Studies of Aerosols in The Marine Boundary Layer in the Coastal Area During The EOPACE’99 Campaign

STUDIES OF AEROSOLS IN THE MARINE BOUNDARY LAYER IN

THE COASTAL AREA DURING THE EOPACE’99 CAMPAIGN

T. ZIELINSKI* and J. PISKOZUBInstitute of Oceanology, Polish Academy of Sciences U1, Powstancow Warszawy 55, 81-712

Sopot, Poland

(Received in final form 08 August 2004)

Abstract. Results are presented based on measurements taken using an FLS-12 lidar systemand laser particle counters only on the Atlantic coast of the U.S.A. during a campaign withinthe scope of the international EOPACE experiment. The objectives of the EOPACE

(Electro-optical Propagation Assessment in Coastal Environments) effort, which wasconducted in Duck, N.C. (U.S.A.) between 25 February and 11 March 1999, involvedinvestigating, developing and evaluating ocean and coastal aerosol models and their effects on

visibility; integrating and developing simple, realistic models for infrared propagation near theocean surface and developing a consistent chemical/optical model for aerosol particles suitablefor inclusion in navy meteorological models.

Keywords: Aerosol, Coastal area, EOPACE, Lidar.

1. Introduction

In the coastal area the marine aerosol size distribution is complicated anddepends heavily on weather conditions in the marine boundary layer, espe-cially on wind speed, wind duration and direction as well as on relativehumidity.

Atmospheric models need a reliable and easily obtainable indication of airmass characteristics in order to determine aerosol size distributions, con-centrations and extinction in coastal areas, where a wide variety of rapidlychanging atmospheric conditions have an impact on the aerosol ensemble.

The analyses of the EOPACE/Duck’99 field campaign focused on threeareas of marine aerosol measurements and modelling: the flux of sea-saltparticles from both the surf-line and whitecaps; the transition of the conti-nental boundary layer into a marine boundary layer during periods of off-shore and along-shore flow; the variability of salt particle size distributionsunder varying wind speeds (Jensen et al., 2001).

* E-mail: [email protected]

Boundary-Layer Meteorology (2005) 116: 533–541 � Springer 2005DOI 10.1007/s10546-005-0904-6

2. Experimental Site and Instrumentation

The data presented in this paper were collected during an internationalexperiment known as EOPACE conducted in Duck, N.C. (U.S.A.). All of theaerosol concentration measurements were taken with an FLS-12 lidar systemand laser particle counters (CSASP-200 and CSASP-100-HV). The aerosoloptical thickness derived from the FLS-12 lidar measurements was comparedwith the values obtained from measurements with a CIMEL CE 318 auto-matic sun tracking photometer.

The FLS-12 lidar and the CIMEL sunphotometer were located at the endof a measurement pier, still within the range of the surf zone and pointedoffshore, while several laser particle counters were placed along the pier atdifferent altitudes.

The marine boundary layer was sounded using a reflecting mirror. Alti-tudes of up to 100 m a.s.l. (above sea level) were sounded and the horizontalrange of lidar soundings was up to 400 m. The FLS-12 lidar system used forthe measurements is a tunable laser system designed for remote sensing of theatmosphere in the visible (VIS) range of the electromagnetic spectrum (400–670 nm). A detailed description of the principal parameters of the FLS-12lidar was given in Zielinski and Zielinski (2002).

During the campaign, the measurements were carried out using wave-lengths k = 443 and 566 nm. During the experimental period (1–10 March1999) the temperature, relative humidity and wind speed varied over rangesT 2 [)0.2, 19 �C], RH 2 [44, 99%], and v 2 [2, 15 m s)1], respectively.

3. Methodology

The aerosol concentration at an arbitrary altitude above sea level may bedetermined using the Potter (1987) procedure and the Mie algorithm, if apredetermined function is assumed. This procedure allows for the determi-nation of extinction at an arbitrary point zi located on the sounding pathwith lidar radiation of wavelengths k1 and k2.

