This chapter contains a description of the specific approaches that will be used to implement the research strategies outlined in chapter 3. This implementation plan is organized around the six scientific questions that form the general focus of the ACE-1 science plan.
2. Measured hygroscopic response of the aerosol to changes in RH and the hygroscopic response calculated from the measured aerosol number and chemical mass size distributions, RH and published functional relationships between chemical composition and water uptake.
3. Measured aerosol scattering and calculated aerosol scattering derived from Mie theory applied to measured number and chemical mass size distributions.
4. Measured increase in aerosol scattering due to hygroscopic growth and the calculated increase based on measured number and chemical mass size distributions, RH, and published functional relationships between particle chemical composition and water uptake.
5. Measured CCN supersaturation spectrum and the CCN supersaturation spectrum derived from Khler theory applied to the measured aerosol number and chemical mass size distributions and the CCN supersaturation spectrum derived from cloud droplet number concentration.
6. Measured cloud droplet number concentration just above cloud base and the cloud droplet number concentration calculated from the measured updraft velocity, aerosol number size distribution, aerosol chemical composition, and CCN supersaturation below cloud base.
7. Measured index of refraction and index of refraction derived from the measured aerosol number and chemical mass size distributions, RH and published functional relationships between chemical composition and water uptake.
The C-130 will have roughly the same suite of instruments on board to permit local closure calculations. The growth of aerosol mass with humidity will also be determined from a controlled-humidity microbalance technique and only two size-cuts will be available for the chemical composition (to improve the time resolution of the measurement). A thermally-conditioned OPC will be used to infer the size resolved chemical composition of the dry aerosol (30% RH) over rapid collection intervals (5-10 min.). Information derived in clean regions includes the sulfate size fraction, the ammonium to sulfate molar ratio, and the refractory component including sea-salt, dust and soot (Clarke, 1990). Results will be used to characterize and model variability in aerosol refractive index and scattering coefficients at ambient conditions. Modelling of the refractive index will be compared with the direct measurements obtained from the MASP. Modelling of coarse and fine fractions will be compared to coarse and fine dry scattering coefficients obtained from the “dry” nephelometer. Comparisons of variability in the fine particle refractory component (often soot) will also be compared to soot estimates obtained from the aethalometer. The aircraft will also measure the cloud droplet number concentration for comparison with the CCN size spectrum measured at cloud base.
Macquarie Island, although small and logistically more difficult, will have the advantage of almost constant remote marine air. Cascade impactors will be used to collect aerosols in multiple size ranges for analysis of major ions, trace metals, elemental carbon, organic carbon, and total mass. The total number population of aerosols with Dp > 3 nm and Dp >15 nm, the number concentration of aerosols in several size bins (20 < Dp < 600 nm), and the CCN supersaturation will be measured continuously. These measurements will be intercompared to assess their internal consistency.
The ACE-1 experiment is being designed so that the various platforms can make the necessary measurements to perform local closure calculations for virtually the entire experiment. Thus, even when allowances for undesirable sampling conditions and instrument downtime are made, it is reasonable to expect a statistically significant number of closure calculations (100 or more) from each platform over the course of the ACE-1 campaign. This is important both for establishing valid measures of instrumental uncertainty and for exploring, within the context of the entire ACE-1 data set, which factors significantly influence the difference between measured and predicted quantities. Although the C-130 will not achieve the same temporal coverage, it will be able to extend the closure experiment to higher altitudes, to a wide range of latitudes, to the Kilauea volcanic plume (a higher concentration case), and perhaps to some continental pollution during the transit flights. Since the backscatter and aerosol absorption in the cleanest parts of the Southern Hemisphere will be near the limit of detection of the nephelometers and aethelometers, respectively, the higher concentration cases will provide data with a better signal to noise ratio for validating the overall experimental strategy.
Critical measurements that will be made aboard the C-130 include: aerosol size distributions, size-resolved aerosol chemical composition (impactors), scatter and hemispheric backscatter (nephelometer), light absorption (aethalometer), aerosol optical depth spectra and solar and infrared upwelling and downwelling radiation (up/down shadowband radiometer), aerosol vertical distributions (lidar), and meteorological/state parameters at a variety of altitudes. The changes in radiation levels between altitudes will be correlated with the size, concentration, composition, and optical thickness spectra of the aerosols in each altitude interval. While the ship and ground based sunphotometers will determine optical depths by looking upward at the sun, both the shadowband radiometer and the water-leaving radiance/satellite combination will permit optical depths to be calculated by looking downward at the ocean's surface. The experiment will seek closure between all these methods, as well as seeking local closure, as described above, at each altitude.
