Can cormorants be used as indicators of local fish abundances?
A
diet study of cormorants on Gotland
Oliver Hansen
Degree project inbiology, Master ofscience (2years), 2021 Examensarbete ibiologi 45 hp tillmasterexamen, 2021
Biology Education Centre and Institutionen för biologisk grundutbildning vid Uppsala universit, Uppsala University
Supervisors: Richard Svanbäck and Karl Lundström External opponent: Alessandro Culicchi
Abstract
Human wildlife conflicts can represent missed opportunities for ecological monitoring, including tracking invasive species. The great cormorant Phalacrocorax carbo sinensis is the centre such a conflict, where the lack of concrete scientific evidence is often replaced by anecdotal evidence, leading to the vilification of these birds. The primary aim of this study was to assess the extent of the overlap between cormorant diet and the fish the fishermen are allowed to catch on the North West coast of Gotland, the Baltic seas´ biggest island. To assess cormorant diet, the otoliths in the cormorant pellets were analysed. Secondary aims included assessing the potential to use cormorant diet as a proxy for local fish abundances by comparing it to monitoring fisheries in the same area. Highly contentious species only included cod, herring and flounder, none of which were commonly consumed by cormorants. Cormorants and the monitoring fisheries found comparable proportions of all species except for flatfish herring sprat, sculpin. We conclude that the cormorant poses a relatively low risk to the fishing industry on the North Western coast of Gotland and that they could potentially be used as a sentinel for local fish abundances, including tracking invasive species such as the round goby.
Introduction
Conflicts between living organisms occur when there is a finite pool of shared vital resources.
Through innovation and exploitation efficiency, humans have now become a major ecological force (Waters et al. 2016). As a dominant force, we often see wildlife as stealing resources that should ‘rightfully’ belong to us. Such conflicts between wildlife and us are called human- wildlife conflicts. These conflicts tend to be very complex, both on the ‘human’ and ‘wildlife’
side of the equation. On the wildlife side, the severity of the conflict will depend on the size of the animal, its behaviour, the population size and the shared resource in question. The relationships between these factors and the severity of the conflict are not linear, but rather dynamic. For example, sensitizing macaques through hazing modified their behaviour and reduces conflict, without impacting population size (Honda et al. 2019). Thus, we cannot make blanket statements such as ‘large animals with large population sizes will be the cause of the most severe human-wildlife conflict’. On the human side, we have a similar situation. The conflict between human actors can be as severe and challenging to solve as the conflict between humans and wildlife itself (Rauschmayer & Behrens, 2008), because often scientific judgment is replaced by emotional responses, especially if scientific evidence is lacking and livelihoods or hobbies/passions are at stake (Carss D et al. 2009). In addition, social factors such as religion, ethnicity and affiliation further complicate human wildlife conflicts, as these play a role in determining the outcome of these conflicts (Dickman 2010).
One of the most polarizing birds in Europe is the great cormorant (Phalacrocorax carbo sinensis), henceforth cormorant. This generalist piscivorous predator employs both social and solitary fishing strategies, both on pelagic and benthic prey and in fresh and salt water (Grémillet et al. 1998, Grémillet et al. 2004). Their vision is surprisingly bad (White et al. 2007)
for an animal who achieved the highest known foraging yield for a marine predator (Grémillet et al. 2004). Uncommonly for pelecaniform birds, cormorants can nest almost anywhere, only requiring protection from ground predators and a structure large enough to support the nest and its inhabitants (Causey & Padula 2019). Cormorants have a day roost for resting and digesting and night roost for sleeping. They are both are in close proximity to their breeding and feeding grounds (Causey & Padula 2019). A stark population increase in Europe has raised fears of negative impacts on aquatic as well as terrestrial ecosystems (Klimaszyk & Rzymski 2016).
Like in many parts of Europe, the Swedish population of cormorants has increased from 100- 150 pairs in 1965 (Havs och Vatten myndigheten 2015a) to 40 000 in 2012 (Naturvårdsverket 2013). The exact reasons for this expansion are unclear. It seems that the cormorant population is benefiting from a number of factors. They are protected by law, no longer limited by food supply (Carss DN & Marzano 2005), favoured by environmental change (Marzano et al., 2013) including benefiting from reduced environmental toxins such as DDT and PCB and have gained a novel food source in invasive species such as the round goby (Neogobius melanstomus) (Bzoma & Meissner 2005).
Although evidence suggests cormorants inhabited Sweden 9 000 years ago, they probably did not breed in Sweden, or at least were not very common, during the 17th and 18th century (Engström 2001a). Cormorants were present in Sweden during the 19th century (Engström 2001a) and human persecution in the form of nest destruction and hunting has led to their local extinction within Sweden in the beginning of the 20th century (Ericson & Carrasquilla 1997). It is also though that pollutants such as DDT and PCB contributed to their decline (Hermann et al. 2018). Once protective measures were in place, including the ban of said pollutants, the number of breeding pairs started to increase in the latter half of the 1970´s and cormorants subsequently recolonized the Baltic in the 1980´s (Hermann et al. 2018). On Gotland, evidence for a subspecies of the Great cormorant breeding is 2 500-5 000 years old (Ericson &
Carrasquilla, 1997). The first breeding pairs since local extinction were found in 1992 and the population grew to 10 500 breeding pairs in 2008, declined to around 8 000 pairs from 2010- 2013, increased to around 10 000 pairs in 2014 and has been declining slowly since then to around 7 500 pairs in 2018 (Hermann et al. 2018). Cormorants have caused elevated nitrogen, ammonium and phosphorous levels in at least one historically nutrient poor lake on Gotland, Ajkesträsk, at which hunting has now been allowed, as the protection of the lake outweighs the protection of the cormorant according to Länsstyrelsen Gotlands Län (2020).