The Potter algorithm was used since it is the default in the FLS-12 lidarsoftware. The optical depth within the distance range the authors used in theresearch is so small, that, in fact, extinction could be disregarded and mul-tiplication by square distance would be a sufficient algorithm. Therefore, anydeviations from the Potter algorithm do not significantly influence the results.

Non-linear minimization was used to determine the parameters thatdetermine the size distribution and the total aerosol concentration at par-ticular altitudes hi for particles in the radius range of r 2 [r1, r2]. The valuesof radii r1 and r2 were derived theoretically using the Mie code and with theassumption that the aerosol ensemble consists of spherical particles with

T. ZIELINSKI AND J. PISKOZUB534

known optical properties (refraction index). It was assumed that the mixtureof marine and continental particles of unknown physical, optical andchemical properties, as well as unknown advection sources in the marineboundary layer over the coastal area in Duck, N.C., had a light refractiveindex in the VIS: m = 1.5 – 0.005i (Ivlev and Popov, 1973; Sokolik et al.,1993; Kim et al., 1995) with wind directions of 150� through south to 300�.

When the wind direction was from 330� through north to 120� the aerosolswere regarded as marine aerosols, both generated in the surf zone andtransported from the open Atlantic Ocean. Thus, the light absorption coef-ficient in the VIS range was assumed to be zero and the light refractive indexwas chosen for sea water with a salinity of S = 36 psu and a temperature ofT = l �C. This gave m = 1.351312 for k=404.7 nm and m = 1.339606 fork=643.8 nm (Popov et al., 1979). The value of this index was derived the-oretically when calculating distribution parameters a and b in the lidarequation; this facilitated the best agreement of the aerosol concentration Nc

derived from lidar measurements and the data obtained from the laser par-ticle counter.

The values of aerosol concentrations were derived from the backscatteredlidar signal and have errors of ±30%, comparable with the error whenaerosol concentrations are measured with the laser particle counter used tocalibrate the results obtained with the lidar (Zielinski et al., 1998). The de-tailed description of the methodology can be found Zielinski and Zielinski(2002).

4. Results and Discussion

The values of the aerosol concentrations obtained from the FLS-12 lidar atan altitude of 15 m a.s.l. as a function of wind speed and direction are pre-sented in Figure 1. The aerosol concentration increases independently ofwind direction as wind speed increases. When marine aerosols were generatedby onshore flow, with wind speeds of v > 10 m s)1, these concentrations areapproximately one third of those with offshore winds.

Figure 1 shows that the range of measurement values obtained during theEOPACE experiment is much greater than the measurement error of themethod applied. This is caused by the highly non-homogeneous density in themarine boundary layer resulting from significant and rapid changes inmeteorological conditions that occurred daily (Jensen et al., 2001).

The aerosol size distribution function for offshore wind speedsv = 9.4 m s)1 and onshore wind speeds v = 7.3 m s)1 are presented inFigure 2a. For the purpose of comparison, the experimentally derived dataobtained from the laser particle counter were included. The figure shows thatfor onshore winds the function variations are similar throughout the entire

( ) p

STUDIES OF AEROSOLS 535

range of sizes r 2 [0.5, 5 lm] regardless of the method used. For offshorewinds the similarity is clear for particle sizes r 2 [0.5, 2 lm]. The increase ofparticle sizes causes differences in function values, and for particle sizesr = 5 lm the values of the size distribution function derived from the lidarare one order of magnitude lower than those obtained from the laser particlecounter. Such significant differences for continental aerosol particle sizesr > 2.5 lm are caused by large particle fallout in the marine boundary layer

( )

Figure 1. Variations of aerosol concentration with wind speed for two directions, onshoreand offshore, at Duck, N.C., U.S.A.

Figure 2. Comparison of size distribution functions derived from the lidar method andfrom the laser particle counter for (a) winds from different directions but of different speedsand (b) winds from different directions and the same speeds.

T. ZIELINSKI AND J. PISKOZUB536

between the measurement positions of the lidar and the laser particle coun-ters (separated by 640 m), as well as the difference in the measurement alti-tude (lidar – 15 m a.s.l. and particle counters – 4 m a.s.l.).

