back to the book content

 

Advances in Measurement Methods for Atmospheric Chemistry

Lead authors :

G.S. Tyndall, D. Winker, T. Anderson, F. Eisele
To download the text of this chapter in PDF format, click here, and for the figures click here.

Table of Contents

SECTION 1: LABORATORY METHODS RECENT ADVANCES Ozone/HOx Photochemistry
Isoprene / Terpene Oxidation Mechanisms
Aerosol Formation / Heterogeneous Chemistry
Instrument Intercomparisons FUTURE TRENDS Optical Spectroscopy
Hydrocarbon Oxidation
DMS Oxidation
Heterogeneous Chemistry
Intercomparisons SUMMARY

SECTION 2: IN SITU AEROSOL MEASUREMENTS

INTRODUCTION
RECENT ADVANCES Parameters requiring measurement
Observational strategies
Method development Light scattering
Mass scattering (and absorption) efficiencies
Single scattering albedo
Aerosol composition and size distribution
Sampling inlets Field deployment Long-term monitoring
Transects and vertical profiles
Intensive campaigns SUMMARY AND FUTURE TRENDS

SECTION 3. SATELLITE INSTRUMENTS FOR TROPOSPHERIC CHEMISTRY

INTRODUCTION
RECENT ADVANCES Ozone and other trace gases
UV/Visible Techniques
Infrared Techniques FUTURE TRENDS Observations from Geostationary and Lagrange Orbits
Active Sensing of Aerosols and Trace Gases SUMMARY
ACRONYMS:

SECTION 4. IN SITU GAS PHASE MEASUREMENTS

RECENT ADVANCES
FUTURE TRENDS

REFERENCES

LABORATORY METHODS
IN SITU AEROSOL MEASUREMENTS
SATELLITE INSTRUMENTS FOR TROPOSPHERIC CHEMISTRY
IN SITU GAS PHASE MEASUREMENTS

 

Section 1: Laboratory Methods

Geoffrey S. Tyndall, John J. Orlando and Jack G. Calvert, NCAR contact: tyndall@ucar.edu

Recent advances

Our knowledge of the details of tropospheric chemistry has increased tremendously over the last ten years. This increase is in part related to the sheer volume of work done, but is also a consequence of the incorporation of new measurement techniques. The largest source of data has been the environmental or smog chambers, which provide integrated pictures of the overall chemistry. However, the methodology often suffers from an inability to isolate the particular reaction of interest. Furthermore, wall reactions are often irreproducible, and differ from chamber to chamber. On the other hand, direct, time-resolved spectroscopic methods can be designed to study a single species undergoing a single reaction. However, the conditions under which that species can be studied are often far removed from those encountered in the atmosphere, and care must be exercised when extrapolating the results to the real world. Clearly, the most progress has been made where direct and indirect methods have been integrated to test our understanding at both the microscopic and macroscopic levels. Some of the major advances in laboratory techniques are highlighted below, along with future trends.

Ozone/HOx Photochemistry
The photolysis of ozone to produce O(1D), followed by the reaction of O(1D) with H2O, provides the largest source of OH in the lower atmosphere. The study of the quantum yields for ozone photolysis (in particular in the energy-deficient region beyond 310 nm) has long fascinated physical chemists and has even been the subject of some controversy over the years [Ravishankara et al., 1998]. However, in the last several years, major advancements have been made in our understanding of the processes involved in ozone photolysis. The techniques used have included traditional methods such as resonance fluorescence or laser induced fluorescence, but have also included more advanced techniques borrowed from the chemical physics community. For example, the excited O(1D) and O2(1D) photofragments have been detected using a Resonance Enhanced Multi-Photon Ionization (REMPI) technique [Ball et al., 1993] or by vacuum ultraviolet laser induced fluorescence [Takahashi et al., 1996]. Studies conducted in molecular beams have allowed the ozone to be prepared with a minimal amount of internal energy, so that the wavelength of the threshold can be determined with high accuracy. The direct detection of all possible photofragments, O(3P), O(1D), O2(1D), and O2(3S), in all possible combinations has now been accomplished and the presence of a clear tail in the O(1D) quantum yields beyond 310 nm (the result of both the photolysis of internally excited ozone and the occurrence of spin-forbidden photolysis pathways) has firmly been established. Corroborative evidence for the existence of this long wavelength tail has been obtained from in situ measurements of the O(1D) production rate using chemical actinometry [Shetter et al., 1996]. Calculated photolysis rates only show a good agreement with such actinometer measurements (particularly at high solar zenith angle) when the long-wavelength tail is incorporated into the calculation.
A great deal of improvement has been made in the database for the kinetics of the reactions of OH with organic species (including hydrocarbons, halocarbons, and organosulfur compounds [Atkinson, 1997a]. Improvements in measurement techniques have allowed OH to be measured in pressures up to 600 Torr O2/N2 in both flash photolysis [Hynes et al., 1995; Brown et al., 1999] and flow tube systems [Abbatt et al., 1990], in spite of the rapid quenching of the fluorescence that occurs under these conditions. The versatility of flow tube experiments has been extended by the use of high pressure turbulent flow systems [Abbatt et al., 1990, Seeley et al., 1996]. The database has particularly been improved at low temperature, in large part due to the care taken in sample purification [Vaghjiani and Ravishankara, 1991].

Isoprene / Terpene Oxidation Mechanisms
Much of the progress in isoprene chemistry has occurred using traditional smog chamber techniques, or variations thereupon. While the major pathways to methylvinyl ketone and methacrolein have been quantified by Atkinson and coworkers using Fourier Transform Infrared Spectrometry/GC [Tuazon and Atkinson, 1990] and by Paulson et al. [1992] using GC , there still exist many uncertainties in the yields of minor products, particularly under conditions of low NOx. The atmospheric pressure ionization (API) mass spectrometric technique has been used by Atkinson's group to identify the presence of hydroxycarbonyls not detected by traditional techniques (Kwok et al., 1995). Jeffries and co-workers [Yu et al., 1995] have implemented the technique of derivatization using O-(2,3,4,5,6-pentafluorobenzyl) hydroxylamine hydrochloride (PFBHA) which enables the identification of many reactive carbonyls and acids which otherwise would not be amenable to detection by standard chromatographic methods [Yu et al., 1998; 1999]. Both these new techniques need some development before they can be considered quantitative, but offer the possibility of quantifying multifunctional products. The oxidation of terpenes has been less well studied, with each of the terpenes subject to only 2 or 3 investigations. The range of reaction products is wide, and the mass balances obtained are poor. The chemistry of the oxygenated reaction products remains poorly quantified.
Much work in the last few years has been dedicated to understanding the chemistry of alkoxy radicals [Atkinson, 1997b]. Alkoxy radicals provide a branching point in the oxidation of most hydrocarbons, and the competition between the various reaction pathways determines the rate at which the carbon chain gets shortened, and influences the degree of oxidation, and ultimately the solubility, of the stable reaction products formed. The majority of studies have been performed in smog chambers by studying the relative rates of formation of the various products. However, recent chamber studies at lower temperatures have indicated that the oxidation chemistry of even simple organics such as ethene may be more complicated than inferred from room temperature studies [Orlando et al., 1998]. A few studies using direct spectroscopic detection have been performed, but these are limited to small alkoxy radicals derived from alkanes [Devolder et al., 1999; Blitz et al., 1999].