Diets of generalist and opportunist carnivores have been said to be a good indicator of temporal as well as spatial variation in prey availability (see Moorhouse-Gann et al. 2020), therefore, predator diet could possibly be used as proxy for prey abundances. Combining seabird diet and fish monitoring data is likely to reflect a more accurate state of nearshore fish recruitment (Robinette et al. 2018) and changes in cormorant diet has been shown to coincide with similar changes in fish community (Boström et al 2012). Mecenero et al. (2007) suggested that seal diet could be a proxy for future fish catches.
One way to assess the diet of cormorants is based on the pellets that they regurgitate. The pellets are the undigested remains of fish (mostly bones and corneas) covered in a membrane. Pellets are a suitable method of assessing diet because most hard parts such as spinal bones but also ear bones, also known as otoliths, remain identifiable. Pellets easy to find and non-invasive to collect (Duffy and Laurenson 1983). Although cormorants are a highly mobile species being able to migrate from the Baltic Sea to the Mediterranean Sea (Ericson & Carrasquilla 1997), this should not impact our ability to predict local fish abundances based on their pellets. During the breeding season, cormorants seem to stay within 25km of their nesting site (Paillisson et al.
2004), although foraging at a distance of 35km from the nest has also been recorded (Gremillet 1997). The feeding range of cormorants seems to also stay fairly limited to the nest site when fish abundances are declining, and respond with increased number of foraging trips as opposed to fewer, but further trips (Gremillet 1997). However, Gremillet (1997) also postulates that cormorants will commit to further feeding trips if a more energetic resource is available there.
In addition, cormorants are constrained by not having waterproof feathers and thus need to stop and dry after being in water (Srinivasan et al. 2014). Because of this drying, the distance they can travel between meals is reduced. Therefore, a pellet represents the diet within the feeding range of the cormorant.
Cormorants have the ability impact local fish communities thorough both top down (predation) and bottom up (eutrophication) (Piotr et al. 2015, Gagnon et al. 2015). Thus a growing population of cormorants has caused a polarizing debate between those who deem this a conservation success and those who see cormorants a nuisance to ‘their’ fishing industry (Žydelis & Kontautas 2008, Bregnballe et al. 2015). This debate is known as the ‘cormorant problem’ (Marzano et al. 2013b). Cormorants have a bad reputation as a bird that consumes vast amounts of fish and this can lead to a perceived conflict, when in reality the conflict is likely to be less severe than public perception would suggest (Žydelis & Kontautas 2008). Other factors contributing to the bad reputation include damaging trees (Petersson 2020), causing unpleasant odour, eutrophication, and being seen as invasive species.
This however does not mean that they have no effect, but their impact is exaggerated, with speculation substituting evidence as the basis of arguments against cormorants. The problem is that we more often than not do not know the impact cormorants have on the local environment.
Cormorants have been known to catch between 3-13% of total valued commercial and recreational catches and between 10% and 44% of fish biomass available in certain areas (Östman et al. 2013). However in the Baltic Sea as a whole, humans almost harvest 4 times more fish in total than seals and cormorants combined (Hansson et al. 2018). Humans consume the highest number of fish for all species in the Baltic Sea, with the exception of perch, of which cormorants are estimated to consume the highest number (Hansson et al. 2018). A lack of data on fish consumption by both predators and fisheries and their effects on local fish populations further fuels the emotionally laden debate between different stakeholders (Hansson et al. 2018).
Fishermen across the Baltic Sea perceive the stark population increase of a bird that can consume 200-600g of fish per individual per day (Grémillet et al. 1995) as a threat and thus warrants a pan European culling effort (Marzano et al. 2013a).
The debate of the ‘cormorant problem’ has, like many human-wildlife debates, become highly politicized, fuelled by emotions and decision have not always been made on the basis of science (Marzano et al. 2013a). What makes the ‘cormorant problem’ such a difficult problem to solve, is that the impacts that the cormorants have are hard to generalize because they are opportunistic generalists. The fishes they consume have varying commercial and ecological importance, and the proportion of these fish in the cormorants diet varies with feeding grounds, local fish community dynamics, and breeding status of the birds (Boström et al., 2012; Hansson et al., 2018; Lehikoinen, 2005; Lehikoinen et al., 2011; Östman et al., 2013).
Cormorants however are not always viewed as villains, as they are used as allies in some traditional fishing techniques (Inoue‐Murayama et al. 2002) and have the potential to be of scientific value. Evidence suggests that coastal bird diet could be supplement fisheries monitoring data (Robinette et al. 2018), giving us a better idea of local fish abundances.
A handful of effects can be observed by looking at the diet of a predator, and one of them is prey switching. Prey switching is the phenomenon where predators change their diet either due to altered abundances in their usual prey due to environmental conditions (Wells et al. 2017), invasive prey species (Mutze 2017), novel prey items (Dilley et al. 2016), environmental anomalies (Xavier et al. 2018), time of year (Čech et al. 2008) and breeding season. For example, frugivorous hornbills have been shown to adjust their diet to the changing nutritional needs, selecting a diet with high calcium content during egg laying, and focusing on energy, protein and iron during the subsequent breeding period (Lamperti et al. 2014). Diet can also be impacted by breeding status, as seen in the wood duck, where egg laying females consumed more invertebrates than pre- or post-laying females (Drobney & Fredrickson 1979). Seasonal diet variation has also been documented in cormorants. Breeding phase has been shown to affect the diet of cormorants; however, the effect is small, with only explaining 3-10% of variation (Boström et al. 2012b). Diet variations within the breeding season likely occur due to chicks growing up and thus being able to eat larger prey items (Lehikoinen 2005).
Novel diet changes due to invasive species has been observed before. The round goby is one of the most impactful invasive species in the Baltic Sea, in terms of predation, competition but also being food for others (Ojaveer & Kotta, 2015). The goby now predates on mussels in the Gulf of Gdańsk (Skora & Rzeznik 2001) and competes with flounder (Platichthys flesus) for food (Karlson et al. 2007). It can also be considered a major food source for cod (Gadus morhua), perch (Perca fluviatilis) (Almqvist et al. 2010) and cormorants (Bzoma 1998, Johnson et al. 2010) and have thus changed the food web in the Baltic Sea. The round goby has been known to prey on fish eggs of commercial species, at least in a lab setting (Fitzsimons et al.