The results obtained for offshore and onshore flow, v = 5 m s)1, whichare presented in Figure 2b, justify the application of the aerosol size distri-bution function. In this case even better agreement was obtained for resultsfrom the lidar and the laser particle counters for both wind directions in thesize range of r 2 [0.5, 5 lm]. The differences between the functions increasefor particle sizes r > 4.5 lm and this increase is greater for continentalaerosols. This can be explained for the same reasons as in Figure 2a. Themixing of the cold air masses with the existing warm marine air massesaffected the refractive index of the aerosol particles for the aerosol ensembleinvestigated on 5 March (Figure 2b – offshore), which, in accordance withthe Arctic Monitoring and Assessment Programme (AMAP) recommenda-tions, could not be regarded as either purely marine or purely continental.

Figure 3 presents the results of the measurements (lidar) and calculationsof the size distribution function using the Coastal Aerosol Transport (CAT)model (Vignati et al., 2001), which was also tested on the experimental dataobtained during the EOPACE campaign in Duck N.C. The aerosol sizedistribution function was derived for offshore wind speeds of v = 5 and9 ms)1 (Figure 3a), for onshore winds at v = 5.5 and v = 7 m s)1 (Figure 3b)and with the particle size range used in the FLS-12 lidar measurements. Theresults indicate that there is good agreement between the values obtainedusing both approaches for aerosol particle sizes r 2 [3, 5 lm]. The differ-ences between the experimental and theoretical data in onshore winds aremuch smaller for aerosol particle sizes r 2 [0.5, 2.8 lm] than with offshore

Figure 3. Comparison of the size distribution function derived from the lidar method inDuck, N.C, and the CAT Model for two different wind speeds and directions (a) offshoreand (b) onshore.

STUDIES OF AEROSOLS 537

winds (Figure 3a). The theoretical values (CAT) are lower than thosedetermined experimentally for aerosol particle sizes 1 lm < r < 3 lm.

The relatively high differences between the theoretical (CAT) and exper-imental size distribution functions for aerosol particle sizes r < l lm may berelated to the erroneous estimation of the source of sea spray particles(deLeeuw et al., 2000) or by the relatively inaccurate measurement system(significant measurement error) in this range of particle sizes.

The experimental (lidar) and theoretical (CAT) values (Figure 3a) of thesize distribution function for aerosol particle sizes r > 3 lm are similar andthey do not significantly differ from those obtained from the laser particlecounters (Figure 2a).

Figure 4 presents the values of aerosol extinction derived during theEOPACE experiment using the both the lidar and the laser particle counters.The values obtained from the laser particle counter as functions of windspeed have a similar variation to those obtained from the lidar method. Thevalues of eA obtained from both methods are very similar for both onshoreand offshore flow for speeds v < 6 m s)1 and the difference is within themeasurement error range. However, for wind speeds v > 6 m s)1 the dif-ferences are greater than the measurement error. The differences between theeA values obtained with the two methods may be reduced if the relativehumidity and the height of the measurement are taken into consideration(Fitzgerald, 1989). Since the relative humidity varied from 45 to 96% duringthe EOPACE experiment, and the measurements using the lidar and the laserparticle counter were made at altitudes of 1 m and 15 m a.s.l., the derived eA

values at the same wind speeds should vary. According to the Fitzgerald(1989) theory regarding the marine boundary layer, the extinction values

g , p

Figure 4. Comparison of aerosol extinction at 550 nm obtained from the lidar and laserparticle counter for winds of various speeds and two directions, (a) offshore and (b) on-shore.

T. ZIELINSKI AND J. PISKOZUB538

obtained at an altitude of 15 m with 90% relative humidity and at a windspeed of 10 m s)1 are about 10% higher than those obtained at an altitude of1 m with 70% relative humidity at a wind speed of 5 m s)1. However, thiscannot be the ultimate explanation for these differences; they are also causedby changes in the optical parameters of aerosols suspended above the seasurface at an altitude of 15 m produced by changes in the intensity of mixingprocesses.