The flight profiles (Figure 7) will include several long (up to an hour) legs prior to a satellite scene and several after the overpass to maximize the signal-to-noise ratio for collection methods with long integration times (impactors and the aethalometer). Each of the legs between 60 and 540 minutes in Figure 7 will be centered under the expected AVHRR scene and above the R/V Discoverer. A rapid sequence of five short (15 min) legs in the middle of the flight will permit the radiation spectra, scattering coefficients, thermodynamic variables, and physical size-distributions to be measured vs altitude almost simultaneously with the scene. A volume-scanning NCAR aerosol lidar will be used to map aerosols above and below the aircraft. This will be used to choose sampling altitudes, to identify any strata of aerosols that might contribute a significant amount of scattering, and to scale the observations over the altitudes that were not included in the extended in-situ sampling legs.
This exercise must be conducted in cloud-free air, so that clouds do not swamp the more subtle aerosol scattering signal. Flights will be timed so that the mid-flight short legs occur during an AVHRR scene, preferably with the R/V Discoverer R/V Southern Surveyor making surface measurements of water-leaving radiance beneath the flight pattern. This will not be a trivial task, since it means finding a section of cloud-free air within an AVHRR scene and rapidly deploying the ship to the same clear patch as the aircraft. In practice, the ship will be located near the center of the target scene well in advance of the overpass and then real-time satellite imagery will be used to direct both platforms to the largest clear patch that the ship can reach by the time of the scene. High-quality communication links between platforms are clearly essential to achieving this goal.
5.2.1 Low optical depth case
The aerosol optical depths south of Tasmania will likely be in the range 0.04 to 0.05 much of the time (Forgan, 1990; Durkee et al., 1991). Although these low optical depths will push the optical instrumentation to near their limits of detection, the up/down shadowband radiometer is capable of making high-quality measurements in this range, provided certain stringent experimental conditions are met. These conditions include the need for:
2 excellent knowledge of interfering optical depths (e.g., ozone, nitrogen dioxide, Raleigh scattering, cloud),
3 measurements at low solar elevation angles (to increase airmass and hence sensitivity of output to optical depth).
5.5.2 High optical depth case
The Kilauea Volcano plume, which stretches west-southwest from the Island of Hawaii, will be used to test the column-closure methodology. Optical depths through this plume reach 0.1 and greater which is well above the instrumental noise levels. Two flights will be flown from Oahu during the southbound transect and both will do vertical profile measurements on the volcano plume.
The flights will start with a volume-imaging lidar run parallel to the plume, to identify the structure and the location and altitude of the highest aerosol concentration. Legs will then be flown beneath the plume center, within it, and above it to characterize its physical, chemical and radiative properties and their radiative impact. This pattern will be flown twice per flight at separate locations where the preflight satellite-derived optical depths differ by a factor of two or more. A clean-air stack will be flown adjacent to the plume. The contrast between radiances at these three locations within and outside the plume will provide a range of conditions suitable to link the key measurements required for physical, chemical, and radiative closure. Characterization of the response of the diverse instrumentation over this wide range of sulfate-driven variability will provide a well defined basis for data intercomparison of the significantly lower concentrations expected during the ACE-1 intensive.
The goals of these measurements will be to:
2. Identify and quantify parameters (e.g. phytoplankton speciation, nutrient concentration, photic zone depth, meteorological variables) and/or processes (e.g. zooplankton grazing, aeolian input of micro-nutrients) that can be related to DMSP and DMS production and consumption rates.
3. Provide data for marine pelagic food web models (Gabric et al., 1993).
There are still many uncertainties in the oxidation pathway of atmospheric DMS (section 3.2.2). ACE-1 will use both Eulerian and Lagrangian strategies to try to answer this question.
5.4.1 Eulerian observations
Time-series data can be very useful in elucidating sulfur chemistry, particularly if the air is sufficiently horizontally homogeneous that advection is not the primary cause of concentration changes. For instance, the diurnal cycles of DMS and SO2 can be used to estimate the efficiency with which DMS is converted to SO2. During ACE-1 both Cape Grim and the R/V Discoverer will support extensive measurements of sulfur chemistry that will in part be used for the analysis of diurnal variations. Measurements will include gas phase DMS and SO2 and particulate phase anions and cations on time scales of several hours or less. All sites will have 12 or 24 hour observations of size-dependent aerosol composition from which information about formation and removal processes can be derived. Regular radiosonde launches at all sites will support estimates of mixing depths and vertical structure.
The sites at Baring Head and Macquarie Island will collect time series of smaller sets of species, but with enough spatial separation to make them useful for comparison with regional models of sulfur chemistry. Mass size distributions of the major cations and anions will be collected on a daily basis. Extensive physical measurements of aerosols will be made continuously.
The four surface sites will compile time series of concentrations for use in characterizing sulfur chemistry in airmasses with different histories.
