THE ROLE OF WATER WAVE DYNAMIC PROCESSES IN THE EXCHANGE

OF GASEOUS MERCURY AT THE AIR-SEA INTERFACE

TROMBINO, G., HEDGECOCK, I., FORLANO, Ll.,, Hedgecock, I. AND PIRRONE, N.

CNR-Institute for Atmospheric Pollution, c/o: UNICAL, 87036 Rende, Italy

ABSTRACT

The overall goal of this is work work is to evaluate addresses the role of pollutant exchange mechanism of pollutants exchanges at the air-water interface during the over-water transport of air masses over large bodies of waterin the overall scheme of pollutant cycling. The focus of this paper is to present Tthe preliminary results of the Pollutant Exchange at Air-Water Interface (PEAWI) model developed in the frame of the MAMCS project are presented. model Ddeveloped tofor estimateing the exchange rate of semi-volatile contaminants at the air-water interface, the PEAWI model takes into account with changing meteorological conditions and chemistry ofin the lower atmosphere and top-seawater microlayer. The PEAWI model modelsaccounts for the sea spray formation and bubble ejection with changing meteorological conditions and the consequent effect on pollutant exchange as a function of the pollutant characteristics. during the over-water transport of air parcels. It ishas been A new approach to the problem has been formulated and particular attention has been paid to the role of wind velocity in the exchange process. The result of these studies is the Pollutant Exchange at Air-Water Interface model elaborated in this paper. The most important characteristic of this model designed is its capability to calculate air-water exchange rates for all wind speeds velocities in the range 0-20 m/ s^-1. The PEAWI model is This model was linked to the Regional Atmospheric Modelling System (RAMS) which provides aallow the use of a detailed description of the microphysics of the atmosphere above the seawater. As first attempt, the The first use of PEAWI model has been was applied to calculate the exchange rates of Hg at the air-water interface of the Mediterranean Sea.Thus, using the balance of mercury exchanged at the air-water interface a regional scale assessment of mercury released from the water surface to the atmosphere with changing meteorological conditions has been calculated.

INTRODUCTION

Pollutant exchange at the air-water interface is a problem of some complexity and is not always considered to be of major importance in the overall scheme of pollutant cycling. Actually, tThere are not many models that describe this process and those that exist they are mostly used only to study the air-water exchange of POPs. The two-film model (Liss and Slater, 1974; Schwartzenbach et al., 1993) is the most commonly used tofor describing e air-water exchange, although surface renewal (Daenkwerts, 1951; Asher and Pankow, 1991) and boundary layer (Deacon, 1977; Kerman, 1984) models are sometimes employed. This study presents the preliminary results of a new sea spray aerosol generation model describes a new approach to the problem using the relation between wind speed and aerosol generation. This approach is used to assess the cycle of based on the transport of semi-volatile contaminants via the sea spray formation in the Mwhich is component of marine Baerosol in the marine boundary Llayer (MBL). Wind stress at the sea surface generates droplets in the size range of 37.5 mm to 400 mm. The droplet radius is of great importance as it determines whether gas phase elemental mercury is removed from or released to the atmosphere. In order to model pollutant exchange over a range of wind speeds it proved necessary to combine models which describe the mechanics The most important input parameters in this approach are the wind velocity at 10 m above the mean sea level, and the concentration of pollutant species in air and water. The principle problem of this approach is that it cannot work at wind speeds lower than 7.5 m/s or greater than 20 m/s. A two-film model is used for low wind velocity. If the wind speed is greater than 20 m/s it is very difficult to describe the atmospheric conditions.of the exchange processes at a given wind speed.

