-- Jack A. Kaye (jkaye@hq.nasa.gov), Office of Earth Science, NASA Headquarters, Washington, DC
-- David Rind (drind@giss.nasa.gov), NASA/GISS, New York, NY
A meeting of representatives of instrument teams and interdisciplinary investigators in the Earth Observing System (EOS) program, as well as scientists in other NASA research programs, working in the areas of global modeling and analysis of tropospheric chemistry and tropospheric aerosols was held at the Goddard Institute for Space Studies (GISS) in New York City on May 14-15, 1998. The meeting brought together some 40 investigators, including many added to the EOS program in 1996 following the NASA Research Announcement (NRA) soliciting additional interdisciplinary investigations issued that year. The primary purposes of the meeting were to bring together the tropospheric chemistry and aerosol modeling communities supported by NASA (regardless of which program provided their support) and to ensure that scientists doing both global-scale modeling and space-based measurements on these quantities were aware of all the work supported by NASA in these areas. David Rind of GISS was the chair for the meeting; Jack Kaye of NASA HQ worked with him in setting up the meeting.
The meeting began with Yoram Kaufman of NASA/GSFC, Project Scientist for the first EOS spacecraft (EOS AM, launch scheduled no earlier than December, 1998) describing aerosol data that would come from two instruments - the Moderate Resolution Imaging Spectroradiometer (MODIS) and the Multi-Angle Imaging SpectroRadiometer (MISR). The latter was given on behalf of MISR aerosol scientist Ralph Kahn of JPL, who was unable to attend the meeting. MODIS will provide data on aerosols over both land and water by use of a wide range of spectral information covering the wavelength range from 0.4 to 1.4 µm. For aerosol measurement over land, the MODIS team will make use of "dark pixels." The aerosols have a relatively small effect at 2.1 and 3.7 µm, allowing surface properties to be detected, and the relationship between the radiances at these wavelengths and those in the visible (0.47, 0.66 µm) is thought to be well understood. Aerosol reflectance decreases with wavelength, but more so for sulfate than sea salt or smoke, allowing for some discrimination among the aerosol types. Cirrus clouds provide a major complication in MODIS analysis; their presence can be detected through use of a channel at 1.37 µm, although this coincides with a water absorption channel. Additional problems involve limitations on the number of different types and modes capable of being retrieved (currently a maximum of two). A key element of MODIS science will be comparison with surface-based measurements, notably those from the Aerosol Robotic Network (AERONET) of ground-based measurements which measure column integrated properties.
Aerosol measurements from MISR rely on its use of angular information, as 9 different directions are sampled. Therefore, over a period of a few minutes, 9 different observations of the same location are observed covering a range of measurement angles from -70 to + 70 degrees. MISR uses four spectral bands; there is therefore excellent complementarity between MISR's emphasis on angular information with MODIS's emphasis on spectral information, although the angular signal may be more difficult to deconstruct. MISR measures with a 360 km swath width allowing for 9-day global coverage. Among the information to be obtained by MISR are aerosol particle types over cloud-free calm ocean locations, aerosol optical depths to 0.05 or 10%, whichever is larger, for all common types of aerosols except soot, some sense of particle size (small, medium, or large), indices of refraction, and some sense of particle shape (spherical vs. nonspherical, which may be sufficient for differentiating large desert dust particles from others).
P. K. Bhartia of NASA/GSFC, Project Scientist for the EOS CHEM spacecraft, first briefly reviewed the three well-known instruments (HIRDLS, MLS, TES) to be launched in December 2002. He then spoke about a fourth instrument, referred to as the "Imaging Spectrometer" to be provided by the government of The Netherlands (with participation from Finland). The commitment from the Dutch government was only recently received. Some of the details of this instrument are yet to be determined, but it is expected that it will continue the data record from the Total Ozone Mapping Spectrometer (TOMS) series of instruments, including daily global measurements of total column ozone, UV flux, volcanic sulfur dioxide and ash, UV absorbing tropospheric aerosols. This new instrument is expected to have many more spectral channels than TOMS, however, and to cover a much broader wavelength region. Excellent spatial resolution should be obtainable through use of an imaging detector, and instrument reliability should be high because of the absence of moving parts. The finer horizontal resolution should reduce the effects of cloud contamination in a single pixel, which is a non-negligible problem given the spatial resolution of the current TOMS instrument. Work in progress includes the attempt to improve the instrument's horizontal resolution to 5 km (in the nadir), and to extend the long wavelength limit from its current 450 nm to 780 nm. Bhartia reviewed recent results from the current TOMS instrument for both tropospheric aerosols and tropospheric ozone; the latter of these requires the use of either some assumption about the distribution of either stratospheric ozone or total ozone, such as using the ozone column above convective regions to define stratospheric ozone amounts. Employing this technique, he concluded that in 1997 tropospheric ozone was greater over the Atlantic and less over the Pacific, anticorrelated with MLS 215 mb water vapor, perhaps due to biomass burning.