5.4.2 Lagrangian observations
Two Lagrangian experiments will be conducted during ACE-1. In each experiment an air mass will be tagged and repeatedly sampled over a 48 hour period. Instruments on the C-130 will measure gas phase DMS, DMSO, DMSO2, SO2, MSA, and H2SO4 and particulate phase MSA and SO4. In addition, the measurement package will include ammonia, light hydrocarbons, halocarbons, ozone, peroxides, CO, nitric acid, NO, NOy, and all the aerosol physical and chemical properties described in the preceding sections. The C-130 will be well-equipped to observe changes in the sulfur, nitrogen, and carbon budgets and gas/particle conversions with time.
The analysis of Lagrangian data will be based on a continuity equation, which has previously been used in various forms for evaluating fluxes of heat, momentum, and ozone in the marine boundary layer (Lenschow, et al., 1981; Kawa and Pearson, 1989). For a substance S (DMS, DMSO, nss sulfate, or any other species) in a Lagrangian parcel,
d<S>/ dt = Jo(S) - Jh(S) + F(S) - D(S) (1).
Here, Jo is its surface flux, Jh is the flux at top of the mixed layer (i.e., entrainment of free tropospheric air
across the capping inversion), F is the column formation rate (such as from chemical reactions), D is the column
destruction rate, and
is the column concentration of S. Implicit in this equation is the Lagrangian assumption
that transfer through the sides of the parcel are not causing significant concentration change. Inert tracers will
be used to test this assumption and evaluate dispersion. Although neither advection nor dispersion appears in the
budget equation above, these terms will still contribute to the uncertainty in quantities derived from Equation
1, and must be explicitly included in the propagation of errors analysis.
The first challenge to implementing a Lagrangian experimental strategy is to know where the air is going. Airmasses will be tagged with several constant-density balloons and one constant-altitude balloon (a "smart" self-compensating design). After these have been launched from the R/V Discoverer (Figure 5), the balloons will float with the air and radio their satellite-derived positions, altitudes, and thermodynamic variables to the C-130 and Hobart Operations Center. Every 5 minutes the balloons will transmit all their data from the last 7 hours, so that the entire trajectory can be reconstructed for those periods when the C-130 is refueling and changing crews. The C-130 will begin sampling the airmass just after the balloons have been launched.
The R/V Discoverer will make a detailed characterization of both the target airmass (during balloon launch) and surrounding airmasses (before launch), with which to evaluate the horizontal heterogeneity of chemical composition (for calculating the effect of dispersion on concentrations). It will also estimate the surface sources of DMS, ammonia, and other species (the first term on the right of Equation 1) based on air and water measurements both during the launch and immediately after launch as it follows the airmass downwind for about 12 hours. Although the ship will not be able to keep up with the air parcel, it can characterize the variability of the surface source term along the initial track of the parcel. During this chase period, the ship's exhaust will be passing forward over the bow, so most of the air sampling operations will be shut down and the C-130 will become the principal source of information on airmass composition.
The C-130 will fly 20 minute "L" shaped patterns centered around the centroid of the balloons. Each flight will include at least two stacks of the type described in Figure 6, to obtain data to evaluate the vertical distribution of species in the boundary layer, as well as the rate of entrainment (the second term on the right of Equation 1) of free-tropospheric air into the parcel under study. The stacks will advect with the wind, thus maintaining the aircraft's position relative to the balloon centroid. One longer leg will be flown per flight, to determine the nature of the air mixing through the sides of the box and the rate of dispersion.
After the first 10 hour flight, the C-130 will spend just two hours on the ground (refueling and changing both flight and scientific crews), before taking off on flight 2 to rejoin the tagged airmass. Crew duty limitations (13 hours rest between flights by the same crew) require that the ground time be extended to three hours before flights 3 and 4. By making four ten-hour flights (Figure 4), the airmass will be under almost continual observation for nearly 48 hours.
The budget terms in Equation 1 will be determined as follows:
2. The surface fluxes, Jo(S), of the dry-depositing aerosol species will be estimated from the Slinn and Slinn (1980) model of marine aerosol deposition, using measured size distributions. Surface fluxes of gases (such as SO2) will be estimated from momentum and heat eddy flux measurements. Precipitation will be avoided as much as possible, since the aircraft is unable to collect a statistically-valid sample of precipitation to quantitate that surface-loss term for soluble species and aerosols in the budget analysis. Emission fluxes will either be estimated based on Liss and Merlivat (1986) with air and water measurements from the ship or be left as unknowns to be derived from the budget analysis.