Aerosol Formation / Heterogeneous Chemistry
Several improvements in instrumentation have led to a better understanding of aerosol formation and growth at the microscopic level. Kulmala and coworkers have performed both experimental and theoretical studies on nucleation of sulfuric acid water mixtures. Ball et al. made use of an ultrafine condensation nuclei counter to detect particles in the early stages of growth (diameter > 3 nm), and thereby infer the nucleation rate. The use of Chemical Ion Mass Spectroscopy (CIMS) made detection of very low concentrations of sulfuric acid possible. Both these groups concluded that the functional dependence of the nucleation rate on water and sulfuric acid was much stronger than given by older theory, but in reasonable accord with field experiments. Nucleation was also strongly enhanced in the presence of NH3. The electrostatic balance (particle levitation) has been used to follow the deliquescence behavior of NH4HSO4 and (NH4)2SO4 particles [Imre et al., 1997; Xu et al., 1998]. In this technique, a charged particle is trapped by an electric field, its mass is measured by the magnitude of the restoring force, and its physical state inferred from measurement of scattered laser light.
A variety of different aerosol detection techniques have been applied to quantify the formation of aerosol from terpene oxidation. While the potential for the formation of aerosol particles from biogenic hydrocarbons has long been recognized, descriptions of particle growth are still largely empirical, describing the yield of particulate matter as a function of the extent of oxidation [Odum et al., 1996; Hoffmann et al., 1997]. However, it is not clear whether this parameterization can be simply transferred to atmospheric conditions. It is thought that particle nucleation, or initial growth on seed nuclei, involves the deposition of low-volatility products, often formed in small amounts, followed by ad- or absorption of more volatile products into the organic layer formed.
Tropospheric heterogeneous chemistry is a relatively new and untapped research area. Atmospheric pressure flow reactors have been used to measure the coefficients for reaction of gaseous N2O5 directly on monodispersed, submicron sulfuric acid aerosols [Mozurkewich and Calvert 1988; Fried et al., 1994; Lovejoy and Hanson, 1995]. The use of the measured reaction coefficients in tropospheric modeling studies [Dentener and Crutzen, 1993] indicates strongly that heterogeneous reactions in the troposphere can be very important; for example, it is estimated that for January in the troposphere north of 45°N, more than 90% of all NOx is removed by the heterogeneous reaction, N2O5 + H2O(aerosol) -> 2HNO3. Much progress has been made in understanding the uptake of organics onto aqueous or acidic aerosol using the falling droplet technique pioneered by Aerodyne/Boston College [Jayne et al., 1992]. These experiments often yield values for both the uptake parameters and the solubility constants, and the suite of molecules investigated includes carbonyls, acids, alcohols and sulfur gases. Less is known about the chemical transformations that occur in aqueous tropospheric aerosols. While the aerosol offers abundant surface area for uptake the small volume may limit the extent to which gas-phase chemistry may be influenced. However, the very concentrated solutions encountered could lead to rapid reaction rates under the right circumstances. The interaction of common atmospheric oxidants with organics, and particularly the effects of ambient light need to be investigated in detail before a full understanding of the impact of aerosol on biogenic carbon (and vice versa) can be determined.

Instrument Intercomparisons
Instrument intercomparisons are vital and indeed as much a part of doing atmospheric research as is gathering data to test a geophysical hypothesis. A reliable estimate of the uncertainty in a measurement is as important as the measurement itself. A great variety of intercomparisons have been carried out during the past ten years. Techniques and methods to measure many of the atmospheric gases have been evaluated quantitatively; for example: NO2 [Fehsenfeld et al., 1990]; CH2O [Gilpin et al., 1997]; and hydrocarbons [Apel and Calvert, 1994; Apel et al., 1999]. Atmospheric scientists should continue to show an honest skepticism about the accuracy of measurements derived from any new atmospheric component measurement technique until such time that it has been proven through adequate testing and evaluation in blind intercomparisons and through other careful laboratory and field testing.

Future trends

Optical Spectroscopy
The use of cavity ring-down spectroscopy has been demonstrated as being a very sensitive way to detect free radicals. Since it is an absorption technique it is not subject to the effects of quenching as encountered with fluorescence techniques, while the sensitivity may approach that of LIF for larger molecules. Other advances in laser technology (diodes, near UV, near IR) could lead to improvements in techniques for detecting peroxy and/or alkoxy radicals. Part of the reason for the improvements in precision in the last ten years is related to the wider availability of better lasers and signal acquisition hardware. It should also be noted that a better detection sensitivity allows lower radical concentrations to be used, which in turn minimizes the effects of secondary reactions. As further developments in instrumentation occur they will undoubtedly be incorporated into laboratory systems, too.

Hydrocarbon Oxidation
Much work still needs to be done on the chemistry of oxygenated reaction products, particularly multifunctional carbonyls. Derivatization techniques for carbonyls are constantly being improved and incorporated into both laboratory and field studies. Mass spectrometric methods (API, CIMS) have rapid time response and high sensitivity, and could be used in time- resolved laboratory studies if appropriate conditions can be found. The sensitivity and selectivity of CIMS has been demonstrated both for stable molecules [Lovejoy and Hanson, 1995] and for free radicals [Villalta and Howard, 1996].
The chemistry of alkoxy radicals remains central to our understanding of hydrocarbon chemistry. Most of the work to date has been done using competitive product studies. Some direct spectroscopic techniques have been used, but at present they tend to be labor intensive. The development of more general techniques, applicable to several different radicals, would be useful.

DMS Oxidation
There still exist many problems in identifying and quantifying the intermediate radicals in DMS oxidation. The intermediates studied directly include CH3S by LIF [Tyndall and Ravishankara, 1989; Turnipseed et al., 1992] and CH3SO by mass spectrometry [Mellouki et al., 1988; Dominé et al., 1990]. Both mass spectrometric and optical techniques should be developed to further investigate the reaction mechanisms. The key branching steps to the formation of SO3 and MSA and their dependence on temperature and NOx [Sørensen et al., 1996; Patroescu et al., 1999] need to be identified through direct laboratory studies in conjunction with chamber studies.

Heterogeneous Chemistry
Continuing laboratory studies should elucidate the details of nucleation of particles, particularly those comprised largely of organic materials. Most studies to date have been on model mixtures on binary mixtures, e.g., sulfuric acid water. A better understanding of the conditions required for nucleation will enable the extent of new particle formation in the atmosphere to be assessed. It is particularly important to isolate the effects of the individual oxidants, e.g., OH, NO3 or O3 since these oxidants tend to be important in different regions of NOx or solar intensity, and each leads to products of differing volatility.
Formation of intermediate clusters. Study by mass spec; see particle forming. Single particle studies.
The question of chemistry inside, and particularly on the surface of, particles needs to be addressed more systematically. Currently, the identity and concentrations of oxidants within aqueous aerosol particles have not been thoroughly studied. The oxidation of partially oxygenated compounds removed from the gas phase could continue inside droplets and influence not only the measured concentrations but also the physical properties of the particles. Finally, the occurrence of reactions on the surface of (organic) aerosol needs to be first demonstrated and then thoroughly investigated. This is a new field which has been the subject of speculation for many years, but which has not been quantified satisfactorily to date.

Intercomparisons
Several different techniques of measurement of the OH radical concentration in the troposphere have been developed, for example the ion- assisted OH measurement [Tanner et al., 1997], laser induced fluorescence [Hofzumahaus et al., 1996; Mather et al., 1987] and long path absorption spectroscopy [Dorn et al., 1996a]. These techniques have been improved greatly during the past ten years, and their use is now extensive throughout the world. It is suggested that the atmospheric scientists plan an IGAC sponsored intercomparison involving measurement of the most important atmospheric transient species, OH radical, using blind tests in air mixtures prepared in a common manifold in the laboratory and in field studies at a common site under a variety of atmospheric conditions.

Summary

Tropospheric homogeneous and multiphase chemistry have both seen impressive expansions in the last ten years. Tropospheric chemistry, particularly relating to biosphere-atmosphere interactions, tends to be complex and, to date, not enough studies have been carried out to fully test the reproducibility of many of the different systems studied in the laboratory. In this sense, the discipline is still immature, and much work remains to be done to validate the current database and further extend it. Our understanding of heterogeneous and multiphase systems is largely based on data from bulk systems at the moment. Further developments in measurement techniques will undoubtedly lead to a continued appreciation of the complexities and subtleties of the chemical reactivity in the atmosphere.

