Turbulent nutrient fluxes in the Iceland Basin

Deep-Sea Research I 63 (2012) 20–35

Contents lists available at SciVerse ScienceDirect

Deep-Sea Research I journal homepage: www.elsevier.com/locate/dsri

Turbulent nutrient fluxes in the Iceland Basin A. Forryan a,n, A.P. Martin b, M.A. Srokosz b, E.E. Popova b, S.C. Painter b, M.C. Stinchcombe b a b

University of Southampton, Waterfront Campus, Southampton SO14 3ZH, UK National Oceanography Centre, Southampton SO14 3ZH, UK

a r t i c l e i n f o

a b s t r a c t

Article history: Received 18 May 2011 Received in revised form 14 December 2011 Accepted 16 December 2011 Available online 24 December 2011

As part of a multidisciplinary cruise to the Iceland Basin in July–August 2007, near to the historical JGOFS Ocean Weather Station India site (  591 N,  191 W), observations were made of vertical turbulent nutrient fluxes around an eddy dipole, a strong mesoscale feature consisting of a cyclonic eddy and an anti-cyclonically rotating mode-water eddy. Investigation of the spatial distribution of vertical turbulent diffusivity around the dipole shows an almost uniform horizontal distribution despite the strong horizontal gradients in water velocity and density observed. An area mean turbulent diffusivity was calculated as 0.21 (95% confidence interval: 0.17–0.26)  10  4 m2 s  1 at the base of the euphotic zone. The vertical turbulent fluxes of three major macro-nutrients into the euphotic zone were calculated as 0.13 (95% confidence interval 0.08–0.22) mmol m  2 day  1 for nitrate, 0.08 (0.05–0.12) mmol m  2 day  1 for silicate and, 8.6 (13.0–5.2 )  10  3 mmol m  2 day  1 for phosphate. The vertical turbulent flux of dissolved iron (dFe) into the euphotic zone was calculated to be 2.6 (95% confidence interval 1.3–4.3)  10  6 mmol m2 day  1. Turbulent macro-nutrient flux is estimated to contribute up to 14% of the deep winter mixing supply of silicate, nitrate and phosphate in the region. The magnitude of turbulent dFe flux is estimated to be at most 8% of the deep winter mixing supply of dFe. Deep winter mixing is hypothesised to supply an adequate amount of iron to fully utilise the deep winter mixed supply of silicate but not the deep winter mixed supply of nitrate. This suggests that while the iron supply may not limit the magnitude of the spring bloom, iron limitation may be occurring post bloom. & 2012 Elsevier Ltd. All rights reserved.

Keywords: Nitrate Silicate Phosphate Dissolved iron Turbulent diffusion North Atlantic

1. Introduction Phytoplankton, the myriad tiny single-celled plants that comprise the vast majority of life in the oceans, live exclusively in the sunlit upper 100–200 m of the water column known as the euphotic zone. In addition to sunlight, phytoplankton require a regular supply of macro-nutrients – nitrate, phosphate and, silicate – as well as trace elements, such as iron, to grow. However, with the exception of some phytoplankton that can fix nitrogen, the principle source of macro-nutrients and trace elements is from the remineralisation of decaying organic matter. Gravity ensures that decaying organic matter sinks, and so this remineralisation occurs almost exclusively in the dark waters below the euphotic zone (e.g., Sarmiento et al., 2006). Consequently, for phytoplankton to grow nutrient-laden water must pass from the deeper, dark sections of the water column up into the euphotic zone. To a large extent, it is the physical mechanisms driving this transfer of water into the euphotic zone which dictates patterns of primary production within the ocean (e.g., Sarmiento et al., 2006; Williams and Follows, 2003). Of the many

n

Corresponding author. Tel.: þ44 23 8059 6666. E-mail address: [email protected] (A. Forryan).

0967-0637/$ - see front matter & 2012 Elsevier Ltd. All rights reserved. doi:10.1016/j.dsr.2011.12.006

physical mechanisms that bring deeper waters up to the surface, three that have received the most attention are deep winter convective mixing (e.g., Williams et al., 2000), turbulent mixing (e.g., Law et al., 2001; Jickells et al., 2008; Lewis et al., 1986) and mesoscale up-welling (e.g., Klein and Lapeyre, 2009; Martin and Pondaven, 2003; McGillicuddy et al., 2003; Oschlies, 2002; Le´vy et al., 2001; Allen et al., 2005). In the north Atlantic sub-polar gyre, of which the Iceland Basin is a part, the primary source of nutrients to the surface ocean is thought to be deep winter convection (Williams and Follows, 2003). In the Iceland Basin, seasonal re-stratification and increasing levels of irradiance lead to a major phytoplankton bloom event in spring when the nutrients brought to the surface by deep winter mixing are largely consumed (Allen et al., 2005; Sanders et al., 2005; Williams and Follows, 2003). The spring bloom in the north Atlantic is dominated by diatoms, rapidly growing phytoplankton that use silicon in the construction of their cell walls (Allen et al., 2005; Sanders et al., 2005). When the supply of silicate in the surface waters is exhausted the spring bloom decays and the phytoplanktonic community shifts towards non-siliceous species (Allen et al., 2005; Sanders et al., 2005). However, significant nitrate and phosphate concentrations can persist post bloom (Sanders et al., 2005). Several reasons for the existence of residual post-bloom nutrient concentrations in the sub-polar gyre have

A. Forryan et al. / Deep-Sea Research I 63 (2012) 20–35

been advanced (Nielsdo´ttir et al., 2009; Sanders et al., 2005; Popova et al., 2002). Nutrient levels supplied through deep winter mixing topped up with the occasional injection of fresh nutrient through summer storms may supply more nutrient than can be used by phytoplankton in the high latitude, relatively low light regime of the sub-polar gyre (Popova et al., 2002). Heavy grazing by zooplankton and species succession as the bloom progresses may result in the dominant phytoplankton groups at the end of the bloom primarily utilising re-cycled nitrogen, such as ammonium, and unable to utilise fully the remaining ‘‘fresh’’ nitrate pool (Sanders et al., 2005). Light levels at high latitudes may be insufficient to allow adequate nitrate uptake by non-siliceous phytoplankton (Sanders et al., 2005). Iron limitation has also been suggested as contributing to the residual post-bloom macronutrient concentrations in the Iceland Basin (Nielsdo´ttir et al., 2009). Iron is an essential trace element for all phytoplankton and a lack of iron has been demonstrated to result in low production despite high concentrations of surface macro-nutrients in the Southern Ocean, sub-polar Pacific Ocean, and equatorial Pacific Ocean (Boyd et al., 2007). In contrast to the seasonal supply of nutrients from deep winter mixing, the supply of nutrient through vertical turbulent mixing is considered to be a relatively small flux (Williams and Follows, 2003; Williams et al., 2000). However, where nutrients are limiting the small but constant supply of limiting nutrient from turbulent mixing may still control levels of primary production. For example, in the oligotrophic eastern Atlantic the vertical nitrate flux associated with vertical turbulent transport from deeper waters matches, within error limits, the integrated rate of nitrate uptake measured, in situ, by 15N-labelled nitrate incorporation (Lewis et al., 1986). In the post-bloom Iceland Basin two of the potential explanations for the residual nutrient pool posit a limiting nutrient, either silicate (Sanders et al., 2005) or more recently iron (Nielsdo´ttir et al., 2009). In both cases the vertical turbulent nutrient flux could potentially play an important role in the post-bloom supply of nutrient. Observations in the presence of strong mesoscale features such as the ones originating from the Gulf Stream suggest that there is little or no observable spatial variation in turbulent mixing at the mesoscale (Gregg and Sanford, 1980; Oakey and Elliott, 1977). However, there is reason to believe that enhanced mixing may be occurring as a product of frontal instability processes associated with phenomena such as frontogenesis (Nagai et al., 2009; Rodrı´guez-Santana et al., 1999, 2001). Additionally geostrophically stable shear flow associated with mesoscale features may set up conditions for vertical turbulent mixing which is then triggered by other processes such as internal waves, wind and tides (Moum et al., 2009; Rippeth et al., 2009). Such mixing events, whether triggered through instability or process interactions, would be both temporally and spatially highly heterogeneous. Over the past 20 years there have been several studies made on the biophysics of the post-bloom Iceland Basin (Jickells et al., 2008; Marra et al., 1995; Harris et al., 1997; Martin et al., 1998). On two separate occasions turbulent nutrient fluxes have been estimated for the base of the mixed layer with turbulent mixing inferred indirectly through tracer release yielding a temporally and spatially integrated value (Law et al., 2001; Jickells et al., 2008). Here we present the first direct measurements of upper ocean turbulent mixing in the Iceland Basin taken using a free-fall microstructure profiler in the presence of strong mesoscale flow. Unlike previous studies from the region that measured turbulent mixing, the use of a microstructure profiler allows an estimation of the spatial distribution of the turbulent diffusivity around strong dynamical mesoscale features to be made and turbulent diffusivity to be calculated simultaneously for multiple depths in the water column. These measurements are combined with contemporary measurements of nutrient concentrations for

21

the three main macro-nutrients, nitrate, silicate and phosphate as well as with contemporary measurements of dissolved iron, to estimate directly turbulent nutrient fluxes. The importance of this turbulent flux to post bloom primary production in the Iceland Basin is discussed. In this paper we focus on turbulent nutrient fluxes across the base of the euphotic zone. The base of the euphotic zone represents the limit of primary production. In the data presented here the depth of the euphotic zone is consistently deeper than the depth of the mixed layer and the maximum in recorded fluorescence (data not presented here) is in the majority of cases at a depth that is between the depth of the mixed layer and the base of the euphotic zone. Hence, measurable amounts of primary production can still occur below the base of the mixed layer. Under these circumstances the turbulent flux of nutrients into the euphotic zone is of more relevance to estimates of total primary production than the flux into the mixed layer. The surface mixed layer, especially in the presence of horizontal density gradients, can be subject to a range different physical processes which can potentially affect mixing across the base of the mixed layer both directly and indirectly through changing local stratification (Hoskins and Bretherton, 1972; Boccaletti et al., 2007; Molemaker et al., 2005; Thomas and Lee, 2005). These surface layer physical processes are potentially highly spatially and temporally heterogeneous making turbulent mixing measured at the base of the surface mixed layer much more variable than that measured deeper in the water column. Hence, for the data presented here, the base of the euphotic zone is preferred for measuring turbulent nutrient flux.

