Microbial and nutrient enrichments in Brine and Nilas during the first 24 h of open water lead refreezing

Abstract

Open water leads in the sea ice covering the Polar regions are becoming more frequent and cover larger areas as a result of the warming and thinning of the ice. Refreezing of the leads advances the occurrence of young and newly formed sea ice (nilas), which has a thickness of < 10 cm and is characterized by a highly saline surface brine. Due to the growth of fragile frost flowers exposed at its surface, nilas can become a significant source of bromine, mercury, sea salt aerosols, bioaerosols, and ice nucleating particles for the atmosphere. Here we report the results of a study where we investigated the initial formation phase of a 3–6 mm–thick nilas in cleared sea ice ponds. Samples for quantifications of macronutrients, microorganisms, Chlorophyll a (Chl a) and photobiology were collected from seawater, nilas, and brine 5 and 24 h, after the experiment commenced. The concentration changes between constituents were scaled relative to changes in salinity with the enrichment index (I). Nilas was enriched in microalgae and Chl a and the brine was highly enriched in bacteria, virus like particles (VLPs) and algae after only 5 h. We did not find a significant further enrichment between 5 and 24 h of nilas formation. This indicates that the initial formation phase of the nilas is an important step in distributing nutrients and microorganisms in nilas and brine. Only particulate matter was enriched in nilas and brine and none of the macronutrients. Photosynthetic microalgae were strongly impaired in the nilas and the brine, where no fluorescence signal could be detected supposedly due to a combination of excessive light at the surface and high brine salinity. The reasons for the observed differences in enrichments are evaluated and discussed including concentration of particulate matter, evaporation of surface brine, microbial growth and concentration of microorganisms in the sea surface microlayer (SML). Further studies are needed to decipher which of the proposed mechanisms were predominantly responsible for the observed enrichments as well as how these enrichments affect firstly the frost flowers and secondly the atmosphere.

Supplementary Information

The online version contains supplementary material available at 10.1038/s41598-025-15516-5.

Introduction

Sea ice plays a crucial role as a physical entity in both the Arctic and Antarctic regions, significantly impacting the global climate due to its high albedo, which reflects radiation back into space¹. This reflective capability is additionally enhanced by the snow cover on top of the sea ice². Sea ice further provides a low temperature and highly saline habitat for a diverse community of microorganisms such as viruses³ fungi⁴ bacteria⁵ and microalgae⁶. We refer to⁷ for a comprehensive review of living conditions for microorganisms in sea ice. The microorganisms along with inorganic solutes and solids are incorporated into the ice during ice growth and their concentrations often exceed those in the underlying water column⁸. Sea ice microorganisms play an important role in biogeochemical cycles, with e.g. sea ice microalgae contributing 10–20% to the total marine primary production in the Arctic^9,10which serves as the main carbon source in the arctic food–web during the late winter/early spring transition⁶. However, the Arctic sea ice ecosystem is rapidly changing as demonstrated by a reduction of 48% in minimum sea ice extent (i.e. during September) from 7.7 × 10⁶ km² in 1979 to 4.0 × 10⁶ km² in 2020 and, predicted that the minimum sea ice extent could approach zero within a few decades^11–13. The average thickness of the Arctic Ocean sea ice has also decreased by about 1 m, from 2.7 m (1990–2006) to an average of 1.7 m (2007–2019), as the first–year ice (FYI) is increasingly replacing the thicker multi–year ice (MYI)^14,15. This shift in thickness promotes longer periods with open water and leads in the ice as well as growth of new and young sea ice such as nilas, which develops during quiescent conditions¹⁶. Nilas is the initial stage in the continuum of freezing ocean water, and starts forming via frazil ice and grease ice^17,18. With further ice growth, nilas develops into young sea ice roughly defined as sea ice thicker than 10 cm¹⁷whereas the thickness of nilas can vary from a few millimeters to 10 cm¹⁸. During nilas desalination, highly saline brine is rejected into the water column below largely by gravity drainage and onto the surface of the nilas by brine expulsion^17,19. The brine on the surface of the nilas, termed brine skim²⁰contains nutrients and microorganisms in significantly higher concentrations compared to underlying seawater^21,22usually quantified by the enrichment index, which scales concentration differences to changes in salinity^8,23. The significance of the brine skim is highlighted by its ability to enrich microorganisms and substances (e.g., nutrients), which can become airborne through two mechanisms, thereby affecting atmospheric processes. First, compounds and particles from brine skim can be incorporated into the overlying snowpack and subsequently transported into the lower atmosphere by wind events²⁴. Second, in the absence of a snowpack and under the right conditions – such as cold, moist and calm weather – fragile frost flowers can develop on the surface of the sea ice²². Frost flowers may also serve as a source of atmospheric compounds and particles, such as microorganisms, ice nucleating particles, and bromine, which have known climate impacts^25–29. The so–called “wet frost flowers”, are covered by the brine skim through the process of wicking due to capillary action^25,30. Nilas is the precursor of the brine skim, and fully developed nilas and its biogeochemical properties have been studied for several decades^8,31. Contrary to the importance of brine skim as a source of biochemical compounds for the lower troposphere, there are only a few studies of the initial developmental phase of nilas and brine skim, which are mainly laboratory based^18,32. Here, we report the results of an in situ 24 h study where nilas with a brine skim layer developed on the seawater surface. The experiment was carried out to resemble the development of nilas in newly opened and refreezing leads in the sea ice¹⁶rather than the autumn freeze–up with no initial ice⁶. We observed significant differences in microbial, nutrient, and Chlorophyll a (Chl a) enrichments between nilas and brine skim and strongly impaired microalgae in the brine. The following questions are accordingly addressed: (1) To what extent are nilas and brine skim enriched with microorganisms and nutrients relative to seawater, (2) What biogeochemical mechanisms or processes are responsible for the observed differences in enrichments, (3) Were there changes in enrichments over time for any of the variables, and, (4) What are the reasons for the strong microalgae impairment in the brine skim?

