The ACP numerical modeling and theoretical studies group has broad based expertise in a range of complementary subjects. The group maintains, develops, tests, and applies a full suite of models, measurement techniques, and field campaigns on many scales to address the effect of energy-related activities on the earth's chemical climatology. Our research addresses topics from the molecular level to regional and global scales.
In particular, the ACP numerical modeling and theoretical studies group's capabilities include:
€ Aerosol properties
€ Boundary layer meteorology and subgrid scale processes
€ Field campaigns/parameterizations/interpretations
€ Radiative forcing
€ Sensitivity/uncertainty analysis
€ Stratospheric chemistry (gas phase, heterogeneous)
€ Tropospheric chemistry (gas phase, heterogeneous on scales of observation based, regional, and global (2D and 3D))
Our recent progress is summarized below and presented in detail in Appendix A. Work by this group has included characterizing aerosol properties (e.g. intermolecular potentials between spherical particles, aspects of heterogeneous condensation, nucleation, etc.). Other advances include development of a new dry deposition parameterization and studies of wet scavenging of nitrogen and sulfur species during cloud formation. A global effort to develop the IGAC emission databases has also involved DOE participants.
Field studies and complementary modeling analysis of both NARE (North Atlantic Regional Experiment) and Nashville/Middle Tennessee Ozone Study have advanced our understanding of tropospheric chemistry. The tropospheric models are now starting to simulate not only gas phase, but also heterogeneous chemistry. Regional tropospheric models have also been used to study the effect of UV-B fluxes on O3 calculations.
The characterization of stratospheric models and subsequent analysis have addressed the effects of Mt. Pinatubo, the chemical response of increased UV-radiation from stratospheric O3 depletion, and the effect of increased temperature and moisture associated with global warming. The radiative forcing due to perturbations from high speed civil transport emissions round out this area. Several sensitivity analysis studies show that we have several viable techniques which we are now beginning to apply to both regional and global models.
Figure 1 shows areas of collaborations between scientists within the modeling group and between the modeling group and other ACP groups. During the December, 1995 ACP meeting, our sessions focused on updating each other as to our progress and identifying areas of potential collaboration and future work. Below we describe the proposed future collaborative efforts, and indicate which subgroups would participate (denoted by bold text). These projects will strengthen both our individual efforts as well as a unified DOE effort.
Proposed future collaborations and future plans:
(NOTE: Proposals (1) and (2) received the most discussion time, and are presented in the most detail).
(1) Identify key variables in ozone sensitivity
(Models [observation based, 2D, 3D, regional and global]<-->Sensitivity analysis)
The goal for project (1) is to couple modern sensitivity analysis with three dimensional global tropospheric models (and perhaps models of other scales, also). The group will then investigate ozone sensitivities at midlatitudes in the northern hemisphere by performing the analysis at the surface, mid-troposphere, and upper troposphere for several different latitudes. This will allow us to identify the key variables in the ozone sensitivity.
(2) Compare model results with measurements such as NARE, SOS, NARSTO-NE, ACE-2, etc.
(Models<-->Field campaigns).
We will use these results to evaluate how well models (observation based, 2D, 3D, global and regional) can evaluate quantities of interest, and the sensitivities of models to rate constants, input parameters, etc. We will also identify what quantities should be measured in field campaigns, and with what accuracy. This will allow us to evaluate the relative effects of emissions of NOx, hydrocarbons, and other important precursors on quantities such as O3, OH, HNO3, etc.