3. The entrainment term, Jh(S), will be estimated from the expression
Jh(S) = -we ([S]FT - [S]BL) (2)
in which the measured concentration difference across the capping inversion is multiplied by an entrainment velocity, -we, to compute the entrainment flux. In many cases this flux will be negative, representing the dilution of boundary layer concentrations with cleaner free tropospheric air. The entrainment velocity will be determined from direct airborne eddy-correlation measurements of ozone, sensible heat, moisture, and momentum fluxes, along with the corresponding differences in their scalar quantities across the inversion. The rate of entrainment can also be estimated from divergence calculations on meteorological model output and from observations of the moisture structure of the atmosphere.
4. The chemical formation and destruction terms in Equation 1, F(S) and D(S), will either be determined from chemical considerations (F(DMS) and D(nss sulfate) are zero, for instance) or be left as unknowns to be determined from the budget analysis.
at least one balloon will survive to the end of the two-day experiment. The expense of the smart balloons precludes launching several of them, so the divergence of five
or six constant density tetroons will be used as another measure of the integrity of our airmass over the 48 hours.
Among the other lessons of ASTEX/MAGE:
2. Protocols will be established for how each platform responds to the loss of any or all of the balloons.
3. The launch point of the balloons will be selected with some knowledge of the homogeneity of the airmass, based on satellite observations, ship measurements, and (if necessary) a survey flight. This will prevent the initiation of an experiment in a region of high concentration gradients.
4. Meteorological data from the ships and surface sites will be fed rapidly into the regional meteorological models, so that the best possible predicted trajectories can be generated to position ships and aircraft for the start of the Lagrangian experiments.
5. If the boundary layer decouples, the budget study will focus on the lower, surface mixed layer rather than the entire boundary layer.
6. At least 10% of the flight hours and ship days will be devoted to instrument intercomparisons between platforms.
Nucleation produces transient ultrafine particles in the 3 to 20 nm size range. These particles generally have relatively short lifetimes, because they readily grow by condensation of gases and/or coagulation to form larger particles. The presence of appreciable numbers of particles ([N]>10,000 cm) in this size range is an indication that a large nucleation event has recently occurred. It is now possible to measure the particle size-distribution in this size range and over time determine a nucleation rate. Comparing this nucleation rate with observations of the time rate of change of the nucleating species concentrations (gas-phase MSA and sulfuric acid) will provide data to test nucleation theories.
These approaches will be used during ACE-1 in both the extremely clean marine boundary layer and in the upper troposphere, the two locations where low existing aerosol surface area makes nucleation most likely. The presence of ultrafine particles (3<Dp<20 nm) can be estimated by comparing the concentration of CN with Dp > 3 nm (as measured by a TSI 3025 or 3027 ultrafine condensation particle counter) with the concentration of CN with Dp > 15 nm (as measured by a TSI 3760 or similar condensation particle counter). Several groups have used a ratio of these concentrations ([N(Dp>15)]/[N(Dp>3)]) as an indicator of nucleation. The C-130, Discoverer, Cape Grim, and Macquarie Island will be equipped to look for high concentrations of ultra-fine particles continuously, thereby generating a time series of their occurrence under a variety of conditions.
In addition to the measurement of total particle concentrations with Dp >3 nm, the C-130, Discoverer and Cape Grim will be equipped to measure the size distribution of the ultra-fine particle mode. The C-130 will use both a system comprised of four ultrafine particle counters tuned to have slightly different cutoff sizes that will generate an aerosol number distribution in four size ranges between 3 and 12 nm and a radial DMA which measures the number distribution from 3 to 500 nm. Discoverer and Cape Grim will use UDMPSs to obtain the size distribution in eight bins from 3 to 20 nm. These instruments should provide data to directly calculate the nucleation rate.
An extremely important corollary to particle production is the simultaneous measurement of the gases that may be responsible for the nucleation: sulfuric acid, water, and perhaps ammonia. Aboard the C-130 gas phase sulfuric acid and MSA will be measured with an API-MS. In addition, ammonia and sulfur dioxide concentrations will be measured aboard the C-130 using a LIF spectrometer and GC-MS, respectively. Aboard Discoverer, ammonia and sulfur dioxide concen-trations will be measured using fluorescence techniques. Thermally conditioned TDMAs will be used on Discoverer and at Cape Grim to quantify the ammonium to nss sulfate molar ratio of these newly formed particles.
The C-130 will conduct flight profiles to thoroughly sample the two regions where aerosol surface area is lowest, thus making nucleation most likely: the clean marine boundary layer south of Australia and the upper free troposphere, where ascending air has recently been stripped of most of its aerosol surface area. During both Pacific transects (Figure 3), the C-130 will search for evidence of nucleation over upwelling and subsiding regions at 9 km altitude. Discoverer will make these same measurements in the boundary layer from Seattle to Hobart.
