variability. The figure shows the 64 CTD station time seriesthat we took in 1991 at about 65 0S 470W.

In 1973, we thought that the large, steplike structureswere due to a combination of the diffusive instability (thisoccurs when the fresher layer is on top due to the greater dif-fusivity of heat than salt) and the cabbeling instability (this isdue to the nonlinearity of the equation of sea water). Fromthe evidence that the density ratio (the ratio of the densitydue to salinity changes to that due to temperature changes) inthe large steps had an average of 1.03 whereas in the smallsteps in the upper region it had an average of 1.39, we hypoth-esized that the main cause of the large steps was probably thecabbeling instability. This was further justified by the fact thatthe transition between these two regions decreased in depth(pressure level) as one proceeded toward the center of theWeddell Gyre due to the changes in the equation of state thatoccur due to pressure.

The 1991 time series showed that internal waves withapproximately semidiurnal frequency (that is, internal tides)move the principal interlaces up and down to different pres-sure levels. The pressure changes at the largest interface byabout 80 decibars and, thus, may trigger an instability due tothe pressure effect on the equation of state. Casts numbered

about 57 to 60 show a near disappearance of the large steps inthe 1991 series. Although it should be pointed out that thepressure effect is slightly out of phase with this event, perhapsthis is due to the time it would take for the cabbeling instabili -ty to initiate convection.

We are presently conducting both numerical and labora-tory experiments in an attempt to sort out the effects of boththe double-diffusion of heat and salt and the nonlinearity ofthe equation of state of sea water. We have seen that as thelayer thickness is increased (and thus the Rayleigh number)the convection intensifies but that the relative penetration ofthe lighter, warmer water decreases. The nonlinearity of theequation of state should affect the convection quite different-ly since, in this case, there is no reason that the convectionwould be confined to a relatively thin layer.

This research was supported by National Science Foun-dation grant OPP 89-15730.

Reference

Foster, T.D., and E.C. Carmack. 1976. Temperature and salinity struc-ture in the Weddell Sea. Journal of Physical Oceanography, 6(1),36-44.

Photochemistry of antarctic waters during the1993 austral spring

DAVID J. KIEBER, Department of Chemistry, State University of New York, College of Environmental Science and Forestry,Syracuse, New York 13210

The dramatic springtime decrease in stratospheric ozoneover the Antarctic results in an increase in the ultravio-

let-B (UV-B) flux penetrating the surface of the ocean. Thehigher levels of UV-B radiation magnify photochemicaltransformations in the euphotic zone, which, in turn, mayhave negative and positive effects on antarctic organisms. Inparticular, cellular damage can occur, although very little isknown about the extent of this damage to marine organisms.It is reasonable to assume this damage will partly be due tothe photochemistry of the surrounding sea water throughthe production of reactive species [e.g., hydroxyl radical(OH) and hydrogen peroxide]. Increased UV-B fluxes mayhave subtle positive impacts on the marine ecosystem aswell. The UV-B photolysis of dissolved organic matter(DOM), much of which is presumably biologically refractory,will yield biological substrates (e.g., pyruvate). Higher pro-duction rates of these substrates should increase secondaryproductivity in carbon-limited antarctic waters (Kirchman1990).

The importance of photochemical transformations onthe health and growth of marine plankton under enhancedUV-B conditions has not been evaluated because virtuallynothing is known about the photochemistry of antarctic

waters. This prompted us to initiate a laboratory and fieldstudy to examine the photochemistry of antarctic waters inresponse to changing UV-B fluxes. The initial results of thisresearch are reported here.

We participated in an oceanographic cruise aboard theR/V Nathaniel B. Palmer from 10 October to 10 November1993, occupying several stations in the confluence of the Sco-tia and Weddell Sea. Analytes were determined onboardaccording to published procedures: OH radical (Mopper andZhou 1990), hydrogen peroxide (Miller and Kester 1988),flavins (Vastano et al. 1987), and formaldehyde (Kieber, Zhou,and Mopper 1990).

Filtered sea-water samples [0.2 micrometer (rim)] forphotochemical experiments were placed in stoppered quartztubes. Deckboard irradiations were performed by placingquartz tubes in a surface sea-water bath located on the heli-copter deck. The UV-B flux was determined by valerophenoneactinometry (Zepp et al. 1992, pp. 282-285), and the total lightflux was determined with an ILC model 1700 researchradiometer. Laboratory irradiations were conducted using amerry-go-round (MGR) system with a 450-watt medium pres-sure mercury lamp in a borosiicate immersion well (Mopperand Zhou 1990).