 

Section 2: In situ Aerosol Measurements

Theodore L. Anderson and David S. Covert, University of Washington, Seattle, WA contact: tadand@u.washington.edu

Introduction

In this section we examine the role of in-situ aerosol measurements1 in addressing global-scale scientific questions. To do this in a coherent way, we focus on a single, well-defined problem - direct climate forcing by anthropogenic aerosols2,3 . As a result of this focus, only a small subset of aerosol measurement techniques will be mentioned. Readers interested in a general review of aerosol measurements are referred to the excellent treatment by McMurry (1998).
A major development in climate research over the last decade has been the recognition that anthropogenic aerosol particles - and not just greenhouse gases - are significant players in altering the Earth's energy budget (IPCC, 1996). For example, Hansen et al. (1998) have referred to a "paradigm change" wherein "uncertainties in climate forcings have supplanted global climate sensitivity as the predominant issue." Figure 2 from that paper indicates that the sign of the net global forcing over the industrial era is unknown, within current uncertainties, and that the largest uncertainties involve anthropogenic aerosols. Thus, it is appropriate that improved quantification of aerosol climate forcings is a major focus of current research. Even for the direct aerosol effect (the best-posed problem), this is a challenging observational undertaking, given the spatial non-uniformity of aerosol particles and their complex and highly variable optical properties. How well has the measurement community responded to this challenge over the last decade? What are the prospects and challenges for the future? We will address these questions not only in relation to instrumental techniques but also in terms of the development of a practical observational strategy.


Figure 2. Global distribution of CO total column for June 16-19, 1997 as retrieved from cloud-filtered IMG spectra. Dots corresponding to the measured pixels were enlarged on the plot for clarity. Source: C. Clerbaux, personal communication.

Recent advances

Parameters requiring measurement
A cogent argument for the importance of climate forcing by anthropogenic sulfate aerosols was put forward in 1990 (Charlson et al., 1990). Subsequent papers confirmed and refined this argument (e.g. Kiehl and Briegleb, 1993) and extended it to other anthropogenic aerosol components, primarily biomass smoke (Penner et al., 1992), mineral dust from disturbed regions (Sokolik and Toon, 1996), and light-absorbing soot (Haywood and Shine, 1995). As a result of these and many other studies, direct climate forcing by anthropogenic aerosols was shown to be important as well as highly uncertain. Improved quantification was framed in terms of a set of critical, measurable properties of the aerosol (Penner et al., 1994; Ogren, 1995). Table 1, adapted from Ogren (1995), lists the most important of these parameters, separating them into extensive and intensive categories 4 .

Table 1: Measurable parameters for quantifying direct aerosol forcing
a. Extensive parameters

ssp(l), sbsp(l),
sap(l)

Mi

t(l)

bp(l)		 Scattering, hemispheric backscattering, and absorption components of
the aerosol light extinction coefficient, sep(l) (also called the volume
extinction cross-section), at wavelength l, all with units of m-1 .
Mass concentration of aerosol component i (or of the total aerosol), with
units of mg m-3 .
Aerosol optical depth, defined as the vertical integral of sep(l),
dimensionless.
180° backscattering coefficient, with units of m-1 sr-1 .


b. Intensive parameters
b(l)
w(l)
å

CFi(l)


fj(RH)

ai,j(l)

mi*(r)

Sa(l)		 The hemispheric backscatter fraction, defined as sbsp(l)/ssp(l).
Aerosol single scattering albedo, defined as ssp(l)/sep(l).
Ångström exponent describing the wavelength dependence of
extinction, defined as the exponent in the functional relationship, sj ~
l-å where sj can refer to ssp(l), sbsp(l), or sap(l).
The fraction of sj due to coarse mode particles, where the coarse
mode is operationally defined as all particles with low-RH
aerodynamic diameters larger than 1 µm and sj can refer to ssp(l),
sbsp(l), or sap(l).
Hygroscopic growth factor for parameter j, a multiplier that describes
dependence of the parameter on relative humidity (RH). Parameter j
can be ssp(l), sbsp(l), sap(l), b(q,l), Mi, or mi*(r).
Mass extinction efficiency for component i and parameter j, defined as
, si,j = fsj / fMi,
where sj can be ssp(l), sbsp(l), or sap(l), with units of m2 g-1
Normalized distribution of the mass of component i (or of the total
aerosol) with respect to particle radius, r, defined as,
mi*(r) =1/Mi fMi / flog(r)
Aerosol extinction-to-backscatter ratio, defined as sep(l)/bp(l).

Observational strategies
Table 1 begins to show the complexity of the observational challenge. The dimensions over which aerosol properties display important variations include space (from turbulent to continental scales), time (from turbulent to interannual scales), height, relative humidity, chemical composition, size parameter (particle size relative to wavelength), and scattering angle. Thus, the observational task amounts to filling in a seven-dimensional matrix of required information, where five of the dimensions have of order 10 elements, the geographical/spatial dimension has 4 million elements (for 10 km global resolution), and the time dimension has 9000 elements (for hourly time resolution over one year.) Clearly, a brute force approach is not an option. Rather, a feasible measurement program must be highly strategic in nature and must involve tight coupling with both satellite observations and numerical models. The measurement strategy has two basic elements: the development of suitable measurement methods and their deployment in the field.

Method development
In-situ instruments and sampling techniques must be developed to enable demonstrably accurate measurements of a small but highly leveraged subset of optical properties. These measurements are needed for initializing and testing aerosol optical models, chemical transport models, and satellite retrieval algorithms. Two difficult and frequently neglected aspects of method development are critical to the direct forcing problem. Standardization of instruments, sampling methods, and data reduction procedures is required since measurements from different times and places must be comparable. In addition, robust procedures for determination of uncertainties are required both to allow measurement error to be distinguished from ambient variability and so that the degree of constraint imposed by the measurements on models and satellite retrievals is known. Much of the progress during the 1990's has been in these two areas.

Light scattering
A large fraction of the parameters listed in Table 1 flow from the measurement of light scattering by nephelometry. Nephelometer technology has steadily evolved over the past 50 years (Heintzenberg and Charlson, 1996), with demonstrated accuracy improving from roughly a factor of two in the 1960's to around 10% in the 1990's. Recognizing the need for a standardized instrument, a group of scientists collaborated with TSI, Inc. in 30 the early 1990's to develop a high-sensitivity, commercial version. This instrument (TSI model 3563) has been subjected to rigorous noise and calibration tests (Anderson et al., 1996) and protocols for sampling, field calibration, and data reduction have been published (Anderson and Ogren, 1998). Since becoming available, it has been broadly embraced by the measurement community and is now routinely deployed in field experiments by a large number of research groups.
The model 3563 nephelometer measures ssp and sbsp at three wavelengths spanning the visible range (450, 550, and 700 nm). It therefore allows a direct calculation of backscatter fraction, b, which is a proxy for the asymmetry parameter, g, used in many radiative transfer models (Marshall et al., 1995). In addition, it allows å to be calculated over two discrete wavelength ranges.
When combined with size segregation at 1 µm diameter as recommended by Anderson and Ogren (1998), the nephelometer allows a calculation of the coarse-mode fraction of scattering, CFsp. This parameter is a useful predictor of many other optical properties, since the optical characteristics of sub- and super-micron particles are quite distinct. In addition, it is strongly related to aerosol composition and source, since super- micron aerosol mass is generally dominated by mechanically-produced particles (e.g. dust, seasalt, flyash, or vegetative debris) while sub-micron mass is generally dominated by secondary or combustion-derived particles (e.g. sulfates, organics, or soot).
Because of its enclosed chamber, the nephelometer is suitable for controlled-RH measurements (Covert et al., 1972), which allows a direct determination of fsp(RH). Nephelometer measurements also contribute to determining component scattering efficiencies, ai,sp, and single scattering albedo, w, discussed below.
Recently, nephelometer technology has been extended to the measurement of the 180° backscatter coefficient (Doherty et al., 1999), permitting a tight coupling of in-situ and lidar techniques. In particular, it permits a calibrated, in-situ measurement of the extinction-to-backscatter ratio, Sa, which is the primary aerosol property required for quantitative lidar retrievals.

