Diatom flux in McMurdo Sound, Antarctica

Marine Micropaleontology, 12 (1987): 49-64

49

Elsevier Science Publishers B.V., Amsterdam - - Printed in The Netherlands

DIATOM FLUX IN MCMURDO SOUND, ANTARCTICA AMY L E V E N T E R and ROBERT B. D U N B A R Department of Geology and Geophysics, Rice University, Houston, Texas 77251 (U.S.A.) ( Revised and accepted August 15, 1986)

Abstract Leventer, A. and Dunbar, R.B,, 1987. Diatom flux in McMurdo Sound, Antarctica. Mar. Micropaleontol., 12: 49-64. Quantitative floral analyses have been performed on sea ice, sediment trap, and surface sediment samples collected from six sites in McMurdo Sound, Antarctica. Although diatom abundances in the sea ice reach 108-109 cells m 2, maximum diatom fluxes between October-December, 1984, as measured in the sediment trap samples, are only 10~- 107 individuals m -2 day -1. The cumulative flux of diatoms for this three month period is approximately 107-10 ~ individuals m-2, possibly an order of magnitude less than the sea ice abundances, implying that diatom and opal flux must significantly increase when the sea ice melts and releases those particles previously trapped in the ice. Five species dominate the sea ice assemblage - Amphiprora sp., Pleurosigma sp., Nitzschia stellata, Pinnularia quadratarea, and Nitzschia curta. These species are also common in the water column, along with Thalassiosira spp., a genus rarely found within the sea ice. Within the upper 250 m of the water column, at a site in Granite Harbor, diatom flux decreases between 47-79% from 34 to 220 m. Opal flux, however, decreases by only 13-40% over this same interval, indicating that dissolution of thinly silicified diatom frustules has occurred. At mid-water depths, increases in absolute diatom flux of two species in particular, Nitzschia curta and Thalassiosira spp., may indicate their transport from other areas. At all sites, the greatest increase in diatom flux occurs just above the sea floor. Resuspension of diatom tests and/or lateral advection creates a near-bottom nepheloid layer in which significant modification of the diatom assemblage occurs. A combination of preferential dissolution of those forms which dominate the sea ice and dilution of the assemblage with resuspended and/or advected diatoms representative of many year averages of the flora may be responsible for the production of a sediment assemblage primarily made up of Thalassiosira spp., Nitzschia curta, and other robust forms of Nitzschia.

Introduction Sea ice cover in the Southern Ocean ranges from a minimum of 3×106 km 2 during late summer to a maximum of 20 × 106 km 2 during late winter (Zwally et al., 1979). The area of the Southern Ocean covered by late winter sea ice is roughly equivalent to the size of the Antarctic continent. Although the presence of annual sea ice reduces the amount of light available for primary production (Thomas, 1963 ), the sea ice harbors a rich microbial community (Bunt, 1963, 1964; Burkholder and

0377-8398/87/$03.50

Mandelli, 1965; Horner, 1976; Sullivan et al., 1983). Primary production within the annual sea ice in M c M u r d o Sound may exceed 40 g carbon m -2 yr -1 (McGrath-Grossi et al., in press). Estimates of annual primary production for Antarctic waters range from 16 to 100 g carbon m -2 yr -1 (Ryther, 1963; H o l m - H a n sen et al., 1977). Despite the presence of extensive seasonal ice cover in the Antarctic, and its influence on primary productivity, little is known about vertical transport processes in such a setting. In this paper, we describe the first use of diatom fluxes

© 1987 Elsevier Science Publishers B.V.

50 to study systematically the dynamics of sinking particulate matter in an area covered by annual sea ice. We have focused on two principal questions: (1) What are the major sites of silica production, dissolution, and recycling on the Antarctic continental shelf?. Diatom production within the photic zone is the dominant mechanism for the uptake of dissolved silica from the oceanic water column (Heath, 1974). Annual silica production on the Antarctic shelf ranges between 100-500 g SiO2 m -2 yr -1, with diatoms accounting for > 99% (by weight) of the total ( Lisitzin, 1972 ). In lower latitude oceanic gyre settings, annual silica production is generally much lower, <100 g SiO2 m -2 yr -1, and diatoms may account for < 35% of the total (Lisitzin, 1972). Removal of silica from the ocean to long-term storage reservoirs occurs via burial of opal-bearing sediments, particularly in the Southern Ocean (DeMaster, 1981). Ledford-Hoffman et al. (1986) have estimated that as much as one-third of the dissolved silica supplied to the oceans is ultimately sequestered on the Antarctic shelf. A study of silica production, transport, dissolution, and sedimentation on the Antarctic margin is esential to an understanding of the global silica cycle. (2) To what extent does the sedimentary microfossil assemblage reflect the processes occurring in the sea ice, water column, and at the sediment-water interface? These processes include production, release, settling, dissolution, winnowing, resuspension, and deposition. The answer to this question is particularly important considering the already extensive use of diatoms in paleoceanographic and stratigraphic studies of the Southern Ocean (Jousd et al, 1963; McCollum, 1975; Fenner et al., 1976; Gombos, 1976; Schrader, 1976; Kellogg and Truesdale, 1979; Truesdate and Kellogg, 1979; DeFelice and Wise, 1981; Burckle et al., 1982; and many others). Truesdale and Kellogg (1979), for example, observed that the epontic species Nitzschia curta dominates surface sediments in areas of the Ross Sea where undisturbed sedimentation occurs, and have

interpreted the downcore presence of this assemblage in Ross Sea cores as indicative of Holocene conditions similar to those presently occurring (Kellogg and Truesdale, 1979). We note, however, that most of the commonly reported sea ice diatoms from this region (McGrath-Grossi, 1985; McGrath-Grossi and Sullivan, 1985) are not recorded in these sediments. To interpret downcore changes in the sedimentary floral assemblage requires a prior knowledge of the interplay of factors involved in producing that assemblage. Our strategy in answering these questions has been to examine variability in the character and composition of biogenic particulate debris collected from within the sea ice, throughout the water column, and at the sea floor. During October through December, 1984, sea ice and sediment trap samples were collected from 12 sites in McMurdo Sound (Fig. 1). The sites were chosen to encompass a variety of glacial subenvironments and a wide range of water depths.