2. Material and methods 2.1. Description of the survey site plus background to cruise. The measurements presented here were taken as part of UK RRS Discovery cruise D321 to the Iceland Basin between 24th July and 23rd August 2007 (Fig. 1). The purpose of this cruise was to examine controls on export production in the region. On arrival at the survey site it was found that within the survey area there was an eddy dipole, consisting of a cyclonic eddy and an anticyclonically rotating mode-water eddy. The cyclonic eddy is similar to that described by Harris et al. (1997) and is characterised by a doming up of the isopycnals displacing the permanent thermocline upwards, cyclonic rotation, and a reduction in sea-surface height. The cyclonic eddy has an elevated sea-surface temperature signal (Figs. 2 and 3). The mode-water eddy is similar to eddies described by Martin et al. (1998) and Read and Pollard (2001). The mode-water eddy consists of a lens-shaped water mass at mid depth (centred on  550 dbar) displacing the seasonal thermocline upwards and the permanent thermocline downwards. The mode water eddy has strong anti-cyclonic rotation, elevated sea-surface height and a strong barotropic component to the flow. The mode-water eddy has a reduced sea-surface temperature signal (Figs. 2 and 3). Within the dipole the two eddies interact producing a region of high current speed (  0:7 ms1 ) between the eddy cores. The influence of this high speed region between the eddy cores is apparent where high chlorophyll concentration water has been drawn in from north of the survey region forming a filament (Fig. 2). 2.2. Hydrographic measurements and ADCP During a 3-week period, the eddy dipole was mapped using conductivity, temperature and depth measurements (CTD) from a combination of towed vehicle (Sea-Soar, maximum depth

22

A. Forryan et al. / Deep-Sea Research I 63 (2012) 20–35

16247 16242

16241 16285

16283 16288 16295

64°N

16286

16289 16292 16269 16260

16232

62°N

16222

Lat. ( °N

)

16226

60° N

58° N 0°

25°W

20°W

10° W

15° W

5° W

Lon. (°W) Fig. 1. Map showing the location of the area surveyed during UK RRS Discovery cruise D321 to the Iceland Basin July to Aug. 2007. The D321 cruise track is marked in red on the main figure. The black dashed rectangle represents the survey area and is expanded in the inset. The positions of the stations where turbulence measurements were taken are plotted on the insert, diamonds indicate stations with both turbulence and nutrient concentration measurements, crossed diamonds stations with turbulence, nutrient, and iron concentration measurements (see Table 1). The station numbers are colour coded to match the region that the station has been assigned to (Table 2; Section 2.7), with black representing the background region, green the edge region, blue the core region, and red the jet region. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

of  400 mÞ and conventional full depth (  2900 mÞ and partial depth (  1000 mÞ vertical profiles. Three surveys of the eddy dipole were carried out; survey one from 5 August 2007 to 10 August 2007 using a combination of Sea-Soar and CTD with ADCP, survey two from 10 August 2007 to 15 August 2007 using conventional CTD with ADCP and survey three from 15 August 2007 to 22 August 2007 again using ADCP with Sea-Soar (Allen, 2007). As part of each survey, direct measurements of turbulent mixing were made and nutrient samples taken. A summary of all the measurements taken at each of the stations where turbulent mixing was measured is given in Table 1. Current velocity down to  300 m was measured using a shipmounted 150 kHz RDI Acoustic Doppler Current Profiler (ADCP). The instrument was configured to sample over 120 s intervals with 96 depth intervals of 4 m thickness starting at 14 m depth using pulse length 4 m and blank beyond transmit of 4 m. Calibration of the ADCP was carried out over the continental shelf on route to the survey site. Values of the misalignment angle (14.41), which corrects for the rotational position of the ADCP on the ship’s hull relative to the ship’s axis, and the amplitude factor (0.9683), which corrects for the fore-aft tilt of the instrument relative to the horizontal plane, were applied to all subsequent ADCP water velocity estimates (Allen, 2007).

2.3. Calculation of mixed layer and euphotic depths The euphotic depth was calculated as 1% of surface irradiance, where irradiance was measured using a 4p downwelling Photosynthetically Available Radiation (PAR) sensor attached to the main shipboard CTD frame. A mean euphotic depth was calculated using PAR data from all CTD casts taken during the D321 cruise. Mixed layer depths were calculated for each station using the method of Kara et al. (2000).

2.4. Nutrient concentration measurements 2.4.1. Macro-nutrient concentrations Analysis for micro-molar concentrations of nitrate and nitrite (referred to hereafter collectively as nitrate), phosphate, and silicate was carried out using a scalar Sanplus autoanalyser following the methods described in Kirkwood (1996) with the pump rates through the phosphate line increased by a factor of 1.5. Samples were analysed within 24 h of being taken and were kept refrigerated at  4 1C until analysed. An artificial seawater matrix (ASW) of 40 g L  1 sodium chloride was used as the inter-sample wash and standard matrix. The nutrient free status of the ASW solution was checked by running Ocean Scientific International (OSI) nutrient free seawater on every run of the autoanalyser. Data processing was done using Skalar proprietary software and was carried out within 72 h of the sample analysis run being finished (Allen, 2007). 2.4.2. Iron concentrations Seawater samples to be analysed for dissolved iron (dFe) were collected using a titanium frame CTD with designated ‘‘ironclean’’ sample bottles. Samples were pressure filtered using nitrogen free oxygen through 0:4 mm and 0:2 mm filters and acidified to a pH  1:8 with ultra pure HCl. Dissolved iron concentration was measured using flow-injection chemiluminescence methods where samples are buffered with ammonium acetate to pH 4 and pre-concentrated on a resin column during analysis. Each sample was run in triplicate. The dissolved iron (dFe) data collected as part of UK RRS Discovery cruise D321 has been published separately by Nielsdo´ttir et al. (2009). 2.4.3. Estimating mean nutrient profiles An area mean nutrient profile for all turbulence stations was estimated, for each nutrient, by first linearly interpolating individual

A. Forryan et al. / Deep-Sea Research I 63 (2012) 20–35

60

14

result of turbulent motions as described by the equation

13.8

@C @2 C ¼K 2 @t @z

13.6

59.8

13.4 59.6

13.2 13

59.4

12.8 59.2

12.6 12.4

59 12.2 −21.5

−21

−20.5

−20

−19.5

−19

−18.5

12 1

60

0.9 0.8

59.8

0.7 59.6

0.6 0.5

59.4

0.4 59.2

0.3 0.2

59 0.1 −21.5

−21

−20.5

−20

−19.5

−19

−18.5

0

Fig. 2. Satellite images of AVHRR sea surface temperature (1C upper panel) and MODIS chlorophyll concentration (mg m  3 lower panel) from the 5th and 6th August 2007 respectively for the D321 survey site (see Fig. 1). Processed satellite image data for sea surface temperature from  1 km resolution AVHRR data and chlorophyll concentration from 1 km resolution MODIS data were downloaded from the NERC Earth Observation Data Acquisition and Analysis Service (NEODAAS). Estimated positions, and flow direction for the eddy dipole and jet regions are marked. The warm and cold sea-surface temperature signals of the cyclone and mode-water eddies respectively can be seen, with the cyclone centred approximately at 59.8N 19.8W and the mode-water centred approximately at 59.5 1N 20.4 1W. The white areas are cloud.

station nutrient profiles onto a common set of pressures. Nutrient samples were taken at different pressures on each cast in order to satisfy the requirements for other biological studies conducted on the cruise. The set of common pressures used for all macro-nutrients were 14, 25, 32, 37, 52, 81, 131, 206, 408, 610, 811 and 1014 dbar and for iron were 7, 12, 22, 29, 34, 49, 78, 128, 403, 598, 801 and 1010 dbar. Common pressures, for each nutrient, were chosen to minimise the differences in pressure between the original sample pressures and the nearest common profile pressure. Following interpolation nutrient concentrations were then averaged, at each pressure of the common profile, over all profiles. The averaged profiles for each nutrient at each station were checked for accuracy by visually comparing to the original nutrient profiles.

2.5. Turbulence measurement The magnitude of turbulent mixing can be characterised as a turbulent diffusivity (K) which, in analogy to molecular diffusivity, describes the rate at which scalar properties disperse as a

23

ð1Þ

(Tennekes and Lumley, 1972) where C is a scalar property, t is time and z is depth. Turbulent diffusivity is related to the dissipation of turbulent kinetic energy (e), using the relationship

e¼

KN 2

G

ð2Þ

(Osborn, 1980) where G is a constant mixing efficiency and N is the buoyancy frequency, N2 ¼ 

g dr r dz

(Gill, 1982) where g is acceleration due to gravity and r potential density. In line with previous studies (Rippeth et al., 2003; Moum et al., 1995; Osborn, 1980) a value of 0.2 was used for the mixing efficiency (G). Turbulent kinetic energy dissipation can be estimated directly from measurements of microstructure velocity shear made using a microstructure shear profiler. The microstructure profiler used here was an MSS90L free-fall microstructure profiler (serial number 35) produced by Sea and Sun Technology GmbH and ISS Wassermesstechnik. The profiler is cylindrical in shape with two velocity microstructure shear probes as well as standard high precision CTD sensors mounted at the descending end protected by a guard ring. The two shear probes are on slim shafts,  150 mm in front of the CTD sensors, measuring the velocity fluctuations in the ‘‘clean’’, undisturbed, water in advance of the other sensors. A vibration control sensor and a two component tilt sensor provide data to remove noise contamination from the signal. The profiler has buoyant foam rings at the opposite end from the sensor array where a light tether is attached for data and power transmission. On deployment the profiler is allowed to free-fall vertically through the water, the sensor array downwards, by maintaining sufficient slack in the tether. This also isolates the profiler from the motions of the ship and minimises contamination of the signal by vibrations caused through cable tension (pseudo-shear). Data from the sensors are recorded continuously while the profiler is falling by a P.C., connected via the tether, using software provided by Sea and Sun Technology GmbH (Prandke, 2008). The calibration of the CTD sensors was carried out by Sea & Sun Technology GmbH using standard calibration equipment and procedures for CTD probes. The vibration control sensor, the tilt sensors and the shear sensors were calibrated by ISW Wassermesstechnik. Each station consisted of approximately 10 casts of the microstructure shear profiler taken over the course of approximately 1 h. To obtain robust estimates of dissipation rates, the measurements from all casts of a station were combined. This is necessary because the mixing processes involved are intermittent, in both time and space. Consecutive casts often show considerably different, yet genuine, structure with the distribution of turbulent diffusivities at the same pressure for different casts usually being strongly non-Gaussian. Turbulent diffusivity and turbulent kinetic energy dissipation data were averaged into 8 dbar vertical pressure intervals to ensure that the turbulence data were averaged over a vertical scale greater that the estimated Thorpe scale (Thorpe, 1977) for the turbulent overturning. The distribution of the turbulent diffusivity and turbulent kinetic energy dissipation data for each station was tested using the Kolmogorov–Smirnov test (Press et al., 1989) over pressure intervals of 8 dbar and found, in all cases to be lognormally distributed. Consequently, when averaging turbulent diffusivity and dissipation data into 8 dbar

24

A. Forryan et al. / Deep-Sea Research I 63 (2012) 20–35

50 cms

1624916264 16265

16248

60

16251

16265 1027.5

59.8

16266

16278

16263

16252

16266 1027.4

Lat. (°N)

59.6

16267

16277

16262

16253

16267 1027.3

59.4

16268

16275

16254

16261

16268 1027.2

16274

59.2

16269

16260

16255

16269 1027

59

16270

16273

16259

16270

16256 1026.5

16272

58.8 −20.1

16271 −19.6

16258

16271

16257

−19.1

−18.6

100

200

Lon. (°E)

300

400

500

600

700

800

press (dbar)

Fig. 3. Acoustic Doppler Current Profile (ADCP) current velocity and the position of conductivity, temperature, and depth (CTD) measurements made during UK RRS Discovery cruise D321 survey two from 10th August 2007 to 15th August 2007 (left hand panel). The position of the two eddies in the eddy dipole, estimated as described in Section 2.7, are marked with the mode-water eddy (centered on 59.2 1N–19.8 1E) in red and the cyclone (centered on 59.7 1N–19.4 1W) in black. A contour cross section of density at 0.1 kg m  3 intervals along  19.5 1E between stations 16265 and 16271 shows the 1027.3 kg m  3 isopycnal doming up at approximately 59.2 1N and 59.7 1N coincident with the estimated positions of the eddy cores (right hand panel). Table 1 The position, date and number of casts taken for each turbulence measurement station. Station no.