Methods

Field site, sampling, and experimental setup

The study site is Malene Bight (64.179 N, 51.706 W) near Nuuk, southwest Greenland (Fig. 1A). First year sea ice covers the bight from about start–February and lasts until about mid–April with a thickness of 30–40 cm^33,34. Tidal range in Malene Bight is 1–5 m depending on the phase of the tide³⁵. In this experiment the snow cover was removed manually from the ice in a rectangle of 13 × 4 m in which 6 ponds (1 × 1 m) of open water were created by removing the ice at midday on 9 March 2022 to initiate the experiment (Fig. 1B). Bulk water samples were collected at the start of the experiment from a depth of 0.5 m in three of the ponds by submerging sterile blue cap flasks, opening them at the target depth, and sealing them. After approximately 5 h, a thin (3–6 mm) layer of nilas had formed, along with a layer of brine skim on top. The brine skim, termed “brine” in the following for brevity, was collected using sterile Corning Falcon transfer pipettes, into sterile tubes. There were no visible impurities or ice crystals in the collected brine samples. The nilas ice was carefully broken into pieces by hand while wearing sterile gloves, ensuring avoidance of brine and water film on the bottom of the nilas. Samples of nilas were collected in sterile bags. To investigate the effect of time on the development of the brine three of the remaining ponds were left undisturbed overnight for 24 h. Upon return, we collected the brine from the remaining ponds. All samples were placed in cooling boxes at dark conditions and brought to the laboratory within 30 min. The collected nilas ice was thawed at 4 °C in sterile bags before transferring samples to sterile falcon tubes.

Fig. 1.

Experimental site in Malene Bight, Nuuk, Greenland (A) with snow cleared area with the six 1 × 1 m ponds (B). The map (A) was created in Adobe Creative Suite 6 (CS6), https://www.adobe.com.

Air temperature, PAR, salinity, Chl a and nutrients

Continuous recordings of photosynthetic active radiation (PAR), air temperature, and relative humidity were obtained from a meteorological station in Nuuk (https://www.asiaq.gl/en/). Salinity was determined on thawed nilas samples using an YSI Pro30 CTD, which was also applied for measuring salinity in seawater under the ice. Brine salinities (small volumes) were measured with a calibrated digital refractometer PAL–06 S ATAGO. Samples for Chl a and nutrients were determined using standard methods described in detail³³. In brief, for determining Chl a concentrations, a known volume was filtered onto 25–mm Whatman GF/F filters, which were then extracted in the dark at room temperature for 18 h in 10 ml 96% ethanol and analyzed fluorometrically (TD–700, Turner Designs) before and after addition of 200 µl of a 1 M HCl solution. The fluorometer was calibrated against a pure Chl a standard (Turner Designs). Melted sea ice and under–ice seawater samples were filtered through 0.45 μm Q–Max CA syringe filters Filtrates were stored frozen (− 20 °C) until analysis on a Seal QuAAtro autoanalyzer using standard colorimetric methods to obtain concentrations of ammonium (NH₄⁺), nitrate (NO₃^–), nitrite (NO₂^–), phosphate (PO₄^3–) and silicate (Si(OH)₄). Mean values of triplicate measurements are reported where NO_x is the sum of NO₃^– + NO₂^– concentrations.

Ice microalgae photobiology

Photobiological parameters were derived from variable chlorophyll fluorescence using a Walz Phyto–PAM (Pulse Amplitude Modulated) variable fluorometer fitted with a Phyto–ED emitter–detector unit (Heinz Walz GmbH, Effeltrich, Germany) as described previously³⁴. The Phyto–PAM instrument was operated in a darkened and cooled laboratory, and thawed nilas samples with ice microalgae were dark–acclimated (12 h) before measurements. The 12 h is the time it approximately takes for ice samples to thaw in the dark at 4 °C³⁶ in accordance with the 8 h recommended for algae to be total dark-acclimated³⁷. Samples were corrected for background F_o by filtering each sample through a syringe filter (Millex MCE 0.22 µM) and measure the background fluorescence, which was automatically set as background for the unfiltered samples in the instrument³⁴. The instrument provides Φ_{PSII_max} (the maximum dark-acclimated quantum yield of Photosystem II), and its Rapid Light Curve (RLC) protocol models the standard light response cardinal points using the Phyto-PAM software based on the Jassby & Platt et al. (1976) model³⁸.