Study area McMurdo Sound is located in the southwestern corner of the Ross Sea. It is an inlet approximately 70 km wide by 80 km long bounded by the volcanic edifice of Ross Island to the eas~. the Transantarctic Mountains to the west, the Ross Ice Shelf to the south and the open Ross Sea to the north. McMurdo Sound can be divided into four bathymetric regions including a western shelf with an average depth of 200 m. a gentle western slope (1 ° ). two deep basins and a steep slope (6 ° ) off Ross Island ( Barrett e~ al., 1983; Pyne et al., 1985 ). Compositionally, modern sediments in the sound are highly variable. Basaltic fragments are supplied from the McMurdo volcanics. The Transantarctic province supplies late Precambrian to early Paleozoic granitic and metamorphic debris, Devonian to Triassic sandstones of the Beacon supergroup, and Jurassic dolerites. This terrigenous debris enters the sound as aeolian particles and as supraglacial and englacial debris (Barrett et al., 1983 ). Biogenically

51

, 77oS

77o15 ,

77o30 '

77o45 '

162 °

163 °

164 o

165 ° E

166 °

167°

Fig. 1. Locationof study area and samplesites. Dashedline indicatesapproximatelimit of Octobersea ice. produced sediments are widespread, with biogenic silica percentages as high as 35% found in surface sediments from the deeper areas of the sound (Dunbar et al., 1984). Although sponge spicules may form a significant component of the siliceous debris at depths less than 100 m, diatom tests predominate in the deeper, finegrained sediments of the sound. Circulation in McMurdo Sound is generally cyclonic with High Salinity Shelf Water (HSSW) entering the sound from the northeast, turning westward and then exiting northward along the western side (Heath, 1977; Jacobs et al., 1985; Lewis and Perkin, 1985). HSSW is characterized by salinities > 34.75%0 and temperatures near the surface freezing temperature (Jacobs et al., 1985). HSSW is

concentrated near McMurdo Sound by both the general cyclonic circulation pattern of the Ross Sea and by the presence of sills to the north which act to "trap" water in the sound (Lewis and Perkin, 1985). Water advected from under the Ross Ice Shelf contributes to the generally northward flow along the western sound (Heath, 1977; Barry and Dayton, 1985; Lewis and Perkin, 1985). The upper 20 m of the water column exhibit large seasonal temperature and salinity fluctuations in response to the annual cycle of freezing and melting of the sea ice. Beginning in March or April, congelation ice begins to form over the sound. The ice continues to thicken until about November, when it reaches an average thickness of 2 m. Under-ice irradiance is

52 typically < 1% of surface irradiance and varies according to both ice thickness and snow cover (Sullivan et al., 1984). Levels of sub-ice primary productivity are unknown. In the summer (January-February), high melting rates in the eastern Sound have been attributed to the southward advection of relatively warm oceanic water ( Tressler and Ommundsen, 1962; Heath, 1977; Mitchell and Bye, 1985). The process is one of positive feedback during which melting is enhanced by the production of a shallow pycnocline which favors the propagation of internal waves. The waves bring the warm subboundary water in contact with the sea ice, causing additional melting. In the western Sound, melting is thought to be related to the surface intrusion of coastal meltwater, which raises the temperature of the sea ice (Mitchell and Bye, 1985). We have examined particulate matter from the sea ice, water column, and sediment in three regions of McMurdo Sound ( Table I). Site I is located in the northwestern portion of the Sound, in Granite Harbor. The study site is approximately 20 km from the floating terminus of the MacKay Glacier, one of the fastest moving glaciers providing ice drainage through the Transantarctic Mountains ( > 200 m yr--1, Barrett et al., 1983). Sites E and F are both located in southwest McMurdo Sound, in New Harbor. Site E is located i km from the terminus of the slowly advancing Ferrar Glacier ( < 10 m yr -~, Barrett et al., 1983), while Site F is farther east of the terminus, Sites B, D and L are located in southeastern McMurdo Sound. Site B is located approximately 75 m west of the tip of the Erebus Ice Tongue, site D is situated about 5 km west of McMurdo Station,and site L is located 100 m north of Hut Point Peninsula. Methods During October through December, 1984, 45 sediment traps were deployed on 12 moorings suspended beneath the annual sea ice. Our traps are open-cone, single-cup sediment traps simi-

lar in design to those of Soutar ( Dymond et al., 1981). Large traps (1950 cm 2 collecting area) were deployed through 91 cm holes drilled with the USARP mobile drill. Smaller traps (400 cm ~ collecting area) were used in remote locations where 30 cm holes were drilled with a hand-held auger. At each site, between 2-8 traps were suspended beneath the sea ice on wire rope or braided nylon. Traps were hauled out, serviced, and redeployed approximately every two weeks from mid-October to late-December. To enhance sample preservation, collection tubes were filled with a dense (40%~.) solution of NaCI, approximately 1%~ HgC12, and approximately 1% gluteraldehyde. Sea ice cores were collected at most study sites using a Sipre Auger. The lower 20 cm of each ice core was returned to the lab and melted at room temperature. Samples were preserved with approximately 1% gluteraldehyde. Surface sediment samples from McMurdo Sound were obtained from the collections stored at Victoria University of Wellington, New Zealand. Sediment trap and sea ice samples were split to provide working and archive aliquots. Prior to splitting, zooplankton were hand-picked from the samples and processed separately. These zooplankton included pteropods, copepods, and amphipods. Total particulate flux was determined by filtering a 1/16 sample split onto a preweighed 0.4 pm Nuclepore filter. Filters were rinsed with deionized water, dried, and weighed. Aliquots for biogenic silica analyses were filtered, rinsed, gently washed off the filter paper with deionized water and oven-dried. The dried sample was ground and homogenized with a mortar and pestle and analyzed following a NaOH dissolution procedure modified from DeMaster (1979, 1981). Counts of both relative diatom species abundance and absolute diatom flux were performed on sediment trap samples. For both types of determinations, quantitative splits were filtered through 0.4 #m Nuclepore filters, rinsed with deionized water and dried. Dried filters were mounted on glass slides with Hyrax

53 TABLE I Mooring location information and trap servicing schedule Site B

D

E

F

I

L

Latitude Longitude Water depth (m) Trap depths (m)

77°42 ' 166°21 ' 415 25 51 77 103

77°52 ' 166°30 ' 172 28" 57 109 161

77°40 ' 163°36 ' 264 47 99 151 203 255

77°38 ' 163°46' 250 32 84 186 238

76o56' 163°13' 715 34 127 220 313 406 499 592 685

77°51 ' 166°37' 41 15 37

Deployment Service Deployment Service Deployment Service Deployment Service Deployment Service Deployment Service Deployment Service

10/21

10/26 11/12 11/14 11/29 11/29 12/6 12/6

10/28 11/18 11/18 12/13 12/13 12/26

10/28 11/19 11/19 12/10 12/10 12/27

11/3 11/24 11/24 12/17

11/29 12/5 12/5 12/11 12/11 12/21

11/7 11/10 11/15 11/15 11/23 11/23 11/30 11/30 12/7 12/7 12/15 12/15 12/21

12/18

"Until 11/12.

(refractive index = 1.63). Relative species abundances were d e t e r m i n e d by counting t ransects at 1000 X, until 500-600 specimens had been counted. Absolute diatom fluxes were determined by scanning and counting the entire slide at 625 X. For the sea ice and surface sedimen t samples only relative species abundances were determined. S m e a r slides of t he surface sediment were utilized in order to alter the assemblage as little as possible and to obtain results comparable to t he sea ice and sediment trap data. At least 500 diatoms were c o u n t e d along transects on each slide. For all sea ice and sediment trap samples only diatoms with more t h a n one half of the frustule intact were counted

in order to avoid counting the same specimen twice. For sediment samples, valves rat her t h a n frustules were counted.