Date

Position (deg. min.)

No. of casts

Max. depth of profile (m)

16222 16226 16232 16241 16242 16247 16260 16269 16283 16285 16286 16288 16289 16292 16295

02/08/07 05/08/07 06/08/07 09/08/07 09/08/07 10/08/07 12/08/07 13/08/07 16/08/07 18/08/07 19/08/07 20/08/07 20/08/07 20/08/07 20/08/07

581 581 591 591 591 591 591 591 591 591 591 591 591 591 591

10 10 10 10 12 10 10 9 10 11 10 10 10 10 10

141 152 139 135 130 138 134 133 139 134 129 204 138 133 130

50 N 50 N 01 N 52 N 52 N 56 N 10 N 12 N 36 N 41 N 17 N 30 N 26 N 22 N 18 N

191 211 211 191 201 201 191 191 201 181 191 191 191 191 191

51 W 00 W 00 W 37 W 07 W 26 W 08 W 28 W 38 W 42 W 47 W 02 W 16 W 26 W 40 W

pressure intervals the method of Baker and Gibson (1987) was used. The mean (M) of the lognormally distributed data is given by M ¼ exp



mþ

s2



2

Dissolved Iron

Y Y

Y Y Y Y Y Y

Y

Y

ADCP

Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y

where qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi Zb ¼ ½s2 =n þ s4 =2ðn1Þ and n is the number of data points (Baker and Gibson, 1987).

ð3Þ

where m and s2 are the arithmetic mean and variance of the log transformed data (Baker and Gibson, 1987). The 95% confidence intervals to the mean are then given by M n exp ð 71:96nZb Þ

Macro-nutrients

2.5.1. Calculation of turbulent diffusivity Vertical microstructure shear was calculated from the measurements taken using the microstructure profiler following the method of Stips (2005). Assuming isotropic turbulence (Yamazaki and Osborn, 1990) the rate of turbulent kinetic energy dissipation can be calculated from the variance of the vertical microstructure

A. Forryan et al. / Deep-Sea Research I 63 (2012) 20–35

calculation of turbulent dissipation (Oakey, 1982; Moum et al., 1995; Rippeth et al., 2003).

shear  0 2 15 du e¼ v 2 dz

ð4Þ 2.6. Calculating nutrient fluxes

where n is the molecular viscosity of water, u0 is the turbulent velocity fluctuations and the overbar indicates a spatial mean value (Lueck et al., 2002). The variance of the vertical microstructure shear is determined by integration of the vertical microstructure shear power spectrum (F(k)), where k is the wavenumber, estimated using the Welch modified periodigram method (Welch, 1967) from the vertical microstructure shear fluctuations. Hence Eq. (4) can be represented as Z kc 15 e¼ v FðkÞ dk 2 kl (Stips, 2005; Rippeth et al., 2003; Moum et al., 1995). The measured vertical microstructure shear power spectrum was used to scale and dimensionalise a non-dimensional analytical form of the empirical Naysmith universal turbulence spectrum (FNas ðkÞ) 1=3

FNas ðkÞ ¼

8:05k

Nutrient flux is conventionally modelled by analogy with molecular diffusion (Eq. (1)). The vertical flux of nutrient into the waters above a depth z is therefore given by the product of the turbulent diffusivity and the nutrient gradient at depth z FðzÞ ¼ KðzÞ

@C @z

ð5Þ

where C is the concentration of nutrient. Vertical nutrient gradients were calculated by first order differencing nutrient concentration profiles. For the stations where nutrient measurements were taken on CTDs before or following turbulence stations (see Table 1) the calculated turbulent diffusivity for the sample pressure interval was combined with the nutrient gradient for the station at the corresponding pressure to give an estimate of turbulent nutrient flux. 2.7. Estimating horizontal distribution

3:7

1þ ð20kÞ

(Roget et al., 2006). The universal spectrum was scaled by curve fitting, using a least squares fit, to the measured power spectra between the limits 2–30 cpm (cycles per meter) for each 1 s segment of recorded data (  1024 data points representing 0.5 dbar with a configured probe drop speed of 0.5 dbar s  1). The lower limit wavenumber of 2 cpm, the smallest wavenumber resolvable within a pressure interval of 0.5 dbar, eliminates low frequency noise from the probe tumbling during descent (Prandke, 2007). The maximum upper limit of the integration was selected to be below the resonant frequency of the shear probe guard ring (Prandke and Stips, 1998). The rate of kinetic energy dissipation was then calculated by integration of the fitted universal spectrum between 2 cpm and the Kolmogorov wavenumber (kc). The Kolmogorov wavenumber, the reciprocal of the Kolmogorov microscale, is given by kc ¼

25

 e 1=4 v3

and represents the smallest scale of turbulent motions unaffected by the dissipative effects of molecular viscosity. A correction for the attenuation of the shear probe response as the wavelength of the velocity fluctuations decreases was applied to the dissipation estimate using an empirical polynomial function derived for the shear probe (Prandke, 2008, 2007). The kinetic energy dissipation rates calculated for each of the two independent shear sensors were combined, to provide a single estimate of dissipation for each cast, following the method described in Prandke (2008). Turbulent diffusivity was calculated from the measured rate of dissipation of turbulent kinetic energy using Eq. (2). Errors in calculating estimates of the dissipation rate arise from a number of sources. Calibration of the shear sensors is to within 75%, and the influence of non-isotropic turbulence is estimated to add up to 35% error to calculations (Yamazaki and Osborn, 1990). In addition to these, uncertainties in the flow speed past the shear probe, estimated to be  7 5%, adds an additional  20% error to the calculation (Oakey, 1982; Moum et al., 1995), as the calculated dissipation depends on the variance of flow shear squared. Lesser o 10% (Dewey and Crawford, 1988) errors arise from drift in shear probe calibration and uncertainties in the estimates of viscosity. Combining all the estimates of error together gives a generally accepted estimate of 750% error in the

In an attempt to characterise spatial variation in turbulent mixing due to the influence of the eddy dipole, the location of each of the turbulence stations was grouped according to their relationship to the dipole. The positions and core diameters of the two eddies were estimated using the ADCP data from the three surveys made during cruise D321 (Section 2.2) by least squares fitting of the ADCP data, recorded at 63 m depth (63.6 dbar pressure the closest ADCP depth interval to the euphotic depth), to azimuthal velocity profiles of the form    r  1 r2 VðrÞ ¼ V 0 exp 1 2 R 2 R (Martin and Richards , 2001) where V(r) is the azimuthal velocity at radius r from the eddy centre, V0 is the maximum azimuthal velocity, and R is the radius of maximum azimuthal velocity. As the position of the dipole changed from survey to survey, three separate estimates of the positions of the two eddies were made. For each ADCP survey, values of V0 and R for both eddies and the positions of the eddy centres were fitted to the ADCP velocity data for 63 m depth by minimising the root mean square difference between the calculated velocity field and the ADCP velocity data. The position of turbulence stations conducted as part of a survey were then compared to the estimated position of the eddies in the dipole for that survey. Stations were grouped into four distinct dynamical regions: the jet region between the two eddy cores, the eddy cores, the eddy edges away from the jet region, and the background waters. Once assigned, the location of the station was checked by consideration of both the current magnitude and direction during the station, taken from the shipboard ADCP. The location of each of the turbulence stations relative to the dipole is given in Table 2. The four dynamical regions were distinguished primarily by consideration of position relative to the eddies in the dipole and by water movement (Appendix A). The jet and edge regions are characterised by approximately constant direction flow with speeds greater than 0.2 m s  1, while the background and core regions are characterised by low speed flow ( o 0:2 ms1 ) with more variable direction. For station 16283, the identification of appropriate region is difficult. The station is sited away from the estimated positions of the known eddy dipole cores. The magnitude of the current velocity while station 16283 was in progress was low ( o 0:2 ms1 ) and the current direction was erratic, showing no

26

A. Forryan et al. / Deep-Sea Research I 63 (2012) 20–35

Table 2 Turbulence measurement stations grouped according to location with respect to the eddy dipole structure. Region

Jet

Core

Edges

Background

Station number

16286 16288 (cyclone) 16289 16292 16295 16269 (mode  water)

16283 (uncertain see Section 2.7) 16285 (cyclone)

16241 (cyclone) 16242 (cyclone) 16247 (cyclone)

16222 16226 16232 16260

constant direction. This is consistent with either known background stations or with station 16285, known to be within the cyclone eddy core. The mixed layer depth of 45.5 m (46 dbar) for station 16283 is greater than observed for known background stations and most closely compares to station 16285 in the cyclonic eddy core (51.5 m). This would suggest that the station 16283 was in the core of a third unidentified eddy within the survey region. However, current direction on arriving and leaving station 16283 did not show the changes which would be expected on entering and exiting an eddy with a rotational current flow. As a result of the uncertainty in the location of this station it has not been included in subsequent comparisons of mixing distribution between the four regions. The density profile calculated from CTD station 16286 taken on 19 August 2007 at 591 170 N 191 470 W suggests that this station ought to be within the mode-water eddy core. However, analysis of the ADCP data recorded while the subsequent turbulent diffusivity measurements were taken shows a near constant water velocity of a magnitude comparable to that recorded during stations known to be in the jet region. This would suggest that the ship had, by this time, drifted out of the eddy core and into the jet region between the two eddies. Consequently when considering the regional distribution of turbulent mixing station 16286 is included within the jet region (Table 2). However, when considering nutrient concentrations the station is referred to as within the mode-water eddy. Profiles of buoyancy, turbulent diffusivity and turbulent kinetic energy dissipation from all turbulence stations within each dynamical region were first averaged into 8 dbar pressure intervals. Station profiles of buoyancy, originally calculated from microstructure profiler CTD data at 0.5 dbar pressure intervals, were averaged into 8 dbar pressure intervals by taking a mean for each pressure interval. Turbulence station profiles of turbulent kinetic energy dissipation and turbulent diffusivity were averaged into 8 dbar pressure intervals using Eq. (3). The 8 dbar pressure interval turbulence station profiles of buoyancy were combined into representative profiles for the four dynamical regions by taking a mean, and an estimate of standard error, at each pressure interval. For turbulent kinetic energy dissipation and turbulent diffusivity, the 8 dbar pressure interval station profiles were combined into regional profiles by taking a mean of the log transformed data at each pressure interval and then reversing the log transform. Confidence intervals of the regional profiles were calculated by taking a mean of the log transformed upper and lower confidence limits for the individual station profiles at each pressure interval.

