Importance of Fucus vesiculosus (bladderwrack) for coastal fish communities in the Baltic Sea

Emma Mattsson

Department of Ecology, Environment and Plant Sciences

Degree 45 HE credits

Marine biology

Marine biology (120 credits)

Spring term 2019

Supervisor: Charlotte Berkström

Co-supervisor: Maria Eggertsen and Nils Kautsky

Importance of bladderwrack for coastal fish communities in the Baltic Sea

Emma Mattsson

Abstract Within temperate coastal seascapes, macroalgae provide habitats for different organisms such as invertebrates and fish. In analogy to seagrass meadows, macroalgae beds are known for their importance as fish nurseries. However, within the Baltic Sea the importance of the canopy forming macroalgae Fucus vesiculosus for coastal fish communities, especially the juveniles, is unclear. In order to address this knowledge gap, fish communities in areas with and without F. vesiculosus were investigated around Askö, an island in the archipelago of the Baltic Sea. Sites were subjected to different exposures (sheltered, exposed or very exposed) and three different methods were used for sampling (underwater visual census (UVCs), beach seine netting and remote underwater videos (RUVs)). Overall, fish community composition differed significantly among locations and fish abundance and fish biomass were significantly higher in sites with F. vesiculosus than sites without. There was no significant relationship between algae cover or habitat complexity and fish abundance/biomass in sites with F. vesiculosus. Fish behaviour differed between sites with and without F. vesiculosus, with fish feeding more in sites with F. vesiculosus and traveling more in sites without F. vesiculosus. Only one location, Knabberskär, had significantly higher species richness in F. vesiculosus than in sites without F. vesiculosus. There were no differences in juvenile abundance among sites with or without F. vesiculosus and abundance of adult fish was higher than juvenile fish, regardless of location, site or species. Mean invertebrate abundance was a twice as high in the sheltered location Husbåtsviken than in Knabberskär. Higher fish abundance, fish biomass and species richness in sites with F. vesiculosus compared to sites without, suggest that macroalgae may play an important role in the Baltic Sea, however it might not be as important for juvenile fishes as predicted. The three different sampling methods provided similar results for fish abundance, but not for fish biomass. Continued studies where the relationship between fish communities and aspects of F. vesiculosus structure (such as canopy height) as well as linkage with other habitats is recommended for further understanding and better protection of F. vesiculosus habitats.

Keywords

Baltic Sea, Fucus vesiculosus, fish communities, fish nurseries, climate change

Popular summary Seaweed habitats are an important part of the coastal seascape and provide shelter and food for fish, especially young fish worldwide. However, less is known within the Baltic Sea region. Therefore, this study focused on the importance of the seaweed bladderwrack for fish communities along the coast of the Baltic Sea. The study took place around Askö, an island located outside Trosa, Sweden where three different methods were used to sample fish. The results showed that the fish species differed between locations and that there was a higher amount of fish in sites where bladderwrack was present, compared to where it was lacking. Contrary to predictions, the amount of young fish was not higher in bladderwrack and lower than that of the adult fish. The behaviour of fish was also studied and showed that fish travel more in sites where bladderwrack is lacking, while they eat more in sites where it is present. In areas with bladderwrack, there was also twice as many small animals (often seen as fish food) in plants gathered from sheltered sites than less sheltered sites. The higher amount of fish and fish species at sites with bladderwrack, suggests that the seaweed is an important habitat for fish communities along the coast of the Baltic Sea and needs to be further studied, for a better understanding of its importance.

Ethical and social aspects Beach seine nets were used to sample fish in sites with and without Fucus vesiculosus. To minimize impact on the fish community only the necessary amount of fish was gathered, enough for a scientific analysis. It is documented that fish feel stress and pain (Braithwaite & Boulcott, 2007) so to diminish the pain and stress of individual fish, fish were anaesthetized directly after capture by placing them in a bucket with MS22 solution. This sampling was necessary to complement the other two chosen methods (underwater visual census and remote underwater video), in order to capture the whole fish community. Regarding the social aspects of the study, it is important to understand which ecosystem services are provided by F. vesiculosus habitats and if there would be consequences if F. vesiculosus disappeared? Hence, it is important to gather information about the importance of macroalgae for fish communities within the Baltic Sea. Especially since the Baltic Sea is subjected to a high amount of anthropogenic pressure and climate change.

Contents Introduction .............................................................................................................................. 1 Material and methods .............................................................................................................. 2

Study sites and field sampling ................................................................................................ 2 Underwater visual point census (UVCs) ............................................................................. 4 Remote Underwater Videos (RUVs) .................................................................................. 5 Beach seine net ................................................................................................................... 5

Data handling (and standardization) ....................................................................................... 5 Statistical analysis of data ....................................................................................................... 6

Results ....................................................................................................................................... 7 Fish communities .................................................................................................................... 7 Fish abundance and biomass .................................................................................................. 8 Species richness .................................................................................................................... 11 F. vesiculosus as juvenile habitat ......................................................................................... 13 Habitat and associated invertebrates ..................................................................................... 13

Discussion ................................................................................................................................ 16 Fish community differences in areas with and without F. vesiculosus ............................. 16 Fish community differences among locations .................................................................. 17 F. vesiculosus as juvenile habitat ...................................................................................... 17 Food availability and fish behaviour ................................................................................. 18 Comparison between sampling methods........................................................................... 18 Future perspectives and management ............................................................................... 19

Acknowledgements ................................................................................................................. 20 References ............................................................................................................................... 20 Appendix ................................................................................................................................. 24

1

Introduction A seascape is composed by a mosaic of diverse habitats, which associated flora and fauna depend on and support the exchange of nutrients and energy between habitats (Banks et al., 2007; Berkström et al., 2012). One important, but perhaps less acknowledged habitat, is the canopy-forming macroalgae that not only contribute to primary production along the coast, but also is an important component (ecological engineer) in shaping the ecosystem (Steneck et al., 2002). Studies within temperate habitats have shown that there is a higher fish and invertebrate species richness and abundance in habitats with higher complexity, such as macroalgae habitats (Christie et al., 2009; Levinl & Hay, 1996; Wikström & Kautsky, 2007). In addition, macroalgae habitats are suggested to be important for the survival of juvenile fish (Carr, 1994; Keats et al., 1987; Sundblad et al., 2014) which make them contribute to the linkage between the coast and open waters through ontogenetic movements and habitat shifts. Habitats where macroalgae, such as kelp, have been removed mechanically by humans have shown slow recovery (Lilley & Schiel, 2006) and much lower abundances of juvenile fish (Lorentsen et al., 2010) and thus indicating a possibility of trophic cascades in the coastal seascape following a disturbance in macroalgae habitats. The importance of canopy forming macroalgae for fish in coastal seascapes have been studied worldwide, however, its importance for fish in the Baltic Sea is less clear. The Baltic Sea is a temperate, brackish inland sea in the northern hemisphere and in comparison to other inland seas, it has a low diversity of species (Ojaveer et al., 2010). This, combined with low connectivity to the North Sea (Hiddink & Coleby, 2012), makes the Baltic Sea rather unique. One major factor that affects the seascape within the Baltic Sea is its salinity gradient. The Baltic Sea has a salinity range from > 30 PSU (practical salinity units) in the south to < 5 PSU in the far north (Hiddink & Coleby, 2012), which affects the distribution of inhabiting species. Hence, species in the south might not be able to survive in the north and vice versa. Examples are the two fleshy macroalgae toothed wrack (Fucus serratus) and bladderwrack (Fucus vesiculosus) which have different distribution pattern along the coast of the Baltic Sea. Fucus serratus can only be found in the south part of the Baltic Proper (Malm et al., 2001) while F. vesiculosus stretches from the Baltic Proper to the Gulf of Bothnia, with the majority of the population in the Baltic Proper and Bothnian Sea (Torn et al., 2006; Wikström & Kautsky, 2007). Within the Baltic Proper, F. vesiculosus is one of the larger macroalgae found on hard substratum creating a complex structural habitat (e.g. Kalvas & Kautsky, 1993; Kautsky et al., 1992). It is acknowledged for being a key canopy-forming species along the coastlines, creating important habitats for organisms such as invertebrates (Nilsson et al, 2004; Wikström & Kautsky, 2007). The soft canopy complexity in fleshy macroalgae, has been found to have a positive effect on fish abundance, fish biomass and diversity of marine fauna on a local scale (Levin & Hay, 1996; van Lier et al., 2018). Lilley & Schiel (2006) found a decline in the epifaunal community when coverage of macroalgae canopy was lost, indicating the importance of complexity within these kinds of habitats. High complexity combined with a wide distribution of F. vesiculosus within the Baltic Sea may therefore create several niches where different fauna, including fish, may thrive and co-exist. The importance of F. vesiculosus for coastal fish communities in the Baltic Sea is, however, less clear. Compared to seagrass meadows, which often are portraited as a crucial habitat for juvenile fish communities (e.g. Perry et al., 2018; Staveley et al., 2016), information on the relationship between macroalgae,