Mass scattering (and absorption) efficiencies
The scattering efficiency of an aerosol component is the essential parameter linking its mass concentration (which can be predicted from source emissions and chemical transport models) to the scattering of visible light for calculations of its direct radiative forcing. In a review of this topic, focused on the example of sulfate scattering efficiency, Charlson et al. (1999) describe a number of important and often subtle complications. In part, these arise from inherent difficulties in any attempt to apportion a population effect among members of that population5 . However, Charlson et al. show that consistent and useful results can emerge from a strategy based on practical realities of both aerosol chemistry and aerosol instruments. Approaches to measuring scattering efficiencies divide into two, complementary categories: (1) multiple linear regression of the scattering coefficient against aerosol component masses and (2) Mie calculations based on the size distribution of chemical components combined with other information on refractive index and state of mixture. For sulfate, Charlson et al. show that consistent results emerge from these approaches and support about a factor-of-two range in values for anthropogenically-perturbed aerosols. This consistency requires, however, that the sulfate component includes the masses of chemically bound cations and water but not but masses of other aerosol compounds that are merely correlated with sulfate due to collocated sources or common sinks. In this sense, the sulfate ion - which can be measured readily and with high accuracy - is used as an indicator species for the sulfate component but must not inadvertently be taken as an indicator for pollution aerosol in general.
Other issues discussed by Charlson et al. include relative humidity effects, the need to treat ammonia species differently from sulfates, and the need to assess covariances between scattering efficiencies and other parameters that control direct radiative forcing.
For the case of light absorption, an alternative to the mass-based approach has been proposed by Bond et al. (1998). These authors show that current estimates of mass emission rates for black carbon (the dominant aerosol absorber) are fraught with difficulties and are likely to be seriously in error. As a complementary (and possibly preferable) approach, they propose that the emission rate of light absorption be measured directly for various types of combustion and that transport models use light absorption, not black carbon mass, as the prognostic variable. Demonstration of this technique at a coal-burning power plant revealed emission rates of light absorption that were an order of magnitude lower than what would be inferred from published mass-based emission factors.

Single scattering albedo
The radiative effects of atmospheric aerosols are critically sensitive to the relative amounts of scattering and absorption, expressed in the parameter w (Haywood and Shine, 1995). While the measurement of ssp is relatively mature and accurate (see 1.1), the measurement of sap is not. The latter, therefore, is what limits the accuracy of current measurements of w.
The measurement of the aerosol light absorption coefficient, sap, by differential transmission photometry has a long history (dating back at least to Lin et al., 1973), but a standardized, calibrated, real-time measurement has only recently become available. This instrument (the Radiance Research, Inc. PSAP) has been subjected to noise, calibration and intercomparison tests (Bond et al., 1999; Reid et al., 1998) and a detailed set of field use and data reduction protocols have been published (Anderson et al., 1999). Many research groups have adopted this instrument, bringing a measure of standardization to the field.
Significant problems remain, however, regarding the accurate determination of sap (and, thus, w). First, the optical properties of particles on a filter cannot fully mimic those of suspended particles. If the response depends on particle morphology or chemistry, the calibration with laboratory- generated particles may be inaccurate. Recognizing this possibility, Bond et al. (1999) estimate that the 95% confidence uncertainty of the PSAP is at least +/-20%. Similarly, intercomparison tests by Reid et al. (1998) for biomass smoke aerosols were inconclusive but suggested a PSAP uncertainty of +/-40%. Calibration experiments with a variety of absorbing aerosol types are needed. The PSAP is also limited in that it measures at only one wavelength (the calibration applies to 550 nm), although the technique could readily be applied to other wavelengths. Finally, by collecting particles on a filter prior to measurement, the PSAP does not permit studies of the humidity dependence of light absorption, fap(RH). Indeed, there is no available method for measuring this important parameter.
Recent improvements in the photoacoustic method (Moosemüller et al., 1997; Arnott et al., 1999), which senses light absorption by suspended particles in an enclosed volume, may offer a route to both improved accuracy and the measurement of fap(RH).

Aerosol composition and size distribution
The distributions of particle number, mass and composition as functions of size are needed to build aerosol optical models for closure tests and for extrapolating measurements across wavelength and scattering angle. Regional- to global-scale versions of these models are presently based on a very limited number of measurements with large uncertainties (see, e.g., Pan et al., 1997). Considerable improvements in measurement accuracy have occurred in the 1990's. Sources of uncertainty have been compiled by Quinn et al. (1996). Measurements with these newer methods have been made in many regions, revealing large-scale patterns (Covert et al., 1996; Porter and Clarke, 1997) that could form the basis for improved, global-scale models. Some caution, however, is warranted. Already there is a common tendency to fit size distributions with analytical functions without evaluating the resulting errors. These errors can be large, particularly when converting between moments of the distribution (e.g. from number to mass). A much needed, but difficult, first step is to compile the existing measurements (from many different campaigns and investigators) into a common data base for general use.

Sampling inlets
Coarse mode aerosols (largely mineral dust and seasalt) dominate the variation in optical properties over many regions and are known to play significant roles in the cycling of secondary aerosol species (e.g. biogenic or anthropogenic sulfate.) Thus, a major problem for in-situ measurements is the loss of coarse mode aerosols during sampling from high-speed aircraft (Huebert et al., 1990; Huebert et al., 1998). The losses are significant for diameters above 1 or 2 µm and may be 90% or more for diameters larger than 5 µm. No solution has yet been found (other than sampling at low speed) - indeed, a definitive means of calibrating inlet efficiency as a function of size has yet to be developed. A solution to the calibration problem might lie in the use of identical, non-aspirated, particle sizing probes both outside the aircraft and inside the sample plumbing (see, e.g., Noone et al., 1992). A promising approach to solving the inlet loss problem is currently being pursued. This approach involves removing the turbulent surface layer in the deceleration zone by sucking a large fraction of the flow through porous walls. The resulting inlet would produce size-dependent, over-sampling of large particles, which would require accurate calibration. However, this would be a much more manageable problem than under-sampling and would have the advantage of improving large particle counting statistics.

Field deployment
The measurement strategy must select a judicious combination of long- term and intensive deployments from ground-based, ship-based, and airborne platforms. Goals are to characterize the major aerosol types and dominant scales of variability. In addition, the measurement plans should be optimized to provide rigorous tests of internal consistency (e.g. among methods and platforms) and to detect errors in transport models and satellite retrievals. Intensive field campaigns have been the focus of effort in the 1990's and will, accordingly, be discussed in greatest detail.

Long-term monitoring
Measurements at ground-based stations can be used to develop instrumental methods, to study temporal variations in aerosol properties over all time scales, and to serve as baseline stations for intensive campaigns. There are presently about a dozen such stations (e.g. Cape Grim baseline station) that conduct routine measurements of aerosol optical properties intended for studying direct climate forcing. Ideally, the number and location of these stations would span the full range of regional aerosol types and the sampling and measurement protocols would be identical at all stations such that their results could be strictly compared. It is important, but difficult, to assess the vertical and regional representativeness of in-situ measurements from fixed surface stations. One promising approach to this problem involves coupling to collocated lidar measurements.

Transects and vertical profiles
Ship-based measurements have been used extensively to study spatial variations in aerosol optical properties. As a result, the marine boundary layer aerosol is one of the best understood components of the tropospheric aerosol system. Oceanic measurements have the advantage that they can be coupled to reasonably well understood satellite aerosol retrievals. On the other hand, being confined to the ocean surface - where aerosol mass and optical properties are generally dominated by seasalt - is a serious limitation for studying the effects of anthropogenic aerosols.
Airborne aerosol measurements have also been used extensively to study spatial variations. The overwhelming majority of in-situ data on vertical variations comes from airplanes. Their main limitations are expense, payload capacity, difficulty of operating and calibrating instruments in flight, and losses of coarse-mode particles in sample inlets (see 1.4, above.)

Intensive campaigns
During intensive field campaigns, an integrated suite of observational and modeling methods are focused on one region over a several-week time period. These campaigns serve both an exploratory function - characterizing the aerosol and the factors controlling it - and as opportunities for testing the internal consistency among various measurement methods and platforms.
Ideally, the outcomes of an intensive campaign would be
(1) rigorous field-assessments of aerosol measurement methods, showing where improvements are most needed,
(2) rigorous tests of key hypotheses regarding the processes that control aerosol properties,
(3) a thorough characterization of a regional aerosol at one specific time, and
(4) an optimal, long-term measurement strategy for achieving representative sampling and monitoring changes to that region.
Results from two of the many intensive campaigns conducted during the 1990's will now be used to illustrate the type of progress that has been achieved. 18 ACE-1(6) was the first large-scale, international field campaign explicitly devoted to quantifying aerosol properties with regard to their climate forcing effects (Bates et al., 1998). It was conducted during one month in the late Spring of 1995 in the Pacific Ocean, south of Australia. A wealth of data on aerosol optical and chemical properties relevant to the direct effect was acquired from three platforms - a coastal ground station, a ship, and an airplane. Notable accomplishments include the following.