The remaining recorded euphotic depths varied approximately normally between 49 and 83 dbar with no detectable regional variation. The mean depth of the euphotic zone across the whole D321 survey area, for the duration of the survey, was 64710 dbar (63.4710 m mean 7standard deviation). Throughout the survey area, for the duration of the survey, density profiles show a strong seasonal thermocline with the mixed layer depth varying between 18 dbar and 52 dbar (17.8 m and 51.5 m depth respectively). The shallowest mixed layer depths are associated with measurements taken in the jet region and the deepest (52 dbar) in the core of the cyclonic eddy. The mean mixed layer depth for the survey area was 30710 dbar (29.7710 m mean 7standard deviation). 3.2. Nutrient concentrations

3. Results

Macro-nutrients, nitrate, phosphate and silicate, all exhibit typical summertime macro-nutrient vertical profiles, with concentrations increasing with pressure. Surface macro-nutrient measurements, averaged over the top 30 dbar (29.7 m the area mean mixed layer depth) for all stations not within the modewater eddy, show above detection limit concentrations for all three macro-nutrients. The observed surface concentration of silicate is 0.5270.04 mmol m  3 (all stations mean 7standard error), of nitrate is 3.77 70.14 mmol m  3 and of phosphate is 0.370.01 mmol m  3. Within the mode-water eddy (station 16286) surface macro-nutrient concentrations, averaged over the top 30 dbar, are higher than outside the mode-water eddy. Within the mode-water eddy surface concentrations are 1.39 7 0.15 mmol m  3(mean 7standard error) for silicate, 6.85 7 0.31 mmol m  3 for nitrate and 0.4470.01 mmol m  3 for phosphate (Fig. 4). The sharpest gradients in nutrient concentrations in all cases are between the base of the mixed layer and the euphotic depth. Where there is a measurable nutrient concentration gradient across the base of the euphotic zone, the gradient is positive (concentration lower above and higher below). The mean nutrient gradients for all stations across the base of the euphotic zone are 49 75 mmol m4 (mean 7standard error) for nitrate, 32 72 mmol m4 for silicate and 2:57 0:4 mmol m4 for phosphate (Fig. 4). The vertical profiles for dissolved iron (dFe) were more variable than macro-nutrient profiles across the survey region. However, all profiles showed an increase in concentration at the largest scale increasing from 0.1 70.8  10  3 mmol m  3 (all stations mean 7standard deviation) in the mixed layer to 0.3470.19  10  3 mmol m  3 at 400 dbar pressure (Nielsdo´ttir et al., 2009). Gradients in iron concentration across the base of the euphotic zone vary from positive to negative (Fig. 5).

3.1. Euphotic and mixed layer depth

3.3. Nutrient ratios

Minimum and maximum recorded depths of the euphotic zone (euphotic depth) were 33 dbar and 101 dbar respectively.

The molar ratios of macro-nutrients in the surface waters, above the euphotic zone, show a reduction in both silicate and

A. Forryan et al. / Deep-Sea Research I 63 (2012) 20–35

Nitrate

press (dbar)

10

Phosphate

10 20

20

30

30

30

40

40

40

50

50

50

60

60

60

70

70

70

80

80

jet background edge core90

90 100

0

5

10

15

100

20

Silicate

10

20

80

27

90

0

0.5

(mmol m−3)

1

100

0

4

2

(mmol m−3)

(mmol m−3)

Fig. 4. Profiles of macro-nutrient concentrations for the surface 100 dbar for stations in the four regions (Section 2.7). Vertical dashed lines show 7 standard error to the mean concentration for the region where the regional profile is a mean value (jet and background). Here, station 16286 in the mode-water eddy core is included in the jet region (Section 2.7). The mixed layer depth is marked as a horizontal dashed red line (30 dbar). The euphotic depth is marked as a blue dashed line (64 dbar). The gap in the nutrient data between 30 and 50 dbar for the core region (station 16285) is due to contamination of the water sample taken at 36 dbar. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

press (dbar)

station 16236

station 16260

station 16282

10

10

10

20

20

20

30

30

30

40

40

40

50

50

50

60

60

60

70

70

70

80

80

80

0

0.1

0.2

0.3

0.4

0

0.1

0.2

0.3

0.4

0

0.1

0.2

dFe (x10

press (dbar)

10

20

20

30

30

40

40

50

50

60

60

70

70 0

0.1

0.2

0.3

0.3

0.4 −3

mmol m )

station IB16

station 16286 10

80

−3

0.4

dFe (x10−3 mmol m−3)

80

dFe euphotic depth

0

0.1

0.2

0.3

0.4

dFe (x10−3 mmol m−3)

Fig. 5. Profiles of dissolved iron (dFe) concentrations (x10  3 mmol m  3) for the surface 80 m for all stations published in Nielsdo´ttir et al. (2009). Station IB16 is to the North of the D321 survey area at position 61.5 1N 20.00 1W. The euphotic depth is marked as a blue dashed line (64 dbar).

nitrate when compared to deeper waters in all cases (Table 3). The area mean molar ratio of Si:N:P at 610 dbar is 7.66:15.55:1 calculated from the averaged nutrient concentrations from all

stations. The area mean molar ratio of Si:N:P in the euphotic zone is 2.44:13.16:1 calculated for all stations and all pressures above 64 dbar.

28

A. Forryan et al. / Deep-Sea Research I 63 (2012) 20–35

Table 3 Macro-nutrient molar rations in the mixed layer, euphotic zone and at depth. Molar ratios are expressed as Si:N:P from average nutrient concentrations calculated from all nutrient stations. Background (mol:mol)

Core (mol:mol)

Edge (mol:mol)

Jet (mol:mol)

Mixed layer (above 30 dbar) Euphotic zone (above 65 dbar) 610 dbar

1.1:11.1:1 2.1:13:1 7.3:17:1

1.3:10.5:1 4.8:13:1 9.2:14:1

1.6:14:1 2.5:14.5:1 7.6:15.8:1

2.4:14.1:1 3.7:15:1 6.8:15.4:1

20

20

30

30

40

40

50

50 press (dbar)

press. (dbar)

Pressure

60

70

60

70

80

80 edge core jet background

90

100 10−7

10−6

background jet edges core

90

10−5

10−4

10−3

100 10−9

N2 (s−2) 2

10−8

10−7

ε (W kg ) −1

Fig. 6. Profiles of N for the four regions. Note the cyclonic eddy core region is represented by a single station (16285). Error bars mark standard errors for the combined regional values. The depth of mixed layer is marked as a dashed red line (52 dbar for core, 30 dbar for the remaining regions). The euphotic depth is marked as a dashed blue line (64 dbar). (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

Fig. 7. Profiles of turbulent kinetic energy dissipation (e) for the four regions. Note the cyclonic eddy core region is represented by a single station (16285). Dashed lines mark the upper and lower 95% confidence limits for each regional profile. The depth of mixed layer is marked as a dashed red line (52 dbar for core, 30 dbar for the remaining regions). The euphotic depth is marked as a blue dashed line (64 dbar). (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

The ratio of dFe to nitrate for all published iron stations at depths below the euphotic depth, calculated from averaged nutrient concentrations for all pressure levels and all stations, is 270.3  10  5:1 (mean 7standard error).

3.5. Turbulent mixing

3.4. Buoyancy Buoyancy frequency, for all regions, shows a peak just below the mixed layer in the seasonal thermocline where the density gradients are steepest, with N2 between 1 and 10  10  4 s  2 N2 reduces with pressure to between 1 and 2  10  5 s  2 at the euphotic depth for all stations except 16285 (N2 ¼5  10  5 s  2) in the core of the cyclonic eddy (Fig. 6). Below the sub mixed layer peak, buoyancy frequency for the cyclonic eddy core is higher than for the other regions throughout the pressure range, with N2 between 2  10  5 s  2 and 3  10  4 s  2 (Fig. 3).

Turbulent kinetic energy dissipation for the jet, background, and edge regions is between 1 to 3  10  9 W kg  1 at the euphotic depth. The 95% confidence limits on turbulent kinetic energy dissipation for the three regions overlap at all pressures below the mixed layer (Fig. 7). Turbulent kinetic energy dissipation in the (cyclonic eddy) core region is higher than the other regions (above 3  10  9 W kg  1) and outside the 95% confidence limits of the other regions at all pressures above 100 dbar (Fig. 7). At the euphotic depth the core region turbulent kinetic energy dissipation is 1.1  10  8 (95% confidence 0.9–1.3) W kg  1. For the core region, represented by a single station (11 casts), the confidence limits quoted are for averaging the station into 8 dbar vertical pressure intervals. The turbulent diffusivity for all regions is of similar magnitude at the euphotic depth, with turbulent diffusivity between 0.9 to 3  10  5 m2 s  1 (Fig. 8). The core region has

A. Forryan et al. / Deep-Sea Research I 63 (2012) 20–35

An area mean nutrient flux was calculated for the base of the euphotic zone from the area mean nutrient profiles (calculated as described in Section 2.4.3) and the area mean turbulent diffusivity using Eq. (5). The mean nitrate flux is 0.13 (95% confidence interval 0.08–0.22) mmol m  2 day  1, the silicate flux is 0.08 (0.05–0.12) mmol m  2 day  1 and, the phosphate flux is 8.6  10  3 (5.2–13.0) mmol m  2 day  1. An area mean profile of dissolved iron was constructed as described in Section 2.4.3 for all the published iron measurements (Nielsdo´ttir et al., 2009) from the cruise. An area mean dissolved iron flux was calculated for the base of the euphotic zone of 2.6 (95% confidence interval 4.3–1.3)  10  6 mmol m  2 day  1.