Results D i at om fluxes range from < l x 1 0 5 to > 4 X 1 0 7 individuals (frustules) m -2 day -1 (Fig. 2). In general, diatom fluxes increase slightly with depth, except at the Granite Harbor site ( S i t e / , Fig. 3 ), where an initial decrease in diatom flux is observed in the upper water column. During the earlier part of the season, diatom flux decreases by 62% while opal flux decreases by 49%, between 34 m and 220 m. T h e

54

DIATOM

FLUX ('#,/m2/day

0

1 .

0 t

!

2

.

.

.

3 ,¢

~

DIATOM

4

I k

100

x 10 6)

~

-i

12/6-12/t8

1

¢

100-

~ .~

,oo -

,,,,,,-,

,,

I

/

11/3-11/24

\

....

\

12113

400-

12/13-12/26

oo, ........

SITE E

600-

1

700-

'

0

5

L

300"

----. / z z , , ,

4

i

~-~ 11/24-12/17

HI0/28-11/18

lOO

3

i

\

3 ~

A---dL11/18

E

2

200-

2 ---J- .....

x lO 7)

/+

/

.......

SITE D

1 ,

1

~, 11112.-11129

i.

2ooI,. :...... .~ ............~ : z

0

0

FLUX (,~/m2/day

,,t,,lll',.,Irr,/,.,,¢t,r,,r.r,r,,,,,,t,,,,,,rr.1

H11119-12/1Q b...~A10/28

SITE I

11119

100 0

"

'

*

200

_= ./zz,..,

~

~\\

o

4

~./,.z,z,,

-= 1 1 / 2 9

12/5

4~--.4k 1 2 / 5 - 1 2 1 1 1

i.-.!

12/11

12]21

l, ~,,,,,,

SITE L Fig. 2. Flux of diatoms versus depth for five mooring sites. Note that the scale for the x-axis for sites D, E, and F is ( × 10 ~) ,but is ( × 10 7) for sites I and L. Also note the expanded depth scale for mooring L.

OPAL FLUX (mg/m2/day) 0

100 I

i

200 i

i

300 =

i

400 ,

DIATOM FLUX (÷/m2/day x 10 6) 0

10

i

i

i

20 i

I

p . J

A

vE

/

200-

-1-

a. IJJ 4 0 0

600

;

~ 1113-11124

~-~ 11t24-12/17

Fig. 3. Opal and diatom fluxes for site I, Granite Harbor.

11/3-1

1/24

,~-'~ 1 1 / 2 4 - 1 2 / 1 7

30 I

I

40 i

i

55 TABLE II

Mooring B

Mooring D

Mooring F

1001

Opal and diatom fluxes at 34 and 220 m, for two sediment trapping intervals, at site I. The % decrease in both opal and diatom flux from 34 to 220 m is also listed

LU

Depth (m)

11/3-11/24

11/24-12/17

Diatom flux (no.

Opal flux (mgm 2

Diatom flux (no.

m -2 day-1

day-')

m -2 day-1

× 106) 34 220

1.9 1.0

>

Opal flux (rag m - ~ day 1)

,..J

X 106 4.7 2.8

12.3 2.6

14.2 12.3

79

13

.~o % Decrease

47

40

@

[ ] Amphiprora sp.

late season diatom flux at 220 m is one fifth the flux at 34 m, while the opal flux decreases by less than 15% over this same interval (Table II). The greatest increases in vertical flux always occur in the bottom-most sediment traps (Fig. 2). The near bottom diatom flux at site E, for example, is as much as thirty times the flux to shallower depths. This near bottom increase in diatom flux is mirrored by increases in the biogenic silica flux (Fig. 3). A progressive seasonal increase in diatom flux, excluding bottom traps, is exhibited at moorings D, I and L. At the three deep water sites ( > 200 m) illustrated in Fig. 2 (moorings E, F, I), the early season near-bottom increase in diatom flux exceeds that occurring later in the season. At site I, for example, the average flux at 685 m from 11/3 to 11/24 1984 is 1.7 times the flux from 11/24 to 12/17 1984. Generally, the floral composition of the shallow sediment trap assemblage is quite similar to that found in the sea ice, with Amphiprora sp., Pleurosigma sp., Nitzschia stellata, and Nitzschia curta as the dominant diatoms (Fig. 4). A similar congelation ice assemblage has been found by McGrath-Grossi (1985) in McMurdo Sound, by Fukushima and Meguro (1966) in Lutzow-Holm Bay (Queen Maud Land), and by McConville and Weatherbee (1983) in Vincennes Bay (Wilkes Land).

.@~ ~

.~

~,

~J Pinnularia quadratarea

[ ] Pleurosigrna sp.

[ ] Thalassiosira sp.

[ ] N i t z s c h i a stellata

[ ] Nitzschia curta

m Fragilaria islandica

[ ] Berkeleya sp.

Fig. 4. Cumulative percentage of d o m i n a n t species of diatoms for sea ice a n d corresponding shallow sediment traps. Mooring B data, from the tip of the Erebus Ice Tongue, is averaged over the period 11/15-12/16 1984. Mooring D data, 5 k m west of M c M u r d o Station, is averaged over the period 11/29-12/18 1984. Data from Mooring F, in New Harbor, is averaged over the period 10/28-12/10 1984.

The occurrence of similar assemblages in the sea ice and water column has been reported previously (Hoshiai and Kato, 1961; Meguro, 1962; Burkholder and Mandelli, 1965; Fukushima and Meguro, 1966; Krebs, 1977; Ackley et al., 1978; Garrison and Buck, 1985). This similarity may result from a seasonal cycle of entrapment of diatoms during sea ice formation followed by their release as the ice melts (Garrison and Buck, 1985 ). Distinctly different floral assemblages observed in both the sea ice and sediment trap samples occur in different regions of the Sound. For example, at Moorings B and D, in the southeastern Sound, Fragilaria islandica var. adeliae is found in both the sea ice and shallow traps. This species is not found at site F, 65 km across the Sound in New Harbor. Pinnularia quadratarea is relatively abundant in the samples from site F, but not seen in significant numbers outside of the western Sound. It is not

56 assemblage - Amphiprora sp., Pleurosigma sp., and Nitzschia stellata. Togethter they compose an average of 82% of the sea ice flora. Pinnularia quadratarea and Nitzschia curta are less abundant, constituting another 10%. Thalassiosira spp. are rare in the sea ice samples, <0.5%. From 11/3-11/24 1984 (Fig. 5.A,B), the water column can be divided into four sections (0-127, 127-313; 313-592, 592-715 m), according to the distinct trends observed and the processes occurring within these depth ranges. Within the uppermost water column, 34-127 m, total diatom flux decreases from 1.9-0.7 × 106 individual tests m - 2 day 1, possibly the result of near-surface dissolution. Dissolution does not appear to disproportionately affect particular species. Although absolute fluxes of Amphiprora sp., Pleurosigma sp., Nitzschia 8tetlata and :Pinnularia quadratarea remain approximately constant between 127 and 592 m, the relative

known what factors cause these geographic differences in diatom assemblage or if the assemblages remain constant from year to year. Thalassiosira spp. are rarely found in the sea ice samples, reaching maximum abundances of < 0.5%. This group is common in the sediment trap samples however, as best illustrated by the data from site D (Fig. 4 ) and site I ( Figs. 5 and 6). Significant production may occur within the sub-ice water column and/or this species may be advected laterally from areas of open water. At site I in Granite Harbor, modification of the floral assemblage from the sea ice, through the water column, to the sea floor can be followed for two sediment trapping intervals, 11/3-11/24 1984 (Fig. 5) and 11/24-12/171984 (Fig. 6). Surface sediment data is from NZ8113. Comparison of early versus late season data illustrates that although similar processes operate throughout the summer season, variability is present. Three species of diatoms dominate the sea ice A .o

B

CUMULATIVE % .