2

fish abundance and species richness is scarce. There are limited studies available on the importance of F. vesiculosus for fish in the seascape on the Swedish west coast (e.g. Pihl & Wennhage, 2002; Pihl et al., 1994), however very few exist on the importance of F. vesiculosus for fish assemblages in the Baltic Sea (Jansson et al., 1985). Even though the Baltic Sea is one of the most studied seas, it is still unclear how it is affected by disturbances. There was a decline of F. vesiculosus during 1960-1980 due to eutrophication (Schramm, 1996), anthropogenic pollution (Kautsky et al., 1992) and to some extent grazing of herbivores, specifically Baltic isopod (Idotea baltica) (Nilsson et al., 2004). However, F. vesiculosus has recovered and at depths where it was previously scarce, re-established macroalgae has been observed along the east coast (Karlsson et al., 2009). Apart from anthropogenic disturbances, there are also the increasing threats from the effects of climate change. It is hypothesized that the Baltic Sea will experience a decline in salinity, which in turn could affect the distribution of marine species (Hiddink & Coleby, 2012). Also, the risk of increased temperature could in turn affect the reproduction of F. vesiculosus (Graiff et al., 2017), possibly changing its ecological functions. Sundblad et al. (2014) highlighted that F. vesiculosus might be crucial as nursery habitat for specific predatory fishes and consequently reductions/disappearance in cover could trigger a negative cascade. Thus, if F. vesiculosus will decline again, it is unsure whether it would be able to recover due to the future disturbances by climate change. In order to understand the importance of F. vesiculosus for coastal fish communities in the Baltic Sea and the possible consequences of its disappearance in the light of climate change, the following hypothesis were tested:

(1) Fish communities in habitats with F. vesiculosus have higher fish abundances, biomass, species richness, and more juveniles than habitats without F. vesiculosus,

(2) There is a positive relationship between environmental variables such as macroalgae cover and hard structural complexity and fish density, abundance, biomass and species richness within F. vesiculosus,

(3) There is a difference between abundance and species richness of invertebrates (fish food/prey) among locations with F. vesiculosus,

(4) Fish behaviour differs among sites where F. vesiculosus is present or not, (5) There is no difference in results between the three different methods used in the study

(beach seine, visual survey and remote underwater videos (RUVs)).

Material and methods Study sites and field sampling The survey was performed in the archipelago outside Trosa, Sweden, around the island of Askö (58°49′0″N 17°39′0″E). Three locations (Husbåtsviken, Knabberskär and Vrångskär) were sampled during 10-17th August in 2018. A follow-up was conducted during 10-11th September the same year, including an additional fourth site named Knekten. The locations were chosen because of different wave exposure and had sites where Fucus vesiculosus was present and not.

3

Table 1. Summary of locations per used method. Sites cleared from F. vesiculosus have had the macroalgae mechanically removed, while sites clean from F. vesiculosus are sites where the macroalgae were naturally absent. Numbers represent the quantity of sites with or without F. vesiculosus within the location. S = sheltered site, E = exposed site and VE = very exposed site.

Husbåtsviken was the least exposed location and there were sites both cleared from F. vesiculosus as well as sites with natural occurrence of F. vesiculosus. Knabberskär was more exposed and had sites with naturally occurring F. vesiculosus, some sites where F. vesiculosus was naturally absent and sites where F. vesiculosus was removed. Knekten had fewer sites in total, all with F. vesiculosus naturally occurring, while Vrångskär, the most exposed location, had 1 site where F. vesiculosus was naturally absent and 5 sites with F. vesiculosus naturally occurring (Table 1). Husbåtsviken and Knabberskär had all 3 different types of surveys conducted: underwater visual census (UVCs, see Berkström et al., 2013 and references therein), remote underwater videos (RUVs) and beach seine netting, while Knekten and Vrångskär only had UVCs and RUVs conducted. The depth of chosen sites ranged from 0.4-1.6 m.

Method Samples (n) Husbåtsviken Knabberskär Knekten Vrångskär

Underwater visual point census

Sites with F. vesiculosus

18 4S + 2E 5S

3S 2E + 3VE

Sites cleared from F. vesiculosus

14 6S + 3E 4S +1E

NA NA

Sites clean of F. vesiculosus

6 NA 3S + 2E

NA 1VE

Beach seine net

Sites with F. vesiculosus

16 5S + 3E 6S + 2E

NA NA

Sites cleared from F. vesiculosus

8 2S + 1E 3S + 2E

NA NA

Sites clean of F. vesiculosus

NA NA NA NA NA

Remote underwater video

Sites with F. vesiculosus

21 3S + 5E 6S 3S

3E + 1VE

Sites cleared from F. vesiculosus

9 4S + 2E 2S + 1VE

NA NA

Sites clean of F. vesiculosus

5 NA 1E + 3VE NA 1VE

4

Figure 1. Map of the Askö area and study locations. (A) Askö area where arrow 1-4 show the locations. Map 1-4 display each location and its sampling sites, which are marked by black points.

Water temperature was noted at all locations and varied between 17°C and 23°C. In addition to the surveys, individual fronds of F. vesiculosus were collected at Husbåtsviken and Knabberskär in order to investigate the invertebrate communities as potential fish food (abundance and species richness). The fronds collected from Husbåtsviken were from a sheltered habitat, while fronds from Knabberskär were from a more exposed habitat. Each frond was put in a marked container and transported to the Askö laboratory for further analysis where they were shaken and rinsed with fresh water above a tray to collect the epifauna. Each F. vesiculosus frond was weighed (wet weight), invertebrates were identified to lowest taxonomical level possible (usually family), counted and weighed (wet weight).

Underwater visual point census (UVCs) Underwater visual point census was conducted at all locations, by snorkelling, during both sampling occasions (August and September). In Husbåtsviken and Knabberskär, 15 UVCs were conducted at each location, 3 UVCs were conducted in Knekten and 5 UVCs were conducted in Vrångskär. Fish abundance was estimated based on a standardized method of counting fish within a point census (Bohnsack et al., 2005). The point census had a diameter of 2 meters, and the recorder was positioned either above the centre of the circle or to its side, depending on the visibility. The recorder would turn around slowly within the circle and document all fish (number, species and size) for 3 minutes, followed by swimming in a “floral” pattern back and forth to the centre to search for more cryptic individuals that might be hiding in macrophytes, cracks or under ledges. If needed, the recorder would stir or move the macrophytes to make sure that all fishes where observed and counted. For each census, notes were taken regarding depth, time, and exposure (sheltered, exposed or very exposed) as well as bottom substrate (flat rock, flat rocks with some boulders, rocks/boulders). The coverage of F. vesiculosus was estimated (%) in addition to hard bottom complexity, according to a 6-grade scale (0 = flat, 1 low relief: <10 cm, 2 low/moderate relief: 10-30 cm, 3 moderate relief: 30 – 60 cm, 4

A

5

high/moderate relief: 60 cm – 1 m, 5 high relief: > 1m) following Van Lier et al. (2018) and Wenger et al. (2018).

Remote Underwater Videos (RUVs) Remote underwater videos were performed by using GoPro cameras submerged within the four locations. Husbåtsviken, Knabberskär and Vrångskär were sampled both in August and September, while Knekten was only sampled in September. At each site, one square meter was marked with a line, and the depth of the cameras position was noted. Within every site, a random subsample from each video recording was observed for five minutes. 14 samples were collected in Husbåtsviken, 13 in Knabberskär, 3 in Knekten and 5 samples in Vrångskär. All fish within the square meter were counted and identified to species level and behaviour was noted as follows: eating (the mouth of the fish was in contact with a substrate, food or snatching), searching (the fish was swimming head down, hovering over the vegetation but not seemingly snapping at anything), travelling (constantly moving), hunting (actively following and attacking prey), resting (staying at the same spot without any major movements), hiding (swimming in to vegetation and not being seen again).