2006).
Originally from the Ponto-Caspian Sea, the round goby came to the Baltic probably through ballast water in the 1990s when it was found in the Gulf of Gdańsk and is now a dominant species (Sapota & Skóra 2005). They have probably been dispersing at around 30km a year (Azour et al. 2015). They show site fidelity (in rivers), most likely preferring rocky habitats and dispersing from sandy ones, with relatively small home ranges (Ray & Corkum 2001). When
colonizing new river habitat, they use long and short-term dispersal (Bronnenhuber et al. 2011) which might undermine control efforts through culling and increasing predation pressure (Ojaveer et al. 2015). The round goby also seems to experience natural selection promoting dispersal. Gobies of newer and sparse populations have better body condition than those in denser and old populations (Azour et al. 2015). Old populations also have a broad range of age classes, although new populations are male biased and have intermediate age (Azour et al.
2015). Round gobies are relatively well established on Gotland. They were first seen in Visby harbour in 2010 and have colonized some streams in 2016 (Puntila et al. 2018). Since cormorants are known to feed on gobies, a healthy cormorant population could help suppress the dominance of the invasive goby, as well as reveal the extent of the invasive fish’s range.
Following the rapid population increase of cormorants, it seems the population size has stabilized on Gotland (Hermann et al. 2018). However, the population as well as the resulting consumption potential remains large. Therefore, it is important to present sound facts concerning the extent to which a human-wildlife conflict is occurring. Because of the impacts of cormorants can vary so widely by area, this study will only focus on assessing the impact on the North- Eastern coast of Gotland. This study will investigate if cormorant diet varies with season and region, if cormorant diet is a viable proxy for local fish abundance as well as if cormorants show a seasonal preference to fish sizes and to what extent individual diet specialization occurs. Finally, to assess the potential for a human-wildlife conflict, I will compare to what extent cormorants consume fish of a harvestable size class. This will attempt to fill a gap in literature, as studies comparing cormorant predation to independent fish monitoring are currently lacking (Ovegård 2017).
Methods Data collection
Pellet collection occurred at eight different locations on the north-eastern coast of Gotland between August 2018 to November 2019 (Fig.1). Pellets collected in June and August were classified as coming from summer and pellets collected in September to November from winter.
Every collection consisted of gathering approximately 30 pellets at known night-roosting sites.
Upon locating a pellet, they were individually wrapped in plastic and frozen once returned to the field station.
To prepare them for analysis, the pellets were washed in a large plastic bowl to rid them of the encapsulating membrane and any other organic matter. The remaining fish hard parts were then placed in a petri dish and left to dry. Once dried, the content was inspected using a stereomicroscope. Distinct fish bones were noted and later used to confirm otolith species identification. Three-spined stickleback (Gasterosteus aculeatus) numbers were estimated by counting the amount of dorsal spines and dividing the total by three for every pellet.
Otoliths were separated from the rest of the content and individually photographed using a Leica M165C stereo microscope and the Infinity analyze (Release 6.1 Lumenera corporation) software. Otolith pictures were measured using the software ImageJ (version 1.52a).
Estimating the amount of fish per pellet can be challenging, time consuming (Boström et al.
2012a) and problematic (MoKay et al. 2003). To get around some issues, the relative proportions of fish per pellet were calculated from absolute values.
Figure 1 Map of study area showing Gotland with sample locations, represented by black dot and name. No samples were taken at Visby, but was added for context. Filled grey circles represent locations where the monitoring fisheries took place.
Wear classes and species
Size correction factors (SCF) were used in an attempt to make fish length calculations more accurate by accounting for erosion through digestion (Boström et al. 2012a). Otolith digestion was classified according to guides from North Sea otoliths (Leopold et al. 2001). I assigned one of four wear classes to the otoliths depending on how worn down they were. Class 1 otoliths were those that were minimally eroded. Otoliths that were the most eroded were assigned to class 4. Wear class 1 was characterized by a well-defined surface, distinct sulcus and lobate margins as per uneroded otolith of that species. Wear class 2 and 3 showed decreasingly well- defined characteristics. Wear class 4 was highly eroded, showing no sign of sulcus and an exceedingly smooth surface and margin. In the case that otoliths were so eroded that they were unrecognizable, they were assigned to unknown.
In order to calculate the original size of an eroded otolith, size correction factors (SCF) were calculated for each species and wear class. To calculate what the average erosion between wear classes is, for example the average size difference between wear class 1 and 2, I divided the average size of wear class 1 by the average size of wear class 2. By doing this, I got an SCF for wear class 2. A class 2 otolith can then be transformed into a class 1 otolith by multiplying the length of the class 2 otolith by the SCF. The result is the theoretical, uneroded size. This process was repeated for every species.
Assigning a species to otoliths is a challenging and sometimes impossible task. Otoliths belonging to possibly either black goby (Gobius niger) or round goby (Neogobius melanostomus) were assigned to goby. The single confirmed turbot (Scophthalmus maximus) otolith was assigned to flatfish. In the case where flounder and turbot were indistinguishable, I assigned them to flatfish.
Fish length regressions
I accounted for the size reduction effect of eroded otolith by employing SCF mentioned above.
Fish length was calculated using regressions and size-corrected otolith measurements. The width of goby otoliths was used to calculate the fish length using the regression of Azour et al.
(2015) For all other species, otolith length and regressions from Otoliths of North Sea Fish (Leopold et al. 2001) were used. Otoliths that could not be confidently assigned to a species were labelled unknown. Calculated fish lengths were capped at the known maximum length according to Otoliths of North Sea fish.
Monitoring fishing
Fishing was carried out by Swedish Agricultural University (SLU) as part of the coastal fish monitoring program on behalf of the County Administrative Board of Gotland. It occurs every year during a 2–3-week period in August. The nets are set between 2 and 5pm and stay out for a night, being collected the following day between 7 and 10 am (Havs och Vatten myndigheten 2015a, Havs och Vatten myndigheten 2015b).
