<?xml version="1.0" encoding="UTF-8"?>
<!DOCTYPE article PUBLIC "-//NLM//DTD Journal Publishing with OASIS Tables v3.0 20080202//EN" "journalpub-oasis3.dtd">
<article xmlns:xlink="http://www.w3.org/1999/xlink" xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:oasis="http://docs.oasis-open.org/ns/oasis-exchange/table" xml:lang="en" dtd-version="3.0">
  <front>
    <journal-meta><journal-id journal-id-type="publisher">BG</journal-id><journal-title-group>
    <journal-title>Biogeosciences</journal-title>
    <abbrev-journal-title abbrev-type="publisher">BG</abbrev-journal-title><abbrev-journal-title abbrev-type="nlm-ta">Biogeosciences</abbrev-journal-title>
  </journal-title-group><issn pub-type="epub">1726-4189</issn><publisher>
    <publisher-name>Copernicus Publications</publisher-name>
    <publisher-loc>Göttingen, Germany</publisher-loc>
  </publisher></journal-meta>
    <article-meta>
      <article-id pub-id-type="doi">10.5194/bg-15-2125-2018</article-id><title-group><article-title>OUTPACE long duration stations: physical variability, context of biogeochemical sampling, and evaluation of sampling strategy</article-title><alt-title>LD OUTPACE</alt-title>
      </title-group><?xmltex \runningtitle{LD OUTPACE}?><?xmltex \runningauthor{A.~de~Verneil et al.}?>
      <contrib-group>
        <contrib contrib-type="author" corresp="yes" rid="aff1 aff4">
          <name><surname>de Verneil</surname><given-names>Alain</given-names></name>
          <email>ajd11@nyu.edu</email>
        </contrib>
        <contrib contrib-type="author" corresp="no" rid="aff1">
          <name><surname>Rousselet</surname><given-names>Louise</given-names></name>
          
        </contrib>
        <contrib contrib-type="author" corresp="no" rid="aff1">
          <name><surname>Doglioli</surname><given-names>Andrea M.</given-names></name>
          
        <ext-link>https://orcid.org/0000-0003-1309-9954</ext-link></contrib>
        <contrib contrib-type="author" corresp="no" rid="aff1">
          <name><surname>Petrenko</surname><given-names>Anne A.</given-names></name>
          
        </contrib>
        <contrib contrib-type="author" corresp="no" rid="aff2">
          <name><surname>Maes</surname><given-names>Christophe</given-names></name>
          
        </contrib>
        <contrib contrib-type="author" corresp="no" rid="aff3">
          <name><surname>Bouruet-Aubertot</surname><given-names>Pascale</given-names></name>
          
        </contrib>
        <contrib contrib-type="author" corresp="no" rid="aff1">
          <name><surname>Moutin</surname><given-names>Thierry</given-names></name>
          
        <ext-link>https://orcid.org/0000-0003-1297-8893</ext-link></contrib>
        <aff id="aff1"><label>1</label><institution>Aix Marseille Univ., Université de Toulon, CNRS, IRD, MIO UM 110, 13288, Marseille, France</institution>
        </aff>
        <aff id="aff2"><label>2</label><institution>Université de Bretagne Occidentale (UBO), Ifremer, CNRS, IRD, Laboratoire d'Océanographie Physique et Spatiale (LOPS), IUEM, 29280, Brest, France</institution>
        </aff>
        <aff id="aff3"><label>3</label><institution>Sorbonne Université (UPMC, Univ Paris 06)-CNRS-IRD-MNHN, LOCEAN, Paris, France</institution>
        </aff>
        <aff id="aff4"><label>a</label><institution>now at: The Center for Prototype Climate Modeling, New York University in Abu Dhabi, Abu Dhabi, UAE</institution>
        </aff>
      </contrib-group>
      <author-notes><corresp id="corr1">Alain de Verneil (ajd11@nyu.edu)</corresp></author-notes><pub-date><day>10</day><month>April</month><year>2018</year></pub-date>
      
      <volume>15</volume>
      <issue>7</issue>
      <fpage>2125</fpage><lpage>2147</lpage>
      <history>
        <date date-type="received"><day>26</day><month>October</month><year>2017</year></date>
           <date date-type="rev-request"><day>2</day><month>November</month><year>2017</year></date>
           <date date-type="rev-recd"><day>14</day><month>March</month><year>2018</year></date>
           <date date-type="accepted"><day>15</day><month>March</month><year>2018</year></date>
      </history>
      <permissions>
        
        
      <license license-type="open-access"><license-p>This work is licensed under the Creative Commons Attribution 4.0 International License. To view a copy of this licence, visit <ext-link ext-link-type="uri" xlink:href="https://creativecommons.org/licenses/by/4.0/">https://creativecommons.org/licenses/by/4.0/</ext-link></license-p></license></permissions><self-uri xlink:href="https://bg.copernicus.org/articles/15/2125/2018/bg-15-2125-2018.html">This article is available from https://bg.copernicus.org/articles/15/2125/2018/bg-15-2125-2018.html</self-uri><self-uri xlink:href="https://bg.copernicus.org/articles/15/2125/2018/bg-15-2125-2018.pdf">The full text article is available as a PDF file from https://bg.copernicus.org/articles/15/2125/2018/bg-15-2125-2018.pdf</self-uri>
      <abstract>
    <p id="d1e155">Research cruises to quantify biogeochemical fluxes in the ocean require
taking measurements at stations lasting at least several days. A popular
experimental design is the quasi-Lagrangian drifter, often mounted with in
situ incubations or sediment traps that follow the flow of water over time. After initial
drifter deployment, the ship tracks the drifter for continuing measurements
that are supposed to represent the same water environment. An outstanding
question is how to best determine whether this is true. During the
Oligotrophy to UlTra-oligotrophy PACific Experiment (OUTPACE) cruise, from 18
February to 3 April 2015 in the western tropical South Pacific, three
separate stations of long duration (five days) over the upper 500 <inline-formula><mml:math id="M1" display="inline"><mml:mi mathvariant="normal">m</mml:mi></mml:math></inline-formula>
were conducted in this quasi-Lagrangian sampling scheme. Here we present
physical data to provide context for these three stations and to assess
whether the sampling strategy worked, i.e., that a single body of water
was sampled. After analyzing tracer variability and local water circulation
at each station, we identify water layers and times where the drifter risks
encountering another body of water. While almost no realization of this
sampling scheme will be truly Lagrangian, due to the presence of vertical
shear, the depth-resolved observations during the three stations show most
layers sampled sufficiently homogeneous physical environments during OUTPACE.
By directly addressing the concerns raised by these quasi-Lagrangian sampling
platforms, a protocol of best practices can begin to be formulated so that
future research campaigns include the complementary datasets and analyses
presented here to verify the appropriate use of the drifter platform.</p>
  </abstract>
    </article-meta>
  </front>
<body>
      

<sec id="Ch1.S1" sec-type="intro">
  <title>Introduction</title>
      <p id="d1e172">Biogeochemical cycles dictate the global distribution and fluxes of the
chemical elements. Quantifying the mechanisms that mediate the various forms
key elements take in these cycles, especially in the midst of ongoing climate
change in the ocean, is vital to understanding the future evolution of the
Earth system <xref ref-type="bibr" rid="bib1.bibx15 bib1.bibx9 bib1.bibx18" id="paren.1"/>. Considering the wide diversity of environments where
biogeochemical processes take place, it is not surprising that each
sub-discipline has its own challenges with regards to collecting and
processing samples. The sampling protocols put in place thus need to ensure
the mechanisms of interest are isolated and put into their proper context.</p>
      <?pagebreak page2126?><p id="d1e178">In the world's surface oceans, a dominant difficulty is the medium itself:
water. Sampling in a fluid that is always liable to move normally requires
that one of two approaches be taken. In the first approach, a geographic
location is chosen and then repeatedly sampled. This produces an Eulerian
perspective, and this methodology is employed by definition at permanent
mooring platforms. Set geographic locations are also often used to define
time series or recurrent sampling locations, for example stations ALOHA,
BATS, CalCOFI, DYFAMED, and PAPA <xref ref-type="bibr" rid="bib1.bibx19 bib1.bibx40 bib1.bibx43 bib1.bibx3 bib1.bibx28 bib1.bibx17" id="paren.2"/>. These sites can be combined into worldwide networks and
initiatives such as OceanSITES <xref ref-type="bibr" rid="bib1.bibx41" id="paren.3"/>. While this
strategy makes no attempt to actually follow a given water parcel, if
currents are relatively weak during a single field campaign then the
variability due to advection can be ignored. Unfortunately, the
spatio-temporal scales of shipboard station sampling (in time, days to weeks;
in space, 1–100 <inline-formula><mml:math id="M2" display="inline"><mml:mi mathvariant="normal">km</mml:mi></mml:math></inline-formula>) overlap with a multitude of physical phenomena
ubiquitously found in the ocean, ranging from internal waves, submesoscale
turbulence, up to mesoscale eddies <xref ref-type="bibr" rid="bib1.bibx13" id="paren.4"/>. All of these
motions can easily transport water such that instantaneously observed
temperature, salinity, and by extension the organisms and chemical
environments mediating biogeochemical processes, are markedly different from
some mean value or state.</p>
      <p id="d1e197">One way to rectify physical displacements is the second sampling approach,
namely to follow the water during ongoing experiments. This approach creates
a Lagrangian point of view. A common implementation of this strategy is with
quasi-Lagrangian drifting moorings <xref ref-type="bibr" rid="bib1.bibx22 bib1.bibx32" id="paren.5"/>. These drifters are structured so that a vertical
line with sampling devices (e.g., incubation bottles and/or sediment traps)
drifts along with the flow. This approach has been in routine use for decades
across the globe; some examples of French campaigns known to the authors
include the OLIPAC (1994), PROSOPE (1999), BIOSOPE (2004), and BOUM (2008)
experiments (all data and metadata accessible from the LEFE CYBER website, <uri>http://www.obs-vlfr.fr/proof/cruises.php</uri>, last access: 6 April 2018).</p>
      <p id="d1e206">Naturally, the question arises whether the trajectory undertaken by the
drifting mooring in the quasi-Lagrangian approach accurately represents the
water movements at each of the sampling sites. If the drifter is successful
in following the water, then indeed a single biogeochemical setting will have
been sampled; if it is not successful, then the risk grows that a different
environment has been brought in via advection. Previous efforts by physicists
to make floats Lagrangian show the effort needed to make an instrument
neutrally buoyant, and these floats have been instrumental in demonstrating
complicated flow regimes <xref ref-type="bibr" rid="bib1.bibx14" id="paren.6"/>. In contrast, the
quasi-Lagrangian platform, with a variable distribution of incubation
bottles, will necessarily fail to be Lagrangian in finite time outside of a
barotropic flow where currents are the same throughout the water column. As a
result, ensuring the success of this strategy requires taking into account
how different currents potentially shorten the timespan of validity. In
fast-moving flows with strong vertical shear and possible vertical motions,
such as zones of enhanced mesoscale turbulence near fronts and filaments, the
drifter will not be Lagrangian for long. Alternatively, if a drifter is trapped
inside a coherent eddy, it can follow a similar water mass for a long time
over great distances. Therefore, before deployment the selection of sampling
sites needs to be carefully considered. Unless the focus of study, fronts and
filaments need to be avoided because shearing will quickly separate water
parcels at different depths in the direction of the structure's alignment;
finding signs of their presence has become more feasible with satellite data.
An eddy can be targeted because of its coherence, and there are ways to
confirm that sampling is indeed inside of it <xref ref-type="bibr" rid="bib1.bibx31" id="paren.7"/>. In
other words, if a physical structure is targeted or identified, its
particular nature supersedes other considerations. These structures are not
necessarily representative of the World Ocean, and so many biogeochemical
measurements will be taken elsewhere. For the campaigns where sites are far
(possibly by design) from obvious, organized mesoscale structures, there is
still a need to conduct an independent, post-cruise validation of the
drifter's success, which is the focus of the present study.</p>
      <p id="d1e216">Before proceeding into the description of our methodology, a few
remarks are needed regarding its applicability. We already mentioned that we
will consider regions away from strong, organized mesoscale structure.
Additionally, the method relies upon independent physical, not
biogeochemical, measurements to indicate a change of water mass due to the
drifter not being Lagrangian. This approach does not detect the existence of
biogeochemical gradients in water that might exist on smaller scales, so
application of our method requires the user to apply contextual knowledge of
their sampling region and keep this possibility in mind. For this study, a
regional biogeochemical gradient was expected <xref ref-type="bibr" rid="bib1.bibx33" id="paren.8"/> and
rationales for this method's application will be provided.</p>
      <p id="d1e222">The Oligotrophy to UlTra-oligotrophy PACific Experiment (OUTPACE) cruise
provided an opportunity to assess the success of the quasi-Lagrangian
sampling strategy. Conducted from 18 February to 3 April 2015 in the western
tropical South Pacific (WTSP), one of the goals of OUTPACE was to assess the
regional contribution of nitrogen fixation as a biogeochemical process to the
biological carbon pump <xref ref-type="bibr" rid="bib1.bibx33" id="paren.9"/>. During the cruise, three long
duration (LD) stations employed the quasi-Lagrangian strategy. In the
subsequent discourse regarding these stations, we proceed as follows. Section 2
describes how the drifting mooring was deployed, our methodological strategy,
how concurrent data were collected, and the analyses undertaken to answer our
central question of whether we sampled a single environment. We then present
the data and results in Sect. 3, followed by a discussion in Sect. 4. The
paper finishes in Sect. 5 with a summary of our recommendations regarding
future implementations of this sampling strategy.</p>

      <?xmltex \floatpos{t}?><fig id="Ch1.F1" specific-use="star"><caption><p id="d1e230">Satellite surface <bold>(a)</bold> chl <italic>a</italic> and <bold>(b)</bold> SST for the OUTPACE
cruise. Pixel data are weighted by the normalized inverse distance squared
between each pixel and the RV <italic>L'Atalante</italic>'s daily position over the
42 <inline-formula><mml:math id="M3" display="inline"><mml:mi mathvariant="normal">days</mml:mi></mml:math></inline-formula> of OUTPACE. Ship track shown in white. LD station locations
shown with black <inline-formula><mml:math id="M4" display="inline"><mml:mo>+</mml:mo></mml:math></inline-formula>'s. Domains used in Fig. 2 are shown by color-coded
rectangles, with green for LDA, red for LDB, and blue for LDC.</p></caption>
        <?xmltex \igopts{width=455.244094pt}?><graphic xlink:href="https://bg.copernicus.org/articles/15/2125/2018/bg-15-2125-2018-f01.pdf"/>

      </fig>

<?xmltex \floatpos{t}?><table-wrap id="Ch1.T1" specific-use="star"><caption><p id="d1e269">Start and stop times for the time series data in each LD station of
OUTPACE. Times expressed in DD/MM/YYYY HH:MM:SS (GMT) format. When multiple
instruments were deployed or multiple discrete observations made, this is
also noted.</p></caption><oasis:table frame="topbot"><oasis:tgroup cols="5">
     <oasis:colspec colnum="1" colname="col1" align="left"/>
     <oasis:colspec colnum="2" colname="col2" align="left"/>
     <oasis:colspec colnum="3" colname="col3" align="left"/>
     <oasis:colspec colnum="4" colname="col4" align="left"/>
     <oasis:colspec colnum="5" colname="col5" align="left"/>
     <oasis:thead>
       <oasis:row rowsep="1">
         <oasis:entry colname="col1">STATION</oasis:entry>
         <oasis:entry colname="col2"/>
         <oasis:entry colname="col3"/>
         <oasis:entry colname="col4"/>
         <oasis:entry colname="col5"/>
       </oasis:row>
     </oasis:thead>
     <oasis:tbody>
       <oasis:row>
         <oasis:entry colname="col1">LDA</oasis:entry>
         <oasis:entry rowsep="1" colname="col2">Instrument</oasis:entry>
         <oasis:entry rowsep="1" colname="col3">CTD rosette</oasis:entry>
         <oasis:entry rowsep="1" colname="col4">SVP</oasis:entry>
         <oasis:entry rowsep="1" colname="col5">AQUADOPP</oasis:entry>
       </oasis:row>
       <oasis:row>
         <oasis:entry colname="col1">(19.21<inline-formula><mml:math id="M5" display="inline"><mml:msup><mml:mi/><mml:mo>∘</mml:mo></mml:msup></mml:math></inline-formula> S,</oasis:entry>
         <oasis:entry colname="col2">Number</oasis:entry>
         <oasis:entry colname="col3">46 casts</oasis:entry>
         <oasis:entry colname="col4">3 drifters</oasis:entry>
         <oasis:entry colname="col5">6 deployed</oasis:entry>
       </oasis:row>
       <oasis:row>
         <oasis:entry colname="col1">164.69<inline-formula><mml:math id="M6" display="inline"><mml:msup><mml:mi/><mml:mo>∘</mml:mo></mml:msup></mml:math></inline-formula> E)</oasis:entry>
         <oasis:entry colname="col2">Start</oasis:entry>
         <oasis:entry colname="col3">25/02/2015 14:09:18</oasis:entry>
         <oasis:entry colname="col4">25/02/2015 20:00:00</oasis:entry>
         <oasis:entry colname="col5">26/02/2015 22:40:00</oasis:entry>
       </oasis:row>
       <oasis:row>
         <oasis:entry colname="col1"/>
         <oasis:entry colname="col2">Stop</oasis:entry>
         <oasis:entry colname="col3">02/03/2015 16:10:10</oasis:entry>
         <oasis:entry colname="col4">02/03/2015 22:00:00</oasis:entry>
         <oasis:entry colname="col5">02/03/2015 16:10:00</oasis:entry>
       </oasis:row>
       <oasis:row>
         <oasis:entry colname="col1"/>
         <oasis:entry rowsep="1" colname="col2">Duration</oasis:entry>
         <oasis:entry rowsep="1" colname="col3">5 days 2 h 0 min 52 s</oasis:entry>
         <oasis:entry rowsep="1" colname="col4">5 days 2 h</oasis:entry>
         <oasis:entry rowsep="1" colname="col5">3 days 17 h 30 min</oasis:entry>
       </oasis:row>
       <oasis:row>
         <oasis:entry colname="col1"/>
         <oasis:entry rowsep="1" colname="col2">Instrument</oasis:entry>
         <oasis:entry rowsep="1" colname="col3">SADCP 150</oasis:entry>
         <oasis:entry rowsep="1" colname="col4">SADCP 38</oasis:entry>
         <oasis:entry rowsep="1" colname="col5">SedTrap position</oasis:entry>
       </oasis:row>
       <oasis:row>
         <oasis:entry colname="col1"/>
         <oasis:entry colname="col2">Start</oasis:entry>
         <oasis:entry colname="col3">25/02/2015 14:09:57</oasis:entry>
         <oasis:entry colname="col4">25/02/2015 14:10:26</oasis:entry>
         <oasis:entry colname="col5">25/02/2015 19:01:13</oasis:entry>
       </oasis:row>
       <oasis:row>
         <oasis:entry colname="col1"/>
         <oasis:entry colname="col2">Stop</oasis:entry>
         <oasis:entry colname="col3">02/03/2015 16:09:51</oasis:entry>
         <oasis:entry colname="col4">02/03/2015 16:08:38</oasis:entry>
         <oasis:entry colname="col5">02/03/2015 22:00:00</oasis:entry>
       </oasis:row>
       <oasis:row rowsep="1">
         <oasis:entry colname="col1"/>
         <oasis:entry colname="col2">Duration</oasis:entry>
         <oasis:entry colname="col3">5 days 1 h 59 min 54 s</oasis:entry>
         <oasis:entry colname="col4">5 days 1 h 58 min 12 s</oasis:entry>
         <oasis:entry colname="col5">5 days 2 h 58 min 47 s</oasis:entry>
       </oasis:row>
       <oasis:row>
         <oasis:entry colname="col1">LDB</oasis:entry>
         <oasis:entry rowsep="1" colname="col2">Instrument</oasis:entry>
         <oasis:entry rowsep="1" colname="col3">CTD rosette</oasis:entry>
         <oasis:entry rowsep="1" colname="col4">SVP</oasis:entry>
         <oasis:entry rowsep="1" colname="col5">AQUADOPP</oasis:entry>
       </oasis:row>
       <oasis:row>
         <oasis:entry colname="col1">(18.24<inline-formula><mml:math id="M7" display="inline"><mml:msup><mml:mi/><mml:mo>∘</mml:mo></mml:msup></mml:math></inline-formula> S,</oasis:entry>
         <oasis:entry colname="col2">Number</oasis:entry>
         <oasis:entry colname="col3">47 casts</oasis:entry>
         <oasis:entry colname="col4">6 drifters</oasis:entry>
         <oasis:entry colname="col5">6 deployed</oasis:entry>
       </oasis:row>
       <oasis:row>
         <oasis:entry colname="col1">189.14<inline-formula><mml:math id="M8" display="inline"><mml:msup><mml:mi/><mml:mo>∘</mml:mo></mml:msup></mml:math></inline-formula> E)</oasis:entry>
         <oasis:entry colname="col2">Start</oasis:entry>
         <oasis:entry colname="col3">15/03/2015 12:04:44</oasis:entry>
         <oasis:entry colname="col4">15/03/2015 10:00:00</oasis:entry>
         <oasis:entry colname="col5">15/03/2015 23:10:00</oasis:entry>
       </oasis:row>
       <oasis:row>
         <oasis:entry colname="col1"/>
         <oasis:entry colname="col2">Stop</oasis:entry>
         <oasis:entry colname="col3">20/03/2015 14:16:13</oasis:entry>
         <oasis:entry colname="col4">20/03/2015 23:00:00</oasis:entry>
         <oasis:entry colname="col5">20/03/2015 14:15:00</oasis:entry>
       </oasis:row>
       <oasis:row>
         <oasis:entry colname="col1"/>
         <oasis:entry rowsep="1" colname="col2">Duration</oasis:entry>
         <oasis:entry rowsep="1" colname="col3">5 days 2 h 11 min 29 s</oasis:entry>
         <oasis:entry rowsep="1" colname="col4">5 days 13 h</oasis:entry>
         <oasis:entry rowsep="1" colname="col5">4 days 15 h 5 min</oasis:entry>
       </oasis:row>
       <oasis:row>
         <oasis:entry colname="col1"/>
         <oasis:entry rowsep="1" colname="col2">Instrument</oasis:entry>
         <oasis:entry rowsep="1" colname="col3">SADCP 150</oasis:entry>
         <oasis:entry rowsep="1" colname="col4">SADCP 38</oasis:entry>
         <oasis:entry rowsep="1" colname="col5">SedTrap position</oasis:entry>
       </oasis:row>
       <oasis:row>
         <oasis:entry colname="col1"/>
         <oasis:entry colname="col2">Start</oasis:entry>
         <oasis:entry colname="col3">16/03/2015 08:51:53</oasis:entry>
         <oasis:entry colname="col4">15/03/2015 23:06:31</oasis:entry>
         <oasis:entry colname="col5">15/03/2015 12:15:48</oasis:entry>
       </oasis:row>
       <oasis:row>
         <oasis:entry colname="col1"/>
         <oasis:entry colname="col2">Stop</oasis:entry>
         <oasis:entry colname="col3">20/03/2015 14:15:54</oasis:entry>
         <oasis:entry colname="col4">20/03/2015 14:14:50</oasis:entry>
         <oasis:entry colname="col5">20/03/2015 21:00:00</oasis:entry>
       </oasis:row>
       <oasis:row rowsep="1">
         <oasis:entry colname="col1"/>
         <oasis:entry colname="col2">Duration</oasis:entry>
         <oasis:entry colname="col3">4 days 5 h 24 min 1 s</oasis:entry>
         <oasis:entry colname="col4">4 days 15 h 8 min 19 s</oasis:entry>
         <oasis:entry colname="col5">5 days 8 h 44 min 12 s</oasis:entry>
       </oasis:row>
       <oasis:row>
         <oasis:entry colname="col1">LDC</oasis:entry>
         <oasis:entry rowsep="1" colname="col2">Instrument</oasis:entry>
         <oasis:entry rowsep="1" colname="col3">CTD rosette</oasis:entry>
         <oasis:entry rowsep="1" colname="col4">SVP</oasis:entry>
         <oasis:entry rowsep="1" colname="col5">AQUADOPP</oasis:entry>
       </oasis:row>
       <oasis:row>
         <oasis:entry colname="col1">(18.42<inline-formula><mml:math id="M9" display="inline"><mml:msup><mml:mi/><mml:mo>∘</mml:mo></mml:msup></mml:math></inline-formula> S,</oasis:entry>
         <oasis:entry colname="col2">Number</oasis:entry>
         <oasis:entry colname="col3">46 casts</oasis:entry>
         <oasis:entry colname="col4">4 drifters</oasis:entry>
         <oasis:entry colname="col5">6 deployed</oasis:entry>
       </oasis:row>
       <oasis:row>
         <oasis:entry colname="col1">194.06<inline-formula><mml:math id="M10" display="inline"><mml:msup><mml:mi/><mml:mo>∘</mml:mo></mml:msup></mml:math></inline-formula> E)</oasis:entry>
         <oasis:entry colname="col2">Start</oasis:entry>
         <oasis:entry colname="col3">23/03/2015 23:10:57</oasis:entry>
         <oasis:entry colname="col4">23/03/2015 12:00:00</oasis:entry>
         <oasis:entry colname="col5">23/03/2015 23:25:00</oasis:entry>
       </oasis:row>
       <oasis:row>
         <oasis:entry colname="col1"/>
         <oasis:entry colname="col2">Stop</oasis:entry>
         <oasis:entry colname="col3">28/03/2015 14:32:30</oasis:entry>
         <oasis:entry colname="col4">28/03/2015 22:00:00</oasis:entry>
         <oasis:entry colname="col5">28/03/2015 14:30:00</oasis:entry>
       </oasis:row>
       <oasis:row>
         <oasis:entry colname="col1"/>
         <oasis:entry rowsep="1" colname="col2">Duration</oasis:entry>
         <oasis:entry rowsep="1" colname="col3">4 days 15 h 21 min 33 s</oasis:entry>
         <oasis:entry rowsep="1" colname="col4">5 days 10 h</oasis:entry>
         <oasis:entry rowsep="1" colname="col5">4 days 15 h 5 min</oasis:entry>
       </oasis:row>
       <oasis:row>
         <oasis:entry colname="col1"/>
         <oasis:entry rowsep="1" colname="col2">Instrument</oasis:entry>
         <oasis:entry rowsep="1" colname="col3">SADCP 150</oasis:entry>
         <oasis:entry rowsep="1" colname="col4">SADCP 38</oasis:entry>
         <oasis:entry rowsep="1" colname="col5">SedTrap position</oasis:entry>
       </oasis:row>
       <oasis:row>
         <oasis:entry colname="col1"/>
         <oasis:entry colname="col2">Start</oasis:entry>
         <oasis:entry colname="col3">23/03/2015 12:08:12</oasis:entry>
         <oasis:entry colname="col4">23/03/2015 12:06:55</oasis:entry>
         <oasis:entry colname="col5">23/03/2015 12:19:55</oasis:entry>
       </oasis:row>
       <oasis:row>
         <oasis:entry colname="col1"/>
         <oasis:entry colname="col2">Stop</oasis:entry>
         <oasis:entry colname="col3">28/03/2015 14:30:31</oasis:entry>
         <oasis:entry colname="col4">28/03/2015 14:31:39</oasis:entry>
         <oasis:entry colname="col5">26/03/2015 03:31:34</oasis:entry>
       </oasis:row>
       <oasis:row>
         <oasis:entry colname="col1"/>
         <oasis:entry colname="col2">Duration</oasis:entry>
         <oasis:entry colname="col3">5 days 2 h 22 min 19 s</oasis:entry>
         <oasis:entry colname="col4">5 days 2 h 24 min 44 s</oasis:entry>
         <oasis:entry colname="col5">2 days 15 h 11 min 39 s</oasis:entry>
       </oasis:row>
     </oasis:tbody>
   </oasis:tgroup></oasis:table></table-wrap>