5.5.2 Aerosol growth and processing
The number and mass size-distributions of each chemical constituent of an aerosol are the result of its formation and processing mechanisms. Observed size distributions and condensable vapor concentrations will be used as inputs for aerosol models to derive information about aerosol growth and processing (Eulerian observations). In addition, the Lagrangian studies will be used to observe changing aerosol number and mass size-distributions as a function of time.
Multijet impactor samples from Cape Grim, Macquarie Island, and Discoverer will be used to determine the mass size distributions of the major anions and cations, aerosol mass, and organic and elemental carbon. A series of instruments, including DMPS's, APSs, and OPCs will measure the number size distribution of the aerosol for comparison to the chemical analyses. The condensable vapors, sulfur dioxide and ammonia, will be measured at Cape Grim and aboard the R/V Discoverer.
On the C-130 only two chemical size cuts will be available, but the complete number size distribution will be measured. A thermally-conditioned OPC will be used to infer size information for several species, including sulfate, soot, and sea salt. Several condensable vapors will be measured, including sulfuric acid, sulfur dioxide, MSA, and ammonia. Although most flights will be flown in clear air, the cloud local closure experiments will provide data in-cloud and along the edges of cloud decks that are needed to evaluate the effect of cloud processing on aerosol size distributions.
5.5.3 Aerosol removal
Wet removal of aerosols usually dominates over dry deposition, but it is difficult to collect statistically valid samples of precipitation over broad marine regions. Wet removal terms will be approximated by collecting and analyzing precipitation from all surface sites. Most of the C-130 boundary layer patterns will be flown in nonprecipitating air to isolate the effect of dry removal. Budget analyses during the Lagrangian studies will be coupled with the Slinn and Slinn (1980) model and measured size distributions to estimate the magnitude of dry surface deposition of each aerosol species.
5.6.1 Process-oriented models
GIT Sulfur/Aerosol Modules (P. Kasibhatla, W. Chameides, & D. Davis)
The Georgia Institute of Technology process-oriented model of the natural sulfur/aerosol cycle in the MBL is being developed with the goal of examining the linkages between various components in the natural sulfur/aerosol system. In its initial configuration, this model will include coupled gas-phase chemistry, seasalt aerosol chemistry, and aerosol microphysics modules. The gas-phase chemistry module is currently based on a standard CH4-CO-NxOy-HxOy-O3 kinetic mechanism. The seasalt chemistry module is based on the box model of Chameides and Stelson (1992). The model is currently being extended to make it fully time-dependent and one-dimensional in nature, and to incorporate the effect of HNO3 uptake by the seasalt aerosols. The aerosol microphysics module describes the nucleation and subsequent growth of H2SO4-water particles and is based on the model of Lin et al. (1992). The next phase of model development will include the coupling of cloud microphysics and chemistry (Hegg et al., 1992) in order to study the influence of cloud processing on the physico-chemical properties of aerosols in the MBL. These models will then form the basis for developing appropriate parameterizations for use in global chemical/transport models.
AERO2 (F. Raes and R. Van Dingenen)
AERO2, from the CEC Environment Institute in Ispra, Italy, is a box model of the MBL that describes homogeneous gas phase chemistry, aerosol dynamics and in-cloud chemistry (Raes et al., 1993). The model focuses mainly on aerosol dynamical processes like nucleation, condensation, coagulation and cloud processing of sulfuric acid aerosols. It yields the development of the full aerosol size distribution with time. Comparison with ACE-1 data will constrain the uncertainties that exist in the description of the processes mentioned above and how they are affected by the meteorological setting (precipitation, entrainment, etc.). As a box model it is best applied to results of the Lagrangian experiments.
UW- Boundary layer aerosol evolution model (D. Hegg and M. Baker)
A University of Washington research team proposes a coordinated program of theoretical investigations (including numerical modeling) and data analysis to elucidate aerosol evolution in the MBL. The MBL has been selected as a relatively simple scenario for the study of such evolution.
The model for SO2, for example, includes vertical and horizontal advection, DMS flux, dry deposition, and three chemical sinks from reaction with OH, sea salt, and cloud drops. Preliminary assessment of the magnitude of the various terms reveals that, for plausible vertical and horizontal gradients of SO2, the advection terms are comparable to the source term. Hence, determination of the time variation of SO2 will require measurement of horizontal and vertical gradients of SO2. Similarly, because SO2 is a major source term for H2SO4 (and thus particle nucleation and condensational growth), these gradients will impact the particle concentrations in the MBL, even without consideration of likely gradients in H2SO4 and particle number concentrations, which could prove at least equally important.
