Transactions on Ecology and the Environment vol 4, © 1994 WIT Press, www.witpress.com, ISSN 1743-3541
Parameterization of the dry deposition velocity of submicronic aerosol particles E. Lamaud, J. Font an, A. Lopez, A. Druilhet Laboratoire d'Aerologie, U.A. associee an CNRS n° 358, Universite Paul Sabatier, Toulouse, France Abstract Dry deposition velocity of submicronic aerosol particles has been measured during 3 campaings corresponding to different ground cover. The eddy correlation method cannot be computed blindly, because of low frequency fluctuations. After a properfiltrationthe dry deposition velocity can be obtained and parametrized. During unstable situations, dry deposition velocity exhibit high values. 1 Introduction Aerosol particles are removed from the atmosphere by dry and wet deposition. The knowledge of these phenomena is necessary to calculate the aerosol residence time in the troposphere, to modelize the transport at different scales, to determine the possible impact of aggressive aerosol particles on the biosphere, etc. We have measured over different types of soils or vegetations, the dry deposition of submicronic aerosol and we propose a parametrization of the dry deposition velocity, not very different of that given by Wesely et al. (6) for sulphate aerosol. 2 Method of measurements Aerosol is charged by corona effect and the electric charges carried by particles are collected and measured with an electrometer, e.g.El Bakkali (1). To avoid statistical fluctuations a high flow sampling rate (3 1/s) is necessary. The time constant of the sensor is about 0,2 s. For a log-normal particle size distribution (geometrical mean 0.044 urn, standard deviation « 2,5) representative of the experimental distributions obtained at our sites, the analyser has a response
Transactions on Ecology and the Environment vol 4, © 1994 WIT Press, www.witpress.com, ISSN 1743-3541
434
Pollution Control and Monitoring
centered at 0,15 pm, with the same standard deviation. 90% of the detected particles have a size between 0,05 and 1 um Vertical fluxes over the ground of aerosol, ozone, momentum, sensible heat are measured by the eddy correlation method, e.g.Lamaud et al, (4) Wind is obtained with a 3D sonic anemometer. 3 Experimental sites The measurements have been made during 3 campaigns on different sites In the Landes forest located in the South West of France; the vegetation canopy consists of maritime pine (Pinus pinaster Ait) with a short (less than 1 m) gramineous under storey on a sandy soil. The canopy high is 15 m and the leaf area index of the trees is about 3. The displacement distance zj = 11 m, the roughness length ZQ = 1.4. It is equipped with a 25 m high tower. Fluxes were measured at the top, e.g. Lamaud et al. (4) The second site is the Ebre valley near Zaragoza in Spain and the campaign was called Stars II experiment. The ground is a bare soil in the north-west direction with a roughness length ZQ = 0,25 cm, a bare soil also in the East direction with ZQ ^ 1 cm and an almond tree plantation in the south-west with ZQ = 10 cm. The altitude measurement is z = 7 m. The third place is in Sahel near Tilabery in Niger (Africa). The campaign is call Stars I. The ground consists of a bare clay soil with ZQ = 0,20 cm. 4 About the methology of thefluxmeasurements The different fluxes are measured by the eddy correlation method Fluctuation of low frequencies are much more important for aerosol than for other quantities like the vertical wind velocity, the temperature T, or also ozone measured at the same time and the same altitude. Figure 1 gives the energy spectra for W, T, Og, and aerosol (C) and we can observe the high contribution of the lower frequency range in the case of aerosol. These fluctuations are not produced by the dry deposition, but they give however a contribution to the vertical flux. Figure 2 gives mean cospectra WC obtained during the Stars II experiment. These cospectra are represented with the classical adimensional frequency f = nz./u (u : horizontal wind speed) for different values of the Monin-Obukhov parameter z/L in unstable situations. The low frequencies are abundant (less than 10'2) jn f^e case of aerosol but in the range 10~2 - 1, cospectra can be considered as universal with only a very weak shift of the maxima, in function of the instability To obtain the vertical flux produced by the interaction with ground and vegetation, it is necessary to eliminate this low frequency contribution. After filtration we can compare the variation of the normalized standard deviation for aerosol and temperature in function of the Monin-Obukhov parameter The variations are quite the same, as indicated by the figure 3 In the same way, the correlation coefficients between vertical wind component and aerosol concentration, vertical wind and temperature have equivalent variations. During
Transactions on Ecology and the Environment vol 4, © 1994 WIT Press, www.witpress.com, ISSN 1743-3541
Pollution Control and Monitoring
435