(3) Develop and incorporate new dry deposition techniques into models
(Boundary layer meteorology/subgrid scale processes<-->Models)
(4) Study 2D and 3D sensitivities due to perturbations
(Models<-->Models)
(5) Identify and study UV-flux O3 effects
(Models<-->UV-B group)
(6) Model intercomparisons and comparisons to observations
(Models<-->Models<-->Field campaigns)
(7) Identify important heterogeneous reactions and incorporate them into models
(Laboratory measurements<-->Models<-->Field campaigns<-->Aerosol properties)
€ Aerosol Properties
Marlow
The long-range intermolecular potential between pairs of spherical particles has been calculated as an essential component of studies of irregular aerosol particle aggregation rates. This calculation utilizes recent work from this laboratory which for the first time gives the van der Waals energy at very close approach including contact without assumptions of an equilibrium separation at contact. This work should also find application in colloidal science important for clean-up studies and elsewhere in addition to air chemistry. In another study, the interaction energy of water on a nanometer substrate has for the first time been computed and compared with the energy on a bulk substrate of the same composition. Those results show a significant decrease in the energy well depth for the nanometer particles due to the small size of the particle. Quantitatively accounting for heterogeneous condensation on insoluble substrates composed of the finest atmospheric particles is the long-range objective of this work.
Schwartz
Calculations of the rate of binary nucleation in sulfuric acid - water mixtures indicate substantial flux across the free-energy ridge at high supersaturations, rather than through the free-energy minimum as previously assumed, substantially increasing nucleation rates. Examination of nucleation using an exact expression for the evaporation rate of individual clusters rather than the customary canonical average over an equilibrium ensemble shows considerable reduction in evaporation rates, with consequent increase in nucleation rate.
€ Boundary Layer Meteorology/Subgrid scale Processes
Wesely, Doskey, and Gao
A significant accomplishment during calendar year 1995 was the development and testing of methods to use surface spectral reflectances remotely sensed by environmental satellites to improve parameterization of surface processes in modeling dry deposition over large geographical regions. This work, described in three peer-reviewed articles published in 1995, showed that the Argonne dry deposition module can be effectively coupled with information from satellites to provide more realistic estimates of surface uptake rates of trace chemicals; rates are based on observed surface conditions rather than tied only to fixed land use and seasonal categories. Highlights of efforts in 1996 will be experimental evaluation of the rates of NO emission rates from soils in a cooperative effort with the U.S. Environmental Protection Agency, use of the resulting data to test and refine an existing model combining turbulence and fast chemical reactions involving NOx and O3 in the lower atmosphere, development of methods to model NO soil emission rates over regional scales, and use of experimental data to improve parameterization of the dry deposition rate of PAN.
Schwartz
Examination of scavenging of nitrate and HNO3 during cloud formation shows significant differences versus the sulfate/SO2 system because of different solubilities and oxidation kinetics, for example nitrate being in larger drops than sulfate. Nonetheless the main solute mass is present in the main water mass, as for sulfate.
€ Field Campaigns and Observational based modeling
Berkowitz et al.
Several detailed investigations evaluating the role of boundary layer processes on resulting chemical profiles have been pursued at PNNL. The stratification of ozone over the western North Atlantic was the subject of several investigations, using a variety of observations to evaluate model performance. Much of the layering observed during the 1992 NARE campaign can be explained through natural mixing processes. Frequency distributions of parcel age as a function of height were developed for the 1993 aircraft observations and found to be in good agreement with a limited set of field observations evaluating plume aging. Another study provided a detailed analysis of the mixing and turbulent transport of plumes sampled during the 1993 campaign, relating many of the observations to the convective boundary layer of major urban areas. A number of questions relating to accepted ways of evaluating source regions through back-trajectories were raised during these studies and addressed in another analysis.
Kleinman et al.
Properties of low and high NOx atmospheric states and reasons for the existence of these two qualitatively different modes of tropospheric photochemistry were determined by constructing a transparently simple photochemical model of radical sources and sinks. Connections to multiple steady states, and the problem of NOx versus hydrocarbon control of O3 were explored.
Observation based photochemical calculations have yielded a mechanistic description of O3 production in air masses that are advected eastward from North America to the North Atlantic Ocean. These calculations are based on concentrations of O3, NO, CO, and hydrocarbons that were measured during the 1993 NARE experiment conducted in southern Nova Scotia. Although air masses in that region can be very polluted, calculations indicate only a small potential for further increases in O3 concentration.
In FY 96, observation based calculations will be performed using data collected from the ACP G-1 aircraft during the 1995 Nashville/Middle Tennessee Ozone Study. The focus of these calculations will be on determining O3 production rates and their sensitivity to NOx and hydrocarbon emissions.
