Can selective export processes be explained in a regional context?

Stations M2 and M6 were intended to be representative of the HNLC areas within the vicinity of the Crozet Plateau. The exported particles at M2 and M6 shared very similar Si:C molar ratios (Fig. 3.6) consistent with the similarities between these two control sites demonstrated by other properties. The diatom cells exported at M6 were

dominated by F. kerguelensis, 66% of exported diatom frustules. Exported particles at

M2 also contained F. kerguelensis frustules, but to a smaller extent (34%), and were

mainly dominated by T. nitzschoides (55% of the exported diatom frustules). In general,

the particles exported south of the plateau appear to be enriched in F. kerguelensis

relative to the surface community which at M2 made up 14%, and 13% at M6 of the surface community (Table 3.6).

The dominance of T. nitzschioides, Pseudo-nitzschia spp. together with Chaetoceros species and F. kerguelensis in the surface (5m) samples is consistent with other studies of summer diatom floras in the region of the Polar Front in the Indian Sector of the

Southern Ocean (Kopczynska et al., 1998). The almost complete absence of Pseudo-

nitzschia spp. from the PELAGRA samples is not surprising. This delicate and lightly silicified diatom rarely contributes to flux due to dissolution and/ or grazing (Abrantes and Moita, 1999; Lange et al., 1994), and probably also low aggregation and sinking rates of smaller diatoms, compared to larger species. Thus the increase in relative abundance of T. nitzschioides is to be expected although the greater proportional

increase in F. kerguelensis may result from enhanced abundances of this species at

depths greater than 5m in the surface mixed layer. During the SOIREE experiment aggregates captured by sediment traps after iron enrichment were enriched in F.

kerguelensis relative to measured water column abundances (Waite and Nodder, 2001). The large diatoms C. criophilium and D. antarcticus that were present in the surface

community (Poulton et al., 2007) were not found in PELAGRA samples south of the

plateau. Based on this we must conclude that these organisms were not dominant in export processes at this locality, however previous work in the Southern Ocean has

demonstrated that C. criophilum can play a significant role in bloom sedimentation

(Crawford, 1995).

The species composition of the export assemblage was different at M8E and M8W; cell counts (Table 3.3) revealed that 91% of the diatom cells exported at M8W was of a single species, E. antarctica, with smaller contributions (8%) from F. kerguelensis, D. antarcticus and C. criophilium at M8W. The community structure of the diatom

population at 10m (Table 3.6) showed that E. antarctica makes up <1% of diatom cells

and F. kerguelensis ~4% (see also Poulton et al., 2007). This result demonstrates that there is selective export of not only one functional group, i.e. diatoms, but just one or two key species within that group. The architecture of heavily silicified diatom shells offer mechanical protection (Hamm et al., 2003) against grazing thus enriching the export of these species and facilitating their accumulation in Southern Ocean diatom oozes (Verity and Smetacek, 1996).

The situation at M8E was somewhat different; T. nitzschoides accounted for 80% of the

exported diatom cells, and E. antarctica for only 16% (Table 3.3). However, the surface

community structure at the time of sampling contained fewer E. antarctica cells (Table

3.6), and thus the export was again enriched with E. antarctica. Moore et al. (2007)

reported that E. antarctica responded most favourably to Fe-addition experiments, and

results from the sediment trap shows it is a major exporter of both POC and biogenic silica from iron-fertilised regions, implying that it plays a key role in the biogeochemical cycling of POC and silica in this study area. Previous laboratory experiments and

mesoscale enrichment experiments in the Southern Ocean (e.g. Gall et al, 2001a) have

Table 3.6 Percentage contributions of individual diatom species to the surface assemblage (5m) from each station calculated from absolute cell abundance. Number in brackets is the proportion of the species’ frustules that were observed to be empty, no

brackets indicates 100% of frustules contained cellular material.