Flow cytometry

To prepare samples for flow cytometry analysis, we concentrated 500 mL of the environmental sample using 0.2 μm Millipore^® SterivexTM filter units. The concentrated material was subsequently eluted into 2 mL of filtered sample water. Due to the limited availability of brine, it was collected directly into sterile Falcon tubes and analyzed without further concentration. All samples were fixed in 0.2 μm filter sterilized gluteraldehyde (0.5% final concentration, incubated at 4 °C 20 min. For quantification of bacterial cells and virus–like particles (VLPs), half of the sample was stained with SYBR GREEN (10,000 ×, Sigma–Aldrich) at a 5 × final concentration to enable nucleic acid detection. For quantification of chlorophyll–containing cells (microalgae and cyanobacteria), the remaining half of the sample was left unstained to analyze autofluorescence associated with natural pigments present in the chlorophyll–containing cells. Filtered (0.2 μm, PES) seawater and MQ water served as a negative control for both stained and unstained conditions. Flow cytometric analysis was conducted using a NovoCyte flow cytometer (Agilent), which has a scatter resolution of 0.2 μm and can detect particles ranging in size from 0.2 to 50 μm. The 488 nm blue laser with a 530/30 nm bandpass filter was used to detect SYBR GREEN fluorescence, and the 640 nm red laser with a 675/30 nm bandpass filter was used to detect Chl a fluorescence. The detection threshold was set to 700 on the 530/30 nm fluorescence signal (stained samples) or to 700 on the 675/30 nm fluorescence signal (unstained samples). For bacterial and VLP analysis, 50 µL of sample or 50,000 events were processed, while for Chl a containing cells, 100 µL or 50,000 events were analyzed. Data analysis and gating were performed using FlowJo™ v10.8 Software (BD Life Sciences). The gating strategy for bacterial cells and VLPs followed the protocol described by Haynes et al.³⁹. However, due to the low number of events, the gating protocol for microalgae was modified. Specifically, all events which surpassed 10³ on the Chl a axis were considered to be Chl a positive events, as all negative control events fell below this threshold (Supplementary Figure. S1). All Chl a positive events had a low forward and side scatter. Forward scatter is associated with cell size and side scatter to internal complexity of the cells. Based on this, we propose that the Chl a positive events were primarily reflecting the presence of cyanobacteria.

Microalgae enumeration

For algal abundance measurements, samples of melted nilas and under–ice seawater were fixed with acidic Lugol’s solution (final concentration of 1%) and stored dark and cold (3 ± 1 °C). Algal cells were identified and counted from light–microscopy analyses using the Utermöhl sedimentation technique⁴⁰. Counting chambers with a volume of 10 ml were used and the counts were carried out using 100–400 × magnification applying appropriate subsampling units using an ocular counting grid. The following subsampling units were used: window, diagonal and total chamber. A total number of cells of > 400 per sample were counted. Microalgae cells were analyzed within 2 months of collection, average values of duplicate samples are reported and data are located in Supplementary Table S1.

Enrichment index (I)

The enrichment index (I) was calculated following Eq. 1.

1

where S_B is salinity of seawater and S_A salinity of the ice, and C_A the concentration of a substance in nilas or brine, and C_B the concentration of the substance in the seawater below the ice⁸. The index expresses the increase in concentrations of a variable in spite of any losses related to desalination where I < 1, I = 1, or > I correspond to a depletion, no enrichment, or enriched relative to bulk salinity, respectively^23,31.

Data analysis and statistics

Due to high heterogeneity of variances in the raw data, non–parametric Kruskal–Wallis tests were performed to assess any significant differences in concentrations of variables followed by Dunn’s post hoc comparisons to recognize which variables were different (p < 0.05) in each of the segments: seawater, nilas, brine 5 h, and brine 24 h. Enrichment index data displayed normal distributions and homogeneity of variances, so were tested with one–way ANOVA and post hoc Tukey tests. All statistical analysis was performed using JASP Version 0.11.1 (The JASP Team, University of Amsterdam, The Netherlands).

Results

PAR and air temperatures

The experimental period was 9th–10th March during a period of cold and clear weather with air temperatures between − 10 and − 13 ˚C, and maximum PAR recordings of 300 µmol photons m^{– 2} s^–1 on the 9th at noon and 600 µmol photons m^–2 s^–1 on the 10th at noon (Fig. 2).

Fig. 2.

PAR and air temperature between 5th and 12th March in Nuuk with the experimental period marked in red.