Joe Waters of the Jet Propulsion Laboratory discussed plans for the Microwave Limb Sounder (MLS) planned for EOS CHEM. This instrument will carry out "radio astronomy of the Earth" by passively measuring the microwave radiation emitted by trace constituents in the Earth's atmosphere. This instrument is a follow on to the current MLS instrument flying aboard the Upper Atmosphere Research Satellite (UARS), but has several improvements, including design to include upper tropospheric measurements (water, temperature, pressure, ozone, carbon monoxide, cirrus clouds), measurement capability for additional trace constituents (OH, HO2, HCl, BrO, HNO3, CO, HCN, and HOCl, as well as a dynamical tracer N2O), better global coverage (measure daily between 82 degrees N and S), improved vertical resolution (2-3 km, roughly a factor of two better than the UARS MLS), and improved horizontal coverage (limb scans every ~170 km instead of ~500 km as for UARS). Results of the UARS MLS instrument for upper tropospheric water vapor were also shown, including clear evidence for an El Niño signal in upper tropospheric water vapor. A limiting factor anticipated for quantitative analysis of EOS CHEM MLS tropospheric data is the lack of knowledge of continuum absorptions for dry air and water vapor needed to more accurately obtain tropospheric parameters (current empirical expressions are thought accurate to around 20%). The ability to measure upper tropospheric temperature is resolution dependent: with 3 km resolution, zonal temperature retrievals of 0.1° C are thought possible, but this degrades to 1° C when 2 km retrievals are employed. For the sake of measuring trends, it was suggested that the former choice is preferable.
David Edwards of the National Center for Atmospheric Research spoke about the High Resolution Dynamics Limb Sounder (HIRDLS) instrument planned for EOS CHEM. The key aspect of HIRDLS is its ability to determine trace constituent profiles with ~ 1 km vertical resolution. Vertical profile information should extend from the stratosphere down into the upper troposphere. There are several aerosol channels in spectral window regions in order to get good aerosol information, as well as the trace constituent measurement capability (emphasizing long-lived tracers but also including a fairly comprehensive measurement of nitrogen-containing species). HIRDLS can make up to six azimuth scans in short order as it moves along the orbital track providing essentially global coverage in 24 hours. This type of coverage is unique to HIRDLS. The HIRDLS instrument also has programmable science modes, including one with one degree by one degree horizontal resolution that can be obtained if one is willing to sacrifice the cross-orbital track measurements.
The Measurements of Pollution in the Troposphere (MOPITT) instrument to fly aboard the EOS AM spacecraft was described by instrument PI, Jim Drummond of the University of Toronto. MOPITT is to fly on the EOS AM-1 spacecraft and is provided by the Canadian Space Agency (CSA). The objectives of MOPITT are the determination of the vertical profile of carbon monoxide (CO) in the troposphere and a total column measurement of methane. CO accuracy is estimated at 10% (tropospheric variation of about 5x). Methane accuracy is 1% (total column variation of about 5%). For CO, channels at both 2.2 and 4.4 µm will be used - the former, mostly affected by surface reflection of solar radiation, can be used only in daylight, but provides column integrals (for both CO and CH4). The latter uses thermal emission and can thus be useful for both day and night retrievals. Due to the difficulty of measuring the profile near the surface, near surface values of CO are obtained by comparing the column integral (at 2.2 µm) with the vertical profile (at 4.4 µm). The CO profile in the 0 to 15 km region should be able to be measured with vertical resolution of approximately 3-4 km (3 levels in troposphere). The horizontal resolution will be 22 x 22 km, with successive measurements being made at 0.4 second intervals. Nearly global coverage will be made every 4 days, but clouds will interfere with measurements reducing the potential viewing opportunities.
The Stratospheric Aerosol and Gas Experiment (SAGE) instruments were described by Pat McCormick of Hampton University. The currently operating SAGE II instrument, flying since 1984 aboard the Earth Radiation Budget Satellite (ERBS), has provided long-term measurements of the distributions of ozone, water vapor, aerosols, and nitrogen dioxide. These data have been particularly useful for helping to define the long-term trend in stratospheric ozone, especially in the lower stratosphere (perhaps down to 17 km), as well as long-term variations in stratospheric aerosol, loading. Stratospheric aerosol column loadings seen recently are the lowest yet observed by SAGE II, including the pre-Pinatubo values. The new SAGE III instrument planned for EOS was also described. This will have numerous enhancements over the SAGE II, including added wavelengths, much higher spectral resolution, and direct measurement of temperature and pressure so that external information is not needed to convert from the measured number density vs. altitude to the mixing ratio vs. pressure more commonly used by atmospheric scientists. SAGE III also has a lunar occultation capability designed to measure OClO and NO3 at night. The first SAGE III instrument is planned to fly aboard a Russian Meteor-3M spacecraft in 1999 and the second aboard the International Space Station in 2002. A third SAGE III instrument is being constructed but is currently unmanifested as a "flight of opportunity" is sought. A flight aboard a Russian RESURS spacecraft in 2000 is a distinct possibility, although no commitments have been made for this flight at this time.