MODEL DESCRIPTION

1. High wind speed

The main purpose of tThis moduleel is to calculates investigate pollutantthe exchange rates of semi-volatile contaminants between the atmosphere and seawater that for wind speeds above the 5-6 m s^-1 . Above this speed, droplet production, depends primarily by is connected with sea spray formation and bubble ejection mechanisms at the air-water interface, constitutes the major factor influencing pollutant exchange. For the purpose of model development we have assumed until now that the time required for the aqueousliquid phase concentration of Hg^o to reach equilibrium with the gas phase concentration is extremely small, i.e. that it happens effectively instantaneously, for the whole droplet size spectrum. This may not necessary be true for all droplet sizes and is currently being investigatedthe subject of another paper presented at this meeting.(Hedgecock et al., 2000). The energy and momentum transferred by the wind to the sea surface produces liquid droplets. Thus:, bursting bubbles give rise to spray droplets and; wind tears off wave crests and blows spume directly into the air. To evaluate how the sea spray droplets contribute to the air-sea flux of mercury it is necessary to estimate the rate at which droplets of any given size are produced. That is, we must estimate the so-called sea spray generation function, which involves to quantifying the spray droplet production mechanism. The droplet radii are described by a spectral generation function, F, such that Fdr is equal to the number of particles produced per m² in unit time with radius (r, dr). The radius range is 15 µm < r < 400 µm. Droplets with a radius greater than 37.5 µm before falling to the sea surface remain in the atmosphere for a fraction of second, we assume this is long enough for them to reach chemical-physical equilibrium with the gas phase mercury present in air. Droplets with a radius less than 37.5 µm start to evaporate, (until thermodynamic equilibrium between liquid water and water vapour, given the salt concentration and relative humidity, is obtained), and release part of their mercury content to the atmosphere.

Aerosol Production

A The key to more accurate estimates of the magnitude of the spray generation function lies in identifying the physical mechanisms that produce the spray droplets of various sizes. For example, wave breaking causes entrainment of air. Droplets may be produced directly or indirectly. In the first case wWave breaking can produce droplets directly: spume and splash droplets. These droplets have radii greater than 20 mm. Spume droplets result from mechanical tearing of the sharpened wave crest by the wind ( Monahan et al., 1983a; Monahan 1986). Splash droplets are a consequence of the vigorous spilling, or curling over of the crests of breaking waves. Droplets are also produced indirectly, wave breaking causes entrainment of air and as a result with the mediation of bubbles; two distinct types of droplet form,: film and jet droplets. When the upper, protruding surface of a bubble rising to the air-sea interface thins sufficiently (as it typically does within a second of a bubble’s arrival at the sea surface) it shatters, producing anywhere from a few to a few hundred film droplets. Jet droplets are pinched off the end of the microscopic column of water that rises out of the centre of the collapsing cavity left after the bubble cap ruptures. This column is called the Rayleigh jet, it results from the rebound, or overshoot, of the sea surface caused by its surface tension.

Generation Function

The model uses a generation function whichcan calculates the number of particles droplets produced per square surface area, per second, per micron increment in r. The radius range is 150-400 mm because particles of greater diameter need wind velocities higher than 20 m/s, and the residence time in the atmosphere is extremely brief. The generation functions used are the Blanchard-Gathman, the Andreas and that proposed by Wu. Blanchard (1963) was one of the first to estimate the spray generation function. Using the sea-salt aerosol distribution fromof Woodcock (1953) (measured at 600 meters above sea level in the marine boundary layer near Hawaii) Blanchard deduced the spray generation function for wind speeds between 5 and 15 m/s. Gathman (1982) fitted Blanchard’s curves with mathematical functions for the Navy Aerosol Model. We have used a mathematically modified version of the Blanchard-Gathman function. In the Blanchard-Gathman model, a relative hHumidity of 91.4% is assumed.