2 0

4o

.

so

,

so,

CUMULATIVE 4,/m 2 / d a y x 10 3

ioo

0

o

100

lOJ

200,

~a. 20(

300'

30(

400'

401

500'

50C

0

2000

4000

6000

8000

10000

600, :4,943 70(

700"

SURFACE SEDIMENT

~mll)l)(lllillIiH HllHi)IIil)ll OTHER

[] Amphlp. . . . . p.

[] ~tzsctliastellata

[] Th,la,,iosirasp.

[]

[]

[]

Pleurosigma sp.

Pinnulariaqu{tdratarea

[] Other

Nitzschia curta

Fig. 5.A. Cumulative percentage of dominantspecies of diatoms fbr sea ice, sediment trap and surface sediment samples from site I, Granite Harbor. The sea ice data is from 11/8 1984. The sediment traps were deployed from 11/8-11/24 1984. Surface

sediment data is from NZARP core 81-13. B. Absolute cumulative flux of diatoms, in number of individuals settling through the water column m -2 day-1, for Site I. The sediment trap data was collected over the period 11/8-11/24 1984.

57

A 0 ~

CUMULATIVE % B 20 40 60 80 100 ~"~[~[I~U~II~I[~I'I!'!]II~I~I!!!,L~[ ' I SEA ICE 0 0 ~-

CUMULATIVE 4 , / m 2 / d a y

2000

4000

6000

8000

1o000

x 10 3 12.ooo

14,000

16,000

18,000

"1" I"i

,,=, C~

7oo~ ~.~

'

OTHER

SURFACESEDIMENT

N A ............

[ ] ............. ,a,a

[]P

[ ] P i n n u l a r i a quadratarea

leurosigma sp.

Fig. 6.A. Cumulative percentage of dominant species of diatoms for sea ice, sediment trap and surface sediment samples from site I, Granite Harbor. The sea ice data is from 12/17 1984. The sediment traps were deployed from 11/24-12/17 1984. Surface sediment data is from NZARP core 81-13. B. Absolute cumulative flux of diatoms, in number of individuals settling through the water column m -2 day 1, for Site L The sediment trap data was collected over the period 11/24-12/17 1984. abundances of these species remain constant only until 313 m, and then decrease steadily below the 313 m trap. At 685 m, the cumulative diatom flux increases by a factor of five, with absolute flux of all species increasing. Later in the season at site I (11/24-12/17 1984; Fig. 6.A,B ), the relatively simple patterns observed earlier are overprinted by additional events. The water column again can be divided into four discrete units, which roughly correspond to those described from earlier in the season. Total flux decreases from 34 to 220 m, reflecting dissolution. All species are affected approximately equally. Total flux begins to increase slightly below the 220 m trap, and more distinctly below the 313 m trap. Compositionally this lower water column assemblage is very different from that observed earlier in the season. Instead of increases in only Thalassiosira spp., and Nitzschia curta, absolute flux of all species increases. A decrease in absolute flux is observed at 592 m. As earlier in the season, the largest increase in total flux occurs just above the sea floor.

Discussion Flux data The average vertical flux of diatoms, from October-December, is 10~-10 v individual frustules m -e day -1. The cumulative flux for this three month period is thus approximately 107-109 individuals m -2. Although the sub-ice diatom flux cannot be compared directly with the concentration of diatoms in the sea ice, the two can be related. McGrath-Grossi et al. (1984) have found the concentration of diatoms in the sea ice of M c M u r d o Sound to range between 10s-109 individuals m -z, about the same order of magnitude as the total cumulative diatom flux from October-December, as measured in the sediment traps. This rough comparison alone implies that when the sea ice melts in J a n u a r y - F e b r u a r y , diatom and opal fluxes must significantly increase, as a result of the relatively rapid release of biogenic components previously trapped in the ice. In addition, ice edge blooms (E1-Sayed, 1971; Alexander and Niebauer, 1981; E1-Sayed and Taguchi, 1981;

58 Stifling, 1982; Smith and Nelson, 1985) may provide another source for an increased flux of particulates while the sea ice is melting. Wilson and Smith (1984), for example, found 109-101° cells m -3 in an ice edge bloom in the Ross Sea. Strong temporal differences are undoubtedly present in levels of particulate flux. Seasonal increase in sub-ice primary productivity and/or increased release of diatoms from the sea ice may account for the progressive increase in diatom flux observed at sites D, I and L (Fig. 2 ). Several studies have documented a logarithmic increase in the number of diatom cells in the bottom few cm of the sea ice during the austral spring (Bunt and Lee, 1970; Hsiao, 1980; McGrath-Grossi et al., 1984). Increased primary production is stimulated by increased levels of light at this time ( McGrath-Grossi et al., 1984). Current velocities also influence particulate flux. Increased near-bottom diatom flux (Fig. 2 ) and silica flux (Fig. 3 ) suggest either bottom resuspension and/or lateral advection of particulate material at depth, both resulting in the production of a near-bottom nepheloid layer. The presence of a bottom nephetoid layer associated with a concurrent increase in the near bottom concentration of particulate silica in the southern Pacific Ocean has also been noted by Nelson and Gordon (1982) and by Jacobs et al. (1970) in the Ross Sea, tn the deep water traps at sites E, F, and I (Fig. 2), the early season near-bottom increase in diatom flux exceeds that occurring later in the season, Data from year-long current meter moorings near the Ross Ice Shelf indicate maximum current velocities in the mid and deep water column generally occur during the austral winter, July-August. Velocities then decrease through the spring and summer with m i n i m u m velocities in December-January (Pillsbury and Jacobs, 1985). High near-bottom diatom fluxes during the early austral spring may reflect resuspension events linked to greater current velocities at that time. Comparison of the average sub-ice diatom