</sec>
<?pagebreak page2127?><sec id="Ch1.S2">
  <title>Materials and methods</title>
      <p id="d1e840">In this section, we begin by describing the general manner in which the three
LD stations were conducted during the OUTPACE cruise. Following an outline of
the methodological strategy, we present the different data sources and their
processing. Additionally, we describe in detail the analyses needed to answer
our central question regarding sampling in a coherent environment.</p>
<sec id="Ch1.S2.SS1">
  <title>Sampling strategy</title>
      <p id="d1e848">The OUTPACE cruise occurred aboard the RV <italic>L'Atalante</italic> from 18 February to 3 April
in late austral summer, starting in New Caledonia and
finishing in Tahiti, traveling over 4000 <inline-formula><mml:math id="M11" display="inline"><mml:mi mathvariant="normal">km</mml:mi></mml:math></inline-formula>. Stations were conducted
in a mostly zonal transect traveling west to east, with the ship track
averaging near 19<inline-formula><mml:math id="M12" display="inline"><mml:msup><mml:mi/><mml:mo>∘</mml:mo></mml:msup></mml:math></inline-formula> S. The three LD stations, denoted as LDA, LDB, and
LDC, and lasting 5 <inline-formula><mml:math id="M13" display="inline"><mml:mi mathvariant="normal">days</mml:mi></mml:math></inline-formula> each, were designed to resolve a regional
zonal gradient in oligotrophy, the existence of which is reflected in the
surface chlorophyll <italic>a</italic> (chl <italic>a</italic>) data (Fig. 1a). As described
in the introductory article of this special issue <xref ref-type="bibr" rid="bib1.bibx33" id="paren.10"/>, site
selection for the LD stations involved identifying physical structures by use
of the SPASSO software package (<uri>http://www.mio.univ-amu.fr/SPASSO/</uri>, last access: 6 April 2018) using
near real-time satellite imagery, altimetry, and Lagrangian diagnostics
<xref ref-type="bibr" rid="bib1.bibx12 bib1.bibx13 bib1.bibx36" id="paren.11"/>.</p>
      <p id="d1e893">Before starting each LD station, surface velocity program (SVP;
<xref ref-type="bibr" rid="bib1.bibx26" id="author.12"/>, <xref ref-type="bibr" rid="bib1.bibx26" id="year.13"/>) drifters
were deployed adjacent to the site. The number of drifters deployed are
summarized in Table 1, and their mean initial positions were 1.1, 1.6, and
0.9 <inline-formula><mml:math id="M14" display="inline"><mml:mi mathvariant="normal">km</mml:mi></mml:math></inline-formula> away from the first CTD of station LDA, LDB, and LDC,
respectively. At the start of each station, two quasi-Lagrangian drifting
moorings were deployed during the OUTPACE LD stations with surface floats.
The first drifting mooring, hereafter referred to as the SedTrap Drifter, had
a “holey sock” attached at 15 <inline-formula><mml:math id="M15" display="inline"><mml:mi mathvariant="normal">m</mml:mi></mml:math></inline-formula> depth. It was followed actively by the
ship and is the emphasis of this study. It had three sediment traps
(Technicap PPS5/4) fixed at 150, 250, and 500 <inline-formula><mml:math id="M16" display="inline"><mml:mi mathvariant="normal">m</mml:mi></mml:math></inline-formula> depth, along with
onboard conductivity–temperature–depth (CTD) sensors and current meters,
described below in Sect. 2.4.1 and 2.6, respectively. The SedTrap Drifter
was deployed at the beginning of each station and was left in the water until
the station's completion. The second drifting mooring, referred to as the
production line, housed in situ incubation platforms for measuring primary
production, nitrogen fixation, oxygen, and other biogeochemical measurements
(see <xref ref-type="bibr" rid="bib1.bibx30" id="author.14"/>, <xref ref-type="bibr" rid="bib1.bibx30" id="year.15"/>, for more
documentation). The production line was redeployed on a daily basis close to
the SedTrap Drifter. While no telemetry exists for the production line, the
CTD casts from which incubation water was drawn ranged from 300 <inline-formula><mml:math id="M17" display="inline"><mml:mi mathvariant="normal">m</mml:mi></mml:math></inline-formula> to
5.7 <inline-formula><mml:math id="M18" display="inline"><mml:mi mathvariant="normal">km</mml:mi></mml:math></inline-formula> from the SedTrap drifter. After 5 <inline-formula><mml:math id="M19" display="inline"><mml:mi mathvariant="normal">days</mml:mi></mml:math></inline-formula>, the SedTrap
Drifter was recovered, and the LD station completed. Occasions when the exact
implementation of this general strategy was not realized will be mentioned in
following sections for the relevant measurements. A summary of time duration
for each data source can be found in Table 1.</p>
      <?pagebreak page2128?><p id="d1e951">Between the LD stations, 15 short duration (SD) stations lasting
approximately 8 <inline-formula><mml:math id="M20" display="inline"><mml:mi mathvariant="normal">h</mml:mi></mml:math></inline-formula> each were interspersed along the ship's trajectory
in roughly equidistant sections (Fig. 1b). Among the measurements made, CTD
casts from SD stations will figure into the validation process in this study.
Most casts (both LD and SD) were at least 200 <inline-formula><mml:math id="M21" display="inline"><mml:mi mathvariant="normal">m</mml:mi></mml:math></inline-formula>, with at least one
2000 <inline-formula><mml:math id="M22" display="inline"><mml:mi mathvariant="normal">m</mml:mi></mml:math></inline-formula> cast for all stations. These casts were conducted with the same
CTD rosette platform described more fully below in Sect. 2.3.1.</p>
      <p id="d1e975">Throughout the cruise, surface conductivity–temperature measurements from the
thermosalinograph (TSG) and currents from shipboard acoustic Doppler current
profilers (SADCP) were collected. Their processing is described in Sect. 2.4.1 and 2.6, respectively.</p>
</sec>
<sec id="Ch1.S2.SS2">
  <title>Post-validation method</title>
      <p id="d1e984">The goal of this study is to evaluate whether the three LD stations sampled in a homogeneous body of water during
OUTPACE. In order to achieve this goal, a number of steps were undertaken:
<list list-type="bullet"><list-item>
      <p id="d1e989"><italic>Validity of application and environmental context.</italic> As mentioned in
the introduction, if a physical structure such as an eddy or front is
present, its dynamics will dominate and must be taken into account.
Additionally, since we used physical water properties in this study, we must
determine whether biogeochemical gradients existed at smaller scales. For
this purpose, we looked at remote sensing data.</p></list-item><list-item>
      <p id="d1e995"><italic>Establishment of statistical baseline.</italic> To evaluate whether station sampling
remained in one water mass, the water mass itself needed to characterized.
This was<?pagebreak page2129?> achieved by initializing a baseline within the time series of
hydrographic properties. The subsequent time evolution of these properties
within the defining dataset served as a first test for whether sampling
stayed in one environment.</p></list-item><list-item>
      <p id="d1e1001"><italic>Spatial scale determination and baseline context.</italic> If time series analysis showed
no change in water properties, complementary data from farther away were
compiled to evaluate the spatial scale at which the water mass did change.
These data were also used to contextualize whether the statistical baseline
over-estimated or under-estimated water mass variability.</p></list-item><list-item>
      <p id="d1e1007"><italic>Currents analysis and Lagrangian risk.</italic> The spatial scale of the water
mass already
determined, water trajectories were used to evaluate at what point the
observed flow regime might have brought another water mass into contact with
LD station sampling near the SedTrap drifter.</p></list-item></list>
The following sections in the methods are organized around these steps, detailing the data and analyses involved for each step.</p>
</sec>
<sec id="Ch1.S2.SS3">
  <title>Validity of method application and environmental context</title>
      <p id="d1e1019">Detection of physical structures and biogeochemical gradients used satellite
measurements of sea surface temperature (SST), surface chl <italic>a</italic>, and
sea surface height with its associated geostrophic currents. These data were
also used in the LD site selection phase (Sect. 2.1). All processed satellite
data were provided by CLS with support from CNES. SST was derived from a
combination of AQUA/MODIS, TERRA/MODIS, METOP-A/AVHRR, METOP-B/AVHRR sensors,
with the daily product produced being a weighted mean spanning 5 <inline-formula><mml:math id="M23" display="inline"><mml:mi mathvariant="normal">days</mml:mi></mml:math></inline-formula>
(inclusive) previous to the date in question. Weighting was greater for more
recent data. Similar to SST, chl <italic>a</italic> was a 5 <inline-formula><mml:math id="M24" display="inline"><mml:mi mathvariant="normal">day</mml:mi></mml:math></inline-formula> weighted mean
produced by the Suomi/NPP/VIIRS sensor. The SST and chl <italic>a</italic> products
had a 0.02<inline-formula><mml:math id="M25" display="inline"><mml:msup><mml:mi/><mml:mo>∘</mml:mo></mml:msup></mml:math></inline-formula> resolution, equivalent to <inline-formula><mml:math id="M26" display="inline"><mml:mo>∼</mml:mo></mml:math></inline-formula> 2 <inline-formula><mml:math id="M27" display="inline"><mml:mi mathvariant="normal">km</mml:mi></mml:math></inline-formula>. These
satellite products spanned from 1 December 2014 to 15 May 2015. In order to
compress the daily satellite products, weighted temporal means were
calculated. For each grid location, the weight for a given day was inversely
proportional to the distance from the grid point to the ship's daily
position.</p>
      <p id="d1e1069">Temporal fluctuations of SST and surface chl <italic>a</italic> were determined by
producing time series of both variables within a given spatial range
surrounding the starting position of the three LD stations. The spatial range
consisted of a <inline-formula><mml:math id="M28" display="inline"><mml:mrow><mml:mn mathvariant="normal">120</mml:mn><mml:mo>×</mml:mo><mml:mn mathvariant="normal">120</mml:mn></mml:mrow></mml:math></inline-formula> <inline-formula><mml:math id="M29" display="inline"><mml:mi mathvariant="normal">km</mml:mi></mml:math></inline-formula> box centered on each LD station. Satellite
pixels falling within this region were used to create a probability
distribution function. The 120 <inline-formula><mml:math id="M30" display="inline"><mml:mi mathvariant="normal">km</mml:mi></mml:math></inline-formula> square size was chosen because 60 <inline-formula><mml:math id="M31" display="inline"><mml:mi mathvariant="normal">km</mml:mi></mml:math></inline-formula>
is a typical size of the Rossby radius of deformation for the
region <xref ref-type="bibr" rid="bib1.bibx6" id="paren.16"/>. Sudden changes in SST and
chl <italic>a</italic> distributions indicated strong surface forcing or the passage
of gradients, which could invalidate the applicability of the method.</p>
      <p id="d1e1115">Local surface currents derived from altimetry were also provided by CLS with
support from CNES. These data come from the Jason-2, Saral-AltiKa, Cryosat-2,
and HY-2A missions, cover a domain from 140 to 220<inline-formula><mml:math id="M32" display="inline"><mml:msup><mml:mi/><mml:mo>∘</mml:mo></mml:msup></mml:math></inline-formula> E, and
30<inline-formula><mml:math id="M33" display="inline"><mml:msup><mml:mi/><mml:mo>∘</mml:mo></mml:msup></mml:math></inline-formula> S to the Equator, covering the yearlong period of June 2014 to
May 2015. The velocity grid had a <inline-formula><mml:math id="M34" display="inline"><mml:mrow><mml:mn mathvariant="normal">1</mml:mn><mml:mo>/</mml:mo><mml:msup><mml:mn mathvariant="normal">8</mml:mn><mml:mo>∘</mml:mo></mml:msup></mml:mrow></mml:math></inline-formula> resolution, applying
the FES2014 tidal model and CNES_CLS_2015 mean sea surface. Ekman effects
due to wind were also added using ECMWF ERA INTERIM model output.</p>
</sec>
<sec id="Ch1.S2.SS4">
  <title>Establishment of statistical baseline</title>
      <p id="d1e1158">Water mass characterization depended upon observations of conservative
temperature (<inline-formula><mml:math id="M35" display="inline"><mml:mrow><mml:msub><mml:mi>C</mml:mi><mml:mi mathvariant="normal">T</mml:mi></mml:msub></mml:mrow></mml:math></inline-formula>) and absolute salinity (<inline-formula><mml:math id="M36" display="inline"><mml:mrow><mml:msub><mml:mi>S</mml:mi><mml:mi mathvariant="normal">A</mml:mi></mml:msub></mml:mrow></mml:math></inline-formula>), or <inline-formula><mml:math id="M37" display="inline"><mml:mi>T</mml:mi></mml:math></inline-formula>-<inline-formula><mml:math id="M38" display="inline"><mml:mi>S</mml:mi></mml:math></inline-formula> measurements. The
statistical baseline, which served as the reference for each LD station,
needed to reflect the initial state of the water near the SedTrap drifter.
While the SedTrap drifter itself had CTD sensors onboard, these were fixed in
depth and did not resolve the full variability of the water column.
Additionally, although the SedTrap drifter served as the moving station's
location, water derived from the shipboard CTD rosette ultimately served as
the source material for the biogeochemical measurements taken during the
cruise. The shipboard casts were always positioned near the SedTrap drifter,
averaging 1.2 <inline-formula><mml:math id="M39" display="inline"><mml:mi mathvariant="normal">km</mml:mi></mml:math></inline-formula> over the entire cruise. For these reasons the CTD
cast data were chosen for the baseline calculation, while both SedTrap
drifter and CTD cast data were included in the time series analysis.</p>
<sec id="Ch1.S2.SS4.SSS1">
  <title>CTD data for time series analysis</title>
      <p id="d1e1209">The shipboard CTD employed during OUTPACE was a Seabird SBE 9+ CTD rosette
with two CTDs installed. Data from each cast were calibrated and processed
post-cruise using Sea-Bird Electronics software into 1 <inline-formula><mml:math id="M40" display="inline"><mml:mi mathvariant="normal">m</mml:mi></mml:math></inline-formula> bins. All CTD
data from other instruments mentioned later were likewise processed using
Sea-Bird Electronics software. <inline-formula><mml:math id="M41" display="inline"><mml:mrow><mml:msub><mml:mi>S</mml:mi><mml:mi mathvariant="normal">A</mml:mi></mml:msub></mml:mrow></mml:math></inline-formula>, <inline-formula><mml:math id="M42" display="inline"><mml:mrow><mml:msub><mml:mi>C</mml:mi><mml:mi mathvariant="normal">T</mml:mi></mml:msub></mml:mrow></mml:math></inline-formula>, and potential density
(<inline-formula><mml:math id="M43" display="inline"><mml:mrow><mml:msub><mml:mi mathvariant="italic">σ</mml:mi><mml:mi mathvariant="italic">θ</mml:mi></mml:msub></mml:mrow></mml:math></inline-formula>) were calculated using the TEOS-10 standard
<xref ref-type="bibr" rid="bib1.bibx29" id="paren.17"/>. In total, over 200 CTD casts were performed
during OUTPACE. Most SD stations had three or four casts, except for SD13,
which had time for only one cast owing to a medical emergency. The LDA, LDB,
and LDC stations had 46, 47, and 46 casts, respectively, each approximately 3 <inline-formula><mml:math id="M44" display="inline"><mml:mi mathvariant="normal">h</mml:mi></mml:math></inline-formula>
apart. During LDA, the two drifting moorings accidentally collided
and, due to the time necessary to disentangle them, there is a gap of 9 <inline-formula><mml:math id="M45" display="inline"><mml:mi mathvariant="normal">h</mml:mi></mml:math></inline-formula>
in the time series. The majority of CTD casts were to a 200 <inline-formula><mml:math id="M46" display="inline"><mml:mi mathvariant="normal">m</mml:mi></mml:math></inline-formula>
depth, with at least one 2000 <inline-formula><mml:math id="M47" display="inline"><mml:mi mathvariant="normal">m</mml:mi></mml:math></inline-formula> cast per station. Mixed layer depth
was determined using <xref ref-type="bibr" rid="bib1.bibx10" id="text.18"/>'s method, by finding the depth where
density has changed more than 0.3 <inline-formula><mml:math id="M48" display="inline"><mml:mrow><mml:mi mathvariant="normal">kg</mml:mi><mml:mspace linebreak="nobreak" width="0.125em"/><mml:msup><mml:mi mathvariant="normal">m</mml:mi><mml:mrow><mml:mo>-</mml:mo><mml:mn mathvariant="normal">3</mml:mn></mml:mrow></mml:msup></mml:mrow></mml:math></inline-formula> from a reference
value, which was chosen to be the value at a 10 <inline-formula><mml:math id="M49" display="inline"><mml:mi mathvariant="normal">m</mml:mi></mml:math></inline-formula> depth. The 10 <inline-formula><mml:math id="M50" display="inline"><mml:mi mathvariant="normal">m</mml:mi></mml:math></inline-formula>
reference was<?pagebreak page2130?> chosen because post-processed CTD casts did not always include
the surface.</p>
      <p id="d1e1319">The SedTrap Drifter had six SBE 37 Microcat CTDs on board. Their depths, as
determined by mean in situ pressure, were <inline-formula><mml:math id="M51" display="inline"><mml:mo>∼</mml:mo></mml:math></inline-formula> 14, 55, 88, 105, 137, and
197 <inline-formula><mml:math id="M52" display="inline"><mml:mi mathvariant="normal">m</mml:mi></mml:math></inline-formula>. These instruments yielded data every 5 <inline-formula><mml:math id="M53" display="inline"><mml:mi mathvariant="normal">min</mml:mi></mml:math></inline-formula> during their
deployments. As mentioned in the previous paragraph, during LDA the SedTrap
Drifter tangled with the production line and so the data presented here from
LDA came from its redeployment until the end of LDA. No gap in the data
occurred for LDB or LDC.</p>
</sec>
<sec id="Ch1.S2.SS4.SSS2">
  <title>Tracer analysis and baseline definition</title>
      <p id="d1e1349">The need for a baseline within the OUTPACE dataset can be shown by comparison
of the CTD data with climatologies such as the World Ocean Atlas (Fig. S1 in the Supplement;
<xref ref-type="bibr" rid="bib1.bibx5" id="author.19"/>, <xref ref-type="bibr" rid="bib1.bibx5" id="year.20"/>). While OUTPACE
observations were consistent with these previous observations, when metrics
of variability were available they produced envelopes of max/min <inline-formula><mml:math id="M54" display="inline"><mml:mi>T</mml:mi></mml:math></inline-formula>-<inline-formula><mml:math id="M55" display="inline"><mml:mi>S</mml:mi></mml:math></inline-formula> values
large enough to preclude distinguishing between different stations. Since no
other sufficiently fine data were available to compare <inline-formula><mml:math id="M56" display="inline"><mml:mi>T</mml:mi></mml:math></inline-formula>-<inline-formula><mml:math id="M57" display="inline"><mml:mi>S</mml:mi></mml:math></inline-formula> measurements, data
from within each station were used to create a reference baseline of <inline-formula><mml:math id="M58" display="inline"><mml:mi>T</mml:mi></mml:math></inline-formula>-<inline-formula><mml:math id="M59" display="inline"><mml:mi>S</mml:mi></mml:math></inline-formula>
variability. Given the lack of fine variability data and the need to work
within the dataset of a single cruise, another approach was needed to
condense <inline-formula><mml:math id="M60" display="inline"><mml:mi>T</mml:mi></mml:math></inline-formula>-<inline-formula><mml:math id="M61" display="inline"><mml:mi>S</mml:mi></mml:math></inline-formula> variability so that physical environments can be distinguished.</p>
      <p id="d1e1415">Generally, over the upper 200 <inline-formula><mml:math id="M62" display="inline"><mml:mi mathvariant="normal">m</mml:mi></mml:math></inline-formula> of the water column, the depth range
of most of our CTD casts, a given profile of <inline-formula><mml:math id="M63" display="inline"><mml:mi>T</mml:mi></mml:math></inline-formula>-<inline-formula><mml:math id="M64" display="inline"><mml:mi>S</mml:mi></mml:math></inline-formula> values will vary along a
curve <xref ref-type="bibr" rid="bib1.bibx44" id="paren.21"/>. This reflects how each profile is made up of
increasingly denser layers over depth, each with distinct histories. In some
sense, these layers could be considered their own physical micro-environment,
and their ensemble constitutes the physical environment. Assuming that the
density layers were not subject to strong forcing, such as diapycnal mixing
events or atmospheric effects, their values should have remained constant
until isopycnal exchange or diffusion could occur over longer timescales.
Treating these density layers as separate entities, variations of <inline-formula><mml:math id="M65" display="inline"><mml:mi>T</mml:mi></mml:math></inline-formula>-<inline-formula><mml:math id="M66" display="inline"><mml:mi>S</mml:mi></mml:math></inline-formula> along
isopycnal surfaces can provide an approach to distinguish physical
environments, which is the goal of our analysis.</p>
      <p id="d1e1457">Using density as one variable, another is needed to fully describe a water
parcel's characteristics, ideally one which is independent of density. Spice,
a variable constructed from <inline-formula><mml:math id="M67" display="inline"><mml:mi>T</mml:mi></mml:math></inline-formula>-<inline-formula><mml:math id="M68" display="inline"><mml:mi>S</mml:mi></mml:math></inline-formula>, is well suited for this purpose. Spice is
defined such that hot and salty water is “spicy”, a convention dating to
<xref ref-type="bibr" rid="bib1.bibx34" id="text.22"/>. In the formulation proposed by
<xref ref-type="bibr" rid="bib1.bibx16" id="text.23"/>, its isopleths are perpendicular to
isopycnals everywhere, and it effectively both encapsulates and accentuates <inline-formula><mml:math id="M69" display="inline"><mml:mi>T</mml:mi></mml:math></inline-formula>-<inline-formula><mml:math id="M70" display="inline"><mml:mi>S</mml:mi></mml:math></inline-formula>
variability at a given density into a single value. Therefore, in our
analysis, spice variability in a given density layer was used to define the
statistical baseline and determine whether a physical environment changed
during OUTPACE.</p>
      <p id="d1e1495">The statistical baseline was defined as a functional fit between
<inline-formula><mml:math id="M71" display="inline"><mml:mrow><mml:msub><mml:mi mathvariant="italic">σ</mml:mi><mml:mi mathvariant="italic">θ</mml:mi></mml:msub></mml:mrow></mml:math></inline-formula> and spice measurements at the beginning of each LD station in
the upper 200 <inline-formula><mml:math id="M72" display="inline"><mml:mi mathvariant="normal">m</mml:mi></mml:math></inline-formula> of the water column spanning the euphotic zone. The
period of time used for defining the baseline was chosen to be the local
inertial period, so that internal wave effects would be minimized. For each
station, this meant that the first 13, 15, and 15 casts were used for LDA,
LDB, and LDC, respectively. A regular grid of density values was created,
with one fourth of the number of values as the total number of observations. The
fit of baseline spice, or <inline-formula><mml:math id="M73" display="inline"><mml:mrow><mml:msub><mml:mi>S</mml:mi><mml:mi mathvariant="normal">base</mml:mi></mml:msub><mml:mo>(</mml:mo><mml:mi mathvariant="italic">ρ</mml:mi><mml:mo>)</mml:mo></mml:mrow></mml:math></inline-formula>, was calculated inside a moving
window of <inline-formula><mml:math id="M74" display="inline"><mml:mo>±</mml:mo></mml:math></inline-formula>0.1 <inline-formula><mml:math id="M75" display="inline"><mml:mrow><mml:mi mathvariant="normal">kg</mml:mi><mml:mspace width="0.125em" linebreak="nobreak"/><mml:msup><mml:mi mathvariant="normal">m</mml:mi><mml:mrow><mml:mo>-</mml:mo><mml:mn mathvariant="normal">3</mml:mn></mml:mrow></mml:msup></mml:mrow></mml:math></inline-formula> along with the standard error in
spice, <inline-formula><mml:math id="M76" display="inline"><mml:mrow><mml:mi>S</mml:mi><mml:msub><mml:mtext>Err</mml:mtext><mml:mi mathvariant="normal">base</mml:mi></mml:msub><mml:mo>(</mml:mo><mml:mi mathvariant="italic">ρ</mml:mi><mml:mo>)</mml:mo></mml:mrow></mml:math></inline-formula>. Only values corresponding to windows with at
least 50 observations were kept.</p>
      <p id="d1e1578">Comparisons between the baseline and new <inline-formula><mml:math id="M77" display="inline"><mml:mrow><mml:msub><mml:mi mathvariant="italic">σ</mml:mi><mml:mi mathvariant="italic">θ</mml:mi></mml:msub></mml:mrow></mml:math></inline-formula>-spice measurements
were made using a <inline-formula><mml:math id="M78" display="inline"><mml:mi>Z</mml:mi></mml:math></inline-formula>-Score, following the general formula as follows:
              <disp-formula id="Ch1.E1" content-type="numbered"><mml:math id="M79" display="block"><mml:mrow><mml:mstyle class="stylechange" displaystyle="true"/><mml:mi>Z</mml:mi><mml:mo>(</mml:mo><mml:msub><mml:mi mathvariant="italic">ρ</mml:mi><mml:mi mathvariant="normal">obs</mml:mi></mml:msub><mml:mo>)</mml:mo><mml:mo>=</mml:mo><mml:mstyle displaystyle="true"><mml:mfrac style="display"><mml:mrow><mml:msub><mml:mi>S</mml:mi><mml:mi mathvariant="normal">obs</mml:mi></mml:msub><mml:mo>-</mml:mo><mml:msub><mml:mi>S</mml:mi><mml:mi mathvariant="normal">base</mml:mi></mml:msub></mml:mrow><mml:mrow><mml:mi>S</mml:mi><mml:msub><mml:mtext>Err</mml:mtext><mml:mi mathvariant="normal">base</mml:mi></mml:msub></mml:mrow></mml:mfrac></mml:mstyle><mml:mo>,</mml:mo></mml:mrow></mml:math></disp-formula>
            where, for a density observation <inline-formula><mml:math id="M80" display="inline"><mml:mrow><mml:msub><mml:mi mathvariant="italic">ρ</mml:mi><mml:mi mathvariant="normal">obs</mml:mi></mml:msub></mml:mrow></mml:math></inline-formula>, <inline-formula><mml:math id="M81" display="inline"><mml:mrow><mml:msub><mml:mi>S</mml:mi><mml:mi mathvariant="normal">obs</mml:mi></mml:msub></mml:mrow></mml:math></inline-formula> is the observed
spice with <inline-formula><mml:math id="M82" display="inline"><mml:mrow><mml:msub><mml:mi>S</mml:mi><mml:mi mathvariant="normal">base</mml:mi></mml:msub></mml:mrow></mml:math></inline-formula> and <inline-formula><mml:math id="M83" display="inline"><mml:mrow><mml:mi>S</mml:mi><mml:msub><mml:mtext>Err</mml:mtext><mml:mi mathvariant="normal">base</mml:mi></mml:msub></mml:mrow></mml:math></inline-formula> being the linearly interpolated
functional baseline spice value and standard error. The assumption applied in
this analysis is that while a continuous curve in <inline-formula><mml:math id="M84" display="inline"><mml:mi>T</mml:mi></mml:math></inline-formula>-<inline-formula><mml:math id="M85" display="inline"><mml:mi>S</mml:mi></mml:math></inline-formula>, or
<inline-formula><mml:math id="M86" display="inline"><mml:mrow><mml:msub><mml:mi mathvariant="italic">σ</mml:mi><mml:mi mathvariant="italic">θ</mml:mi></mml:msub><mml:mo>-</mml:mo><mml:mi>S</mml:mi></mml:mrow></mml:math></inline-formula>, is to be expected and can be fit to a function, the
isopycnal layers were independent of each other, and represented different
physical sub-populations. Keeping track of variability through <inline-formula><mml:math id="M87" display="inline"><mml:mi>Z</mml:mi></mml:math></inline-formula>-score tied
to a functional <inline-formula><mml:math id="M88" display="inline"><mml:mrow><mml:msub><mml:mi mathvariant="italic">σ</mml:mi><mml:mi mathvariant="italic">θ</mml:mi></mml:msub><mml:mo>-</mml:mo><mml:mi>S</mml:mi></mml:mrow></mml:math></inline-formula> relationship produced a flexible metric.
For sensors fixed at a certain depth, such as for the SedTrap drifter, a
<inline-formula><mml:math id="M89" display="inline"><mml:mi>Z</mml:mi></mml:math></inline-formula>-score could be computed irrespective of whether internal waves were
shoaling or deepening isopycnals.</p>
</sec>
</sec>
<sec id="Ch1.S2.SS5">
  <title>Spatial scale determination and baseline context</title>
      <p id="d1e1757">The <inline-formula><mml:math id="M90" display="inline"><mml:mi>Z</mml:mi></mml:math></inline-formula>-scores derived from the CTD and SedTrap time series provided a
first-order evaluation of physical variability in the immediate surroundings
of the SedTrap drifter as it moved through time. If large variability (<inline-formula><mml:math id="M91" display="inline"><mml:mrow><mml:mo>|</mml:mo><mml:mi>Z</mml:mi><mml:mo>|</mml:mo><mml:mo>&gt;</mml:mo><mml:mn mathvariant="normal">2</mml:mn></mml:mrow></mml:math></inline-formula>,
in the traditional <inline-formula><mml:math id="M92" display="inline"><mml:mrow><mml:mi mathvariant="italic">α</mml:mi><mml:mo>=</mml:mo><mml:mn mathvariant="normal">0.05</mml:mn></mml:mrow></mml:math></inline-formula>) was observed, then the physical
environment likely had changed. However, if <inline-formula><mml:math id="M93" display="inline"><mml:mrow><mml:mo>|</mml:mo><mml:mi>Z</mml:mi><mml:mo>|</mml:mo><mml:mo>&lt;</mml:mo><mml:mn mathvariant="normal">2</mml:mn></mml:mrow></mml:math></inline-formula> this was not a guarantee
that the physical environment had not changed. Since the functional fit of
<inline-formula><mml:math id="M94" display="inline"><mml:mrow><mml:msub><mml:mi mathvariant="italic">σ</mml:mi><mml:mi mathvariant="italic">θ</mml:mi></mml:msub></mml:mrow></mml:math></inline-formula> and spice was based only upon the data from OUTPACE, <inline-formula><mml:math id="M95" display="inline"><mml:mi>Z</mml:mi></mml:math></inline-formula>-score
is a relative measure of variability. In order to test whether the
<inline-formula><mml:math id="M96" display="inline"><mml:mrow><mml:msub><mml:mi mathvariant="italic">σ</mml:mi><mml:mi mathvariant="italic">θ</mml:mi></mml:msub></mml:mrow></mml:math></inline-formula>-spice relationship was robust, it was necessary to extend the
<inline-formula><mml:math id="M97" display="inline"><mml:mi>Z</mml:mi></mml:math></inline-formula>-score analysis farther in space to include complementary density and spice
measurements. If <inline-formula><mml:math id="M98" display="inline"><mml:mi>Z</mml:mi></mml:math></inline-formula>-scores remained low for large distances, then the
<inline-formula><mml:math id="M99" display="inline"><mml:mrow><mml:mi>S</mml:mi><mml:msub><mml:mtext>Err</mml:mtext><mml:mi mathvariant="normal">base</mml:mi></mml:msub></mml:mrow></mml:math></inline-formula> was too large. By compiling independent <inline-formula><mml:math id="M100" display="inline"><mml:mi>Z</mml:mi></mml:math></inline-formula>-scores over larger
distances, we can test whether there is a relationship between <inline-formula><mml:math id="M101" display="inline"><mml:mi>Z</mml:mi></mml:math></inline-formula>-score and
distance.</p>
      <?pagebreak page2131?><p id="d1e1883">During OUTPACE, complementary <inline-formula><mml:math id="M102" display="inline"><mml:mrow><mml:msub><mml:mi mathvariant="italic">σ</mml:mi><mml:mi mathvariant="italic">θ</mml:mi></mml:msub></mml:mrow></mml:math></inline-formula>-spice observations stemmed from
neighboring stations, the SD stations (Fig. 1). Additionally, surface
measurements for the entire cruise were provided from a Seabird SBE 21 SeaCAT
thermosalinograph (TSG), with SBE 38 thermometer using the ship's continuous
surface water intake. Subsequent to post-cruise processing of TSG data as
detailed in <xref ref-type="bibr" rid="bib1.bibx2" id="text.24"/>, the time series was available in 2 <inline-formula><mml:math id="M103" display="inline"><mml:mi mathvariant="normal">min</mml:mi></mml:math></inline-formula> intervals.</p>
      <p id="d1e1907">The relationship between <inline-formula><mml:math id="M104" display="inline"><mml:mi>Z</mml:mi></mml:math></inline-formula>-score vs. distance was used to evaluate the
baseline. Distances were calculated from the ship position of observation and
the initial CTD cast position for the LD station. For the SD stations, the
<inline-formula><mml:math id="M105" display="inline"><mml:mi>Z</mml:mi></mml:math></inline-formula>-scores found by the functional fit of spice for each cast were binned by
density, in a regular grid with bins of 0.25 <inline-formula><mml:math id="M106" display="inline"><mml:mrow><mml:mi mathvariant="normal">kg</mml:mi><mml:mspace width="0.125em" linebreak="nobreak"/><mml:msup><mml:mi mathvariant="normal">m</mml:mi><mml:mrow><mml:mo>-</mml:mo><mml:mn mathvariant="normal">3</mml:mn></mml:mrow></mml:msup></mml:mrow></mml:math></inline-formula> width.
TSG data from during the LD stations were excluded and <inline-formula><mml:math id="M107" display="inline"><mml:mi>Z</mml:mi></mml:math></inline-formula>-scores were binned
by distance from the station, in 10 <inline-formula><mml:math id="M108" display="inline"><mml:mi mathvariant="normal">km</mml:mi></mml:math></inline-formula> increments for the first 100 <inline-formula><mml:math id="M109" display="inline"><mml:mi mathvariant="normal">km</mml:mi></mml:math></inline-formula>,
and then 20 <inline-formula><mml:math id="M110" display="inline"><mml:mi mathvariant="normal">km</mml:mi></mml:math></inline-formula> afterwards until 500 <inline-formula><mml:math id="M111" display="inline"><mml:mi mathvariant="normal">km</mml:mi></mml:math></inline-formula>. The spatial
scale <inline-formula><mml:math id="M112" display="inline"><mml:mrow><mml:msub><mml:mi>R</mml:mi><mml:mi>Z</mml:mi></mml:msub></mml:mrow></mml:math></inline-formula> was determined where <inline-formula><mml:math id="M113" display="inline"><mml:mrow><mml:mi>Z</mml:mi><mml:mo>≥</mml:mo><mml:mn mathvariant="normal">2</mml:mn></mml:mrow></mml:math></inline-formula> and used to evaluate the ability
of the statistical baseline to discern gradients in physical properties. A
natural spatial scale to serve as a useful reference to the empirical
distance is the first Rossby radius of deformation, approximated via
Wentzel–Kramers–Brillouin (WKB) method by <xref ref-type="bibr" rid="bib1.bibx6" id="text.25"/>
as follows:
            <disp-formula id="Ch1.E2" content-type="numbered"><mml:math id="M114" display="block"><mml:mrow><mml:mstyle displaystyle="true" class="stylechange"/><mml:msub><mml:mi>R</mml:mi><mml:mi>D</mml:mi></mml:msub><mml:mo>=</mml:mo><mml:mstyle displaystyle="true"><mml:mfrac style="display"><mml:mn mathvariant="normal">1</mml:mn><mml:mrow><mml:mi mathvariant="italic">π</mml:mi><mml:mi>f</mml:mi></mml:mrow></mml:mfrac></mml:mstyle><mml:msubsup><mml:mo movablelimits="false">∫</mml:mo><mml:mrow><mml:mo>-</mml:mo><mml:mi>H</mml:mi></mml:mrow><mml:mn mathvariant="normal">0</mml:mn></mml:msubsup><mml:mi>N</mml:mi><mml:mo>(</mml:mo><mml:mi>z</mml:mi><mml:mo>)</mml:mo><mml:mi mathvariant="normal">d</mml:mi><mml:mi>z</mml:mi><mml:mo>,</mml:mo></mml:mrow></mml:math></disp-formula>
          where <inline-formula><mml:math id="M115" display="inline"><mml:mi>f</mml:mi></mml:math></inline-formula> is the local coriolis parameter and <inline-formula><mml:math id="M116" display="inline"><mml:mrow><mml:mi>N</mml:mi><mml:mo>(</mml:mo><mml:mi>z</mml:mi><mml:mo>)</mml:mo></mml:mrow></mml:math></inline-formula> is the depth-dependent
Brunt–Väisälä frequency. <inline-formula><mml:math id="M117" display="inline"><mml:mrow><mml:msub><mml:mi>R</mml:mi><mml:mi>D</mml:mi></mml:msub></mml:mrow></mml:math></inline-formula> was calculated for each LD station using
the deepest cast available: 2000 <inline-formula><mml:math id="M118" display="inline"><mml:mi mathvariant="normal">m</mml:mi></mml:math></inline-formula> casts for LDA and B, and a deep
5000 <inline-formula><mml:math id="M119" display="inline"><mml:mi mathvariant="normal">m</mml:mi></mml:math></inline-formula> cast for LDC. <inline-formula><mml:math id="M120" display="inline"><mml:mi>N</mml:mi></mml:math></inline-formula> was calculated with centered differences of
the 1 <inline-formula><mml:math id="M121" display="inline"><mml:mi mathvariant="normal">m</mml:mi></mml:math></inline-formula> binned density profiles.</p>
</sec>
<sec id="Ch1.S2.SS6">
  <title>Currents analysis and Lagrangian risk</title>
      <p id="d1e2119">The previous step analyzed the relationship between <inline-formula><mml:math id="M122" display="inline"><mml:mi>Z</mml:mi></mml:math></inline-formula>-score and distance,
providing an estimated distance over which the physical environment changed.
In order to evaluate the risk that the SedTrap drifter encountered different
water masses, an analysis of the local currents was needed. Since it is clear
that the SedTrap drifter was not perfectly Lagrangian and that vertical
shear could transport layers at different rates, it was necessary to see if
water at specific depths could have advected the distance over which
different water masses appear.</p>
      <p id="d1e2129">The in situ velocities for each LD station were derived from the shipboard
acoustic Doppler current profilers (SADCP), two ocean surveyors at 150 and 38 <inline-formula><mml:math id="M123" display="inline"><mml:mi mathvariant="normal">kHz</mml:mi></mml:math></inline-formula>. Time series data for the SADCPs were
post-processed using the CASCADE software package <xref ref-type="bibr" rid="bib1.bibx23 bib1.bibx25" id="paren.26"/> and binned into 2 <inline-formula><mml:math id="M124" display="inline"><mml:mi mathvariant="normal">min</mml:mi></mml:math></inline-formula> intervals. The 150 <inline-formula><mml:math id="M125" display="inline"><mml:mi mathvariant="normal">kHz</mml:mi></mml:math></inline-formula>
SADCP provided a depth resolution of 8 <inline-formula><mml:math id="M126" display="inline"><mml:mi mathvariant="normal">m</mml:mi></mml:math></inline-formula>, with bins
starting from 20 <inline-formula><mml:math id="M127" display="inline"><mml:mi mathvariant="normal">m</mml:mi></mml:math></inline-formula>, and reliable data coverage down to 200 <inline-formula><mml:math id="M128" display="inline"><mml:mi mathvariant="normal">m</mml:mi></mml:math></inline-formula>
depth. Since the SedTrap Drifter had sediment traps extending down to 500 <inline-formula><mml:math id="M129" display="inline"><mml:mi mathvariant="normal">m</mml:mi></mml:math></inline-formula>
depth, the 38 <inline-formula><mml:math id="M130" display="inline"><mml:mi mathvariant="normal">kHz</mml:mi></mml:math></inline-formula> data was also used, albeit with reduced
depth resolution of 24 <inline-formula><mml:math id="M131" display="inline"><mml:mi mathvariant="normal">m</mml:mi></mml:math></inline-formula> bins, extending from 52 <inline-formula><mml:math id="M132" display="inline"><mml:mi mathvariant="normal">m</mml:mi></mml:math></inline-formula> down to 1000 <inline-formula><mml:math id="M133" display="inline"><mml:mi mathvariant="normal">m</mml:mi></mml:math></inline-formula>.
Additional in situ velocities came from six Nortek AQUADOPP current
meters, positioned at 11, 55, 88, 105, 135, and 198 <inline-formula><mml:math id="M134" display="inline"><mml:mi mathvariant="normal">m</mml:mi></mml:math></inline-formula> on the SedTrap
Drifter. The post-processed AQUADOPP time series provided observations every 5 <inline-formula><mml:math id="M135" display="inline"><mml:mi mathvariant="normal">min</mml:mi></mml:math></inline-formula>.</p>
      <p id="d1e2228">Velocities were integrated using a first-order Euler method to calculate the
theoretical trajectories of water subsequent to the beginning of each LD
station. Since the object of these calculations was to see whether water
could have traveled a critical spatial scale, for each dataset the maximal
amount of time was given for the time integration. SADCP time series spanned
between the first and last CTD of the LD station, using the ship position as
the initial position. The AQUADOPP integrations spanned the entirety of valid
data and used the corresponding SedTrap Drifter satellite fix for an initial
position.</p>
      <p id="d1e2231">To compare the integrated velocity positions with the realized positions of
the SedTrap drifter and SVP drifters, GPS positioning was achieved by use of
Iridium telemetry. Positions were successfully found for LDA before and after
the SedTrap Drifter's redeployment, along with all of LDB. During LDC, the
battery of the positioning antenna ran out and so the time series for LDC
positions of the SedTrap Drifter was shortened. Since only the initial
position is needed for the velocity integration, the AQUADOPP integration was
continued beyond this positioning failure until the SedTrap Drifter was
recovered. Positions of the SVP drifters deployed at each station were
successfully retrieved for all three LD stations. Satellite fixes were
available spaced about 1 <inline-formula><mml:math id="M136" display="inline"><mml:mi mathvariant="normal">h</mml:mi></mml:math></inline-formula> apart for both datasets. Both SedTrap
Drifter and SVP positions were interpolated to hourly time series. SVP
positions were used to compute relative dispersion (Fig. S2 in the
Supplement),
using the definition for N particles as follows
<xref ref-type="bibr" rid="bib1.bibx21" id="paren.27"/>:
            <disp-formula id="Ch1.E3" content-type="numbered"><mml:math id="M137" display="block"><mml:mrow><mml:mstyle class="stylechange" displaystyle="true"/><mml:mi>D</mml:mi><mml:mo>(</mml:mo><mml:mi>t</mml:mi><mml:mo>)</mml:mo><mml:mo>=</mml:mo><mml:mstyle displaystyle="true"><mml:mfrac style="display"><mml:mn mathvariant="normal">1</mml:mn><mml:mrow><mml:mn mathvariant="normal">2</mml:mn><mml:mi>N</mml:mi><mml:mo>(</mml:mo><mml:mi>N</mml:mi><mml:mo>-</mml:mo><mml:mn mathvariant="normal">1</mml:mn><mml:mo>)</mml:mo></mml:mrow></mml:mfrac></mml:mstyle><mml:msubsup><mml:mo movablelimits="false">∑</mml:mo><mml:mrow><mml:mi>i</mml:mi><mml:mo>≠</mml:mo><mml:mi>j</mml:mi></mml:mrow><mml:mi>N</mml:mi></mml:msubsup><mml:mo>[</mml:mo><mml:msub><mml:mi>x</mml:mi><mml:mi>i</mml:mi></mml:msub><mml:mo>(</mml:mo><mml:mi>t</mml:mi><mml:mo>)</mml:mo><mml:mo>-</mml:mo><mml:msub><mml:mi>x</mml:mi><mml:mi>j</mml:mi></mml:msub><mml:mo>(</mml:mo><mml:mi>t</mml:mi><mml:mo>)</mml:mo><mml:msup><mml:mo>]</mml:mo><mml:mn mathvariant="normal">2</mml:mn></mml:msup><mml:mo>,</mml:mo></mml:mrow></mml:math></disp-formula>
          where <inline-formula><mml:math id="M138" display="inline"><mml:mi>N</mml:mi></mml:math></inline-formula> here is the total number of SVP drifters and <inline-formula><mml:math id="M139" display="inline"><mml:mi>x</mml:mi></mml:math></inline-formula> the time series of the drifter <inline-formula><mml:math id="M140" display="inline"><mml:mi>i</mml:mi></mml:math></inline-formula>'s <inline-formula><mml:math id="M141" display="inline"><mml:mrow><mml:mi>x</mml:mi><mml:mo>,</mml:mo><mml:mi>y</mml:mi></mml:mrow></mml:math></inline-formula> position.</p>