instability RwT is between + 0.05 and + 0.08 and RwC between - 0.04 and - 0.6, (figure 4 ). During stability, z/L > 0, RwT « RwC « 0.3. These results are in good agreement with the similarity theory of the surface layer, but only after filtration of the low frequencies, in the case of aerosol. The reasons for the presence of low frequencies are not quite clear These non stationary events are important and more difficult to eliminate in the case of large instability, with low wind speed For, the range of high frequencies, linked to the interaction with ground, shifts toward the low frequencies It becomes obvious that turbulent fluxes must not be computed "blindly" The simple observation of time changes in mean concentration is not sufficient for detecting detrimental low-frequency trends and other methods like flux-gradient or eddy accumulation cannot be applied to the measurement of the dry deposition velocity of aerosol particles. We have observed this phenomena also in the case of ozone, and if we generalize for a constituent X, to apply methods like gradient or eddy accumulation, it is necessary to demonstrate that the low frequency contribution to the correlation coefficient WX is weak. 5 Parametrization of the dry deposition velocity Wesely et ai, (5, 6), have proposed in the case of sulphate aerosol a parametrization of the dry deposition velocity Vj For stable or neutral situations, L > 0 Yd / u* = 0.002 In the case of unstable cases, L < 0 Vd / u* = 0.002 { I + ( - 0 3 zj / L )2/3 } with zj the high of the convective boundary layer Our experimental results show that when L > 0 , V^ is function of the friction velocity and by dimensional analysis we can write Vj / u* = kj = Cte . In the Landes experiment kj = 0.004 and kj = 0.002 for the Stars I and II campaigns. There is however a large variation of ZQ in function of wind direction during Stars. The main difference between Landes and Stars is that there is a substantive air flow through the forest canopy In convective situations, L < 0, the dry deposition velocity is function of the friction velocity but also of the convective velocity scale W*, and we can write dimensionally : Vd / u* = ki + k2 W*2/ u*2 = ki + k2 (0 H z;)2/3 / u*% = ki + k2 (-z; / L)%/3 H is the sensible heat flux, /J the buoyancy parameter. In all our experiments, Figure 5 gives the results of the Stars II experiment. Solid lines represent Vd/u* = 0002 + 0.0009 (zj/ L)2/3 for different values of zj We don't have measured the zj values, but zj increases during the morning with the
Transactions on Ecology and the Environment vol 4, © 1994 WIT Press, www.witpress.com, ISSN 1743-3541
436
Pollution Control and Monitoring
development of the convection . If we represent the experimental results in function of the time of the day (figure 6 ) we have then a realistic distribution of the values with the curves corresponding to different z\. During the day with unstable conditions Vj reach 0.8 cm/s and is lower than 0.1 cm/s during the night in neutral or stable conditions. This type of variation is not given by the different theories of the dry deposition of aerosols particles. Values during unstable conditions are higher than that obtained for fine particles in wind tunnel measurements but confirm results previously obtained in situ experiments, e.g. Hicks et al. (2, 3), Wesely et al (5), Hicks et al (3). 6 Conclusions Dry deposition velocity of submicronic aerosol to be measured need a filtration of low frequency fluctuation of concentration and only eddy correlation method can be applied. This dry deposition velocity can be parametrized in function of the friction velocity and the atmospheric stability The convection in the boundary layer influences the deposition at ground surface and the phenomena is not piloted only in the surface layer The turbulent kinetic energy at ground level, partly produced by convective eddies in the ABL seem to be the leading parameter. Acknowledgements. This work has been supported by the scientific committee " Phase Atmospherique des Cycles Biogeochimiques" of the French CNRS Environmental Programme . References 1. El Bakkali Y. "Etude et realisation d'un analyseur pour la mesure rapide des fluctuations de 1'aerosol atmospherique" Thesis, Universite Paul Sabatier, Toulouse, France, 1991. 2. Hicks, B.B., ML Wesely, J.L. Durham, and M.A. Brown. Some direct measurements of atmospheric sulfur fluxes over a pine plantation. Atmos. Environ. 17,2899-2903, 1982. 3 Hicks, B.B., D.D., Baldocchi, T.P. Meyers, R P Hosker, and D.R Matt. A preliminary multiple resistance routine for deriving dry deposition velocities from measured quantities. Water, Air, and soil Pollution,36 ,311-330, 1987. 4. Lamaud E , Y. Brunet, A. Labatut, A. Lopez, J; Fontan, A; Druilhet. The Landes experiment. Biosphere-atmosphere exchanges of ozone and aerosol particles above a pine forest, J Geosphys. Res. (In press), 1994. 5. Wesely M.N., D.R. Cook, R.L. Hart Fluxes of gases and particles above a decideous forest in winter time. Boundary layer meteorol, 27, 237-255, 1983 6. Wesely M.N., D.R. Cook, R.L. Hart.Measurements and parametrization of particulate sulfur dry deposition over grass. J. Geophys. Res. 90, 2131-2143, 1985.