flux in McMurdo Sound to diatom fluxes in other regions of the world is hindered by the scarcity of similar data. Takahashi (1986a,b), however, has assessed the flux of diatoms over biweekly intervals for two years, at a sediment trap site in the open waters of the subarctic Pacific (Station PAPA, sediment trap depths of 1000 and 3800 m). Maximum fluxes of approximately 25 × 108 individuals m-2 day occur during October, and minimum fluxes of' about 0.5X10 s individuals m -2 day -~ occur during January through March. These data are of similar magnitude to the McMurdo fluxes of 105-107 individuals m -2 day --1, although assemblages are totally different from the pres ent study. Spatially, diatom fluxes range over three orders of magnitude, with highest fluxes observed in Granite Harbor (I) and off Hut Point Peninsula. Lowest fluxes occur at sites E and F, in New Harbor, in southeastern McMurdo Sound. Previous workers have characterized the western Sound as oligotrophic and the eastern Sound as eutrophic in terms of biomass density (Bunt, 1964; Damon and Oliver. 1977; Holm-Hansen et al., 1977; F u h r m a n and Azam, 1980; Hodson et al., 1981). Overall low productivity in the western Sound has been attributed to two factors: (1) more persistent presence of sea ice cover in the western Sound and (2) influx into the western Sound of oligotrophic water originating from the Ross Ice Shelf (Littlepage, 1965; Dayton and Oliver, 1977 ). Based on this data, New Harbor appears to be less productive than the two sites sampled from the eastern Sound. Granite Harbor, however, does not follow this trend, possibly because of its greater distance from the Ross Ice Shelf and its closer proximity to open water. Floral and flux data We have observed a significant decrease of diatom frustules in the upper water column. At site L for example, there is a decrease of approximately 45-80% of the diatom flux

59 between 34 and 220 m (Table II, Fig. 3). A similar upper water column decrease in diatom flux is not seen at the other sites but this may be a consequence of the mooring setup at sites D, E, and F, with a significant reduction in diatom numbers perhaps occurring above the shallowest sediment traps at those moorings. This decrease in diatom flux is tentatively attributed to dissolution of fragile diatom frustules in the upper water column. Our findings at Mooring I are consistent with previous observations from open water portions of the Southern Ocean. Kozlova (1961, 1964), based on suspended diatom counts, estimated that > 80% of the diatom frustules produced in near surface waters of the Indian and Pacific sectors of the Antarctic dissolve by 100 m. It is also possible that the flux of frustules decreases in the upper water column due to fragmentation of tests via grazing organisms. Extremely low zooplankton populations are reported, however, from the Ross Sea (Biggs et al., 1984) and similarly we have found low zooplankton fluxes in our sediment traps. Preliminary results indicate that fecal pellet fluxes are extremely low in McMurdo Sound as compared to elsewhere in the Southern Ocean (Dunbar, 1984; Dunbar et al., 1985a,b). For example, maximum pellet fluxes measured in McMurdo Sound between October-December, 1984, are only 130 pellets m -2 day -1. This is quite small as compared to pellet fluxes of 5 × 105 pellets m -2 d a y - ' measured in the Bransfield Strait (Antarctic Peninsula) ( Dunbar, 1984). Low zooplankton and pellet fluxes make it unlikely that fragmentation by grazers is principally responsible for the decrease in diatom flux in the upper water column. Nelson and Gordon (1982) estimate that as little as 18% to as much as 58% (by weight) of the biogenic silica produced in surface waters of the Pacific sector of the Southern Ocean dissolves in the upper 100 m. We find similar results based on the sediment trap data in McMurdo Sound, where opal flux decreases between 13-40% within the upper column. In

contrast, in an upwelling region off Northwest Africa few of the diatom tests escaped dissolution in the upper 60 m (Nelson and Goering, 1977). Nelson and Gordon (1982) suggest that the relatively low amount of silica dissolution in the Southern Ocean water column, as compared to lower latitude regions, may be a function of the very low water temperature, since the dissolution rate constant for silica is strongly temperature dependent (Hurd, 1972; Kamatami and Riley, 1980). In areas covered by sea ice, like McMurdo Sound, it is also likely that the presence of sea ice inhibits wind mixing and subsequent particulate resuspension in the upper water column. As a result of the shorter residence time of particulates within the upper water column, less dissolution of biogenic silica may occur. Based on the finding that absolute diatom flux (in number of frustules) decreases more significantly than biogenic silica mass flux in the upper water column (Table II), it appears that large numbers of delicate, thin-walled frustules are rapidly recycled in the upper water column. These fragile tests do not contribute to the silica mass flux as much as suggested by numbers alone. Comparison of the sea ice and shallow sediment trap samples provides insight into the effect of upper water column dissolution on the composition of the floral assemblage (Fig. 4). Two species in particular, Fragilaria islandica var. adeliae and Berkeleya sp., show a great decrease in relative abundance between the sea ice and shallow sediment trap samples. Both have quite fragile tests and apparently undergo rapid dissolution in the uppermost water column. Modification of the diatom assemblage also occurs throughout the middle part of the water column. The floral data trends shown in Figs. 5 and 6 indicate that dissolution does not appear to have a significant effect over this interval as absolute abundances of most species remain approximately constant (Fig. 5.B ) or increase

60

(Fig. 6.B ). During the early part of the season, above 313 m, the diatom assemblage appears to be composed of individuals settling vertically through the water column. These specimens have principally originated in the local sea ice, or possibly in the case of Thalassiosira spp., a species rarely found in the ice, from low levels of in-situ sub-ice primary productivity or lateral advection. The increased flux of Thalassiosira spp. and Nitzschia curta below 313 m (Fig. 6.A) cannot be accounted for by in-situ primary productivity, as this is well below the photic zone. These diatoms have been laterally advected, quite likely from regions of open water conditions, where Thalassiosira spp., in particular, are expected to be more common, Fryxell et al. (1984) and Buck et al. (1985), have found Thalassiosira gravida, a species which has not been recorded in the ice, to be a :common component of the open water assemblage found near the ice edge zone in the WeddeU ~ a . Similarly, E1-Sayed (1971) found Thalassiosira tumida in overwhelming abundance in a 1968 ice edge bloom in the Weddell Sea. Nitzschia curta, though common in the sea-ice, has also been reported from an ice edge bloom in the Ross Sea (Smith and Nelson, 1985; Wilson et al., 1986). Deep advective input and subsequent dilution of the original assemblage, rather than dissolution is responsible for the major modification of the diatom assemblage between 313 and 592 m.