Desired measurements: Spatial gradients in DMS, SO2, MSA, H2SO4, CN and CCN concentrations in both the horizontal and vertical are needed to model secular trends in CN and CCN concentrations in the MBL. Measurements are also needed under both clear and cloudy conditions (cloudy or clear on at least the mesoscale) since the preliminary calculations hint at quite different time scales for particle formation and loss under these disparate conditions.
MIT MBL Photochemistry Model (R. Prinn, A. Pszenny, and X. Shi)
The Massachusetts Institute of Technology MBL photochemistry model (Donahue and Prinn, 1990) will be used with the ACE-1 data set to investigate the role of reactive NMHCs in atmospheric photochemistry (Donahue and Prinn, 1993). The code is currently being modified to include gas phase sulfur chemistry.
CIT/CMU Marine Sulfur/Aerosol/CCN Model (L. Russell, S. Pandis, and J. Seinfeld)
The California Institute of Technology/Carnegie Mellon University model incorporates the primary sulfur cycle in the marine boundary layer, from the emission of DMS to the atmosphere to the fate of particulate sulfate. The model assumes a well-mixed boundary layer, topped with intermittent stratus clouds. In the gas phase, the processes considered are homogeneous oxidation of biogenically derived sulfur species to products that may then be transferred to the particulate phase by mass transfer and homogeneous nucleation. Aqueous-phase chemistry in both cloud droplets and sea-salt aerosol provides additional oxidation pathways. Processing of the air parcel through non-precipitating stratus clouds results in additional transfer of sulfur dioxide to activated cloud droplets, followed by evaporation. Precipitating clouds are the main means for removal of particulate sulfate from the air parcel. The modeling framework proposed in Pandis et al. (1994) has been used to follow the variation of the key reactants, the DMS gas-phase concentration, the sulfur dioxide gas-phase concentration, the sulfuric acid gas-phase concentration, the nucleation mode number concentration, and the CCN number concentration.
DMS Production Model (A. Gabric)
A time-dependent non-equilibrium model of the marine pelagic food web (Gabric et al., 1993) will use ACE-1 data to study the mechanisms that control the seawater concentration of DMS. The present model is currently being adapted to Southern Ocean conditions.
UPRC Water Nucleation Model (B. Gorbunov, R. Hamilton & K. Sabelfeld)
The Urban Pollution Research Centre (Middlesex University, London) model was developed to study the relationship between physical and chemical characteristics of aerosol particles and their ability to form cloud water droplets in atmospheric conditions. The model is based on a new three stage theory of water embryo formation on aerosol particles with non-uniform surfaces, which contains both soluble and insoluble fraction of particulate matter. The Monte Carlo techniques are used to calculate nucleation rates in a real non-uniform atmosphere, where local supersaturation depends on spatial coordinates and time. The model takes into account coagulation of aerosol particles, adsorption, evaporation and condensation of water vapor, nitric acid, ammonia and some other atmospheric gases, sedimentation of particles and interaction of particles with water droplets. It allows one to calculate the transformation of aerosol particle size distribution spectra in a real atmosphere as well as formation of cloud water droplet spectra.
UU CEM (S. Krueger & S. Siems)
The University of Utah cumulus ensemble model (UU CEM) is a two-dimensional model governed by the anelastic equations in which turbulence other than the large eddies is parameterized with a third-moment turbulence closure. The UU CEM has been used to simulate stratocumulus, cumulus and a Lagrangian scenario over a number of days in which stratocumulus are followed through the transition into trade cumulus. These simulations have provided insight into the circulation of the marine boundary and the response to various large scale forcing.
At present the model employs a delta-four-stream approximation for radiative transfer of both the infrared and solar spectra. Work is underway to include drizzle processes.
5.6.2 Column radiative transfer models
GFDL Radiative Transfer Models (V. Ramaswamy)
1. The high-spectral resolution `benchmark' solar and longwave radiative transfer models at NOAA/GFDL are extremely appropriate for computing the spectrally-dependent radiances and fluxes within the atmosphere and at the surface. The models may be used in two ways. First, aerosol single-scattering parameters as a function of height can be used as inputs to the model to calculate highly accurate solar and longwave fluxes. Second, and what would constitute an important verification of the aerosol radiative effect, the model-computed radiative quantities could be directly compared with available measurements. This would afford a strict test of the multiple scattering effects due to the aerosols and would also enable a systematic column closure experiment. As such, this verification process should enable the narrowing of uncertainties of the marine aerosol radiative effects.
2. The GFDL `benchmark' models are premier radiative transfer tools and have been recognized so in the international scientific community. The solar radiative transfer model is the only one of its kind in existence for computations across the entire solar spectrum. These models have been widely used to calibrate climate and weather prediction model parameterizations. Comparisons of the rigorous models' output with measurements would enhance our confidence in the accuracy of aerosol radiative processes.
