Oxygenation of anoxic sediments triggers hatching of zooplankton eggs

Abstract

Many coastal marine systems have extensive areas with anoxic sediments and it is not well known how these conditions affect the benthic–pelagic coupling. Zooplankton lay their eggs in the pelagic zone, and some sink and lie dormant in the sediment, before hatched zooplankton return to the water column. In this study, we investigated how oxygenation of long-term anoxic sediments affects the hatching frequency of dormant zooplankton eggs. Anoxic sediments from the brackish Baltic Sea were sampled and incubated for 26 days with constant aeration whereby, the sediment surface and the overlying water were turned oxic. Newly hatched rotifers and copepod nauplii (juveniles) were observed after 5 and 8 days, respectively. Approximately 1.5 × 10⁵ nauplii m⁻² emerged from sediment turned oxic compared with 0.02 × 10⁵ m⁻² from controls maintained anoxic. This study demonstrated that re-oxygenation of anoxic sediments activated a large pool of buried zooplankton eggs, strengthening the benthic–pelagic coupling of the system. Modelling of the studied anoxic zone suggested that a substantial part of the pelagic copepod population can derive from hatching of dormant eggs. We suggest that this process should be included in future studies to understand population dynamics and carbon flows in marine pelagic systems.

1. Introduction

Coastal waters are among the most productive systems in the world and sustain large populations of fish. The current paradigm is that availability of organic carbon is mainly fuelled through the pelagic pathways from nutrients to primary and secondary producers, but the importance of the benthic–pelagic coupling has recently received more attention [1]. A typical example of benthic–pelagic coupling is re-suspension of nutrients from the sediments because of storms that stimulates bacterial, phytoplankton and zooplankton growth [2–5]. Another example of benthic–pelagic coupling is the occurrence of organisms that have both benthic and pelagic life stages, such as benthic resting stages among phytoplankton that are present in a variety of taxa [6–9]. A wide range of zooplankton have also been shown to have benthic resting stages including rotifers, cladocerans and copepods [10]. It is suggested that zooplankton egg banks are analogous to the plant seed banks that allow long-term survival through periods of harsh environmental conditions [11]. By producing different types of eggs the individual can decrease egg loss in fluctuating environments and ensure hatching of at least some of the eggs [10,12,13]. Hence, dormant life stages and a benthic–pelagic coupling are important for understanding variations and seasonal dynamics in plankton communities.

While zooplankton resting stages have been intensively studied in freshwater lakes and ponds, less is known from marine systems [10]. Many of these systems are eutrophicated with large areas of hypoxic or anoxic sediments where zooplankton may deposit their eggs. Oxygen deficient sediments are rich in hydrogen sulfide which is toxic to many organisms at high concentrations [14] and potentially destroys buried zooplankton eggs. A number of different types of zooplankton resting eggs have been described, for example, diapause and subitaneous quiescent eggs [15]. Diapause eggs have an obligatory pause in development and they require a refractory phase before hatching. During this resting stage they can withstand extreme physical conditions such as low or high temperature and low oxygen conditions [10,12]. On the other hand, quiescent eggs are produced from normal, subitaneous eggs but their development stops if environmental conditions are unfavourable. Development of the eggs can resume without a refractory state if conditions improve [15]. These eggs are less studied but are generally less tolerant to harsh physical conditions.

In this study, we sampled sediment cores from a semi-enclosed brackish Baltic Sea bay which is heavily affected by eutrophication and has extensive anoxic zones in offshore areas and along the coast [16,17]. The hatching frequency of mesozooplankton eggs was counted from anoxic sediments that were oxygenated by bubbling with air. The effects of oxygen on sediment biology is especially interesting to study in systems affected by eutrophication as large-scale oxygenation has been proposed as a measure to reduce internal nutrient loading from anoxic sediments [18–22]. We hypothesize that anoxic sediments contain a large store of zooplankton resting stages deriving from surface dwelling populations. These eggs can be activated upon oxygenation contributing to an increased benthic–pelagic coupling.