Later in the season, increased absolute flux of all species occurs in the mid water depths (Fig. 6.B), as opposed to the early season increases in only Thalassiosira spp. and Nitzschia curta. It is not clear why this material is of different composition than that advected earlier in the season. This material may represent the sinking products of an extensive diatom bloom event which has partially settled through the water column but not yet reached the sea floor, as evidenced by the decreased diatom flux at 592 m. Holm-Hansen et al. (1982) have observed post bloom sinking of phyto-

plankton standing crops to depths greater than 100 m in the Ross Sea. It is also possible that this material was advected laterally, but the composition of the source material has changed from the dominantly Thalassiosira spp. and Nitzschia curta assemblage seen earlier in the season. At 685 m, the absolute diatom flux increases by a factor of 6.5 (Fig. 5.B) and 3 (Fig. 6.B), most likely the result of local resuspension from the seafloor. Resuspension of diatom frustutes in a near bottom nepheloid layer maintains tests in silica undersaturated waters and contributes to additional dissolution and modification of the assemblage. The predominance of Thatassiosira spp. and Nitzschia curta in the sediment floral assemblage develops primarily in this near-bottom portion of the water column. Modification probably occurs through a combination of two processes. First, the dissolution of those forms which dominate the sea ice and much of the water column may occur in this near bottom nepheloid layer. Edmond (1974) suggests that most dissolution of biogenic opal occurs at the sea floor rather than in the water column. Takahashi and Honjo (1981) and Takahashi {1981 ) have shown this to be the case for radiolarians from the western tropical Atlantic, the Panama Basin and the tropical central Pacific and Takahashi (1986a, b) similarly finds that most of the modification of the diatom assemblage in his subarctic Pacific study, occurs at the sea floor. Second, our sediment trap data only cover three months of the year. Sedimentation accumulation rates in the southwestern Ross Sea average 1.2 mm yr-1 (Ledford-Hoffman et al., 1986) so our surface sediment samples of the top centimeter of the sediment column represent an average of many years of accumulation. During the ice edge bloom and open water conditions occurring later in the season the composition of the water column diatom assemblage may correlate more closely to that of the surface sediment, with higher relative abundances of Nitzschia curta and Thalassiosira spp. Both

61

groups have been reported to dominate ice edge blooms (El-Sayed, 1971; Fryxell et al., 1984; Buck et al., 1985; Smith and Nelson, 1985; Wilson et al., 1986). Nelson and Smith (1986) suggest that the dominance of Nitzschia curta in surface sediments of the Ross Sea results from the quantitative importance of the ice edge bloom diatom community.

Conclusions (1) Dissoluton of large numbers of diatom frustules may occur in the upper 200 m, decreasing the number of settling diatom tests. Silica dissolution in the upper water column, however, has a much smaller effect on the opal flux, the result of dissolution of thinly silicified tests which do not contribute greatly to the total mass of silica settling through the water column. Two species which readily dissolve within the upper 25 m are Fragilaria islandica var. adeliae and Berkeleya sp. Other species common in the sea ice do not appear to be disproportionately dissolved in the upper water column. (2) Lateral advection plays a large role in the modification of the water column diatom assemblages in Granite Harbor. Relative abundances of two species in particular, Thalassiosira spp. and Nitzschia curta, increase at midwater depths, signaling their transport from other areas. ( 3 ) The most significant site of modification of the diatom assemblage appears to be within a near-bottom nepheloid layer, where diatoms are either advected laterally and/or resuspended off the sea floor. A combination of preferential dissolution of those forms which dominate the sea ice (Amphiprora sp., Pleurosigma sp., Nitzschia steUata, and Pinnularia quadratarea) and dilution of the assemblage described with resuspended and/or advected diatoms representative of many year averages of the assemblage, may be responsible for the near bottom alteration of the flora. These two processes work together to produce an assemblage primarily made up of Thalassiosira spp.,

Nitzschia curta, and other robust forms of Nitzschia. (4) It is apparent that despite the presence of a highly productive sea ice microbial community within McMurdo Sound, a relatively low percentage of these diatoms is represented in the surface sedimentary record. The extremely low relative abundance of the common sea ice diatoms in the surface sediments of McMurdo Sound has important paleoceanographic implications. Many studies have attempted to correlate microfossil assemblages of Southern ocean surface sediments with oceanographic parameters of the overlying water column (Truesdale and Kellogg, 1979; DeFelice and Wise, 1981; for example). In this case, any attempt to correlate surface sediment floral composition to oceanographic parameters would be hindered by the extremely poor sedimentary representation of the dominant sea ice diatoms.

Acknowledgements We thank Richard Marty and many members of I T T / A N S for field assistance. Greta Fryxell and Sarah McGrath-Grossi aided in diatom identification. Connie Sancetta and Kozo Takahashi provided critical reviews of this manuscript. This work was supported by National Science Foundation Grant DPP-8312486 and a Sigma Xi grant-in-aid of research.

References Ackley, S.F., Taguchi, S. and Buck, K.R., 1978. Primary productivity in sea ice of the Weddell region. U.S. Army Cold Region Res. Eng. Lab. Tech. Rep., 78-19, 17 pp. Alexander, V. and Niebauer, H.J., 1981. Oceanography of the eastern Bering Sea ice-edge zone in spring. Limnol. Oceanogr., 26(6): 1111-1125. Barrett, P.J., Pyne, A.R. and Ward, B.L., 1983. Modern sedimentation in McMurdo Sound, Antarctica. In: R.L. Oliver, P.R. James and J.B. Jago (Editors), Antarctic Earth Science. Australian Academy of Sciences, Canberra, pp. 550-554. Barry, J.P. and Dayton, P.K., 1985. Under-ice oceanography in McMurdo Sound, Antarctica. EOS. Trans. Am. Geophys. Union, 66 (51) : 1286.

62 Biggs, D.C., Amos, A.F. and Holm-Hansen, 0., 1985. Oceanographic studies of epipelagic ammonium distributions: the Ross Sea NH~ flux experiment: In: W.R. Siegfried, P.R. Condy, R.M. Laws (Editors), Antarctic Nutrient Cycles and Food Webs. Springer, Berlin, pp. 93-103. Buck, K., Garrison, D.L. and Fryxell, G.A, 1985. Algal assemblages in sea ice and in the ice-edge zone of the Weddell Sea: An AMERIEZ study. EOS Trans. Am. Geophys. Union, 66 ( 51 ): 1287. Bunt, JS., 1963. Diatoms of Antarctic sea-ice as agents of primary production. Nature, 199: 1255-1257. Bunt, J.S., 1964. Primary productivity of undersea ice in Antarctic waters. 2. Influence of light and other factors on photosynthetic activities of Antarctic marine microalgae. Antarct. Res. Ser., 1: 27-31. Bunt, J.S. and Lee, C.C., 1970. Seasonal primary production in the Antarctic sea ice at McMurdo Sound in 1967. J. Mar. Res., 28: 304-320. Burckle, L.H., Robinson, D. and Cooke, D.W., 1982. Reappraisal of sea-ice distribution in Atlantic and Pacific sectors of the Southern Ocean at 18,000 yrs B.P. Nature, 299: 435-437. Burkholder, P,R. and Mandelli, G.E., 1965. Productivity of microalgae in Antarctic sea ice. Science, 149: 872-874. Dayton, P.K and Oliver, J.S., 1977. Antarctic soft bottom benthos in oligotrophic and eutrophic environments. Science, 148: 872-874. DeFelice, D. and Wise, S., 1981. Surface lithofacies, biofacies and diatom diversity patterns as models for delineation of climatic change in the southeast Atlantic Ocean. Mar. Micropaleontol., 6: 29-70. DeMaster, D.J,, 1979. The marine budgets of silica and Si32. Thesis. Yale University, New Haven, Conn. DeMaster, D.J., 1981. The supply and accumulation of silica in the marine environment. Geochim. Cosmochim. Acta, 45: 1715-1732. Dunbar, R.B., 1984. Sediment trap experiments in the Antarcticcontinental margin. Antarct. J. U.S., 19 (5): 70-71. Dunbar, R.B., Dehn, M. and Leventer, A., 1984. Distribution of biogenic components in surface sediments from the Antarctic continental shelf.Antarct. J. U.S., 19 (5) : 71-73. Dunbar, R.B., MacPherson, A.J. and Wefer,G. 1985a. Water column particulate flux and seafloor deposits in the Bransfield Strait and southern Ross Sea, Antarctica. Antarct. J. U.S., 20(5): 98-100. Dunbar, R.B., Leventer, A.R. and Marty, R.C., 1985b. Vertical sediment flux beneath annual sea ice, McMurdo Sound, Antarctica. Anterct. J. U.S., 20(5): 109-110. Dymond, J., Fisher, K., Clauson, M., Cobler, R., Gardner, W., Richardson, M.J., Berger, W., Soutar, A. and Dunbar, R., 1981. A sediment trap intercomparison study in the Santa Barbara Basin. Earth. Planet. Sci. Lett., 53: 409-418. Edmond, J.M, 1974. On the dissolution of carbonate and