20

30

40

50 press (dbar)

29

4. Discussion

60

4.1. Nutrient concentrations 70

80 core background edges jet

90

100 10−6

10−5

10−4 K

10−3 (m2

10−2

10−1

s−1)

Fig. 8. Profiles of turbulent diffusivity (K) for the four regions (note the cyclonic eddy core region is represented by a single station 16285). Dashed lines mark the upper and lower 95% confidence limits for the regional profile. The depth of mixed layer is marked as a dashed red line (52 dbar for core, 30 dbar for the remaining regions). The euphotic depth is marked as a blue dashed line (64 dbar). (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

the lowest turbulent diffusivity at the euphotic depth, 8.8 (95% confidence interval: 7–11)  10  6 m2 s  1, which may be due to the close proximity of the mixed layer base to the euphotic depth in the core region (Fig. 8). Combining the results of all turbulence stations the area mean turbulent diffusivity for 65 dbar (64.4 m depth just below mean euphotic depth) is 0.21 (95% confidence interval: 0.17– 0.26)  10  4 m2 s  1 and for 33 dbar (32.7 m depth just below the mean mixed layer depth) is 0.13 (95% confidence interval: 0.09–0.21)  10  4 m2 s  1. 3.6. Turbulent nutrient fluxes Fluxes of nitrate at the base of the euphotic zone (64 dbar) vary between zero and 0.34 (95% confidence interval: 0.241– 0.469) mmol m  2 day  1. Fluxes of silicate at the base of the euphotic zone vary between 0.21 (0.018–0.023) and 0.094 (0.067–0.131) mmol m  2 day  1. Fluxes of phosphate at the base of the euphotic zone vary between zero and 0.035 (0.025–0.050) mmol m  2 day  1. There are only two turbulence stations with accompanying measurements of dissolved iron (stations 16260 and 16286). For both of these stations the dissolved iron flux at the base of the euphotic zone (64 dbar) is negative, i.e. downwards, with the fluxes for each station  5.2  10  7 (95% confidence interval 4.75 to 5.70) mmol Fe m  2 day  1 and  1.2  10  5 (9.0 to 1.5) mmol Fe m  2 day  1 respectively.

The observations presented here of nutrient concentrations for the Iceland Basin made during July and August are consistent with observations from previous studies made post-bloom at approximately the same time of year. Observations of surface macro-nutrient concentrations from the D321 survey area show silicate concentration of 0.5270.04 mmol m  3 (all stations mean 7standard error), nitrate concentration 3.77 70.14 mmol m  3 and phosphate concentration 0.3 70.01 mmol m  3. Previous studies report surface silicate concentrations of 0.9–1.2 mmol m  3 (Harris et al., 1997) in July with accompanying surface nitrate concentration of 1–2 mmol m  3 (Harris et al., 1997). Slightly earlier in May, at the tail end of the bloom, nutrient concentrations are observed to be higher with surface silicate concentrations of 2–6 mmol m  3 (Marra et al., 1995) and surface nitrate concentrations of 10–12 mmol m  3 (Marra et al., 1995). The observation that surface concentrations of all three macronutrients are higher within the D321 mode-water eddy than in the surrounding waters is also consistent with the observations of previous studies made during June and July. Within mode-water eddies near to the D321 survey site (591 100 N, 201 150 W, Law et al., 2001; Woodward and Rees, 2001); 601 N, 211 W, Jickells et al., 2008, surface silicate concentrations have been reported between 0.3 and 1 mmol m  3 (Jickells et al., 2008) and approximately 3.35 mmol m  3 (Woodward and Rees, 2001), surface nitrate concentrations between 7 and 8 mmol m  3 (Jickells et al., 2008) and between 4 and 6 mmol m  3 (Woodward and Rees, 2001), and surface phosphate concentrations between 0.3 and 1 mmol m  3 (Jickells et al., 2008) and approximately 0.46 mmol m  3 (Woodward and Rees, 2001). Macro-nutrient concentrations in the surrounding surface waters, recorded at the same time as the observations of nutrient concentration within a mode-water eddy, were observed to be between 0.05 and 0.09 mmol m  3 for silicate (Woodward and Rees, 2001), between 6 and 8 mmol m  3 for nitrate (Woodward and Rees, 2001) and approximately 0.29 mmol m  3 for phosphate (Woodward and Rees, 2001). The ratio of nitrate and silicate concentrations to phosphate concentration is lower in the surface waters than at depth in all cases (Table 3). The area mean molar ratio of Si:N:P reduces from  8 : 16 : 1 at 610 dbar to  2 : 13 : 1 for the euphotic zone. This suggests that there has been a removal of both nitrate and silicate in the surface waters. Observed surface concentrations of silica are below  2 mmol m3 which is considered to be limiting to diatom growth (Dugdale and Wilkerson, 1998). This would suggest that the observations presented here are consistent with post diatom bloom conditions. There is a large amount of variability associated with the published surface concentrations of dissolved iron (dFe), with

30

A. Forryan et al. / Deep-Sea Research I 63 (2012) 20–35

surface concentrations of between o0:010 2 0:218  103 mmol m3 for the data used here (Nielsdo´ttir et al., 2009). However, there is considered to be a strong relationship between dFe and nitrate concentrations below the seasonal nitrocline (Johnson et al., 1997). Previous studies of iron biogeochemistry suggest a ratio of 1  10  5:1 between dFe and nitrate (Johnson et al., 1997; Fung et al., 2000). This is half of the value calculated here from the data in Nielsdo´ttir et al. (2009) (2.070.3  10  5:1). 4.2. Turbulent mixing There appears to be little, if any, measurable variation in turbulent diffusivity at all pressures between the four regions identified around the eddy dipole. Of the four regions sampled, only the core of the cyclonic eddy shows any consistent deviations outside the 95% confidence interval of the area mean values with higher turbulent kinetic energy dissipation and higher buoyancy frequency. However, the resultant turbulent diffusivity in the cyclonic eddy core at the euphotic depth is of similar magnitude to the area mean due to the greater dampening effect of the elevated buoyancy frequency. This would suggest that the area mean profile of turbulent diffusivity, being consistent with both the regional profiles and the individual station profiles, is representative of the area as a whole within the confidence limits. The spatial consistency of the turbulent diffusivity appears initially surprising given the strong horizontal variation in water speeds, from o 10 cm s1 in the core of the eddies to  1 ms1 in the region between the two eddies. However, the observations here are consistent with microstructure observations made near western boundary currents, e.g. the Gulf Stream, which show only moderate levels of mixing of similar magnitude to that measured in the ocean main thermocline (Gregg and Sanford, 1980; Oakey and Elliott, 1977). The value of the area mean turbulent diffusivity reported here of 0.21 (95% confidence interval: 0.17–0.26)  10  4 m2 s  1 for the base of the euphotic zone is comparable to those reported within a mode-water eddy core in the Sargasso Sea of 0.3570.05  10  4 m2 s  1 (Ledwell et al., 2008), measured by tracer release at the base of euphotic zone, and consistent with values reported elsewhere for the open ocean of between 0.1270.02  10  4 m2 s  1 and 0.1770.02  10  4 m2 s  1 (Ledwell et al., 1998), measured using tracer release at 300 m depth for the south eastern part of the subtropical gyre in the North Atlantic. The value of the area mean turbulent diffusivity reported here for the base of the mixed layer is 0.13 (95% confidence interval: 0.09–0.21)  10  4 m2 s  1 is lower than recorded in previous studies from the Iceland basin, where turbulent diffusivity has been reported to be between 0.9770.3  10  4 m2 s  2 (Jickells et al., 2008) and 1.51 70.29  10  4 m2 s  2 (Law et al., 2001). In both cases the turbulent diffusivity was measured by tracer release at the base of the mixed layer,  15 m depth, within the core of mode-water eddies located near the D321 survey site (locations given above). In the measurements reported here, for all regions, the turbulent diffusivity at the base of the mixed layer is consistently lower that the recorded at the euphotic depth (Fig. 8). This is due to the increased buoyancy associated with the strong density gradients at the base of the mixed layer (Fig. 6) which suppress mixing. The tracer release technique allows the calculation of time and space integrated estimates of mixing which can reduce the statistical uncertainties of instantaneous measurements. Turbulent diffusivity measured by tracer release compares favourably with diffusivity calculated using microprofilers for measurements in the deep ocean (Ledwell et al., 2000; Polzin et al., 1997). However, during this study no measurements of turbulent diffusivity were taken within the core of the mode-water eddy part of

the dipole (see Section 2.7) and as a result more direct comparison with previous studies cannot be made. The mixed layer and the region directly below (up to 20 m deeper) can be subject to wind generated inertial motions resulting in high, time varying, vertical shears (D’Asaro, 1985). Such time varying vertical shears may likewise result in time varying levels of turbulent diffusivity at the base of the mixed layer. Near-inertial internal waves can break in the high buoyancy region below the mixed layer generating enhanced regions of time-variant turbulent kinetic energy dissipation and potentially enhanced mixing (Gregg et al., 1986). The enhanced mixing generated by inertial features can persist for several days, but is limited horizontally to the regions near the inertial feature and vertically to within a few meters of the base of the mixed layer (Gregg et al., 1986; D’Asaro, 1985). In the presence of horizontal gradients in density and vorticity on the order of 10 km, the surface mixed layer can be subject to a range of sub-mesoscale physical processes, such as strain driven frontogenesis (Hoskins and Bretherton, 1972), mixed layer instability (Boccaletti et al., 2007), loss of geostrophic balance (Molemaker et al., 2005), and wind-frontal interactions (Thomas and Lee, 2005). These processes may result in potentially high vertical velocities at the base of the mixed layer and enhanced convective mixing. Such sub-mesoscale physical processes are both temporally and spatially highly heterogeneous. As with wind generated inertial motions, the effects of these sub-mesoscale physical processes are limited to the maximum depth of wind penetration, likely to be within a few meters of the base of the mixed layer (Thomas and Lee, 2005; Boccaletti et al., 2007). The inertial period for the Iceland Basin is approximately 14 h. The relatively instantaneous (station duration  1 h) measurement technique used here is capable of detecting enhanced mixing due to inertial features and sub-mesoscale processes if deployed where and when the mixing is occurring. However, it is unlikely that the coarse temporal and spatial resolution of the D321 turbulence measurements would be able to adequately resolve any enhanced mixing due to inertial or sub-mesoscale processes over an inertial period. Hence it is possible that the increased levels of turbulent diffusivity reported previously for the Iceland Basin (Jickells et al., 2008; Law et al., 2001) are as a result of the effects of short lived wind driven inertial motions or sub-mesoscale processes in the surface layer. 4.3. Turbulent nutrient flux The turbulent macro-nutrient fluxes into the euphotic zone calculated here, of 0.13 (95% confidence interval 0.08–0.22) mmol m  2 day  1 for nitrate, 0.08 (0.05–0.12) mmol m  2 day  1 for silicate and 8.6  10  3 (13.0 to 5.2) mmol m  2 day  1 for phosphate, are approximately an order of magnitude lower that those previously reported for the Iceland Basin. Previous studies, calculating fluxes into the mixed layer, have recorded nitrate fluxes of 1.8 (Law et al., 2001) and 1.5 mmol m  2 day  1 (Jickells et al., 2008), silicate fluxes of 0.9 mmol m  2 day  1 (Jickells et al., 2008) and phosphate fluxes of 1.25 mmol m  2 day  1 (Law et al., 2001). Previous studies report nutrient gradients at the base of the mixed layer of 107:2 mmol m4 for nitrate, 34:3 mmol m4 for silicate and 7:41 mmol m4 for phosphate (Law et al., 2001). These gradients are of the same magnitude as the area mean nutrient gradients observed here at the euphotic depth, of 49 mmol m4 for nitrate, 32 mmol m4 for silicate and 2:5 mmol m4 for phosphate. This would suggest that the order of magnitude differences in the nutrient fluxes are mainly due to the difference in the observed turbulent diffusivity and not only to differences in nutrient concentrations. Turbulent fluxes into the mixed layer estimated here show a much greater variability than those into the euphotic zone (Fig. 9).