2. Material and methods

(a). Field sampling

Sampling was conducted in a Baltic Sea bay close to the town of Loftahammar, Sweden (WGS 84 coordinates: lat. 57 53.531, lon. 16 35.165) on 14 May 2014. A gravity corer was used to sample a set of 13 sediment cores (polymethylmethacrylate; length 60 cm, inner diameter 70 mm and outer diameter 80 mm) at a depth of 31 m below the sea surface. Collected sediment cores contained sediment with a length of 46–49 cm and 440–560 ml of overlaying benthic water. The benthic water (approximately 20 cm over the sediment surface) had a dissolved oxygen concentration of 0.2 mg l⁻¹ (in situ oxygen sensor, WTW Multiline). The sediment at the sampling site had experienced long-term hypoxia–anoxia and is black over the entire year (E Broman 2013-2015, unpublished data, collected on multiple previous occasions during April, May, October and November for the years 2013–2015) and the sediment as well as the bottom water had a strong smell of hydrogen sulfide. Once the sediment core reached the surface waters it was closed at the bottom. Three cores were immediately sliced in the field at intervals of 0–3 cm, 3–6 cm and 6–9 cm for hatching experiments with sediments deriving from different depths. These intervals were decided in order to collect a sufficient volume of sediment for the experiment, deeper sediment was not sliced considering that most buried and viable copepod eggs have previously been observed in the top 10 cm [23]. These sediment slices were stored in 50 ml sterile polypropylene centrifuge tubes (Thermo Scientific) during transportation to the laboratory on the same day. The other 10 sediment cores were transported back to the laboratory (maintained at approx. 10°C during transportation) and stored at 8°C for 1 day in the dark before the incubation experiment was initiated.

(b). Incubation set-up

The sediment cores containing sediment and overlying bottom water were incubated for 26 days in darkness at 8°C. Half of the cores were oxygenated by continuous gentle bubbling with air in the water phase without disturbing the sediment surface (flow rate: 20 cm⁻³ s⁻¹; number of replicates (n) = 5). Over the course of the incubation experiment the sediment surface in this treatment changed from black to light brown, indicating that the sediment became oxygenated. The other five cores acted as anoxic controls (n = 5) and were bubbled with nitrogen gas for 40 min without disturbing the sediment surface at the start of the incubation and for 15–20 min after each sampling and filtration. To ensure the anoxic controls were devoid of oxygen, dissolved oxygen was measured directly above the sediment surface (O₂ dissolved oxygen meter, Innovative Instruments). The anoxic controls were sealed during the incubation between sampling times. A polytetrafluoroethylene tube (Sarstedt) containing four neodymium magnets (each with a diameter of 1 cm) and gravel (to add weight) was submersed in the water phase of all sediment cores, i.e. cores turned oxic and anoxic controls, and used to create a gentle circulation of the water phase via external rotating neodymium magnets.

Sampling of the water phase was carried out by removing 250–300 ml of the water overlying the sediment, the water was filtered through a 15 µm nylon net to catch recently hatched zooplankton and the zooplankton were immediately inspected to see if they were alive. One part of the caught nauplii was cultivated into adulthood in 0.2 µm filtered seawater (polypropylene filter cartridge, Roki Techno) and fed with Rhodomonas salina ad libitium. Another portion of the caught nauplii was preserved in Lugol's solution (final concentration per sample was approx. 0.5% (v/v) Lugol) for microscopic quantification of zooplankton species. The water filtered for zooplankton was then returned back to the corresponding sediment core creating as little turbulence as possible. When needed, 0.2 µm filtered seawater was added to adjust the volume to ensure that the same volume of water was present in the cores. A subsample of 2 ml was also filtered through a 25 mm, 0.7 µm pore size glass fibre filter (GF/F filter, Whatman) and sulfate measured using the Hach–Lange LCK 353 kit on a DR 5000 Hach–Lange spectrophotometer. To confirm that the sediment cores were kept either oxic or anoxic dissolved oxygen levels in the water were measured at the start of the incubation and at every subsampling using an oxygen electrode (O₂ dissolved oxygen meter, Innovative Instruments). Samples for nutrient analysis in the water were also collected at the final sampling occasion in acid washed 500 ml polyethylene bottles. Phosphate and nitrate (in combination with nitrite) were measured on the DR 5000 Hach–Lange spectrophotometer according to Valderrama [24]. Results are presented as means of replicate incubations ± 1 s.d.

(c). Incubation of slices from different sediment depths

Sediments from varying depths (0–3, 3–6 and 6–9 cm) were incubated (n = 3 for each depth) by transferring the sediment slices collected in the field to 600 ml polypropylene containers (VITLAB) containing 400 ml of filtered seawater (0.2 µm polypropylene filter cartridge, Roki Techno). The containers were incubated for 26 days without lids in darkness at 8°C and were continuously bubbled with air until the end of the experiment and fed with Rhodomonas salina ad libitium. The water phase was then filtered and preserved in Lugol's solution. Results are presented as means of replicate incubations ± 1 s.d.