      <?xmltex \floatpos{t}?><fig id="Ch1.F2"><caption><p id="d1e2363">Satellite chl <italic>a</italic> around <bold>(a)</bold> LDA, <bold>(b)</bold> LDB, and <bold>(c)</bold> LDC. SD
stations shown by black crosses, land is shaded gray, and coastlines and
reefs are
plotted in black. LD stations shown by plus symbols following the color code from Fig. 1.
Squares with 120 <inline-formula><mml:math id="M142" display="inline"><mml:mi mathvariant="normal">km</mml:mi></mml:math></inline-formula> to a side plotted around each LD station to
represent approximate Rossby radius <inline-formula><mml:math id="M143" display="inline"><mml:mrow><mml:msub><mml:mi>R</mml:mi><mml:mi>D</mml:mi></mml:msub></mml:mrow></mml:math></inline-formula>.</p></caption>
          <?xmltex \igopts{width=236.157874pt}?><graphic xlink:href="https://bg.copernicus.org/articles/15/2125/2018/bg-15-2125-2018-f02.pdf"/>

        </fig>

      <?xmltex \floatpos{t}?><fig id="Ch1.F3" specific-use="star"><caption><p id="d1e2405">Time series of surface chl <italic>a</italic> and SST, respectively, for
<bold>(a, b)</bold> LDA, <bold>(c, d)</bold> LDB, and <bold>(e, f)</bold> LDC. Intervals of LD sampling shown with gray
dashed lines. Mean values are plotted in black, with darker shades
representing the 25–75 <inline-formula><mml:math id="M144" display="inline"><mml:mi mathvariant="normal">%</mml:mi></mml:math></inline-formula> interval and lighter shades for 1–99 <inline-formula><mml:math id="M145" display="inline"><mml:mi mathvariant="normal">%</mml:mi></mml:math></inline-formula>.
Subpanels above each time series depict the <inline-formula><mml:math id="M146" display="inline"><mml:mi mathvariant="italic">%</mml:mi></mml:math></inline-formula> of pixels with data. All data
come from within the 120 <inline-formula><mml:math id="M147" display="inline"><mml:mi mathvariant="normal">km</mml:mi></mml:math></inline-formula> <inline-formula><mml:math id="M148" display="inline"><mml:mo>×</mml:mo></mml:math></inline-formula> 120 <inline-formula><mml:math id="M149" display="inline"><mml:mi mathvariant="normal">km</mml:mi></mml:math></inline-formula> squares shown in
Fig. 2.</p></caption>
          <?xmltex \igopts{width=398.338583pt}?><graphic xlink:href="https://bg.copernicus.org/articles/15/2125/2018/bg-15-2125-2018-f03.pdf"/>

        </fig>

</sec>
</sec>
<sec id="Ch1.S3">
  <title>Results</title>
<sec id="Ch1.S3.SS1">
  <title>Satellite data, temporal context, and method applicability</title>
      <p id="d1e2481">The regional distributions of SST and surface chl <italic>a</italic> as seen during
the OUTPACE cruise are shown in Fig. 1. The data in Fig. 1 are weighted
means, with the weight being the inverse square of the ship's daily distance
to each pixel. A north–south meridional gradient was found in SST, with
warmer water near the equator (<inline-formula><mml:math id="M150" display="inline"><mml:mo lspace="0mm">∼</mml:mo></mml:math></inline-formula> 30 <inline-formula><mml:math id="M151" display="inline"><mml:mrow><mml:msup><mml:mi/><mml:mo>∘</mml:mo></mml:msup><mml:mi mathvariant="normal">C</mml:mi></mml:mrow></mml:math></inline-formula>) and cooler water
poleward (<inline-formula><mml:math id="M152" display="inline"><mml:mo lspace="0mm">∼</mml:mo></mml:math></inline-formula> 25 <inline-formula><mml:math id="M153" display="inline"><mml:mrow><mml:msup><mml:mi/><mml:mo>∘</mml:mo></mml:msup><mml:mi mathvariant="normal">C</mml:mi></mml:mrow></mml:math></inline-formula>). This gradient was uninterrupted for the
duration of the OUTPACE cruise. Due to the zonal ship track the surface
thermal conditions observed by the ship during OUTPACE were relatively
homogeneous. Furthermore, no strong temperature gradients, indicative of
frontal or eddy structures, were visible in the vicinity of the stations.
While a north–south regional gradient was found in SST, the opposite was
found in chl <italic>a</italic>. Chl <italic>a</italic> values were around 0.3 <inline-formula><mml:math id="M154" display="inline"><mml:mrow><mml:mi mathvariant="normal">mg</mml:mi><mml:mspace linebreak="nobreak" width="0.125em"/><mml:msup><mml:mi mathvariant="normal">m</mml:mi><mml:mrow><mml:mo>-</mml:mo><mml:mn mathvariant="normal">3</mml:mn></mml:mrow></mml:msup></mml:mrow></mml:math></inline-formula>
in the western portion of the domain, west of 190<inline-formula><mml:math id="M155" display="inline"><mml:msup><mml:mi/><mml:mo>∘</mml:mo></mml:msup></mml:math></inline-formula> E.
Stations<?pagebreak page2132?> LDA and LDB were in this region, with LDB positioned inside a bloom
with values near 1 <inline-formula><mml:math id="M156" display="inline"><mml:mrow><mml:mi mathvariant="normal">mg</mml:mi><mml:mspace width="0.125em" linebreak="nobreak"/><mml:msup><mml:mi mathvariant="normal">m</mml:mi><mml:mrow><mml:mo>-</mml:mo><mml:mn mathvariant="normal">3</mml:mn></mml:mrow></mml:msup></mml:mrow></mml:math></inline-formula>. More details concerning the LDB
bloom can be found in <xref ref-type="bibr" rid="bib1.bibx11" id="text.28"/>. Chl <italic>a</italic> values dropped
precipitously, over an order of magnitude to 0.03 <inline-formula><mml:math id="M157" display="inline"><mml:mrow><mml:mi mathvariant="normal">mg</mml:mi><mml:mspace width="0.125em" linebreak="nobreak"/><mml:msup><mml:mi mathvariant="normal">m</mml:mi><mml:mrow><mml:mo>-</mml:mo><mml:mn mathvariant="normal">3</mml:mn></mml:mrow></mml:msup></mml:mrow></mml:math></inline-formula>,
just east of LDB near LDC. The low value of chl <italic>a</italic> was indicative of
the South Pacific Gyre (SPG; <xref ref-type="bibr" rid="bib1.bibx7" id="author.29"/>,
<xref ref-type="bibr" rid="bib1.bibx7" id="year.30"/>).</p>
      <p id="d1e2608">Since SST was relatively unchanging during OUTPACE, Fig. 2 provides zoomed-in
views of the chl <italic>a</italic> data for the three LD stations, with domains
chosen to include the nearest SD stations. The spacing of the SD stations was
relatively regular along the OUTPACE transect (Fig. 1b). In Fig. 2a the
enhanced chl <italic>a</italic> was distributed evenly inside the domain, so no clear
surface gradients are present. In Fig. 2b, the chl <italic>a</italic> was
concentrated in the aforementioned bloom, with values higher than those seen
in Fig. 2a near LDA. The size of the bloom was large enough to cover most of
the <inline-formula><mml:math id="M158" display="inline"><mml:mrow><mml:mn mathvariant="normal">120</mml:mn><mml:mspace linebreak="nobreak" width="0.125em"/><mml:mi mathvariant="normal">km</mml:mi><mml:mo>×</mml:mo><mml:mn mathvariant="normal">120</mml:mn><mml:mspace linebreak="nobreak" width="0.125em"/><mml:mi mathvariant="normal">km</mml:mi></mml:mrow></mml:math></inline-formula> region shown in Fig. 2b, so the bloom edges
were far away from station LDB's initial position. By contrast, waters
outside the bloom had chl <italic>a</italic> values lower than in Fig. 2a. The low
chl <italic>a</italic> values near LDC in Fig. 2c were typical of the SPG, and no
visible patches of chl <italic>a</italic> indicated sharp gradients.</p>

      <?xmltex \floatpos{t}?><fig id="Ch1.F4" specific-use="star"><caption><p id="d1e2650"><inline-formula><mml:math id="M159" display="inline"><mml:mi>T</mml:mi></mml:math></inline-formula>-<inline-formula><mml:math id="M160" display="inline"><mml:mi>S</mml:mi></mml:math></inline-formula> diagrams of <bold>(a)</bold> LDA, <bold>(b)</bold> LDB, and <bold>(c)</bold> LDC and surrounding
stations. LD stations are color coded, and SD stations different shades of
gray. Isopycnals are displayed in black, with isopleths of spice shown in
red.</p></caption>
          <?xmltex \igopts{width=426.791339pt}?><graphic xlink:href="https://bg.copernicus.org/articles/15/2125/2018/bg-15-2125-2018-f04.pdf"/>