Daniel Jacob of Harvard University presented a summary of the Tropospheric Emission Spectrometer (TES) planned for the EOS CHEM spacecraft. TES is a Fourier Transform Spectrometer, measuring in the 650 to 3000 cm-1 wavelength region using both nadir and limb viewing geometries; spectral resolution is 0.025 cm-1 in the limb mode and 0.1 cm-1 in the nadir mode. The primary focus of the TES instrument is the measurement of ozone and its precursors in the troposphere, although TES has excellent capability for detecting a broad range of species because of its high resolution. In its nadir viewing mode, TES should be able to get 3-4 km vertical resolution if single pixels are considered; if averaging over multiple pixels, better vertical resolution should be possible. Besides ozone, carbon monoxide and water vapor should be measurable in nadir mode. In limb mode, with 2-3 km vertical resolution, these species as well as some nitrogen-containing species should also be measurable. Nitric acid should be measurable through most of the troposphere in this mode, while nitrogen dioxide is likely only to be measurable where it is present in elevated amounts (e.g. due to air pollution), and nitric oxide retrieval is only likely in the upper troposphere where concentrations exceed approximately 100 pptv, which corresponds to the standard background level (if one is willing to average over successive spectra, increased sensitivity can be obtained). Ozone measurements can conceivably be made down to the surface, and carbon monoxide measurements should be usable down to the planetary boundary layer. However, in most cases profiles will not extend below approximately 5 km. An airborne version of the TES instrument has flown several times, notably as part of the Southern Oxidant Study in 1995. Species detected included ammonia, methanol, and formic acid; high CO was observed over downtown Nashville. It was noted that validation of some species may be difficult due to lack of independent measurements.
Vertical profiles of ozone, water vapor, and clouds obtained by the Differential Lidar (DIAL) technique were described by Bill Grant of the Langley Research Center, on behalf of the principal investigator Edward Browell. In particular, observations of water vapor from the LASE instrument originally designed for ER-2 use but now being modified for the DC-8 (including both upward and downward viewing) were described. There is long term interest in evolving the DIAL technique for space-based measurements, but it is expected that this may take 5-10 years before being practical. Wavelength regions suitable for stratospheric and tropospheric ozone are 305-315 nm and 308-300 nm, respectively. For aerosol and cloud studies, the 820 and 940 nm spectral regions are of greatest potential use. In initial application, space-based DIAL would probably involve purely nadir pointing, although it would be desirable to have side scanning in order to obtain increased horizontal coverage. Specific case studies discussed include an O3 bulge off of Africa, and low O3 values in both the boundary layer and upper troposphere over the Pacific.
An EOS interdisciplinary effort on the integrated impact of urban areas on tropospheric trace gas burdens was described by Mark Zahniser of Aerodyne Research Inc. on behalf of investigation principal investigator Chuck Kolb. The investigation focuses on several different types of species, including gaseous smog precursors, greenhouse gases which provide for direct and indirect radiative forcing, photochemical oxidants, aerosol precursors, and particulates. Geographical information system (GIS) techniques are used to help relate known information about land surface properties (including land use characteristics), knowledge about population and industry distributions, and topographic and meteorological information to observed trace constituent concentrations. Measurement capability for Aerodyne's mobile laboratory include a HeNe methane monitor, an infrared carbon dioxide monitor, an electron capture detector gas chromatograph for sulfur hexafluoride, a condensation nucleus counter, and a tunable diode laser which can be used for simultaneous measurements of several species (e.g. nitric oxide, nitrogen dioxide, ozone). Results shown for Manchester, New Hampshire using SF6 as a tracer, indicated the buildup for CO2 (with levels up to 450 ppm) as well as high methane values over a landfill and a sewage treatment plant. Several campaigns are planned for Manchester in the coming few months. Subsequent field studies are planned for the larger urban area of Boston, Massachusetts.
Jim Hansen of the Goddard Institute for Space Studies reviewed the sources of changes in radiative forcing that have occurred and what is believed to be their effect on climate. Key information needed includes the exact nature of changes in the vertical distribution of ozone (especially to see if increases in tropospheric ozone may have countered decreases in stratospheric ozone in their contributions to radiative forcing) and the nature of aerosol particles (especially single scattering albedo) as well as their height. With reasonable estimates of aerosol absorption, aerosol cooling is reduced from previous assessments. Although changes in methane amounts are well characterized, their variability is not well understood, and accurate future predictions require that future methane amounts be well modeled. The observed long-term change in diurnal temperatures is best modeled in terms of corresponding changes in cloudiness, although these have not been detected. The effect of changes in vegetation and land use on the climate system may be significant, although this has not been well characterized. It is also possible that when looking at the last 150 years the net effect of changes in solar irradiance on the climate system may not be negligible relative to the sum of all other changes. In the longer term, changing aerosol distributions due to increased Asian development are likely to be important to the climate system.
