Decreasing salinity improves the feeding conditions of the Baltic herring (Clupea harengus membras) during spring in the Bothnian Sea, northern Baltic

Abstract

The stomach contents of the herring (Clupea harengus membras L.) from the Bothnian Sea, northern Baltic, were analysed during spring 2011 and 2013. The stomachs were full of Limnocalanus macrurus in May and June, and an improvement in the condition of herring was observed when fish started to feed on this prey. The analyses showed that Limnocalanus is currently an important link between lower trophic levels and Baltic herring in the Bothnian Sea.

Baltic herring, Bothnian Sea, food, Limnocalanus, salinity

Introduction

The global change in climate has affected marine ecosystems worldwide. In the Baltic Sea (Figure 1a), the most significant change is the decrease in salinity (Hänninen et al., 2000), which has caused a decline in marine zooplankton (Möllmann et al., 2005; Suikkanen et al., 2013) and a deterioration in the feeding conditions of herring (Clupea harengus membras) and sprat (Sprattus sprattus) (Casini et al., 2004). Low abundance of preferred prey and the concomitant increase in competition for food among clupeids have contributed to a reduction in the herring’s nutritional status and condition, growth rate, and possibly even stock size (Möllmann et al., 2004; Casini et al., 2006, 2010).

Figure 1.

(a) The Baltic Sea and the ICES subdivisions (SD30 = the Bothnian Sea; SDs 25–29 = the Baltic Proper); (b) mean surface salinity in the southern Bothnian Sea and the spawning stock of the herring (SSB in tonnes) in SD30 and SDs 25–29 and 32 (combined) in 1975–2011. (SSB according to ICES (2012); salinity according to the Helcom database [http://www.ices.dk/marine-data/dataset-collections/Pages/HELCOM.aspx (last accessed 5 May 2013)]; (c) sampling station of salinity (SR5; black circle) and approximate sampling area of fish (grey area).

The Bothnian Sea (Figure 1a), which is a ∼70 000 km² wide basin north of the Baltic Proper, has also become less saline during the past 30 years. However, the environmental conditions in the Bothnian Sea differ from those in the central Baltic, as water salinity is lower in general and marine copepods occur in lower abundances, even in periods of higher salinity (Lindqvist, 1959). In the Bothnian Sea, the most abundant copepod species in spring is Limnocalanus macrurus, which is found in brackish water ecosystems and deep lakes in the northern hemisphere. Limnocalanus is the largest copepod in the Baltic Sea (adult length 2–3 mm), but it is abundant only in areas where salinity is low. In the Bothnian Sea, Limnocalanus appears in zooplankton early in spring when other zooplankton is sparse (Lindqvist, 1959), and its abundance has significantly increased during the salinity decline (Postel et al., 2011). Limnocalanus contains plenty of lipids (Dahlgren et al., 2012), which makes it optimal food for the herring in spring, when the energy requirements of fish are high due to an ongoing reproduction period.

Although the spawning stock of herring in the Baltic Proper (Figure 1b) has decreased over the past 30 years, the trend is the opposite in the Bothnian Sea (ICES, 2012; Figure 1c). The reason for this difference is not known, but recently, Lindegren et al. (2011) put forward the idea that increased abundance of Limnocalanus and other zooplankton prey could explain the successful stock development in the Bothnian Sea by decreasing fish mortality in the adult phase. This has not been verified, however, as even the current food of herring is poorly known in the Bothnian Sea.

In this work, we investigated the stomach contents of herring collected from the Bothnian Sea in order to find out the role of Limnocalanus in the herring’s diet. The study was focussed on spring and early summer (March–June), when Limnocalanus is found in the plankton, and, moreover, this season is potentially the bottle-neck in the survival of adult herring as a result of the poor feeding opportunities and high energy requirements of fish during the spawning season.

Material and methods

Herring samples were collected monthly from March–June 2011 from the catches of commercial trawlers operating in the Bothnian Sea (Figure 1c). The data were supplemented with additional samples collected in 2013 in order to detect short-time variations in the herring's diet at the onset of the active feeding period. Of the catch, a random sample of fish was taken and stored at −20°C until analysis. In the laboratory, fish were measured for total length (mm) and weight (g at 0.1 g accuracy), and the sex, gonad weight (g at 0.01 g accuracy) and developmental stage of gonads were determined (stages: 2–4 = developing; 5 = ripe; 6 = spawning; 7 = spent). Fulton’s condition factor (CF) (Bagenal and Tesch, 1978) was calculated using somatic weight (total weight − gonad weight) instead of fish total weight: CF = 100 × somatic weight/length³. The gut and stomach were removed and preserved individually in glass tubes containing 70% ethanol. The stomachs were opened; the contents were then carefully removed to a Petri dish and investigated under a dissecting microscope. The fullness of the stomach was estimated by an index (stomach fullness index; SFI), which was described as follows:

0 = Empty; stomach not swollen; no remains of food found (whole, parts or digested)

1 = Stomach not swollen; some remains of food found

2 = Stomach slightly swollen; food takes up about half of the full volume

3 = Stomach clearly swollen; food takes up about two-thirds of the full volume

4 = Full; stomach swollen to its full volume to “bursting”

Macroscopic prey items were identified under the dissecting microscope. When food consisted only of zooplankton, the stomach contents were mixed with 70% ethanol on a Petri dish; from the mixture a sample was taken (using a wide-mouthed pipette) and placed on an object slide. Organisms were carefully separated from one another using a thin needle; a drop of glycerol was added and the sample was covered with a cover glass. The sample was investigated with a light microscope (magnification 100× and larger). About 100 prey items on each slide were identified (to a species level whenever possible) and their number counted. Frequency of occurrence (F%) of the prey species in stomachs was expressed as the percentage of stomachs in which each prey type was recorded out of the total number of stomachs examined, excluding the empty stomachs and those where the contents were too digested to be identifiable. The total number of fish examined was 302, of which the stomach contents were identifiable in 234 fish.

Comparisons of stomach fullness index were performed using Kruskall–Wallis non-parametric ANOVA and the Mann-Whitney U-test. Somatic condition factor was compared using the t-test, after normality of data and equality of variances had been examined using the Kolmogorov–Smirnov one-sample test and Levene’s test, respectively. Correlations between fish length and somatic condition factor were calculated with Pearson’s correlation test.

Results

In 2011, the SFI varied significantly among the samples collected between March and June (Kruskall–Wallis H = 71.863; p < 0.001; df = 4). The mean SFI was <2 until the middle of May (Figure 2), but increased significantly between 18 and 25 May (U = 424.00; p < 0.001; n₁ = 47; n₂ = 39). Also, the frequency of SFI-value 4 (stomachs full to “bursting”) increased from 0 to 33%. In June 2011, 43% of all stomachs investigated were full (SFI = 4) and 23% almost full (SFI = 3). Additional samples collected in 2013 showed that the mean SFI was again 2.9–3.1 at the end of May (Figure 2), although somewhat lower in the sample taken in early June, when 21% of fish had SFI-values of 3–4. Totally empty stomachs were found only in March 2011 (7% of the total at that time) and on 3 June 2013 (4% of total; Figure 2).

Figure 2.

Prevalence of empty stomachs (ES%) and stomach fullness index (SFI; mean ± SD) in the herring of the Bothnian Sea in 2011 and 2013. Number of fish studied is shown in parenthesis, vertical line separates the study years.

In March and April (2011), herring fed on Mysis spp. (F = 100% and 88%, respectively; Figure 3a) and occasionally amphipods (Monoporeia affinis). The frequency of mysids in herring stomachs decreased in May to 15–24% then further to 5% in June, when only a few fish (2 out of 42 examined) had fed on them (Figure 3a). In the middle of May (2011), herring started to feed on zooplankton, which was found in all stomachs (F = 100%). Zooplankton consisted of adults of two calanoid species, Eurytemora affinis and Limno-calanus macrurus. At first, Eurytemora was numerically more important, but at the end of May the stomachs contained no other prey but Limnocalanus (Figure 3b). Later in June, Eurytemora started to increase in the diet again. The samples collected within a week in 2013 indicated little variation in the stomach contents; Limnocalanus was the only prey at the end of May and the main prey also at the beginning of June (Figure 3a and b). It was eaten by herring of a range of sizes, and even small fish (13–14 cm in length) had hundreds of adult individuals in their stomachs. In May and June, the stomach contents were also characterized by plenty of oil globules having orange colour.

Figure 3.

(a) Frequency of occurrence (%) of main prey items in herring stomachs in the Bothnian Sea during spring months 2011 and 2013. Number of fish with identifiable prey is shown in parenthesis; (b) species composition of zooplankton in the stomachs of the Bothnian Sea herring in May–June 2011 and 2013, estimated as percentage of total number of prey identified (%N). Number of samples studied is shown in parenthesis; vertical lines separate the study years.

The size of herring in the samples varied between 7.8 cm and 20.9 cm, with some variation in the mean lengths among the dates (Table 1). Juveniles and adults with developing gonads were found in all samples, but spawning fish were found only in May and June. During spring 2011, CF of herring decreased significantly from March to 18 May (t = 3.438; df = 74; p < 0.01; Figure 4), but started to increase again, being significantly higher on 17 June than on 18 May (2011; t = −5.590; p < 0.001; df = 82). CF had no correlation with fish length on 18 May (Pearson’s correlation coefficient r = −0.097; p > 0.05; n = 47), on 17 June (r = −0.290; p > 0.05; n = 37), or on 26 March (r = −0.343; p > 0.05; n = 26).