€ Radiative forcing
Grossman
The tropospheric radiative forcing has been calculated for ozone and water vapor perturbations caused by a realistic High Speed Civil Transport aircraft emission scenario. Atmospheric profiles of ozone and water vapor were obtained using the LLNL 2-D chemical-radiative-transport model of the global stratosphere and troposphere. Radiative forcing calculations were made using the LLNL correlated k-distribution IR radiative transfer model and the LLNL two stream solar radiative transfer model. Major results were that the radiative forcing obtained for this aircraft scenario is too small to cause serious greenhouse effects on climate. Work has started on a new 3-D LLNL chemistry-transport model which includes both the stratosphere and the troposphere and which utilizes the LLNL capability for parallel process computation. This model will be used to compare 2-D vs. 3-D results for ozone change scenarios.
Wang
We used atmospheric models to evaluate the indirect effects on radiative forcing due to increasing atmospheric CH4. Specifically, we study chemical interactions involving tropospheric O3 and CH4. Because the tropospheric chemistry is sensitive to OH and thus the moisture content, we have also studied these effects when increases in temperature and moisture associated with the enhanced greenhouse effect due to increasing atmospheric CO2, N2O, CH4, and CFCs are considered.
The results suggest that tropospheric O3 can be substantially increased in response to direct CH4 increase, which is, however, in sharp contrast to the O3 decrease when the global warming effect is included in the CH4 increase calculations. In addition, CH4 increase can also be substantially reduced when enhanced OH associated with water vapor increases is taken into account. The decreases in O3 and CH4 provide a negative feedback, about 10%, to the total radiative forcing attributed to all the greenhouse gases. Results are also presented concerning the potential effect of CH4 increase to stratospheric water vapor and its subsequent climate implications.
€ Sensitivity/Uncertainty Analysis
Carmichael
The concept of automatic differentiation was applied to tropospheric chemistry. A box model chemistry study was conducted using automatic differentiation to calculate sensitivities to initial conditions and rate constants. The application study examined all of the IPCC chemistry scenarios, and the subsequent paper was submitted to Atmospheric Environment.
One result of note was that sensitivity calculations put added emphasis on chemical integrators. This led to the development of new methods, with results in papers submitted to J. Comp. Physics and SIAM J. Scientific Computations.
NOTE: See also tropospheric chemistry section
Shorter and Rabitz
We have developed the new Guided Monte Carlo (GMC) technique. It provides an efficient and accurate method for calculating model output probability distributions, from which output uncertainties and potential risks can be assessed. This technique makes Monte Carlo analysis on large complex models possible. Using the GMC, we calculated the model output uncertainty of ozone predictions due to chemical reaction rate and photolysis coefficient uncertainties for the region of 45 N latitude between 20-40 km altitude. The analysis was performed using a series of 0-D box models and the ozone prediction uncertainties were on the order of 30%. We transferred the technology of the GMC to a NASA program. They have implemented the GMC technique in their GSFC 2-D atmospheric chemistry model and presented results from the work at the 1995 Fall AGU meeting. We presented the GMC technique and the uncertainty results at the 1995 Fall AGU meeting and we are preparing papers for journal submission.
Smith
We have adapted the Sandia Senkin 0-D direct sensitivity analysis code and LLNL 2-D atmospheric photochemistry model to perform localized box model sensitivity calculations. Our ozone sensitivity-uncertainty analysis has examined latitudes 2N, 32N, and higher altitudes at 47N thus far. Methods were developed and explored to deal with species drifts in the computations which are traceable to transport of species and chemistry from nearby boxes and times.
We have begun to explore several issues using this sensitivity-uncertainty analysis tool:
o Identifying rate processes and measurements where improvements would reduce ozone model uncertainty.
o Examining effects of aerosol loading and reactions on ozone-destroying radicals.
o Computing the uncertainty of local ozone difference predictions for a HSCT perturbation scenario.
o Calculating sensitivities and uncertainties for observations not well matched by current models, such as high altitude ozone, HO2/OH, and ClO/HCl ratios.