Beach seine netting Fish collection using a small beach seine net (6 x 2 m with 5 mm mesh size and a cod end with a diameter of 2.25 m and 2 mm mesh size) at sites with and without F. vesiculosus was done at Husbåtsviken and Knabberskär during the sampling in August, 11 hauls in Husbåtsviken and 13 hauls in Knabberskär. All fish that were caught in the net were transported in buckets with sea water to the Askö Laboratory. Fish were sedated by carefully transferring them to a bucket with MS22 solution (250 mg MS22 + 500 mg NAHCO3 per l1 seawater), and when no movements of individuals could be detected, each fish was decapitated. Samples were identified to species, measured in mm (total length) and a subsample of ¼ was weighed at Askö while the rest of the samples were frozen and transported to the Department of Ecology, Environment and Plant Sciences (DEEP) at Stockholm University where they were identified and measured.

Figure 2. Pictures of the beach seine net. Left: the beach seine net on land. Right: usage of beach seine net at a site with F. vesiculosus. Pictures taken by Nils Kautsky.

Data handling (and standardization) To enable comparison between the different methods, abundance per square meter (1m2) for each site was calculated for all data. Sample areas differed between the methods, where the area fished by the beach seine net differed from sample to sample and the area covered by UVCs where all the same (3.14 m2). Data from the RUVs were always 1 m2 and therefore did not need any standardization afterwards. All locations, except Knekten, were used in the location-specific analyses since Knekten had too few samples to make comparisons. However, when conducting analysis and graphs of the total data, Knekten was included.

6

In order to estimate biomass, calculations of wet weight per each individual fish was done by using the formula: W = a * Lb, where a and b is geometric species-specific mean values, a function of the fish shape based on height and thickness (Kulbicki et al., 2005), while L is the measured length (TL) for each individual. Species where a and b values could not be found were given the mean values from their family level. This included the fish species two-spotted goby (Gobiusculus flavescens), common goby (Pomatoschistus microps) and three-spined stickleback (Pungitius pungitius). Respective value of a and b where found at www.Fishbase.org. Biomass (g) per square meter was calculated. Biomass could only be calculated for UVCs and beach seine net data, since length of the individual fishes was not possible to determine from RUVs. Fishes from data gathered by UVCs and beach seine netting had individuals identified as adults or juveniles. This was not possible for the RUV data because of mentioned problems to estimate individual fish length. By using each species maximum length, each individual was classified as juvenile or adult (juvenile being 1/3 of the maximum length) following Nagelkerken & Velde (2002), where the maximum length of each species was taken from www.artfakta.artdatabanken.se. Within the two methods used, some fish individuals that either belonged to P. microps or sand goby (Pomatoschistus minutus) were identified as Pomatoschistus spp., since they could not be identified to species level. To determine the life stage of these individuals, a mean value of maximum length was calculated from the species-specific maximum values (P. microps = 9.0 cm and P. minutus = 11.0 cm).

Statistical analysis of data Linear mixed-effect models were conducted for each survey method using the packages “lme4” (Bates et al., 2013) and “nlme” (Bates et al., 2018), comparing abundances and biomasses of fish in different habitat types (presence and absence of F. vesiculosus) at the different locations (Husbåtsviken, Knabberskär and Vrångskär). Habitat was used as a fixed factor and location as random factor. Sites where F. vesiculosus was manually removed (cleared) were pooled together with sites naturally absent from F. vesiculosus (clean). Transformations using square-root or log10 were used when needed to achieve a better approximation of normality towards the mean. In addition to the presence of F. vesiculosus, exposure of the sites was added as a predictor variable within the linear mixed-effect models. Within each method, subsets of data were separated to enable ANOVAs on a location-specific level. If the models were statistically significant (p < 0.05), a post-hoc Tukey HSD was conducted. Relationships between complexity and coverage of F. vesiculosus and fish abundance and biomass were tested by linear regressions. A relationship between invertebrate abundance (richness and fish abundance were not possible to test due to low sample size. However, differences in invertebrate species richness and abundances between locations with different exposure was tested using a one-way ANOVA. Non-metrical multidimension scaling (nMDS) with Bray-Curtis dissimilarity from the package “vegan” (Oksanen et al., 2019) was conducted in order to explore the differences in fish communities between sites with and without F. vesiculosus. This was done for all methods pooled and for each method separately. To avoid rare species influencing the result, species that occurred ≤ 3 where removed from the analysis. To statistically test for differences, a Permutational Multivariate Analysis of Variance (PERMANOVA), Bray-Curtis dissimilarity (99 permutations) was done. In addition, similarity percentages tests (SIMPER) were conducted, identifying which fish species that influenced the community-differences the most. Both functions were taken from the package “vegan”, with the function of “adonis” in the PERMANOVA (Oksanen et al., 2019). All analysis was performed in R version 3.5.2.

7

Results In total, 4015 individual fishes were recorded within the four locations around Askö. 1224 fishes were recorded by UVCs, 2055 fishes by beach seine netting and 736 by RUVs. In total, 13 species were found, with three-spined stickleback (Gasterosteus aculeatus) and nine-spined stickleback (Pungitius pungitius) being the most abundant fish species and dominating in all locations except for Vrångskär, where species from the family Gobiidae such as two-spotted goby (Gobiusculus flavescens) had the highest abundances (Fig. 9). With all methods pooled, the total abundance and biomass of fish communities was higher in habitats with F. vesiculosus than without (Fig. 4 and Fig. 6). There was no significant difference in fish abundance and biomass between mechanically cleared and naturally cleared sites of F. vesiculosus and data was therefore pooled.

Fish communities There were significant differences in fish species composition between all locations, both when using UVCs (PERMANOVA, R2 = 0.28) and RUVs (PERMANOVA, R2 = 0.27) as well as all methods pooled (PERMANOVA, R2 = 0.18, Fig. 3). Additionally, fish species composition differed significantly between locations and between sites with and without F. vesiculosus when using the beach seine net method (PERMANOVA, R2

Location = 0.223, R2Habitat = 0.203, Fig 3c).

In the pooled data for all methods, Husbåtsviken, Knabberskär and Knekten had more similar fish communities and formed a cluster while fish communities in Vrångskär seemed more different and scattered (Fig. 3a). The beach seine net method as well as RUVs showed a similar pattern but was weaker within and between the different locations and habitats. However, the UVCs had a more distinguished pattern, where habitats without F. vesiculosus were more clustered together than those habitats with F. vesiculosus (Fig. 3b). Knekten and Vrångskär were the locations within RUVs that had the most diverse fish communities. For all methods, separate and pooled, P. pungitius and G. aculeatus were the two species that influenced differences in the community composition the most between habitats (with and without F. vesiculosus). Within the pooled data as well as UVCs, G. aculeatus had the highest influence (pooled = 23%; UVC = 34%) while P. pungitius influenced the most (26%) within RUVs. Within the beach seine net data, P. pungitius and G. aculeatus together influenced more than 50% of the dissimilarities between habitats with and without F. vesiculosus.

8

Figure 3. Non-metrical multidimension scaling (nMDS, Bray-Curtis) of fish species communities for each method used, distance based on location and habitat with or without F. vesiculosus. (A) Pooled data (stress = 0.20) from the three methods; (B) data from UVCs (stress = 0.18, ntotal = 37); (C) data from beach seine nets (stress = 0.18, ntotal = 29); (D) data from RUVs (stress = 0.14, ntotal = 28).

Fish abundance and biomass With all methods and locations pooled, fish abundance was significantly higher in sites with F. vesiculosus compared to sites without F. vesiculosus (F = 8.29, p = 0.005, Fig. 4). When separating methods and locations, patterns of fish abundance were similar to the pooled data (Fig. 5). Data separated by location but with all methods pooled, Knabberskär (F = 4.15, p = 0.046) and Vrångskär (F = 10.5, p = 0.012) had significantly higher abundances of fish in sites with F. vesiculosus than sites without (Fig. 5a). UVCs had no significant differences, but the fish abundance was higher in sites with F. vesiculosus than without in Husbåtsviken and Vrångskär, while Knabberskär had higher fish abundances in sites without F. vesiculosus (Fig. 5). There were no significant differences in fish abundance in beach seine samples, regardless of location or sites with F. vesiculosus. Similarly, there were no significant difference in fish abundance in RUV samples between sites with or without F. vesiculosus, but Knabberskär had slightly higher fish abundance at F. vesiculosus sites. Husbåtsviken had the over-all highest fish abundances in both habitats with and without F. vesiculosus.