Transactions on Ecology and the Environment vol 4, © 1994 WIT Press, www.witpress.com, ISSN 1743-3541
Pollution Control and Monitoring
437
nS(n) -i.zaL -1.53 _
c
r,i
T w
los n
Figure 1: Spectra of the variance of vertical wind speed component temperature T, ozone concentration Og, and aerosol concentration C.
w,
Co(vvC)
-3.
Figure 2 : Mean cospectra ^ with the adimensional frequency in the case of z/L (Starts II Experiment). The low frequency contribution is abundant in the case of aerosol.
Transactions on Ecology and the Environment vol 4, © 1994 WIT Press, www.witpress.com, ISSN 1743-3541
438
Pollution Control and Monitoring Figure a
Sc/C
' St/T i r
i
i
i
;
,
;
i
4.5
Figure b i l l !
i
T* = iHi /u*
"~
0 1 3 July 4- 1 4 July * 1 7 July
~
A 18
~
X 2(Huf
3.0
2.0 1.5
i
f
4.0 3.5
2.5
i
1.0 .5 fl 3.a
A X* *< IT ^ $ •
X + xXX H H
4^
z/L
1 % ! 1 ^ i >^ ! ^ 1 • ^ I f \ 1 ^ .3 -2.0
'°" "°* " ^"* "
-
i .a
°f 'k normalized standard
Transactions on Ecology and the Environment vol 4, © 1994 WIT Press, www.witpress.com, ISSN 1743-3541
Pollution Control and Monitoring RwC 1.03 ~~T I 1 .90
439
Figure a
1
T
.63 .48 .20 .0
-.28 -.40 -.60 -.80 -I .00
1 -1.0
1.93
.0
'
z/L I ! ' i .a
RwT i i i—i—i—r
I I 2.2
Figure b
.80
.40 .20 .0
-.20 -.40 -.63 -.80 -1 .00 -2.8
Z/L-. -i.a
.9
i .0
2.3
Figure 4 : Variation of the correlation coefficient between w and C and w and T in fucnction de z/L.
Transactions on Ecology and the Environment vol 4, © 1994 WIT Press, www.witpress.com, ISSN 1743-3541
440
Pollution Control and Monitoring
-.48
-.35 -.30 -.25 -.28 - . 1 5
o 4-
July]] July 14
*
July 17
A X
July 18 July 20
.65
-.18
. 18
Figure 5 : Results of the Stars II experiment. Solid lines represent for different values of zj, relation : Vj / u* = 0.002 + 0.0009 (- z;/ L)2/3
= 1500 m 4- 06hOO-09hOO 09hOO-12hOO 12hOO-15hOO
.835 .838 .ees
zi = 500 m
15hOO-18hOO
.829 .815
zi = 200m
.am
.825
1/L
.8
-.«9
-.38
.a
.13
Figure 6 : Variation of V d / u * = f(l/L) for the day 20 of July (Stars II) in unstable situations. Experimental results are represented by period of 3 hours and solid lines represent the same relation as in figure 5.