3. Results

(a). Zooplankton hatching in response to oxygenation

Newly hatched nauplii appeared after 5 days of incubation but significant hatching occurred at the second sampling occasion on day 8 when they reached a maximum of 592 ± 223 l⁻¹(figure 1). Hatching continued during the following sampling times but the rate decreased after 15 days. Adult copepods were not detected in incubations but some nauplii were cultivated to adulthood outside the incubation after each sampling occasion (fed Rhodomonas salina ad libitium). From these cultured animals, Acartia spp. and Eurythemora spp. were the only detected copepod species with Acartia spp. constituting approximately 80% of the total abundance. As observed for copepods, rotifers were absent at the start of the incubation. Syncheata spp. and Keratella spp. appeared in the oxygenated water column on day 5 and peaked on day 12 at concentrations of 271 ± 70 l⁻¹ and 34 ± 15 l⁻¹, respectively (figure 1). Syncheata spp. and Keratella spp. were virtually absent in the anoxic incubations. Plagiopyla spp. ciliates were present in low numbers (237 ± 33 l⁻¹) in the anoxic cores turned oxic at the start of the experiment but quickly disappeared when the cores were oxygenated by bubbling with air (figure 1). In contrast, Plagiopyla spp. increased in abundance from 160 ± 30 l⁻¹ at day 0 to a maximum of 1610 ± 390 l⁻¹ after 5 days in the anoxic incubations (figure 1). Other ciliate species were sporadically observed in the oxygenated cores but were not quantified.

Figure 1.

Number per litre of hatched nauplii (a), Syncheata spp. (b), Keratella spp. (c) and Plagiopyla spp. (d) from anoxic sediment cores turned oxic by bubbling with air (open circle, n = 5) and control sediments that were maintained anoxic (filled circle, n = 5).

The oxygen concentration in the anoxic incubations was constantly less than 0.2 mg l⁻¹ and often below the detection limit (0.1 mg l⁻¹) while the aerated incubations contained 5.5 ± 0.6 mg l⁻¹ oxygen. Sulfate levels were initially 4.9 ± 0.01 mM in all incubations and were relatively stable in the oxygenated treatment but declined in the anoxic treatment with final concentrations of 4.5 ± 0.08 and 3.2 ± 0.05 mM in oxic and anoxic treatments, respectively (t = 14.3, d.f. = 8, p < 0.001). Nutrient levels were not measured at the start of the incubations but phosphate was lower in oxygenated cores compared with anoxic cores after 26 days of incubation (6.9 ± 1.2 and 27.9 ± 1.0 µM, respectively; t = 14.6, d.f. = 7, p < 0.001). In contrast, nitrate plus nitrite levels were higher in the oxic compared with the anoxic treatments at the end of the experiment (3.2 ± 0.4 and 1.1 ± 0.2 µM, respectively; t = 5.0, d.f. = 7, p = 0.01).

(b). Hatching in relation to sediment area and volume

A total of 1.55 ± 0.31 × 10⁵ nauplii m⁻² emerged from the oxygenated sediments compared with 0.02 ± 0.01 × 10⁵ m⁻² that emerged from the anoxic sediment control (figure 2a). Of the incubated sediments sliced in the field, 2320 ± 213 nauplii l⁻¹ sediment hatched from the 0–3 cm sediment (figure 2b) compared with 49 ± 15 l⁻¹ and 9 ± 4 l⁻¹ that hatched from the sediments taken from 3–6 cm and 6–9 cm below the surface, respectively.

Figure 2.

Accumulated number of hatched nauplii per square metre of sediment calculated as the total of the anoxic sediment cores turned oxic (n = 5) and anoxic controls (n = 5) after 26 days of incubation (a) and the accumulated total of hatched nauplii from the different sediment layers per litre of sediment (n = 3 for each layer) cut in the field and incubated for 26 days.