Statistical analysis
All statistics were done in R version 4.0.3 (R Core Team 2020) and data was tidied using the tidyverse group of packages (Wickham et al. 2019). The northeastern coastline of Gotland was separated into two regions. The southern region included cormorant pellets from Östergarnsholm, Storholmrn, Laus Holmar, Nabben and the southern monitoring fisheries data.
The northern region included cormorant pellets from Sarvagrund, Furillen, Grauten and Smöjen and the northern monitoring fisheries data. The two season considered in this analysis was summer and winter. The pellets collected between July and August were assigned to summer, and pellets collected between September and November were assigned to winter.
Diet variation
Data was prepared for these analyses by calculating the numerical proportions of each fish species found in every individual pellet. Using the metaMDS function in the vegan package (Oksane et al. 2020), compositional diet variations were visualized with a non-metric
multidimensional scaling (NMDS) using the Bray-Curtis dissimilarity indices. NMDS was plotted using ggplot2 (Wickham 2009). A PERMANOVA was then used to test diet composition variation between regions and seasons. The Adonis function in the vegan package was used to execute the PERMANOVA. In total, I had to remove four outliers. The four outliers all came from Furillen during the winter season. They almost all exclusively contain very rare prey items such as sculpin and perch.
Regional and seasonal variations
Stacked boxplots from ggplot2 were used to visualize diet variation by regions and season. Two barplots were created by filtering the data for the two regions and plotting the difference between seasons. The other two barplots were created by filtering the data for the two seasons and plotting the differences between regions. For every barplot, a quasibinomial generalized linear model for each species was used to test for significant differences in proportion between seasons or region. Detecting seasonal size preferences was done by creating a grouped violin plot by season in ggplot2.
Comparing cormorants and fisheries
Fisheries data was prepared by calculating the proportions at which species were caught. The fishing data was subsetted by region, producing a dataset for the north and a dataset for the east.
The proportions of fish caught were then calculated for every fishing trip and the mean proportions per region were compared. The north only had one trip, meaning the proportions for the trip and the mean for the region was the same. Species proportions per region were also calculated for cormorant data. Only pellets collected the same month the monitoring fishery took place were used in this comparison. Here, the mean proportion of species in cormorant pellets was used for the barplot. Seasonal differences were not possible to test since fishing only occurred in August, which falls in the cormorant post-breeding season. For every barplot, a quasibinomial generalized linear model pair for each species was used to test for significant differences in proportion between regions. Following fish were added to the category “other”:
ide (Leuciscus idus), common bleak (Alburnus alburnus), crucian carp (Carassius carassius) , rudd (Scardinius erythrophthalmus), whitefish (Coregonus lavaretus), mackerel (Scomber scombrus), sea trout (Salmo trutta), the Syngnathidae family (pipefish) and smelt (Osmerus eperlanus)
Individual specialization
The same proportion data used for the NMDS was used in this analysis. To calculate the individual specialization, the dataset was first subsetted into four datasets: north summer, north winter, east summer, east winter. Using the PSicalc in the RInSp package (Zaccarelli et al. 2013) package, the degree of specialization was determined. The specialization scores from the different datasets were then compared with an ANOVA.
Human- wildlife conflict
To visualize the potential for this conflict, a violin plot contrasting fish sizes caught by cormorants with fishes caught by the monitoring fishery was created. Lines were added to show minimum allowable catch sizes according to (Länsstyrelsen, Gotlands Län 2019). To detect
significant differences in mean fish size per species pair, a Wilcoxon signed-rank test from the rstatix package (Kassambara 2021) was performed for all species. I chose a Wilcox test to be as consistent as possible because not all data was normally distributed.
Results
In total 357 pellets were analysed of which 68 were empty. 4 929 otoliths were photographed and analysed, of which 1 948 were classified as unknown. Species were caught in the following amounts: sticklebacks (1 948), goby (1 181), eelpout (716), flatfish (391), cod (241), sandeel (194), sculpin Taurulus bubalis (99), herring Clupea harengus (71), perch (19) and sprat Sprattus sprattus (1).
Figure 2. Non-metric multidimensional scaling (NMDS) ordination plot comparing the diet of cormorants in different regions (north and east) and seasons (summer and winter) on Gotland.
Points that are closer together are more similar in composition (both species and number of individuals per species). Ellipses are drawn for combination of region and season with every ellipse representing a different combination.
Cormorant diet does differ significantly by region and season (Fig 1, Fig 3, Fig 2, Appendix Table 2). However, the amount of variation explained by these factors remains low, with seasons, region and their interaction only explaining 6.9%, 2.6% and 1.7% of the variation respectively (Appendix Table2).
Differences in cormorant diet occur regionally within the summer season. Significantly more goby and stickleback were caught in the north and significantly more flatfish in the east (Fig 3,
Appendix Table 3). Regional differences also occurred outside the summer season (Fig 3, Appendix Table 3). Significantly more goby and sandeel were found in the north and flatfish and herring in the east. The amount of empty pellets was significantly higher in the north (Fig 3, Appendix Table 3). Seasonal differences were also detected (Fig 3, Appendix Table 3). In the north, significantly more cod and sculpin were found in the winter season. The northern summer season saw a significantly higher proportion of eelpout, goby and stickleback (Fig 3, Appendix Table 3). On the east coast of Gotland, the summer season had significantly larger proportions of sculpin (Fig 3, Appendix Table 3). Cod, goby and sandeel were significantly higher in proportion during the winter season (Fig 3, Appendix Table 3).
Figure 3 Diet proportions of cormorants on North-eastern coast of Gotland. Barplots show the numerical prey composition in cormorant pellets during a) summer, b) winter, and in the c) north and d) east. Asterisk represents significant differences between mean proportions of fish in the accompanying plot.
When comparing fisheries and cormorant catch in the north, herring, sprat, flatfish and sculpin were significantly underrepresented in the cormorant diet (Figure 4, Appendix Table 3). Not enough data was available to produce a Generalized linear model (GLM) for perch, however a visual inspection seems to indicate that perch is also underrepresented in cormorant diet.