9

Figure 4. Fish abundance (median and quartile) per square meter for all methods and locations pooled in habitats with and without F. vesiculosus. Outliers in Knabberskär (value 183) and Husbåtsviken (value 102) was discarded due to y-axis limits. Boxes show median (thick line), the 25th and 75th percentile and error bars of 95% confidence interval.

Figure 5. Fish abundance (median and quartile) per square meter in habitats with and without F. vesiculosus from all locations. (A) pooled data all methods; (B) underwater visual census; (C) beach seine nets and (D) remote underwater videos. Site “Knekten” is excluded from the graph due to its samples being in only one kind of habitat (F. vesiculosus). Boxes show median (thick line), the 25th and 75th percentile and error bars of 95% confidence interval.

10

Overall fish biomass (methods and locations pooled) was significantly higher in sites with F. vesiculosus (F = 8.19, p = 0.005, Fig. 6). When separated by locations but pooled methods, Knabberskär had a significantly higher biomass in sites with F. vesiculosus than sites without (F = 5.53, p = 0.023, Fig. 7a). Husbåtsviken (F = 6.36, p = 0.025) and Knabberskär (F = 12.8, p = 0.002) had significantly higher fish biomass in sites with F. vesiculosus than sites without in beach seine samples. Additionally, Husbåtsviken had significantly higher fish biomass than Knabberskär (F = 7.97, p = 0.009). There were no significant differences in fish biomass between sites with and without F. vesiculosus in UVCs, but fish biomass was slightly higher in habitats with F. vesiculosus (Fig. 7c). Overall, Husbåtsviken had the highest fish biomass irrespective of methods and sites with and without F. vesiculosus.

Figure 6. Total fish biomass (median and quartile) per gram (wet weight) with beach seine nets, underwater visual census and locations pooled, in habitats with or without F. vesiculosus. Outliers (values 28, 54) were discarded due to y axis limits. Boxes show median (thick line), the 25th and 75th percentile and error bars of 95% confidence interval.

Figure 7. Fish biomass (median and quartile) per gram (wet weight) for beach seine nets and underwater visual census. (A) Methods pooled per location; (B) beach seine nets and (C) UVCs. Knekten is excluded from A and C due to it only having samples from the same habitat (with F. vesiculosus present) and sampling at Vrångskär were only conducted with UVCs. Boxes show median (thick line), the 25th and 75th percentile and error bars of 95% confidence interval.

11

There was no significant relationship between fish abundance and hard structural complexity or coverage of F. vesiculosus in either kind of habitat (presence or absence of F. vesiculosus). There were no significant differences in fish abundance between sites with different exposure, but all locations had highest fish abundance in sites with F. vesiculosus with some variation amongst exposure (Fig. A1, Appendix). Via RUVs, fish behaviour was surveyed but no location had significantly different behaviour between habitats with or without F. vesiculosus (Fig. A2, Appendix). However, fishes that were travelling and searching for food were more common, but not significantly higher, in sites without F. vesiculosus while fish feeding were more common in sites with F. vesiculosus. In addition, Husbåtsviken had fish resting in sites without F. vesiculosus, a behaviour not observed in Knabberskär, while at sites with F. vesiculosus in Knabberskär, all behaviours were observed.

Species richness Knabberskär had a significantly higher species richness in sites with F. vesiculosus (F = 7.54, p = 0.006), otherwise no significant difference in species richness among sites with and without F. vesiculosus was found (all methods pooled, Fig. 8). Husbåtsviken had a higher, but not significant, species richness in sites without F. vesiculosus, compared to Knabberskär and Vrångskär that had higher species richness in sites with F. vesiculosus (Fig. 8). Gasterosteus aculeatus and Pungitius pungitius were the most abundant species within Husbåtsviken and Knabberskär, irrespective of F. vesiculosus being present or not, and Vrångskär had the only sightings of Gobiusculus flavescens in sites with F. vesiculosus (Fig. 9).

Figure 8. Species richness (median and quartile) per location where F. vesiculosus either is present or not, all methods pooled. Note: Vrångskär only has one sample in a site without F. vesiculosus. Boxes show median (thick line), the 25th and 75th percentile and error bars of 95% confidence interval.

12

Figure 9. Species abundance (median and quartile) per square m

eter in (A) sites w

ith F. vesiculosus and (B) sites

without F. vesiculosus. The location of K

nekten is not included in the graph due to sampling only being conducted in

one habitat type (with F. vesiculosus present).

13

F. vesiculosus as juvenile habitat Abundance of adult fish was higher than juvenile fish, regardless of location or site, while juvenile fish abundance was higher in sites without F. vesiculosus than in sites with F. vesiculosus (Fig. 10). Knekten and Vrångskär was excluded because of too few samples (nKnekten = 3 and nVrångskär = 6). There were no significant differences between adult and juvenile fish abundances, when separated into species, in sites with or without F. vesiculosus in either location (Fig. A3, Appendix).

Figure 10. Abundance (median and quartile) of adult and juvenile fish in habitats with and without F. vesiculosus at (A) Husbåtsviken and (B) Knabberskär.

Habitat and associated invertebrates Mean coverage of F. vesiculosus was higher in the more exposed site Knabberskär than the more sheltered site Husbåtsviken (75% and 45% respectively), as well as the mean number of holdfasts (13.6 and 6.8 respectively) per m2. Mean filamentous algae cover was higher in Husbåtsviken than Knabberskär (75% and 61% respectively), in sites without F. vesiculosus (table 2). Husbåtsviken also had a higher mean height of F. vesiculosus plants compared to Knabberskär. Invertebrate species richness was significantly higher in Husbåtsviken than in Knabberskär (F = 12.7, p = 0.004, Fig. 11). Mean invertebrate abundance (per gram wet weight of macroalgae) was twice as high in Husbåtsviken than in Knabberskär (32.2 ± 0.35, 14.9 ± 0.81 respectively, Fig. 11). Marine bivalves (Cerastoderma/Cardium sp.) and the river nerite (Theodoxus fluviatilis) were the most common species. The locations differed in invertebrate species richness, where Nementa spp. was unique to only exposed sites, while sheltered sites had 3 unique species (Table A1, Appendix).

14

Table 2. Data from habitat quadrants of m2 from Husbåtsviken and Knabberskär. Coverage = estimated cover of F. vesiculosus within the quadrant; filamentous algae cover = coverage of filamentous algae on F. vesiculosus in sites with F. vesiculosus and coverage of filamentous alga on rock in sites without F. vesiculosus; height = height in cm of F. vesiculosus fronds and no. of holdfasts = no. of root structures (holdfasts) found in the quadrant.

Figure 11. Mean number of invertebrates per gram wet weight of F. vesiculosus grouped by species gathered from F. vesiculosus fronds in Husbåtsviken and Knabberskär.

Husbåtsviken

Mean ± SE

Knabberskär

Mean ± SE

Habitat with F. vesiculosus

Coverage (%) 45 ± 7.4 75 ± 6.8

Filamentous algae cover (%) 54 ± 8.8 17 ± 5.3

Height (cm) 37 ± 2.4 36 ± 5.8

Number of holdfasts 6.8 ± 0.9 13.6 ± 0.7 Habitat without F. vesiculosus

Coverage (%) 4 ± 1.2 3 ± 1.1

Filamentous algae cover (%) 75 ± 7.3 61 ± 10

Height (cm) NA NA

Number of holdfasts 5.8 ± 1.3 8.6 ± 3.2

15

Table 3. Invertebrates gathered from F. vesiculosus fronds in Knabberskär and Husbåtsviken. The table does not include the fish species that were collected during the sampling. Species from the Radix and Palaemon families are grouped within their respective families. Insect larvae = mosquito and dragonfly larvae.

