Tidal influence on particulate organic carbon export fluxes around a tall seamount

As tall seamounts may be ‘stepping stones’ for dispersion and migration of deep open ocean fauna, an improved understanding of the productivity at and food supply to such systems needs to be formed. Here, the 234Th/238U approach for tracing settling particulate matter was applied to Senghor Seamount – a tall sub-marine mountain near the tropical Cape Verde archipelago – in order to elucidate the effects of topographically-influenced physical flow regimes on the export flux of particulate organic carbon (POC) from the near-surface (topmost ⩽ 100 m) into deeper waters. The comparison of a suitable reference site and the seamount sites revealed that POC export at the seamount sites was ∼2–4 times higher than at the reference site. For three out of five seamount sites, the calculated POC export fluxes are likely to be underestimates. If this is taken into account, it can be concluded that POC export fluxes increase while the passing waters are advected around and over the seamount, with the highest export fluxes occurring on the downstream side of the seamount. This supports the view that biogeochemical and biological effects of tall seamounts in surface-ocean waters might be strongest at some downstream distance from, rather than centred around, the seamount summit. Based on measured (vessel-mounted ADCP) and modelled (regional flow field: AVISO; internal tides at Senghor: MITgcm) flow dynamics, it is proposed that tidally generated internal waves result in a ‘screen’ of increased rates of energy dissipation that runs across the seamount and leads to a combination of two factors that caused the increased POC export above the seamount: (1) sudden increased upward transport of nutrients into the euphotic zone, driving brief pulses of primary production of new particulate matter, followed by the particles’ export into deeper waters; and (2) pulses of increased shear-driven aggregation of smaller, slower-settling into larger, faster-settling particles. This study shows that, under certain conditions, there can be an effect of a tall seamount on aspects of surface-ocean biogeochemistry, with tidal dynamics playing a prominent role. It is speculated that these effects can control the spatiotemporal distribution of magnitude and nutritional quality of the flux of food particles to the benthic and benthic-pelagic communities at and near tall seamounts.


305
Before results for 234 Th, POC, PN and export can be discussed the physical-oceanographic context 306 needs to be scrutinised to clarify which fluid-dynamic features are of most relevance to this study. This 307 physical-oceanographic context will, therefore, be developed in Section 3.1. The results for 234 Th, for POC 308 and PN, and for POC export will then be presented and discussed in Section 3.2, 3.3 and 3.4, respectively.  on the processes at the seamount (Fig. 2c,d). VM-ADCP data show that the eddies reached no deeper than 349~ 150-200 m (Fig. 5,6).

350
During this study NPP was consistently low near and south of the seamount and indicates tropical non-351 bloom conditions . It also appears that S-Ref has been lying in the same type of low-NPP 352 waters that then turned back onto and impinged on the seamount from the NE. By contrast, the wider

367
If there was a weak Taylor column, then its vertical extent was limited to waters deeper than ~ 250 m.

368
This limitation may be due to the strong mass-density stratification in the topmost ~ 100 m. Composite towards the summit in the upper ~ 150 m of the water column (see Fig. 9). Doming was also observed by

392
In contrast to the near-inertial oscillations tidal oscillations are continuous and therefore probably 393 more important for nutrient redistribution and/or particle dynamics. Near Senghor in the far field, the 394 TPXO barotropic-tide model of Egbert and Erofeeva (2002) predicts the ellipse of the current vector of the 395 predominantly semidiurnal barotropic tide to be strongly elongated in the NNW and SSE direction (Fig. 1a).

422
The MITgcm run also predicts that, in addition to beam formation, tidally oscillating flow that is forced 423 over the summit plateau generates soliton-like internal waves that propagate horizontally in the topmost 424~ 100 m and north-and southwards away from the seamount. This leads to increased maximum horizontal 425 and vertical flow velocities (Fig. S3) and increased rates of kinetic energy dissipation (Fig. 4) around the 426 summit plateau. Real-world evidence for these waves was also found in the VM-ADCP data: internal waves 427 occurred in the topmost ~ 100 m, have a period of ~ 0.5 hr and appear to propagate predominantly 428 horizontally (they can be seen in Fig. 7, but are more noticeable in the example given in westward-moving anticlockwise eddy (not shown). Therefore, the location of Senghor Seamount relative to 441 the mesoscale eddies seems to be unlikely to play an important role for these vertically phase-propagating 442 higher-frequency internal waves.

443
The prevailing upward phase propagation indicates downward energy propagation and an energy 444 source higher up in the water column. These waves became particularly prominent near / during spring 445 tides when peak semidiurnal barotropic current speeds were ≥ 4.5 cm s -1 (see Fig. 6e: after day 288.7, when 446 peak barotropic speeds start exceeding 4.5 cm s -1 (Fig. 6a), the waves become more obvious and can be 447 seen up to day 290.7 in Fig. 7; the waves are also a constant feature in February 2013 when barotropic tidal 449 above the seamount slopes ( Fig. 6-8), may reach northwards to at least the waters above the seamount rise 450 (Fig. 7), did not occur at N-Ref (Fig. 5) and only occurred very faintly at S-Ref (Fig. 7), further support the 451 notion that the waves were generated at Senghor Seamount.