The TARFOX campaign7 investigated the direct radiative effects of North American industrial pollution aerosols as they advected off the East coast of the United States (Russell et al., 1999). Vertical profiles of aerosol chemical and optical properties were measured from an aircraft during 18 flights in July of 1996, yielding several important results.

Summary and future trends

The problem of direct forcing by anthropogenic aerosols is an excellent example (by no means unique) in which (i) a theory has been developed primarily in relation to numerical models, (ii) the models are largely empirical in nature, and (iii) the models are only weakly constrained by observational data. These features characterize research in an exploratory mode. Having identified an important problem and the key parameters needed for improved quantification, a major effort is now underway to move from this exploratory mode to a definitive one.
This transition requires a careful measurement strategy, given the extremely complex and variable nature of the atmospheric aerosol phenomenon. Direct measurements must be effectively coupled to aerosol optical models, chemical transport models, and satellite observations. This coupling must take into account the respective virtues and limitations of each approach. In-situ measurements can never sample a significant fraction of the atmosphere, but they do offer detailed and, above all, calibrated measurements of aerosol properties. Their key advantage is that uncertainties can be quantified with minimal assumptions; if this is done rigorously, they can provide a firm anchor to the reality of aerosol optics, physics, and chemistry.
Over the last decade, the aerosol research community has made substantial efforts to develop and standardize measurement techniques, to better quantify uncertainties (e.g. through closure experiments), and to measure aerosol properties from many different regions. In-situ data relevant to direct forcing have been acquired from long-term monitoring activities, geographical transects, and intensive field campaigns, with a strong emphasis on the last. The resulting large data sets are rapidly reaching the point where they can be used, in conjunction with models, to characterize important features of the aerosol on a global scale.
Weaknesses do exist, however, that undermine the ability of in-situ measurements to constrain optical properties for models and satellite retrievals. Uncertainties in the measurement of aerosol light absorption are high, even for the case of low relative humidity, and there is no available method for measuring the humidity dependence of light absorption. In addition, there is no effective method at present for sampling coarse mode aerosols from high-speed, research airplanes. At the level of experimental design, there is need for more rigorous uncertainty analysis and for more active testing of hypotheses.
Looking ahead to the next decade, a major challenge will be to devise optimal ways to couple in-situ measurements to satellite observations. We are only beginning to learn how to do this. A shift is clearly underway (see Section 3) whereby satellites will map out aerosol spatial and temporal variations while targeted in-situ measurements will establish the limits and uncertainties associated with satellite data interpretation. An important step in this direction will be to assess the scales of variability for various extensive and intensive aerosol properties. For example, there is an implicit assumption in all quantitative aerosol retrievals from satellites that certain intensive aerosol properties are conserved over broad regions, even though extensive properties undergo wide variations. Existing in-situ data sets are already ripe for testing this and many related assumptions. Future in-situ measurements should make these tests ever more definitive.

 

Section 3. Satellite Instruments for Tropospheric Chemistry

David M. Winker, NASA Langley Research Center contact: d.m.winker@larc.nasa.gov

Introduction

It is possible to adequately monitor long-lived tropospheric species with a limited number of measurement stations distributed around the globe. Satellite-based measurements are necessary, however, to obtain global observations of short-lived species and species with sources that are variable in space and time. Satellite-based measurements must, of necessity, use remote sensing techniques. Current techniques are based on the passive sensing of scattered or transmitted sunlight, or emitted thermal or microwave radiation. These passive techniques have employed viewing along limb (horizontal) paths, nadir viewing, or a combination of the two to perform the required measurements. Observations from instruments such as TOMS, SBUV, and SAGE, since the late 1970's, and the UARS satellite during the 1990's have provided the scientific community with a comprehensive picture of the chemical composition of the stratosphere and mesosphere. These measurements will be continued during the next decade by the NASA EOS- CHEM and ESA ENVISAT satellites. Table 1 summarizes a selection of past, present, and future satellite instruments designed for sensing atmospheric composition.
Initially, the remote sensing of atmospheric composition by satellite sensors was limited to the stratosphere. The primary difficulty in satellite remote sensing of trace gases and aerosols in the troposphere is the limited number of techniques available. A persistent problem has been limited vertical resolution and difficulty in separating tropospheric concentration from the total column. Tropospheric chemistry will be advanced in the coming decade by satellite sensors with new capabilities for observing the troposphere: from passive sensors with improved spectral coverage; from passive sensors in geostationary and L1 orbits providing regional and continental-scale observations with high spatial and temporal resolution; and from lidars, active sensors based on lasers, which offer the potential for observations of aerosols and some trace gases with greatly improved vertical resolution, at the cost of restricted geographic coverage. Development of these new sensors has been enabled by technologies which have matured during the 1990's. These new technologies and the advances expected from them are discussed in this section. Several satellite instruments flown during the 1990's which are forerunners of the next generation of satellite sensors are used to illustrate the discussion.

Recent advances

Ozone and other trace gases
Satellite measurements of aerosols and chemical species in the troposphere are much more difficult than in the upper atmosphere and measurement capabilities from space have been much more limited. The limb viewing geometry is characterized by horizontal resolutions on the order of 400 km, which is inconsistent with the spatial variability of many tropospheric trace species. Additionally, tropospheric clouds interfere with limb viewing measurements, so observations of tropospheric composition must rely primarily on nadir- or near-nadir-viewing instruments. Passive nadir-viewing instruments have limited vertical resolution, however, and separation of the tropospheric amount from the total column concentration can be very difficult.

Table 1. Summary of instruments discussed in this section. The list is not comprehensive but gives an example of past, present, and future capabilities. Only measurements of primary significance to atmospheric chemistry are listed.
Instrument Satellite Agency Relevant Measurements References
AVHRR TIROS-N, NOAA-6 through NOAA-14, 1978 present NOAA aerosols Husar et al., 1997
TOMS Nimbus 7, 1979-1992; Meteor, 1992 1994; ADEOS, 1996 1997; Earth Probe, 1996 present NASA ozone, aerosols Herman et al., 1997
SAGE I/II Atmospheric Explorer, 1979 1981; Earth Radiation Budget Satellite, 1984 present NASA O3, NO2, and aerosols in stratosphere and upper troposphere
MAPS Space Shuttle, 1981, 1984, 1994 NASA CO Neil and Fishman, 2000
LITE Space Shuttle, Sept. 1994 NASA aerosols Winker et al 1996
GOME ERS-2, 1995 -present ESA O3, NO2, H2O, BrO, OClO, SO2, HCHO, aerosols Burrows et al. 1999
IMG ADEOS, 1996 1997 NASDA O3, N2O, H2O, CH4, CO Ogawa et al 1994
MISR Terra, 12/99 present NASA aerosols Diner et al. 1991 Martonchik, 1997
MODIS Terra, 12/99 present; Aqua, 12/2000 NASA aerosols Kaufman et al 1997 Tanre et al 1997
MOPITT Terra, 12/99 present CSA CO, CH4 ??
SCIAMACHY ENVISAT, 2000 ESA O3, NO, NO2, N2O, H2O, BrO, OClO, CO, CH4, SO2, HCHO aerosols Bovensmann et al. 1999
EPIC Triana/2001 NASA ozone, aerosols ??
OMI EOS Chem, 2002 ESA O3, SO2, NO2, aerosols ??
TES EOS Chem, 2002 NASA O3, NO, N2O, CO, CH4, HNO3, etc. ??
GIFTS commercial satellite, 2003 NASA O3, CO, H2O, etc. ??
lidar PICASSO-CENA, 2003 NASA aerosols Winker and Wielicki, 1999
GeoTRACE TDRS NASA O3, CO, aerosols ??