Salinity, nutrients, microorganisms and Chl a in seawater, Nilas, and Brine

Salinities in the brine sampled after 5 h (brine 5 h) and after 24 h (brine 24 h) were both significantly higher compared to seawater and nilas (Fig. 3A). Ammonium concentrations were similar in seawater and brine 5 h, but significantly lower in the nilas. Nitrate and nitrite (NO_x) were significantly lower in nilas compared to seawater and significantly higher in both brine 5 h and brine 24 h with a significant higher concentration in the brine 24 h (Fig. 3B). Phosphate concentrations were low in seawater and nilas and significantly higher in both brine 5 h and 24 h. Silicate was significantly higher in both brine 5 h and brine 24 h compared to seawater and nilas, and especially in the 24 h brine. In general, nilas nutrient concentrations were all lower compared to seawater and significantly higher in both brine 5 h and 24 h, but with no differences in concentrations between 5 h and 24 h except for silicate and NO_x (Fig. 3B). Microalgae, VLPs and bacteria were all present at significantly higher concentrations in both brine 5 h and brine 24 h compared to both nilas and seawater, and with no significant differences in concentrations between brine 5 h and brine 24 h (Fig. 3C). Chl a concentrations were significantly higher in the nilas, brine 5 h and brine 24 h as compared to seawater (Fig. 3D). The abundance of dinoflagellates was similar in nilas (2.7 × 10³ cells L^–1) and under ice (3.0 × 10³ cells L^–1) but diatoms were about 2 times higher in nilas (3.3 × 10⁴ cells L^{– 1}) compared to seawater (1.9 × 10⁴ cells L^{– 1}) though not statistically significant. Diatom relative abundances reached 92% in the nilas and 86% in the seawater identified to genera though with a high number (1.55 × 10³ cells L^{– 1}) of unidentified flagellates in the nilas and similarly in the seawater samples (1.54 × 10³ cells L^–1). Dinoflagellates were dominated by both athekate and thekate species, whereas diatom species composition was dominated (40%) by small cell–sized (15–20 μm) Thalassiosira spp., Cylindrotheca closterium/longissima (6%), and Bacillariophyceae (9%). Brine samples were not analyzed for species composition due to limited sample volumes.

Fig. 3.

Concentrations (n = 3) ± s.d. for salinity (A), of ammonium, nitrite + nitrate (NO_x), phosphate, and silicate (B), Bacteria, VLP, and microalgae where “right” refers to right y–axis scale (C), and Chl a (D) in seawater, nilas, brine 5 h, and brine 24 h. Letters indicate statistical significant (p < 0.05) differences in concentrations when letters differ between sampling sites for a parameter as given by the post hoc analyses.

Enrichment index (I)

Nilas enrichment index (I) showed that nilas was not enriched except for microalgal cells (2.48) and Chl a (2.97), but the low enrichment indices may be within the measurement error (Fig. 4). NO_x and PO₄ were not enriched in the brine 5 h and were significantly lower in the nilas (Table 1). The brine 5 h was, in comparison, highly enriched with microorganisms with enrichment indices of 15.5 (bacteria), 38.2 (VLP), and 27.4 (microalgae). Microorganisms had significantly lower enrichment indices in nilas compared to the brine. There was no Chl a enrichment in the brines with I = 0.63 (5 h) and I = 1.0 (24 h) given by the higher brine salinities compared to seawater. Brine 24 h enrichment indices were just slightly higher for NO_x and PO₄ but with enrichment index < 1. There were no significant differences in enrichment index between brine 5 h and brine 24 h for any of the parameters, which demonstrates that parameters did not change significantly between 5 h and 24 h.

Fig. 4.

Average (n = 3) ± s.d. enrichment index (I) for nitrite + nitrate (NO_x), phosphate, silicate, ammonium, bacteria, VLPs, microalgae, and Chl a, in nilas, brine 5 h, and brine 24 h.

Table 1.

Average (n = 3 ± s.d.) concentrations of salinity, NO_x (nitrate + nitrite), phosphate, silicate, ammonium, bacteria, VLP, and microalgae, Chl a in nilas, Brine 5 h, and Brine 24 h, and enrichment in nilas, Brine 5 h, and Brine 24 h in bold. Nilas, Brine 5 h, and Brine 24 h are scaled to seawater in the enrichment index. Statistically significant differences in mean enrichment index (p < 0.05) between Nilas and Brine 5 h (Brine 5 h) and between Brine 5 h and Brine 24 h (Brine 24 h) are shown with different letter superscripts within rows.

	Concentration				Enrichment index
	Seawater	Nilas	Brine 5 h	Brine 24 h	Nilas	Brine 5 h	Brine 24 h
Salinity	32.6 ± 0.0	21.5± 0.0	72.0 ± 0.1	77.0 ± 1.63	–	–	–
NOx (µmol L–1)	10.14 ± 0.3	2.53± 1.1	18.9 ± 0.46	20.9 ± 4.1	0.38a ± 0.16	0.84b ± 0.02	0.87b ± 0.17
PO4 (µmol L–1)	0.74 ± 0.1	0.33 ± 0.33	1.53 ± 0.01	1.68 ± 0.12	0.67a ± 0.05	0.92b ± 0.01	0.96b ± 0.07
SiO2 (µmol L–1)	5.73 ± 0.2	2.39 ± 1.62	11.7 ± 0.15	22.44 ± 16.4	0.63a ± 0.42	0.93b ± 0.01	1.66b ± 1.24
NH4 (µmol L–1)	4.42 ± 0.1	1.46 ± 0.22	4.21 ± 1.63	3.38 ± 0.59	0.50a ± 0.07	0.43b ± 0.17	0.32b ± 0.06
Bacteria (cells mL–1)	1.89∙104 ±3.01∙103	5.28∙103 ±5.66∙102	6.70∙105 ± 1.17∙105	9.40∙105 ± 2.63∙105	0.42a ± 0.09	15.5b ± 3.74	21.3b ± 8.82
VLP (cells mL–1)	6.14∙103 ± 3.86∙102	3.75∙103 ±4.66∙102	5.11∙105 ± 2.66∙104	6.7∙105 ± 2.29∙105	0.89a ± 0.18	38.2b ± 4.18	44.6b ± 20.8
Micro– algae (cells mL–1)	1.65∙102 ±2.81∙101	2.5∙102 ± 4.63∙101	1.1∙104 ± 5.12∙103	9.14∙103 ± 4.14∙103	2.48a ± 0.7	27.4b ± 12.3	22.4b ± 6.21
Chl a (µg Chl a L–1)	0.09 ± 0.005	0.18 ± 0.00	0.13 ± 0.04	0.22 ± 0.06	2.97a ± 0.15	0.63b ± 0.16	1.08b ± 0.34