A. Forryan et al. / Deep-Sea Research I 63 (2012) 20–35

press (dbar)

Nitrate

Silicate

20

20

40

40

60

60

80

80

100

0

1

0.5

1.5

2

100

−0.2

−0.1

press (dbar)

Phosphate 20

40

40

60

60

80

80

0

0.01

0.02

0.03

0.04

0

0.1

0.2

0.3

dFe

20

100

31

0.05

0.06

(mmol m−2 day−1)

100 −5

0 (mmol m−2 day−1)

5 x 10−5

Fig. 9. Area mean profiles of nutrient flux for nitrate, silicate, phosphate and dissolved iron. Fluxes were calculated from the area mean nutrient profile (Section 2.4.3) and the area mean turbulent diffusivity profile (Section 2.7). Dashed lines mark the upper and lower 95% confidence limits. The area mean depth of mixed layer is marked as a dashed red line (30 dbar). The euphotic depth is marked as a blue dashed line (64 dbar). (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

This reflects the greater variability in the turbulent flux at the base of the mixed layer. The mixed layer fluxes are consistently more than twice the fluxes reported for the base of the euphotic zone, reflecting partly the steeper nutrient concentration gradient across the base of the mixed layer (Fig. 4). However, the mixed layer fluxes are still approximately an order of magnitude lower than those reported previously (Jickells et al., 2008; Law et al., 2001). There is a high degree of uncertainty associated with the turbulent flux of dFe, estimated to be 2.6 (95% confidence interval 4.3–1.3)  10  6 mmol m2 day  1. This is due largely to the amount of variability in the observed vertical profiles for dissolved iron. Gradients in iron concentration across the euphotic depth are positive and negative for different profiles. The area mean vertical profile for dissolved iron suggests a concentration minima between the mixed layer and the euphotic depth though this region is poorly resolved in the iron concentration measurements. Dissolved iron profiles are considered to be constant at depth with dissolved iron concentrations of  0:6 nM below 500 m (505.5 dbar in the Iceland basin) in all ocean waters away from the continental shelf (Johnson et al., 1997). Dissolved iron concentrations for the Iceland Basin used here are  0:4 nM at pressures of  400–600 dbar increasing to 4 0:6 nM at pressures 41000 dbar (Nielsdo´ttir et al., 2009). Deep winter mixing in the region penetrates to approximately 600 m (Section 4.6) which suggests that dissolved iron concentrations above 1000 dbar are reduced by mixing with lower concentration dissolved iron waters from the near surface during winter. There is potentially greater variability in surface concentrations of dissolved iron where the resultant surface concentration of iron is a balance between supply and utilisation (Luther III and Wu, 1997). The major supply route for iron into the surface waters is considered to be through aeolean deposition (Fung et al., 2000), while the majority of the chemical processes responsible

for the production of dissolved iron occur in sunlit surface waters (Weber et al., 2005). This would suggest that a combination of production of dissolved iron in the sunlit upper mixed layer combined with aeolean deposition might result in locally higher concentrations of dissolved iron in the mixed layer compared to those immediately below the euphotic depth. The presence of poorly resolved vertical profiles not capable of correctly representing the concentration minima in the dissolved iron profiles above the euphotic depth may well be the cause of the negative concentration gradients for some of the dissolved iron stations which contributes to the high levels of uncertainty associated with the turbulent iron flux.

4.4. Nitrate uptake The mean uptake of nitrate, integrated over the surface 45 m (45.5 dbar pressure, initial shipboard estimate for the depth for the euphotic zone), for the D321 survey area was approximately 2 mmol m  2 day  1 (M. Lucas, personal communication). Hence, the turbulent nitrate flux is  7% of the observed nitrate uptake. Along with the observed vertical turbulent nitrate flux, nitrate uptake is smaller than previously observed. Mean nitrate uptake 1 rates of  6:7 mmol m2 day (Jickells et al., 2008) and 8.1 mmol m  2 day  1, (Woodward and Rees, 2001) both integrated over the surface 30 m have been reported in the core of mode-water eddies located near to the D321 survey site (location given above). In both cases the coincident vertical turbulent nitrate flux reported was  22% of the reported nitrate uptake (Jickells et al., 2008; Law et al., 2001; Woodward and Rees, 2001). Taking the figures for turbulent nitrate flux reported here in combination with all observations of nitrate uptake suggests that vertical turbulent flux in the region represents between 1 and 10% of the nitrate requirements in the euphotic zone.

32

A. Forryan et al. / Deep-Sea Research I 63 (2012) 20–35

4.5. Phytoplankton stoichiometry Considering the concentrations of nitrate and silicate relative to the concentration of phosphate for all samples from below the euphotic depth (Fig. 10), the concentrations of both silicate and nitrate increase in approximately constant proportion to the concentration of phosphate, which suggests remineralisation is occurring in the ratio of approximately 16:16:1 (Si:N:P Fig. 10). This would suggest a stoichiometric composition for the Iceland Basin phytoplankton community which is consistent with the stoichiometric composition of marine diatoms (Brzezinski, 1985). In a similar manner, the relative concentrations of dFe and phosphate for all samples below the euphotic depth suggests a stoichiometric composition ratio for the Iceland Basin phytoplankton community of approximately 1  10  3:1 (dFe:P, Fig. 10). Marine phytoplankton have differing requirements for dFe with green algae, particularly the chlorophytes, having a relatively high dFe requirement compared with diatoms and coccolithophores (Ho et al., 2003). The dFe:P stoichiometric ratio estimated here of approximately 1  10  3:1 is at the lower end of the published estimates of between 1 and 15  10  3:1 (Ho et al., 2003). However, this mean value was estimated for a mixture of phytoplankton from both the open ocean and coastal or estuarine waters. Coastal and estuarine phytoplankton are considered to have higher trace metal requirements than open ocean species with the dFe:P stoichiometric ratio for the coastal diatom T. halassiosira weissflogii being 1.7  10  3:1 compared to a dFe:P stoichiometric ratio of 0.3  10  3:1 for

NO3 (mmol m−3)

The ratio of turbulent macro-nutrient fluxes into the euphotic zone, 9Si:15N:1P, is approximately equal to the ratio of the concentrations of the nutrients at depth (610 dbar, Table 3). The ratio of dFe turbulent flux to phosphate turbulent flux is 3  10  4:1. Hence the vertical turbulent flux is lacking in both dFe and silicate when compared to the elemental composition of the Iceland Basin phytoplankton community based on remineralisation. This, in combination with the observations of nitrate uptake given above, suggests that unlike in the sub-tropical gyre where turbulent nitrate flux and production appear to be closely linked (Lewis et al., 1986) this is not the case in the Iceland Basin the flux is accompanied by too little silicate. Primary production involving nitrate uptake in the Irminger Basin (approximately 21 North of the D321 survey site) is considered to be negligible after August (Sanders et al., 2005), typical of post-bloom conditions. However the steepest vertical nutrient gradients are associated with post-bloom conditions when the surface nutrient concentrations are at their lowest (Sanders et al., 2005). Hence, the turbulent supply of nutrients into the Iceland Basin euphotic zone is potentially at its largest at

NO = 16.3 PO −0.97 (R = 0.97)

20 15 10 5 0.4

0.5

0.6

0.7

0.8

0.9

1

1.1

1.2

1.3

1.4

1

1.1

1.2

1.3

1.4

1

1.1

1.2

1.3

1.4

Silicate

25 SiO3 (mmol m−3)

4.6. Annual nutrient supply

Nitrate

25

SiO = 15.95 PO −8.33 (R = 0.83)

20 15 10 5 0 0.4 10

dFe (mmol m−3)

Ditylum brightwellii an open-ocean diatom (Ho et al., 2003). Although there is considerable uncertainty this would suggest that the stoichiometric ratio estimated here may well be reasonable for an open-ocean diatom dominated regime.

0.5

0.6

0.7

x 10−4

0.8

0.9

Dissolved iron

dFe = 9.45x10

PO − 6.1x10

(R = 0.55)

5

0

−5 0.4

0.5

0.6

0.7

0.8

0.9

PO4 (mmol

m−3)

Fig. 10. Scatterplots of nutrient concentrations, nitrate vs phosphate (upper panel) silicate vs phosphate (middle panel) and dFe vs phosphate for all samples taken at depths below the base of the euphotic zone (64 dbar). A line of best fit, calculated by minimising least square differences is shown in red, with 7standard estimates of error marked as dashed red lines. The equation of the line of best fit and the R2 value for the fit are indicated. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