silicate in the deep ocean. Deep-Sea Res., 21: 455-480. E1-Sayed, S.Z., 1971. Observations on phymplankton bloom in the Weddetl Sea. In: G.A. Llano and E. Wallen (Editom). Biology of the Antarctic Seas IV. Antarct. Res. Ser., 17: 301-312. E1-Sayed, S.Z. and Taguchi, S., 1981. Primary production and standing crop of phytoplankton along the ice-edge in the Weddell Sea. Deep-Sea IRes., 28A: 1017-1032. Fenner, J., Schrader, H.J. and Wienigk, H.. 1976. Diatom phytoplankton studies in the southern Pacific Ocean. composition and correlation to the Antarctic convergence and its paleoecological significance. In: Initial Reports of the Deep Sea Drilling Project. 35. U.S. Government Printing Office, Washington, D.C., pp. 757-813. Fryxell, G.A., Theriot, E.C. and Buck, K.R., 1984. Phy~oplankton, ice algae, and choanoflagellates from AMERIEZ, the southern Atlantic and Indian Oceans. Antarct. J. U.S., 19(5): 107-109. Fuhrman, J.A. and Azam, F., 1980. Bacterioptankton secondary production estimates for coastal waters of British Columbia, Antarctica, and California. Appl. Environ. Micro, 39: 1085-1095. Fukushima, H. and Meguro, H.. 1966. The plankton ice as a basic factor of the primary productivity of the Antarctic ocean. Antarct. Rec., 27: 99-109. Garrison, D.L. and Buck, K.R., 1985. Sea-ice algal communities in the Weddell Sea: species composition in ice and plankton assemblages, In: J.S. Gray and M.E. Christiansen (Editors), Marine Biology of Polar Regions and Effects of Stress on Marine Organisms. Wiley, New York, N.Y., pp. 103-122. Gombos, A., 1976. Paleogene and Neogene diatoms from the Falkland Plateau and Malvinas Outer Basin. In: Initial Reports of the Deep Sea Drilling Project, 36. U.S. Government Printing Office, Washington. D.C.. pp. 575-695. Heath, A.R., 1977. Circulation across the ice edge shelf in McMurdo Sound, Antarctica. In: M.J. Dunbar (Editor), Polar Oceans, Proc. Polar Oceans Conf., McGill Univ., Montreal, pp. 129-139. Heath, G.R., 1974. Dissolved silica and deep-sea sediments. In: W.W. Hay (Editor), Studies in Paleo-oceanography, Soc. Econ. Paleontol. Mineral. Spec. Publ., 20: 77-93. Hodson, R.E., Azam, R., Carlucci, A.F., Fuhrman, J.A., Karl, D.M. and Holm-Hansen, O., 1981. Microbial uptake of dissolved organic matter in McMurdo Sound, Antarctica. Mar. Biol., 61: 89-94. Holm-Haneen, 0., Et-Sayed, S,A., Franceschini, B.A. and Cuhel, R., 1977. Primary production and the factors controlling phytopiankton growth in the Antarctic seas. In: G.A. Llano (Editor), Adaptations within Antarctic Ecosystems. Gulf, Houston, Tex., pp. 11-50. Holm-Hansen, 0., Neori, A. and Koike, I., 1982. Phytoplankton distribution, biomass, and activity in the southwest Ross Sea. Antarct. J. U.S., 17 (5) : 150-152.

63 Horner, R.A., 1976. Sea ice organisms. Oceanogr. Mar. Biol. Annu. Rev., 14: 167-182. Hoshiai, T. and Kato, M., 1961. Ecological notes on the diatom community in the sea ice in Antarctica. Bull. Biol. Stn. Asamushi, 10: 221-230. Hsiao, S.I.C., 1980. Quantitative composition, distribution, community structure and standing stock of sea ice microalgae in the Canadian Arctic. Arctic, 33: 768-793. Hurd, D.C., 1972. Factors affecting solution rate of biogenic opal in seawater. Earth Planet. Sci. Lett., 15: 411-417. Jacobs, S.S., Amos, A.F. and Bruchhausen, P.M, 1970. Ross Sea oceanography and Antarctic Bottom Water formation. Deep-Sea Res., 17: 935-962. Jacobs, S.S., Fairbanks, R.G. and Horibe, Y., 1985. Origin and evolution of water masses near the Antarctic continental margin: Evidence from H21SO/H2160 ratios in seawater. In: S.S. Jacobs (Editor), Oceanology of the Antarctic Continental Shell Antarct. Res. Ser. Am. Geophys. Union, 43: 59-85. Joust, A.P., Koroleva, G.S. and Nagaeva, G.A., 1963. Stratigraphic and paleogeographic research in the Indian section of the Antarctic. Okeanol Issled., 8: 137-160. Kamatani, A. and Riley, J.P., 1980. Rate of dissolution of diatom silica walls in water. Mar. Biol., 55: 29-35. Kellogg, T.B. and Truesdale, R., 1979. Late Quaternary paleoecology and paleoclimatology of the Ross Sea: the diatom record. Mar. Micropaleontol., 4: 137-158. Kozlova, O.G., 1961. Quantitative content of diatoms in the waters of the Indian sector of the Antarctic. Akad. Nauk. S.S.S.R. Dokl., 138: 207-210. Kozlova, O.G., 1964. Diatoms of the Indian and Pacific sectors of the Antarctic. Akad. Nauk U.S.S.R.., Moscow, 191 pp. (translated by the Crail Program for scientific translations, U.S. Dep. Commer. C.F.S.T.I.). Krebs, W.N., 1977. Ecology and preservation of neritic diatoms, Arthur Harbor, Antarctica. Thesis, Univ. California, Davis, 216 pp. Ledford-Hoffman, P.A., DeMaster, D.J. and Nittrouer, C.A., 1986. Biogenic-silica accumulation in the Ross Sea and the importance of Antarctic continental shelf deposits in the marine silica budget. Geochim. Cosmochim. Acta, 50: 2099-2110. Lewis, E.L. and Perkin, R.G., 1985. The winter oceanography of McMurdo Sound, Antarctica. In: S.S. Jacobs (Editor), Oceanology of the Antarctic Shelf. Antarct. Res. Ser. Am. Geophys. Union, 43: 145-165. Lisitzin, A.P., 1972. Sedimentation in the World Ocean. Soc. Econ. Paleontol. Mineral. Spec. Publ., 17:218 pp. Littlepage, J.L., 1965. Oceanographic investigations in McMurdo Sound, Antarctica. Antarct. Res. Ser., 5: 1-37. McCollum, D.W., 1975. Diatom stratigraphy of the Southern Ocean. In: Initial Reports of the Deep Sea Drilling Project, 28. U.S. Government Printing Office, Washington, D.C., pp. 515-571. McConville, M.C. and Weatherbee, R., 1983. The bottom-