        </fig>

      <p id="d1e2681">The time series of chl <italic>a</italic> and SST for the three stations are shown in
Fig. 3. Comparing the three LD stations, a few patterns emerge. SST showed
similar trends across the three LD stations. All stations experienced warming
trends from December 2014 to mid-March 2015, consistent with summer heating.
The lack of data from cloud cover sometimes led to abrupt drops in the
distribution of daily SST shown. However, the timing of maximum temperature and the
magnitude of that warming did differ between LD stations. A rapid
heating in December 2014 occurred around LDA's position, which then slowly
continued until early March 2015, at which point temperatures began to drop.
Towards the end of sampling at station LDA the SST rises, possibly indicating
another warming event occurred or the arrival of a warm patch of water.
Depth-resolved application of our method in the later sections will evaluate
this possibility. The overall evolution in LDA's temperature during the
period shown, from <inline-formula><mml:math id="M161" display="inline"><mml:mo>∼</mml:mo></mml:math></inline-formula> 26 to 30 <inline-formula><mml:math id="M162" display="inline"><mml:mrow><mml:msup><mml:mi/><mml:mo>∘</mml:mo></mml:msup><mml:mi mathvariant="normal">C</mml:mi></mml:mrow></mml:math></inline-formula>, represented a 4 <inline-formula><mml:math id="M163" display="inline"><mml:mrow><mml:msup><mml:mi/><mml:mo>∘</mml:mo></mml:msup><mml:mi mathvariant="normal">C</mml:mi></mml:mrow></mml:math></inline-formula>
change. LDB showed a slight cooling in December 2014, but this may have been
an artifact of cloud cover. Station sampling for LDB occurred immediately
after the maximum heating, though the values seen at LDB were relatively
stable and slightly warmer than at LDA. The maxima in temperature for LDA and
LDB seemed to overlap in time in early March 2015. LDB's change in
temperature, from <inline-formula><mml:math id="M164" display="inline"><mml:mo>∼</mml:mo></mml:math></inline-formula> 27 to 30 <inline-formula><mml:math id="M165" display="inline"><mml:mrow><mml:msup><mml:mi/><mml:mo>∘</mml:mo></mml:msup><mml:mi mathvariant="normal">C</mml:mi></mml:mrow></mml:math></inline-formula>, was a 3 <inline-formula><mml:math id="M166" display="inline"><mml:mrow><mml:msup><mml:mi/><mml:mo>∘</mml:mo></mml:msup><mml:mi mathvariant="normal">C</mml:mi></mml:mrow></mml:math></inline-formula> change. LDC had
the smallest change in temperature, from <inline-formula><mml:math id="M167" display="inline"><mml:mo>∼</mml:mo></mml:math></inline-formula> 27 to 29 <inline-formula><mml:math id="M168" display="inline"><mml:mrow><mml:msup><mml:mi/><mml:mo>∘</mml:mo></mml:msup><mml:mi mathvariant="normal">C</mml:mi></mml:mrow></mml:math></inline-formula> for a
2 <inline-formula><mml:math id="M169" display="inline"><mml:mrow><mml:msup><mml:mi/><mml:mo>∘</mml:mo></mml:msup><mml:mi mathvariant="normal">C</mml:mi></mml:mrow></mml:math></inline-formula> change. Sampling for LDC coincided with the warmest period
observed in the satellite data, in late March 2015, and was stable for the LD
sampling period.</p>
      <p id="d1e2782">The timing of temperature maxima is important to note for biological reasons,
since <inline-formula><mml:math id="M170" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">N</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula> fixation by <italic>Trichodesmium</italic> spp. is known to occur in
warm, stratified waters (specifically, a <inline-formula><mml:math id="M171" display="inline"><mml:mo>∼</mml:mo></mml:math></inline-formula> 25 <inline-formula><mml:math id="M172" display="inline"><mml:mrow><mml:msup><mml:mi/><mml:mo>∘</mml:mo></mml:msup><mml:mi mathvariant="normal">C</mml:mi></mml:mrow></mml:math></inline-formula>
threshold, <xref ref-type="bibr" rid="bib1.bibx45" id="author.31"/>, <xref ref-type="bibr" rid="bib1.bibx45" id="year.32"/>) and
one of the goals of OUTPACE was to observe this biogeochemical process. Since
SST was above 25 <inline-formula><mml:math id="M173" display="inline"><mml:mrow><mml:msup><mml:mi/><mml:mo>∘</mml:mo></mml:msup><mml:mi mathvariant="normal">C</mml:mi></mml:mrow></mml:math></inline-formula> for all stations throughout this period,
the thermal conditions during OUTPACE would not have limited <inline-formula><mml:math id="M174" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">N</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula> fixation.</p>
      <p id="d1e2848">In between December 2014 and January 2015, the region around LDA had higher
chl <italic>a</italic> concentrations than LDB. The period between February and May 2015
showed a remarkable increase in chl <italic>a</italic> near the LDB site. This
was due to advection of the surface bloom, which subsequently collapsed and
advected away, as documented in another study in this special issue
<xref ref-type="bibr" rid="bib1.bibx11" id="paren.33"/>. The downward trend of chl <italic>a</italic> during this period
is more indicative of in situ evolution, rather than advection of the bloom's
boundaries, and does not invalidate subsequent use of our method. Near<?pagebreak page2133?> LDC,
chl <italic>a</italic> was systematically low, a reflection of the goals of OUTPACE
to sample in the SPG.</p>
      <p id="d1e2866">Besides the increase in SST at the end of LDA and the decrease in
chl <italic>a</italic> during LDB, both SST and chl <italic>a</italic> for the LD stations
were stable, providing evidence that no surface gradients, physical or
biological, immediately invalidate the application of our strategy. Though
the change in chl <italic>a</italic> at LDB has been argued to be due to endogenous
dynamics in the aforementioned study, application of our post-validation
method provides an independent test of whether advection of gradients could
be responsible. Likewise, the method will also determine whether the surface
increase in SST during LDA was reflective of changes at depth. As a final
note, <xref ref-type="bibr" rid="bib1.bibx39" id="text.34"/> found with satellite-derived surface
velocities that during LDC a coherent structure was present, despite the lack
of surface SST and chl <italic>a</italic> gradients. Since tracer gradients are
non-existent at the surface, we find that this is not a strong structure, and
does not invalidate the applicability of our approach. Rather, the
application of depth-resolved in situ measurements of tracers and velocities
will serve as an independent evaluation of this finding.</p>
</sec>
<?pagebreak page2134?><sec id="Ch1.S3.SS2">
  <title>In situ properties, statistical baseline, and time series analysis</title>
      <p id="d1e2890">The hydrographic variability during the three LD stations and surrounding SD
stations are shown in the <inline-formula><mml:math id="M175" display="inline"><mml:mi>T</mml:mi></mml:math></inline-formula>-<inline-formula><mml:math id="M176" display="inline"><mml:mi>S</mml:mi></mml:math></inline-formula> diagrams of Fig. 4. All three stations followed
a general pattern, where surface water near the 1022 <inline-formula><mml:math id="M177" display="inline"><mml:mrow><mml:mi mathvariant="normal">kg</mml:mi><mml:mspace width="0.125em" linebreak="nobreak"/><mml:msup><mml:mi mathvariant="normal">m</mml:mi><mml:mrow><mml:mo>-</mml:mo><mml:mn mathvariant="normal">3</mml:mn></mml:mrow></mml:msup></mml:mrow></mml:math></inline-formula>
isopycnal and 29 <inline-formula><mml:math id="M178" display="inline"><mml:mrow><mml:msup><mml:mi/><mml:mo>∘</mml:mo></mml:msup><mml:mi mathvariant="normal">C</mml:mi></mml:mrow></mml:math></inline-formula> temperature (though LDB had warmer surface water,
Fig. 4b) dropped in temperature and increased in salinity until a subsurface
salinity maximum near the 1025 <inline-formula><mml:math id="M179" display="inline"><mml:mrow><mml:mi mathvariant="normal">kg</mml:mi><mml:mspace linebreak="nobreak" width="0.125em"/><mml:msup><mml:mi mathvariant="normal">m</mml:mi><mml:mrow><mml:mo>-</mml:mo><mml:mn mathvariant="normal">3</mml:mn></mml:mrow></mml:msup></mml:mrow></mml:math></inline-formula> isopycnal. The
increase in salinity maximum from LDA to LDC reflects the high salinity
tongue of the South Pacific <xref ref-type="bibr" rid="bib1.bibx20" id="paren.35"/>. The surface water
in LDA (Fig. 4a) showed a bifurcation. This change was manifest in the
satellite data time series, as well. Whether this is due to a heating event or
the arrival of new water at the end of LDA will be addressed in the
time series analysis below. For LDA, neighboring stations SD2, 3, and 4
largely overlapped with the LDA profile. SD3, the station closest to LDA,
almost entirely overlapped the LD profile, except for a subsurface salinity
deviation below the 1024.5 <inline-formula><mml:math id="M180" display="inline"><mml:mrow><mml:mi mathvariant="normal">kg</mml:mi><mml:mspace width="0.125em" linebreak="nobreak"/><mml:msup><mml:mi mathvariant="normal">m</mml:mi><mml:mrow><mml:mo>-</mml:mo><mml:mn mathvariant="normal">3</mml:mn></mml:mrow></mml:msup></mml:mrow></mml:math></inline-formula> isopycnal. SD2 and SD4
showed greater deviations, with SD4 being saltier than LDA for almost the
entire profile. Similar overlaps occurred with LDB and its surrounding
stations, SD12 and SD13 (Fig. 4b). SD12 showed lower salinity near the
surface, with a kink in salinity at the 1025 <inline-formula><mml:math id="M181" display="inline"><mml:mrow><mml:mi mathvariant="normal">kg</mml:mi><mml:mspace linebreak="nobreak" width="0.125em"/><mml:msup><mml:mi mathvariant="normal">m</mml:mi><mml:mrow><mml:mo>-</mml:mo><mml:mn mathvariant="normal">3</mml:mn></mml:mrow></mml:msup></mml:mrow></mml:math></inline-formula>
isopycnal. The salinity offsets of SD4 and SD12 at depth are within
climatological variability (Figs. S1 and S2). SD13 had similar surface
structure to LDB, but higher salinities from the 1023.5 to 1025 <inline-formula><mml:math id="M182" display="inline"><mml:mrow><mml:mi mathvariant="normal">kg</mml:mi><mml:mspace width="0.125em" linebreak="nobreak"/><mml:msup><mml:mi mathvariant="normal">m</mml:mi><mml:mrow><mml:mo>-</mml:mo><mml:mn mathvariant="normal">3</mml:mn></mml:mrow></mml:msup></mml:mrow></mml:math></inline-formula>
isopycnal. The LDC, SD13, and SD14 (Fig. 4c) profiles nearly
entirely overlapped except near the surface when the SD stations were at
first less salty at the surface and then became more salty. Additionally, the
saltier nature of LDC relative to LDA and LDB, especially between 1024 and
1025 <inline-formula><mml:math id="M183" display="inline"><mml:mrow><mml:mi mathvariant="normal">kg</mml:mi><mml:mspace linebreak="nobreak" width="0.125em"/><mml:msup><mml:mi mathvariant="normal">m</mml:mi><mml:mrow><mml:mo>-</mml:mo><mml:mn mathvariant="normal">3</mml:mn></mml:mrow></mml:msup></mml:mrow></mml:math></inline-formula>, was visible. The variability in <inline-formula><mml:math id="M184" display="inline"><mml:mi>T</mml:mi></mml:math></inline-formula>-<inline-formula><mml:math id="M185" display="inline"><mml:mi>S</mml:mi></mml:math></inline-formula> values
between stations was within the range seen in the climatology of the region
(Figs. S1 and S2).</p>

      <?xmltex \floatpos{t}?><fig id="Ch1.F5"><caption><p id="d1e3042">Statistical LD baseline of spice versus potential density for
<bold>(a)</bold> LDA, <bold>(b)</bold> LDC, and <bold>(c)</bold> LDC. Mean values plotted in between envelope of <inline-formula><mml:math id="M186" display="inline"><mml:mo>±</mml:mo></mml:math></inline-formula>2
<inline-formula><mml:math id="M187" display="inline"><mml:mrow><mml:mi>S</mml:mi><mml:msub><mml:mtext>Err</mml:mtext><mml:mi mathvariant="normal">obs</mml:mi></mml:msub></mml:mrow></mml:math></inline-formula>. </p></caption>
          <?xmltex \igopts{width=199.169291pt}?><graphic xlink:href="https://bg.copernicus.org/articles/15/2125/2018/bg-15-2125-2018-f05.pdf"/>

        </fig>

      <?xmltex \floatpos{t}?><fig id="Ch1.F6" specific-use="star"><caption><p id="d1e3082">SedTrap drifter <inline-formula><mml:math id="M188" display="inline"><mml:mi>Z</mml:mi></mml:math></inline-formula>-score time series for <bold>(a)</bold> 14, <bold>(b)</bold> 55, <bold>(c)</bold> 88,
<bold>(d)</bold> 105, <bold>(e)</bold> 137, and <bold>(f)</bold> 197 <inline-formula><mml:math id="M189" display="inline"><mml:mi mathvariant="normal">m</mml:mi></mml:math></inline-formula> depth. End of inertial period timeframe
for baseline definition plotted in magenta. <inline-formula><mml:math id="M190" display="inline"><mml:mrow><mml:mi>Z</mml:mi><mml:mo>=</mml:mo><mml:mo>-</mml:mo><mml:mn mathvariant="normal">2</mml:mn></mml:mrow></mml:math></inline-formula> and 2 plotted in black.</p></caption>
          <?xmltex \igopts{width=341.433071pt}?><graphic xlink:href="https://bg.copernicus.org/articles/15/2125/2018/bg-15-2125-2018-f06.pdf"/>

        </fig>

      <?xmltex \floatpos{t}?><fig id="Ch1.F7" specific-use="star"><caption><p id="d1e3141">Same as Fig. 6 but for LDB.</p></caption>
          <?xmltex \igopts{width=341.433071pt}?><graphic xlink:href="https://bg.copernicus.org/articles/15/2125/2018/bg-15-2125-2018-f07.pdf"/>

        </fig>

      <?xmltex \floatpos{t}?><fig id="Ch1.F8" specific-use="star"><caption><p id="d1e3152">Same as Fig. 6 but for LDC.</p></caption>
          <?xmltex \igopts{width=341.433071pt}?><graphic xlink:href="https://bg.copernicus.org/articles/15/2125/2018/bg-15-2125-2018-f08.pdf"/>

        </fig>

      <?xmltex \floatpos{t}?><fig id="Ch1.F9"><caption><p id="d1e3163">Time series of CTD observations for LDA. Observations with <inline-formula><mml:math id="M191" display="inline"><mml:mrow><mml:mo>|</mml:mo><mml:mi>Z</mml:mi><mml:mo>|</mml:mo><mml:mo>&lt;</mml:mo><mml:mn mathvariant="normal">2</mml:mn></mml:mrow></mml:math></inline-formula>
plotted in green, <inline-formula><mml:math id="M192" display="inline"><mml:mrow><mml:mo>|</mml:mo><mml:mi>Z</mml:mi><mml:mo>|</mml:mo><mml:mo>&gt;</mml:mo><mml:mn mathvariant="normal">2</mml:mn></mml:mrow></mml:math></inline-formula> in black. End of statistical baseline definition
period shown in magenta.</p></caption>
          <?xmltex \igopts{width=236.157874pt}?><graphic xlink:href="https://bg.copernicus.org/articles/15/2125/2018/bg-15-2125-2018-f09.pdf"/>

        </fig>

      <?xmltex \floatpos{t}?><fig id="Ch1.F10"><caption><p id="d1e3206">Same as Fig. 9 for LDB, with <inline-formula><mml:math id="M193" display="inline"><mml:mrow><mml:mo>|</mml:mo><mml:mi>Z</mml:mi><mml:mo>|</mml:mo><mml:mo>&lt;</mml:mo><mml:mn mathvariant="normal">2</mml:mn></mml:mrow></mml:math></inline-formula> observations plotted in red.</p></caption>
          <?xmltex \igopts{width=236.157874pt}?><graphic xlink:href="https://bg.copernicus.org/articles/15/2125/2018/bg-15-2125-2018-f10.pdf"/>

        </fig>

      <?xmltex \floatpos{t}?><fig id="Ch1.F11"><caption><p id="d1e3234">Same as Fig. 9 for LDC, with <inline-formula><mml:math id="M194" display="inline"><mml:mrow><mml:mo>|</mml:mo><mml:mi>Z</mml:mi><mml:mo>|</mml:mo><mml:mo>&lt;</mml:mo><mml:mn mathvariant="normal">2</mml:mn></mml:mrow></mml:math></inline-formula> observations plotted in blue.</p></caption>
          <?xmltex \igopts{width=236.157874pt}?><graphic xlink:href="https://bg.copernicus.org/articles/15/2125/2018/bg-15-2125-2018-f11.pdf"/>

        </fig>

      <p id="d1e3259">The LD statistical baselines of spice in density space, with means and
intervals of two standard errors, are shown in Fig. 5. These standard error
intervals, representing the inherent variability in the baseline, show the
values wherein a <inline-formula><mml:math id="M195" display="inline"><mml:mi>Z</mml:mi></mml:math></inline-formula>-score of <inline-formula><mml:math id="M196" display="inline"><mml:mo>≤</mml:mo></mml:math></inline-formula> 2 was achieved. LDB and LDC overlapped for
essentially their entire profiles. All stations are missing observations near
the surface and mixed layer due to the<?pagebreak page2135?> intense stratification which left
several density bins with less than 50 observations, the threshold used in
the spice analysis. LDA was noticeably less spicy than the other two LD
stations for density less than 1024 <inline-formula><mml:math id="M197" display="inline"><mml:mrow><mml:mi mathvariant="normal">kg</mml:mi><mml:mspace linebreak="nobreak" width="0.125em"/><mml:msup><mml:mi mathvariant="normal">m</mml:mi><mml:mrow><mml:mo>-</mml:mo><mml:mn mathvariant="normal">3</mml:mn></mml:mrow></mml:msup></mml:mrow></mml:math></inline-formula>. At the highest
densities, all three LD stations overlapped. The envelope of two standard
errors, or <inline-formula><mml:math id="M198" display="inline"><mml:mi>Z</mml:mi></mml:math></inline-formula>-score <inline-formula><mml:math id="M199" display="inline"><mml:mo>≥</mml:mo></mml:math></inline-formula> 2, show that variability has some dependence on
depth. The LDA baseline shows high variability near the surface, with a thin
envelope below down to 1024 <inline-formula><mml:math id="M200" display="inline"><mml:mrow><mml:mi mathvariant="normal">kg</mml:mi><mml:mspace linebreak="nobreak" width="0.125em"/><mml:msup><mml:mi mathvariant="normal">m</mml:mi><mml:mrow><mml:mo>-</mml:mo><mml:mn mathvariant="normal">3</mml:mn></mml:mrow></mml:msup></mml:mrow></mml:math></inline-formula>, and widening at depth
down to 1025 <inline-formula><mml:math id="M201" display="inline"><mml:mrow><mml:mi mathvariant="normal">kg</mml:mi><mml:mspace linebreak="nobreak" width="0.125em"/><mml:msup><mml:mi mathvariant="normal">m</mml:mi><mml:mrow><mml:mo>-</mml:mo><mml:mn mathvariant="normal">3</mml:mn></mml:mrow></mml:msup></mml:mrow></mml:math></inline-formula> and beyond. LDB did not have high
surface variability, but the envelope widens shortly below 1023 <inline-formula><mml:math id="M202" display="inline"><mml:mrow><mml:mi mathvariant="normal">kg</mml:mi><mml:mspace linebreak="nobreak" width="0.125em"/><mml:msup><mml:mi mathvariant="normal">m</mml:mi><mml:mrow><mml:mo>-</mml:mo><mml:mn mathvariant="normal">3</mml:mn></mml:mrow></mml:msup></mml:mrow></mml:math></inline-formula>.
Variability in LDC shows similar widening as in LDB, with a
noticeable pinch in the envelope near 1024.5 <inline-formula><mml:math id="M203" display="inline"><mml:mrow><mml:mi mathvariant="normal">kg</mml:mi><mml:mspace linebreak="nobreak" width="0.125em"/><mml:msup><mml:mi mathvariant="normal">m</mml:mi><mml:mrow><mml:mo>-</mml:mo><mml:mn mathvariant="normal">3</mml:mn></mml:mrow></mml:msup></mml:mrow></mml:math></inline-formula>.</p>
      <?pagebreak page2136?><p id="d1e3376">The <inline-formula><mml:math id="M204" display="inline"><mml:mi>Z</mml:mi></mml:math></inline-formula>-score time series for the SedTrap Drifter sensors are shown in Figs. 6–8.
During LDA, at 14 <inline-formula><mml:math id="M205" display="inline"><mml:mi mathvariant="normal">m</mml:mi></mml:math></inline-formula> depth (Fig. 6a), after the inertial period
baseline determination the <inline-formula><mml:math id="M206" display="inline"><mml:mi>Z</mml:mi></mml:math></inline-formula>-score first descended, increased, and then
leveled off after the first two and a half days. Afterwards, the <inline-formula><mml:math id="M207" display="inline"><mml:mi>Z</mml:mi></mml:math></inline-formula>-score
increased, reaching 2, decreased, and surpassed <inline-formula><mml:math id="M208" display="inline"><mml:mrow><mml:mi>Z</mml:mi><mml:mo>=</mml:mo><mml:mn mathvariant="normal">2</mml:mn></mml:mrow></mml:math></inline-formula> before falling again. The
SedTrap drifter at 55 <inline-formula><mml:math id="M209" display="inline"><mml:mi mathvariant="normal">m</mml:mi></mml:math></inline-formula> showed no trend (Fig. 6b), and a single
<inline-formula><mml:math id="M210" display="inline"><mml:mi>Z</mml:mi></mml:math></inline-formula>-score was seen below <inline-formula><mml:math id="M211" display="inline"><mml:mo>-</mml:mo></mml:math></inline-formula>2. <inline-formula><mml:math id="M212" display="inline"><mml:mi>Z</mml:mi></mml:math></inline-formula>-scores for 105, 137, and 197 <inline-formula><mml:math id="M213" display="inline"><mml:mi mathvariant="normal">m</mml:mi></mml:math></inline-formula> (Fig. 6d–f)
showed no temporal trends and were always less than 2 in magnitude. The
time series at 88 <inline-formula><mml:math id="M214" display="inline"><mml:mi mathvariant="normal">m</mml:mi></mml:math></inline-formula> showed no trend but the variability in <inline-formula><mml:math id="M215" display="inline"><mml:mi>Z</mml:mi></mml:math></inline-formula>-score
increased over time, with some observations surpassing <inline-formula><mml:math id="M216" display="inline"><mml:mrow><mml:mo>|</mml:mo><mml:mi>Z</mml:mi><mml:mo>|</mml:mo><mml:mo>=</mml:mo><mml:mn mathvariant="normal">2</mml:mn></mml:mrow></mml:math></inline-formula>.</p>
      <p id="d1e3486">LDB SedTrap drifter <inline-formula><mml:math id="M217" display="inline"><mml:mi>Z</mml:mi></mml:math></inline-formula>-scores (Fig. 7) showed similar patterns to LDA. The
surface sensor (Fig. 7a) decreased and increased over the first two days,
then leveled with temporary departures below <inline-formula><mml:math id="M218" display="inline"><mml:mo>-</mml:mo></mml:math></inline-formula>2. The sensors at 55, 105, 137,
and 197 <inline-formula><mml:math id="M219" display="inline"><mml:mi mathvariant="normal">m</mml:mi></mml:math></inline-formula> (Fig. 7b, d–f), similar to LDA showed no trends and low
variability. A few observations below <inline-formula><mml:math id="M220" display="inline"><mml:mo>-</mml:mo></mml:math></inline-formula>2 occurred at 55 <inline-formula><mml:math id="M221" display="inline"><mml:mi mathvariant="normal">m</mml:mi></mml:math></inline-formula>. The
<inline-formula><mml:math id="M222" display="inline"><mml:mi>Z</mml:mi></mml:math></inline-formula>-scores at 88 <inline-formula><mml:math id="M223" display="inline"><mml:mi mathvariant="normal">m</mml:mi></mml:math></inline-formula> showed no trend, similar to LDA with enhanced
variability and some <inline-formula><mml:math id="M224" display="inline"><mml:mrow><mml:mo>|</mml:mo><mml:mi>Z</mml:mi><mml:mo>|</mml:mo><mml:mo>&gt;</mml:mo><mml:mn mathvariant="normal">2</mml:mn></mml:mrow></mml:math></inline-formula> but no time-dependence.</p>
      <p id="d1e3555">The LDC <inline-formula><mml:math id="M225" display="inline"><mml:mi>Z</mml:mi></mml:math></inline-formula>-scores were large at more depths than the other LD stations (Fig. 8).
Values at 14 <inline-formula><mml:math id="M226" display="inline"><mml:mi mathvariant="normal">m</mml:mi></mml:math></inline-formula> started with <inline-formula><mml:math id="M227" display="inline"><mml:mrow><mml:mi>Z</mml:mi><mml:mo>&gt;</mml:mo><mml:mn mathvariant="normal">2</mml:mn></mml:mrow></mml:math></inline-formula>, but that dropped before rising
again after a day, before slowly dropping and eventually decreasing to
<inline-formula><mml:math id="M228" display="inline"><mml:mo>∼</mml:mo></mml:math></inline-formula> <inline-formula><mml:math id="M229" display="inline"><mml:mo>-</mml:mo></mml:math></inline-formula>2 at the end of LDC. <inline-formula><mml:math id="M230" display="inline"><mml:mi>Z</mml:mi></mml:math></inline-formula>-scores at 55, 137, and 197 <inline-formula><mml:math id="M231" display="inline"><mml:mi mathvariant="normal">m</mml:mi></mml:math></inline-formula> showed no
trends in <inline-formula><mml:math id="M232" display="inline"><mml:mi>Z</mml:mi></mml:math></inline-formula>-score, and had limited observations with <inline-formula><mml:math id="M233" display="inline"><mml:mrow><mml:mo>|</mml:mo><mml:mi>Z</mml:mi><mml:mo>|</mml:mo><mml:mo>&gt;</mml:mo><mml:mn mathvariant="normal">2</mml:mn></mml:mrow></mml:math></inline-formula>. At 88 <inline-formula><mml:math id="M234" display="inline"><mml:mi mathvariant="normal">m</mml:mi></mml:math></inline-formula>,
no trend was seen, and for the first two days there were few observations
with <inline-formula><mml:math id="M235" display="inline"><mml:mrow><mml:mi>Z</mml:mi><mml:mo>&lt;</mml:mo><mml:mo>-</mml:mo><mml:mn mathvariant="normal">2</mml:mn></mml:mrow></mml:math></inline-formula>. Toward the end of LDC, two spikes with <inline-formula><mml:math id="M236" display="inline"><mml:mrow><mml:mi>Z</mml:mi><mml:mo>&gt;</mml:mo><mml:mn mathvariant="normal">2</mml:mn></mml:mrow></mml:math></inline-formula>, with <inline-formula><mml:math id="M237" display="inline"><mml:mrow><mml:mi>Z</mml:mi><mml:mo>∼</mml:mo><mml:mn mathvariant="normal">4</mml:mn></mml:mrow></mml:math></inline-formula>–5,
occurred with returns back to <inline-formula><mml:math id="M238" display="inline"><mml:mrow><mml:mo>|</mml:mo><mml:mi>Z</mml:mi><mml:mo>|</mml:mo><mml:mo>&lt;</mml:mo><mml:mn mathvariant="normal">2</mml:mn></mml:mrow></mml:math></inline-formula>. The <inline-formula><mml:math id="M239" display="inline"><mml:mi>Z</mml:mi></mml:math></inline-formula>-scores at this depth ended near
<inline-formula><mml:math id="M240" display="inline"><mml:mrow><mml:mi>Z</mml:mi><mml:mo>=</mml:mo><mml:mn mathvariant="normal">2</mml:mn></mml:mrow></mml:math></inline-formula>. Observations at 105 <inline-formula><mml:math id="M241" display="inline"><mml:mi mathvariant="normal">m</mml:mi></mml:math></inline-formula> started around <inline-formula><mml:math id="M242" display="inline"><mml:mrow><mml:mo>-</mml:mo><mml:mn mathvariant="normal">2</mml:mn><mml:mo>&lt;</mml:mo><mml:mi>Z</mml:mi><mml:mo>&lt;</mml:mo><mml:mn mathvariant="normal">0</mml:mn></mml:mrow></mml:math></inline-formula>, but spikes with
<inline-formula><mml:math id="M243" display="inline"><mml:mrow><mml:mi>Z</mml:mi><mml:mo>∼</mml:mo><mml:mn mathvariant="normal">2</mml:mn></mml:mrow></mml:math></inline-formula> occurred. Over time, <inline-formula><mml:math id="M244" display="inline"><mml:mi>Z</mml:mi></mml:math></inline-formula>-scores trended upward with more oscillations,
with a shift to <inline-formula><mml:math id="M245" display="inline"><mml:mrow><mml:mi>Z</mml:mi><mml:mo>&gt;</mml:mo><mml:mn mathvariant="normal">2</mml:mn></mml:mrow></mml:math></inline-formula> becoming dominant during and following 27 March 2015.</p>
      <?pagebreak page2137?><p id="d1e3774">CTD <inline-formula><mml:math id="M246" display="inline"><mml:mi>Z</mml:mi></mml:math></inline-formula>-score time series are shown for LDA, LDB, and LDC in Figs. 9, 10, and
11, respectively. LDA <inline-formula><mml:math id="M247" display="inline"><mml:mi>Z</mml:mi></mml:math></inline-formula>-scores were generally <inline-formula><mml:math id="M248" display="inline"><mml:mrow><mml:mo>|</mml:mo><mml:mi>Z</mml:mi><mml:mo>|</mml:mo><mml:mo>&lt;</mml:mo><mml:mn mathvariant="normal">2</mml:mn></mml:mrow></mml:math></inline-formula>, but <inline-formula><mml:math id="M249" display="inline"><mml:mi>Z</mml:mi></mml:math></inline-formula>-scores for
densities <inline-formula><mml:math id="M250" display="inline"><mml:mrow><mml:msub><mml:mi mathvariant="italic">σ</mml:mi><mml:mi mathvariant="italic">θ</mml:mi></mml:msub><mml:mo>&lt;</mml:mo><mml:mn mathvariant="normal">1022</mml:mn></mml:mrow></mml:math></inline-formula> <inline-formula><mml:math id="M251" display="inline"><mml:mrow><mml:mi mathvariant="normal">kg</mml:mi><mml:mspace width="0.125em" linebreak="nobreak"/><mml:msup><mml:mi mathvariant="normal">m</mml:mi><mml:mrow><mml:mo>-</mml:mo><mml:mn mathvariant="normal">3</mml:mn></mml:mrow></mml:msup></mml:mrow></mml:math></inline-formula> were greater than 2
starting 1 March, and continued for the rest of the station. The increasing
trend in <inline-formula><mml:math id="M252" display="inline"><mml:mi>Z</mml:mi></mml:math></inline-formula>-score near the surface was also reflected in the SedTrap drifter.
LDB CTD <inline-formula><mml:math id="M253" display="inline"><mml:mi>Z</mml:mi></mml:math></inline-formula>-scores showed almost no observations with <inline-formula><mml:math id="M254" display="inline"><mml:mrow><mml:mo>|</mml:mo><mml:mi>Z</mml:mi><mml:mo>|</mml:mo><mml:mo>&gt;</mml:mo><mml:mn mathvariant="normal">2</mml:mn></mml:mrow></mml:math></inline-formula>. These occurred at
the surface with low densities and a few near <inline-formula><mml:math id="M255" display="inline"><mml:mrow><mml:msub><mml:mi mathvariant="italic">σ</mml:mi><mml:mi mathvariant="italic">θ</mml:mi></mml:msub><mml:mo>∼</mml:mo><mml:mn mathvariant="normal">1023.25</mml:mn></mml:mrow></mml:math></inline-formula> <inline-formula><mml:math id="M256" display="inline"><mml:mrow><mml:mi mathvariant="normal">kg</mml:mi><mml:mspace width="0.125em" linebreak="nobreak"/><mml:msup><mml:mi mathvariant="normal">m</mml:mi><mml:mrow><mml:mo>-</mml:mo><mml:mn mathvariant="normal">3</mml:mn></mml:mrow></mml:msup></mml:mrow></mml:math></inline-formula>.
All these observations occurred before or around 19 March and no temporal trend in <inline-formula><mml:math id="M257" display="inline"><mml:mrow><mml:mo>|</mml:mo><mml:mi>Z</mml:mi><mml:mo>|</mml:mo><mml:mo>&gt;</mml:mo><mml:mn mathvariant="normal">2</mml:mn></mml:mrow></mml:math></inline-formula> was seen. The <inline-formula><mml:math id="M258" display="inline"><mml:mi>Z</mml:mi></mml:math></inline-formula>-scores for LDC
showed similar trends to the SedTrap drifter. Near the surface close to
<inline-formula><mml:math id="M259" display="inline"><mml:mrow><mml:msub><mml:mi mathvariant="italic">σ</mml:mi><mml:mi mathvariant="italic">θ</mml:mi></mml:msub><mml:mo>∼</mml:mo><mml:mn mathvariant="normal">1022</mml:mn></mml:mrow></mml:math></inline-formula> <inline-formula><mml:math id="M260" display="inline"><mml:mrow><mml:mi mathvariant="normal">kg</mml:mi><mml:mspace linebreak="nobreak" width="0.125em"/><mml:msup><mml:mi mathvariant="normal">m</mml:mi><mml:mrow><mml:mo>-</mml:mo><mml:mn mathvariant="normal">3</mml:mn></mml:mrow></mml:msup></mml:mrow></mml:math></inline-formula>, <inline-formula><mml:math id="M261" display="inline"><mml:mrow><mml:mo>|</mml:mo><mml:mi>Z</mml:mi><mml:mo>|</mml:mo><mml:mo>&gt;</mml:mo><mml:mn mathvariant="normal">2</mml:mn></mml:mrow></mml:math></inline-formula> was seen early in
the time series, but then dropped to <inline-formula><mml:math id="M262" display="inline"><mml:mrow><mml:mo>|</mml:mo><mml:mi>Z</mml:mi><mml:mo>|</mml:mo><mml:mo>&lt;</mml:mo><mml:mn mathvariant="normal">2</mml:mn></mml:mrow></mml:math></inline-formula> until another increase around
27 March. This pattern was similar to the SedTrap drifter's observations at
14 <inline-formula><mml:math id="M263" display="inline"><mml:mi mathvariant="normal">m</mml:mi></mml:math></inline-formula>. Regions of <inline-formula><mml:math id="M264" display="inline"><mml:mrow><mml:mo>|</mml:mo><mml:mi>Z</mml:mi><mml:mo>|</mml:mo><mml:mo>&gt;</mml:mo><mml:mn mathvariant="normal">2</mml:mn></mml:mrow></mml:math></inline-formula> appeared in the 1024–1025 <inline-formula><mml:math id="M265" display="inline"><mml:mrow><mml:mi mathvariant="normal">kg</mml:mi><mml:mspace width="0.125em" linebreak="nobreak"/><mml:msup><mml:mi mathvariant="normal">m</mml:mi><mml:mrow><mml:mo>-</mml:mo><mml:mn mathvariant="normal">3</mml:mn></mml:mrow></mml:msup></mml:mrow></mml:math></inline-formula> range, primarily during 27 March. A small density range near
1024.5 <inline-formula><mml:math id="M266" display="inline"><mml:mrow><mml:mi mathvariant="normal">kg</mml:mi><mml:mspace linebreak="nobreak" width="0.125em"/><mml:msup><mml:mi mathvariant="normal">m</mml:mi><mml:mrow><mml:mo>-</mml:mo><mml:mn mathvariant="normal">3</mml:mn></mml:mrow></mml:msup></mml:mrow></mml:math></inline-formula> showed <inline-formula><mml:math id="M267" display="inline"><mml:mrow><mml:mo>|</mml:mo><mml:mi>Z</mml:mi><mml:mo>|</mml:mo><mml:mo>&gt;</mml:mo><mml:mn mathvariant="normal">2</mml:mn></mml:mrow></mml:math></inline-formula> during 26 March, but as time went
on a larger swath of density had <inline-formula><mml:math id="M268" display="inline"><mml:mrow><mml:mo>|</mml:mo><mml:mi>Z</mml:mi><mml:mo>|</mml:mo><mml:mo>&gt;</mml:mo><mml:mn mathvariant="normal">2</mml:mn></mml:mrow></mml:math></inline-formula> and this change was largely
permanent. Near 1025 <inline-formula><mml:math id="M269" display="inline"><mml:mrow><mml:mi mathvariant="normal">kg</mml:mi><mml:mspace linebreak="nobreak" width="0.125em"/><mml:msup><mml:mi mathvariant="normal">m</mml:mi><mml:mrow><mml:mo>-</mml:mo><mml:mn mathvariant="normal">3</mml:mn></mml:mrow></mml:msup></mml:mrow></mml:math></inline-formula>, a separate series of large
<inline-formula><mml:math id="M270" display="inline"><mml:mi>Z</mml:mi></mml:math></inline-formula>-scores appeared on 27 March and lasted for most of the rest of LDC.</p>