UV/Visible Techniques
Total column ozone has been derived since 1979 from ultraviolet measurements of scattered sunlight by the TOMS instrument. In the late 1980's, a technique was developed to derive tropospheric column of ozone as the residual of the total column measured by TOMS and the stratospheric column from co-located SAGE II measurements (Fishman et al., 1990). The drawback of applying this technique to SAGE II data is that only climatological studies can be done, as SAGE II longitudinal coverage is sparse and full latitudinal coverage is obtained only over about 40 days.
The Global Ozone Monitoring Experiment (GOME) has operated since its launch in April 1995 as part of the ERS-2 satellite payload, and flies in a 780 km sun-synchronous orbit. GOME is a nadir-viewing spectrometer designed for trace gas measurements by observing reflected sunlight in the ultraviolet and visible (Burrows, et al., 1999). In contrast to TOMS and SBUV which operate at a number of discrete wavelengths, GOME observes the entire spectrum between 240 nm and 790 nm with an improved spectral resolution which varies from 0.2 nm in the UV to 0.4 nm in the red. The improved coverage of the ozone spectrum by GOME and the improved signal to noise ratio of the measurements provide GOME with enhanced capabilities relative to TOMS and SBUV. In the troposphere and lower stratosphere, GOME is able to resolve the vertical profile of ozone with an effective resolution of 10 km. For other gases, the tropospheric column is determined by subtracting the stratospheric column from the total column, assuming local longitudinal homogeneity of the stratosphere. Using this residual technique, GOME has been used to observe tropospheric NO2 (Figure 1), formaldehyde (HCHO), and SO2 resulting from pollution and biomass burning. GOME is the first satellite instrument suitable for studying halogen oxides in the lower troposphere and has been used to observe the global distribution of BrO.



Figure 1. Global tropospheric column amount (molecules/cm2 ) of NO2 for the month of September 1997 derived from GOME data. Elevated levels of NO2 can be seen over the industrial areas of Europe, the United States, China, and the Middle East oil fields. The production of NO2 from biomass burning in the southern hemisphere and its long range transport is also evident. Source: J. P. Burrows, personal communication, and Burrows (1999).

The observations begun by GOME will be continued by three GOME-2 instruments scheduled to be flown on the EUMETSAT/ESA METOP series of satellites, as well as by the conceptually similar OMI instrument, scheduled to be flown on EOS-Chem. The capabilities of GOME will be extended by SCIAMACHY, planned for launch on ESA-ENVISAT in early 2000. SCIAMACHY is similar to GOME in design concept but with spectral coverage extended into the near infrared, between 240 nm and 2380 nm, and with three viewing geometries: nadir viewing, limb-scanning, and occultation (Bovensmann, et al., 1999). Extended spectral coverage and matched limb/nadir measurements allow the tropospheric column amounts of a variety of trace gases to be determined: O3, NO2, N2O, BrO, CO, CH4, H2O, and, under polluted conditions: HCHO and SO2. Additionally, SCIAMACHY will measure vertically resolved profiles of O3, H2O, N2O, and CH4.

Infrared Techniques

Many trace species exhibit spectral signatures in the thermal infrared, and infrared techniques hold the potential to observe many species which cannot be observed using UV/visible techniques. Satellite measurements of tropospheric gases have been performed using the gas filter correlation technique and by inversion of the radiance spectrum observed at high spectral resolution. In the gas filter correlation technique, an onboard gas cell filled with the target gas provides measurements of a specific gas without requiring high spectral resolution and tight spectral requirements. This technique does not require the absorption lines to be resolved and is characterized by high SNR. The first satellite measurements of CO were obtained by the MAPS (Measurement of Air Pollution from Satellites) instrument using this technique in the mid-IR. MAPS flew on the space shuttle in 1981, 1984, and 1994 and provided the scientific community with valuable insights into the importance of biomass burning as a source of CO (Neil and Fishman, 2000). The utility of MAPS was limited by the short duration of shuttle missions, however. Global observations of CO on a continuing basis are planned to be provided by MOPITT on the NASA Terra platform in low Earth orbit, launched in December 1999.
High spectral resolution infrared spectrometers hold the potential to observe a broader range of species than is feasible using the gas filter correlation technique. Limited vertical resolution of CO, CH4, and O3 is possible if spectral resolution is high enough to resolve the absorption line structure. The TES (Tropospheric Emission Spectrometer) instrument, part of the EOS-Chem payload, is a Fourier Transform Spectrometer observing the spectral region between 3.2 microns and 15.4 microns at a resolution of 0.025 cm-1 . TES is expected to provide tropospheric measurements of O3, CO, CH4, HNO3, NO, and N2O with vertical resolutions of 2 km to 6 km. The IMG (Interferometric Monitor for Greenhouse Gases) instrument, which operated between October 1996 and June 1997 on the Japanese ADEOS satellite, illustrates the potential of high spectral resolution observations from instruments such as TES. IMG is a nadir-viewing Fourier Transform Spectrometer (FTS) operating between 3.3 microns and 14.0 microns with 0.1 cm-1 spectral resolution (Ogawa, et al., 1994). The instrument has an 8x8 km2 footprint and 32 km swath. As shown in Figure 2, IMG data has provided global measurements of tropospheric CO, using a neural net retrieval approach (Clerbaux et al. 1999; Hadji-Lazaro et al. 1999).

Future Trends

Observations from Geostationary and Lagrange Orbits
Among the issues in tropospheric chemistry which cannot be addressed by instruments in low Earth orbit are those processes which take place on small scales and require measurements with spatial resolutions on the order of 10 km and temporal resolutions on the order of fractions of an hour to a few hours. Satellites in low earth orbit require a full day or more to provide full global coverage and any given region is revisited no more than twice a day. Instruments on satellites in geostationary or Lagrange orbits are required in order to document the spatial distribution, temporal trends, and variability of a number of environmentally significant shorter-lived atmospheric species. A satellite in geosynchronous orbit - at an altitude of about 36000 km - orbits the Earth once per day and so has a stable and continuous view of a large portion of the Earth's surface. Four to five satellites in geosynchronous orbit, spaced longitudinally, are required to obtain full coverage of the Earth up to a latitude of 50-60 degrees. A Lagrange orbit is an orbit at a point between the Sun and the Earth where the gravitational pulls from both bodies are balanced such that the satellite maintains a constant distance from both the Sun and the Earth. The L1 and L2 points lie along a line between the Earth and the Sun. The L1 point is between the Earth and the Sun, at a distance of about 1.6 million km from the Earth, and the L2 point is beyond the Earth's orbit. The L1 point provides a continuous view of the full solar-illuminated Earth disk, and the L2 point provides a view of the night side of the Earth.
The use of sensors in geosynchronous and L1 orbits to observe 5 tropospheric composition has been enabled by the development of large format detector arrays, and their associated low-noise readout electronics, high bandwidth satellite communications links, and high volume data handling facilities. These large arrays enable sensors to make quasi-continuous measurements over the full Earth disk with high spatial, spectral, and temporal resolution. These new satellite platforms will allow the investigation of a whole new class of atmospheric phenomena, including processes occurring at small spatial scales and the transport of pollutants from local source regions to the global atmosphere. Instruments in these orbits can also give the synoptic context for measurements from ground networks or field campaigns.
GIFTS is a mission recently selected for NASA's New Millenium program for launch in 2003 to demonstrate new technologies and the potential for high spectral resolution infrared observations from geosynchronous orbit. GIFTS is an imaging Fourier Transform Spectrometer which will provide spectra in two bands in the thermal infrared. The spectral resolution is programmable, from 18 cm-1 to 0.6 cm-1 . The primary focus of GIFTS is to demonstrate an improved capability for temperature and humidity sounding to improve the accuracy of weather forecasts. However, observations at the highest spectral resolution will provide vertically resolved profiles of ozone, CO, and other constituents. Views of mesoscale-sized regions can be obtained every half hour. Triana is a NASA mission scheduled to launch in early 2001 which will fly several instruments at the L1 point to provide continuous observations of the sun-illuminated side of the Earth from this unique vantage point. A 10- band imaging filter radiometer called EPIC will provide hourly measurements of ozone over the full sunlit disk of the Earth at a spatial resolution of 8-14 km. This is done using the UV-backscatter technique developed for TOMS.
GeoTrace is an instrument which is somewhat similar to the Triana EPIC camera, but with higher spectral resolution. GeoTrace has been proposed to the NASA New Millenium Program to fly in a geosynchronous orbit and would provide measurements similar to GOME but with continuous full-disk coverage. GeoTrace gives atmospheric column amounts, is aimed at monitoring, distribution, diurnal changes, and transport. 6 Instruments such as GIFTS, GeoTrace, and Triana will provide unique insights into the temporal behavior of atmospheric composition and dynamics at small to medium scales. Rapid changes in tropospheric constituents will be observed, providing a means of separating the tropospheric burden from the relatively slowly changing stratospheric burden.