Microalgae photobiology

Maximum quantum yield of Photosystem II (Φ_{PSII_max}) of the microalgae was significantly higher in seawater compared to nilas which also applies for α, whereas there were no differences in E_k between seawater and nilas (Fig. 5A–B). Note that no signal could be detected (n.s.d.) for any of the photobiological parameters in either 5–24 h brines.

Fig. 5.

(Φ_{PSII_max}) (A) and α (B) in seawater, nilas, brine 5 h, brine 24 h. n.s.d = no fluorescence signal detected.

Discussion

Nilas and new ice

Nilas is a thin, elastic crust which develops at the water surface in quiescent⁴¹ and cold conditions, forming a continuous layer of initial mm thick ice, reaching up to 10 cm while still being classified as nilas^18,42. The fact that nilas forms as a continuous layer of ice differentiates it from grease and frazil ice that are unconsolidated particles floating on the surface⁴³. In this study, nilas developed on the surface of the cleared ponds and reached a thickness of 3–6 mm after 5 h of exposure during a cold period with temperatures between − 10 ˚C and − 13 ˚C with calm wind conditions. A mm–sized layer of brine developed on the surface of the nilas during its development. The decreased bulk salinity of the nilas as compared to the seawater (21.5 vs. 32.6) (Fig. 3A), demonstrated an active desalination process, likely since the initial formation of the ice. Laboratory studies have shown that desalination of nilas is linear and that bulk salinity reaches a near constant and lowest level after about 8 h¹⁸. Desalination of the ice takes place during ice formation where salts and nutrients are expelled both downwards into the water column and upwards whereby a mm–sized high saline brine layer is formed^17,44. It has been demonstrated that upwards permeabilities of salts in the ice are generally higher than downwards, but both depend on surface temperatures, i.e. the colder the lower the permeabilities³². Present nilas bulk salinity of 21.5 is, however, somewhat higher than the 12.5 given by the experimentally derived desalination rate of 2.1 g kg^{– 1} h^1,18. This might relate to differences in laboratory and in situ experimental conditions as for the higher laboratory temperatures between − 5 ˚C and − 10 ˚C compared to present in situ temperatures between − 10 ˚C and − 13 ˚C. Nilas bulk salinities of 16–18 were reached after 6 h of exposure in a comparable in situ study in young Antarctic sea ice²³, which supports that there are differences between in situ and laboratory experiments. Traces of brine could have endured on the surface of the nilas after sampling of the nilas, although brine was carefully drained from the nilas. Nevertheless, previously reported nilas bulk salinities of 5–8 in nilas in the Greenland Sea⁸ are about 3 times lower than present 21.5. The bulk salinity difference emphasizes that present nilas were formed within 5 h whereas the Greenland Sea samples were collected in September–October with an ice age of at least 0.5–1.0 months. A comparable experiment carried out in Saroma–Ko Lagon showed that young sea ice attained a bulk salinity around 13.4 within 24 h compared to a seawater salinity of 25.0⁴⁵ though with no differentiation between nilas and brine. Anyway, results emphasize time between formation of the nilas and time of sampling as an important parameter in enrichment studies.

Impaired photophysiology of microalgae in Nilas and Brine

The photobiological parameters ΦPSII_max and E_k were all reduced in the nilas compared to the seawater, and no signal could be detected in the brines 5 h and 24 h (Fig. 5A–B). The fact that ΦPSII_max is lower in the nilas is opposite to the general observation of a relatively higher ΦPSII_max at the bottom of the sea ice compared to seawater^36,46. Microalgae fixed to the bottom of the ice will in this extreme low light environment receive a relatively higher number of photons compared to phytoplankton moving vertically up and down along a light gradient, and thereby enhance their photosynthetic potential expressed as a higher ΦPSII_max⁴⁷. In contrast, the microalgae are, in the present study, fixed in a thin nilas layer and fully exposed to the high PAR intensities of 300–600 µmol photons m^–2 s^–1 at the surface (Fig. 2), which will impair their photosynthetic potential in terms on photoinhibition^37,48. The low nilas ΦPSII_max compared to seawater shows that newly incorporated microalgae in the nilas have not fully acclimated to the higher light intensities experienced at a fixed position in the nilas at the water surface. The microalgae which were present just before the start of the experiment were adapted to under ice PAR intensities of 2.20 µmol photons m^–2 s^–1 on 10 March at noon³⁴ as compared to PAR noon surface intensities of 300–600 µmol photons m^–2 s^–1. Pigments and photoprotective compounds were not measured in this study, but a comparable study³⁶ in Kangerlussuaq (west Greenland) showed that 4 to 6 days were required to develop photoprotective pigments (diadinoxanthin and diatoxanthin) and mycosporine–amino acids (MAAs) following a sudden high light exposure. There was no detectable variable chlorophyll fluorescence signal in either brine 5 h or brine 24 h, where microalgae were exposed to high salinities of 72–77 and potential osmotic stress in the brine as compared to the seawater with a salinity of 32.6. Ice microalgae are generally adapted to high (50–70) salinities in the brine⁶. The dominant diatom species that we observed in the nilas and was also dominant in seawater was Thalassiosira spp., which has been found in sea ice by several studies in Greenland^34,49. In a comparable 24 h ice pond study there was a high abundance of Navicula transitans (33%) and Achnanthes taeniata (12%) in the young sea ice, which also originated from the seawater⁴⁵. However, the impaired photophysiology of microalgae was most likely attributed to a combination of a possible lack of photoprotective pigments and compounds, along with the inability of microalgae to rapidly adapt to increased salinities in the brine.