Although the proportion of stickleback seems very large in northern cormorant populations compared to the monitoring fisheries, the difference is not significant. In the east, flatfish, herring and sprat seem significantly underrepresented in cormorant diet (Figure 4, Appendix Table 4). Although the proportion of cod and eelpout seem clearly larger in the cormorant diet compared to the monitoring fisheries, they do not seem to be significantly different (Appendix, Table 4). Not enough perch were found in the cormorant diet for the GLM to function. The
differences are thus statistically insignificant only because no statistics were run. They are still likely to reflect real differences.
Figure 4 Proportions of species caught by fisheries and cormorants during the month of August. Asterisks in legend indicate significant differences in proportion found in pellets between the two seasons, based on a GLM (Appendix Table 4) not enough perch were found in the diet of cormorants to do a GLM.
Figure 5 The degree of individual specialization (IS) for both regions (left) and both seasons (right) based on cormorant pellets composition Lower scores indicate higher levels of individual specialization. Significant differences between boxplots is indicated above the barplot.
Cormorants on Gotland do seem to specialize on certain foods, however the overlap between the populations diet and the individuals diet still remains fairly large with higher levels of specialization during the winter season for both regions (North summer: IS = 0.49, p = 0.001, North winter: IS = 0.29; East summer IS = 0.48, p = 0.001; East winter IS = 0.31, p = 0.001).
An ANOVA explaining the proportional similarity index PSI by region and season, including their interaction, suggests that only season is a significant factor affecting the degree of individual specialization (Appendix Table 1). I subsequently performed a linear model on the
same data that resulted in an estimate of PSi being around 17% lower during the winter season compare to the summer season (Appendix Table 8).
Figure 1. Violin plot showing distribution of fish sizes caught by cormorants during the summer season (red) and winter season (blue). The wider the plot, the more frequent the fish at that size was observed. Significant differences are noted above the plots. Points represent minimum allowable catch size. In flatfish, two dots are found; the lower dot represents the minimum size for flounder, the upper for turbot.
Generally, fish sizes in the diets of cormorants were similar between seasons. However, cod, flatfish, herring, and sculpin were significantly larger in the diet during the winter season (Figure 6, Appendix Table 4). The length distribution of cormorant diet did overlap with the minimum allowable catch size, with all three commercially harvested species (cod, flatfish, and herring) (Figure 6).
Figure 2 Violin plot showing distribution of fish sizes caught by cormorants (blue) and the monitoring fishery (red). As the monitoring fishery happened in August, only cormorant data from August was used to create this plot. Wider plots represent higher frequencies.
Cormorants seem to catch similar sizes to the monitoring fisheries, except for flatfish, herring, and sandeel, whose mean sizes were significantly smaller in the cormorant diet compared to the sizes from the fishing (Figure 7, Appendix Table 5).
Discussion
Although cormorants are one of the most studied birds in Europe, too little is known about the assessment of potential for human-wildlife conflicts in different areas. This study focusses on discerning this potential on the West coast of Gotland, the largest island in the Baltic Sea.
According to pellets collected between August and November in 2018 and 2019, the most common species by number were three-spined stickleback, gobies (Gobiidae) and eelpout. I found significant yet small differences between seasons and regions and limited size preference between summer and winter season. Cormorant diet did overlap with the minimum allowable catch size of the three commercially harvested species (cod, flatfish, herring). However, the bulk of fish consumed was smaller than the minimum allowable catch size for commercial harvesting (Fig 6). The likelihood of a severe human wildlife conflict between the fishing industry and cormorants seems low.
Cormorant pellet composition
In relation to previous studies, the abundance of fish in cormorant pellets found in this study was partially different. In contrast to my study, Boström, Lunneryd, et al., (2012) found that in Lövstabukten in the southern Bothnian Sea, herring, perch and eelpout were the most numerous species found. My study supports continued reliance of post-breeding cormorants on gobies in the north of Gotland (Larsson 2017), however the overall reliance on sculpins and cod remains low, unlike in Larsson (2017). Cormorants in Kalmarsund, on the southeaster coast of Sweden, seemed to also rely on sticklebacks, eelpout and herring in an earlier study (Boström et al.
2012b). Compared to the Swedish mainland, cormorants on Gotland seem to rely less on herring during our study period, possibly due to a difference in local herring abundance. Cormorants also seem to be able to thrive on a completely different set of fishes, like in the Curonian spit, where they fed on mostly ruffe, perch and roach (Pūtys & Zarankaitė 2010) and cormorants in the Gulf of Finland have been shown to rely heavily on eelpout, roach and perch (Lehikoinen 2005).
Cormorant pellet composition in my study varied significantly between regions and seasons (Figure 2), however the amount of variation explained by these factors remains low (Appendix, Table 2). Thus other factors are likely more important in determining the pellet compositions.
Individual variation, even within the same breeding phase, has been previously hypothesized as the reason for the low explanatory power of season and region (Boström et al. 2012b). It has been shown that even on a population level, if behavioural repeatability is low, there can be large differences on the individual level (Potier et al. 2015). Individuals can be anywhere on a spectrum between highly flexible and very consistent in their foraging behaviour. A study on white-tailed eagles suggests that feeding habits can differ within and between populations, among individuals (generalist or specialist individuals), and between seasons (Nadjafzadeh et al. 2015). Diet specialism in a mostly generalist predator has also been linked to higher foraging success rate, possibly due to limiting their response to stimuli from different prey items to the ones which usually are more easily captured (Terraube et al. 2014).
Individual variation is likely to occur when the feeding grounds consist of diverse habitats with a diverse set of potential prey items, as opposed to homogenous habitat that is dominated by a few very abundant prey with short handling time and high attack success rate (see Rosenblatt et al. 2015). This could possibly be a reason why cormorant diets can be so different at different locations.