A. Forryan et al. / Deep-Sea Research I 63 (2012) 20–35

the time of the survey. This would suggest, from the fluxes reported in here, a potential maximum annual turbulent nutrient supply into the euphotic zone in the Iceland Basin of the order of 48 mmol m  2 year  1 for nitrate, 28 mmol m  2 year  1 for silicate and, 3 mmol m  2 year  1 for phosphate. Using the combined area mean dissolved iron profile and the area mean estimate of turbulent diffusivity, turbulent mixing in the Iceland Basin is estimated to supply 1 mmol Fe m2 year1 (95% confidence interval 0.2–2) of dissolved iron into the euphotic zone. This is consistent with an estimate of a dissolved iron flux of 0:96 mmol Fe m2 year1 calculated from the nitrate flux and the dissolved Fe:NO3 concentration ratio of 2  10  5:1 for water depths below the euphotic zone. Modelling the supply of nutrients from deep wintertime mixing in the North Atlantic sub-polar gyre gives an estimate of 1.470.2 mol m  2 for the annual supply of nitrate (Williams et al., 2000). Modelling is likely to overestimate of the convective supply of nutrients due to both inaccuracy in determining the depth of the wintertime mixed layer (Fox-Kemper et al., 2008; Fox-Kemper and Ferrari, 2008) and the use of climatological nutrient concentrations. We can be more specific for the Iceland Basin. Mixed layer depths calculated from observations from ARGO floats (BODC, 2011) in the North Atlantic (Section 2.3) suggest wintertime mixing during 2007 penetrated to a depth of  600 m. Hence, waters at depths of  600 m, with an observed nitrate concentration of  16 mmol m3 , are representative of the end of winter surface waters. Taking the surface nitrate concentration observed here of  4 mmol m3 as typical of the post bloom residual nitrate concentration that persists up to the start of winter mixing, a summertime mixed layer of between 30 and 40 m depth with an initial nitrate concentration of  16 mmol m3 contains a volume integrated total of between 360 and 480 mmol m  2 of nitrate supplied by deep wintertime mixing. This would suggest that turbulent mixing in the Iceland Basin provides a supply of nitrate equivalent to between 10 and 13% of the convective nitrate supply. By a similar argument deep winter mixing is estimated to supply between 240 and 320 mmol m  2 of silicate and between 21 and 28 mmol m  2 of phosphate assuming the post bloom residual concentration of silicate is close to zero and of phosphate is  0:3 mmol m2 . The turbulent flux of silicate is estimated to be between 9 and 12% of the convective silicate supply and of phosphate to be between 10 and 14% of the convective phosphate supply. The observed concentration of dFe at 600 dbar pressure (593.5 m depth) is 4.2  10  4 mmol m  3. Hence the estimated deep winter supply of dFe, using the same method as above, is between 13 and 17 mmol m2 . The turbulent supply of 1 mmol Fe m2 year1 represents an additional source of iron of between 6 and 8% of the estimated convective supply. Atmospheric deposition is estimated to add a further  5 mmol m2 year1 of iron (Nielsdo´ttir et al., 2009) giving the estimated total annual supply of dFe as between 19 and 23 mmol m2 year1 . Using the estimated stoichiometric ratios of Si:N:dFe:P of 16:16:1  10  3:1 the complete utilisation by phytoplankton of 360–480 mmol m  2 year  1 of nitrate is calculated to require between 22 and 30 mmol m2 year1 of dFe while complete utilisation of silicate is calculated to require between 15 and 20 mmol m2 year1 of dFe. This would suggest that the annual supply of dFe is more likely to be sufficient for the complete utilisation of the convective silicate supply than for the convective supply of nitrate. This would lend support to the suggestion that conditions of iron limitation are possibly occurring from the latter part of the spring bloom onward. The potential for iron limitation in the Iceland Basin post spring bloom has previously been demonstrated using in vitro iron addition experiments (Nielsdo´ttir et al., 2009).

33

5. Conclusions The area mean turbulent diffusivity reported here for the Iceland Basin is comparable to expected open ocean background levels (Ledwell et al., 1998) and turbulent diffusivity reported for a mode-water eddy core in the Sargasso Sea (Ledwell et al., 2008), but lower than the values reported for mode-water eddy cores in the Iceland Basin measured using tracer release techniques (Law et al., 2001; Jickells et al., 2008). The discrepancy between the results reported here and previous studies in the same area is potentially due to time variant mixing events caused by processes in the upper mixed layer which were not captured in measurements taken using a microstructure profiler. Investigation of the spatial distribution of turbulent diffusivity shows an almost uniform horizontal distribution of diffusivity across the survey area. This observation is quite surprising given the strong horizontal gradients in water velocity and density observed between the different regions. This suggests that such time variant mixing events are likely to be highly spatially or temporally heterogeneous. When compared to the convective supply of nutrient from deep winter mixing, observations of turbulent nutrient flux reported here would tend to support the view that, for the Iceland Basin, vertical turbulent flux is a minor source of nutrient into the surface waters (Williams and Follows, 2003; Williams et al., 2000). The magnitude of turbulent macro-nutrient flux is estimated to be at most 14% of the estimated supply of macronutrient by deep winter mixing in the region for silicate, nitrate and phosphate. Turbulent macro-nutrient fluxes calculated here are an order of magnitude lower than previous estimates for the region. This is due to the lower order of magnitude estimate of turbulent diffusivity. Observations of the vertical turbulent flux of iron into the surface waters of the Iceland Basin are, at best, tentative. The likely existence of an iron concentration minima between the base of the mixed layer and the base of the euphotic zone that is poorly resolved in the concentration measurements leads to an order of magnitude sized 95% confidence limit on dissolved iron fluxes. Deep winter mixing is estimated to supply an adequate amount of iron to utilise the deep winter mixed supply of silicate but not the deep winter mixed supply of nitrate. This would suggest that while the iron supply may not limit the magnitude of the spring bloom, iron limitation may be occurring post bloom. This would tend to support the proposal that the residual concentrations of nitrate detectable post bloom are due to iron limitation.

Acknowledgements The authors wish to thank the officers, crew, and entire scientific compliment aboard the R.R.S. Discovery during cruises D321 and the NERC Earth Observation Data Acquisition and Analysis Service for the satellite data. We are particularly grateful to Mike Lucas for calculating the nitrate uptake rates.

Appendix A. Assigning turbulence stations to regions The primary method of determining which region to assign a turbulence station to was by estimating the position of the station relative to the position of the eddies in the dipole. The positions of the two eddies were estimated for each of the three surveys of the region (Section 2.1) from recorded ADCP data (Section 2.7) and the positions of the individual turbulence stations plotted. Figs. A.11–A.13 show the ADCP current vectors, the estimated

34

A. Forryan et al. / Deep-Sea Research I 63 (2012) 20–35

Survey 3 (15th August − 22 nd August 2007)

60.2

60.2

60

60

59.8

59.8

59.6

59.6

Lat. (°N)

Lat. (°N)

Survey 1 (5th August − 10th August 2007)

59.4

59.4

59.2

59.2

59

59 58.8

58.8 −21.5

−21

−20.5

−20

−19.5 Lon. (°E)

−19

−18.5

−18

Fig. A.11. Acoustic Doppler Current Profile (ADCP) current vectors, the estimated position of the eddy cores, and the positions of the relevant turbulence stations colour coded depending on region for survey 1 (Section 2.1). The jet region is red, background, black, edge green and the core blue. The cyclone is marked in black and the mode-water eddy in red. The black dashed rectangle marks the boundary of the satellite images presented in Fig. 2.

−21.5

−21

−20.5

−20

−19.5 Lon. (°E)

−19

−18.5

−18

Fig. A.13. Acoustic Doppler Current Profile (ADCP) current vectors, the estimated position of the eddy cores, and the positions of the relevant turbulence stations colour coded depending on region for survey 3 (Section 2.1). The jet region is red, background, black, edge green and the core blue. The mode-water eddy in marked in red. There were insufficient ADCP measurements to accurately resolve the position of the cyclone. The black dashed rectangle marks the boundary of the satellite images presented in Fig. 2. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

Survey 2 (10th August − 15th August 2007) 60.2

for station 16286 showed a high speed ( 4 0:2 ms1 ) constant direction flow more consistent with a station in the jet region. Hence this station was assigned to the jet region.

60

Lat. (°N)

59.8

References

59.6 59.4 59.2 59 58.8 −21.5

−21

−20.5

−20

−19.5

−19

−18.5

−18

Lon. (°E) Fig. A.12. Acoustic Doppler Current Profile (ADCP) current vectors, the estimated position of the eddy cores, and the positions of the relevant turbulence stations colour coded depending on region for survey 2 (Section 2.1). The jet region is red, background, black, edge green and the core blue. The cyclone is marked in black and the mode-water eddy in red. The black dashed rectangle marks the boundary of the satellite images presented in Fig. 2. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

position of the eddy cores, and the positions of the relevant turbulence stations colour coded depending on region. In the case of survey 3 all stations (except 16283) appear to be in the jet region. However, there were insufficient ADCP measurements to accurately resolve the position of the cyclone. Consequently, the water movement while the turbulence stations were being conducted was also considered. Stations 16285 and 16286 were regarded in the cruise report as having been conducted within the cores of the cyclone and the mode-water eddies respectively. Examination of the water movement for station 16285 showed a low speed ( o 0:2 ms1 ) erratic flow consistent with being within an eddy core. Hence this station was assigned to the core region. However, examination of the water movement

Allen, J., 2007. Cruise Report No. 23 RSS Discovery Cruise 321 24th July–23 August 2007. Biophysical interactions in the Iceland Basin 2007. Technical Report. National Oceanography Centre, Southampton. Allen, J., Brown, L., Saunders, R., Moore, M., Mustard, A., Fielding, S., Lucas, M., Rixen, M., Savidge, G., Henson, S., Mayor, D., 2005. Diatom carbon export enhanced by silicate upwelling in the Northeast Atlantic. Nature 437, 728–732. Baker, M., Gibson, C., 1987. Sampling turbulence in the stratified ocean-statistical consequences of strong intermittency. J. Phys. Oceanography 17, 1817–1836. Boccaletti, G., Ferrari, R., Fox-Kemper, B., 2007. Mixed layer instabilities and restratification. J. Phys. Oceanography 37, 2228–2250. BODC, 2011. British oceanographic data centre ARGO float data. /https://www. bodc.ac.uk/data/online_delivery/argo/S. Boyd, P., Jickells, T., Law, C., Blain, S., Boyle, E., Buesseler, K., Coale, K., Cullen, J., de Baar, H., Follows, M., Harvey, M., Lancelot, C., Levasseur, M., Owens, N., Pollard, R., Rivkin, R., Sarmiento, J., Schoemann, V., Smetacek, V., Takeda, S., Tsuda, A., Turner, S., Watson, A., 2007. Mesoscale iron enrichment experiments 1993– 2005: synthesis and future directions. Science 315, 167–612. doi:10.1126/ science.1131669. Brzezinski, M.A., 1985. The si: c:n ratio of marine diatoms: Interspecific variability and the effect of some environmental variables. J. Phycol. 21, 347–357. D’Asaro, E., 1985. Upper ocean temperature structure, inertial currents, and Richardson Numbers observed during strong meteorological forcing. J. Phys. Oceanography 15, 943–962. Dewey, R., Crawford, W., 1988. Bottom stress estimates from vertical dissipation rate profiles on the continental shelf. J. Phys. Oceanography 18, 1167–1177. Dugdale, R., Wilkerson, F., 1998. Silicate regulation of new production in the equatorial pacifc upwelling. Nature 391, 270–273. Fox-Kemper, B., Ferrari, R., 2008. Parameterization of mixed layer eddies. Part II: prognosis and impact. J. Phys. Oceanography 38, 1166–1179. Fox-Kemper, B., Ferrari, R., Hallberg, R., 2008. Parameterization of mixed layer eddies. Part i: theory and diagnosis. J. Phys. Oceanography 38, 1145–1165. Fung, I., Meyn, S., Tegen, I., Doney, S., John, J., Bishop, J., 2000. Iron supply and demand in the upper ocean. Global Biogeochem. Cycles 14, 281–295. Gill, A., 1982. Atmosphere-Ocean Dynamics. Academic Press, London, United Kingdom. Gregg, M., D’Asro, E., Shay, T., Larson, N., 1986. Observations of persistent mixing and near-inertial internal waves. J. Phys. Oceanography 16, 856–885. Gregg, M., Sanford, T., 1980. Signatures of mixing from the Bermuda slope, the Sargasso sea and the Gulf stream. J. Phys. Oceanography 10, 105–127.