4. Discussion

Many marine systems have large areas with anoxic sediments [16,17] and this study demonstrated that anoxic sediments can harbour a large pool of viable zooplankton eggs that were activated after turning the sediments oxic. Even though it is possible that mixing caused by the gentle bubbling of air triggered egg hatching rather than the increase in oxygen concentration, this seems unlikely considering that the controls were also bubbled (with N₂ gas) at the start of the experiment and after every sampling point. It is also unlikely that the N₂ gas prevented hatching of eggs in the anoxic controls, considering that these cores were not continuously bubbled and that N₂ gas is inert. To our knowledge, no studies have been published to link N₂ gas to interference with zooplankton egg hatching. On average, 1.5 × 10⁵ nauplii copepods emerged per square metre of sediment and this is in the range of previous measurements of copepod resting egg densities that are of the order of 10³–10⁷ eggs m⁻² [10,25–28]. However, most estimates of resting egg frequency are derived from active sieving and hatching of eggs from the sediment. In contrast, we incubated intact sediment and studied actual emergence of zooplankton with minimal disturbance of the sediment surface which was more similar to the natural situation. Previously, temperature and photoperiodicity have been shown to be important factors in terminating diapause and inducing hatching [10]. In contrast, although a relationship has been observed between oxygen and hatching of zooplankton eggs [29], it is not generally regarded as an important cue for hatching [10]. Most hatching occurred from the top layer of the sediment but a few hatchlings also emerged from deeper sediments. Neither age estimates nor sedimentation rates have been estimated in this system and therefore, the age of the zooplankton eggs cannot be determined. In other systems, copepod eggs at a depth of approximately 10 cm are in the order of a few decades old [30] and marine copepod eggs as old as 70 years can be viable [30–32]. The long-term survival of resting eggs is limited by the storage of energy and the metabolic rate and hence can be prolonged by oxygen deficiency and low temperature because anoxia can completely arrest metabolism [33–35].

The only copepod species detected in the sediments were Acartia spp. and Eurytemora spp. that along with the rotifer genera Synchaeta spp. and Keratella spp. are common species at nearby sampling sites [36]. Copepods mainly use abiotic cues for initiation and termination of diapause, whereas rotifers and cladocerans also use biotic cues [10]. In contrast, cladocerans were absent which could indicate that their eggs were less tolerant of anoxia or that these eggs were not activated upon oxygenation. As the morphology of the eggs was not studied, it cannot be determined if the hatched eggs were subitaneous eggs in quiescent state or true diapause eggs. Acartia bifilosa is a very common species in the study area and appears to only produce normal subitaneous eggs that can then go into a quiescent state in response to low oxygen levels [29,37]. In general, the quiescent state is argued to only last for a few months before the eggs die [38,39]. For example, egg hatching success in A. tonsa is heavily reduced after 30 days of exposure to anoxic water [40]. This suggests that the hatched eggs in this study were true diapausing eggs, and not subitaneous quiescent eggs, as the sediments were sampled in early spring before any major zooplankton production. However, Baltic Sea A. bifilosa subitaneous eggs are very tolerant to low oxygen levels as hatching frequency is maintained at approximately 40% after the eggs had been stored in anoxic water up to 10 months [29]. Buried zooplankton eggs and their species origin were not investigated initially before the experiment, and further studies are needed to determine which eggs survive during long-term anoxic conditions. In shallow systems with anoxic sediments, eggs would probably sink to the bottom before they hatch [41,42] and selection for eggs that are resistant to anoxia would hence be strong. Sinking rates of marine copepod eggs have been estimated to range between 15 and 35 m⁻¹ d⁻¹ [41,43], and in shallow water ( <20 m) it is not uncommon that nearly all eggs are settled on the sediment before hatching [41,42]. Considering that the studied site had a depth of 31 m with cold anoxic bottom water, it is plausible that a substantial part of the eggs settle on the sediment before hatching. The Baltic Sea is a relatively shallow system with a mean depth of 55 m. Hence, eggs ending up in anoxic sediments in shallow systems can be trapped if oxygen conditions are poor.

In addition to copepods, a number of rotifers also emerged from the sediment. Rotifer resting stages have not been greatly studied in marine systems [10] but diapausing amictic eggs have been observed from Synchaeta pectinata [44]. Temperature and light are important factors initiating hatching of rotifer resting stages [45] and this is, to our knowledge, the first example of oxygen as a cue for rotifer resting stage hatching. Ciliates were not included in this study but one species belonging to Plagiopyla spp. was resistant to the preservative (Lugol's solution) and had the opposite preference compared with the other zooplankton groups and thrived in anoxic water. Several ciliate species are unable to use oxidative phosphorylation in their energy metabolism and may be sensitive to oxygen [46]. Hence, low oxygen zones are likely niches where other zooplankton cannot exist creating a competitive benefit for ciliates.