Studies that were conducted in different regions reveal that air massesfrom various directions have a significant impact on this parameter (Hoppelet al., 1985; Hoyningen-Huene and Wendisch, 1994), and especially on theirchemical composition (Hegg et al., 1997; Quinn et al., 2001) and physicalproperties (Villevalde et al., 1989; Bates et al., 1998, 2001).

The aerosol extinction coefficients derived from the lidar method asfunctions of wind speed and direction facilitate deriving the aerosol opticalthickness sA. The results presented in Figure 5 are consistent with the rangeof variability as reported by Smirnov et al. (1995) and Villevalde et al. (1994).However, these results must be regarded only as temporary variations in thisparameter in air where physical and chemical mixing processes occur. Thishypothesis is confirmed by comparing data regarding sA obtained from lidarmeasurements and measurements using other methods during the EOPACEexperiment. The sA values determined for onshore flow and speeds ofv = 7 m s)1 using the lidar and the Cimel sunphotometer differ by about200%.

If, in the derivation of eA, the relative humidity and altitude are taken intoconsideration, then the differences in sA obtained from the lidar and theCimel sunphotometer could be decreased by about 50%.

Figure 5. Comparison of lidar and CIMEL obtained aerosol optical thickness derived forwavelength 440 nm in Duck, N.C. for winds of various speeds and two directions.

STUDIES OF AEROSOLS 539

5. Conclusions

The results of lidar-based aerosol investigations in the coastal area ofDuck, N.C. conducted within the scope of the EOPACE effort allow for thefollowing conclusions to be drawn:

1. Aerosol concentrations depend on the type of air mass advected fromdifferent directions and at different speeds. The aerosol concentration wasapproximately three-fold higher in offshore flow than onshore flow. It wasobserved that aerosol extinction and optical thickness in the coastal areaincrease with higher wind speeds, and that the increase is more rapid forcontinental than for marine aerosols.

2. The differences in aerosol optical thickness did not exceed 30% for windspeeds v < 6 m s)1 and 200% for wind speeds v > 10 m s)1. The valuesof aerosol extinction obtained from the lidar correspond (within the rangeof the measurement error) with those obtained using other methods.

3. It was confirmed that the lidar-based results agree with those obtained fromthe CAT model for particle sizes r > 3 lm and with the laser particlecounter throughout the range of particle sizes investigated r2 [0.5, 5 lm] forboth onshore and offshore flow for wind speeds v < 6 m s)1. Despite thedifferences between the experimental and theoretical data, theCATmodel isa very good quantitative and qualitative tool for describing aerosols in thecoastal atmospheric boundary layer, especially for onshore winds.

Acknowledgements

The authors would like to acknowledge the support of the Office of Naval Re-search under grantNo.N00014-99-1-4024.Theywould also like toacknowledgethe work of the SPAWAR, NRL,Washington and NASAGSFC teams, whoseresults have been compared with the lidar results presented in this work.

References

Bates, T. S., Huebert, B. J., Gras, J. L., Griffiths, F. B., and Durkee, P. A.: 1998, ‘International

Global Atmospheric Chemistry (IGAC) Project’s First Aerosol Characterization Experi-ment (ACE 1): Overview’, J. Geophys. Res. 103, 16,297–16,318.

De Leeuw, G., Neele, F. P., Hill, M., Smith, M. H., and Vignati, E.: 2000, ‘Production of Sea

Spray Aerosol in the Surf Zone’, J. Geophys. Res. 105, 29,397–29,409.Fitzgerald, W. J.: 1989, ‘Model of the Aerosol Extinction Profile in a Well-mixed Marine

Boundary Layer’, Applied Optics 28(15), 3,534–3,538.

Hegg, D. A., Livingston, J., Hobbs, P. V., Novakov, T., and Russell, P.: 1997, ‘ChemicalApportionment of Aerosol Column Optical Depth off the Mid-Atlantic Coast of theUnited States’, J. Geophys. Res. 102, 25,293–25,303.

T. ZIELINSKI AND J. PISKOZUB540

Hoppel, W. A., Fitzgerald, J. W. and Larson, G. E.: 1985, ‘Aerosol Size Distributions in AirMasses Advecting off the East Coast of the United States’, J. Geophys. Res. 90, 2,365–2,379.