Station M2 M6 M8E M8W M3.3 M3.7 M3.8

Centric diatoms

Small Chaetoceros sp. 0.65 - - - - 6.61 16.35

Chaetoceros "peruvianum" - 0.84 - 1.52 0.17 - -

Chaetoceros 'mitra' / debilis 1.18 2.35 - 24.24 3.61 - -

Chaetoceros curvitus - 2.01 - - - - -

Chaetoceros debilis - - - -

Chaetoceros resting spores - - - -

Centric 20 um 11.19 (32.55) 12.39 (4.05) 1.12 (16.67) 2.12 (21.43) 3.66 (21.88) 6.93 (1.16) 21.61 (0.54) Centric 40 um - 1.68 (0) - - - - - Corethron crophilium 2.09 1.01 (20.0) - 1.21 0.74 (18.18) - - Dactyliosolen antarcticus 0.92 1.34 - - 0.11 - - Eucampia antarctica - - - 0.91 (83.33) 0.46 (25.0) - - Guinardia spp. 1.18 0.34 - 3.18 0.40 0.04 - Leptocylindricus spp. 1.31 0.84 - 1.82 0.11 - - Rhizolsolenia spp. - 0.17 - - - - - Proboscia alata - - - 0.15 - - - Pennate diatoms Cylindrotheca closterium 2.35 1.01 - 5.30 1.26 0.14 0.23 "Thalassionema nitzschoides" 35.42 (38.40) 44.89 (39.93) 93.68 (14.09) 24.70 (24.54) 51.40 (23.27) 60.35 (9.89) 47.65 (34.56) Fragilariopsis kerquelensis 14.51 13.07 0.74 3.94 4.12 0.48 (12.50) 1.40 (20.83) Thalassiothrix antarctica - - - - 0.23 - - Pseudo-nitzschia spp. 28.50 (0.92) 16.42 4.46 29.85 33.60 22.42 5.85 Small pennates - 1.17 - - - 3.02 6.89 Manginea spp (M. rigida) - - - 0.15 0.06 - 0.01 Haslea trompii 0.65 0.50 - 0.91 0.06 - -

F. kerguelensis. The importance of E. antarctica in the CROZEX study area may be related to the formation of winter resting spores that are advected from the shallow topography of the Crozet Plateau, along with dissolved Fe, northwards into the bloom area. Subsequently, when light limitation is relieved, E. antarctica may possess an ecological advantage over other heavily silicified diatoms. Significantly the presence of E. antarctica, usually associated with ice marginal zones has also been documented near other Southern Ocean Islands in the vicinity of the Polar Front (e.g. (Froneman et al., 1997).

Vertically integrated production at M3.7 and M3.8 (Fig 3.4) was an order of magnitude higher than all the other stations, the bloom probably being at its peak with a small vertical loss term. T. nitzschoides was the diatom species dominating the export, accounting for 92% of the diatom export assemblage at M3.7, relative to 54% in the surface community, and 97% at M3.8, relative to 48% in the surface. At these stations there did not seem to be the selective export of large diatoms and this is associated with very small calculated export ratios(100m) of 1%. At station M3.7, ~10% of the T.

nitzschoides cells in the surface were present as empty frustules, whereas at station M3.8 this proportion had increased to ~35%. This suggests that the small vertical loss term measured at M3.7 and M3.8 is probably dictated by individual cell deaths and slow gravitational settling potentially caused by self-shading or low cell concentrations (Jackson, 1990; Kiorboe et al., 1990) rather than the formation of marine aggregates. North of the plateau at M8E and M8W there seems to be selective export of large diatom species, whereas south of the plateau the large diatoms present in the water column were not exported. These phytoplankton community differences between north and south of the plateau are likely to be reflected in the characteristics of export flux with south HNLC waters dominated by small picophytoplankton and pennate diatoms due to the iron-limited conditions, resulting in lower export rates. Fv/Fm values, an indication of

physiological stress, were much lower at stations M2 and M6 (Moore et al, 2007) relative

to the other stations, suggesting that the cells were suffering from nutrient limitation,

significant effect on the magnitude of export from a diatom bloom, as observed by

comparing stations M8E and M8W north of the plateau. Phaeocystis sp. cell abundance

at M8W was an order of magnitude higher than M8E, 1.2 x 106 _{and 2.4 x 10}5 _{cells l}-1_,

respectively. There is also a strong linear correlation (r2 _{0.74, n = 7, P<0.05) between}

Phaeocystis sp. cell abundance and the export ratio(100m) suggesting it plays a role in

regulating export, possibly through the production of transparent exopolymer particles

(TEP) (Mari et al., 2005), which have been demonstrated to be important mediators for

the formation of large marine aggregates (Passow et al., 2001; Passow, 2002). Whilst

previous studies (Wassman, 1994; DiTullio et al, 2000) have documented the direct involvement of Phaeocystis sp in rapid flux events, possibly related to iron supply, our sediment trap observations do not suggest this to be the case around the Crozet Plateau.

It remains unclear why we detected a small contribution of Phaeocystis cells in the M8E

deployment compared to zero at M8W, despite the increased cell abundance at the surface. Based on a strong power-law correlation (r2 _{0.74) between Phaeocystis sp. and}

Si(OH)4 concentrations, it is suggested the observed correlation between export ratio(100m)

and Phaeocystis sp. cell abundance is an artefact caused by Si depletion, which induces export, and the subsequent community shift to Phaeocystis sp., rather than the organism having a direct role in the export process.