€ Stratospheric Chemistry
Brasseur and Tie
1. Impact of Mt Pinatubo's eruption on stratospheric ozone
We have used a coupled chemical radiative, dynamical and microphysical 2-D model to assess the response of stratospheric ozone to the injection of sulfur following the eruption of Mt Pinatubo. Model calculation suggests that, during the first year (July/1991 to June/1992) following the volcanic eruption, the observed changes in the ozone amount integrated between 65oS and 65oN have been caused primarily by changes in the meridional circulation (associated with heating by the volcanic cloud in the tropics) and in the photolysis rate of molecules such as ozone (associated with backscattering of light by the cloud). During the second year after the eruption, as the aerosol has been dispersed at all latitudes and, in particular, has reached the polar region, the largest contribution to ozone reduction results from the heterogeneous chemical conversion of N2O5 and ClONO2 on the surface of the aerosol particles. Model calculations also suggest that the ozone decrease observed a few years after the eruptions of Mt. Pinatubo and El Chichon may have been unique in the Earth's history, and is directly linked to the emission in the atmosphere of industrially manufactured chlorofluorocarbons. For chlorine loadings typical of the pre-1980 period, the ozone column abundance should have increased after a large volcanic eruption. After 1980, as a result of growth in chlorine loading, the response of ozone became negative in winter at mid- and high latitudes. In the future, the response of ozone is expected to become positive again, if the production of chlorofluorocarbons is sufficiently reduced.
2. Three-dimensional model of chemical constituents in the stratosphere
We have developed a global three-dimensional transport/chemical model of the stratosphere, which includes a representation of the formation of Polar Stratospheric Clouds (PSCs) and detailed heterogeneous reactions on the surface of PSCs and sulfate aerosols. The formation of the observed springtime "Antarctic ozone hole" is well reproduced by the model. A maximum of 40% total ozone depletion occurs in October. Calculated ozone and chlorine concentrations are consistent with satellite observations. After the breakdown of the polar vortex in December, air with depleted ozone is transported to mid-latitudes in the Southern Hemisphere, resulting in a 2-4% ozone decrease at 50oS in December and a 1% decrease in the tropics. Ozone-poor airmasses are also transported to the troposphere, and produce a significant decrease (20-30%) in upper tropospheric ozone. Ice particles (Type II PSCs) sediment into the troposphere, producing a large decrease in the concentrations of stratospheric HNO3 and NO2. As a result, the conversion of ClO into ClONO2 after the evaporation of PSCs is reduced. The model shows that this process could enhance the ozone depletion by 20% during the month of October. The model is also able to reproduce the ozone minimum observed in Antarctica when the chlorine loading was as low as 0.6 pptv. Under these conditions, the polar ozone depletion caused by chemical processes is very small (maximum of 3%) in October. In November, the ozone concentration even increases above 22 km in response to PSC processes.
Grossman
Work has started on a new 3-D LLNL chemistry-transport model which includes both the stratosphere and the troposphere and which utilizes the LLNL capability for parallel process computation. This model will be used to compare 2-D vs. 3-D results for ozone change scenarios.
€ Tropospheric Chemistry (gas phase and heterogeneous)
Brasseur and Tie
We have implemented in our 2-D model a chemical scheme describing more accurately the methane oxidation chain in the troposphere, and have estimated the change in the tropospheric ozone abundance due to a doubling in the methane concentration. We have also completed the development of the IMAGES model, which describes the three-dimensional distribution of approximately 50 chemical compounds from the surface to the 50 mbar level. The model has been used to assess the impact of aircraft emissions on tropospheric ozone and to investigate potential causes for the recently observed decrease in CO abundances. Most recently, we have completed the development of a global three-dimensional model (MOZART) which simulates the distribution of ozone and its precursors in the troposphere and lower stratosphere. The model, which includes approximately 40 chemical species and 120 chemical and photochemical reactions, is driven by winds and temperatures provided by the NCAR Community Climate Model (CCM-2). The spatial resolution is 2.8 degrees in longitude and latitude, with 18 levels in the vertical from the surface to 1 mb. Boundary layer exchanges, cloud convection, and aqueous phase chemistry are included in the model. Surface emissions are based on pre-established inventories.