_{DF/dr91.4 = (2.152*104C2/r91.4) exp[C3(ln(C4/r91.4))2] (1)}

_{This gives the number of spray droplets produced per square meter of surface per second, per micrometer increment of r91.4.} It is very difficult to use a generation function in this form, so we have used the following modifications:

dF₀/dr₀ = (dr_91.4/dr₀)(dF/dr_91.4) (1a2)

r_91.4 = 0.627r₀^1.002 (53)

dr_91.4/dr₀ = 0.628r₀^0.002 (64)

Wu‘s model is used to calculate aerosol with radii in the range 37.5mm<r₀<400mm. Wu’s generation function is based on measurements with a lower cut-off radius of 60 mm. In Wu’s model the total production of spume drops is adjusted to include production in the radius range 37.5-60 mm. The total rate of spume drop production per unit of area of the sea surface is:

P_s = 8.7*10^-5exp(0.875U₁₀) (75)

The spectral production function of spume drops can be expressed by:

dP_s/dr = 9.06*10^-1 P_s*r ^-1 37.5mm < r < 75 mm (86)

dP_s/dr = 5.10*10³ P_s*r ^-3 75mm < r < 150 mm (97)

dP_s/dr = 3.87*10¹⁴ P_s*r ^-8 r > 150 mm (108)

in which dP_s/dr (m^-2s^-1mm^-1) is the number of spume drops produced per unit area of sea surface area per second per unit radius bin. The last generation function considered is the Andreas function:

dF/dr₈₀=C₁(U₁₀)r₈₀^-1 15mm £ r₈₀ £ 37.5 mm (119)

dF/dr₈₀=C₂(U₁₀)r₈₀^-2.8 37.5mm £ r₈₀ £ 100 mm (1210)

dF/dr₈₀=C₃(U₁₀)r₈₀^-8 100mm £ r₈₀ (131)

Where we can convert dF/dr₈₀ in this way:

dF/dr = (dr₈₀/dr₀)dF/dr₈₀ (142)

and

r₈₀ = 0.518r₀^0.976 (153)

thus

dr₈₀/dr₀ = 0.506r₀^-0.024 (164)

2. Low wind speed

The two-film model considers mass transfer to be limited by the rate of molecular diffusion through thin films of air and water on either side of the surface. The net flux (F, ng/m²/day ):

F = K_ol (C_w - C_aRT/K_H) (175)

where

1/K_ol = 1/K_L + RT/HK_g (186)

where K_ol, K_l, and K_g are the overall liquid and gas mass transfer coefficients (m/s), C_w and C_a are the concentration of dissolved and gaseous pollutant in bulk water and air (both ng/m³), K_HH is the Henry’s law constant (atm ´m^-3 ´mol^-1), R is the gas constant (82*´10^-6atm´ m³/(mol^-1´°K^-1)). The critical variables are K_HH and the transport terms K_l, and K_g. The Henry’s law constant is a function of temperature and salinity and it may be obtained from solubility and vapour-pressure data or may be measured directly by gas stripping. K_l and K_g are strongly influenced by wind speed, atmospheric stability and sea surface conditions (breaking waves, bubble injection). Mackay and Yeun (1983) proposed an equation for K_l and K_g as a function of friction velocity (U_*) and the Schmidt number of the pollutant compound.

K_g = 1.0*10^-3 + 42.6 + 10^-3 U^* Sc_g^-0.67 (197)

K_l = 1.0*10^-6 + 34.1*10^-4 U^* Sc_l^-0.5 (U^* > 0.3) (2018)

K_l = 1.0*10^-6 + 144*10^-4 U^*2.2 Sc_l^-0.5 (U^* < 0.3) (2119)

These relationships were used to estimate gas fluxes of POPs in the Great Lakes (Baker and Eisenreich, 1990; MacConnel et al., 1993a; Pirrone et al., 1995) and the worlds oceans (Cotham and Bidleman, 1991; Hinckleyet al., 1991; Iwata et al., 1993)