An interdisciplinary study to characterize aerosol forcing over the Atlantic Basin was summarized by Brent Holben of NASA/GSFC. This investigation focuses on the improved knowledge of direct aerosol forcing through analysis of satellite measurements, focused in situ and remote sensing measurement campaigns, long-term ground based measurements, and aerosol models. Key data sets include the TOMS and Advanced Very High Resolution Radiometer (AVHRR) satellite data sets, the ground-based AERONET and Baseline Surface Radiation Network measurements, and aircraft campaigns such as TARFOX, SCAR-A/SCAR-B, and ACE II. The biggest uncertainty in interpreting existing data is the lack of information on the vertical distribution of atmospheric aerosols. Observations of aerosol thickness for some 800 days at Goddard found a distinct seasonal cycle, with values ranging up to 0.5 (at 0.5 µm) amid much variability in summer, and 0.1 in winter.
Studies of the use of satellite measurements to determine the effect of smoke particles on clouds and climate were reviewed by Yoram Kaufman of NASA/GSFC. Over the ocean, spectral information can be used to help get size information on aerosol particles, but over land the aerosols affect the outgoing radiation and the problem is much tougher. Cloud information can be obtained in the infrared (3.75 µm) but this may only pertain to some parts of the cloud. The indirect effects of smoke on clouds remains a difficult problem. The obvious response of clouds is to brighten in the vicinity of fires, especially in regions of high precipitable water. At relative humidities in excess of 90%, some particles are very effective as cloud condensation nuclei, although not as good as sulfates. It is clear that cloud effects of smoke can occur far from the areas of emissions, given an approximate 5 day time constant for smoke.
A newly initiated investigation of tropospheric trace gas budgets and their interactions with aerosols was described by Joyce Penner of the University of Michigan. In prior work, three dimensional distributions and specified size distributions were calculated for different types of aerosols (aerosol sulfate, organic carbon/black carbon, dust, sea salt), from specified initial distributions for ammonium ion, ammonia, and water vapor. Future work will involve the calculations of partitioning of materials between the gas phase and aerosols, and understanding the implications of this partitioning for particle size distributions which govern dry deposition rates. Significant data needs include measurements of aerosol composition, especially if organic aerosols are to be understood and well simulated.
Bob McGraw of Brookhaven National Laboratory summarized work being done on the representation of aerosol microphysics in regional to global scale models. This work involves two parts - treatments of aerosol microphysics, including both basic nucleation and aerosol dynamics in complex flow fields, and the representation of aerosol transport and transfer in large-scale models, notably the Brookhaven three-dimensional model. A particular focus has been the study of sulfuric acid-water hydrates, including the temperature dependence of vapor pressure of sulfuric acid in these systems. The computational work centers on the use of the method of moments, in which particle size distribution information is carried without having an explicit representation of the nature of the size distribution. As implemented numerically, a quadrature method of moments approach is used. This approach is being implemented in the Brookhaven 3D transport and transformation aerosol model, which covers the latitude range of 81N-81S with a horizontal resolution of 1°x1° and 27 vertical levels from the surface to 100 hpa.
Steve Ghan of the Pacific Northwest National Laboratory (PNNL) described the PNNL MIRAGE (Model for Integrated Research for Atmospheric Global Exchange), a global model developed to estimate direct and indirect forcing by anthropogenic sulfate aerosol. MIRAGE couples PNNL's version of the Community Climate Model (CCM2) (with cloud microphysics coming from the Colorado State University's RAMS model), with the PNNL global chemistry model. Results for July-August, 1994 suggest that the total radiative forcing due to aerosols is -2.7 W m-2 (-0.8 direct, -1.9 indirect), although there is significant temporal variability in this number because of variability in clouds. Comparison of simulated and observed "aerosol radiance" suggests the direct forcing estimate is too high; the explanation for the bias has been identified as a bias in the simulated relative humidity, which influences water uptake on soluble particles.