Location Species Number of individuals

Mean abundance (wet weight (g)) ± SE

Knabberskär Balanus improvisus 6 0.04 ± 0.003 Bithynia tentaculata 0 NA Cerastoderma 2013 8.13 ± 0.8 Dendrocoelum lacteum 5 0.02 ± 0.002 Gammarus sp. 74 0.33 ± 0.02 Hydrobia sp. 22 0.06 ± 0.004 Idotea sp. 129 0.61 ± 0.04 Insect larvae 4 0.008 ± 0.0007 Jaera sp. 0 NA Mytilus edulis 100 0.19 ± 0.02 Nementa spp. 1 0.005 ± 0.0008 Palaemon sp. 35 0.13 ± 0.007 Radix spp. 92 0.46 ± 0.03 Theodoxus fluviatilis 1186 4.98 ± 0.27 Total 3667 14.9 Husbåtsviken Balanus improvisus 85 0.24 ± 0.01 Bithynia tentaculata 14 0.04 ± 0.006 Cerastoderma 6532 21.0 ± 1.66 Dendrocoelum lacteum 20 0.04 ± 0.002 Gammarus sp. 111 0.27 ± 0.01 Hydrobia sp. 410 1.47 ± 0.19 Idotea sp. 351 0.95 ± 0.05 Insect larvae 32 0.08 ± 0.005 Jaera sp. 7 0.09 ± 0.001 Mytilus edulis 34 0.10 ± 0.004 Nementa spp. 0 NA Palaemon sp. 67 0.14 ± 0.005 Radix spp. 92 0.78 ± 0.05 Theodoxus fluviatilis 2810 6.96 ± 0.2 Total 10785 32.2

16

Discussion As hypothesised, fish abundance and biomass, and species richness were higher in sites with F. vesiculosus than those without. Results from the beach seine netting displayed a significant difference in community composition between sites with and without F. vesiculosus. However, contradictory to the hypotheses, the abundance of juvenile fish was not significantly higher in sites with F. vesiculosus compared to sites without and although it was expected that environmental factors, such as F. vesiculosus coverage and complexity, would affect the fish abundance and biomass, no such relationship was found. Although it was not possible to test the relationship between the presence of invertebrates and fish abundance due to limited sample size, there was a significant difference in the number of invertebrates between the more sheltered Husbåtsviken and the exposed location Knabberskär, which may affect fish abundance, biomass and behaviour. The number of fishes spending time eating was higher in sites with F. vesiculosus in Husbåtsviken than in Knabberskär, where the number of invertebrates was half of that in Husbåtsviken. Furthermore, a higher number of fishes were found resting in sites without F. vesiculosus in Husbåtsviken, suggesting that fish behaviours somewhat differ between sites with and without F. vesiculosus. There were no significant differences among methods, however beach seine samples had higher abundance of fish per m2 than UVCs and RUVs combined.

Fish community differences in areas with and without F. vesiculosus All locations had a higher fish abundance in F. vesiculosus sites and overall, fish biomass was significantly higher in sites with F. vesiculosus. Therefore, this study suggests that canopy forming macrophytes, like F. vesiculosus, are important habitat for some fish communities within the Baltic Sea. Sites that had F. vesiculosus mechanically removed during the study were rather small patches within belts of F. vesiculosus, which could explain why the fish abundance did not differ significantly between those two habitats. Hence, for future studies, separated sites with a larger area without F. vesiculosus surrounded by smaller F. vesiculosus sites, might give a more clear result between sites with and without F. vesiculosus, as in the study by Perry et al. (2018) where macrophyte habitats (i.e. seaweed and seagrass) had higher fish abundances than those without. Likewise, an increase in the number of samples per location could give clearer results. Within locations, Husbåtsviken had the highest fish abundance and biomass in sites with F. vesiculosus. Knabberskär (UVCs) and Husbåtsviken (beach seine netting) however, had higher fish abundances in sites without F. vesiculosus, in contrast to fish biomass that was higher in sites with F. vesiculosus regardless of location or method. There might be a higher abundance of larger, but fewer, fish in sites with F. vesiculosus where the fish biomass exceeds the abundance. This in turn could depend on the availability of food (Wikström & Kautsky, 2007) and shelter (Kraufvelin & Salovius, 2004) in F. vesiculosus sites. The overall high fish abundance and biomass within Husbåtsviken could also depend on high production, where a lower exposure enabled a higher coverage of filamentous green algae compared to Knabberskär, where small invertebrates (fish food) may feed and hide. Indeed, higher abundances and species richness of invertebrates were found in the sheltered Husbåtsviken. Fish species richness was also higher in sites with F. vesiculosus in Knabberskär and Vrångskär, while Husbåtsviken had higher fish species richness in sites without F. vesiculosus. Overall species richness was rather low in the locations, but Husbåtsviken and Knabberskär had the highest abundances of G. aculeatus and P. pungitius. Byström et al. (2015) highlighted that a higher abundance of G. aculeatus resulted in a lower abundance of common perch (Perca

17

fluviatilis), where predation on P. fluviatilis eggs and larvae was believed to be the reason (Bergström et al., 2015; Byström et al., 2015). Therefore, since both P. fluviatilis and G. aculeatus prefer shallow vegetated habitats, such a relationship between mesopredators and piscivores could explain the very low abundance and single appearance of P. fluviatilis during this study. In comparison, the highest fish abundance in Vrångskär consisted of species from the family Gobiidae, both in sites with and without F. vesiculosus. Gobiusculus flavescens, a less benthic Gobiidae species, had the highest abundance within F. vesiculosus sites while benthic Pomatoschistus spp. were the species found in sites without F. vesiculosus. This could be because of Vrångskär was the most exposed location, restricting some species of fish.

Fish community differences among locations Throughout the study, the most protected location, Husbåtsviken had the highest variance in fish abundance, biomass and richness. Similarly, Mustamäk et al. (2015) found that in a protected basin, without internal physical barriers, fish abundance and species composition could vary significantly over shorter distances. In contrast, Vrångskär had the highest exposure and the lowest fish abundance. Studies have found F. vesiculosus to be sensitive to wave exposure, which affect the morphology by making the macroalgae smaller and less branched (Kalvas & Kautsky, 1993; Kautsky et al., 1992), however this was not visually noticeable at Vrångskär. Perhaps the waves restrict some fish species swimming ability, which could explain the low fish abundance both within sites with and without F. vesiculosus at Vrångskär. Vahteri et al. (2009) suggested that fish assemblages within the Baltic Sea differ depending on spatial zones, where fish communities consisting of Gasterosteidae and Cyprinidae preferred outer areas in the archipelago where F. vesiculosus, amongst other vegetation, is present. This may suggest that fish communities within the present study might depend more on the actual placement of habitats of F. vesiculosus than on the macroalgae itself. It could also explain the lack of a relationship between fish abundance and macroalgae cover and hard structural complexity in the present study. The lack of a relationship may, however, also be due to low sample size since data from UVCs could not be used.

F. vesiculosus as juvenile habitat Macroalgae habitats are often considered to act as nurseries for juvenile fish (e.g. Carr, 1994; Heck Jr et al., 2003; Kraufvelin et al., 2018), however this was not supported by the present study since juvenile fish abundances were higher in sites without F. vesiculosus than sites with F. vesiculosus. Contradictory to other reports (e.g. Anderson 1994), the present study found that adult fish abundances were higher than those of juvenile fish abundances regardless of the presence or absence of F. vesiculosus. Although Gasterosteus aculeatus was the most abundant species and 83% of this species was classified as juveniles, no differences in juvenile abundance between sites with and without F. vesiculosus was found. However, there was an uneven distribution of juvenile G. aculeatus within Husbåtsviken and Knabberskär which might contribute to the non-significant result. Seasonal and ontogenetic movements by fishes is an important process within the seascape, and studies have shown that G. aculeatus moves into shallow, coastal areas during early summer to spawn (Bergström et al., 2015), where F. vesiculosus might act as a possible nursery. In addition to large-scale variation, it is important to also include the variation on a small-scale, or some fish species might be missed (Mustamäki et al., 2015). According to Staveley et al. (in press), Z. marina did not play a major role as a habitat for fish communities within the Baltic Sea, and suggested that it was just one of many habitats, linked by fish movements (Perry et al., 2018), where F. vesiculosus might be one of them. Also, the height of macroalgae has been found to have a positive effect on juvenile fish abundances in both tropical (Eggertsen et al., 2017) and

18

temperate seascapes (Anderson, 1994), but unfortunately this could not be tested within the present study and should therefore be included in future studies.

