Antarctic coastal polynyas are persistent and recurrent regions of open water located between the coast and the drifting pack-ice. In spring, they are the first polar areas to be exposed to light, leading to the development of phytoplankton blooms, making polynyas potential ecological hotspots in sea-ice regions. Knowledge on polynya oceanography and ecology during winter is limited due to their inaccessibility. This study describes i) the first in situ chlorophyll fluorescence signal (a proxy for chlorophyll-a concentration and thus presence of phytoplankton) in polynyas between the end of summer and winter, ii) assesses whether the signal persists through time and iii) identifies its main oceanographic drivers. The dataset comprises 698 profiles of fluorescence, temperature and salinity recorded by southern elephant seals in 2011, 2019–2021 in the Cape-Darnley (CDP;67˚S-69˚E) and Shackleton (SP;66˚S-95˚E) polynyas between February and September. A significant fluorescence signal was observed until April in both polynyas. An additional signal occurring at 130m depth in August within CDP may result from in situ growth of phytoplankton due to potential adaptation to low irradiance or remnant chlorophyll-a that was advected into the polynya. The decrease and deepening of the fluorescence signal from February to August were accompanied by the deepening of the mixed layer depth and a cooling and salinification of the water column in both polynyas. Using Principal Component Analysis as an exploratory tool, we highlighted previously unsuspected drivers of the fluorescence signal within polynyas. CDP shows clear differences in biological and environmental conditions depending on topographic features with higher fluorescence in warmer and saltier waters on the shelf compared with the continental slope. In SP, near the ice-shelf, a significant fluorescence signal in April below the mixed layer (around 130m depth), was associated with fresher and warmer waters. We hypothesize that this signal could result from potential ice-shelf melting from warm water intrusions onto the shelf leading to iron supply necessary to fuel phytoplankton growth. This study supports that Antarctic coastal polynyas may have a key role for polar ecosystems as biologically active areas throughout the season within the sea-ice region despite inter and intra-polynya differences in environmental conditions.

The dataset is composed of in situ profiles of temperature, salinity and fluorescence recorded by nine post-moulting male southern elephant seals (SES) equipped with CTD-Fluo-SRDLs (Conductivity-Temperature-Depth-Fluorescence Satellite Relay Data Loggers) (Boehme et al., 2009; Guinet et al., 2013) between 2011 and 2021. All SES visit a coastal polynya in East Antarctica at least once. Individual seals were anesthetized with an intravenous injection of a 1:1 combination of tiletamine and zolazepam (Zoletil 100) (McMahon et al., 2000; Field et al., 2002) to attach the instrumentation. The data loggers were glued to the seals heads using quick-setting epoxy (Araldite AW 2101, Ciba; Field et al., 2012). Data on seal’s diving behavior as well as in situ hydrographic conditions were transmitted when the seal surfaces to breathe through communication with polar-orbiting Argos satellites (Harcourt et al., 2019). 

Fluorescence measurements are used as a proxy to estimate the concentration of chlorophyll-a in the water (Guinet et al., 2013). In this study, we have chosen not to convert chlorophyll-a fluorescence into chlorophyll-a concentration, as conversion ratios are highly variable in the Southern Ocean (Schallenberg et al., 2022), which could affect the validity of results interpretations. However, although quantitative assessments of phytoplankton biomass can hardly be reached with the sole use of fluorometers (Roesler et al., 2017; Petit et al., 2022), in vivo chlorophyll-a fluorescence is a commonly used method to detect presence of living phytoplanktonic organisms and study their dynamics in terms of concentration (IOCCG, 2011). 

Each profile of fluorescence, salinity and temperature had a vertical resolution of one meter after interpolation. No fluorescence data were available above 10 m and below 170 m. The mixed layer depth (MLD) was computed from temperature and salinity profiles using a density criteria Δρ = 0.03kg.m^-3, with density at 10 m depth used as the reference value following de Boyer Montégut et al. (2004).

Fluorescence profiles can be corrected for non-photochemical quenching (NPQ) with a correction algorithm developed by Xing et al. (2018). The NPQ effect is induced by photo-inhibition in phytoplankton cells exposed to high light levels. This method aims at correcting the depression observed in the fluorescence signal (generally in the surface layer) resulting from the NPQ effect. As all profiles start at 10 m depth and the study focuses on autumn and winter, the effect of NPQ is negligible. However, when light was available, which is the case for profiles recorded in 2020 and 2021 (236 profiles out of 766), a correction was still applied. The correction is based on the definition of a “NPQ-layer”. Upper and lower boundaries are, respectively, the ocean surface, and the shallowest value between MLD and a light-threshold depth fixed at 15 μmol. m^-2s^-1 (PAR15, expressed in m). The maximum fluorescence value in the so-called NPQ-layer is then extrapolated up to 10 m depth which, as a result, discards the fluorescence depression due to NPQ (for details, see Xing et al. (2018)). Moreover, background noise from the fluorescence sensor detected at 170 m depth (where it is assumed that the fluorescence signal should be null) was removed from all the 766 fluorescence profiles.

The geographical positions transmitted via the Argos system use the Doppler shift of the signal frequency received during the passing of one of the polar-orbitting satellites and depending on the position of the satellite relative to the transmitting tags, errors from 0.5 to 10 km may be observed. To provide the best location estimates we (i) filter the trajectories of individuals in order to (ii) interpolate the position of the CTD profiles from the filtered trajectory. The method used is based on a continuous-time state-space model presented in Jonsen et al. (2020) and implemented using the R package foieGras (now known as Animotum (Jonsen et al., 2019; Jonsen et al., 2020; Jonsen et al., 2023)).

Each profile was assigned to a broad regional area based on its location. These three areas: the continental shelf, the continental slope and the pelagic zone, were determined from the GEBCO_2021 Grided bathymetry (Gebco, 2021). The latitudinal limits of the continental slope were determined by graphic visualization every 0.5˚ longitudes (between -0.5˚E and 158°E) by plotting bathymetry against latitude. The beginning of the slope corresponds to the latitude where bathymetry starts to decrease drastically while the end of the slope corresponds to the first latitude where bathymetry is deeper than 2,900 m.

The data are in a .txt format and the codes are in R and Matlab. See Readme.md for details.

Please do not hesite to contact the corresponding author (Lucie Bourreau, lucie.bourreau.1@ulaval.ca) if you have any questions.