      <?xmltex \floatpos{t}?><fig id="Ch1.F12"><caption><p id="d1e4115">TSG <inline-formula><mml:math id="M271" display="inline"><mml:mi>Z</mml:mi></mml:math></inline-formula>-score over distance for LDA (left, green), LDB (center, red),
and LDC (right, blue). <inline-formula><mml:math id="M272" display="inline"><mml:mrow><mml:mo>|</mml:mo><mml:mi>Z</mml:mi><mml:mo>|</mml:mo><mml:mo>&gt;</mml:mo><mml:mn mathvariant="normal">2</mml:mn></mml:mrow></mml:math></inline-formula> is shaded black. Rossby radii <inline-formula><mml:math id="M273" display="inline"><mml:mrow><mml:msub><mml:mi>R</mml:mi><mml:mi>D</mml:mi></mml:msub></mml:mrow></mml:math></inline-formula> distance
plotted in horizontal dashed lines, color coded to the LD stations.</p></caption>
          <?xmltex \igopts{width=113.811024pt}?><graphic xlink:href="https://bg.copernicus.org/articles/15/2125/2018/bg-15-2125-2018-f12.pdf"/>

        </fig>

      <?xmltex \floatpos{t}?><fig id="Ch1.F13" specific-use="star"><caption><p id="d1e4160">SD station <inline-formula><mml:math id="M274" display="inline"><mml:mi>Z</mml:mi></mml:math></inline-formula>-score over distance for <bold>(a)</bold> LDA, <bold>(b)</bold> LDB, and <bold>(c)</bold> LDC.
<inline-formula><mml:math id="M275" display="inline"><mml:mrow><mml:mo>|</mml:mo><mml:mi>Z</mml:mi><mml:mo>|</mml:mo><mml:mo>&gt;</mml:mo><mml:mn mathvariant="normal">2</mml:mn></mml:mrow></mml:math></inline-formula> shaded black. Rossby radii <inline-formula><mml:math id="M276" display="inline"><mml:mrow><mml:msub><mml:mi>R</mml:mi><mml:mi>D</mml:mi></mml:msub></mml:mrow></mml:math></inline-formula> distance plotted in horizontal lines,
color coded to the LD stations.</p></caption>
          <?xmltex \igopts{width=426.791339pt}?><graphic xlink:href="https://bg.copernicus.org/articles/15/2125/2018/bg-15-2125-2018-f13.pdf"/>

        </fig>

</sec>
<sec id="Ch1.S3.SS3">
  <title>Spatial scale and baseline context</title>
      <?pagebreak page2138?><p id="d1e4219">The TSG <inline-formula><mml:math id="M277" display="inline"><mml:mi>Z</mml:mi></mml:math></inline-formula>-scores for the three LD stations are shown in Fig. 12. For LDA,
<inline-formula><mml:math id="M278" display="inline"><mml:mrow><mml:mi>Z</mml:mi><mml:mo>&gt;</mml:mo><mml:mn mathvariant="normal">2</mml:mn></mml:mrow></mml:math></inline-formula>
occurred at 150 <inline-formula><mml:math id="M279" display="inline"><mml:mi mathvariant="normal">km</mml:mi></mml:math></inline-formula>. <inline-formula><mml:math id="M280" display="inline"><mml:mi>Z</mml:mi></mml:math></inline-formula>-scores were consistently large farther away from
this point. The LDB TSG surpassed <inline-formula><mml:math id="M281" display="inline"><mml:mrow><mml:mi>Z</mml:mi><mml:mo>=</mml:mo><mml:mn mathvariant="normal">2</mml:mn></mml:mrow></mml:math></inline-formula> at 55 <inline-formula><mml:math id="M282" display="inline"><mml:mi mathvariant="normal">km</mml:mi></mml:math></inline-formula>, though <inline-formula><mml:math id="M283" display="inline"><mml:mi>Z</mml:mi></mml:math></inline-formula>-score
diminished again 300 <inline-formula><mml:math id="M284" display="inline"><mml:mi mathvariant="normal">km</mml:mi></mml:math></inline-formula> away. For LDC, TSG <inline-formula><mml:math id="M285" display="inline"><mml:mi>Z</mml:mi></mml:math></inline-formula>-score reached 2 at 35 <inline-formula><mml:math id="M286" display="inline"><mml:mi mathvariant="normal">km</mml:mi></mml:math></inline-formula>
distance, and <inline-formula><mml:math id="M287" display="inline"><mml:mrow><mml:mo>|</mml:mo><mml:mi>Z</mml:mi><mml:mo>|</mml:mo></mml:mrow></mml:math></inline-formula> oscillated between larger and less than 2 farther
away. Therefore, at the surface layer, 150, 55, and 35 <inline-formula><mml:math id="M288" display="inline"><mml:mi mathvariant="normal">km</mml:mi></mml:math></inline-formula> were the
spatial scales. Since at least some <inline-formula><mml:math id="M289" display="inline"><mml:mi>Z</mml:mi></mml:math></inline-formula>-scores were found to be greater than 2,
the baseline was sensitive enough to determine gradients over a 500 <inline-formula><mml:math id="M290" display="inline"><mml:mi mathvariant="normal">km</mml:mi></mml:math></inline-formula>
scale. Since <inline-formula><mml:math id="M291" display="inline"><mml:mi>Z</mml:mi></mml:math></inline-formula>-scores for LDB and LDC were not consistently <inline-formula><mml:math id="M292" display="inline"><mml:mrow><mml:mo>|</mml:mo><mml:mi>Z</mml:mi><mml:mo>|</mml:mo><mml:mo>&gt;</mml:mo><mml:mn mathvariant="normal">2</mml:mn></mml:mrow></mml:math></inline-formula>, then the
baseline's sensitivity was perhaps not as great as LDA. The Rossby radii for
the three stations were 46.5, 48.8, and 60 <inline-formula><mml:math id="M293" display="inline"><mml:mi mathvariant="normal">km</mml:mi></mml:math></inline-formula>. The spatial scales for
the TSG data at LDB and LDC matched up to the Rossby radii, whereas LDA's TSG
data indicated a larger scale.</p>
      <p id="d1e4367"><inline-formula><mml:math id="M294" display="inline"><mml:mi>Z</mml:mi></mml:math></inline-formula>-scores from the SD stations are presented in Fig. 13. For LDA, the
<inline-formula><mml:math id="M295" display="inline"><mml:mrow><mml:mo>|</mml:mo><mml:mi>Z</mml:mi><mml:mo>|</mml:mo><mml:mo>&gt;</mml:mo><mml:mn mathvariant="normal">2</mml:mn></mml:mrow></mml:math></inline-formula>
distances demonstrated density dependence. Near 1022 <inline-formula><mml:math id="M296" display="inline"><mml:mrow><mml:mi mathvariant="normal">kg</mml:mi><mml:mspace linebreak="nobreak" width="0.125em"/><mml:msup><mml:mi mathvariant="normal">m</mml:mi><mml:mrow><mml:mo>-</mml:mo><mml:mn mathvariant="normal">3</mml:mn></mml:mrow></mml:msup></mml:mrow></mml:math></inline-formula>,
<inline-formula><mml:math id="M297" display="inline"><mml:mrow><mml:mo>|</mml:mo><mml:mi>Z</mml:mi><mml:mo>|</mml:mo><mml:mo>&gt;</mml:mo><mml:mn mathvariant="normal">2</mml:mn></mml:mrow></mml:math></inline-formula> immediately, at <inline-formula><mml:math id="M298" display="inline"><mml:mo>∼</mml:mo></mml:math></inline-formula> 45 <inline-formula><mml:math id="M299" display="inline"><mml:mi mathvariant="normal">km</mml:mi></mml:math></inline-formula>, though this did not occur at the
surface. Approaching 1000 <inline-formula><mml:math id="M300" display="inline"><mml:mi mathvariant="normal">km</mml:mi></mml:math></inline-formula> distance, <inline-formula><mml:math id="M301" display="inline"><mml:mrow><mml:mo>|</mml:mo><mml:mi>Z</mml:mi><mml:mo>|</mml:mo><mml:mo>&gt;</mml:mo><mml:mn mathvariant="normal">2</mml:mn></mml:mrow></mml:math></inline-formula> occurred from the surface
to 1024 <inline-formula><mml:math id="M302" display="inline"><mml:mrow><mml:mi mathvariant="normal">kg</mml:mi><mml:mspace width="0.125em" linebreak="nobreak"/><mml:msup><mml:mi mathvariant="normal">m</mml:mi><mml:mrow><mml:mo>-</mml:mo><mml:mn mathvariant="normal">3</mml:mn></mml:mrow></mml:msup></mml:mrow></mml:math></inline-formula>. By 3500 <inline-formula><mml:math id="M303" display="inline"><mml:mi mathvariant="normal">km</mml:mi></mml:math></inline-formula>, all density layers show
<inline-formula><mml:math id="M304" display="inline"><mml:mrow><mml:mo>|</mml:mo><mml:mi>Z</mml:mi><mml:mo>|</mml:mo><mml:mo>&gt;</mml:mo><mml:mn mathvariant="normal">2</mml:mn></mml:mrow></mml:math></inline-formula>. LDB showed large <inline-formula><mml:math id="M305" display="inline"><mml:mi>Z</mml:mi></mml:math></inline-formula>-scores in some density layers at the closest SD
station located 189 <inline-formula><mml:math id="M306" display="inline"><mml:mi mathvariant="normal">km</mml:mi></mml:math></inline-formula> away. Past 750 <inline-formula><mml:math id="M307" display="inline"><mml:mi mathvariant="normal">km</mml:mi></mml:math></inline-formula>, <inline-formula><mml:math id="M308" display="inline"><mml:mi>Z</mml:mi></mml:math></inline-formula> from 1022–1024 <inline-formula><mml:math id="M309" display="inline"><mml:mrow><mml:mi mathvariant="normal">kg</mml:mi><mml:mspace width="0.125em" linebreak="nobreak"/><mml:msup><mml:mi mathvariant="normal">m</mml:mi><mml:mrow><mml:mo>-</mml:mo><mml:mn mathvariant="normal">3</mml:mn></mml:mrow></mml:msup></mml:mrow></mml:math></inline-formula> was consistently high. For densities greater than
1024 <inline-formula><mml:math id="M310" display="inline"><mml:mrow><mml:mi mathvariant="normal">kg</mml:mi><mml:mspace width="0.125em" linebreak="nobreak"/><mml:msup><mml:mi mathvariant="normal">m</mml:mi><mml:mrow><mml:mo>-</mml:mo><mml:mn mathvariant="normal">3</mml:mn></mml:mrow></mml:msup></mml:mrow></mml:math></inline-formula>, <inline-formula><mml:math id="M311" display="inline"><mml:mi>Z</mml:mi></mml:math></inline-formula>-scores were enhanced between 1000 and 1500 <inline-formula><mml:math id="M312" display="inline"><mml:mi mathvariant="normal">km</mml:mi></mml:math></inline-formula>,
but then decreased farther away. LDC's <inline-formula><mml:math id="M313" display="inline"><mml:mi>Z</mml:mi></mml:math></inline-formula>-scores show that <inline-formula><mml:math id="M314" display="inline"><mml:mrow><mml:mo>|</mml:mo><mml:mi>Z</mml:mi><mml:mo>|</mml:mo></mml:mrow></mml:math></inline-formula> was
greater than 2 from the first observations at 310 <inline-formula><mml:math id="M315" display="inline"><mml:mi mathvariant="normal">km</mml:mi></mml:math></inline-formula>. All density
layers showed enhanced <inline-formula><mml:math id="M316" display="inline"><mml:mi>Z</mml:mi></mml:math></inline-formula> values, with the majority of all observations being
larger<?pagebreak page2139?> than 2. Similar to the TSG data, the SD stations showed that the
baselines were sufficiently sensitive to detect physical gradients on large
scales, with some detecting changes immediately. Putting together the
near-surface TSG data and CTD data from the SD stations, LDA showed smaller
<inline-formula><mml:math id="M317" display="inline"><mml:mrow><mml:mo>|</mml:mo><mml:mi>Z</mml:mi><mml:mo>|</mml:mo><mml:mo>=</mml:mo><mml:mn mathvariant="normal">2</mml:mn></mml:mrow></mml:math></inline-formula> scales at depth, whereas LDB and LDC both showed variability both near
the surface and at depth at smaller scales. In order to be the most
conservative in our velocity and trajectory estimates, we will use the
smallest spatial scale of <inline-formula><mml:math id="M318" display="inline"><mml:mrow><mml:mo>|</mml:mo><mml:mi>Z</mml:mi><mml:mo>|</mml:mo><mml:mo>=</mml:mo><mml:mn mathvariant="normal">2</mml:mn></mml:mrow></mml:math></inline-formula> to determine the spatial scale <inline-formula><mml:math id="M319" display="inline"><mml:mrow><mml:msub><mml:mi>R</mml:mi><mml:mi>Z</mml:mi></mml:msub></mml:mrow></mml:math></inline-formula> and
evaluate Lagrangian risk, namely 45 <inline-formula><mml:math id="M320" display="inline"><mml:mi mathvariant="normal">km</mml:mi></mml:math></inline-formula> for LDA, 55 <inline-formula><mml:math id="M321" display="inline"><mml:mi mathvariant="normal">km</mml:mi></mml:math></inline-formula> for LDB,
and 35 <inline-formula><mml:math id="M322" display="inline"><mml:mi mathvariant="normal">km</mml:mi></mml:math></inline-formula> for LDC.</p>

      <?xmltex \floatpos{t}?><fig id="Ch1.F14" specific-use="star"><caption><p id="d1e4681">Time series of 38 <inline-formula><mml:math id="M323" display="inline"><mml:mi mathvariant="normal">kHz</mml:mi></mml:math></inline-formula> SADCP <inline-formula><mml:math id="M324" display="inline"><mml:mi>u</mml:mi></mml:math></inline-formula> and <inline-formula><mml:math id="M325" display="inline"><mml:mi>v</mml:mi></mml:math></inline-formula> components over depth for <bold>(a, d)</bold> LDA, <bold>(b, e)</bold> LDB, and <bold>(c, f)</bold> LDC. Comparison of 55 <inline-formula><mml:math id="M326" display="inline"><mml:mi mathvariant="normal">m</mml:mi></mml:math></inline-formula>
Aquadopp <inline-formula><mml:math id="M327" display="inline"><mml:mi>u</mml:mi></mml:math></inline-formula> and <inline-formula><mml:math id="M328" display="inline"><mml:mi>v</mml:mi></mml:math></inline-formula> components with 52 <inline-formula><mml:math id="M329" display="inline"><mml:mi mathvariant="normal">m</mml:mi></mml:math></inline-formula> SADCP <inline-formula><mml:math id="M330" display="inline"><mml:mi>u</mml:mi></mml:math></inline-formula> and <inline-formula><mml:math id="M331" display="inline"><mml:mi>v</mml:mi></mml:math></inline-formula> for
<bold>(g, j)</bold> LDA, <bold>(h, k)</bold> LDB, and <bold>(i, l)</bold> LDC. Aquadopp measurements shown in black,
SADCP measurements follow the color code. Units are in <inline-formula><mml:math id="M332" display="inline"><mml:mrow><mml:mi mathvariant="normal">m</mml:mi><mml:mspace width="0.125em" linebreak="nobreak"/><mml:msup><mml:mi mathvariant="normal">s</mml:mi><mml:mrow><mml:mo>-</mml:mo><mml:mn mathvariant="normal">1</mml:mn></mml:mrow></mml:msup></mml:mrow></mml:math></inline-formula>.
</p></caption>
          <?xmltex \igopts{width=426.791339pt}?><graphic xlink:href="https://bg.copernicus.org/articles/15/2125/2018/bg-15-2125-2018-f14.png"/>

        </fig>

      <?xmltex \floatpos{t}?><fig id="Ch1.F15" specific-use="star"><caption><p id="d1e4793">Observed and calculated trajectories for currents analysis. Data
from LDA, LDB, and LDC, shown in left, center, and right columns,
respectively. Top row <bold>(a–c)</bold>: observed trajectory of SedTrap drifter plotted
in magenta. Time-averaged altimetry-derived surface currents shown with black
arrows. Rossby radius <inline-formula><mml:math id="M333" display="inline"><mml:mrow><mml:msub><mml:mi>R</mml:mi><mml:mi>D</mml:mi></mml:msub></mml:mrow></mml:math></inline-formula> traced as a color-coded circle, and <inline-formula><mml:math id="M334" display="inline"><mml:mrow><mml:msub><mml:mi>R</mml:mi><mml:mi>Z</mml:mi></mml:msub></mml:mrow></mml:math></inline-formula>, the
calculated spatial scale, in black. Starting position of SedTrap drifter
shown with a star. Middle row <bold>(d–f)</bold>: calculated SADCP 38 <inline-formula><mml:math id="M335" display="inline"><mml:mi mathvariant="normal">kHz</mml:mi></mml:math></inline-formula> trajectories of
water at each depth down to 600 <inline-formula><mml:math id="M336" display="inline"><mml:mi mathvariant="normal">m</mml:mi></mml:math></inline-formula>. Bottom row <bold>(g–i)</bold>: observed SVP drifter
trajectories, with mean trajectory plotted in black.</p></caption>
          <?xmltex \igopts{width=426.791339pt}?><graphic xlink:href="https://bg.copernicus.org/articles/15/2125/2018/bg-15-2125-2018-f15.pdf"/>