Efforts to develop an "event oriented" study of atmospheric aerosols using global models were summarized by Brian Toon of the University of Colorado and members of his research group. Particular foci of his group's efforts include algorithm development for aerosol microphysics, studies of direct aerosol effects, examining of the relationship of atmospheric chemistry to aerosol formation, and investigations of indirect effects of aerosols on both stratus and cirrus clouds. A variety of numerical models are used for these investigations, including the MATCH model from NCAR, the NASA/Ames Research Center stratospheric dynamics model, the NOAA/NCAR two dimensional stratospheric chemistry model, and a 3D local event simulation model used for stratus cloud studies. Key data sets utilized include SAGE II, AVHRR, and UARS. Concern is also focused on the effects of optical properties of dust (including, for instance, the role of iron on visible absorption, which is variable in concentration but dominant when present, so that no set of optical constants is appropriate everywhere). Non-volcanic stratospheric aerosol studies were summarized by Mike Mills, who found through use of a version of the NOAA/NCAR 2D model modified to include sulfur chemistry and aerosols (45 size bins) that observed aerosol distributions in the lower stratosphere could not be understood solely on the basis of carbonyl sulfide but rather required the considering of sulfur dioxide and tropospheric aerosols. Andy Ackerman reported on comparisons between theory and measurements of cloud susceptibility from the Monterey Area Ship Track (MAST) experiment, and also used large-eddy simulations with explicit microphysics to show that boundary layer depth and cloud fracitonal coverage depends strongly on cloud droplet concentrations under very clean conditions.
Ina Tegen of Columbia University described work carried out on soil dust aerosol modeling. After sulfates, soil dust, especially mineral desert dust, is the second largest contributor to atmospheric aerosols. Long-term studies on the troposphere should look at long-term changes in dust loading, including whether or not there is any dynamical feedback between dust concentrations and dust lofting. Interannual variability in dust can be simulated through the use of meteorological models simulating phenomena such as El Niño and the North Atlantic Oscillation that may affect atmospheric dust distributions. The GISS GCM contains four size classes for aerosol particles, and assumes a fairly simple relationship between deflated dust mass and surface wind velocity, with consideration of effects of soil types and surface wetness. Dust concentrations calculated with the model were compared to those of observations. Both the model and observations showed significant interannual variability, but the model did not get the seasonal cycle on dust right in all locations. In particular, the model did a poor job in estimating the seasonal cycle of Saharan desert dust, underestimating summertime dust in Izana and Barbados, perhaps due to underestimation of surface winds in coarse GCM grids. If one includes the radiative feedback on atmospheric dust, there is a 10-15% decrease in dust loading, especially over Asia, due to stabilization of atmospheric temperatures. Estimation of Saharan dust radiative forcing can range from +1 to -1 W/m2 by changing the single scattering albedo by ±10%.
Studies of subvisible cirrus clouds using the interactive two-dimensional model developed at NASA/GSFC were described by Joan Rosenfield. Subvisible cirrus clouds have optical depths of approximately 0.01 and are seen in satellite data (notably the SAGE II climatologies developed by Pi-Huan Wang and coworkers). The GSFC 2D interactive model used allows for creation of subvisible ice clouds below the tropical tropopause when appropriate temperature and water vapor distributions exist for supersaturation to occur. The zonal variations in temperature that may be important in cloud formation are included based on analysis of meteorological data sets. Radiative heating of ice particles and changes in temperature and circulation are calculated with the model. A log normal size distribution is assumed. The net result of the calculations is that the subvisible cirrus lead to a heating of the tropopause region that allows more water vapor to enter the stratosphere. The smaller the size of the particles, the more heating takes place and the larger stratospheric response occurs. A 1-2° K temperature change led to an increase of 0.1-1 ppmv in stratospheric water vapor.
Radiative and chemical modeling of the troposphere and stratosphere carried out for NASA's Atmospheric Effects of Aviation Project were described by Jose Rodriguez of Atmospheric and Environmental Research, Inc. The effects of both a projected fleet of high speed civil transport (HSCT aircraft) and the current and projected future fleet of subsonic aircraft are considered. Both chemical effects due to aircraft-emitted nitrogen oxides and sulfate particles, as well as the climate effect due to these emissions and those of water vapor are considered. The numerical models designed to represent the response of the stratosphere to emissions of nitrogen oxides give a range of responses, but the ozone column changes tend to be fairly small (typically 0.1% in mid-latitudes, slightly higher in the polar regions), although some models actually suggest increases in ozone. A range of responses (~0.2 to 0.6 ppmv increases) in stratospheric water vapor was calculated. The models in which the largest increase in total inorganic nitrogen (NOy) were calculated also showed the largest increases in water vapor; curiously, models with the most realistic age of air in the lower stratosphere seem to have excessive NOy. Tropospheric phenomena being investigated include the production of ozone from nitrogen oxides emitted by the aircraft, particle production (from sulfate and soot), contrail formation, and the role of other greenhouse gases, notably carbon dioxide. Nitrogen oxide levels in aircraft corridors can be enhanced by up to 50%, and maximum ozone increases of 10-16 ppbv are calculated for the aircraft corridor (maximum in summer, minimum in winter). There are significant limitations in the models used for these simulations, including relatively coarse resolution of the tropopause region, the uncertain magnitude of other sources of nitrogen oxides, such as lightning and convective transport of surface-produced gases, and unidentified sources of odd hydrogen in the upper troposphere, which may be important for converting nitrogen dioxide to nitric acid. Less is known about the impact of aircraft generated aerosols. The models suggest that the effect of aircraft-generated sulfur in plumes may be large, although the best estimate is that it is at the 5-10% level. It seem that aircraft generated particles will constitute only a small perturbation to total mass loading near the tropopause and to direct radiative forcing, although the small size of the particles means that the contribution to total particle surface area may not be negligible, meaning the contributions to heterogeneous chemistry and indirect radiative effects cannot be ruled out. An estimate of contrail extent is that they provide a global coverage of about 0.1%. The best estimate to date of the total radiative forcing due to aircraft effects is now ~0.02 W m-2, although this could increase to something like 0.1 W m-2 by 2050. It is not clear whether any observable impacts have yet occurred.
