During the 38 months between December 2018 and January 2022,
multiparameter hydrographic measurements were taken at 31 sites
within Admiralty Bay, King George Island, Antarctica. These records
consisted of water column measurements (down to 100 m) of temperature,
conductivity, turbidity, and pH as well as the dissolved oxygen, dissolved
organic matter, chlorophyll-
When freshwater from glaciers is introduced to marine environments, it mixes with ambient ocean water masses, leading to the formation of new glacially modified water (GMW; Straneo, 2012). In this way, freshwater export has been shown to influence properties of the coastal ocean, with impacts on the hydrodynamics and thermodynamics (Bendtsen et al., 2015; Chauché et al., 2014). Therefore, there are significant justifications to investigate water quality properties in glacial bays and fjords and to track their variability in order to potentially predict future changes.
While the majority of studies examining the influence of glacial meltwater on the marine ecosystem have been performed in the Northern Hemisphere, the importance of the effect of glacial meltwater for the functioning of coastal Antarctic waters has long been hypothesized. Nevertheless, widely available data that describe water quality in glacial bays beyond seasonal timescales at high sampling resolutions and that examine multiple variables remain non-existent. In fact, such datasets are scarce for the Arctic and Alaska as well.
To address this deficiency, an intricate investigation campaign was designed with the intention of comprehensively observing the seasonal oscillations and long-term trends in water quality variability in Admiralty Bay (AB), King George Island, Western Antarctica. The goal of this project was to widen the scope of previously gathered observations by expanding the overall duration of monitoring, increasing the frequency and number of measured parameters, and to acquire data across all seasons of the year.
AB is a 177.04 km
Map of Admiralty Bay showing the measurement points in the following three distinct zones: the main body of Admiralty Bay (pink), glacial coves (blue inset boxes) and the Ezcurra Inlet (yellow lines based upon Gerrish et al., 2021). Bright yellow denotes the current position of the ice–water coastline, and the bright blue insets show the position of the coastline on 10 March 2018. (Sourced from Sentinel imagery, 29 December 2021.)
The dataset as a whole was split into three different zones within the AB,
identified based on distinct seawater properties and proximity to both
glacial fronts and the mouth of the bay (i.e. proximity to open ocean source
waters). These include the following:
These areas are shown in Fig. 1 and are used as separate, although deeply
interrelated, regions for further study. To that end, measurement points
were chosen, and their locations are marked on the map in Fig. 1; their
details (location, depth, number of measurements performed at a given point,
and, in the case of glacial cove points, distance from the water–ice
boundary) are summarized in Table 1.
Details of the measurement sites. The depth measurements are based
on the YSI EXO sonde depth sensor; depths
Measurements in the glacial coves and Admiralty Bay were taken from December 2018 until January 2022, whereas measurements in Ezcurra Inlet took place from October 2019 until January 2022.
Due to the proximity to glaciers and the harsh Antarctic weather, sampling in this region was especially strenuous. Each measurement campaign lasted only a few hours and was performed from the decks of small Zodiac boats (Fig. 2) that provided little comfort to the crew. Moreover, getting to the assigned sites often involved manoeuvring through moving ice packs and bits of icebergs coming from calving glaciers. Sampling during the winter months required working in the dark, in extremely cold temperatures and with continuous contact to freezing water.
The images in the foreground (sourced from
Measurements were performed with two professional YSI multiparameter EXO sondes (EXO1 and EXO2); these instruments have been designed for simultaneous investigation of multiple water quality properties and have also been used and tested by researchers worldwide (Snazelle, 2015). EXO1 consists of five sensor ports, and EXO2 contains seven ports; therefore, the water properties measured varied between the particular campaigns. Of the 3045 measurements collected, 2069 were acquired using the EXO1 sonde, and the remaining 976 were acquired with EXO2 and its larger sensor capacity (details seen in Fig. 2).
The list of the sensors and the properties investigated by each are summarized in Table 2. Some hydrographic properties are derived from direct sensor measurements (e.g. turbidity from light scatter). In these cases, the sondes automatically calculated the additional related values based on universally accepted formulas (Table 2).
