on seasonal time scales. Changes in air pollution observed on weekly and decadal time scales may be more readily linked to changes in emissions (refs 38-39, 47-48, 52, 56). We use receptor-based models together with online measurement methods to infer, for example, temperature effects on pollutant emissions (ref 49). This is important to understanding the role of day-to-day meteorological variability in affecting air pollution levels, and also in considering possible effects of climate change.
An example of research on this topic is the paper by Rubin et al. (2006), Temperature Dependence of Volatile Organic Compound Evaporative Emissions from Motor Vehicles. Journal of Geophysical Research 111, D03305, doi: 10.1029/2005JD006458.
See also Marr and Harley (2002), Spectral Analysis of Weekday-Weekend Differences in Ambient Ozone, Nitrogen Oxide, and Non-methane Hydrocarbon Time Series in California. Atmospheric Environment 36, 2327-2335.
The transportation sector involves movement of both passengers and freight. This sector currently relies on petroleum-derived fuels such as gasoline and diesel. My research group has made a series of field measurements at a California highway tunnel (Caldecott, hwy 24) that document emission trends over time and, in particular, the effects of improved emission control technologies and gasoline reformulation on vehicle emissions (refs 11, 20-21, 52, 55, 59). While there has been major progress in control of emissions from gasoline-powered cars and light trucks, efforts to control emissions from large trucks and off-road diesel engines have not yet advanced so far.
I have also been a lead developer of the "fuel-based" approach to estimating vehicle emissions, in which vehicle activity is measured by fuel consumption, and emission rates are expressed per unit of fuel burned rather than per km traveled or per unit time (refs 10, 15, 17, 22, 31, 33, 48, 67, 73). Emission rates for many pollutants (e.g., CO, NOx, as well as CO2 of course) vary less over wide ranges of vehicle weight and driving conditions when normalized to fuel consumption (ref 42). This line of research has contributed to policy-relevant revisions in national and state-level air pollution emission inventories.
The freight transport sector relies heavily on diesel fuel, in contrast to passenger travel where gasoline is the dominant fuel. My research group has been working to describe the amount of diesel-related air pollution (previous estimates were flawed in various ways), and the spatial and temporal patterns of those emissions (refs 16-17, 24-25, 31, 48, 55, 59, 62, 71, 72, 76). Advanced diesel emission controls such as continuously regenerating particle traps have been installed on new engines as standard equipment since 2007, and have been retrofit on some older engines as well. Control of diesel emissions in the freight transport sector continues to raise many challenging technical and policy questions.
Mobile lab used to measure Port truck emissions
We are now using the Port of Oakland as a laboratory for studying coming statewide and national changes in diesel engine emissions. Because of the large numbers of marine, locomotive, and heavy truck engines operating in and around the Port, there are extra concerns about exposure to diesel exhaust in the surrounding community. We reported on the effects of an early clean-up effort to replace or install exhaust filters on Port trucks (Dallmann et al., ref. 76). A non-technical summary providing more background on this issue and a summary of research findings is available: see Berkeley Transportation Letter (Winter 2012 edition).
A paper by Millstein and Harley (ref. 67) has influenced policy on the control of emissions from off-road diesel-powered construction equipment such as bulldozers and backhoes. The scale of impact is on the order of one billion dollars statewide in California. Previous emission estimates for nitrogen oxides (NOx) and exhaust particulate matter (PM) from this sector were too high (see our paper for details). Rules that would have required retrofit or replacement of older in-use construction equipment/engines have been revised, and will take effect more gradually. These policy changes followed from different counting methods being used to estimate emissions, and also due to the weak economy.