The status of the Global Modeling Initiative (GMI) developed by AEAP to help support assessments of atmospheric effects of aviation was reviewed by Doug Rotman of the Lawrence Livermore National Laboratory. The GMI model is being developed for both supersonic and subsonic assessment applications. The GMI model is designed to provide several different representations of key model components and input data sets so that the sensitivity of calculations to different approaches can be understood. For instance, meteorological fields used as inputs include the data assimilation products developed by the Data Assimilation Office at NASA/GSFC, and winds from both the CCM2 from NCAR, and general circulation models (GCMs) from GISS (GISS model II/II'). Advection schemes include the flux-form semi-Lagrangian model of Lin and Rood, the semi-Lagrangian technique of Rasch and Williamson, and the second order moments scheme developed by Prather; the age of air in the upper stratosphere can vary by 2 years depending upon which scheme is used. Chemical schemes include those of Brasseur and Lamarck, ONERA, and a parameterization developed by Malcolm Ko and Clarisa Spivakovsky. Photolysis rates use a look up table technique developed by Randy Kawa, and "cold chemistry" driven by the presence of polar stratospheric clouds is simulated with the approach developed by David Considine. Comparisons of model results with satellite and ozone data, including an ozone climatology being developed by Rich McPeters and Jennifer Logan, are important efforts. Some of the modules have been modified to run on parallel computers in order to increase computational speed.
Guy Brasseur of NCAR described simulations of ozone and related chemical tracers being done with the Model for Ozone and Related Chemical Tracers (MOZART) at NCAR. This makes use of an off-line version of the most recent version of the NCAR Community Climate Model (CCM3), including T42 horizontal resolution (corresponding to 2.8 x 2.8 degrees), 25 vertical levels, a time step of 20 minutes, and 50 transported species. Comparisons with available data have been carried out and show that, for instance, calculated carbon monoxide distributions agree well with observations, but nitric acid is much less well modeled (model-calculated ratios HNO3/NOx are much greater than those observed). A simulation of changes in surface ozone since preindustrial time has been carried out, and results show that changes in ozone amounts of some 30-40 ppbv in July over the U.S. and much of Europe along with those of 20-30 ppbv over most of Eurasia have occurred. The average radiative forcing coming from these increases should be approximately 0.45 W m-2 (higher in the northern hemisphere, smaller in the southern hemisphere). Predictions for the future (2050) show the largest changes in surface ozone are likely to come in the tropics, with some seasonal dependence (July changes of 30-40 ppmv are largest in central America, the Middle East, and Asia, while January changes are larger in South America and Africa). In the upper troposphere, increases of close to 20 ppbv are suggested over the northern India/Himalayan region. When aircraft are included, similar increases are seen at high latitudes, so it is likely that the actual evolution of the troposphere will reflect a broad range of gases. The evaluation of tropospheric chemistry models is complicated by the limited information in the troposphere coupled together with its significant variability; a "climatology" of trace constituent measurements in the free troposphere is being assembled at NCAR, although even this may be sufficiently limited that it cannot fully represent the range of actual variability.
The Chemistry, Aerosols, & Climate Tropospheric Unified Simulation (CACTUS) effort was described by Daniel Jacob of Harvard University. The purpose of CACTUS is to improve our understanding of chemistry-aerosol-climate interactions through development of modeling tools incorporating fundamental representation of chemistry and aerosol microphysics. The present targets of the effort include the climate effect of changes in emissions of precursors to tropospheric ozone formation, the feedback of climate change on tropospheric ozone, the climate effects of changes in emissions of sulfate precursors, and the response of sulfate aerosols to climate change. The model used 24 on-line tracers of 90 species, the Fast-J photolysis code of Prather, and the SMV GEAR chemical solver. The strategy for cactus involves using the same chemical models on-line in the GISS Model II' GCM for climate change studies, and in a CTM using data assimilation fluxes from the GSFC DAO GEOS system for chemical model validation. Identical emission inventories, chemical reaction sets and solvers, deposition models, etc. would be used with both models. Calculations have been carried out for emissions corresponding to the 19th century, but the model cannot simulate the low ozone observed at European stations. The need for high vertical resolution in the tropopause region is being actively considered; a 23-layer version of the GISS GCM appears to have realistic tropospheric/stratospheric exchange but runs some six times slower than the current 9 layer model. The simulation of aerosol sulfate is being carried out on-line at GISS, and the importance of including the ammonia is being investigated at the California Institute of Technology. Data from the SUCCESS mission carried out by AEAP provide an important data set for the analysis of the aerosol models.