452
Overall, the evidence suggests that these higher-frequency internal waves were tidally generated at 453 the seamount. Given the short wave periods and short vertical wave lengths, these waves might be (related 454 to) higher harmonics of the internal-tide beams that emanate from the uppermost parts of Senghor 455 Seamount (this is especially likely to be the case for largely supercritical seamounts such as Senghor isohalines, isotherms and oxygen isopleths across the seamount (Fig. 9). Due to the paucity of the 466 hydrographic dataset it is impossible to tell whether this is temporal or spatial variability or both. In any 467 case, the existence of these abrupt shifts and the evidence for temporally highly variable mixed-layer and 468 pycnocline thicknesses suggest that fluid dynamics around the upper seamount can be vigorous, probably 469 due to the higher-frequency seamount-generated internal waves.
where λ is the decay constant of 234 Th ( = ln2 / t 1/2 ≈ 0.02876 d -1 ) and z is the depth at which 234 Th / 238 U 550 equilibrium is met. An implicit assumption of Eq. 1 is that there is no net physical-oceanographic transport 551 of 234 Th into or out of a sampling site due to turbulent diffusion and / or advection that is superimposed on 552 spatial A t Th gradients.

634
The results of all these calculations are compiled in Fig. 12b and Tab. 4. Uncertainties were propagated 635 from the 234 Th export estimates and POC / 234 Th ratios and given as ±1SD and with the relative 1SD 637 12b), providing confidence in the export estimates and allowing for a comparison of the different sampling 638 stations. In the following, the distribution of POC export fluxes will be described (Objective O2). the respective surface-ocean area on the southern side of the seamount (Fig. 2a,b, 5, S3). The sea-surface 665 reflection areas are likely to be associated with enhanced energy dissipation that could translate into 666 increased shear-driven aggregation and export. This is also further evidence to support the notion that S-

667
Ref is better suited as a reference for the seamount sites than N-Ref.

668
Finally, we include a note of caution that relates to the possible effect of atmospheric dust input. In  1). They also demonstrated that this abrupt increase was followed by an equally rapid reduction in the 708 nitrate concentration and a concomitant drop in total 234 Th concentrations, suggesting that the sudden 709 nutrient supply was rapidly converted into a moderate export flux of particulate matter.

710
In analogy, at Senghor, surface waters passing through the screen of enhanced mixing could have been 711 associated with a similarly sudden and moderate increase in primary productivity and particle export. (examples can be seen in Fig. 10,11). Depth-integration of the amount of DO in this subsurface peak above 724 the 'background' concentration in the surface mixed layer yields an estimate of the 'excess' DO inventory in 725 this water layer which can be interpreted as an approximate measure of photosynthetic activity.

726
Excess DO inventories together with the integration depths and maximum DO concentrations for each 727 profile are given in Tab. 6. The highest excess DO inventories were found at station 891 (seafloor at 705 m 728 water depth) that lies near the streamline that connects N-Slope and W-Slope, and at station 920 (2805 m 729 water depth at the seafloor) that also lies on the NW-Slope but further away from the streamline that 730 connects N-Slope and W-Slope (Fig. 1b,3). At sites where repeat CTD casts were performed some temporal 731 variability in DO inventories was displayed which may be due to the aforementioned passing internal 732 waves. However, even if this variability is taken into account, the DO inventories on the NW slope stand out 733 (Tab. 6). The high DO inventories above the NW slope, therefore, support the notion of pulsed productivity 734 in waters passing from the northern to the western slope areas.

735
Unfortunately, there is no downstream station for S-Slope to scrutinise whether a signal comparable to 736 the one of W-Slope formed above the SW side of the seamount. However, the remotely sensed NPP

800
The comparison between the southern reference site and seamount sites revealed what is interpreted 801 as a detectable seamount effect on POC export: calculated POC export at the seamount sites was 802~2-4 times higher than at the southern reference site. Therefore, the core hypothesis of this study that a 803 tall seamount can trigger enhanced localised POC export is accepted. It can also be concluded that the POC 809 The tidally generated screen of high rates of energy dissipation that runs across the seamount is 810 proposed to result in a combination of two main factors that led to the increased POC export above the 811 seamount: (1) increased upward transport of nutrients into the euphotic zone, driving sudden, brief pulses 812 of primary production of new particulate matter, followed by the particles' export into deeper waters; and

815
It can be speculated that a variant of the shear-based mechanism may also affect deeper seamounts

1220
Each plot displays salinity (shaded background), potential temperature (solid lines: °C) and dissolved oxygen 1221 (dashed lines: mg O 2 L -1 ). CTD station numbers used to produce these plots are displayed above their cast 1222 locations in each plot. Created in ODV with VG-gridding interpolation (Schlitzer, 2002). The bathymetry was 1223 constructed by interpolating between station bottom depths, which were determined via a swath 1224 multibeam survey carried out during the cruise (Fig. 1b).    ), expressed as disintegrations per minute per litre of seawater (dpm L -1 ). Uncertainties are given as one propagated standard deviation (1SD).