        </fig>

      <p id="d1e4848">Having evaluated the ability of the statistical baselines to sense physical
gradients over large scales, we are now ready to analyze the currents and
trajectories.</p>
</sec>
<sec id="Ch1.S3.SS4">
  <title>Velocities and Lagrangian trajectories</title>
      <p id="d1e4857">Time series of the 38 <inline-formula><mml:math id="M337" display="inline"><mml:mi mathvariant="normal">kHz</mml:mi></mml:math></inline-formula> SADCP and AQUADOPP data are presented in Fig. 14.
The LDA time series of SADCP <inline-formula><mml:math id="M338" display="inline"><mml:mi>u</mml:mi></mml:math></inline-formula> and <inline-formula><mml:math id="M339" display="inline"><mml:mi>v</mml:mi></mml:math></inline-formula> components (Fig. 14a, d) showed strong
near-inertial oscillations in the upper 200 <inline-formula><mml:math id="M340" display="inline"><mml:mi mathvariant="normal">m</mml:mi></mml:math></inline-formula>, with velocities
reaching magnitudes of 0.6 <inline-formula><mml:math id="M341" display="inline"><mml:mrow><mml:mi mathvariant="normal">m</mml:mi><mml:mspace linebreak="nobreak" width="0.125em"/><mml:msup><mml:mi mathvariant="normal">s</mml:mi><mml:mrow><mml:mo>-</mml:mo><mml:mn mathvariant="normal">1</mml:mn></mml:mrow></mml:msup></mml:mrow></mml:math></inline-formula>. A weaker tidal component
was also present in this layer: below 200 <inline-formula><mml:math id="M342" display="inline"><mml:mi mathvariant="normal">m</mml:mi></mml:math></inline-formula>, vertical columns of
alternating velocity sign indicated the semi-diurnal tide. These tidal
signatures were also the dominant signal in the LDB and LDC time series (Fig. 14b–c, e–f).
The mixed layer, which, for most of the cruise, was <inline-formula><mml:math id="M343" display="inline"><mml:mo>≤</mml:mo></mml:math></inline-formula> 20 <inline-formula><mml:math id="M344" display="inline"><mml:mi mathvariant="normal">m</mml:mi></mml:math></inline-formula>,
was not resolved by either SADCP. So, the near-surface velocities
were only captured by the 11 <inline-formula><mml:math id="M345" display="inline"><mml:mi mathvariant="normal">m</mml:mi></mml:math></inline-formula> AQUADOPP and the SVP drifters drogued
at 15 <inline-formula><mml:math id="M346" display="inline"><mml:mi mathvariant="normal">m</mml:mi></mml:math></inline-formula>. Comparing the 55 <inline-formula><mml:math id="M347" display="inline"><mml:mi mathvariant="normal">m</mml:mi></mml:math></inline-formula> AQUADOPP time series with the 52 <inline-formula><mml:math id="M348" display="inline"><mml:mi mathvariant="normal">m</mml:mi></mml:math></inline-formula>
SADCP, the two data sources displayed similar trends for LDA. The
strong near-inertial oscillations led to correlations between the AQUADOPP
and 38 <inline-formula><mml:math id="M349" display="inline"><mml:mi mathvariant="normal">kHz</mml:mi></mml:math></inline-formula> time series of 0.75 and 0.76 for the <inline-formula><mml:math id="M350" display="inline"><mml:mi>u</mml:mi></mml:math></inline-formula> and <inline-formula><mml:math id="M351" display="inline"><mml:mi>v</mml:mi></mml:math></inline-formula> components,
respectively. During LDB and LDC, the weaker currents did not correlate as
well, leading to <inline-formula><mml:math id="M352" display="inline"><mml:mi>u</mml:mi></mml:math></inline-formula>, <inline-formula><mml:math id="M353" display="inline"><mml:mi>v</mml:mi></mml:math></inline-formula> correlations of <inline-formula><mml:math id="M354" display="inline"><mml:mo>-</mml:mo></mml:math></inline-formula>0.0137, <inline-formula><mml:math id="M355" display="inline"><mml:mo>-</mml:mo></mml:math></inline-formula>0.0554 (LDB), and 0.30, 0.37
(LDC), respectively. For comparison, the 150 <inline-formula><mml:math id="M356" display="inline"><mml:mi mathvariant="normal">kHz</mml:mi></mml:math></inline-formula> 52 <inline-formula><mml:math id="M357" display="inline"><mml:mi mathvariant="normal">m</mml:mi></mml:math></inline-formula>
time series produced <inline-formula><mml:math id="M358" display="inline"><mml:mi>u</mml:mi></mml:math></inline-formula>, <inline-formula><mml:math id="M359" display="inline"><mml:mi>v</mml:mi></mml:math></inline-formula> correlations with the AQUADOPP of 0.83 and 0.80
(LDA);
0.00 and 0.02 (LDB); and 0.68 and 0.68 (LDC). Vector correlations using the method
of <xref ref-type="bibr" rid="bib1.bibx8" id="text.36"/> for the three time series (not reported)
similarly showed a maximum for LDA, minimum near-zero for LDB, and low values
for LDC. These differences likely result from higher frequency fluctuations
of the currents, at the inertial and tidal frequencies. The fact that a
higher correlation is obtained at LDA is probably partly the consequence of
the larger horizontal scales of the near-inertial signal dominant at LDA
compared to the baroclinic tidal signal, e.g., resulting from the dispersion
relation <xref ref-type="bibr" rid="bib1.bibx1" id="paren.37"/>. These oscillations, and their implications
for turbulent mixing, are analyzed in greater detail in
<xref ref-type="bibr" rid="bib1.bibx4" id="text.38"/>.</p>
      <p id="d1e5044">The disagreement between the two velocity data sources had an impact on the
integrated trajectories. Take, for example, a closer inspection of the SADCP
and AQUADOPP during LDA, which had the strongest currents. The initial
positions of the ship and the SedTrap Drifter were 1.46 <inline-formula><mml:math id="M360" display="inline"><mml:mi mathvariant="normal">km</mml:mi></mml:math></inline-formula> apart.
After 3 days and 2 <inline-formula><mml:math id="M361" display="inline"><mml:mi mathvariant="normal">h</mml:mi></mml:math></inline-formula>, the AQUADOPP integration had traveled 67.75 <inline-formula><mml:math id="M362" display="inline"><mml:mi mathvariant="normal">km</mml:mi></mml:math></inline-formula>,
the SADCP 60.71 <inline-formula><mml:math id="M363" display="inline"><mml:mi mathvariant="normal">km</mml:mi></mml:math></inline-formula> with a final separation of 10.89 <inline-formula><mml:math id="M364" display="inline"><mml:mi mathvariant="normal">km</mml:mi></mml:math></inline-formula>.
The result was a positional drift of <inline-formula><mml:math id="M365" display="inline"><mml:mo>∼</mml:mo></mml:math></inline-formula> 3 <inline-formula><mml:math id="M366" display="inline"><mml:mrow><mml:mi mathvariant="normal">km</mml:mi><mml:mspace width="0.125em" linebreak="nobreak"/><mml:msup><mml:mi mathvariant="normal">day</mml:mi><mml:mrow><mml:mo>-</mml:mo><mml:mn mathvariant="normal">1</mml:mn></mml:mrow></mml:msup></mml:mrow></mml:math></inline-formula>, or an average increase in position difference of
147 <inline-formula><mml:math id="M367" display="inline"><mml:mi mathvariant="normal">m</mml:mi></mml:math></inline-formula> for each <inline-formula><mml:math id="M368" display="inline"><mml:mi mathvariant="normal">km</mml:mi></mml:math></inline-formula> traveled. A similar analysis for the LDB
time series, with weaker currents but essentially no correlations over 4 days
and 15 <inline-formula><mml:math id="M369" display="inline"><mml:mi mathvariant="normal">h</mml:mi></mml:math></inline-formula>, resulted in a positional drift of 3.19 <inline-formula><mml:math id="M370" display="inline"><mml:mrow><mml:mi mathvariant="normal">km</mml:mi><mml:mspace linebreak="nobreak" width="0.125em"/><mml:msup><mml:mi mathvariant="normal">day</mml:mi><mml:mrow><mml:mo>-</mml:mo><mml:mn mathvariant="normal">1</mml:mn></mml:mrow></mml:msup></mml:mrow></mml:math></inline-formula>, with an increased position difference of 318 <inline-formula><mml:math id="M371" display="inline"><mml:mi mathvariant="normal">m</mml:mi></mml:math></inline-formula> for
each <inline-formula><mml:math id="M372" display="inline"><mml:mi mathvariant="normal">km</mml:mi></mml:math></inline-formula> traveled. Thus, a lower correlation time series, but with lower
magnitudes, resulted in similar misfit in the integrated trajectory.</p>
      <p id="d1e5160">The trajectories of the integrated velocities, as well as observations of
SedTrap Drifter and SVP positions, are presented in Fig. 15. The average
altimetry-derived currents suggested there should be recirculation around the
positions of LDA and LDC, whereas LDB had a mean northward flow (Fig. 15a–c). The
SedTrap Drifter trajectory for LDA did not<?pagebreak page2140?> follow the surface
altimetry currents and their anticyclonic flow, but instead underwent several
oscillations while cruising in a west-northwest direction (Fig. 15a). The SVP
drifters for LDA (Fig. 15g), while also undergoing oscillatory loops, instead
drifted to the south. The 38 <inline-formula><mml:math id="M373" display="inline"><mml:mi mathvariant="normal">kHz</mml:mi></mml:math></inline-formula> SADCP velocities showed a transition
over depth with shallow water moving south-southwest, but with increasing
depth the trajectories flowed northwest in a similar fashion to the SedTrap
Drifter. During LDB, the SedTrap Drifter went north-northeast, in agreement
with the altimetry currents (Fig. 15b). The SVP drifters moved in a similar
fashion, north-northeast, though they ended up undergoing more oscillations
and eventually advected more eastward (Fig. 15h). The 38 <inline-formula><mml:math id="M374" display="inline"><mml:mi mathvariant="normal">kHZ</mml:mi></mml:math></inline-formula> SADCP
velocities demonstrated that shallow depths flowed east like the SVP
drifters, but with depth this advection swung to a more northerly direction
(Fig. 15e). The LDC SedTrap positioning was relatively uninformative, since
the Iridium satellite fix was unavailable for the second half of the LD
station and so showed little displacement (Fig. 15c). The SVP drifters for
LDC (Fig. 15i), similar to LDB, underwent several oscillations and were
advected the farthest, moving in a southeast direction. The SADCP data showed
a shallow flow to the east, similar to the SVPs, but with depth the majority
of trajectories oscillated near the station and even flowed southwest.</p>
      <p id="d1e5177">For all the LD stations, the SedTrap Drifter stayed within a radius of <inline-formula><mml:math id="M375" display="inline"><mml:mrow><mml:msub><mml:mi>R</mml:mi><mml:mi>Z</mml:mi></mml:msub></mml:mrow></mml:math></inline-formula>
and <inline-formula><mml:math id="M376" display="inline"><mml:mrow><mml:msub><mml:mi>R</mml:mi><mml:mi>D</mml:mi></mml:msub></mml:mrow></mml:math></inline-formula> centered at the LD starting position. Integrated velocities of the
SADCP also stay within <inline-formula><mml:math id="M377" display="inline"><mml:mrow><mml:msub><mml:mi>R</mml:mi><mml:mi>Z</mml:mi></mml:msub></mml:mrow></mml:math></inline-formula> and <inline-formula><mml:math id="M378" display="inline"><mml:mrow><mml:msub><mml:mi>R</mml:mi><mml:mi>D</mml:mi></mml:msub></mml:mrow></mml:math></inline-formula>, except for the trajectory nearest the
surface for LDA. The SVP drifters for LDA and LDB also stayed within the
<inline-formula><mml:math id="M379" display="inline"><mml:mrow><mml:msub><mml:mi>R</mml:mi><mml:mi>Z</mml:mi></mml:msub></mml:mrow></mml:math></inline-formula> and <inline-formula><mml:math id="M380" display="inline"><mml:mrow><mml:msub><mml:mi>R</mml:mi><mml:mi>D</mml:mi></mml:msub></mml:mrow></mml:math></inline-formula> distances, though the LDB SVP trajectories come close to
<inline-formula><mml:math id="M381" display="inline"><mml:mrow><mml:msub><mml:mi>R</mml:mi><mml:mi>Z</mml:mi></mml:msub></mml:mrow></mml:math></inline-formula>. For LDC, however, the SVP drifters traveled farther than <inline-formula><mml:math id="M382" display="inline"><mml:mrow><mml:msub><mml:mi>R</mml:mi><mml:mi>Z</mml:mi></mml:msub></mml:mrow></mml:math></inline-formula>, but
shorter than <inline-formula><mml:math id="M383" display="inline"><mml:mrow><mml:msub><mml:mi>R</mml:mi><mml:mi>D</mml:mi></mml:msub></mml:mrow></mml:math></inline-formula> away from the initial position. <inline-formula><mml:math id="M384" display="inline"><mml:mrow><mml:msub><mml:mi>R</mml:mi><mml:mi>D</mml:mi></mml:msub></mml:mrow></mml:math></inline-formula> was larger than
<inline-formula><mml:math id="M385" display="inline"><mml:mrow><mml:msub><mml:mi>R</mml:mi><mml:mi>Z</mml:mi></mml:msub></mml:mrow></mml:math></inline-formula> for all three stations. Since surface SADCP data for LDA crossed <inline-formula><mml:math id="M386" display="inline"><mml:mrow><mml:msub><mml:mi>R</mml:mi><mml:mi>Z</mml:mi></mml:msub></mml:mrow></mml:math></inline-formula>,
we evaluate that the water during LDA might be from a different water mass.
Likewise, the SVP trajectories for LDC crossed <inline-formula><mml:math id="M387" display="inline"><mml:mrow><mml:msub><mml:mi>R</mml:mi><mml:mi>Z</mml:mi></mml:msub></mml:mrow></mml:math></inline-formula>, meaning that surface
water for LDC might be from a<?pagebreak page2141?> different water mass at the end of the station.
As seen in the time series analysis, water at depth for LDC also changed
part way through the station, though the SADCP velocities show little
displacement. As a result, according to our protocol the water observed in
these density layers may have derived from a new physical environment, and
biogeochemical observations at the end of LDA (primarily the surface) and LDC
(surface and at depth) may need to be examined in closer detail for changes.</p>
      <p id="d1e5326">We conclude by noting that the trajectories highlight how water seen during
the LD stations derived from farther away than the SedTrap drifter's
position, as well as from differing directions. In effect, the drifter is
only quasi-Lagrangian. Beyond the layers shown to have different water masses
due to the time series analysis, the conservative length scale <inline-formula><mml:math id="M388" display="inline"><mml:mrow><mml:msub><mml:mi>R</mml:mi><mml:mi>Z</mml:mi></mml:msub></mml:mrow></mml:math></inline-formula> helps
determine the maximum spatial footprint where sheared layers can be said to
be from the same water mass. Therefore the results suggest that LDB's layers
were all from a single physical environment, surface observations for LDA and
LDC were suspect towards the end, and LDC water in the 1024–1025 <inline-formula><mml:math id="M389" display="inline"><mml:mrow><mml:mi mathvariant="normal">kg</mml:mi><mml:mspace width="0.125em" linebreak="nobreak"/><mml:msup><mml:mi mathvariant="normal">m</mml:mi><mml:mrow><mml:mo>-</mml:mo><mml:mn mathvariant="normal">3</mml:mn></mml:mrow></mml:msup></mml:mrow></mml:math></inline-formula>
range should be closely examined after 27 March 2015.</p>
</sec>
</sec>
<sec id="Ch1.S4">
  <title>Discussion</title>
<sec id="Ch1.S4.SS1">
  <title>Tracer analysis and spatial scale determination</title>
      <?pagebreak page2142?><p id="d1e5369">The main goal of this study is to determine whether the quasi-Lagrangian
sampling strategy during the LD stations of OUTPACE was successful. Since the
SedTrap drifter cannot be truly Lagrangian, successful sampling is judged by
whether observations stem from a single physical environment. The motivation
behind this exercise is to independently evaluate if the biogeochemical
measurements of OUTPACE represented a single biogeochemical milieu, rather
than the advection of the SedTrap drifter into a different area, as well as
the advection of different water toward the drifter. Evaluation of the
strategy is grounded in the variability of <inline-formula><mml:math id="M390" display="inline"><mml:mi>T</mml:mi></mml:math></inline-formula>-<inline-formula><mml:math id="M391" display="inline"><mml:mi>S</mml:mi></mml:math></inline-formula> and analysis of water
velocities.</p>
      <p id="d1e5386">Before the spice analysis was conducted, the initial context of SST and
chl <italic>a</italic> variability at the surface in space and time was provided by
satellite products. At the regional scale of the WTSP, the LD stations were
roughly positioned within the zonal gradient of chl <italic>a</italic> and meridional
gradient in SST (Fig. 1). The gradients of surface chl <italic>a</italic> around the
LD stations were minimized in relation to the regional-scale gradients,
partly by design in the process of choosing station locations (Fig. 2). The
temporal trends of SST and chl <italic>a</italic> largely reflected the seasonal
cycle: chl <italic>a</italic> was decreasing at the end of the summer in the MA and
low values dominated in the SPG; SST reached its peak due to late summer
timing (Fig. 3). The timing of temperature maxima is important to note, as
mentioned in Sect. 3.1, since <inline-formula><mml:math id="M392" display="inline"><mml:mrow><mml:msub><mml:mi>N</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula> fixation by <italic>Trichodesmium</italic> spp. is known to occur in warm waters, and one of the goals of OUTPACE was to
observe this biogeochemical process. While these satellite data are
sufficient to identify large-scale structures or temporal trends, the LD
stations by design were in regions where it is difficult to judge whether the
SedTrap Drifter stayed in the same water mass from these surface data alone.
However, they do justify the use of our methodology, ie there were no strong
mesoscale structures in the vicinity, and no nearby chl <italic>a</italic> gradients
were present. As mentioned previously, the use of remote sensing data to help
identify small-scale surface structures during OUTPACE is further explored in
<xref ref-type="bibr" rid="bib1.bibx39" id="text.39"/>. In that study, a possible coherent structure
was found during LDC, though this derives from satellite-derived surface
velocity data; the SedTrap drifter's Lagrangian limitations preclude testing
whether the structure was truly coherent. To continue with the
post-validation strategy evaluating LD sampling, in situ data are now needed.</p>
      <p id="d1e5425">The depth-resolved in situ <inline-formula><mml:math id="M393" display="inline"><mml:mi>T</mml:mi></mml:math></inline-formula>-<inline-formula><mml:math id="M394" display="inline"><mml:mi>S</mml:mi></mml:math></inline-formula> data (Fig. 4) helped to capture some of the
variability present during the LD and surrounding SD stations. The <inline-formula><mml:math id="M395" display="inline"><mml:mi>T</mml:mi></mml:math></inline-formula>-<inline-formula><mml:math id="M396" display="inline"><mml:mi>S</mml:mi></mml:math></inline-formula>
structures showed consistent values during LD stations with deviations
observed in the neighboring SD stations. As with the satellite data, the in
situ data in this form, although informative, provide qualitative
interpretation. In order to identify water masses, traditional quantitative
methods require identified water masses
<xref ref-type="bibr" rid="bib1.bibx27 bib1.bibx37" id="paren.40"/>. In some regions of the ocean, these
methods are difficult to apply. In addition, the full complement of tracer
data (dissolved <inline-formula><mml:math id="M397" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">O</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula>, nutrients, etc.) potentially cannot be used (like in this study)
because they are liable to rapidly change in the euphotic zone due to
biogeochemical processes. Additionally, local mesoscale activity can
contribute to variability, as has been seen in the WTSP
<xref ref-type="bibr" rid="bib1.bibx38" id="paren.41"/>. In that study, <inline-formula><mml:math id="M398" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">O</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula> measurements were the
distinguishing tracer, which we are precluded from using. As a result of all
these factors, another quantitative method that works within the dataset of a
single cruise in the WTSP is needed.</p>
      <p id="d1e5485">The quantitative approach used in this study leverages the large quantity of
in situ <inline-formula><mml:math id="M399" display="inline"><mml:mi>T</mml:mi></mml:math></inline-formula>-<inline-formula><mml:math id="M400" display="inline"><mml:mi>S</mml:mi></mml:math></inline-formula> data available from multiple platforms to condense the physical
variability present in the WTSP during OUTPACE over a 4000 <inline-formula><mml:math id="M401" display="inline"><mml:mi mathvariant="normal">km</mml:mi></mml:math></inline-formula> distance
during austral summer 2015. In order to do this, the statistical baseline in
spice was defined (Fig. 5). In effect, as opposed to an absolute measure
(i.e., specific water mass determination), this provides a relative measure of
variability that can be used. The time series analysis for the latter portion
of the SedTrap drifter and CTD time series data showed density layers where
variability was enhanced, and caution should be applied to the analysis of
biogeochemical data. Since the baseline is a relative measure, observations
from outside the LD station were used to see at what scales physical
gradients appear. The relationship between distance and variability
(summarized as <inline-formula><mml:math id="M402" display="inline"><mml:mi>Z</mml:mi></mml:math></inline-formula>-scores) provided a method by which to establish this scale.
Overall, variability increased with distance, as one would logically expect,
but this was not monotonic across all datasets. Therefore, the first increase
in <inline-formula><mml:math id="M403" display="inline"><mml:mi>Z</mml:mi></mml:math></inline-formula>-score above 2 (using an <inline-formula><mml:math id="M404" display="inline"><mml:mrow><mml:mi mathvariant="italic">α</mml:mi><mml:mo>=</mml:mo><mml:mn mathvariant="normal">0.5</mml:mn></mml:mrow></mml:math></inline-formula> criterion) was used, and the
smallest of these scales across datasets and density layers was
conservatively chosen in determining the cut-off scale <inline-formula><mml:math id="M405" display="inline"><mml:mrow><mml:msub><mml:mi>R</mml:mi><mml:mi>Z</mml:mi></mml:msub></mml:mrow></mml:math></inline-formula>. These distances
were of the same order of magnitude as the Rossby radius <inline-formula><mml:math id="M406" display="inline"><mml:mrow><mml:msub><mml:mi>R</mml:mi><mml:mi>D</mml:mi></mml:msub></mml:mrow></mml:math></inline-formula>.</p>
      <p id="d1e5559">Despite the testing of the baseline on complementary data outside of LD
sampling, there is still the question of its generality. In some instances,
the variability can be partly attributed to an in situ process, such as the
near-surface changes in LDA coinciding with surface heating (though advection
possibly played a role, as seen in the SADCP trajectories). While seeing
increased variability with distance was reassuring, the non-monotonic nature
of some <inline-formula><mml:math id="M407" display="inline"><mml:mi>Z</mml:mi></mml:math></inline-formula> vs. distance relationship raises some questions. Is the water on
the other side, where <inline-formula><mml:math id="M408" display="inline"><mml:mrow><mml:mo>|</mml:mo><mml:mi>Z</mml:mi><mml:mo>|</mml:mo></mml:mrow></mml:math></inline-formula> goes back to below 2, truly the same water mass? Is
<inline-formula><mml:math id="M409" display="inline"><mml:mrow><mml:mo>|</mml:mo><mml:mi>Z</mml:mi><mml:mo>|</mml:mo><mml:mo>&gt;</mml:mo><mml:mn mathvariant="normal">2</mml:mn></mml:mrow></mml:math></inline-formula> truly a change in water mass relative to traditional methods? Is the
<inline-formula><mml:math id="M410" display="inline"><mml:mi>Z</mml:mi></mml:math></inline-formula>-score approach based on the spice–density relationship more, or less,
sensitive than these methods? These questions merit further study, and will
have to be explored using both methods simultaneously.</p>
      <p id="d1e5604">The example of surface heating at LDA introducing unwanted variation brings
up another assumption in our analysis: we used spice hypothesizing that there
was no diapycnal forcing. Clearly, at the surface, atmospheric forcing can
influence the water's <inline-formula><mml:math id="M411" display="inline"><mml:mi>T</mml:mi></mml:math></inline-formula>-<inline-formula><mml:math id="M412" display="inline"><mml:mi>S</mml:mi></mml:math></inline-formula> (and spice) characteristics, and so will impact TSG
measurements as well as observations in the upper mixed layer. Future
applications of this method will have to take this variability into account,
and perhaps make greater use of survey data to fill in the spice variability
below the surface at these spatial scales. However, at depth the greatest
source of along-isopycnal gradients in <inline-formula><mml:math id="M413" display="inline"><mml:mi>T</mml:mi></mml:math></inline-formula>-<inline-formula><mml:math id="M414" display="inline"><mml:mi>S</mml:mi></mml:math></inline-formula>, i.e., density-compensated
features, is mesoscale stirring <xref ref-type="bibr" rid="bib1.bibx42" id="paren.42"/>, and so,
generally, the assumption should be applicable. Part of our assumptions in
developing this method requires that sampling is not near mesoscale fronts
and eddies. However, their residual effects are the most<?pagebreak page2143?> likely to affect
along-isopycnal variability in a given field campaign, and it is probably the
reason that both <inline-formula><mml:math id="M415" display="inline"><mml:mi>Z</mml:mi></mml:math></inline-formula>-scores (especially near the surface) begin to increase and
the <inline-formula><mml:math id="M416" display="inline"><mml:mrow><mml:msub><mml:mi>R</mml:mi><mml:mi>Z</mml:mi></mml:msub></mml:mrow></mml:math></inline-formula>'s were found to be at or near the Rossby radius <inline-formula><mml:math id="M417" display="inline"><mml:mrow><mml:msub><mml:mi>R</mml:mi><mml:mi>D</mml:mi></mml:msub></mml:mrow></mml:math></inline-formula> for each LD
station. Granted, if ship sampling happens to be placed immediately next to a
strong eddy or filament, <inline-formula><mml:math id="M418" display="inline"><mml:mi>Z</mml:mi></mml:math></inline-formula>-scores could increase over much shorter distances,
though as previously stated this situation was expressly avoided. Therefore,
in situations where our starting assumptions are met, at first order the
Rossby radius <inline-formula><mml:math id="M419" display="inline"><mml:mrow><mml:msub><mml:mi>R</mml:mi><mml:mi>D</mml:mi></mml:msub></mml:mrow></mml:math></inline-formula> serves as a default scale at which the integrated history
of previous mesoscale stirring will on average manifest itself. In situ data
and further analysis is needed, however, to verify whether smaller-scale
variability is large through determination of <inline-formula><mml:math id="M420" display="inline"><mml:mrow><mml:msub><mml:mi>R</mml:mi><mml:mi>Z</mml:mi></mml:msub></mml:mrow></mml:math></inline-formula>, as evidenced by the
appearance of salty, spicy water in the mid-water column at the end of LDC.
One recommendation that emerges from these results is that, if future field
measurements are to take place in an area devoid of obvious mesoscale
structures, the Rossby radius <inline-formula><mml:math id="M421" display="inline"><mml:mrow><mml:msub><mml:mi>R</mml:mi><mml:mi>D</mml:mi></mml:msub></mml:mrow></mml:math></inline-formula> quickly calculated from a deep CTD cast
may be useful in determining at which point a quasi-Lagrangian drift array
should be recovered as a precautionary measure.</p>
      <p id="d1e5709">The choice of spice as a variable, though useful, is not a magical
transformation in itself. The similarity in <inline-formula><mml:math id="M422" display="inline"><mml:mi>T</mml:mi></mml:math></inline-formula>-<inline-formula><mml:math id="M423" display="inline"><mml:mi>S</mml:mi></mml:math></inline-formula> between LD stations is still
manifest in spice–density space (Fig. 5), especially at a depth where some
density layers overlap. This means that for these density ranges, the
stations are not distinguishable from one another. The overlap in statistical
baselines further emphasizes the need to compare with other datasets to
highlight and determine the scales at which variability exists. Having enough
data to span a sufficient spatial range is what will determine what
differences in spice are relevant or not. The fact that the SedTrap drifter,
positioned close to the CTD time series, mirrored the same trends,
corroborated both the presence or lack of small-scale gradients shown through
<inline-formula><mml:math id="M424" display="inline"><mml:mi>Z</mml:mi></mml:math></inline-formula>-score. If the SedTrap Drifter time series, representing the smallest scales,
had displayed much larger <inline-formula><mml:math id="M425" display="inline"><mml:mi>Z</mml:mi></mml:math></inline-formula>-scores, then the conclusion would have been that
there was large variability right next to the ship somehow missed by the CTD
baseline. Instead, by showing similar trends, the SedTrap data validated the
baseline and the spatial scales at which the LD CTD time series was sampled,
even if statistical baselines between the far-flung LD stations overlap.</p>
      <p id="d1e5740">Having considered some of the caveats and assumptions implicit in our present
approach, we feel that its application for the OUTPACE campaign was warranted
and subsequently validated by the consistency of the results, both in
consideration of the multiple data sources concerned but also the theory of
mesoscale circulation. In our application, a conservative <inline-formula><mml:math id="M426" display="inline"><mml:mrow><mml:msub><mml:mi>R</mml:mi><mml:mi>Z</mml:mi></mml:msub></mml:mrow></mml:math></inline-formula> was used
based on surface data at intermediate scales. Ideally, future applications of
this method could hopefully use depth-resolved measurements to see when
surface layers diverge as opposed to deeper ones, which might be reflective
of different turbulence regimes such as the submesoscale.</p>
</sec>
<sec id="Ch1.S4.SS2">
  <title>Integrated velocities and drifter trajectories</title>
      <p id="d1e5760">The analysis of the integrated in situ velocities, alongside the SedTrap
Drifter and SVP positions, was used to determine whether certain parcels of
water advected far enough during each LD station's sampling to move beyond
<inline-formula><mml:math id="M427" display="inline"><mml:mrow><mml:msub><mml:mi>R</mml:mi><mml:mi>Z</mml:mi></mml:msub></mml:mrow></mml:math></inline-formula>, the conservative spatial scale. The conclusion for the OUTPACE cruise
is that almost all of these proxies for Lagrangian pathways did not do so, and
hence the changes observed in the upper water column during the LD stations
were probably due to biogeochemical processes alone. Nevertheless, given the
multiple sources of trajectories available, there is no clear interpretation
as to which of them is the most “truly Lagrangian.”</p>
      <p id="d1e5774">As stated in the introduction, the quasi-Lagrangian drift array, spanning the
top 100s of meters of the water column, with various instruments having
varying degrees of drag strewn at different depths, will never be purely
Lagrangian. Unless a drift array is deployed in a flow that is entirely
barotropic, there will likely be sources of shear in the velocity field from
the mesoscale flow and from baroclinic near-inertial waves and internal tides
(see <xref ref-type="bibr" rid="bib1.bibx4" id="author.43"/>, <xref ref-type="bibr" rid="bib1.bibx4" id="year.44"/>). By
definition, vertical shear leads to different velocities at different depths.
The net effect of these velocities results in the drift array's observed
trajectory, with water at some depths moving slower or faster than the array
itself. As a result of this inevitable decoupling of in situ and drifter
velocities, the water passing the drifter can come from entirely different
areas. Take, for example, the near-inertial oscillations during LDA (Fig. 7a, d).
The oscillations influenced the top 200 <inline-formula><mml:math id="M428" display="inline"><mml:mi mathvariant="normal">m</mml:mi></mml:math></inline-formula>, so immediately the
top sediment trap at 150 <inline-formula><mml:math id="M429" display="inline"><mml:mi mathvariant="normal">m</mml:mi></mml:math></inline-formula> experienced vigorous currents whereas the
bottom two at 250 and 500 <inline-formula><mml:math id="M430" display="inline"><mml:mi mathvariant="normal">m</mml:mi></mml:math></inline-formula> did not. This complicates the
interpretation of where the falling organic matter comes from. Strong
vertical structuring of the phytoplankton community further complicates this
picture. In oligotrophic areas, light-adapted organisms are found near the
surface with darkness-adapted organisms near the deep chl <italic>a</italic> maximum (not
to mention the near-surface <inline-formula><mml:math id="M431" display="inline"><mml:mrow class="chem"><mml:msub><mml:mi mathvariant="normal">N</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msub></mml:mrow></mml:math></inline-formula> fixation occurring during OUTPACE). During
LDA, the shear resulted in bands of instantaneously opposing velocities in
layers about 50 <inline-formula><mml:math id="M432" display="inline"><mml:mi mathvariant="normal">m</mml:mi></mml:math></inline-formula> thick. Therefore, organisms sampled by a CTD near
the chl <italic>a</italic> max at 80 <inline-formula><mml:math id="M433" display="inline"><mml:mi mathvariant="normal">m</mml:mi></mml:math></inline-formula> would be traveling in a direction
opposite to the organisms at the surface. Immediately, it seems, the
expectation of being Lagrangian is lost. However, whether this is an
irrevocable loss depends upon the nature of the motions: are they associated
with internal waves, such as LDA, or not? If shearing is wave-induced, then
after one period, water should return to its original position. Thus,
sampling over at least one wave period and within corresponding horizontal
wavelengths could help to preserve the physical environment. If motions are
not related to oscillatory processes, then the only recourse would be that
biological communities (and the state of biogeochemical processes underway)
are contiguous enough in horizontal extent that they reflect<?pagebreak page2144?> the physical
spatial scales found in the tracer analysis of this study.</p>
      <p id="d1e5836">The potential ability of individual shearing layers to remain coherent over
long timescales during OUTPACE is reflected by the SVP drifters during LDC.
While they had traveled far, surpassing <inline-formula><mml:math id="M434" display="inline"><mml:mrow><mml:msub><mml:mi>R</mml:mi><mml:mi>Z</mml:mi></mml:msub></mml:mrow></mml:math></inline-formula> and near <inline-formula><mml:math id="M435" display="inline"><mml:mrow><mml:msub><mml:mi>R</mml:mi><mml:mi>D</mml:mi></mml:msub></mml:mrow></mml:math></inline-formula> by the end of
sampling, they remained close to each other. This is also reflected in the
SVP relative dispersion time series (Fig. S2). Typically, relative dispersion
increases in the first few days in an exponential fashion, until the drifters
begin to be influenced by mesoscale structures <xref ref-type="bibr" rid="bib1.bibx21" id="paren.45"/>.
For the OUTPACE SVP drifters, the exponential phase (a linear increase in
log-space) did not abate in this timeframe. However, the exponential growth
rate was so low that it took over a week for the relative dispersion to
arrive near <inline-formula><mml:math id="M436" display="inline"><mml:mrow><mml:mn mathvariant="normal">1</mml:mn><mml:mo>×</mml:mo><mml:msup><mml:mn mathvariant="normal">10</mml:mn><mml:mn mathvariant="normal">8</mml:mn></mml:msup></mml:mrow></mml:math></inline-formula> <inline-formula><mml:math id="M437" display="inline"><mml:mrow><mml:msup><mml:mi mathvariant="normal">m</mml:mi><mml:mn mathvariant="normal">2</mml:mn></mml:msup></mml:mrow></mml:math></inline-formula>. Thus, while the SVP drifters during
LDC advected past <inline-formula><mml:math id="M438" display="inline"><mml:mrow><mml:msub><mml:mi>R</mml:mi><mml:mi>Z</mml:mi></mml:msub></mml:mrow></mml:math></inline-formula> and near <inline-formula><mml:math id="M439" display="inline"><mml:mrow><mml:msub><mml:mi>R</mml:mi><mml:mi>D</mml:mi></mml:msub></mml:mrow></mml:math></inline-formula>, their spread was less than this,
indicating that the surface layer in LDC was more coherent than comparing
between integrated trajectories over different depths would suggest.</p>
      <p id="d1e5913">In addition to the sheared layers advecting past the drift array, what
determines the trajectory of the array itself is difficult to establish.
During LDA, with vigorous currents in the upper 200 <inline-formula><mml:math id="M440" display="inline"><mml:mi mathvariant="normal">m</mml:mi></mml:math></inline-formula>, the SedTrap
Drifter advected north-west, mirroring the deep SADCP trajectories. By
contrast, during LDB, with weaker currents, the SedTrap Drifter moved in a
similar fashion to the more shallow SADCP and SVP paths. The LDC telemetry,
having been cut short, is not useful here. Perhaps hidden in the data is some
linear combination of SADCP and SVP trajectories that can best explain why
the SedTrap Drifter moved as it did. Although perhaps useful for OUTPACE,
this would be fitting a model to a particular drifter configuration, and
remove future flexibility in drifter design. Hence, we chose not to pursue
this calculation.</p>
      <p id="d1e5924">The different trajectories taken between the AQUADOPP, SADCP, and SVP
demonstrate how trying to follow a Lagrangian perspective can get quickly
complicated within a few days, if not sooner <xref ref-type="bibr" rid="bib1.bibx35" id="paren.46"/>. Even
measurements at similar depths between different instruments do not
correspond, such as between the AQUADOPP and SADCP at 55 and 52 <inline-formula><mml:math id="M441" display="inline"><mml:mi mathvariant="normal">m</mml:mi></mml:math></inline-formula>,
respectively, for LDB. Where they do correspond, such as for LDA, this may be
due to a strong signal, which is also problematic due to the shear involved.
As mentioned in Sect. 3.3, the drift between AQUADOPP and SADCP during LDA
and LDB were comparable, despite the differing correlations, due to the fact
that currents were stronger at LDA. Additionally, while both AQUADOPP and
SADCP integrations started from a chosen initial position, the time series
data reflect measurements taken on moving platforms, i.e., the SedTrap Drifter
and the ship, respectively. All these complicating factors considered, the
conservative route, employed here, is to consider each source of data and
give it the best chance to refute the assumption that advection was weak
enough that the spatial boundary determined by tracer analysis was not
crossed. This procedure is the recommendation the authors provide subsequent
to having done this analysis: while no individual trajectory will likely be
the right one, exploring enough of the variability between them should give
an idea of whether other water masses are present.</p>
</sec>
<sec id="Ch1.S4.SS3">
  <title>Biogeochemical sampling limitations</title>
      <p id="d1e5943">The protocol established in this study has identified the conditions in which
a quasi-Lagrangian drift array can be considered successful in sampling one
water mass, and ostensibly a single biogeochemical environment. The results
from the three deployments during OUTPACE provide an emerging picture of what
is exactly observed during the 3–5 <inline-formula><mml:math id="M442" display="inline"><mml:mi mathvariant="normal">days</mml:mi></mml:math></inline-formula> of sampling. The trajectories
shown in Fig. 15 demonstrate that rather than a single, one-dimensional
profile, vertical shear means that observations are of water from an extended
area, reflecting a spatial sampling footprint. Sometimes water is
sufficiently different that changes can be seen during the in situ
time series, and these layers can be isolated for closer inspection. For the
rest of the layers, the present method produces the <inline-formula><mml:math id="M443" display="inline"><mml:mrow><mml:msub><mml:mi>R</mml:mi><mml:mi>Z</mml:mi></mml:msub></mml:mrow></mml:math></inline-formula> limit for the
sampling footprint.</p>
      <p id="d1e5964">The existence of the sampling footprint requires biogeochemical investigators
to re-consider a few assumptions. As stated in the introduction, this
analysis uses only physical, and not biological, data. Therefore, one must
ask whether horizontal gradients exist at smaller scales. The only way to
characterize this is to make measurements through a gradient, though this can
be difficult with the need to produce measurements over a large area within a
short, quasi-synoptic timescale. Since there exist at least some trajectories
for each LD station that approached <inline-formula><mml:math id="M444" display="inline"><mml:mrow><mml:msub><mml:mi>R</mml:mi><mml:mi>Z</mml:mi></mml:msub></mml:mrow></mml:math></inline-formula>, the risk of encountering gradients
is always present. Thus, for each biogeochemical measurement, the assumption
of whether horizontal <inline-formula><mml:math id="M445" display="inline"><mml:mi>T</mml:mi></mml:math></inline-formula>-<inline-formula><mml:math id="M446" display="inline"><mml:mi>S</mml:mi></mml:math></inline-formula> gradients reflect biogeochemical gradients must be
validated. For OUTPACE, use of a moving vessel profiler (MVP) to survey an
extended area around each LD station was planned, but technical difficulties
prevented their use in the application of this method, though the data were
useful in another study <xref ref-type="bibr" rid="bib1.bibx11" id="paren.47"/>.</p>
      <?pagebreak page2145?><p id="d1e5995">Furthermore, the current approach focuses on individual density layers.
Measurements that reflect time- and depth-integrated processes, such as
particle settlement in the sediment traps, require more assumptions before
they can be considered to represent a single environment. For sediment traps,
depending on sinking particle velocities, the 3–5 <inline-formula><mml:math id="M447" display="inline"><mml:mi mathvariant="normal">day</mml:mi></mml:math></inline-formula> timeframe may
not be long enough to represent the entire water column, i.e., the sinking of
surface particles to the sediment trap depth. Since near-surface trajectories
for all three stations were able to get near <inline-formula><mml:math id="M448" display="inline"><mml:mrow><mml:msub><mml:mi>R</mml:mi><mml:mi>Z</mml:mi></mml:msub></mml:mrow></mml:math></inline-formula> in 5 <inline-formula><mml:math id="M449" display="inline"><mml:mi mathvariant="normal">days</mml:mi></mml:math></inline-formula>, any
process with longer timescales would probably produce measurements
contaminated by a water mass change. Additionally, if water at different
depths arrives from entirely different directions, gradients in any of these
layers will influence the result. For example, particles sinking from the
1024.5–1025 <inline-formula><mml:math id="M450" display="inline"><mml:mrow><mml:mi mathvariant="normal">kg</mml:mi><mml:mspace width="0.125em" linebreak="nobreak"/><mml:msup><mml:mi mathvariant="normal">m</mml:mi><mml:mrow><mml:mo>-</mml:mo><mml:mn mathvariant="normal">3</mml:mn></mml:mrow></mml:msup></mml:mrow></mml:math></inline-formula> density layer during the later part of
LDC will influence subsequent sediment trap collection. While water mass
changes near the sediment trap depths are unlikely due to weaker currents,
the example of LDC suggests the probability is non-negligible. In our study,
since the CTD casts were focused on the upper 200 <inline-formula><mml:math id="M451" display="inline"><mml:mi mathvariant="normal">m</mml:mi></mml:math></inline-formula> of the water
column, there were insufficient <inline-formula><mml:math id="M452" display="inline"><mml:mi>T</mml:mi></mml:math></inline-formula>-<inline-formula><mml:math id="M453" display="inline"><mml:mi>S</mml:mi></mml:math></inline-formula> measurements to conduct the time series
analysis for all the sediment traps. We therefore cannot use this method to
evaluate if the trap data reflect the same upper water column as observed
during each LD station.</p>
      <p id="d1e6062">In summation, the inability of the quasi-Lagrangian drifter to follow the
water in a sheared flow requires the investigator to consider the spatial
scale of horizontal gradients, which might be depth-dependent, and then to
ask whether the measurement they are making involves integrating over
timescales and depths that mean the measurements being made do not reflect
the other in situ measurements.</p>
</sec>
</sec>
<sec id="Ch1.S5" sec-type="conclusions">
  <title>Conclusions</title>
      <p id="d1e6073">The methodology applied in analyzing the tracer data for OUTPACE is
reflective of a few characteristics specific to this region and to
biogeochemical datasets more generally. First, in an ideal situation, the <inline-formula><mml:math id="M454" display="inline"><mml:mi>T</mml:mi></mml:math></inline-formula>-<inline-formula><mml:math id="M455" display="inline"><mml:mi>S</mml:mi></mml:math></inline-formula>
analysis would determine the component water masses. While this might be
achievable for some well-studied regions of the World Ocean, this is not
generally applicable. Second, alternative well-defined water mass properties,
such as dissolved oxygen, tend to be applicable only at depth, such as
through the thermocline. Naturally, this is because photosynthesis,
respiration, air–sea surface flux, and a panoply of other biogeochemical and
physical processes occur near the surface and impact these tracers, making
them not conservative. Unfortunately, we are precisely interested in
analyzing the water properties of the surface ocean where these processes are
most intense. As a result, methodologies need to be developed so that
standard <inline-formula><mml:math id="M456" display="inline"><mml:mi>T</mml:mi></mml:math></inline-formula>-<inline-formula><mml:math id="M457" display="inline"><mml:mi>S</mml:mi></mml:math></inline-formula> measurements, unaffected by biogeochemical processes, can be
used to quantify the effect of the physical circulation upon the biological
environment encountered in a field campaign.</p>
      <p id="d1e6104">The multiple in situ data sources compiled during the LD stations of the
OUTPACE cruise allowed for the determination of whether the ship, and its
associated quasi-Lagrangian drifting array, sampled the same physical
environment, which was divided into density layers. The procedures used to do
this consists of several steps. First, one needs to look for large gradients
(physical or biogeochemical) and structures of circulation (fronts or eddies)
that would both impact the trajectory of the drifter and bring different
water masses into close contact. Then, a statistical baseline is created from
the <inline-formula><mml:math id="M458" display="inline"><mml:mi>T</mml:mi></mml:math></inline-formula>-<inline-formula><mml:math id="M459" display="inline"><mml:mi>S</mml:mi></mml:math></inline-formula> variability seen during the station's occupation and transformed into
spice–density coordinates. Comparison of independent <inline-formula><mml:math id="M460" display="inline"><mml:mi>T</mml:mi></mml:math></inline-formula>-<inline-formula><mml:math id="M461" display="inline"><mml:mi>S</mml:mi></mml:math></inline-formula> data to the baseline
is used to calculate <inline-formula><mml:math id="M462" display="inline"><mml:mi>Z</mml:mi></mml:math></inline-formula>-scores, first to analyze the in situ time series of the
drifter deployment, and then to establish a spatial scale <inline-formula><mml:math id="M463" display="inline"><mml:mrow><mml:msub><mml:mi>R</mml:mi><mml:mi>Z</mml:mi></mml:msub></mml:mrow></mml:math></inline-formula> beyond which
the <inline-formula><mml:math id="M464" display="inline"><mml:mi>T</mml:mi></mml:math></inline-formula>-<inline-formula><mml:math id="M465" display="inline"><mml:mi>S</mml:mi></mml:math></inline-formula> differences amount to a change in the physical environment. While
this scale is depth-dependent due to vertical shear, for the OUTPACE cruise a
conservative <inline-formula><mml:math id="M466" display="inline"><mml:mrow><mml:msub><mml:mi>R</mml:mi><mml:mi>Z</mml:mi></mml:msub></mml:mrow></mml:math></inline-formula> applied to all depths was found to be close to the Rossby
radius <inline-formula><mml:math id="M467" display="inline"><mml:mrow><mml:msub><mml:mi>R</mml:mi><mml:mi>D</mml:mi></mml:msub></mml:mrow></mml:math></inline-formula>, in the 35–55 <inline-formula><mml:math id="M468" display="inline"><mml:mi mathvariant="normal">km</mml:mi></mml:math></inline-formula> range. The last step is to then use all
available data regarding currents and drifter positions to evaluate whether
any water parcel could have ostensibly traveled farther than <inline-formula><mml:math id="M469" display="inline"><mml:mrow><mml:msub><mml:mi>R</mml:mi><mml:mi>Z</mml:mi></mml:msub></mml:mrow></mml:math></inline-formula>. During
OUTPACE, some density layers were shown to be variable enough to represent a
change in water mass, and some trajectories surpassed <inline-formula><mml:math id="M470" display="inline"><mml:mrow><mml:msub><mml:mi>R</mml:mi><mml:mi>Z</mml:mi></mml:msub></mml:mrow></mml:math></inline-formula>. For the majority
of measurements, however, this did not occur.</p>
      <p id="d1e6220">The methodology used in this study provides a framework wherein readily
available <inline-formula><mml:math id="M471" display="inline"><mml:mi>T</mml:mi></mml:math></inline-formula>-<inline-formula><mml:math id="M472" display="inline"><mml:mi>S</mml:mi></mml:math></inline-formula> data can be used to answer the same question (whether a single
physical environment was sampled following a quasi-Lagrangian drifting
mooring) for other oceanographic cruises. More traditional methods, such as
absolute water mass determination, or using alternative tracers such as
dissolved oxygen, require prior knowledge of a given region or are not
applicable in the euphotic zone where biogeochemical measurements are made.
While sampling in a Lagrangian manner is preferable to not attempting to
follow a water parcel at all, the inevitable failure to be truly Lagrangian
with these platforms should be recognized so that experiments are not allowed
to either last too long or be deployed in an inappropriate flow regime.
Regarding the use of this methodology, we give a few recommendations for
future cruise sampling:
<list list-type="bullet"><list-item>
      <p id="d1e6239">Maximize use of remote sensing data during the cruise to identify
possible mesoscale features to either avoid or sample inside of. This can be
achieved with software such as SPASSO <xref ref-type="bibr" rid="bib1.bibx36" id="paren.48"/>.</p></list-item><list-item>
      <p id="d1e6246">Upon arrival at the selected site, a deep CTD cast below the thermocline
can be used to quickly calculate the local <inline-formula><mml:math id="M473" display="inline"><mml:mrow><mml:msub><mml:mi>R</mml:mi><mml:mi>D</mml:mi></mml:msub></mml:mrow></mml:math></inline-formula> Rossby radius in real time
and produce a rough estimate for maximum spatial scale. In patchier
environments, this scale might be too large, and must be reduced.</p></list-item><list-item>
      <p id="d1e6261">Before and after each station, sample with a surveying instrument such as
ISIIS, SWIMS, SeaSoar, or MVP beyond <inline-formula><mml:math id="M474" display="inline"><mml:mrow><mml:msub><mml:mi>R</mml:mi><mml:mi>D</mml:mi></mml:msub></mml:mrow></mml:math></inline-formula> to get depth-resolved data at intermediate scales.</p></list-item><list-item>
      <p id="d1e6276">If possible, mount CTDs and current meters on the quasi-Lagrangian drifting
array (perhaps a sediment trap that does not need to be removed constantly).
Multiple independent observations over a large range of spatial scales are
essential to calculate robust <inline-formula><mml:math id="M475" display="inline"><mml:mrow><mml:msub><mml:mi>R</mml:mi><mml:mi>Z</mml:mi></mml:msub></mml:mrow></mml:math></inline-formula> estimates.</p></list-item><list-item>
      <p id="d1e6291">Deploy surface drifters to compensate for the lack of SADCP observations
near the surface and to provide spatial context beyond the research vessel.</p></list-item></list></p>
</sec>