A simplified but very fast chemical reaction scheme suitable for use with the suite of modeling tools developed for CACTUS was described by Drew Shindell of Columbia University. This scheme uses chemical families and so is sufficiently fast that it can be run rapidly within the general circulation model, allowing for interactive treatments of chemical effects on radiation and dynamics and numerous sensitivity tests. The present version of this scheme involves 9 advected constituents and families and a total of 51 reactions (including photolytic reactions). A major limitation of this scheme is its restriction to single carbon containing species only. Most of the results obtained to date were from a previously used 5-tracer version, which did a reasonable job in simulating the bulk features of tropospheric ozone distributions. This scheme needs to be integrated with other modules (emissions, deposition, lightning) so that it can be run more usefully in long-term simulations. Potential applications include comparisons with other models (e.g. the same GISS II' GCM but different chemistry from CACTUS, or different (ECHAM) GCM/CTM but similar chemistry, at Ultrecht, Holland), calculations of radiative forcing in interactive systems and for surface ozone, and agricultural and human health impacts.
Greg Carmichael of the University of Iowa presented results on the interaction between tropospheric trace gases and aerosol particles, notably possible reactions involving particulate nitrate, particulate sulfate, and dissolved oxidants. Considerable attention has been paid to the atmosphere near Asia, where there is high aerosol loading involving many different types of aerosols. Unfortunately, there are very few data available on size-resolved aerosol composition, although some data do exist near Japan. A regional scale model was used for much of the Asian studies, although a global model has also been used. The role of calcium in the aerosol particles is of special interest, and the fraction of calcium in soils has been shown to be highly variable. Dust sources have been calculated based on models including the effects of vegetation, moisture, and surface winds; different meteorological models have been used as inputs to these efforts. Model simulations of the day-to-day variability in dust loading have been reasonable but not perfect. A particularly interest episode was one measured during the PEM-WEST B mission, in which a small dust storm was sampled by the aircraft. A significant role was found for calcium (or carbonate) in the partitioning of nitrogen oxides - increasing calcium in the fine particle mode shifts nitrogen into nitric acid. Some experimental work has also been done to look at nitrogen dioxide on various surfaces, including elemental carbon and various oxides (aluminum, iron, titanium, calcium, and silicon). This type of reaction could affect the general model problem of excessive HNO3/NOx ratios in the troposphere. The effect of relative humidity on surface reactions with ions has also been studied.
Cloud processes of particles in gases in stratus clouds were described by Graham Feingold of NOAA. Boundary layer models incorporating dynamics, microphysics, and radiation were used for studies of local events. Both Eulerian (Large Eddy Simulation, LES) and Lagrangian (Transformed Ensemble Models, TEM) were used to study dynamics in and near clouds. The TEM model was also used with a cloud parcel model plus chemistry to simulate aerosol size growth. The distribution of sulfate, nitrate, ammonia, and hydrogen peroxide in air parcels were calculated, and the relative concentrations in parcels that went into and did not get into the clouds were studied. Particular application has been made to the North Atlantic Regional Experiment (NARE) aircraft campaign, in which the effect of nitrogen oxides produced in the northeastern U.S. on clouds in the Atlantic was examined. For in-cloud depletion of SO2, mean conditions with one parcel overestimates somewhat the effect compared to the average of 500 parcels. Bulk chemistry calculations without size resolution overestimates SO2 depletion leading to an underestimation of sulfate generation.
Nonlinear interactions involving methane in the troposphere and stratosphere were examined by Don Wuebbles of the University of Illinois. Methane is involved in non-linear chemical cycles in the troposphere, and also is critical in chlorine deactivation in the stratosphere through formation of hydrogen chloride. There are significant questions on how emissions of other trace gases (nitrogen oxides, carbon monoxide, nonmethane hydrocarbons) will affect methane distributions, and how changing climate will affect methane. The effect of methane on ozone is also not completely clear. These effects become particularly important when considering long term scenarios, such as those developed by the Intergovernmental Panel on Climate Change (IPCC) for its deliberations on future climate. Given the change in methane growth rate (reduced from some ~12 ppbv/year to ~5 ppbv) future forecasting of methane growth is complicated. It also depends on future changes of NOx, CO and non-methane hydrocarbons. The IS92A scenario projection used in the most recent IPCC report assumed no changes in these species; if reasonable changes are assumed, it reduces the magnitude of methane increase by 2100 by a factor of two. Hydroxyl will respond to the methane changes, and will affect numerous other gases. If policies are to be developed to try to stabilize methane concentrations at some levels, then methane sources and sinks will clearly need to be better understood.