A. Forryan et al. / Deep-Sea Research I 63 (2012) 20–35

Harris, R., Boyd, P., Harbour, D., Head, R., Pingree, R., Pomroy, A., 1997. Physical, chemical and biological features of a cyclonic eddy in the region of 611100 N 191500 W in the North Atlantic. Deep-Sea Res. Part I 44, 1815–1839. Ho, T., Quigg, A., Finkel, Z., Milligan, A., Wyman, K., Falkowski, P., Morel, F., 2003. The elemental composition of some marine phytoplankton. J. Phycol. 39, 1145–1159. Hoskins, B., Bretherton, F., 1972. Atmospheric frontogenesis models: mathematical formulation and solution. J. Atmos. Sci. 29, 11–37. Jickells, T., et al., 2008. A lagrangian biogeochemical study of an eddy in the Northeast Atlantic. Prog. Oceanography 76, 366–398. Johnson, K., Gordon, R., Coale, K., 1997. What controls dissolved iron concentrations in the world ocean ? Marine Chem. 57, 137–161. Kara, A., Rochford, P., Hurlburt, H., 2000. An optimal definition for ocean mixed layer depth. J. Geophys. Res.—Oceans 105, 16803–16821. Kirkwood, D., 1996. Nutrients: Practical Notes on Their Determination in Seawater. Technical Report 17. International Council for the Exploration of the Seas. Klein, P., Lapeyre, G., 2009. The oceanic vertical pump induced by mesoscale turbulence. Annu. Rev. Marine Sci. 1, 351–375. Law, C., Martin, A., Liddicoat, M., Watson, A., Richards, K., Woodward, E., 2001. A lagrangian SF6 tracer study of an anticyclonic eddy in the North Atlantic: patch evolutional vertical mixing and nutrient supply to the mixed layer. Deep-Sea Res. Part II—Topical Studies Oceanography 48, 705–724. Ledwell, J., McGillicuddy Jr., D., Anderson, L., 2008. Nutrient flux into an intense deep chlorophyll layer in a mode-water eddy. Deep-Sea Res. Part II—Topical Studies Oceanography 55, 1139–1160. Ledwell, J., Montgomery, E., Polzin, K., St Laurent, L., Schmitt, R., Toole, J., 2000. Evidence for enhanced mixing over rough topography in the Abyssal ocean. Nature 403, 179–182. Ledwell, J., Watson, A., Law, C., 1998. Mixing of a tracer in the pycnocline. J. Geophys. Res.—Oceans 103, 21449–21529. Le´vy, M., Klein, P., Treguier, A., 2001. Impact of sub-mesoscale physics on production and subduction of phytoplankton in an oligotrophic regime. J. Marine Res. 59, 535–565. Lewis, M., Harrison, W., Oakey, N., Herbert, D., Platt, T., 1986. Vertical nitrate fluxes in the oligotrophic ocean. Science 234, 870–873. Lueck, R., Wolk, F., Yamazaki, H., 2002. Oceanic velocity microstructure measurements in the 20th century. J. Oceanography 58, 153–174. Luther III, G., Wu, J., 1997. What controls dissolved iron concentrations in the world ocean ? A comment. Marine Chem. 57, 173–179. Marra, J., Langdon, C., Knudson, C., 1995. Primary production, water column changes, and the demise of a phaeocystis bloom at the marine light mixed layers site (591N 210 W) in the Northeast Atlantic ocean. J. Geophys. Res. 100, 6633–6643. Martin, A., Pondaven, P., 2003. On estimates for the vertical nitrate flux due to eddy pumping. J. Geophys. Res. 108 (C11). doi:10.1029/2003JC001841. Martin, A., Richards, K., 2001. Mechanisms for vertical nutrient transport within a North Atlantic mesoscale eddy. Deep-Sea Res. Part II—Topical Studies Oceanography 48, 757–773. Martin, A., Wade, I., Richards, K., Heywood, K., 1998. The PRIME eddy. J. Marine Res. 56, 439–462. McGillicuddy, D., Anderson, L., Doney, S., Maltrud, M., 2003. Eddy driven sources and sinks of nutrients in the upper ocean: results from a 0.1 degrees resolution model of the North Atlantic. Global Biogeochem. Cycles 17. doi:10.1029/ 2002GB001987. Molemaker, M., McWilliams, J., Yavneh, I., 2005. Baroclinic instability and loss of balance. J. Phys. Oceanography 35, 1505–1517. Moum, J., Gregg, M., Lien, R., Carr, M., 1995. Comparison of turbulence kineticenergy dissipation rate estimates from 2 ocean microstructure profilers. J. Atmos. Ocean. Technol. 12, 346–366. Moum, J., Lien, R., Perlin, A., Nash, J., Gregg, M., Wiles, P., 2009. Sea surface cooling at the Equator by subsurface mixing in tropical instability waves. Nat. Geosci. 2, 761–765. Nagai, T., Tandon, A., Yamazaki, H., Doubell, M., 2009. Evidence of enhanced turbulent dissipation in the frontogenetic Kuroshio Font thermocline. Geophys. Res. Lett. 36, L12609. doi:10.1029/2009GL038832. Nielsdo´ttir, M., Moore, C., Sanders, R., Hinz, D., Achterberg, E., 2009. Iron limitation of the postbloom phytoplankton communities in the Iceland Basin. Global Biogeochem. Cycles 23, GB3001. doi:10.1029/2008GB003410. Oakey, N., 1982. Determination of the rate of dissipation of turbulent energy from simultaneous temperature and shear microstructure measurements. J. Phys. Oceanography 12, 256–271.

35

Oakey, N.S., Elliott, J.A., 1977. Vertical temperature gradient structure across the gulf stream. J. Geophys. Res. 82, 1369–1380. Osborn, T., 1980. Estimates of the local rate of vertical diffusion from dissipation measurements. J. Phys. Oceanography 10, 83–89. Oschlies, A., 2002. Can eddies make ocean deserts bloom? Global Biogeochem. Cycles 16. doi:10.1029/2001GB001830. Polzin, K., Toole, J., Ledwell, J., Schmitt, R., 1997. Spatial variability of turbulent mixing in the Abyssal ocean. Science 276, 93–96. Popova, E., Lozano, C., Srokosz, M., Fasham, M., Haley, P., Robinson, A., 2002. Coupled 3d physical and biological modelling of the mesoscale variability observed in the North-East Atlantic in spring 1997: biological processes. DeepSea Res. Part I—Oceanographic Res. Papers 49, 1741–1768. Prandke, H., 2007. Measurement of Small Scale Turbulence—Data Processing, Dissipation Rate Computation, and Error Discussion. Presentation. ISW Wassermesstechnik. Petersdorf. Prandke, H., 2008. MSSpro Microstructure Data Evaluation Tool User Manual. Technical Report. ISW Wassermesstechnik. Petersdorf. Prandke, H., Stips, A., 1998. Test measurements with an operational microstructure-turbulence profiler: Detection limit of dissipation rates. Aquatic Sci. 60, 191–209. Press, W., Flannery, B., Teukolsky, S., Vetterling, W., 1989. Numerical Recipes. The Art of Scientific Computing (FORTRAN Version). Cambridge University Press, Cambridge. Read, J., Pollard, R., 2001. A long-lived eddy in the iceland basin 1998. J. Geophys. Res. 106, 11411–11421. Rippeth, T., Simpson, J., Williams, E., Inall, M., 2003. Measurement of the rates of production and dissipation of turbulent kinetic energy in an energetic tidal flow: Red Wharf Bay revisited. J. Phys. Oceanography 33, 1889–1901. Rippeth, T., Wiles, P., Palmer, M., Sharples, J., Tweddle, J., 2009. The diapcynal nutrient flux and shear-induced diapcynal mixing in the seasonally stratified western Irish Sea. Continental Shelf Res. 29, 1580–1587. Rodrı´guez-Santana, A., Pelegrı´, J.L., Sangra , P., Marrero-Dı´az, A., 1999. Diapycnal mixing in gulf stream meanders. J. Geophys. Res. 104, 912. Rodrı´guez-Santana, A., Pelegrı´, J.L., Sangra , P., Marrero-Dı´az, A., 2001. On the relevance of diapycnal mixing for the stability of frontal meanders. Sci. Marina 65, 259–267. Roget, E., Lozovatsky, I., Sanchez, X., Figueroa, M., 2006. Microstructure measurements in natural waters: methodology and applications. Prog. Oceanography 70, 126–148. Sanders, R., Brown, L., Henson, S., Lucas, L., 2005. New production in the Irminger Basin during 2002. J. Marine Syst. 55, 291–310. Sarmiento, J., Gruber, N., Sundermann, J., 2006. Ocean Biogeochemical Dynamics. Princeton University Press, Princeton and Oxford. Stips, A., 2005. Dissipation measurement. In: Baumert, H., Simpson, J. (Eds.), Marine Turbulence: Theories, Observations, and Models. Results of the CARTUM Project. , Cambridge University Press, pp. 115–126. Tennekes, H., Lumley, J., 1972. A First Course in Turbulence. The MIT Press, Cambridge, Massachusetts and London, England. Thomas, L., Lee, C., 2005. Intensification of ocean fronts by down-front winds. J. Phys. Oceanography 35, 1086–1102. Thorpe, S., 1977. Turbulence and mixing in a Scottish loch. Philo. Trans. R. Soc. Lond. A 286, 125–181. Weber, L., Volker, C., Schartau, M., Wolf-Gladrow, D., 2005. Modeling the speciation and biogeochemistry of iron at the Bermuda Atlantic time-series study site nutrients and SPM. Global Biogeochem. Cycles 19, GB1019. doi:10.1029/ 2004GB002340. Welch, P., 1967. The use of Fast Fourier Transform for the estimation of power spectra: a method based on time averaging over short, modified periodograms. IEEE Trans. Audio Electroacoust. AU15, 70–73. Williams, R., Follows, M., 2003. Physical transport of nutrients and the maintenance of biological production. In: Fasham, M. (Ed.), Ocean Biogeochemistry: The role of the ocean carbon cycle in global change. , Springer, pp. 19–51. Williams, R., McLean, A., Follows, M., 2000. Estimating the convective supply of nitrate and implied variability in export production over the North Atlantic. Global Biogeochem. Cycles 14 (4), 1299–1313. Woodward, E., Rees, A., 2001. Nutrient distributions in an anticyclonic eddy in the Northeast Atlantic ocean, with reference to nanomolar ammonium concentrations. Deep-Sea Res. II 48, 775–793. Yamazaki, H., Osborn, T., 1990. Dissipation estimates for stratified turbulence. J. Geophys. Res.—Oceans 95, 9739–9744.