Enrichment of the Nilas – high in microalgae biomass and low in nutrients

There was an enrichment of the nilas with microalgae cells (2.48) and Chl a (2.97) (Fig. 4). Chl a enrichment in nilas and new ice has also been observed in other studies with enrichment indices of up to 20 for Chl a and 4.6–589 for algal cells^8,31. Laboratory studies showed a comparable Chl a enrichment of up to 53⁴¹. There was no enrichment in the nilas for any of the nutrients NO_x, phosphate, silicate, and ammonium with an average enrichment of 0.55, and also no enrichments of bacteria and VLPs. This strong contrast in enrichments in nilas between phytoplankton and nutrients such as dissolved substances, was also observed in another comparable study of newly formed sea ice²³. Although cyanobacterial and algal cells are relatively small (< 1–30 μm)³⁸ compared to the larger brine channels, which range from several µm to mm in size⁵¹ and would therefore be readily expelled from the ice with the brine along with dissolved nutrients and salts¹⁷, they can become trapped in the extracellular polymeric substances (EPS), that are abundant within these channels⁵². EPS is a range of compounds produced by the microalgae for cryoprotective reasons which forms a sticky matrix^31,52 and EPS is known to bind microalgal cells and helps them attach to the rising frazil ice crystals⁵³ which is later incorporated into the growing ice⁸. For microalgae with EPS attached to cavities inside the brine channels there is a higher chance of remaining there, and not being discharged out of the ice with the brine expulsion. That the nilas is enriched in both algal cells and Chl a relates to the fact that microalgae contain Chl a which they utilize for light harvesting⁵⁵. However, we observed a discrepancy between the enrichment of microalgae and Chl a in the nilas. This may be attributed to technical reasons. Using flow cytometry, we primarily quantified prokaryotic microalgae, which are a minor contributor to the Chl a pool due to their small sizes. The relatively larger microalgae (e.g., Thalassiosira, Nitzschia), which we observed using microscopy, were present at too low abundances to be accurately quantified by flow cytometry. These larger microalgae contribute significantly to the total Chl a pool due to their considerably larger sizes that are between 1 and 2 orders of magnitude bigger than sizes of marine cyanobacteria. It could be suggested that the nutrient enrichments below one unit were related to unusual conditions, if low nutrient concentrations were observed in the water below the ice. The under–ice nutrient concentrations are however very similar to those measured just below the ice in the nearby Kobbefjord at same time of year, i.e. NO_x (9–10 µmol L^–1), phosphate (0.7 µmol L^–1) and silicate (4–5 µmol L^{– 1})⁵⁶, which indicates that the experiment was carried out during representative conditions and that the establishment of the ponds by removing snow and ice did not affect these conditions. A comparable study from Saroma–Ko Lagon showed similar to the present study also a decrease in concentrations of PO₄, NO₃, NO₂, and SiO₂ in young ice developed during 24 h as compared to seawater concentrations⁴⁵. In contrast to these and our results (Fig. 4), other studies found increased nutrient enrichments in nilas, e.g. 1.5–3.8 (PO₄), 3.8 (NO₃), 2.9 (NO₂), and 1.5 (SiO₂) in new, young sea ice at the Mackenzie shelf³¹ and in the Greenland Sea⁸. The Mackenzie shelf study was carried out on ice with a very similar thickness (0.1–3.0 cm) as the present study, which indicates that the initial enrichment of nutrients may be subject to environmental conditions. As for the Greenland Sea⁸ with an ice thickness in the < 22 cm category, age might have played a role for the observed increased nutrient concentrations. An older age gives more time for processes, like desalination to act on the ice. Additionally, the ice could have been flooded by saline and nutrient rich water in periods with negative freeboard due to snow cover²³. All nutrients and salinity were nearly halved in the nilas in our study, which we attribute to the process of desalination¹⁷. The absence of bacterial enrichment of our nilas also contrasts observations from the Greenland Sea⁸ and the Mackenzie shelf³¹ with enrichments of 7 and 5.1, respectively. It is supposedly not related to differences in bacterial concentrations in the seawater from where the nilas developed. Present bacterial concentrations of 1.89 × 10⁴ cells mL^–1 are just slightly lower compared to the Mackenzie shelf (8.0 × 10⁵ cells mL^–1)³¹ and Young Sound, Greenland (2 × 10⁵ cells ml^{– 1})²², and (5.0 × 10⁵ cells mL^–1) in the high Arctic (81° N)⁵⁷. Bacteria were not enriched in the nilas, suggesting that they were not incorporated together with the microalgal cells. As discussed earlier, this discrepancy may be due to differences in methodology. In our study, we used flow cytometry for bacterial quantification, which detects free–living, suspended bacterial cells, as well as bacterial aggregates, which are both counted as single particles despite the fact that aggregates consist of multiple cells. In contrast, Gradinger & Ikävalko (2008)⁸ employed microscopy, which can resolve individual bacterial cells even when they are aggregated or embedded in EPS. This could account for the absence of bacterial enrichment in our study compared to their findings. Additionally, in Gradinger & Ikävalko’s study, bacterial counts were conducted simultaneously with flagellate counts, potentially introducing variability or affecting the specificity of bacterial quantification. These methodological differences likely contribute to the observed discrepancy in bacterial numbers. Epiphytic and endophytic bacterial cells are tightly associated with microalgal cells⁵⁸ and would therefore not be detected by flow cytometry, which analyzes bulk properties of single particles. There was also no VLP enrichment of nilas, which contrasts other Antarctic studies which showed higher concentrations in the ice (4.5 × 10⁶ cells ml^–1) compared to seawater below (1.7 × 10⁶ cells ml^–1)⁵⁹. Present VLP concentrations in the seawater of 6.14 × 10³ cells mL^–1 were quite low compared to other Arctic studies (1.1 × 10⁶ cells mL^–1)⁶⁰ and Young Sound (0.5 × 10⁵ VLP mL^–1)²². Note that the difference in VLP concentrations could also relate to differences in methodology between present flow cytometer and the studies based on microscopy²². There are two possible explanations why VLP and bacteria were not incorporated into the nilas as for the reasons of particle size and sampling of nilas. A study has recently shown that larger particles/cells as microalgae are easier incorporated into nilas as compared to the much smaller bacteria and VLPs²³ which aligns with present observations. The second reason relates to the sampling if nilas was sampled with no differentiation between the nilas ice and the brine layer and sampled as one unit, which is most likely the case in the other studies mentioned above.

