JOURNAL OF GEOPHYSICAL RESEARCH, VOL. 99, NO. C7, PAGES 14,339-14,343, JULY 15, 1994 Comment on "Sea Spray and the Turbulent Air-Sea Heat Fluxes" by Edgar L Andreas Kristina B. Katsaros • Department of Atmospheric Sciences, University of Washington, Seatfie Gerrit de Leeuw TNO Physics and Electronics Laboratory, The Hague, TheNetherlands 1. Introduction Recently, a paper suggesting that sea spray makes a substantial contribution to thenet watervaporflux into the air at high wind speeds waspublished by Andreas[ 1992]. whose conclusion is in conflict with recent measurements. Andreas' [1992] work is a modeling studyof the magnitude of the evaporation by seaspray, especially spume drops, the droplets ripped off thecrests of ocean waves at highwind speeds. The study is carefully done and Andreas [1992] states his assumptions succinctly; however, we take issue with some of these assumptions. The flux of water by sea spray is currently of great interest, but because direct measurements are nonexistent and even aerosol concentration measurements are extremely difficult to make, this field is at a stage of great uncertainty. The many possibleprocesses involved make intuitive understanding unsatisfactory. Debate is therefore essential. It appears that we will have to wait patiently for some time yet beforea clear consensus can be reached. This commentattempts to balancethe resultsof Andreas' [1992] model with evidence from field and laboratory experiments and from other model calculations. We differwith Andreas [1992]onthefollowing points: 1. His choice of a droplet spectrum is strongly tilted toward large drops being important in theaerosol flux. Conflicting ob- servations in theliterature were not fully discussed. 2. He does not allow hissubstantial droplet evaporation (upto 30% of thetotal flux) to influence theshape of thewater vapor profile with appropriate feedback on the evaporation from the air-sea interface. There is somelaboratory evidence for this effect [Mestayer et al., 1990] and some modelingresults [Stramska, 1987; Smith,1990] that,even though unrealistic in some respects, do show thiseffect. 3. He does not acknowledge the experimental results obtained within the Humidity Exchange Over the Sea (HEXOS) program, namelythat the bulk exchange coefficient for net air- sea water flux (Dalton number) isnotwindspeed dependent up to 18 m/s [Katsaros et al., 1987; DeCosmo et al., 1988; DeCosmo and Katsaros, 1991; Katsaros and DeCosmo, 1990; DeCosmo, 1991 ]. 1Also at Departement d'Oc6anographie Spatiale, IFREMER - Centre de Brest, B.P.70, 29280, Plouzan6, France. (e-mail: katsaros@ifremer. fr) Copyright 1994 by theAmerican Geophysical Union. Paper number 94JC00081. 0148-0227/94/94JC-0081505.00 We now elaborate on the points where we differ with Andreas [ 1992]. 2. Choice of Droplet Spectrum 2.1. Previous Work The scientists sampling sprayfrom breaking waves at sea, eventhough they are workingin very difficult circumstances, have generally come to agree that the majority of drops entering the atmosphere have a radius< 100 $tm. The larger drops have a very short life-time and only contribute a small fraction of their volume to the atmosphere beforefalling back to the sea; they are nonetheless responsible for the larger fractionof water vapor enteringthe atmosphere by the spray mechanism [De Leeuw, 1986; Andreas, 1990; Fairall et al., 1990],but the general conclusion is that spray contributes only a small fraction to the net water flux. This conclusion is based on the followingobservations: 1. Comparison of the aerosol volume derived from particle size distributions measured during the HEXOS main experiment [De Leeuw, 1990a] with simultaneously measured seasaltmass concentrations [Marks, 1990] in wind speeds up to about 25 m/s shows that the moisture released by the particles into the atmosphere is negligible and that the associated latent heat contributes only a few percent [De Leeuw, 1989a]. 2. The total liquid water content of the particles is only a small fractionof the total (liquid and vapor) water content in the surface layer. 3. Hasse [1992] madean approximate calculation of the mass of spray water. He concludes that for all large scaleair-sea interaction work, sea-spray droplets need not be considered as a source of evaporation. This conclusion is based on energy considerations, a sea salt mass flux estimate, and added surface area arguments.Hasse's[1992] note requires verification to the sameextent asAndreas [1992]. As opposed to this experimental evidence, some modeling efforts do show the influence of sea spray on theexchange of moisture. Among theseare the work by Ling et al. [ 1980], Bortkovskii [1987], and Stramska [1987]. However, these results arebased on calculations using unrealistic assumptions concerning the seaspray source function. As noted by Fairall and Larsen [ 1984], the droplet production rate recommended byLinget al. [1980] is a factor of 10 ngreater than theirs at the samewind speed. The source functionusedby Bortkovskii [ 1987] is alsotoo large. It is based on the assumption that all air entrained in a breaking wave returns to the atmosphere by the burstingof bubbles and that eachbubbleproduces a jet droplet. Work by, for instance, Blanchard [1983] and Wu [1989] on the bubble mediated production of droplets shows that thelarger bubbles produce nojet droplets. Thisleads us to 14,339