Table 1.

Characteristics of the herring investigated at different dates in 2011 and 2013. Total length (mean, standard deviation and minimum and maximum values in the samples) and maturity stage of the gonads; n = number of fish studied.

	Length cm			Maturity
Date	Mean	s.d.	Min-Max	Stage	n
2011
26 March	15.3	1.8	10.0–19.2	1–5	33
13 April	16.3	1.7	12.5–20.5	1–5	59
18 May	13.8	2.7	7.8–20.9	1–6	47
25 May	15.0	1.7	10.9–17.7	1–6	39
17 June	16.2	1.6	13.0–20.1	1–8	60
2013
29 May	16.7	0.8	15.2–18.1	1–5	15
31 May	16.3	1.1	14.0–18.4	1–5	25
3 June	16.3	1.4	13.8–19.2	2–5	24

Figure 4.

Somatic condition factor (CF; mean ± SD) of the herring in the Bothnian Sea during spring 2011. The results of the t-test (p-value) comparing CF in March vs. 18 May and 18 May vs. 17 June are indicated with dashed lines; number of fish studied is shown in parenthesis.

Discussion

Although herring is the dominant planktivorous fish in the Bothnian Sea and an important target of the commercial fishery, its food is known only broadly in this area. The studies made by Flinkman et al. (1992) and Parmanne et al. (2003) have shown that herring eats mysids and zooplankton in the Bothnian Sea, as in the central Baltic (Gorokhova et al., 2004; Möllmann et al., 2004). However, these studies don't provide information about the herring's food during spring in the open sea as they were carried out in late summer or in the coastal zone. According to Flinkman et al. (1992), herring shows a preference for Limnocalanus and Eurytemora, both of which have substantially increased in the Bothnian Sea plankton since the 1980s (Postel et al., 2011) and which were found to be important prey of herring in the present study.

In March and April 2011, herring fed on mysids, but later in spring only a few fish fed on them, as shown by their low frequency in the stomachs. In the middle of May, herring started to feed on zooplankton, but low SFIs suggest that the abundance of prey was still low in the environment at this stage of spring. At first, fish fed both on Eurytemora and Limnocalanus, but at the end of May and middle of June, Limnocalanus formed the major component of the herring's food. At the same time, the SFI increased significantly, suggesting a peak abundance of Limnocalanus in the plankton. High SFIs were also found in 2013, with the exception of the sample taken in early June, when even some empty stomachs were found. This indicates that there are also spatial and/or temporal differences in the feeding conditions of herring in the Bothnian Sea, possibly caused by prey availability or because of some other factor influencing fish feeding. However, in both years, Limnocalanus was an important prey of herring in late May and June, and its number in fish stomachs was high; in many cases the stomachs were characterized as “bursting”. The lipid stores of Limnocalanus (Dahlgren et al., 2012) could be detected in fish stomachs as oil globules, indicating that the energy content of the food was high. A rich food supply evidently improved the somatic condition of herring, which lost ground during spring as the maturation of the gonads proceeded.

The improvement in somatic condition, associated with increasing stomach fullness and a high amount of Limnocalanus in the diet, suggests that Limnocalanus is the key species responsible for improving the physiological condition of herring in spring and early summer. In our samples, the majority of fish were at the prespawning stage, but no doubt good physical condition helps a fish to recover from the physiological stress caused by spawning. At spawning time, the fat reserves and condition of the Baltic herring are at their lowest level (Rajasilta, 1992), but fish weight and condition can respond rapidly to feeding, especially if the oil content of food is high (Chua and Teng, 1982; Young et al., 2005). It is not known how this affects the herring's survival, but the period after spawning could likely form the bottle-neck when those individuals that are in a poor condition and unable to replenish their energy reserves are eliminated from the population. Thus, it is possible that increased abundance of Limnocalanus contributes to the increase in the herring stock in the Bothnian Sea, as presented by Lindegren et al. (2011). The preconditions for improved fish survival exist at least in late spring, but further studies are needed to detect the spatial and temporal patterns in prey availability and herring feeding, and on their connection to the survival of the Bothnian Sea herring.

Funding

Financial support was received from the Turku University Foundation.

Acknowledgements

We wish to express our thanks to P. Suominen for his assistance and advice during the fieldwork, to the fishermen who kindly provided the herring samples, and to two anonymous referees for their constructive comments on the manuscript.

Author notes

Handling editor: Valerio Bartolino