Hoyningen-Huene, W. and Wendisch, M.: 1994, ‘Variability of Aerosol Optical Parameters by

Advective Processes’, Atmos. Environ. 28, 923–933.International Association for Meteorology and Atmospheric Physics, Radiation Commission,

Aerosol models: 1984, Boulder, Colorado U.S.A., pp. 6–28.

Ivlev, S. L. and Popova, I. S.: 1973, ‘The Complex Refractive Index of the Matter of theDisperse Phase of The atmospheric Aerosol’, Izv. Acad. ANCCR, Fizika atmos. I okeana.9(10), 1,034–1,043.

Jensen, R. D., Gathman, S. G., Zeisse, C. R., McGrath, C. P., deleeuw, G., Smith, M. H.,Fredrickson, P. A., and Davidson, K. L.: 2001, ‘Electro-Optical Propagation Assessment inCoastal Environments (EOPACE): Summary and Accomplishments’, Opt. Eng. 40(8),

1,486–1,498.Kim, Y., Sievering, H., Boatman, J., Wellman, D., and Pszenny, A.: 1995, ‘Aerosol Size

Distribution, Water Content Measurements During Atlantic Stratocumulus TransitionExperiment/Marine Aerosol and Gas Exchange’, J. Geophys. Res. 100(D 11), 23,027–

23,038.Popov, N. J., Fedorov, K. N., and Orlov, W. M.: 1979, ‘Morskaja Voda’ (in Russian),

IZd.Nauka, Moscow, pp. 327.

Potter, J.: 1987, ‘Two-frequency Lidar Inversion Technique’, Appl. Opt. 26, 1,250–1,256.Quinn, P. K., Coffman, D. J., Bates, T. S., Miller, T. L., Johnson, J. E., Voss, K., Welton,

E. J., and Neususs, C.: 2001, ‘Dominant Aerosol Chemical Components and Their

Contribution to Extinction During the Aerosol 99 Cruise Across the Atlantic’, J. Geophys.Res. 106, 20,783–20,809.

Smirnov, A., Villevalde, Y., O’Neill, T. N., Royer, A., and Tarussov, A.: 1995, ‘AerosolOptical Depth Over the Oceans: Analysis in Terms of Synoptic Air Mass Types’, J. Geo-

phys. Res. 100, 16,639–16,650.Sokolik, I., Andronova, A. and Johnson, T. C.: 1993, ‘Complex Refractive Index of Atmo-

spheric Dust Aerosol’, Atmos. Environ. 27A(18), 2,495–2,502.

Vignati, E., de Leeuw, G. and Berkowicz, R.: 2001, ‘Modeling Coastal Aerosol Transport andEffects of Surf-produced Aerosols on Processes in the Marine Atmospheric BoundaryLayer’, J. Geophys. Res. 106, 20,225–20,238.

Villevalde, Y. N., Yakovlev, V. V. and Smyshlaev, P. S.: 1989, ‘Measurements of Atmo-spheric Optical Parameters in the Baltic Sea and the Atlantic Ocean’, in Y. Y. Kulikov(ed.), Studies of the southern part of the Norwegian Sea,Gidrometeoizdat, Moscov, pp. 105–

110.Villevalde, Y. V., Smirnov, A. V., O’Neill, T. N., Yakovlev, V. V., and Smyshlayev, S. P.:

1994, ‘Measurement of Aerosol Optical Depth in the Pacific Ocean and the NorthAtlantic’, J. Geophys. Res. 99, 20,983–20,988.

Zielinski, A., Zielinski, T. and Piskozub, J.: 1998, ‘Dependence of Aerosol Size Distributionon Different Sea Bottom Types’, in RTO Meeting Proc. 1, E-0 Propagation, Signature andSystem Performance Under Adverse Meteorological Conditions Considering Out-of-Area

Operations. 38(1), 38–5.Zielinski, T. and Zielinski, A.: 2002, ‘Aerosol Extinction and Optical Thickness in the

Atmosphere Over the Baltic Sea Determined with Lidar’, J. Aerosol Sci. 33(6), 47–61.

STUDIES OF AEROSOLS 541