      
      </body>
    <back><notes notes-type="dataavailability">

      <p id="d1e6298">All data and metadata are available at the French INSU/CNRS
LEFE CYBER database (scientific coordinator: Hervé Claustre; data manager and
webmaster: Catherine Schmechtig) at the following web address:
<uri>http://www.obs-vlfr.fr/proof/php/outpace/outpace.php</uri> (INSU/CNRS LEFE CYBER,
2017).<?xmltex \hack{\vspace*{-4mm}}?></p>
  </notes><app-group>
        <supplementary-material position="anchor"><p id="d1e6305">The supplement related to this article is available online at: <inline-supplementary-material xlink:href="https://doi.org/10.5194/bg-15-2125-2018-supplement" xlink:title="pdf">https://doi.org/10.5194/bg-15-2125-2018-supplement</inline-supplementary-material>.<?xmltex \hack{\vspace*{-2mm}}?></p></supplementary-material>
        </app-group><notes notes-type="competinginterests">

      <p id="d1e6315">The authors declare that they have no conflict of
interest.<?xmltex \hack{\vspace*{-2mm}}?></p>
  </notes><notes notes-type="sistatement">

      <p id="d1e6322">This article is part of the special issue “Interactions between
planktonic organisms and biogeochemical cycles across trophic and N2 fixation
gradients in the western tropical South Pacific Ocean: a multidisciplinary
approach (OUTPACE experiment)”. It does not belong to a
conference.<?xmltex \hack{\vspace*{-2mm}}?></p>
  </notes><ack><title>Acknowledgements</title><p id="d1e6329">The authors would like to thank two anonymous reviewers and Xavier Capet for
their insightful comments, and the editor Laurent Mémery. The authors would also
like to thank Gilles Rougier for his help in acquiring and processing CTD,
SVP drifter, and SedTrap Drifter data. Marc Picheral is thanked for his help
with sampling and processing CTD data. Olivier Desprez de Géincourt is also
thanked for use of and processing of AQUADOPP and Microcat CTD data. Rick
Lumpkin and Shaun Dolk from NOAA/AOML are greatly thanked for providing and
sharing SVP drifters and data. Satellite SST, chl <inline-formula><mml:math id="M476" display="inline"><mml:mi>a</mml:mi></mml:math></inline-formula>, and altimetry data have
been provided by CLS in the framework of TOSCA/CNES funding; we warmly thank
Marie Isabelle Pujol and Guillaume Taburet for their support in providing
these data. The authors are also indebted to the captain and crew of the R/V <italic>L'Atalante</italic>
for their outstanding assistance and role in the OUTPACE cruise.
This is a contribution of the OUTPACE (Oligotrophy to UlTra-oligotrophy
PACific Experiment) project (Moutin and Bonnet, 2015) funded by the French
national research agency (ANR-14-CE01-0007-01), the LEFE-CyBER program
(CNRS-INSU), the GOPS program (IRD), CNES (BC T23, ZBC 4500048836), and the
European FEDER Fund under project 1166-39417. The OUTPACE cruise was managed
by MIO (OSU Institut Pytheas, AMU) from Marseille (France). Catherine Schmechtig
is thanked for LEFE-CyBER database management. Aurelia Lozingot is
acknowledged for administrative aid for the OUTPACE project.<?xmltex \hack{\newline}?><?xmltex \hack{\newline}?>
Edited by: Laurent Mémery<?xmltex \hack{\newline}?>
Reviewed by:
two anonymous referees<?xmltex \hack{\vspace*{-2mm}}?></p></ack><ref-list>
    <title>References</title>

      <ref id="bib1.bibx1"><label>Alford et al.(2016)</label><mixed-citation>
Alford, M. H., MacKinnon, J. A., Simmons, H. L., and Nash, J. D.: Near-inertial
internal gravity waves in the ocean, Annu. Rev. Mar. Sci., 8,
95–123, 2016.</mixed-citation></ref>
      <ref id="bib1.bibx2"><label>Alory et al.(2015)</label><mixed-citation>
Alory, G., Delcroix, T., Téchiné, P., Divèrres, D., Varillon, D.,
Cravatte, S., Gouriou, Y., Grelet, J., Jacquin, S., Kestenare, E.,
Maes, C., Morrow, R., Perrier, J., Reverdin, G., and Roubaud, F.: The French contribution to the voluntary observing ships network of
sea surface salinity, Deep-Sea Res. Pt. I, 105, 1–18, 2015.</mixed-citation></ref>
      <ref id="bib1.bibx3"><label>Bograd et al.(2003)</label><mixed-citation>Bograd, S. J., Checkley, D. A., and Wooster, W. S.: CalCOFI: A half century of
physical, chemical, and biological research in the California Current
System, Deep-Sea Res. Pt. II, 50, 2349–2353, <ext-link xlink:href="https://doi.org/10.1016/S0967-0645(03)00122-X" ext-link-type="DOI">10.1016/S0967-0645(03)00122-X</ext-link>,
2003.</mixed-citation></ref>
      <ref id="bib1.bibx4"><label>Bouruet-Aubertot et al.(2018)</label><mixed-citation>Bouruet-Aubertot, P., Cuyper, Y., Le Goff, H., Rougier, G., de Verneil, A.,
Doglioli, A., Picheral, M., Yohia, C., Caffin, M., Lefèvre, D., Petrenko,
A., and Moutin, T.: Longitudinal contrast in turbulence along a <inline-formula><mml:math id="M477" display="inline"><mml:mo>∼</mml:mo></mml:math></inline-formula> 19<inline-formula><mml:math id="M478" display="inline"><mml:msup><mml:mi/><mml:mo>∘</mml:mo></mml:msup></mml:math></inline-formula> S
section in the Pacific and its consequences on biogeochemical fluxes, Biogeosciences Discuss., submitted, 2018.</mixed-citation></ref>
      <ref id="bib1.bibx5"><label>Boyer et al.(2013)</label><mixed-citation>Boyer, T., Antonov, J., Baranova, O., Coleman, C., Garcia, H., Grodsky, A.,
Johnson, D., Locarnini, R., Mishonov, A., O'Brien, T., Paver, C., Reagan, J.,
Seidov, D., Smolyar, I., and Zweng, M.: World Ocean Database 2013, NOAA Atlas,
<ext-link xlink:href="https://doi.org/10.7289/V5NZ85MT" ext-link-type="DOI">10.7289/V5NZ85MT</ext-link>, 2013.</mixed-citation></ref>
      <ref id="bib1.bibx6"><label>Chelton et al.(1998)</label><mixed-citation>
Chelton, D. B., Deszoeke, R. A., Schlax, M. G., El Naggar, K., and Siwertz, N.:
Geographical variability of the first baroclinic Rossby radius of
deformation, J. Phys. Oceanogr., 28, 433–460, 1998.</mixed-citation></ref>
      <ref id="bib1.bibx7"><label>Claustre et al.(2008)</label><mixed-citation>Claustre, H., Sciandra, A., and Vaulot, D.: Introduction to the special section
bio-optical and biogeochemical conditions in the South East
Pacific in late 2004: the BIOSOPE program, Biogeosciences, 5, 679–691, <ext-link xlink:href="https://doi.org/10.5194/bg-5-679-2008" ext-link-type="DOI">10.5194/bg-5-679-2008</ext-link>, 2008.</mixed-citation></ref>
      <ref id="bib1.bibx8"><label>Crosby et al.(1993)</label><mixed-citation>
Crosby, D., Breaker, L., and Gemmill, W.: A proposed definition for vector
correlation in geophysics: Theory and application, J. Atmos.
Ocean. Technol., 10, 355–367, 1993.</mixed-citation></ref>
      <ref id="bib1.bibx9"><label>Davidson and Janssens(2006)</label><mixed-citation>
Davidson, E. A. and Janssens, I. A.: Temperature sensitivity of soil carbon
decomposition and feedbacks to climate change, Nature, 440, 165–173, 2006.</mixed-citation></ref>
      <ref id="bib1.bibx10"><?xmltex \def\ref@label{{de~Boyer~Mont{\'{e}}gut et~al.(2004)}}?><label>de Boyer Montégut et al.(2004)</label><mixed-citation>de Boyer Montégut, C., Madec, G., Fischer, A. S., Lazar, A., and Iudicone,
D.: Mixed layer depth over the global ocean: An examination of profile data
and a profile-based climatology, J. Geophys. Res.-Oceans,
109, C12003, <ext-link xlink:href="https://doi.org/10.1029/2004JC002378" ext-link-type="DOI">10.1029/2004JC002378</ext-link>, 2004.</mixed-citation></ref>
      <ref id="bib1.bibx11"><label>de Verneil et al.(2017)</label><mixed-citation>de Verneil, A., Rousselet, L., Doglioli, A. M., Petrenko, A. A., and Moutin, T.:
The fate of a southwest Pacific bloom: gauging the impact of submesoscale vs. mesoscale
circulation on biological gradients in the subtropics, Biogeosciences, 14, 3471–3486, <ext-link xlink:href="https://doi.org/10.5194/bg-14-3471-2017" ext-link-type="DOI">10.5194/bg-14-3471-2017</ext-link>, 2017.</mixed-citation></ref>
      <ref id="bib1.bibx12"><label>Doglioli et al.(2013)</label><mixed-citation>
Doglioli, A. M., Nencioli, F., Petrenko, A. A., Rougier, G., Fuda, J.-L., and
Grima, N.: A software package and hardware tools for in situ experiments in a
Lagrangian reference frame, J. Atmos. Ocean. Tech.,
30, 1940–1950, 2013.</mixed-citation></ref>
      <ref id="bib1.bibx13"><label>d'Ovidio et al.(2015)</label><mixed-citation>d'Ovidio, F., Della Penna, A., Trull, T. W., Nencioli, F., Pujol, M.-I., Rio, M.-H.,
Park, Y.-H., Cotté, C., Zhou, M., and Blain, S.: The biogeochemical structuring
role of horizontal stirring: Lagrangian perspectives on iron delivery downstream
of the Kerguelen Plateau, Biogeosciences, 12, 5567–5581, <ext-link xlink:href="https://doi.org/10.5194/bg-12-5567-2015" ext-link-type="DOI">10.5194/bg-12-5567-2015</ext-link>, 2015.</mixed-citation></ref>
      <ref id="bib1.bibx14"><label>D'saro et al.(2011)</label><mixed-citation>
D'saro, E., Lee, C., Rainville, L., Harcourt, R., and Thomas, L.: Enhanced
turbulence and energy dissipation at ocean fronts, Science, 332, 318–322,
2011.</mixed-citation></ref>
      <ref id="bib1.bibx15"><label>Falkowski et al.(2000)</label><mixed-citation>
Falkowski, P., Scholes, R. J., Boyle, E., Canadell, J., Canfield, D., Elser, J.,
Gruber, N., Hibbard, K., Högberg, P., Linder, S., Mackenzie, F. T., Moore III, B.,
Pdersen, T., Rosenthal, Y., Seitzinger, S., Smetacek, V., and Steffen, W.: The global
carbon cycle: a test of our knowledge of earth as a system, Science, 290,
291–296, 2000.</mixed-citation></ref>
      <?pagebreak page2147?><ref id="bib1.bibx16"><label>Flament(2002)</label><mixed-citation>
Flament, P.: A state variable for characterizing water masses and their
diffusive stability: spiciness, Prog. Oceanogr., 54, 493–501, 2002.</mixed-citation></ref>
      <ref id="bib1.bibx17"><label>Freeland(2007)</label><mixed-citation>
Freeland, H.: A short history of Ocean Station Papa and Line P, Prog.
Oceanogr., 75, 120–125, 2007.</mixed-citation></ref>
      <ref id="bib1.bibx18"><label>Gruber and Galloway(2008)</label><mixed-citation>
Gruber, N. and Galloway, J. N.: An Earth-system perspective of the global
nitrogen cycle, Nature, 451, 293–296, 2008.</mixed-citation></ref>
      <ref id="bib1.bib1"><label>1</label><mixed-citation>INSU/CNRS LEFE CYBER: OUTPACE – Oligotrophy to UlTra-oligotrophy PACific Experiment,
available at: <uri>http://www.obs-vlfr.fr/proof/php/outpace/outpace.php</uri>, last access: 6 April 2018.</mixed-citation></ref>
      <ref id="bib1.bibx19"><label>Karl and Lukas(1996)</label><mixed-citation>
Karl, D. M. and Lukas, R.: The Hawaii Ocean Time-series (HOT) program:
Background, rationale and field implementation, Deep-Sea Res. Pt. II, 43, 129–156, 1996.</mixed-citation></ref>
      <ref id="bib1.bibx20"><label>Kessler(1999)</label><mixed-citation>
Kessler, W. S.: Interannual variability of the subsurface high salinity tongue
south of the equator at 165 E, J. Phys. Oceanogr., 29,
2038–2049, 1999.</mixed-citation></ref>
      <ref id="bib1.bibx21"><label>LaCasce(2008)</label><mixed-citation>
LaCasce, J.: Statistics from Lagrangian observations, Prog. Oceanogr.,
77, 1–29, 2008.</mixed-citation></ref>
      <ref id="bib1.bibx22"><label>Landry et al.(2009)</label><mixed-citation>
Landry, M. R., Ohman, M. D., Goericke, R., Stukel, M. R., and Tsyrklevich, K.:
Lagrangian studies of phytoplankton growth and grazing relationships in a
coastal upwelling ecosystem off Southern California, Prog.
Oceanogr., 83, 208–216, 2009.</mixed-citation></ref>
      <ref id="bib1.bibx23"><label>Le Bot et al.(2011)</label><mixed-citation>
Le Bot, P., Kermabon, C., Lherminier, P., and Gaillard, F.: CASCADE V6. 1:
Logiciel de validation et de visualisation des mesures ADCP de coque, IFREMER, Centre de Brest, France, 2011.</mixed-citation></ref>
      <ref id="bib1.bibx24"><label>LEFE-CYBER()</label><mixed-citation>LEFE-CYBER: Scientific Cruises, available at: <uri>http://www.obs-vlfr.fr/proof/cruises.php</uri>, last access: 6 April 2018.</mixed-citation></ref>
      <ref id="bib1.bibx25"><label>Lherminier et al.(2007)</label><mixed-citation>Lherminier, P., Mercier, H., Gourcuff, C., Alvarez, M., Bacon, S., and
Kermabon, C.: Transports across the 2002 Greenland-Portugal Ovide section and
comparison with 1997, J. Geophys. Res.-Oceans, 112, C07003, <ext-link xlink:href="https://doi.org/10.1029/2006JC003716" ext-link-type="DOI">10.1029/2006JC003716</ext-link>, 2007.</mixed-citation></ref>
      <ref id="bib1.bibx26"><label>Lumpkin and Pazos(2007)</label><mixed-citation>
Lumpkin, R. and Pazos, M.: Measuring surface currents with Surface Velocity
Program drifters: the instrument, its data, and some recent results,
Lagrangian analysis and prediction of coastal and ocean dynamics, Cambridge University Press, Cambridge, UK,  39–67,
2007.</mixed-citation></ref>
      <ref id="bib1.bibx27"><label>Mackas et al.(1987)</label><mixed-citation>
Mackas, D. L., Denman, K. L., and Bennett, A. F.: Least squares multiple tracer
analysis of water mass composition, J. Geophys. Res.-Oceans,
92, 2907–2918, 1987.</mixed-citation></ref>
      <ref id="bib1.bibx28"><label>Marty et al.(2002)</label><mixed-citation>
Marty, J.-C., Chiavérini, J., Pizay, M.-D., and Avril, B.: Seasonal and
interannual dynamics of nutrients and phytoplankton pigments in the western
Mediterranean Sea at the DYFAMED time-series station (1991–1999), Deep-Sea
Res. Pt. II, 49, 1965–1985, 2002.</mixed-citation></ref>
      <ref id="bib1.bibx29"><label>McDougall and Barker(2011)</label><mixed-citation>
McDougall, T. J. and Barker, P. M.: Getting started with TEOS-10 and the Gibbs
Seawater (GSW) oceanographic toolbox, SCOR/IAPSO WG, 127, 1–28, 2011.</mixed-citation></ref>
      <ref id="bib1.bibx30"><label>Moutin and Bonnet(2015)</label><mixed-citation>Moutin, T. and Bonnet, S.: OUTPACE cruise, RV L'Atalante,
<ext-link xlink:href="https://doi.org/10.17600/15000900" ext-link-type="DOI">10.17600/15000900</ext-link>, 2015.</mixed-citation></ref>
      <ref id="bib1.bibx31"><label>Moutin and Prieur(2012)</label><mixed-citation>Moutin, T. and Prieur, L.: Influence of anticyclonic eddies on the Biogeochemistry
from the Oligotrophic to the Ultraoligotrophic Mediterranean
(BOUM cruise), Biogeosciences, 9, 3827–3855, <ext-link xlink:href="https://doi.org/10.5194/bg-9-3827-2012" ext-link-type="DOI">10.5194/bg-9-3827-2012</ext-link>, 2012.</mixed-citation></ref>
      <ref id="bib1.bibx32"><label>Moutin et al.(2012)</label><mixed-citation>Moutin, T., Van Wambeke, F., and Prieur, L.: Introduction to the
Biogeochemistry from the Oligotrophic to the Ultraoligotrophic
Mediterranean (BOUM) experiment, Biogeosciences, 9, 3817–3825, <ext-link xlink:href="https://doi.org/10.5194/bg-9-3817-2012" ext-link-type="DOI">10.5194/bg-9-3817-2012</ext-link>, 2012.</mixed-citation></ref>
      <ref id="bib1.bibx33"><label>Moutin et al.(2017)</label><mixed-citation>Moutin, T., Doglioli, A. M., de Verneil, A., and Bonnet, S.: Preface: The
Oligotrophy to the UlTra-oligotrophy PACific Experiment (OUTPACE cruise,
18 February to 3 April 2015), Biogeosciences, 14, 3207–3220, <ext-link xlink:href="https://doi.org/10.5194/bg-14-3207-2017" ext-link-type="DOI">10.5194/bg-14-3207-2017</ext-link>, 2017.</mixed-citation></ref>
      <ref id="bib1.bibx34"><label>Munk(1981)</label><mixed-citation>
Munk, W. H.: Internal waves and small-scales processes, in: Evolution of Physical
Oceanography, Scientific Surveys in Honor of Henry Stommel, edited by: Warren, B. A. and Wunsch, C., 264–291, MIT Press, Cambridge, Mass., 1981.</mixed-citation></ref>
      <ref id="bib1.bibx35"><label>Ohlmann et al.(2017)</label><mixed-citation>
Ohlmann, J., Molemaker, M., Baschek, B., Holt, B., Marmorino, G., and Smith,
G.: Drifter observations of submesoscale flow kinematics in the coastal
ocean, Geophys. Res. Lett., 44, 330–337, 2017.</mixed-citation></ref>
      <ref id="bib1.bibx36"><label>Petrenko et al.(2017)</label><mixed-citation>
Petrenko, A. A., Doglioli, A. M., Nencioli, F., Kersalé, M., Hu, Z., and
d'Ovidio, F.: A review of the LATEX project: mesoscale to submesoscale
processes in a coastal environment, Ocean Dynam., 67, 513–533, 2017.</mixed-citation></ref>
      <ref id="bib1.bibx37"><label>Poole and Tomczak(1999)</label><mixed-citation>
Poole, R. and Tomczak, M.: Optimum multiparameter analysis of the water mass
structure in the Atlantic Ocean thermocline, Deep-Sea Res. Pt. I, 46, 1895–1921, 1999.</mixed-citation></ref>
      <ref id="bib1.bibx38"><label>Rousselet et al.(2016)</label><mixed-citation>
Rousselet, L., Doglioli, A., Maes, C., Blanke, B., and Petrenko, A.: Impacts of
mesoscale activity on the water masses and circulation in the Coral Sea,
J. Geophys. Res.-Oceans, 121, 7277–7289, 2016.</mixed-citation></ref>
      <ref id="bib1.bibx39"><label>Rousselet et al.(2018)</label><mixed-citation>Rousselet, L., de Verneil, A., Doglioli, A. M., Petrenko, A. A., Duhamel, S., Maes, C.,
and Blanke, B.: Large to submesoscale surface circulation and its implications on
biogeochemical/biological horizontal distributions during the OUTPACE cruise
(SouthWest Pacific), Biogeosciences Discuss., <ext-link xlink:href="https://doi.org/10.5194/bg-2017-456" ext-link-type="DOI">10.5194/bg-2017-456</ext-link>, in review, 2017.</mixed-citation></ref>
      <ref id="bib1.bibx40"><label>Schroeder and Stommel(1969)</label><mixed-citation>Schroeder, E. and Stommel, H.: How representative is the series of Panulirus
stations of monthly mean conditions off Bermuda?, Prog. Oceanogr.,
5, 31–40, <ext-link xlink:href="https://doi.org/10.1016/0079-6611(69)90025-1" ext-link-type="DOI">10.1016/0079-6611(69)90025-1</ext-link>, 1969.</mixed-citation></ref>
      <ref id="bib1.bibx41"><label>Send et al.(2010)</label><mixed-citation>
Send, U., Weller, R.A., Wallace, D., Chavez, F., Lampitt, R., Dickey, T., Honda, M., Nittis, K., Lukas, R., McPhaden, M., and Feely, R.:
OceanSITES, European Space Agency, Noordwijk, the Netherlands, 2010.</mixed-citation></ref>
      <ref id="bib1.bibx42"><label>Smith and Ferrari(2009)</label><mixed-citation>
Smith, K. S. and Ferrari, R.: The production and dissipation of compensated
thermohaline variance by mesoscale stirring, J. Phys.
Oceanogr., 39, 2477–2501, 2009.</mixed-citation></ref>
      <ref id="bib1.bibx43"><label>Steinberg et al.(2001)</label><mixed-citation>
Steinberg, D. K., Carlson, C. A., Bates, N. R., Johnson, R. J., Michaels,
A. F., and Knap, A. H.: Overview of the US JGOFS Bermuda Atlantic Time-series
Study (BATS): a decade-scale look at ocean biology and biogeochemistry, Deep-Sea Res. Pt. II, 48, 1405–1447, 2001.</mixed-citation></ref>
      <ref id="bib1.bibx44"><label>Stommel(1962)</label><mixed-citation>
Stommel, H.: On the cause of the temperature-salinity curve in the ocean,
P. Natl. Acad. Sci. USA, 48, 764–766, 1962.</mixed-citation></ref>
      <ref id="bib1.bibx45"><label>White et al.(2007)</label><mixed-citation>White, A. E., Spitz, Y. H., and Letelier, R. M.: What factors are driving
summer phytoplankton blooms in the North Pacific Subtropical Gyre?, J. Geophys. Res.-Oceans, 112,
C12006,
<ext-link xlink:href="https://doi.org/10.1029/2007JC004129" ext-link-type="DOI">10.1029/2007JC004129</ext-link>, 2007.</mixed-citation></ref>