Enrichment of the Brines – high in microbial organisms and low in nutrients

The salinity (72–77) of the brines (Fig. 3A) is comparable with other studies such as: 90–100⁶¹, 80–100²¹, 70–100²². The enrichment of the 5 h brine was statistically significantly different from the nilas and was higher for nutrients such as NO_x, ammonium, silicate, and phosphate although their enrichment index was < 1. This contrasts with microorganisms, which showed high enrichment indexes, particularly bacteria with an index of 15.5. The bacterial enrichment in this case is about 3 times higher than the 5.1 reported on the Mackenzie Shelf³¹ and 7 in the Greenland Sea⁸. However, the brine is not always enriched, as observed in Young Sound, where the enrichment index is approximately 0.5²². The present VLP enrichment (38.2) contrasts with the findings of Young Sound²² study which showed no enrichment of VLPs in the newly formed brine²². In that study, the brine had lower VLP concentrations (0.5 × 10⁵ cells mL^{– 1}) compared to seawater (1.2 × 10⁵ cells mL^{– 1}). A wide range of microalgae genera–specific enrichments (7–589) in nilas have been reported⁸ which aligns with current observations of significant microalgae enrichment of the brine 5 h. Between the brine 5 h and brine 24 h, there were slight increases in nutrient and microorganism concentrations, except for ammonium. However, there were no significant differences in the enrichment index for bacteria, which align with the results in Young Sound where concentrations remained stable after 24 h²².

Differential enrichments of Nilas and Brine

The main drivers of the enrichment mechanisms that we propose for the nilas and the brine are shown in the conceptual model (Fig. 6).

Fig. 6.

Conceptual model of main drivers in enrichments of nilas and brine.

Frazil ice crystals with microalgae attached via EPS, rise to the surface at the initial state of sea ice development. The microalgae containing Chl a are here incorporated into the nilas which explains the enrichments of the nilas by microalgae and Chl a. As nilas forms, desalination occurs, expelling brine with dissolved substances such as nutrients, resulting in low nutrient concentrations and enrichment indexes in the nilas. The enrichment index of nutrients for the brine 5 h was higher compared to nilas but remained below 1, which implies that the desalination still prevailed in the brine. The most remarkable result is, on the other hand, the high and significant enrichments of microorganisms and the probable reasons for these enrichments are discussed in the following paragraphs.

Assuming a brine volume in the nilas of 5% would give a bacterial concentration of 3.78 × 10⁵ cell ml^{– 1} in the brine expelled out of the ice and equals a bacterial enrichment of 9 in the brine which is somewhat lower compared to the measured bacterial enrichment of 15.5. Enrichment rates of 9–10 were reached for VLPs and microalgae as well, and even if these enrichments were lower it is very likely that up concentration in the brine expelled out in the surface of the nilas, might have influenced the high enrichments. The calculation is based on the assumption that all organisms were expelled upwards with the brine, which is likely not the case as there will also be a loss to the water column. However, upwards permeability in nilas has been shown to be higher compared to the downwards permeability³².