List of sensors and measured water properties (based upon YSI Inc, 2017).
The abbreviations used in the table are as follows: nLF – non-linear function, PSU – practical salinity units, fDOM – fluorescent dissolved organic matter, FNU – formazin nephelometric units, QSU – quinine sulfate units, RFU – relative fluorescence units, Chl – chlorophyll, BGA – blue-green algae and PE – phycoerythrin.
Measurements were conducted from the deck of a Zodiac boat (Fig. 2). When the boat was at the designated point, the sonde was lowered by the cable from the reel to a maximum depth of 100 m. At sites with depth of less than 100 m (see Table 1 for information on sites' depths), the measurements were performed throughout the whole water column (until sea bottom was reached). At sites where the depth surpassed 100 m, data were only collected from the top 100 m; this is a limitation of this study, as data were not obtained from bottom portions of the water column. The sampling rate of the sondes was initially 0.2 Hz until 30 December 2019; after 30 December 2019, the sampling frequency increased to 1 Hz.
The intended descent rate of the instrument was 1 m s
Other obstacles were caused by challenging weather and sea conditions. Waves and surface currents often considerably influenced the position of the boat, making it impossible to remain stationed at the assigned site location for the duration the cast. This can be seen from the position data recorded via handheld GPS during sensor deployment and included within the data file. Currents below the surface moved the sonde and cable horizontally from the initial cast position by an unknown extent.
On numerous occasions, ice prevented scientists from reaching specific sites. This was frequently the case in areas close to glacial fronts, most notably when the water surface froze during the winter months and when glacial calving increased in the summer.
All of the sensors were calibrated in accordance with guidelines found in the YSI EXO manual (YSI Inc, 2017), and they were replaced after the appropriate time or when malfunctions occurred that could not be otherwise resolved. The depth and level sensor was calibrated at the start of every survey day.
Measured data were initially recorded in the YSI proprietary format in software
embedded into all of the sensors. At this stage, data went through real-time
data filtering using a basic rolling filter as well as adaptive filtering and
outlier rejection with the default manufacturer settings (for details, see
YSI Inc, 2017). Gathered data were later downloaded using KorEXO software
and exported to MATLAB where some outliers and extreme values were
extracted due to one of the following reasons:
notes from the measurement crew that indicated malfunctions or some other issues; the sonde showed unrealistic values from all of the sensors after reaching the bottom (caused by the
contact with seafloor) at sites with depths of less than 100 m; the presence of other extreme values and outliers, which were scrutinized individually, such as continuous abnormal values from a particular sensor during a measurement day
(indicating sensor malfunction or decalibration) or
incidental extreme values recorded within otherwise reasonable datasets (indicating momentary disturbances).
Despite this series of steps, the whole dataset did not go through any formalized quality assessment/quality check procedure.
Number of measurements taken at the designated sites
Number of measurement days per month
Optical sensors for total algae and fluorescent dissolved organic matter (fDOM) showed unrealistic negative values
(77.82 % of chlorophyll
The turbidity sensor also showed negative values (19.56 % of the readings), but it was calibrated using a two-point procedure with an appropriate standard, and its FNU values have been confirmed in Admiralty Bay waters via the laboratory procedure explained in detail by Wojcik-Długoborska et al. (2022).
Vertical variability in measured properties divided by zone and season.
Mean values of measured properties dependent on season for the whole water column and the top 5 m of surface water.
Box plot of monthly properties' mean values (excluding properties measured solely by EXO2 sonde devices due to their significantly shorter time series).
The results of the measurement campaign discussed above consist of a large and complex dataset describing the variability in the physical, chemical and biological properties in glacially influenced bays. Figure 3 presents a summary of the total number of investigations performed. This shows that, even at the sites sampled the least, it was possible to gather data during all seasons. However, most studies were performed during summer across all zones, while the fewest measurements were collected in winter. Interestingly, despite the unpredictable conditions in the glacial coves, the number of surveys at each site fluctuates at around 100 per location (average of 98.2 measurements per site), which is promising for future statistical analysis.