Active Sensing of Aerosols and Trace Gases
The global distribution of tropospheric aerosol is highly variable due to the variety and variability of sources and the short residence time of aerosols in the atmosphere. Due to the current limited knowledge of aerosol source strengths and production and transformation mechanisms, aerosol distribution cannot be adequately determined by modeling. Improved satellite measurements are necessary to obtain a more complete picture of the global aerosol.
Aerosols can be detected and characterized from satellite measurements of scattered sunlight in the UV and near-IR, but the retrievable information is limited. The Advanced High Resolution Radiometer (AVHRR), flying on a series of NOAA satellites, has provided aerosol observations since 1979 (Husar et al, 1997). Algorithms using two of the AVHRR wavelengths (0.65, 0.865 um) have been used to retrieve aerosol optical depth and Angstrom coefficient (the spectral coefficient of aerosol scattering). Operational aerosol products, however, have been limited to ocean regions due to the difficulty of retrieving aerosol over inhomogeneous and/or bright land surfaces. More recently (Herman, et al., 1997) aerosol retrieval algorithms have been developed and applied to TOMS UV measurements to detect absorbing aerosols, primarily dust and smoke, over land as well as marine surfaces. This recent work has resulted in new insights into the global sources, production mechanisms, and transport of dust and smoke, but retrieval of quantitative parameters is difficult.
The payload of the Terra satellite, launched in December 1999, includes the MODIS and MISR instruments which provide new capabilities for aerosol sensing. MODIS is a nadir-viewing imaging filter radiometer with 36 spectral channels and improved calibration relative to AVHRR. Seven of those channels, between 550 nm and 2100 nm, will be used for retrievals of aerosol optical depth over dark vegetation as well over ocean (Kaufman et al. 1997; Chu et al. 1998). Over the ocean, it will also be possible to retrieve the mean aerosol size and the relative contribution to the aerosol optical depth of the accumulation and coarse modes (Tanré et al. 1997; Tanré et al. 1999). The Multi-angle Imaging Spectro-Radiometer (MISR) will provide multi-angle observations in 4 spectral bands between 446 nm and 866 nm. Each band is imaged with 9 view angles spread out along the flight path between 70 degrees forward and 70 degrees aft. These multi-angle measurements provide additional information on aerosol angular scattering characteristics and allow the classification of aerosols into one of about a dozen types, distinguished by size, composition, and shape (spherical/non-spherical) as well as measurements of aerosol optical depth (Kahn et al., 1997).
The new observing capabilities provided by the Terra instruments will provide major improvements in knowledge of the global aerosol. None of these instruments, however, is sensitive to aerosols at very low concentrations or provides any information on the vertical profile of aerosols. Active lidar sensors, based on short-pulse lasers, are required.
The Lidar In-space Technology Experiment (LITE) provided the first demonstration of the promise of satellite lidar for aerosol studies. LITE was a lidar system using a high power pulsed laser operating at 355 nm, 532 nm, and 1064 nm (Winker et al. 1996) which flew on the Space Shuttle in September 1994. Figure 3 illustrates the capability of satellite lidar to detect and profile aerosols over land as well as over ocean. Observation of the vertical profile of aerosol also facilitates assimilation of the data by models.



Figure 3. Vertical cross-section of aerosol distribution observed by LITE. A deep haze layer (yellow and red) is seen over the eastern US and extending into the Atlantic Ocean. Yellow lines trace wind back-trajectories at the 850 mb level over the 5 days previous to the LITE overpass.

The relatively recent development of rugged, long-life, and efficient diode-pumped solid-state lasers has enabled the development of lidars which can be operated from long-term satellite platforms in low Earth orbit. PICASSO-CENA is a recently approved NASA mission which will launch in early 2003 and is planned for a three-year mission (Winker and Wielicki, 1999). The PICASSO-CENA instrument suite includes two instruments optimized for aerosol studies: a two-wavelength lidar (532 nm, 1064 nm) and a high-resolution spectrometer viewing the A-band of oxygen. The spectrometer will demonstrate a new approach to the passive sensing of aerosols. Combination of channels with large and small column oxygen optical depth allows the separation of atmospheric scattering from surface scattering, providing improved retrievals of aerosol over land surfaces.
To provide a comprehensive picture of the global aerosol distribution, data from many diverse sources - both satellite and ground-based in situ and remote sensing instruments - will have to be integrated, using model assimilation techniques. The PICASSO-CENA satellite will fly in formation with the Aqua (formerly EOS PM) satellite, allowing the combination of MODIS and lidar observations. The OMI instrument, on EOS-Chem, will be used to detect aerosols using the TOMS retrieval approach (Herman et al. 1997). The unknown vertical distribution of aerosol results in significant uncertainties in quantitative retrievals using this algorithm. Use of near- coincident aerosol profiles from PICASSO-CENA can be used to improve the accuracy of OMI aerosol retrievals.
It should also be mentioned that technologies required for space-based sensing of trace gases by differential absorption lidar (DIAL) systems are being developed. DIAL systems have not yet been flown in space, but have proven their utility through many years of use in ground-based and airborne measurement campaigns. Laser technology has now progressed to the point where spaceborne DIAL systems are in the early planning stages and it is likely that DIAL systems to measure tropospheric ozone and water vapor will be in orbit by the latter part of the decade. Because active sensors provide limited horizontal coverage, assimilation techniques will need to be developed to merge high vertical resolution data from lidar with broad spatial coverage data from passive sensors.

Summary

Global remote sensing of atmospheric constituents is essential to understand the natural processes which determine the composition of the global atmosphere and to assess the impact of human activities on the atmosphere. Following successes in the development of capabilities for satellite measurements of the stratosphere, the next challenge is to develop capabilities for measurements in the troposphere. In contrast to the stratosphere, higher spatial and temporal resolution is required for tropospheric studies. New developments in satellite sensor technologies during the last decade have enabled more capable sensors to meet these challenges. Pathfinder instruments such as GOME, IMG, and LITE have indicated the potential for the new generation of instruments to be flown during the next decade. These new satellite sensors will provide major advances in our understanding of the composition of the troposphere on regional and global scales.