RESULTS AND DISCUSSION

The development of the

The P.E.A.W.I. model model is an effort to create a differentcomprehensive approach to the study the ppollutant exchange at the air-water interface. The main problem is to couplinge two generation functions that can calculate the volumes of liquid aerosol produced with reasonable precision. Thus, we have studied each generation function previously used and we have made two combinations, reported in Fig.1 and Fig.2. We can calculate the volume of liquid droplets by dividing them into evaporated and deposited volume with varying values of wind velocity. The first coupled pair, Blanchard-Gathman and Wu, shown in Fig.1, maycan be related combined very well because theyse show the same trend in aerosol production trend. The other pair is Blanchard-Gathman and Andreas which seems to be the better pairing. In fact these functions have the same trend for the same value of wind speed. The P.E.A.W.I. model strongly depends on pollutant concentration in the sea, see Fig.3. We have considered as test cases studied mercury concentrations in the top-water micro-layer between as a pollutant at. Cconcentrations between, of Hg is between 0.001ng/Ll and <C₀<0.1ng/Ll.. When the concentration is high (C₀=0.1 ng/l) the aerosol that which is eventually re-depositeds on the sea surface, releases the mercury whilst it remains in the air. So we have pollutant emission from droplets (in the range 37.5mm<r₀<400mm) before they falling back to the sea. When the mercury concentration in the sea is low (C₀=0.001 ng/l) the trend is inverted and the aerosol absorbs Hg, whereas . This is an important observation because if the P.E.A.W.I. model is used for a strongly polluted sea it calculates predicts a strong high emission of Hg to the air.and low adsorption so the air-water exchange can be considered like a phenomenon of droplets release. ????Non so se va bene While if we study a sea withat normal low mercury concentrations in the seawater, is studied P.E.A.W.I. model predictsthedescribes air-water -exchange flux asis athe balance between adsorption and re-emission. Actually we are studing Aa new generation function is under development that will be based on experimental data and it willin order to reduce current discrepancies the difference between observed an calculated fluxesdata.

REFERENCES

Andreas E. L., Edson J. B., Monahan E.C.,. Rouaul M.P,. Smith S.D. 1995. The Spray Contribution to Net Evaporation from the Sea: A Review of Recent Progress. Boundary-Layer Meteorology 72: 3-52

Andreas E. (1992) Sea Spray and the Turbulent Air-Sea Heat Fluxes. J. Geophys. Res. 97: 11, 429-11,441.

Bidleman, T. F. and McConnel, L. (1995) A review of field experiments to determine air-water gas exchange of persistent organic pollutants. Sci. Tot. Environ. 159, 101-117.

Pirrone, N., Keeler, G.J. and Holsen, T.M. (1995) Dry Deposition of Semivolatile Organic Compounds to Lake Michigan. Environ. Sci. and Technol., 29, 2123-2132.

Shannon J. D. and Voldner E. C. (1995) Modelling Atmospheric Concentrations of Mercury and Deposition to the Great Lakes. Atmos. Environ, 29; 1649-1661.

Wu, J. (1993) Production of Spume Drops by the Wind Tearing of Wave Crest: The Research for Quantification. J. of Geophys. Res. 98: 18,221-18,227.

Wu, J. (1993) Production of Spume Drops by the Wind Tearing of Wave Crest: The Research for Quantification. J. of Geophys. Res. 98: 18,221-18,227.

Fig.1 - Evaporation [EV(10^-12´m3´m^-2´s^-1)] and Deposition [DV(10^-9´m3´m^-2´s^-1)] fluxes calculated using the generation functions of Blanchard-Gathman and Wu.

Fig.2 - Evaporation [EV(10^-8´m3´m^-2´s^-1)]and Deposition [DV(10^-6´m3´m^-2´s^-1)] fluxes calculated using the generation functions of Blanchard-Gathman and Andreas.

Evaporation and Deposition calculated using generation functions of Blanchard-Gathman and Andreas.

Wu, J. (1993) Production of Spume Drops by the Wind Tearing of Wave Crest: The Research for Quantification. J. of Geophys. Res. 98: 18,221-18,227.

Shannon J. D. and Voldner E. C. (1995) Modelling Atmospheric Concentrations of Mercury and Deposition to the Great Lakes. Atmos. Environ, 29; 1649-1661.

Bidleman, T. F. and McConnel, L. (1995) A review of field experiments to determine air-water gas exchange of persistent organic pollutants. Science of the Total Environment 159, 101-117

Fig. 3 - Temporal and spatial variations of the exchange rate of Hg(0) at the air-water interface along a 72-hr forward 1-hr time step trajectory.