Considering the complete duration of the projects (see Fig. 4), it is noticeable that the number of measurement days fluctuated, with increases during the warmer seasons when there was a maximum of 7 measurement days per month. In Fig. 4b, we observe that the same tendencies apply to all of the zones, and none of them have been more frequently investigated to any degree of significance. The average number of measurements per month was 3.74 in the glacial coves and 2.91 in Admiralty Bay, with the same number of successful measurement days (111) throughout the whole duration of the project, and 2.42 for Ezcurra Inlet over 92 measurement days.
The division of sites into three zones shows how proximity to glacial fronts and open ocean waters alters particular water quality properties. This effect is also notably correlated with seasonal shifts (Figs. 5, 6). In Fig. 5, the vertical distribution of all of the gathered data is presented. It is apparent that temperature, pH, optical dissolved oxygen (ODO), fDOM and phytoplankton pigment values are especially prone to change due to seasonal shifts, whereas salinity and turbidity values remain similar throughout the year. However, Fig. 6 provides a detailed illustration of how different properties vary in surface layers in contrast with the whole water column (limited to 100 m depth), most notably with respect to salinity and turbidity values, although it applies to all measured properties except for pH. This shows the impact of both atmospheric forcing and glacial outflow, which, based on buoyant plume model theory (Kimura et al., 2014; Mankoff et al., 2016; Jenkins, 2011) and observations (Chauché et al., 2014; Osińska et al., 2021), is mainly contained in the top layer of the ocean. Therefore, the results provide information on seasonal changes in water properties and glacier–ocean interactions and can be used for the validation of previously formulated methods of GMW tracking.
The 38-month-long duration of the project allowed for the tracking of
seasonal variability across all measured hydrographic properties and showed
consistency in all cases (Fig. 7). Moreover, this duration permits cautious
predictions regarding long-term shifts in water column properties and
reveals the impact of climate change or other influential conditions in this
region. Using more sophisticated techniques, it is possible to more
precisely determine the nature of this variability. The quantities of
chlorophyll
The described dataset is freely accessible at the PANGAEA repository:
The assembled dataset shared here presents an opportunity to develop a better understanding of Admiralty Bay water characteristics over the 38-month survey period and can be used in further studies exploring the nature of and changes in glacially influenced regions in general. The sheer magnitude of this investigation, with 3045 separate measurements acquired on 142 different days, validates its importance and inspires optimism regarding future work and the application of these data.
The scope of the measured parameters (thermodynamic, physical, chemical and biological) paints a wide and precise picture of AB hydrographic variability during all months of the year and may allow for a multidisciplinary analysis of the complex processes that take place at this location. The varied settings of study sites allow for the tracking and identification of GMW and other water masses (Straneo et al., 2011; Chauché et al., 2014). Additionally, this sizable dataset can be used as a tool for better understanding the general hydrodynamics and thermodynamics of glacial bays and fjords and may be employed for the validation of coupled glacier–ocean modelling (Cowton et al., 2015; De Andrés et al., 2021; Bertino and Holland, 2017).
MO conceptualized the study; curated the data; undertook the formal analysis, investigation and validation; developed the methodology; created figures; and prepared the original draft of the paper. KAW was responsible for carrying out the investigation, developing the methodology, and reviewing and editing the manuscript. RJB acquired funding, carried out the investigation, was responsible for project administration, acquired resources, supervised the study, and reviewed and edited the manuscript.
The contact author has declared that none of the authors has any competing interests.
Publisher's note: Copernicus Publications remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Calculations were made possible by software provided by CI TASK (Centrum Informatyczne TASK) in Gdańsk. The authors also wish to acknowledge the invincible members of the so-called “MorMon” team, part of Polish Antarctic Station's crew, who throughout the whole period of the project, often in trying and almost always in uncomfortable conditions, carried out the measurements presented in this work.
This work was supported by the National Science Centre, Poland (grant no. UMO-2017/25/B/ST10/02092; Quantitative assessment of sediment transport from glaciers of South Shetland Islands on the basis of selected remote sensing methods).
This paper was edited by Salvatore Marullo and reviewed by Mattias Cape, Luca Fiorani, and one anonymous referee.