Food availability and fish behaviour Habitats with macroalgae such as F. vesiculosus offer a high structural complexity which can increase the abundance of invertebrates (Eggertsen et al., 2017; Kautsky et al., 1992; Wikström & Kautsky, 2007). Husbåtsviken had 3 times higher coverage (%) of filamentous algae on F. vesiculosus plants and 2 times higher invertebrate abundance (mean no. of invertebrates per gram wet weight F. vesiculosus) than Knabberskär. Invertebrates likely prefer higher coverage of filamentous algae due to food and shelter access. Indeed, Theodoxus fluviatilis displayed a very high abundance at Husbåtsviken and since they are not known to consume F. vesiculosus (Kautsky et al., 1992) but prefer epiphytic microalgae (Wikström & Kautsky, 2007), this may explain observed patterns. Wikström & Kautsky (2007) found that the invertebrate biomass was half the amount in habitats without F. vesiculosus, compared to those with F. vesiculosus, suggesting that these macrophyte habitats supply fish with high amounts of food. Significant higher species richness of invertebrates was also found in sheltered sites with F. vesiculosus compared to exposed sites. However, Kraufvelin & Salovius (2004) concluded that macrofauna abundance and richness within F. vesiculosus habitats are equal, or even lower, compared to green algal turf such as Cladophora. It was also discussed that there would be no loss of either invertebrates richness or abundance if F. vesiculosus would disappear (Kraufvelin & Salovius, 2004). Wikström & Kautsky (2007) on the other hand implied that the loss of F. vesiculosus could not be replaced by Cladophora, in relation to the canopy loss of F. vesiculosus and the effects on faunal communities. This study cannot be fully compared to these studies but the loss of F. vesiculosus could alter the fish communities and their abundance since invertebrates often are presented as fish food. The behaviour of fish varied depending on F. vesiculosus being present or not, where behaviours were more diverse in sites with F. vesiculosus than sites without. Husbåtsviken had more fish feeding, in total and in sites with F. vesiculosus, compared to Knabberskär, but the proportion of resting fish was higher in sites without F. vesiculosus. This may be due to Husbåtsviken being a sheltered location, or because fish within F. vesiculosus sites are more difficult to see through RUVs and therefore not recorded when motionless. Knabberskär, on the other hand, did not have any resting fish in its sites without F. vesiculosus while sites with F. vesiculosus had all behaviours recorded, hence implying a greater difference between habitats than in Husbåtsviken and possibly linked to the higher exposure in Knabberskär. Also, Helfman (1986) highlighted that most fish species divide their behaviour throughout the day by active food searching or an inactive phase avoiding predators, suggesting that differences of fish behaviour in sites without F. vesiculosus in Knabberskär could depend on a higher predation risk or exposure. General behavioural differences between Husbåtsviken and Knabberskär could also be because of large amounts of epiphytes on F. vesiculosus that protects smaller fish from predators.

Comparison between sampling methods The three different sampling methods provided similar results for fish abundance, but not for fish biomass. Biomass of fish was only calculated for UVCs and beach seine netting and may have skewed results. The three methods have quite different approaches but are all used within marine science (e.g. Bohnsack et al., 2005; Vahteri et al., 2009). UVCs is a non-destructive method where the recorder can gather information on habitat and fish communities, contributing to a wider understanding of the ecosystem. However, it can be difficult to see and record all present fish depending on the structure and complexity of the

19

habitat (Bohnsack et al., 2005) and the recorder may scare fish by its presence, underestimating the number of fish. RUV on the other hand, is a less active method which usually do not distract or disturb the fish. It can be applied on a larger scale, give more of an overview as well as being methodically examined afterwards. However, depending on if the fish of interest are pelagic, benthic or both it is important to know in order to adequately construct and place the camera. Cameras in this study where angled straight forward, giving a good overview of part of the benthos but mainly open water, meaning that many benthic fish species such as Gobiidae spp. probably were missed (Gobiusculus flavescens, a pelagic species, being an exception (Lyons, 2012). An easy way to insure good results from RUVs is to use high-resolution cameras, with perhaps the ability of measurements so also biomass can be conducted from such a method. Following Perry et al. (2018), stereo video could be utilized, where 2 cameras are mounted on a calibrated frame and synchronized to give a simultaneously recording of the same objects. This should then be monitored, using an analysis software, so variables such as fish size could be estimated. Perhaps one of the most common methods used in similar studies is beach seine netting (e.g. Byström et al., 2015; Gagnon et al., 2019; Staveley et al., 2017), where also habitat variables can be recorded. Depending on the mesh size, different fish species can be targeted. Something to keep in mind is the possible chance of destruction of the area where the net is dragged, and it is probably the most disturbing method for fish, which might affect the catches. However, by using a beach seine net it is easier to measure lengths, biomass, and identify species of caught fish. The sampling within this study was only conducted during the day, which exclude the fishes’ nocturnal behaviour. Both within coral reefs, temperate marine kelp-forests and lakes, diurnal fish species lower their activity during the night, seeking shelter, and nocturnal fish species becomes active (Helfman, 1986). If a nocturnal point census, for example, would be implemented in future sampling there might be a larger distinction between the behaviours of the fish than during the day which in turn could give more information about the fishes use of F. vesiculosus and how the macroalgae support fish communities (Azzurro et al., 2007; Helfman, 1986). All three methods have both pros and cons, where the most important part lies in choosing the most appropriate method for the specific study. For this study, UVCs might have been the most suitable method due to its ability to record fish and habitat at all sites with low impact on the environment.

Future perspectives and management Based on the present results, F. vesiculosus should be included within the management of the Baltic Sea. By protecting areas within the coastal seascape where the macroalgae is present, it may improve the survival of fish communities in the future. F. vesiculosus needs to be further acknowledged as an important habitat and protected, for example through marine protected areas (MPAs). However, as implied by Magris et al. (2014), a MPA needs to be quantified with objectives that are grounded in ecological knowledge, both for present and future threats. To keep in mind though, a MPA might not be enough protection for threats as the climate change. It is believed that with climate change, a higher water temperature within oceans may shift organisms’ ranges, where warm species will expand and invade cold water species habitats. The chances of this happening within the Baltic Sea is unsure, but according to Hiddink & Coleby (2012) species expanding their distribution in relation to water salinity is more likely and species ranges are believed to increase. Thus, even if F. vesiculosus show a relatively high tolerance towards lower salinity (Takolander et al., 2017), there are few replacements for higher taxa (Vuorinen et al., 2015) and if F. vesiculosus would decrease, it could be assumed that fish communities would be negatively affected.

20

Even though the present study did not find F. vesiculosus being a fish nursery habitat, it might be because the linkage amongst coastal areas are more important than one individual habitat as suggested by Perry et al., (2018). Therefore, it is vital to conduct more studies on how changes in F. vesiculosus would alter its interactions with other habitats. Additional fish sampling among different seasons may provide better insights of how coastal fish move between habitats including F. vesiculosus. One suggestion for gathering and evaluating possible threats is to use thorough habitat mapping, where pressure variables are sampled (Kraufvelin et al., 2018). Due to future climate changes and the possible threats towards the Baltic Sea ecosystem (e.g. Hiddink & Coleby, 2012; Rugiu et al., 2018; Takolander et al., 2017; Vuorinen et al., 2015), it is important to protect F. vesiculosus and its functions as fish habitat to sustain a stable ecosystem. By combining established knowledge with future studies about F. vesiculosus in the Baltic Sea, hopefully a useful and broad base will be available for better management of F. vesiculosus within of coastal seascapes.

Acknowledgements I would like to thank my supervisors, Charlotte Berkström, Maria Eggertsen and Nils Kautsky, for their support and help during this thesis. Charlotte for her exceptional inputs and reviews, Maria for the invaluable help with the statistical analyses and Nils for all his knowledge about Askö and chosen subject. I would also like to thank family and friends for the support and encouragement, especially Emma Lemos, Felicity Pike and Katarina Rylander, whom I have spent many hours sitting next to, discussing and writing.

References Anderson, T. W. (1994). Role of macroalgal structure in the distribution and abundance of a temperate

reef fish, 113(Womersley 1954), 279–290.