ANTARCTIC JOURNAL - REVIEW 1994

100

Production (nM/hr)27.3 27.4 27.5 27.6 27.7

50E

100-c06) 150

200

0102030

.1'

.j

B OH radical

01234

../

C formaldehyde

A free-floating buoy was deployed with quartz flasks sus-pended at six depths in the water column [2, 4, 6, 10, 15, and20 meters (m)] to determine in situ photochemical produc-tion rates. Six quartz tubes were secured to a platform at eachdepth. The buoy was deployed at dawn and recovered atdusk.

Depth profiles for hydrogen peroxide and the photo-chemically formed flavin, lumichrome, were characterized byhigh concentrations within the surface mixed layer, a rapiddecrease in concentrations across the pycnocline, and verylow values deeper in the water column (figure 1). These pro-files are characteristic of these photochemically producedcompounds. Interestingly, the concentrations of hydrogenperoxide detected in these waters [3-35 nanomolar (nM)]were low compared to other open ocean waters (50-150 nM).This may be due to the very low levels of chromophoric DOM(;^290 nM) that were detected. The surface maximum that wasobserved for hydrogen peroxide was presumably photochem-ical in origin, since no evidence for biological production ofhydrogen peroxide was found during this cruise.

Riboflavin was detected throughout the upper 200 in

1). Concentrations ranged from 3 to 35 picomolar (pM),

or, H202 (nM) Riboflavin (pM)27.3 27.4 27.5 27.6 27.70 5 10 15 20 25010 20 30 40

0

50

I100

CL4)

150

200

Figure 2. Depth profiles for (A) sigma-t, and the photochemical pro-duction of (B) the OH radical, and (C) formaldehyde from sea waterthat was collected at station P (59 055'S 52 0 10'W), julian day 298. Sea-water samples were collected from 12 depths and subsequently irradi-ated in quartz flasks in the MGR irradiation system. Higher productionrates were observed for the OH radical in the MGR system comparedto the in situ rates depicted in figure 3. The higher rates can beaccounted for based on the higher temperatures (approximately 40°C)and light intensities that were used in the MGR system.

with the lowest concentrations generally observed during theday in surface waters due to the photolysis of riboflavin. Asubsurface maximum located within the pycnocline was alsopresent, probably resulting from microbial activity in thisregion.

Photochemical production rates of hydrogen peroxideranged from 2.8 to 6.0 nM per hour. The minor variations thatwere observed reflected differences in the water samplesexamined and daily variations in the UV-B light flux (data notshown). In addition to making these surface production-ratedeterminations, we also examined the photoreactivity of theupper 200 m with respect to OH radical and formaldehydeproduction (figure 2). The photochemical reactivity of thewater in the upper 200 m showed little differences for OH rad-ical production. By contrast, formaldehyde was producedfrom water only in the upper mixed layer. No production wasobserved in water below the pycnocline. We are investigatingthe several possibilities that may account for these interestingresults.

The dark losses of lumichrome and hydrogen peroxidewere very slow, with rates of 1.4 pM per day and 0.9 nM perday, respectively. The loss of lumichrome has not been previ-________________ ously investigated, but it is likely to occur

Lumichrome (pM)through adsorption onto particles.0 50 100 150 200Hydrogen peroxide, on the other hand, is

probably degraded by biological/processes. We plan to test these assump-

tions on a future cruise to the Antarctic.The in situ production rate of the

hydroxyl radical decreased exponentiallybetween 0 and 20 in in each ofthe three buoy studies. Results of thethird buoy study are presented in figure3. Depth profiles of OH production par-alleled the decrease in the UV-B lightflux with depth. This relationship sug-gests that the production of the OH radi-

cal is confined to the UV-B band, a finding that is in agree-ment with previous action spectra results (Mopper and Zhou1990). By contrast, hydrogen peroxide production ratesdecreased more slowly with depth, presumably due to its pro-duction in both the UV-B and LTV-A (Cooper et al. 1989). Sur-prisingly, a subsurface maximum in hydrogen peroxide pro-duction was consistently observed between 2 and 4 in allthree buoy deployments. This suggests that either significantphotochemical loss of hydrogen peroxide in surface waterstakes place or that hydrogen peroxide production decreasesat the surface due to processes competing for its precursor,the superoxide anion. We suspect that these competingprocesses will not be as important below the sea surface dueto the rapid attenuation of the shortest wavelength of UV-B inthe upper 2in.

The photochemical production of hydrogen peroxide wasdetermined for five ultrafiltration fractions. Based on our pre-liminary results, the majority of the hydrogen peroxide pro-duction was from the fraction of DOM with a nominal molec-ular weight less than 1,000 daltons.