My analyses of individual specialization suggest that although the overlap between the population’s diet and the individuals’ diet remains relatively high (between 30 and 50%) depending on the season and region, the specialization is statistically significant (Appendix Table 1). Specialization seems to increase during the winter, suggesting that cormorants become more selective as the season progresses. Region was not significant in the analysis of specialization (Appendix Table 1), suggesting that these trends caused by cormorant preference and not by food availability if fish migration was occurred equally in both regions.
An additional factor influencing observed diet compositions in pellets could be expulsion of otoliths through faeces, making pellets not the sole source of otoliths (Duffy & Laurenson 1983), and it is unknown if there is a consistent bias for otolith that are expelled. Duffy and Laurenson (1983) also reported an otolith recovery of around 30%, and observed that when Cape cormorants were fed twice a day, they would produce less than two pellets a day, making it unlikely that a pellet reliably reflects a single meal. Despite these limitations, Duffy and Laurenson (1983) maintain pellets are suitable to reconstruct diet since most otoliths are not so degraded that they are unidentifiable, they are accessible, non-invasive and easily turned into long term data.
Even though region and season explained a small amount of the diet variation in this study (Figure 2), many significant differences could be observed. Sticklebacks, sandeels and gobies seem to be more abundant in the north, with the latter being significantly higher in both seasons, possibly indicating a higher relative abundance in the north (Figure 3). Herring and flatfish seem to be more abundant in the east, with proportions of flatfish in the east being significantly higher in both summer and winter compared to the same season in north (Figure 3).
Cormorants in my study also changed their diet preferences between seasons, preferring eelpout, gobies, stickleback and sculpins during summer depending on the location, and cod, sculpin and sandeel during winter. These changes in diet suggest that cormorants are flexible predators. This study adds to the body of literature exemplifying variation in cormorant diet (Ovegård 2017). If changes in diet according to season are accounted for and enough data is collected over a long period of time, cormorant diet could be an indicator of changing fish abundances. However, the amount of data must be so large that fish abundances could be detected through the ‘noise’ of cormorant diet variation due to other factors.
Impact on local fishing communities and conflicts with commercial fishing
The results of studies attempting to determine whether cormorants impact local fish communities or not are mixed, with some reporting little or no impact and others severe impact.
Gagnon et al. (2015) found that cormorants did not affect species richness or biodiversity at any trophic level overall, but can have species-specific impacts, e.g. perch and ruffe were less abundant near cormorant colonies. These top-down effects are likely to be spread out over larger areas, as cormorants are able to leave the immediate vicinity of the colony to forage.
Lehikoinen et al. (2011) states that commercial and monitoring fishing catches did not decrease in relation to a growing cormorant population. Further evidence suggests that in the long term, cormorants can have a limited effect on local fish populations in Swedish lakes (Engström 2001b). Engström (2001c) supports a lack of competition between cormorants and fisheries, with cormorants’ predations seeming to work within the range of compensatory mortality and larger negative effects on fish populations being caused by decreasing levels of phosphorous leading to reduced productivity. Schröder et al. (2009) found that culling fish of a certain size class reduces the mortality of other size classes. This occurs because amongst others, resource competition is relaxed, and thus more resources are available for the remaining individuals.
Other factors such as colony size has been negatively associated with abundance of perch (Östman et al. 2012). The impact of colony size however seems to be species specific, as not all species are affected equally (Östman et al. 2012). My study suggests that perch is a small part of the cormorant’s diet on Gotland, even though it was commonly caught in the monitoring fisheries. Östman et al (2012) refrain from suggesting a causational relationship and consider the possibility of changes observed being due to an unaccounted (possibly environmental) factor. In addition to dietary plasticity, factors such as cormorant density, abundance and colony age could be the reason for variable effects on local fish populations (Gagnon et al. 2015).
Because the effects cormorants can have are so variable, efforts to manage their population are unpredictable and area dependent (Östman et al. 2012).
Although variable, cormorants have the potential to affect local fish abundances. Vetemaa et al.
(2010) argue that perch and roach populations have been negatively affected in semi-enclosed bay in the Baltic Sea. However, their monitoring fishing data has been questioned and described as too short-term by Lehikoinen et al. (2011), as it was only collected during 1995 and 2005.
They also did not attempt to look at cormorant diet directly. Nevertheless, other evidence also exists that cormorants feed on fish targeted by commercial fisheries. It has been estimated that cormorants can consume up to 14% of harvestable size fish in the Karlskrona archipelago and 5% in the Mönsterås archipelago, both areas in south-eastern part of Sweden (Östman et al.
2013). However, in terms of biomass consumed in the Karlskrona and Mönsterås archipelago, cormorants consumed between 10% and 44% of commercial and recreational fishing catch of cod, flounder, herring, perch pike and whitefish combined (Ovegård 2017). In a meta-analysis, Ovegård (2017) concluded that in general, cormorants have a negative effect on local fish populations; however, the intensity of this effect was specific to the fish species. Most studies finding negative associations were conducted in closed and simple systems such as farms or dams and evidence of how cormorants affect wild fish populations is still lacking (Ovegård
2017). The fact that there seems to be no clear scientific consensus on how intense the conflict is between fisheries and cormorants emphasizes the need for local long-term monitoring of cormorant diet, cormorant numbers and local fish stocks to track the intensity of the conflict over time.
Of the species of fish targeted by cormorants in my study, only cod, flatfish and herring have minimum allowable catch size (Länsstyrelsen, Gotlands Län 2019) and commercial value.
Cormorants do feed on cod, flatfish and herring of harvestable size (Fig 6). Therefore, there is competition between humans and cormorants for these species. If the cormorant population remains large, a human wildlife conflict seems unavoidable. Is it likely to remain limited, since a many fish are consumed that are under the allowable catch size. The population size at which the conflict could become critical remains unclear, as direct competition still seems low. It is possible that the population will be constrained by other factors before the cormorant pose a risk to commercial fishing, or that negative effects will only be seen once the colony reaches a certain age.
In addition to affecting fish populations through top down control (predation), cormorant also seem to have a bottom up effect on local fish populations though eutrophication (Gagnon et al.