Azzurro, E., Pais, A., Consoli, P., & Andaloro, F. (2007). Evaluating day–night changes in shallow Mediterranean rocky reef fish assemblages by visual census. Marine Biology, 151(6), 2245–2253. https://doi.org/10.1007/s00227-007-0661-9

Banks, S. C., Piggott, M. P., Williamson, J. E., Bove, U., Holbrook, N. J., & Beheregaray, L. B. (2007). Oceanic variability and coastal topography shape genetic structure in a long-dispersing sea urchin. Ecology, 88(12), 3055–3064.

Bates, D. M., & Pinheiro, J. C. (2018). Linear and Nonlinear Mixed-Effects Models. Conference on Applied Statistics in Agriculture. https://doi.org/10.4148/2475-7772.1273

Bates, D., Maechler, M., & Bolker, B. (2013). lme4: Linear mixed-effects models using S4 classes. R Package Version 1.1-7. https://doi.org/citeulike-article-id:1080437

Bergström, U., Olsson, J., Casini, M., Eriksson, B. K., Fredriksson, R., Wennhage, H., & Appelberg, M. (2015). Stickleback increase in the Baltic Sea - A thorny issue for coastal predatory fish. Estuarine, Coastal and Shelf Science, 163(PB), 134–142. https://doi.org/10.1016/j.ecss.2015.06.017

21

Berkström, C., Gullström, M., Lindborg, R., Mwandya, A. W., Yahya, S. A. S., Kautsky, N., & Nyström, M. (2012). Exploring “knowns” and “unknowns” in tropical seascape connectivity with insights from East African coral reefs. Estuarine, Coastal and Shelf Science, 107, 1–21. https://doi.org/10.1016/j.ecss.2012.03.020

Berkström, C., Lindborg, R., Thyresson, M., & Gullström, M. (2013). Assessing connectivity in a tropical embayment: Fish migrations and seascape ecology. Biological Conservation, 166, 43–53. https://doi.org/10.1016/j.biocon.2013.06.013

Bohnsack, J. A., Amaral, A. C. Z., Jablonski, S., Bohnsack, J. A., Bannerot, S. P., Bolker, B. M., … Schlacher, T. a. (2005). A Stationary Visual Census Technique for Quantitatively Assessing Community Structure of Coral Reef Fishes. Megadiversidade, 23(2), 1–9. https://doi.org/10.1007/s10531-013-0611-4

Braithwaite, V. A., & Boulcott, P. (2007). Pain perception, aversion and fear in fish. DISEASES OF AQUATIC ORGANISMS Dis Aquat Org, 75, 131–138.

Byström, P., Bergström, U., Hjälten, A., Ståhl, S., Jonsson, D., & Olsson, J. (2015). Declining coastal piscivore populations in the Baltic Sea: Where and when do sticklebacks matter? AMBIO, 44(S3), 462–471. https://doi.org/10.1007/s13280-015-0665-5

Carr, M. H. (1994). Effects of Macroalgal Dynamics on Recruitment of a Temperate Reef Fish. Ecology, 75(5), 1320–1333.

Christie, H., Norderhaug, K. M., & Fredriksen, S. (2009). Macrophytes as habitat for fauna. Marine Ecology Progress Series, 396, 221–233. https://doi.org/10.3354/meps08351

Eggertsen, L., Ferreira, C. E. L., Fontoura, L., Kautsky, N., Gullström, M., & Berkström, C. (2017). Seaweed beds support more juvenile reef fish than seagrass beds in a south-western Atlantic tropical seascape. Estuarine, Coastal and Shelf Science, 196, 97–108. https://doi.org/10.1016/j.ecss.2017.06.041

Gagnon, K., Gräfnings, M., & Boström, C. (2019). Trophic role of the mesopredatory three-spined stickleback in habitats of varying complexity. Journal of Experimental Marine Biology and Ecology, 510(October 2018), 46–53. https://doi.org/10.1016/j.jembe.2018.10.003

Graiff, A., Dankworth, M., Wahl, M., Karsten, U., & Bartsch, I. (2017). Seasonal variations of Fucus vesiculosus fertility under ocean acidification and warming in the western Baltic Sea. Botanica Marina, 60(3), 239–255. https://doi.org/10.1515/bot-2016-0081

Heck Jr, K. L., Hays, G., & Orth, R. J. (2003). Critical evaluation of the nursery role hypotehsis for seagrass medows. MARINE ECOLOGY PROGRESS SERIES Mar Ecol Prog Ser, 253, 123–136. https://doi.org/10.1080/00396336108440237

Helfman, G. S. (1986). Fish Behaviour by Day, Night and Twilight. In The Behaviour of Teleost Fishes (pp. 366–387). Boston, MA: Springer US. https://doi.org/10.1007/978-1-4684-8261-4_14

Hiddink, J. G., & Coleby, C. (2012). What is the effect of climate change on marine fish biodiversity in an area of low connectivity, the Baltic Sea? Global Ecology and Biogeography, 21(6), 637–646. https://doi.org/10.1111/j.1466-8238.2011.00696.x

Jansson, B.-O., Aneer, G., & Nellbring, S. (1985). Spatial and temporal distribution of the demersal fish funa in a Baltic archipelago as estimated by SCUBA census. Marine, 23(1985).

Kalvas, A., & Kautsky, L. (1993). Geographical variation in fucus vesiculosus morphology in the baltic and north seas. European Journal of Phycology, 28(2), 85–91. https://doi.org/10.1080/09670269300650141

Karlsson, J., Tobiasson, S., & Kautsky, H. (2009). Havets växter berättar. Havet, 44–48.

Kautsky, H., Kautsky, L., Kautsky, N., Kautsky, U., & Lindblad, C. (1992). Studies on the Fucus vesiculosus community in the Baltic Sea. Acta Phytogeographica Suecica (Sweden), 78(JANUARY), 33–48. Retrieved from http://agris.fao.org/agris-

22

search/search.do?recordID=SE9211523

Keats, D. W., Steele, D. H., & South, G. R. (1987). The role of fleshy macroalgae in the ecology of juvenile cod (Gadus morhua L.) in inshore waters off eastern Newfoundland. Canadian Journal of Zoology, 65, 49–53. https://doi.org/10.1139/z87-008

Kraufvelin, P., Pekcan-Hekim, Z., Bergström, U., Florin, A. B., Lehikoinen, A., Mattila, J., … Olsson, J. (2018). Essential coastal habitats for fish in the Baltic Sea. Estuarine, Coastal and Shelf Science, 204, 14–30. https://doi.org/10.1016/j.ecss.2018.02.014

Kraufvelin, P., & Salovius, S. (2004). Animal diversity in Baltic rocky shore macroalgae: Can Cladophora glomerata compensate for lost Fucus vesiculosus? Estuarine, Coastal and Shelf Science, 61(2), 369–378. https://doi.org/10.1016/j.ecss.2004.06.006

Kulbicki, M., Guillemot, N., & Amand, M. (2005). A general approach to length-weight relationships for New Caledonian lagoon fishes. Cybium, 29(2), 235–252. Retrieved from http://sfi.mnhn.fr/cybium/numeros/2005/293/03-Kulbicki 282.pdf

Levin, P. S., & Hay, M. E. (1996). Responses of temperate reef fishes to alterations in algal structure and species composition. MARINE ECOLOGY PROGRESS SERIES, 134, 37–47. Retrieved from https://www.int-res.com/articles/meps/134/m134p037.pdf

Lilley, S. A., & Schiel, D. R. (2006). Community effects following the deletion of a habitat-forming alga from rocky marine shores. Oecologia, 148(4), 672–681. https://doi.org/10.1007/s00442-006-0411-6

Lorentsen, S. H., Sjøtun, K., & Grémillet, D. (2010). Multi-trophic consequences of kelp harvest. Biological Conservation, 143(9), 2054–2062. https://doi.org/10.1016/j.biocon.2010.05.013

Lyons, P. (2012). The Biology of Gobies . Edited by Robert A. Patzner, James L. Van Tassell, Marcelo Kovačić, and B. G. Kapoor. Enfield (New Hampshire): Science Publishers; distributed by CRC Press, Boca Raton (Florida). $139.95. xv + 685 p.; ill.; index. ISBN: 978-1-57808-436-4. 2011. . The Quarterly Review of Biology, 87(3), 271–271. https://doi.org/10.1086/666769

Magris, R. A., Pressey, R. L., Weeks, R., & Ban, N. C. (2014). Integrating connectivity and climate change into marine conservation planning. Biological Conservation, 170, 207–221. https://doi.org/10.1016/J.BIOCON.2013.12.032