Berkowitz and Chapman
The Mass Transfer with Chemical Reaction Model (MTCRM-I) has been developed with detailed representations of gas-phase photochemistry, aqueous-phase chemistry and interphase mass transfer. Initial studies have simulated heterogeneous chemical processes associated with clouds and their impact on photochemistry. A significant reduction in gas-phase pollutant concentrations (9-34%) was found in the presence of heterogeneous processes and aqueous-phase chemistry under typical atmospheric conditions. Cloud water content and droplet size were determined to be the most important variables in determining the effects of the heterogeneous processes and the aqueous phase chemistry.
Work continues on global tropospheric aerosol modeling. The aerosol model, based on PNNL's global chemistry model, has been successfully coupled in core with PNNL's version of the NCAR CCM2 general circulation model. By adding the assimilation of observed winds and temperature via a nudging technique, the CCM2 is able to simulate observed meteorology. Several month-long simulations are underway, and the simulations will be evaluated against available observations using an evaluation protocol developed in the project. Direct radiative forcing by anthropogenic aerosols, and associated uncertainties, will be quantified using aerosol model simulations and model evaluation results.
Carmichael
A 3D regional model of O3 production was used to examine the influence of UV-B changes on predicted quantities. A detailed radiation calculation was coupled "on-line" with the 3D model. By examining a variety of perturbations, it was shown that important 3D effects aren't present in box model results. A paper is in preparation.
Penner
We have further developed and applied a global, three dimensional tropospheric chemistry model. The global model has been modified to account for the latest estimates of emissions of hydrocarbons, CO, and non-methane hydrocarbons from biomass burning as well as CO emissions from fossil fuel use. In addition, we have evaluated the model's capability to reproduce observed O3/CO and O3/(NOy-NOx) ratios by comparison with values for O3/CO measured off of Nova Scotia during the NARE campaign, and values for O3/(NOy-NOx) measured at a variety of places. These results are presented in papers submitted to Atmospheric Environment and the Journal of Geophysical Research. We have begun to develop a numerical treatment for the shorter-lived NMHCs through the development of a sub-grid scale treatment of plumes associated with urban emissions.
Schwartz
A global, seasonal inventory of anthropogenic emissions of sulfur and nitrogen oxides has been compiled on a 1x1 grid. This inventory is being extended to explicitly represent emissions from volcanoes.
A subhemispheric Eulerian model for tropospheric sulfate driven by operational meteorological data has been developed and applied to the North Atlantic and adjacent continental regions for four seasonal months in 1986-1987. Daily-average modeled concentrations accurately represent the magnitudes and spatial and temporal variability of measured surface concentrations. One journal article describing the model and results for October-November 1986 has been published. A second article for three other seasonal months is in preparation. Substantial contribution due to sulfate imported into the model domain is indicated, necessitating extension of the model domain to the entire hemisphere.
Wang
A 2-D tropospheric chemical-transport model was used to study the chemical response to: (1) increased UV radiation from stratospheric ozone depletion and (2) increased temperature and moisture associated with global warming as simulated from our 3-D global climate model. Increased UV radiation increases the photolysis rates for several tropospheric gases, in particular ozone. This leads to enhanced levels of odd hydrogen and reduced concentrations of tropospheric ozone. Increases in temperature and moisture reduce the levels of tropospheric ozone through increased production of odd hydrogen and temperature dependent reaction rates. In both cases, the methane levels are also reduced. The considered mechanisms thus provide negative feedback effects to global warming. Study of the effect of NOx emissions on tropospheric ozone and it subsequent climatic implications was also conducted. These studies were documented in several journal publications. Next, we plan to contrast the above results using the 3-D model. In addition, we will also begin to incorporate the interactive chemistry into the global climate model.