3. The measurements required include aerosol extinction, single-scattering albedo and the phase function as a function of height. While a comprehensive set of these parameters will probably not be available during ACE-I, use can be made of as many parameters as become available. Also, radiation measurements are needed - radiances and fluxes, and for different wave lengths. If these measurements are not available, no check of the model computations will be possible; in this case, the model can be used only in the first mode mentioned under (1).
4. The spatial and temporal resolution will have to extend all the way from the small scales to the larger GCM scales. It should be borne in mind that it is at the GCM scales that the connection will be made to the climatic role of the aerosols. Thus, depending on the flexibility and resources, every attempt should be made to push for measurements all the way to the larger spatial scales. Several samples are required, primarily to address the issue of variability, including measurements at several times during the day to examine specifically the effect of the sun angle.
5. Since the aerosol radiative effect (e.g., albedo) is linear in the amount and the optical depth, any uncertainty in determining the sizes and concentrations would translate into an uncertainty of a similar quantitative nature in the reflection. In addition, uncertainties due to nonsphericity and inhomogeneous mixtures would also have to be reckoned with. From a climate effects perspective, our goal should be to eventually reduce the aerosol uncertainty effects to the same order as for the well-mixed greenhouse gases (estimated to be ~10%). This may seem to be a tall order. However, if we also remember that the current uncertainty in aerosol forcing exceeds a factor of two, then clearly there is scope for substantial improvements as a result of the field campaigns.
6. There has to be a vigorous interaction between observationalists and modelers in the use and interpretation of data and model outputs. Agreements or discrepancies in the closure efforts have to be identified unambiguously and the challenge of explaining the results has to be met jointly by the two communities.
UH/NPGS Satellite/Aerosol Model (A. Clarke, J. Porter, B. Huebert, P. Durkee)
A University of Hawaii/Naval Postgraduate School research group will integrate the aircraft and satellite data to perform column closure calculations. The current satellite algorithm used at the University of Hawaii allows for a variable aerosol phase function to account for different types of aerosol being present in the image. The extensive aerosol characterization during the ACE-1 experiment will provide an excellent opportunity to further test this satellite algorithm. Current plans are to use the satellite images, LIDAR profiles, aerosol size distributions, aerosol scattering coefficients, aerosol optical depths, sky radiance and ocean reflection measurements to model the clear sky aerosol optical properties of the ocean-atmosphere system ranging from very clean regions near Tasmania to high aerosol loading conditions in the Hawaii volcano plume.
5.6.3 Cloud scale models
OU/CIMMS Model (Y. Kogan)
The Oklahoma University/Cooperative Institute for Mesoscale Meteorological Studies (OU/CIMMS) stratocumulus cloud model includes coupled 3-D LES dynamical framework and explicit formulation of the microphysical processes of nucleation, condensation, evaporation, coalescence and drop recycling. The current version of the model includes 19 size categories of cloud condensation nuclei in the size range from 0.076 to 7.6 microns and 25 categories of cloud drops with radii in the size range from 1 to 256 microns (Kogan, 1991; Kogan et al, 1994). The model also includes the long wave radiation module that is formulated according to Herman and Goody (1976) and the short wave radiation module that is based on a 24 band solar radiation package of Slingo and Shrecker (1982), as modified by Slingo (1989).
The OU/CIMMS model has been used to study the aerosol indirect radiative effects by simulating ASTEX cases of clean and polluted marine air. ACE data will be used for further model validation, as well as for process studies identified as ACE Objective 2. Of special interest will be the study of the effects of stratocumulus cloud cycling on the transformation of aerosol spectra, as well as the study of climatic effects of aerosols, in general, and DMS-derived aerosols, in particular.
5.6.4 Regional/global models
ECHAM (J. Lelieveld & F. Dentener)
A research group at the Department of Air Quality, Wageningen Agricultural University, The Netherlands have included a tropospheric chemistry scheme in the GCM of the Max-Planck-Institute for Meteorology in Hamburg. The model is being adapted to include heterogeneous chemistry with a simplified representation of aerosol formation and particle size distributions. The radiation scheme distinguishes between 5 aerosol types: urban, remote continental, marine, wind-blown dust and stratospheric. The model includes representations of stratiform and cumuliform clouds with liquid water as a prognostic variable. The standard model resolution is 5.6° x 5.6°, but can be enhanced to less than l° x l°. The goal is to parameterize the direct sulfate-backscattering-climate relationship as well as the indirect sulfate-cloud-microphysics-climate effects. The efforts are part of the European projects GLOMAC (GLObal Modeling of Atmospheric Chemistry) and SINDICATE (Study of the INdirect and Direct Influences on Climate of Anthropogenic Trace-gas Emissions) that are developing chemistry-transport models and coupled climate-chemistry models. ACE-1 data will be used to validate the model.