Malm, T., Kautsky, L., & Engkvist, R. (2001). Reproduction, recruitment and geographical distribution of Fucus serratus L. in the Baltic sea. Botanica Marina, 44(2), 101–108. https://doi.org/10.1515/BOT.2001.014

Mustamäki, N., Jokinen, H., Scheinin, M., Bonsdorff, E., & Mattila, J. (2015). Seasonal small-scale variation in distribution among depth zones in a coastal Baltic Sea fish assemblage. ICES Journal of Marine Science, 72(8), 2374–2384. https://doi.org/10.1093/icesjms/fsv068

Nilsson, J., Engkvist, R., & Persson, L. E. (2004). Long-term decline and recent recovery of Fucus populations along the rocky shores of southeast Sweden, Baltic Sea. Aquatic Ecology, 38(4), 587–598. https://doi.org/10.1007/s10452-004-5665-7

Ojaveer, H., Jaanus, A., MacKenzie, B. R., Martin, G., Olenin, S., Radziejewska, T., … Zaiko, A. (2010). Status of Biodiversity in the Baltic Sea. PLoS One, 5(9), e12467. https://doi.org/10.1371/journal.pone.0012467

Oksanen, J., Blanchet, F. G., Friendly, M., Kindt, R., Legendre, P., Mcglinn, D., … Maintainer, H. W. (2019). Package “vegan” Title Community Ecology Package. Retrieved from https://cran.r-project.org/web/packages/vegan/vegan.pdf

Perry, D., Staveley, T. A. B., & Gullström, M. (2018). Habitat connectivity of fish in temperate shallow-water seascapes. Frontiers in Marine Science. https://doi.org/10.3389/fmars.2017.00440

Pihl, L., & Wennhage, H. (2002). Structure and diversity of fish assemblages on rocky and soft bottom shores on the Swedish west coast. Journal of Fish Biology, 61(SUPPL. A), 148–166.

23

https://doi.org/10.1006/jfbi.2002.2074

Pihl, Leif, Wennhage, H., & Nilsson, S. (1994). Fish assemblage structure in relation to macrophytes and filamentous epiphytes in shallow non-tidal rocky- and soft-bottom habitats. Environmental Biology of Fishes. https://doi.org/10.1007/BF00005129

Rugiu, L., Manninen, I., Rothäusler, E., & Jormalainen, V. (2018). Tolerance and potential for adaptation of a Baltic Sea rockweed under predicted climate change conditions. Marine Environmental Research, 134, 76–84. https://doi.org/10.1016/j.marenvres.2017.12.016

Schramm, W. (1996). The Baltic Sea and its transition zones. Marine Benthic Vegetation - Recent Changes and the Effects of Eutrophication, 123, 131–164.

Staveley, T. A. B., Perry, D., Lindborg, R., & Gullström, M. (2016). Seascape structure and complexity influence temperate seagrass fish assemblage composition. Ecography, 39(8), 001–011. https://doi.org/10.1111/ecog.02745

Staveley, T., Gullström, M., Björk, M., Lindborg, R., Jackson, E., & Stockholms universitet Naturvetenskapliga fakulteten. (n.d.). Fish in the coastal seascape exploring ecological processes and connectivity for conservation of temperate fish communities.

Steneck, R. S., Bourque, B. J., Tegner, M. J., Erlandson, J. M., Corbett, D., Estes, J. A., & Graham, M. H. (2002). Kelp forest ecosystems: biodiversity, stability, resilience and future. Environmental Conservation, 29(04), 436–459. https://doi.org/10.1017/s0376892902000322

Sundblad, G., Bergström, U., Sandström, A., & Eklöv, P. (2014). Nursery habitat availiability limits adult stock sizes of predatory coastal fish. ICES Journal of Marine Science, 71(3), 528–536. https://doi.org/10.1093/icesjms/fst176

Takolander, A., Leskinen, E., & Cabeza, M. (2017). Synergistic effects of extreme temperature and low salinity on foundational macroalga Fucus vesiculosus in the northern Baltic Sea. Journal of Experimental Marine Biology and Ecology, 495, 110–118. https://doi.org/10.1016/j.jembe.2017.07.001

Torn, K., Krause-Jensen, D., & Martin, G. (2006). Present and past depth distribution of bladderwrack (Fucus vesiculosus) in the Baltic Sea. Aquatic Botany, 84(1), 53–62. https://doi.org/10.1016/j.aquabot.2005.07.011

Vahteri, P., O’Brien, K., & Vuorinen, I. (2009). Zonation and spatial distribution of littoral fish communities from the southwestern Finnish coast (Archipelago and Bothnian Sea, Northern Baltic Sea). Estuarine, Coastal and Shelf Science, 82(1), 35–40. https://doi.org/10.1016/J.ECSS.2008.11.016

van Lier, J. R., Wilson, S. K., Depczynski, M., Wenger, L. N., & Fulton, C. J. (2018). Habitat connectivity and complexity underpin fish community structure across a seascape of tropical macroalgae meadows. Landscape Ecology, 33(8), 1287–1300. https://doi.org/10.1007/s10980-018-0682-4

Vuorinen, I., Hänninen, J., Rajasilta, M., Laine, P., Eklund, J., Montesino-Pouzols, F., … Dippner, J. W. (2015). Scenario simulations of future salinity and ecological consequences in the Baltic Sea and adjacent North Sea areas-implications for environmental monitoring. Ecological Indicators, 50. https://doi.org/10.1016/j.ecolind.2014.10.019

Wenger, L. N., Lier, J. R. Van, & Fulton, C. J. (2018). Microhabitat selectivity shapes the seascape ecology of a carnivorous macroalgae-associated tropical fish, 590, 187–200.

Wikström, S. A., & Kautsky, L. (2007). Structure and diversity of invertebrate communities in the presence and absence of canopy-forming Fucus vesiculosus in the Baltic Sea. Estuarine, Coastal and Shelf Science, 72(1–2), 168–176. https://doi.org/10.1016/j.ecss.2006.10.009

24

Appendix Table A1. Invertebrate species that occurred in one or both of the different exposed F. vesiculosus habitats in Knabberskär (exposed) and Husbåtsviken (sheltered). Species with an asterisk were found in both kind of habitats.

Exposed F. vesiculosus habitat Sheltered F. vesiculosus habitat Balanus improvisus * Cerastoderma/Cardium sp. *

Balanus improvisus * Bythinia tentaculata

Dendrocoelum lactum * Gammarus sp. * Hydrobia sp. * Idotea sp. * Insect larvae * Mytilus edulis * Nementa spp. Palaemon adspersus * Palaemon palaemon * Radix balthica * Theodoxus fluviatilis *

Cerastoderma/Cardium sp. * Dendrocoelum lactum * Gammarus sp. * Hydrobia sp. * Idotea sp. * Insect larvae * Jaera sp. Mytilus edulis * Palaemon adspersus * Palaemon palaemon * Radix balthica * Radix peregra Theodoxus fluviatilis *

Stockholm University

SE-106 91 Stockholm

Phone: 08 – 16 20 00

www.su.se

25

Figure A1. Fish abundance (median and quartile) per square meter in (A) Knabberskär, (B) Husbåtsviken, (C) Vrångskär with all methods pooled. Location Knekten was excluded since it only had samples from habitats with F. vesiculosus. The fish abundance is grouped in habitats with or without F. vesiculosus in relation to the sites being exposed, sheltered or very exposed. (A) had one outlier (value 183) and (B) had six outliers (values 41, 42, 43, 44, 55 and 102) discarded due to y-axis limits.

Figure A2. Fish behaviours in locations with and without F. vesiculosus in Husbåtsviken (sheltered) and Knabberskär (exposed). Knekten &Vrångskär were removed due to only having one sample each.

26

Figure A3. Fish abundance (median and quartile) per square meter divided by species and life stage (juvenile or adult) for data collected with beach seine netting and visual underwater census (pooled data) at Husbåtsviken and Knabberskär in sites without F. vesiculosus (A) and with F. vesiculosus (B). Clupeidae spp. only occurs once in Husbåtsviken as a juvenile in a habitat without F. vesiculosus. Outliers in Husbåtsviken (values 20, 28, 30, 89 and 119) and Knabberskär (values 29, 30 and 45) were discarded due to y-axis limits.