Acronyms:

EPIC - Earth Polychromatic Imaging Camera
GIFTS - Geostationary Imaging Fourier Transform Specrometer
OMI - Ozone Monitoring Instrument
SCIAMACHY - SCanning Imaging Absorption spectroMeter for Atmospheric CartograpHY

 

Section 4. In situ Gas Phase Measurements

Fred L. Eisele, NCAR/ACD contact: eisele@acd.ucar.edu

Recent Advances

Major advances in instrumentation over the past decade, particularly in a few specific areas, have greatly enhanced understanding of gas phase chemical processing and are also providing a whole new view of the production, growth, and interaction of particles in the atmosphere. The rate at which the earth's atmosphere cleanses itself is largely dependent on its ability to produce the hydroxyl radical (OH), yet only a decade ago very few if any credible in situ tropospheric measurements of this compound existed [Crosley, 1993]. By 1991, a sensitive new in situ ground-based OH instrument was successfully intercompared to a long path instrument [Mount and Eisele, 1992; Eisele et al., 1994], and several other successful intercomparisons took place over the next few years [Crosley, 1997; Hofzumahaus et al., 1998]. By the mid 1990's OH measurements were being included on tropospheric aircraft campaigns [Mauldin et al., 1998; Brune et al., 1998]. While still not common place, OH and HO2 measurements have contributed significantly to several major aircraft and ground-based campaigns. Large increases in the sensitivity of long-path absorption instruments have led to new folded-path instruments, which can now also make ground or ship-based in situ measurements [Brandenburger et al., 1998].
Another area of major progress is the measurement of biogenic compounds. Over the past decade or so, the previously held view of vegetation as an emission source of a few fairly passive organic compounds has changed to one in which vegetation plays a central interactive role in ozone production and atmospheric chemistry in general. Its role, however, is extremely complex, and the number of measurements required to fully quantify biogenic emissions and their products is truly huge. Unlike the difficulties encountered in measuring the hydroxyl radical, the compounds of importance are neither in the sub-parts per trillion concentration range, nor so reactive that they are lost on first contact with a surface. Rather, the problem is specificity because of the sheer number of compounds to be measured and the fact that many of the compounds are so similar. In addition, many of the product compounds of interest are quite sticky and can react if deposited on surfaces. Not only are chemical identification and the measurement of concentrations important; flux of the compounds is also of great interest. Much progress has been made using customized GC and GCMS measurement schemes combined with eddy accumulation techniques to inventory both primary biogenic emissions and their products [Guenther et al., 2000]. While such techniques are highly specific they are typically quite slow, requiring 102 -103 seconds. They are thus not well suited to flux or aircraft studies unless accompanied by some type of eddy accumulator. Advances in diode lasers technology have led to several new measurement techniques for making faster measurements of some biogenic and anthropogenic emissions such as: formaldehyde, ammonia, CO, NO2, H2O2 , N2O and SO2 [Silver et al., 1991; Roths et al., 1996; Mackay et al., 1996; Fried et al., 1999]. While these measurements are highly specific and typically have sensitivities allowing measurements into the mid ppt range, it is unlikely that they will be able to sort out and individually identify the bulk of major biogenic emissions and their products (particularly complex compounds with more than 5 atoms) over the next decade. Recently, a modified chemical ionization mass spectrometry technique which operates at low pressures and uses electric fields to enhance collision energies in the ion reactive region was introduced into the atmospheric measurements arena. The instrument is called a Proton Transfer Reaction Mass Spectrometer (PTRMS) [Lindinger et al., 1998a,b] It offers several advantages over existing hydrocarbon measurement techniques: fast measurements (in the 1 second time frame), a reasonable amount of specificity using mass analysis to identify compounds, measurements down to the mid to upper ppt concentration range, and the ability to be used on an aircraft [Crutzen et al., 2000]. To measure the large number of organic compounds present in the atmosphere in an unambiguous way may require more analytical power than is provided by the present PTRMS, but this can be greatly improved through the use of tandem mass spectrometers or ion traps.
Another area in which large strides have been made is sulfur chemistry and its relation to particles. For many years, particle nucleation was thought to occur even in the remote atmosphere but direct quantitative evidence for such processes was lacking. In the past decade, new developments in both the gas and particle side of nucleation have dramatically enhanced observations and understanding. On the particle side the development of ultrafine particle detectors now allows particles as small as 2.5 nm diameter to be measured and size-resolved, even at concentrations below 0.1 particles cm-3 [Stolzenburg and McMurry, 1991]. Gas phase precursors such as sulfuric acid [Eisele and Tanner, 1993] are also now commonly measured in conjunction with ultrafine particles, and this combination for the first time provides a means of directly comparing nucleation theory to field observations [Weber et al., 1995]. Even the transitional growth steps between gas phase molecular clusters and stable particle formation are now observable [Eisele and Hanson, 2000].
Exciting new breakthroughs have also made possible real time chemical composition measurements of particles. A variety of laser evaporation/laser induced ionization time-of-flight mass spectrometry instruments are now being used to chemically analyze individual particles in laboratories and in field studies. The result has been a surprising diversity between individual particles and a complex chemical mixture of many trace constituents, including many metals. These instruments still have limited capabilities for determining absolute chemical concentrations, but rapid progress is being made in this direction.
While lidars are not new, their improved capabilities and extended use are. They are presently included in many aircraft payloads and provide valuable information about the larger environment in which the aircraft in situ measurements are made. In addition, they provide an extended real time survey of local conditions, and as such can be used to direct aircraft flight planning during an ongoing mission.

Future Trends

The better we come to understand atmospheric chemistry, the more complex we realize it is. This complexity takes on several forms, each of which poses a different set of measurement problems. Part of this complexity occurs because observed chemical concentrations are typically a result of dynamics as much as chemistry. In order to understand the influences of each, more chemically speciated lidar and satellite measurements are needed. The development of more rapid measurement techniques which can determine fluxes would also contribute much, as would the measurement of a wider range of compounds, many of which could be used as tracers. The use of multiple measurement platforms could also provide additional insight, especially near localized sources or convective systems. As measurements move down near the surface, particularly in industrialized or urban areas, the complexity increases in part because the sheer number and types of sources increase. If such areas are to be studied in detail, far more compounds need to be measured, and if accomplished on an aircraft, the measurement time needs to be decreased to the order of a few seconds so that individual sources or plumes can be distinguished from each other. This will require new, more sensitive measurement techniques, which will probably be mass spectrometry based. Chemical ionization mass spectrometry offers the required speed and sensitivity and can be applied to large/complex biogenic and anthropogenic molecules. More powerful analytical capabilities, such as tandem mass spectrometers and ion traps will, however, be required. Optical techniques can make fast sensitive measurements for small molecules and continue progress in this area is needed, but these techniques have difficulty, for example, in speciating large organics.
Another area that deserves far more attention in the next decade is aircraft inlets. As an ever larger number of field studies are mounted from aircraft platforms, and as measurement capabilities are extended to include more reactive and/or sticky compounds, sampling becomes a major issue. An instrument that works perfectly on the ground or even in an aircraft will add little to an aircraft mission if the compound of interest can not be transported in a well understood way between ambient outside conditions and the measurement instrument. This problem is particularly acute for aerosols, but affects many gas phase measurements as well. Some of the issues that need to be addressed are: Large air speed changes and the associated temperature changes, even if measurements are made outside the aircraft, temperature changes associated with bringing air into the aircraft, re-equilibration of compounds found both on particles and in the gas phase as temperature and relative humidity change, loss to and evaporation from inlet surfaces as altitude, temperature, and relative humidity change, the influence of pressure changes, aircraft boundary layer thickness, inlet orientation and position relative to other inlets, high speed droplets, droplet shattering, and icing. Finally the influence of heterogeneous reactions on tropospheric gas phase chemistry needs to be explored in far more detail particularly in urban air masses. This will require new more sophisticated measurement techniques probably involving both lasers and mass spectrometers to quantify not only acids and ammonia but also the large number of organic compounds found in both the gas and particle phases.

 

References

Laboratory METHODS

In situ Aerosol Measurements

Satellite Instruments for Tropospheric Chemistry

In situ Gas Phase Measurements

1 Measurements where the instrument or sampling device is collocated with the aerosol; distinct from "remote measurements."
2 A "climate forcing" is a change in global or regional net irradiance (W m-2 ) at the top of the atmosphere (or some other height), imposed by some agent (such as human activity) that is external to the climate system.
3 "Direct" forcing by aerosols refers to the scattering or absorption of solar radiation by the particles themselves and is distinguished from "indirect" forcing, which involves particle-induced changes in cloud brightness or extent. 4 An extensive parameter depends on the amount of aerosol material present while an intensive parameter (frequently the ratio of two extensives) does not. 5 Consider, for example, the shadow cast by a grove of trees. The question, "What fraction of the shadow is due to this individual tree?" is ultimately ambiguous. Other examples, specific to extinction apportionment are discussed by White (1986). 6 The International Global Atmospheric Chemistry (IGAC) Project's first Aerosol Characterization Experiment (ACE-1). 7 Tropospheric Aerosol Radiative Forcing Observational Experiment.
Last modified: Tue May 2 08:25:17 CEST 2000