  </ref-list></back>
    <!--<article-title-html>OUTPACE long duration stations: physical variability, context of biogeochemical sampling, and evaluation of sampling strategy</article-title-html>
<abstract-html><p>Research cruises to quantify biogeochemical fluxes in the ocean require
taking measurements at stations lasting at least several days. A popular
experimental design is the quasi-Lagrangian drifter, often mounted with in
situ incubations or sediment traps that follow the flow of water over time. After initial
drifter deployment, the ship tracks the drifter for continuing measurements
that are supposed to represent the same water environment. An outstanding
question is how to best determine whether this is true. During the
Oligotrophy to UlTra-oligotrophy PACific Experiment (OUTPACE) cruise, from 18
February to 3 April 2015 in the western tropical South Pacific, three
separate stations of long duration (five days) over the upper 500&thinsp;m
were conducted in this quasi-Lagrangian sampling scheme. Here we present
physical data to provide context for these three stations and to assess
whether the sampling strategy worked, i.e., that a single body of water
was sampled. After analyzing tracer variability and local water circulation
at each station, we identify water layers and times where the drifter risks
encountering another body of water. While almost no realization of this
sampling scheme will be truly Lagrangian, due to the presence of vertical
shear, the depth-resolved observations during the three stations show most
layers sampled sufficiently homogeneous physical environments during OUTPACE.
By directly addressing the concerns raised by these quasi-Lagrangian sampling
platforms, a protocol of best practices can begin to be formulated so that
future research campaigns include the complementary datasets and analyses
presented here to verify the appropriate use of the drifter platform.</p></abstract-html>
<ref-html id="bib1.bib1"><label>Alford et al.(2016)</label><mixed-citation>
Alford, M. H., MacKinnon, J. A., Simmons, H. L., and Nash, J. D.: Near-inertial
internal gravity waves in the ocean, Annu. Rev. Mar. Sci., 8,
95–123, 2016.
</mixed-citation></ref-html>
<ref-html id="bib1.bib2"><label>Alory et al.(2015)</label><mixed-citation>
Alory, G., Delcroix, T., Téchiné, P., Divèrres, D., Varillon, D.,
Cravatte, S., Gouriou, Y., Grelet, J., Jacquin, S., Kestenare, E.,
Maes, C., Morrow, R., Perrier, J., Reverdin, G., and Roubaud, F.: The French contribution to the voluntary observing ships network of
sea surface salinity, Deep-Sea Res. Pt. I, 105, 1–18, 2015.
</mixed-citation></ref-html>
<ref-html id="bib1.bib3"><label>Bograd et al.(2003)</label><mixed-citation>
Bograd, S. J., Checkley, D. A., and Wooster, W. S.: CalCOFI: A half century of
physical, chemical, and biological research in the California Current
System, Deep-Sea Res. Pt. II, 50, 2349–2353, <a href="https://doi.org/10.1016/S0967-0645(03)00122-X" target="_blank">https://doi.org/10.1016/S0967-0645(03)00122-X</a>,
2003.
</mixed-citation></ref-html>
<ref-html id="bib1.bib4"><label>Bouruet-Aubertot et al.(2018)</label><mixed-citation>
Bouruet-Aubertot, P., Cuyper, Y., Le Goff, H., Rougier, G., de Verneil, A.,
Doglioli, A., Picheral, M., Yohia, C., Caffin, M., Lefèvre, D., Petrenko,
A., and Moutin, T.: Longitudinal contrast in turbulence along a  ∼ &thinsp;19°&thinsp;S
section in the Pacific and its consequences on biogeochemical fluxes, Biogeosciences Discuss., submitted, 2018.
</mixed-citation></ref-html>
<ref-html id="bib1.bib5"><label>Boyer et al.(2013)</label><mixed-citation>
Boyer, T., Antonov, J., Baranova, O., Coleman, C., Garcia, H., Grodsky, A.,
Johnson, D., Locarnini, R., Mishonov, A., O'Brien, T., Paver, C., Reagan, J.,
Seidov, D., Smolyar, I., and Zweng, M.: World Ocean Database 2013, NOAA Atlas,
<a href="https://doi.org/10.7289/V5NZ85MT" target="_blank">https://doi.org/10.7289/V5NZ85MT</a>, 2013.
</mixed-citation></ref-html>
<ref-html id="bib1.bib6"><label>Chelton et al.(1998)</label><mixed-citation>
Chelton, D. B., Deszoeke, R. A., Schlax, M. G., El Naggar, K., and Siwertz, N.:
Geographical variability of the first baroclinic Rossby radius of
deformation, J. Phys. Oceanogr., 28, 433–460, 1998.
</mixed-citation></ref-html>
<ref-html id="bib1.bib7"><label>Claustre et al.(2008)</label><mixed-citation>
Claustre, H., Sciandra, A., and Vaulot, D.: Introduction to the special section
bio-optical and biogeochemical conditions in the South East
Pacific in late 2004: the BIOSOPE program, Biogeosciences, 5, 679–691, <a href="https://doi.org/10.5194/bg-5-679-2008" target="_blank">https://doi.org/10.5194/bg-5-679-2008</a>, 2008.
</mixed-citation></ref-html>
<ref-html id="bib1.bib8"><label>Crosby et al.(1993)</label><mixed-citation>
Crosby, D., Breaker, L., and Gemmill, W.: A proposed definition for vector
correlation in geophysics: Theory and application, J. Atmos.
Ocean. Technol., 10, 355–367, 1993.
</mixed-citation></ref-html>
<ref-html id="bib1.bib9"><label>Davidson and Janssens(2006)</label><mixed-citation>
Davidson, E. A. and Janssens, I. A.: Temperature sensitivity of soil carbon
decomposition and feedbacks to climate change, Nature, 440, 165–173, 2006.
</mixed-citation></ref-html>
<ref-html id="bib1.bib10"><label>de Boyer Montégut et al.(2004)</label><mixed-citation>
de Boyer Montégut, C., Madec, G., Fischer, A. S., Lazar, A., and Iudicone,
D.: Mixed layer depth over the global ocean: An examination of profile data
and a profile-based climatology, J. Geophys. Res.-Oceans,
109, C12003, <a href="https://doi.org/10.1029/2004JC002378" target="_blank">https://doi.org/10.1029/2004JC002378</a>, 2004.
</mixed-citation></ref-html>
<ref-html id="bib1.bib11"><label>de Verneil et al.(2017)</label><mixed-citation>
de Verneil, A., Rousselet, L., Doglioli, A. M., Petrenko, A. A., and Moutin, T.:
The fate of a southwest Pacific bloom: gauging the impact of submesoscale vs. mesoscale
circulation on biological gradients in the subtropics, Biogeosciences, 14, 3471–3486, <a href="https://doi.org/10.5194/bg-14-3471-2017" target="_blank">https://doi.org/10.5194/bg-14-3471-2017</a>, 2017.
</mixed-citation></ref-html>
<ref-html id="bib1.bib12"><label>Doglioli et al.(2013)</label><mixed-citation>
Doglioli, A. M., Nencioli, F., Petrenko, A. A., Rougier, G., Fuda, J.-L., and
Grima, N.: A software package and hardware tools for in situ experiments in a
Lagrangian reference frame, J. Atmos. Ocean. Tech.,
30, 1940–1950, 2013.
</mixed-citation></ref-html>
<ref-html id="bib1.bib13"><label>d'Ovidio et al.(2015)</label><mixed-citation>
d'Ovidio, F., Della Penna, A., Trull, T. W., Nencioli, F., Pujol, M.-I., Rio, M.-H.,
Park, Y.-H., Cotté, C., Zhou, M., and Blain, S.: The biogeochemical structuring
role of horizontal stirring: Lagrangian perspectives on iron delivery downstream
of the Kerguelen Plateau, Biogeosciences, 12, 5567–5581, <a href="https://doi.org/10.5194/bg-12-5567-2015" target="_blank">https://doi.org/10.5194/bg-12-5567-2015</a>, 2015.
</mixed-citation></ref-html>
<ref-html id="bib1.bib14"><label>D'saro et al.(2011)</label><mixed-citation>
D'saro, E., Lee, C., Rainville, L., Harcourt, R., and Thomas, L.: Enhanced
turbulence and energy dissipation at ocean fronts, Science, 332, 318–322,
2011.
</mixed-citation></ref-html>
<ref-html id="bib1.bib15"><label>Falkowski et al.(2000)</label><mixed-citation>
Falkowski, P., Scholes, R. J., Boyle, E., Canadell, J., Canfield, D., Elser, J.,
Gruber, N., Hibbard, K., Högberg, P., Linder, S., Mackenzie, F. T., Moore III, B.,
Pdersen, T., Rosenthal, Y., Seitzinger, S., Smetacek, V., and Steffen, W.: The global
carbon cycle: a test of our knowledge of earth as a system, Science, 290,
291–296, 2000.
</mixed-citation></ref-html>
<ref-html id="bib1.bib16"><label>Flament(2002)</label><mixed-citation>
Flament, P.: A state variable for characterizing water masses and their
diffusive stability: spiciness, Prog. Oceanogr., 54, 493–501, 2002.
</mixed-citation></ref-html>
<ref-html id="bib1.bib17"><label>Freeland(2007)</label><mixed-citation>
Freeland, H.: A short history of Ocean Station Papa and Line P, Prog.
Oceanogr., 75, 120–125, 2007.
</mixed-citation></ref-html>
<ref-html id="bib1.bib18"><label>Gruber and Galloway(2008)</label><mixed-citation>
Gruber, N. and Galloway, J. N.: An Earth-system perspective of the global
nitrogen cycle, Nature, 451, 293–296, 2008.
</mixed-citation></ref-html>
<ref-html id="bib1.bib19"><label>1</label><mixed-citation>
INSU/CNRS LEFE CYBER: OUTPACE – Oligotrophy to UlTra-oligotrophy PACific Experiment,
available at: <a href="http://www.obs-vlfr.fr/proof/php/outpace/outpace.php" target="_blank">http://www.obs-vlfr.fr/proof/php/outpace/outpace.php</a>, last access: 6 April 2018.
</mixed-citation></ref-html>
<ref-html id="bib1.bib20"><label>Karl and Lukas(1996)</label><mixed-citation>
Karl, D. M. and Lukas, R.: The Hawaii Ocean Time-series (HOT) program:
Background, rationale and field implementation, Deep-Sea Res. Pt. II, 43, 129–156, 1996.
</mixed-citation></ref-html>
<ref-html id="bib1.bib21"><label>Kessler(1999)</label><mixed-citation>
Kessler, W. S.: Interannual variability of the subsurface high salinity tongue
south of the equator at 165&thinsp;E, J. Phys. Oceanogr., 29,
2038–2049, 1999.
</mixed-citation></ref-html>
<ref-html id="bib1.bib22"><label>LaCasce(2008)</label><mixed-citation>
LaCasce, J.: Statistics from Lagrangian observations, Prog. Oceanogr.,
77, 1–29, 2008.
</mixed-citation></ref-html>
<ref-html id="bib1.bib23"><label>Landry et al.(2009)</label><mixed-citation>
Landry, M. R., Ohman, M. D., Goericke, R., Stukel, M. R., and Tsyrklevich, K.:
Lagrangian studies of phytoplankton growth and grazing relationships in a
coastal upwelling ecosystem off Southern California, Prog.
Oceanogr., 83, 208–216, 2009.
</mixed-citation></ref-html>
<ref-html id="bib1.bib24"><label>Le Bot et al.(2011)</label><mixed-citation>
Le Bot, P., Kermabon, C., Lherminier, P., and Gaillard, F.: CASCADE V6. 1:
Logiciel de validation et de visualisation des mesures ADCP de coque, IFREMER, Centre de Brest, France, 2011.
</mixed-citation></ref-html>
<ref-html id="bib1.bib25"><label>LEFE-CYBER()</label><mixed-citation>
LEFE-CYBER: Scientific Cruises, available at: <a href="http://www.obs-vlfr.fr/proof/cruises.php" target="_blank">http://www.obs-vlfr.fr/proof/cruises.php</a>, last access: 6 April 2018.
</mixed-citation></ref-html>
<ref-html id="bib1.bib26"><label>Lherminier et al.(2007)</label><mixed-citation>
Lherminier, P., Mercier, H., Gourcuff, C., Alvarez, M., Bacon, S., and
Kermabon, C.: Transports across the 2002 Greenland-Portugal Ovide section and
comparison with 1997, J. Geophys. Res.-Oceans, 112, C07003, <a href="https://doi.org/10.1029/2006JC003716" target="_blank">https://doi.org/10.1029/2006JC003716</a>, 2007.
</mixed-citation></ref-html>
<ref-html id="bib1.bib27"><label>Lumpkin and Pazos(2007)</label><mixed-citation>
Lumpkin, R. and Pazos, M.: Measuring surface currents with Surface Velocity
Program drifters: the instrument, its data, and some recent results,
Lagrangian analysis and prediction of coastal and ocean dynamics, Cambridge University Press, Cambridge, UK,  39–67,
2007.
</mixed-citation></ref-html>
<ref-html id="bib1.bib28"><label>Mackas et al.(1987)</label><mixed-citation>
Mackas, D. L., Denman, K. L., and Bennett, A. F.: Least squares multiple tracer
analysis of water mass composition, J. Geophys. Res.-Oceans,
92, 2907–2918, 1987.
</mixed-citation></ref-html>
<ref-html id="bib1.bib29"><label>Marty et al.(2002)</label><mixed-citation>
Marty, J.-C., Chiavérini, J., Pizay, M.-D., and Avril, B.: Seasonal and
interannual dynamics of nutrients and phytoplankton pigments in the western
Mediterranean Sea at the DYFAMED time-series station (1991–1999), Deep-Sea
Res. Pt. II, 49, 1965–1985, 2002.
</mixed-citation></ref-html>
<ref-html id="bib1.bib30"><label>McDougall and Barker(2011)</label><mixed-citation>
McDougall, T. J. and Barker, P. M.: Getting started with TEOS-10 and the Gibbs
Seawater (GSW) oceanographic toolbox, SCOR/IAPSO WG, 127, 1–28, 2011.
</mixed-citation></ref-html>
<ref-html id="bib1.bib31"><label>Moutin and Bonnet(2015)</label><mixed-citation>
Moutin, T. and Bonnet, S.: OUTPACE cruise, RV L'Atalante,
<a href="https://doi.org/10.17600/15000900" target="_blank">https://doi.org/10.17600/15000900</a>, 2015.
</mixed-citation></ref-html>
<ref-html id="bib1.bib32"><label>Moutin and Prieur(2012)</label><mixed-citation>
Moutin, T. and Prieur, L.: Influence of anticyclonic eddies on the Biogeochemistry
from the Oligotrophic to the Ultraoligotrophic Mediterranean
(BOUM cruise), Biogeosciences, 9, 3827–3855, <a href="https://doi.org/10.5194/bg-9-3827-2012" target="_blank">https://doi.org/10.5194/bg-9-3827-2012</a>, 2012.
</mixed-citation></ref-html>
<ref-html id="bib1.bib33"><label>Moutin et al.(2012)</label><mixed-citation>
Moutin, T., Van Wambeke, F., and Prieur, L.: Introduction to the
Biogeochemistry from the Oligotrophic to the Ultraoligotrophic
Mediterranean (BOUM) experiment, Biogeosciences, 9, 3817–3825, <a href="https://doi.org/10.5194/bg-9-3817-2012" target="_blank">https://doi.org/10.5194/bg-9-3817-2012</a>, 2012.
</mixed-citation></ref-html>
<ref-html id="bib1.bib34"><label>Moutin et al.(2017)</label><mixed-citation>
Moutin, T., Doglioli, A. M., de Verneil, A., and Bonnet, S.: Preface: The
Oligotrophy to the UlTra-oligotrophy PACific Experiment (OUTPACE cruise,
18 February to 3 April 2015), Biogeosciences, 14, 3207–3220, <a href="https://doi.org/10.5194/bg-14-3207-2017" target="_blank">https://doi.org/10.5194/bg-14-3207-2017</a>, 2017.
</mixed-citation></ref-html>
<ref-html id="bib1.bib35"><label>Munk(1981)</label><mixed-citation>
Munk, W. H.: Internal waves and small-scales processes, in: Evolution of Physical
Oceanography, Scientific Surveys in Honor of Henry Stommel, edited by: Warren, B. A. and Wunsch, C., 264–291, MIT Press, Cambridge, Mass., 1981.
</mixed-citation></ref-html>
<ref-html id="bib1.bib36"><label>Ohlmann et al.(2017)</label><mixed-citation>
Ohlmann, J., Molemaker, M., Baschek, B., Holt, B., Marmorino, G., and Smith,
G.: Drifter observations of submesoscale flow kinematics in the coastal
ocean, Geophys. Res. Lett., 44, 330–337, 2017.
</mixed-citation></ref-html>
<ref-html id="bib1.bib37"><label>Petrenko et al.(2017)</label><mixed-citation>
Petrenko, A. A., Doglioli, A. M., Nencioli, F., Kersalé, M., Hu, Z., and
d'Ovidio, F.: A review of the LATEX project: mesoscale to submesoscale
processes in a coastal environment, Ocean Dynam., 67, 513–533, 2017.
</mixed-citation></ref-html>
<ref-html id="bib1.bib38"><label>Poole and Tomczak(1999)</label><mixed-citation>
Poole, R. and Tomczak, M.: Optimum multiparameter analysis of the water mass
structure in the Atlantic Ocean thermocline, Deep-Sea Res. Pt. I, 46, 1895–1921, 1999.
</mixed-citation></ref-html>
<ref-html id="bib1.bib39"><label>Rousselet et al.(2016)</label><mixed-citation>
Rousselet, L., Doglioli, A., Maes, C., Blanke, B., and Petrenko, A.: Impacts of
mesoscale activity on the water masses and circulation in the Coral Sea,
J. Geophys. Res.-Oceans, 121, 7277–7289, 2016.
</mixed-citation></ref-html>
<ref-html id="bib1.bib40"><label>Rousselet et al.(2018)</label><mixed-citation>
Rousselet, L., de Verneil, A., Doglioli, A. M., Petrenko, A. A., Duhamel, S., Maes, C.,
and Blanke, B.: Large to submesoscale surface circulation and its implications on
biogeochemical/biological horizontal distributions during the OUTPACE cruise
(SouthWest Pacific), Biogeosciences Discuss., <a href="https://doi.org/10.5194/bg-2017-456" target="_blank">https://doi.org/10.5194/bg-2017-456</a>, in review, 2017.
</mixed-citation></ref-html>
<ref-html id="bib1.bib41"><label>Schroeder and Stommel(1969)</label><mixed-citation>
Schroeder, E. and Stommel, H.: How representative is the series of Panulirus
stations of monthly mean conditions off Bermuda?, Prog. Oceanogr.,
5, 31–40, <a href="https://doi.org/10.1016/0079-6611(69)90025-1" target="_blank">https://doi.org/10.1016/0079-6611(69)90025-1</a>, 1969.
</mixed-citation></ref-html>
<ref-html id="bib1.bib42"><label>Send et al.(2010)</label><mixed-citation>
Send, U., Weller, R.A., Wallace, D., Chavez, F., Lampitt, R., Dickey, T., Honda, M., Nittis, K., Lukas, R., McPhaden, M., and Feely, R.:
OceanSITES, European Space Agency, Noordwijk, the Netherlands, 2010.
</mixed-citation></ref-html>
<ref-html id="bib1.bib43"><label>Smith and Ferrari(2009)</label><mixed-citation>
Smith, K. S. and Ferrari, R.: The production and dissipation of compensated
thermohaline variance by mesoscale stirring, J. Phys.
Oceanogr., 39, 2477–2501, 2009.
</mixed-citation></ref-html>
<ref-html id="bib1.bib44"><label>Steinberg et al.(2001)</label><mixed-citation>
Steinberg, D. K., Carlson, C. A., Bates, N. R., Johnson, R. J., Michaels,
A. F., and Knap, A. H.: Overview of the US JGOFS Bermuda Atlantic Time-series
Study (BATS): a decade-scale look at ocean biology and biogeochemistry, Deep-Sea Res. Pt. II, 48, 1405–1447, 2001.
</mixed-citation></ref-html>
<ref-html id="bib1.bib45"><label>Stommel(1962)</label><mixed-citation>
Stommel, H.: On the cause of the temperature-salinity curve in the ocean,
P. Natl. Acad. Sci. USA, 48, 764–766, 1962.
</mixed-citation></ref-html>
<ref-html id="bib1.bib46"><label>White et al.(2007)</label><mixed-citation>
White, A. E., Spitz, Y. H., and Letelier, R. M.: What factors are driving
summer phytoplankton blooms in the North Pacific Subtropical Gyre?, J. Geophys. Res.-Oceans, 112,
C12006,
<a href="https://doi.org/10.1029/2007JC004129" target="_blank">https://doi.org/10.1029/2007JC004129</a>, 2007.
</mixed-citation></ref-html>--></article>