ice microalgal community from annual ice in the inshore waters of East Antarctica. J. Phycol., 19: 431-439. McGrath-Grossi, S.M., 1985. Response of a sea-ice microalgal community to a gradient in under-ice irradiance. Thesis. Univ. Southern California, Los Angeles, Calif., 324 pp. McGrath-Grossi, S. and Sullivan, C.W., 1985. Sea ice microbial communities. V. The vertical zonation of diatoms in and Antarctic fast ice community. J. Phycol., 21 (3) : 401-409. McGrath-Grossi, S.M., Kottmeier, S. and Sullivan, C.W., 1984. Sea ice microbial communities. III. Seasonal abundance of microalgae and associated bacteria, McMurdo Sound, Antarctica. Microb. Ecol., 10: 231-242. McGrath-Grossi, S., Kottmeier, S.T., Moe, R.L., Taylor, G.T. and Sullivan, C.W., 1987. Sea-ice microbial communities. VI. Growth and primary production in bottom ice under graded snow cover. Mar. Ecol., in press. Meguro, H., 1962. Plankton ice in the Antarctic ocean. Antarct. Rec., 14: 1192-1199. Mitchell, W.M. and Bye, J.A.T., 1985. Observations in the boundary layer under the sea ice in McMurdo Sound. In: S.S. Jacobs (Editor), Oceanology of the Antarctic Continental Shelf. Antarct. Res. Ser. Am. Geophys. Union, 43: 167-176. Nelson, D.M. and Goering, J.J., 1977. Near-surface silica dissolution in the upwelling region off nortwest Africa. Deep-Sea Res., 24: 65-73. Nelson, D.M. and Gordon, L.I., 1982. Production and pelagic dissolution of biogenic silica in the Southern Ocean. Geochim. Cosmochim. Acta, 46: 491-501. Nelson, D.M. and Smith, W.O., 1986. Phytoplankton bloom dynamics of the western Ross Sea ice edge. II. Mesoscale cycling of nitrogen and silicon. Deep-Sea Res., 33: 1389-1412. Pillsbury, R.D. and Jacobs, S.S., 1985. Preliminary observations from long-term current meter moorings near the Ross Ice Shelf, Antarctica. In: S.S. Jacobs (Editor), Oceanology of the Antarctic Continental Shell Antarct. Res. Ser. Am. Geopys. Union, 43: 87-107. Pyne, A.R., Ward, B.L., Macpherson, A.J. and Barrett, P.J., 1985. McMurdo Sound Bathymetry. (Misc. Ser.) New Zealand Oceanographic Institute, Wellington, 1st Ed. Ryther, J.H., 1963. Geographic variations in productivity. In: M.N. Hill (Editor), The Sea, vol. 2. Wiley, London, pp. 347-380. Schrader, H.J., 1976. Cenozoic planktonic diatom biostratigraphy of the Southern Pacific Ocean. In: Initial Reports of the Deep Sea Drilling Project, 35. U.S. Government Printing Office, Washington, D.C., pp. 605-671. Smith, W.O. and Nelson, D., 1985. Phytoplankton bloom produced by a receding ice edge in the Ross Sea: Spatial coherence with the density field. Science, 227: 163-166. Stirling, I., 1982. The biological importance of ice edges. EOS. Trans. Am. Geopys. Union, 63: 46-47.

64 Sullivan, C.W., Palmisano, A,C., Kottmeier, S., Grossi, S.M., Moe, R. and Taylor, G.T., 1983. The influence of light on development and growth of sea ice microbial communities (SIMCO) in McMurdo Sound, Antarctica. Antarct. J. U.S., 18: 177-179. Sullivan, C.W., Palmisano, A.C., SooHoo, Beeler, J., 1984. Influence of sea ice and sea ice biota on downweUing irradiance and special composition of light in McMurdo Sound, In: M. Blizard (Editor), Ocean Optics VII. Proc. Int. Soc. Opt. Eng., 489: 159-165. Takahashi, K., 1981. Vertical flux, ecology and dissolution of Radiolaria in tropical oceans: Implications for the silica cycle. Thesis. M.I.T.W.H.O,I., Woods Hole, Mass., 81-103, pp. 461. Takahashi, K., 1986a. Silicoflagellates andActiniscus: vertical fluxes at Pacific and Atlantic sediment trap stations. In: S. Honjo (Editor), Ocean Biocoenosis. Micropaleontology Press, New York, N.Y. Takahashi, K., 1986b. Seasonal fluxes of pelagic diatoms in the subarctic Pacific, 1982-1983. Deep-Sea Res., 33: 1225-1251. Takahashi, K., and Honjo, S., 1981. Vertical flux of radiolaria: A taxon-quantitative sediment trap study from

the western tropical Atlantic. Micropaleontology, 27 (2) : 140-190. Thomas, C.W., 1963. On the transfer of visible radiation through sea ice and snow. J. Glaciol., 4: 481-484. Tressler, W.L. and Ommundse~. A.M, 1962. Seasonal oceanographic studies in McMurdo Sound, Antarctica. U.S. Navy Hydrographic Office. Washington, D.C. Tech. Rep., TR-125, pp. 1-141. Truesdale, R.S. and Kellogg, T.. 1979. Ross Sea diatoms: Modern assemblage distributions and their relationship to ecologic, oceanographic, and sedimentary conditions Mar. Micropaleontol., 4: 13-31. Wilson, D.L. and Smith, W.D., 1984. Species specific productivity in an ice-edge phytoplankton bloom in the Ross Sea. Antarct. J. U.S., 19(5): 132-133. Wilson, D.L., Smith, W.D. and Nelson. D.M., 1986. Phytoplankton bloom dynamics of the western Ross Sea ice edge. L Primary productivity and species-specific production. Deep-Sea Res.. 33: 1375-1387. Zwally, H.C.. Parkinson, C., Carsey, F., Gloersen. P.. Campbell, W.J. and Ramseier. R.O., 1979. Antarctic sea ice variations 1973-1975. In: NASA Weather Clim. Rev.. Pap., 56: 335-340.