Anoxic zones can be sporadically ventilated by natural inflow of oxygen rich water [21] and therefore, the extent of anoxic zones fluctuates over time [16]. This potentially triggers hatching of eggs trapped in anoxic conditions. However, current population and carbon flow models tend not to incorporate the magnitude of this benthic–pelagic coupling [47,48]. Carbon circulation and deposition in the ocean is generally described in a concept called the biological pump [48]. This model assumes that pelagic particles (e.g. faecal pellets, dead cells) aggregate, sink down and are then lost from the system when deposited in the sediment [48]. This model does not include the potential for carbon flows from the sediment back to the open water system. Hence, the magnitude of hatching in terms of carbon flow was modelled in the studied anoxic zone (approx. 60 000 m⁻²). Assuming that copepods at the 31 m deep anoxic site could use the water column down to 20 m (based on oxygen availability) and an average abundance of 10 copepods l⁻¹ (sampled at a nearby site; [36]) there are approximately 7.2 × 10⁹ adult copepods in the water overlying the anoxic sediment. A carbon factor of 0.3 µg per emerged nauplii [49] was then used to calculate a carbon flux of 0.05 g C m⁻² given the number of nauplii that emerged in the incubation experiment. As these nauplii potentially grow to adulthood (20 days at 15°C; [50]), we assume a daily mortality rate of 7.4% [51] and only approximately 20% of all the hatched nauplii survive until adulthood due to factors such as predation mortality and food quantity and quality. Taken together, hatching of the buried egg bank has the potential to contribute to a substantial share of the adult copepod population in the water column (approx. 20%; figure 3). However, the outcome in terms of net adult population size deriving from benthic eggs depends on the limiting factors of the system (e.g. food) and the relative competitive differences between sediment derived and pelagic animals [52,53]. More importantly, as copepod eggs in anoxic Baltic Sea sediments are estimated to survive up to 15–19 years [23,54], hatching of previous generation copepods will add new genetic variation to the pelagic population. This has, for example, been observed among some freshwater zooplankton species where increased genetic polymorphism was observed in synchrony to large sediment hatching events [55,56]. The positive effects of higher genetic and phenotypic variability on population and species performance have been extensively studied in a variety of organisms, e.g. plants, invertebrates and mammals [57]. Some of the observed effects are decreased risk of extinction, lower vulnerability to stress as well as an increase in population size stability [57]. Furthermore, lake studies have shown that a large portion of zooplankton emergence from sediments occurs during spring, when the abundance of pelagic adults is low, suggesting that benthic eggs could contribute to the initiation of the annual population cycle [58]. Hence, this benthic recruitment to the pelagic population would be absent in anoxic sediments. We suggest that the carbon flow of hatched zooplankton from oxygenated sediments to the open water should be added to the concept of the biological pump and to other future studies on benthic–pelagic coupling to quantify how this flow varies seasonally, spatially and with depth of the system.

Figure 3.

Model of carbon flow and zooplankton production in response to oxygenation in the studied anoxic zone. The model suggested that re-oxygenation during summer would induce hatching of nauplii that, when reaching adulthood, would constitute approximately 20% of the copepod population. The number of eggs in the egg bank is according to the literature references [10,25–28].

Coastal zones in the Baltic Sea experiencing hypoxia have increased during the last 50 years with over 100 sites affected [17]. The Baltic Proper is, furthermore, estimated to consist of 70 000–100 000 km² anoxic and hypoxic sediment zones [59] and results from this study implies that a substantial part of the zooplankton lie dormant as eggs in the anoxic sediments. The lack of copepod egg hatching would also limit the capacity of the population to recover from harsh times when adult populations in the water column have decreased, such as in the spring. It has also been proposed that large-scale artificial oxygenation of anoxic sediments could be used as a remediation strategy for eutrophicated systems by reducing the flux of nutrients to the water column [18–22], a strategy that has recently been tested on a smaller fjord with some positive results in terms of reduced phosphate concentrations [60]. However, it is still debatable how feasible artificial oxygenation would be when implemented on a larger scale [19].

To conclude, we show that re-oxygenation of anoxic sediments will activate buried zooplankton eggs, potentially strengthening the benthic–pelagic coupling in these systems with possible effects on trophic interactions as zooplankton are important prey for fish and also can exert top down effects on lower trophic levels such as phytoplankton.

Data accessibility

The datasets supporting this article have been uploaded as part of the electronic supplementary material.

Authors' contributions

E.B. designed and coordinated the overall study, performed laboratory work and helped draft the manuscript; M.B. carried out laboratory work; M.D. helped design the study and draft the manuscript; S.H. helped design the study, performed laboratory work and helped draft the manuscript. All authors gave final approval for publication.

Competing interests

We have no competing interests.

Funding

All authors were financially supported by the Linnaeus University Centre for Ecology and Evolution in Microbial model Systems (EEMiS).