Figure 1. Depth profiles for (A) sigma-t, (B) hydrogen peroxide, (C) riboflavin, and (D)lumichrome at station M (59 035S 49059'W). The CTD cast was deployed at 1600 hours (localtime) on julian day 294 (21 October 1993).

ANTARCTIC JOURNAL - REVIEW 1994101

A H20230

25

20

15

10

5

0

B OH-3025201510

50

0 2 4 6 10 15 20

>..a0

C

C0

.4-C.,

-D0

CL

Depth (m)

Figure 3. In situ photochemical production rates of (A) hydrogen per-oxide and (B) the OH radical as a function of depth in the water col-umn. Error bars denote the 95 percent Cl.

Perhaps the most exciting aspect of this study was thatwe were able to determine in situ photochemical productionrates for hydrogen peroxide and the OH radical. The buoystudy also provided evidence for photochemical destructionand/or decreased production of hydrogen peroxide in surfacewaters. In situ measurements will form the basis for photo-

chemicalmodels that incorporate depth-dependent lightprocesses.

We thank our graduate students (B.M. Brown, J.G. Qian,and B.H. Yocis) and postdoctoral investigator (S.R. Konduru).We also thank the chief scientist, Patrick Neale, and the cap-tain and crew of the R/V Nathaniel B. Palmer for their supportthroughout the cruise. Thanks are also extended to theAntarctic Research Associates and the Chilean agents (Agun-sa) for their logistical support. We gratefully acknowledge thefinancial support from National Science Foundation grantsOPP 93-12767 to David J. Kieber and OPP 92-21598 to Ken-neth Mopper.

References

Cooper, W.J., R.G. Zika, R.G. Petasne, and A.M. Fischer. 1989. Sun-light-induced photochemistry of humic substances in naturalwaters: Major reactive species. In I.H. Suffet and P. MacCarthy(Eds), Aquatic humic substances. Washington, D.C.: AmericanChemical Society.

Kieber, R.J., X. Zhou, and K. Mopper. 1990. Formation of carbonylcompounds from UV-induced photodegradation of humic sub-stances in natural waters: Fate of riverine carbon in the sea. Lim-nology and Oceanography, 35(7), 1503-1515.

Kirchman, D.L. 1990. Limitation of bacterial growth by dissolvedorganic matter in the subarctic pacific. Marine Ecology ProgressSeries, 62, 47-54.

Miller, W.L, and D.R. Kester. 1988. Hydrogen peroxide measurementin seawater by (p-hydroxyphenyl) acetic acid dimerization. Ana-lytical Chemistry, 60(24), 2711-2715.

Mopper, K., and X. Zhou. 1990. Hydroxyl radical photoproduction inthe sea and its potential impact on marine processes. Science, 250,661-664.

Vastano, S.E., P.J. Mime, W.L. Stahovec, and K. Mopper. 1987. Deter-mination of picomolar levels of flavins in natural waters by solidphase ion-pair extraction and liquid chromatography. AnalyticaChimicaActa, 201,127-133.

Zepp, R.G., M.M. Gumz, D.J. Bertino, and W.L. Miller. 1992. Use ofvalerophenone as an ultraviolet-B actinometer for environmentalstudies. American Chemical Society Environmental ChemistrySymposium Volume (203rd National Meeting of the AmericanChemical Society, San Francisco, 1992). Washington, D.C.: Ameri -can Chemical Society.

Analysis of dimethylsulfoniopropionate from Phaeocystispouchetii in the waters of McMurdo Sound, Antarctica

MICHELE K. NISHIGucHI, Department of Biology, University of California, Santa Cruz, Santa Cruz, California 95064BRIAN DUVAL, Department of Microbiology, University of Massachusetts, Amherst, Amherst, Massachusetts 01003

DuANE P. MOSER, Center for Great Lakes Studies, Milwaukee, Wisconsin 53204

Phaeocystis pouchetii is a marine microalga that formsglobular, gelatinous colonies up to several millimeters in

diameter. Phaeocystis also produces dimethyl sulfide (DMS), asulfur-based compound that is released into the atmosphereand is transformed into sulfate aerosols (Stefels and vanBoekel 1993). From these aerosols, cloud condensation nuclei

(CCN) are formed and are the basis of cloud cover and possi-bly global climate change (Sieburth 1979; Andreae 1991). Theimportance of this solute has global implications; DMS con-tributes approximately 30 percent of the total sulfur found onthe Earth today (Andreae 1991; Andreae and Raemdonck1983). Certain species of phytoplankton are responsible for

ANTARCTIC JOURNAL - REVIEW 1994

102