A study of carbon monoxide using a three dimensional chemistry transport model was described by Prasad Kasibhatla of Duke University. The goal of this project is to help get global distributions of CO by filling in the gaps between spatially limited data sets. The ones of most interest will be the Measurement of Air Pollution from Satellites (MAPS) that flew on the Space Shuttle and the upcoming MOPITT instrument. Specific objectives include the development of a comprehensive model of spatial distributions and temporal variability of CO in the troposphere, elucidation of the interplay between sources, transport, and chemistry on tropospheric CO distributions, and combination of in situ and remote sensing measurements to better define the global distributions. The primary modeling tool is a chemistry transport model developed at NASA/GSFC, using input from the GSFC DAO GEOS product. This has already been applied to transport of CO in large-scale turbulent and convective regions. Currently, the chemistry in the model is being modified to provide for interactions between CO and OH using the parameterization developed by Spivakovsky. Statistics for CO distributions in 1994 have been compiled, and preliminary calculations are being done for that year. The model does a good job of simulating the amount and seasonal cycle of surface CO except for the rapidity of the summertime increase (possibly due to northern hemisphere forest fires). The amplitude of the seasonal cycle (~100 ppbv) is well simulated. Both the model and data show a reduced seasonal amplitude at low latitudes, but the model underestimates the seasonal change, perhaps because of an overestimation of biomass burning. There is also too much CO in the tropical upper troposphere, perhaps associated with the model's convective transports. In the southern hemisphere the model calculated amounts are above observed levels (e.g. at Ascension Island). Future work involves development of an improved OH parameterization, update of the biomass burning source, sensitivity studies, and design and implementation of approaches to the assimilation of CO in the model.
Ken Pickering of the University of Maryland reported on tropospheric convection and stratosphere-troposphere exchange (STE) effects in photochemistry, aerosols, and climate. The objective of this effort is to examine the effects of vertical mixing processes such as deep convection and STE on the distributions of photochemically active trace gases and aerosols. The long-term goal of this effort is to help improve the ability of global CTMs to account for vertical mixing under a range of model resolutions. The strategy of the effort is to enhance the vertical resolution of the Goddard chemistry/transport model near the tropopause, and in the longer run couple the CTM with the stretched grid GCM developed at the University of Maryland. The role of chemistry in the Goddard Cumulus Ensemble (GCE) will also be expanded, and simulations of specific events will be carried out. Observations of 222Rn will be important for validation of the new modeling tools. Future work for this effort will involve comparisons with airborne measurements made during the SONEX aircraft campaign sponsored by AEAP in late 1997, as well as earlier aircraft campaigns (PRE-STORM in 1985) and the MAPS shuttle flights (especially the two in 1994).
Areas of mutual interest to oceanographers and atmospheric scientists interested in atmospheric aerosols were summarized by John Marra of Lamont-Doherty Earth Observatory. These include the regulation of ocean (and terrestrial) productivity through supply of aeolian iron and nitrogen to biological systems, and the importance of iron to ocean biogeochemistry. The use of satellite measurements to study both climate and the ocean surface is also of interest to both communities. Finally, the study of DMS fluxes from the ocean to the atmosphere has the potential to help improve understanding of the naturally occurring sources of atmospheric aerosols, especially in remote regions.
There was appreciable discussion throughout the meeting. One item mentioned especially frequently was the importance of having good altitude information about the distribution of tropospheric aerosols. This is very difficult, if not impossible, to obtain from passive instruments that look down (nadir-viewing) from space or up from the ground. The use of active techniques, such as lidar, together with passive instruments to obtain aerosol information was felt to be of critical importance to improving our understanding of aerosols and their effects on climate. The truly significant difficulty of understanding the indirect effect of aerosols on climate because of aerosol-cloud interactions was also made clear to all the meeting participants. It was suggested the cloud radar observations from space would be quite useful in this regard. Observations of the correlation of CCN with aerosol loading (e.g., non sea-salt sulfate, as presented by Jeff Kiehl of the National Center for Atmospheric Research) show great scatter, perhaps due to the lack of constraints on other influences in previous studies; it was further noted that the CCN measurements themselves are quite difficult. Finally, it was felt that laboratory studies of the interaction of aerosols and atmospheric chemical species have to date been quite limited.
We acknowledge the assistance of Drew Shindell and Lee Grenfell in preparing this article.