SALSA (R. Rosset & K. Suhre)
The Laboratoire D'Aerologie, Observatoire Midi Pyrenees mesoscale meteorological model (SALSA) includes detailed parameterizations of cloud physics, radiation, boundary layer and fractional cloudiness, and gas-phase and aerosol chemistry. The model has been run in a variety of different meteorological situations (fronts, breezes, orographic flows, marine (ASTEX) and continental boundary layers) with a resolution of 1 to 50 km in the horizontal and 15 to 80 levels in the vertical. The model is nested in larger-scale models or analyses (e.g. ECMWF products).
The chemistry module in the model includes DMS oxidation in the gaseous phase (85 reactions) and aqueous chemistry ( SO2 oxidation by O3 and H2O2). The aerosol module includes the equilibria between the gaseous, aqueous and solid phases for H2O, H2SO4, HNO3, NH3 and their salts.
ACE-1 data will be incorporated into the model to produce vertical and horizontal profiles of the relevant species and their evolution. To verify the model, measurements are needed of CN, CCN, aerosol composition and size distribution, DMS, SO2, at different altitudes, and below, above and within cloud decks.
Brookhaven Global Chemistry/Aerosol Model (S. Schwartz, C. Benkovitz)
The model (Benkovitz et al, 1994) is derived from the PNL Global Chemistry Model modified to be driven by observation-derived meteorological data, specifically, at present the archived output of the 6-hour forecast model of the European Centre for Medium Range Weather Forecasts (ECMWF), but other sources could be employed. The latitude-longitude grid is 1.125°, and the ECMWF h-coordinate with 15 levels, surface to about 100 hPa, is used as the vertical coordinate. Model time step is 1 hour; resolution of the meteorological data is 6 hours. Vertical velocities are calculated using the horizontal winds and divergence; vertical distribution of condensed moisture is calculated using quasi-steady-state precipitation scavenging models. Material balances are determined by the continuity equations, solved with the application of gradient-transport assumptions. Local abundance of chemical species is represented as mixing ratio for both gaseous and aerosol species. Model domain is subhemispheric (140°W to 62°E; 12°N to 81°N) at present. It is intended to extend this to the entire 0° to 81°N zonal band. A zonally complete model seems necessary to accurately represent inflows. A southern hemisphere model will be developed in conjunction with ACE-1.
At present the model is restricted to sulfur species. Gas- and aqueous-phase reactions are treated. Gas-phase chemical reactions represented in the model are the OH-induced oxidation of SO2 to sulfate and of dimethylsulfide (DMS) to SO2 and methanesulfonic acid (MSA). Clouds within the model provide an environment for aqueous chemical reactions, wet removal, and in the case of convective clouds, mixing and subgrid scale vertical motions. The present model treats aqueous-phase reactions in precipitating clouds only. SO2, H2O2, O3 dissolve in cloudwater based on their solubilities, with a cloudwater pH of 4.5. Aqueous-phase reactions include the oxidation of SO2 to sulfate by H2O2 and O3. Climatological concentrations of O3 are employed; H2O2 is generated in the gas phase at a fixed rate until a seasonally dependent maximum concentration is reached. Oxidation of SO2 by H2O2 is treated as going to completion limited by the lesser of the two concentrations; oxidation by O3 proceeds only if there is any remaining SO2. Future model development will incorporate representation of soil dust, carbonaceous aerosol, nitrates, and ammonium, as well as a representation of particle size.
Dry deposition is represented as the lower boundary condition for the vertical transport. Wet removal is evaluated using a fractional removal for cloudwater determined by precipitation amount and column integral of cloudwater. Time- and location-dependent dry deposition velocities for SO2 and sulfate are calculated using the ECMWF meteorological data, seasonal albedo, land use classifications, and seasonal categories.
Model output is concentration of sulfate aerosol, distinguished by source region and type as a function of latitude, longitude, height, and time. Model output can be used to generate sulfate aerosol optical depth for comparison with contemporaneous measurements.
Desired Measurements: Non-seasalt sulfate concentration; volatile and elemental carbon concentration; aerosol size distribution 0.05 to 1 m radius; aerosol optical depth. Concentrations of sulfate precursors, specifically DMS, SO2, and oxidants H2O2 and to lesser extent O3. Concentration of NH3 will be valuable for evaluation of future version of the model. Enhanced meteorological measurements should be incorporated in real-time inputs to ECMWF to improve model accuracy.
ACE Objectives. Reducing uncertainty in aerosol modeling is central to the ACE objectives. Running this model for time periods of ACE experiments and comparing model with observation will thus make immediate use of ACE data in evaluating model uncertainty.