The experiment was carried out during a period of relatively low (49%) average air humidity during the development of nilas from around 12 noon until about 6 pm on the 9th of March, as compared to the higher relative humidities during the 8th of March (72%) and 10th of March (67%). The low relative air humidity provides a potential for transport of water vapor from the sea ice surface to the atmosphere by evaporation⁶¹ and thereby an increase in concentrations of solutes and particulate matter of the brine. Low ambient air temperature (− 12 °C) strongly reduces evaporation rates⁶³but with a brine layer thickness of only 1–2 mm evaporation might have contributed to the observed high enrichments in the brine. An increase in nutrient concentrations in the brine, potentially attributable to evaporation, was observed, but the disproportionate enrichment of microorganisms remains difficult to explain.

Bacteria, VLPs, and microalgae occur at significant quantities in sea ice where they thrive and are adapted to high salinities³⁴. Therefore, microbial growth may have contributed to the observed enrichments in the brine. The transition of the microorganisms from an ambient seawater salinity of 32.6 to the brine 5 h salinity of 72 would however induce osmotic stress for the organisms and studies have shown that bacteria require a period of one or two days to adjust to the new salinities before growth can occur⁶⁵. Here, the residence time in the brine 5 h was less than 5 hours, which is probably a too short period for a full adjustment and to increase the bacterial concentration about 35 times between seawater (1.89 × 10⁵ cells mL^{– 1}) and brine 5 h (6.7 × 10⁵ cells mL^{– 1}). It is also likely that microalgae would require a longer adjustment time than 5 hours. There were, however, slight increases in both bacterial concentrations in the brine between 5 h and 24 h, and a decrease in microalgae concentrations but changes were not statistically significant. Hence, it is not supposed that growth was the predominant factor contributing to the high brine 5 h concentrations of the microorganisms, but slight growth may have occurred and contributed to the enrichments.

Ocean surfaces are covered by a few µm (1–1000) thick sea surface microlayer (SML) enriched in carbohydrates, lipids, proteins and microorganisms^67,68 with microbial enrichments of 15–25⁶⁸. SML microorganism concentrations were not measured in the present ponds but the enrichment should be 35 in the SML to reach the 5 h brine bacterial concentration of 6.7 × 10⁵ cells ml^{– 1} with a seawater bacterial concentration of 1.89 × 10⁴ cells ml^{– 1}, and enrichment would reach 83 and 68, for VLP and microalgae, respectively. It is assumed that an enriched SML on the surface is incorporated into the brine during brine expulsion though observed values are unrealistically high to be the predominant factor for the high brine enrichment. Thus, we suggest that while the SML can contribute to the high enrichment of the brine 5 h, its impact was likely minor.

Conclusions

In addressing the questions raised, we have shown that nilas was only significantly enriched in microalgae and Chl a, but not in any of the nutrients measured. The brine 5 h showed significant enrichments for all microorganisms (microalgae, bacteria and VLPs), but not for nutrients and Chl a. and with no statistically significant changes in enrichment over time demonstrated by a comparison of the 5 h and 24 h enrichments. The strong impairment of the microalgae was supposedly related to the fact that the microalgae were fixed in position in the nilas, where 100 to 200 higher PAR intensities prevailed, compared to intensities in the water column below the sea ice. Additionally, the microalgae probably suffered from strong osmotic stress along with the high PAR intensities, since no fluorescence signal could be detected in the brine samples. A main question was why we observed such high enrichments of microorganisms and not for nutrients in the brine, and four explanations were discussed – concentration changes, evaporation, growth, and surface microbial layer. Provided that there are not any other unknown processes or mechanisms it appears reasonable to conclude that up concentration offers the best explanation, although the other factors listed above might also contribute to the observed enrichment in the brine. More research is required to elucidate in more detail how particulate matter as the microorganisms could be enriched to the observed degree. This would build a foundation to further investigate firstly, how these enrichments affect frost flowers and secondly the atmosphere.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

TST, DS, LZJ, and LCLH designed the study and TST, LZJ conducted the field work. LZJ, PA, LM, and BS carried out the laboratory analyses. LCLH, LZJ, and TST compiled the data and wrote the manuscript. All authors contributed to the manuscript revision, read, and approved the submitted version.

Funding

This work was supported by the The Danish Agency for Higher Education and Science for the AMINERAL network (1113–00025B), the Villum Fonden (research grants 23175 and 37435), the Novo Nordisk Foundation (NNF19OC0056963), and the Independent Research Fund Denmark (9145–00001B). The project was developed in the frame of the project FACE–IT (The future of Arctic Coastal Systems – Identifying Transitions in Fjord Systems and Adjacent Coastal Areas). FACE–IT has received funding from European Union’s Horizon 2020 research and innovation program under grant number agreement No 869154. We gratefully acknowledge the contributions of the Danish National Research Foundation (DNRF 185) to the Center for Ice-Free Arctic Research (CIFAR), Aarhus University.The Greenland Climate Research Center, GINR, Nuuk, is thanked for financial support and extended logistical support. We are thankful for being able to perform flow cytometry at the FACS Core Facility, Aarhus University, Denmark.

Data availability

Data will be available from authors on request.

Declarations

Competing interests

The authors declare no competing interests.