2015) e.g. filamentous algae and primary productivity seemed to be highest at older and denser colonies, which can have knock on effects on the fish population. Eutrophication has been a problem across the entire Baltic, often having adverse effects such as algal blooms, hypoxia, increased levels of filamentous algae (Rönnberg & Bonsdorff 2004) and shifts in community composition (Bonsdorff et al. 1997). Furthermore, eutrophication caused by cormorants can lead to loss of biodiversity (Gagnon et al. 2016). However, a cormorant colony is likely to only cause shifts in nutrient levels adjacent to the colony (Gagnon et al. 2016) disproportionately affect species living close to the colonies such as flounder while highly mobile species such as herring are likely to be less affected.
However, one should be vary of generalizing results from specific studies and extrapolating them to other areas and/or other time periods or even vilifying cormorants, as they mostly seem to feed on non-commercial species. Our results agree with Östman et al. (2013) that cormorants tend to forage on fish smaller than harvestable stock and with Andersen et al. (2007) in that cormorants mostly prey upon non-commercially valuable fish.
Predicting prey abundance
Seabirds lend themselves to be used as indicators of ecosystem status/change for a number of reasons (Parsons et al. 2008). Seabirds seem to be predominantly affected by bottom up effects (Aebischer et al. 1990), thus changes in their food abundances should be reflected in the predator population. They are highly visible, easily observed, relatively easily captured and often aggregate in large numbers at limited locations which allows scientists to census populations of coexisting species at different trophic levels (Piatt & Sydeman 2007). In fact, breeding failures of seabirds have already signalled fish stock collapses in Peru, Norway, the North Sea and the Barents straight (Piatt & Sydeman 2007). Using seabirds as an indicator for salmons off central California has already been proposed (Roth et al. 2007). Focusing on a
single sentinel species is likely less valuable than focusing on a so called aggregate indicator such as biomass of a class of consumers and in order for predator diet to really become useful, further models need to be developed than can reliably predict change in fish stocks (Piatt &
Sydeman 2007).
Cormorants have the potential to be used as one of these indicator species. Studies have suggested that cormorants eat what is available locally and not move around with the fish (Dias et al. 2012), and their diet change with a changing fish community (Boström et al. 2012b). This is not limited to the great cormorant. The Brand’s cormorants also seems to consume prey according to their relative abundance in their environment (Wells et al. 2017) thus making them potentially useful indicators. Using cormorant data in conjunction to monitoring fisheries is likely to yield the best results. For example, shoal formation is a highly variable event that can be missed by scheduled monitoring efforts. Sea bird diet could help fill gaps that ‘normal’
surveys have because they miss chance events (Robinette et al. 2018). It is probably best to only base predictions of fish populations on long term data, as prey switching not representing a change in local abundance in cormorants can occur (Lehikoinen 2005).
My data suggests that cormorants catch similar proportions of fish species than the monitoring fisheries except for herring, flatfish, sculpin and sprat in the north and herring and sprat in the east (Figure 4). In the north, the mean proportion of flatfish, herring, sprat and herring were significantly higher in the monitoring fisheries than in the pellets of cormorants. In the east, the mean proportion of herring and sprat were significantly higher in the monitoring fisheries compared to cormorant pellets. This suggests that cormorants might be a better indicator of benthic fish species rather than pelagic species, as herring and sprat are the latter. Perch was found almost exclusively in the monitoring fisheries, suggesting that perch is an exceedingly minor part of the cormorants’ diet along the north-eastern coast of Gotland. However, further studies need to be done to assess the use of pellets as indicators for prey abundance. Another possible reason could be that the otoliths of these species just degrade faster, either completely disappearing or end up being classified as unknown (Duffy & Laurenson 1983, Martucci et al.
1993). Cormorants could also produce pellets more than once a day, meaning that pellets are probably representative of numerous meals (Duffy & Laurenson 1983). However, it is generally assumed that one pellet is produced daily (Veldkamp 1995).
Another way that cormorant diet could be very useful is in relation to the round goby invasion in the Baltic Sea. The cormorant could be a valuable ally in locating and perhaps suppressing round goby populations as studies have shown that cormorants readily eat round gobies. Round goby remains in pellets have been found to constitute up to 90% during the gobies spawning period between April and October in the Gulf of Gdanks, possibly even being the reason why there are so many cormorants there today (Bzoma & Meissner 2005). Others have reported a much lower proportion of around 12% (Rakauskas et al. 2013). My study also suggests that Gobidae are an important prey species for cormorants around Gotland. Tracking the round goby otoliths in cormorant diet could help track their invasion around the island and around the Baltic Sea. Because it is very challenging to differentiate between round and black goby otoliths, the species should be determined by using DNA analyses.
Future research
To study effectively and conclusively the effects that cormorants have on commercial fish stocks, scientists and managing authorities should consider a number of aspects. Firstly, colony size, density and age should be taken into account, as they have been shown to impact local conditions. Secondly, substantial effort should be made to assess both the diet of the cormorants, and local fish abundances and tracking both changes over time ideally equating them to environmental conditions.
Future research should take care in determining which species of goby this is, possibly due to DNA analysis. Assessing the role of the cormorant as an indicator for goby invasion or importance as a predator could be extremely useful to help understand and even suppress the round goby invasion. In addition, DNA analyses could help with identifying the otoliths that cannot be classified by morphology and thus reducing the unknown proportion.
Conclusion
The ‘cormorant problem’ in Europe is a very complex issue to solve because the human aspect can be as challenging to address. The debates are often fuelled by emotions and obscured by a lack of science. It is critical that the outcomes of scientific studies are communicated to stakeholders. This will hopefully disband the feeling of inferiority towards the state and remove the antagonizing attitude between policy bodies, scientists, and the people affected by this issue.
This study attempts at expanding the knowledge around the cormorant diet on Gotland. I found that the diet is variable, between both seasons and regions although with low explanatory power.
Therefore, other factors must be co-determining cormorant diet. Cormorant diet proportions seem fairly close to proportions of fish caught by the monitoring fisheries and local fish abundance is likely a factor influencing cormorant diet. Cormorant diet could be used as a proxy for local fish abundances given the following conditions a) frequent yearlong pellet collections need to be made to capture seasonal variation, and b) multiple sample sites should be chosen to increase representation, yet more research is needed to assess how much data is sufficient to imply reliably the relative fish abundances. The cormorant problem on Gotland, at least from a fisheries point of view, seems benign as the population is no longer increasing and the majority of the diet of cormorants seem to be non-commercial species.
Aebischer NJ, Coulson JC, Colebrookl JM. 1990. Parallel long-term trends across four marine trophic levels and weather. Nature 347: 753–755.
Almqvist G, Strandmark AK, Appelberg M. 2010. Has the invasive round goby caused new links in Baltic food webs? Environmental Biology of Fishes 89: 79–93.
Azour F, van Deurs M, Behrens J, Carl H, Hüssy K, Greisen K, Ebert R, Møller P. 2015.
Invasion rate and population characteristics of the round goby Neogobius melanostomus: effects of density and invasion history. Aquatic Biology 24: 41–52.
Bonsdorff E, Blomqvist E, Mattila J, Norkko A. 1997. Long-term changes and coastal eutrophication. Examples from the Aland Islands and the Archipelago Sea, northern Baltic Sea. Oceanolica Acta 20: 11.
Boström MK, Lunneryd S-G, Ståhlberg H, Karlsson L, Ragnarsson B. 2012a. Diet of the Great Cormorant (Phalacrocorax carbo sinensis) at two areas at Lövstabukten, South Bothnian Sea, Sweden, based on otolith size-correction factors. 89: 13.
Boström MK, Östman Ö, Bergenius MAJ, Lunneryd S-G. 2012b. Cormorant diet in relation to temporal changes in fish communities. ICES Journal of Marine Science 69: 175–183.
Bregnballe T, Hyldgaard AM, Clausen KK, Carss DN. 2015. What does three years of hunting great cormorants, Phalacrocorax carbo , tell us? Shooting autumn-staging birds as a means of reducing numbers locally: Shooting autumn-staging cormorants to reduce numbers. Pest Management Science 71: 173–179.
Bronnenhuber JE, Dufour BA, Higgs DM, Heath DD. 2011. Dispersal strategies, secondary range expansion and invasion genetics of the nonindigenous round goby, Neogobius melanostomus, in Great Lakes tributaries: DISPERSAL STRATEGIES OF NEOGOBIUS MELANOSTOMUS, IN GREAT LAKES TRIBUTARIES. Molecular Ecology 20: 1845–1859.
Bzoma S. 1998. The contribution of round goby (Neogobius melanostomus Pallas, 1811) to the food supply of cormorants (Phalacrocorax carbo Linnaeus, 1758) feeding in the Puck Bay. Bulletin of the Sea Fisheries Institute 144: 39–47.
Bzoma S, Meissner W. 2005. Some Results of Long-Term Counts of Waterbirds Wintering in the Western Part of the Gulf of Gdańsk (Poland), with Special Emphasis on the Increase in the Number of Cormorants ( Phalacrocorax Carbo ). Acta Zoologica Lituanica 15:
105–108.
Carss D, Bell S, Marzano M. 2009. Competing and coexisting with cormorants. Landscape, Process and Power: Re-Evaluating Traditional Environmental Knowledge, pp. 99–121.
Berghahn Books, New York.
Carss DN, Marzano M. 2005. Reducing the conflict between cormorants and fisheries on a pan- European scale: REDCAFE-Summary and national overviews. 386.
Causey D, Padula VM. 2019. The Pelecaniform Birds. Encyclopedia of Ocean Sciences, pp.
119–128. Elsevier,
Čech M, Čech P, Kubečka J, Prchalová M, Draštík V. 2008. Size Selectivity in Summer and Winter Diets of Great Cormorant (Phalacrocorax carbo ): Does it Reflect Season- Dependent Difference in Foraging Efficiency? Waterbirds 31: 438–447.
Dias E, Morais P, Leopold M, Campos J, Antunes C. 2012. Natural born indicators: Great cormorant Phalacrocorax carbo (Aves: Phalacrocoracidae) as monitors of river discharge influence on estuarine ichthyofauna. Journal of Sea Research 73: 101–108.
Dickman AJ. 2010. Complexities of conflict: the importance of considering social factors for effectively resolving human-wildlife conflict: Social factors affecting human-wildlife conflict resolution. Animal Conservation 13: 458–466.
Dilley BJ, Schoombie S, Schoombie J, Ryan PG. 2016. ‘Scalping’ of albatross fledglings by introduced mice spreads rapidly at Marion Island. Antarctic Science 28: 73–80.
Drobney RD, Fredrickson LH. 1979. Food Selection by Wood Ducks in Relation to Breeding Status. The Journal of Wildlife Management 43: 109.
Duffy DC, Laurenson LJB. 1983. Pellets of Cape Cormorants as Indicators of Diet. The Condor 85: 305.
Engström H. 2001a. The occurrence of the Great Cormorant Phalacrocorax carbo in Sweden, with special emphasis on the recent population growth. Ornis Svencica 11: 155–170.
Engström H. 2001b. Long term effects of cormorant predation on fish communities and fishery in a freshwater lake. Ecography 24: 127–138.
Engström H. 2001c. Effects of Great Cormorant Predation on Fish Populations and Fishery.
PhD, Uppsala University
Ericson PGP, Carrasquilla FH. 1997. SUBSPECIFIC IDENTITY OF PREHISTORIC BALTIC CORMORANTS PHALACROCORAX CARBOland. 8.
Fitzsimons J, Williston B, Williston G, Bravener G, Jonas JL, Claramunt RM, Marsden JE, Ellrott BJ. 2006. Laboratory Estimates of Salmonine Egg Predation by Round Gobies (Neogobius melanostomus), Sculpins (Cottus cognatus and C. bairdi), and Crayfish (Orconectes propinquus). Journal of Great Lakes Research 32: 227–241.
Gagnon K, Sjöroos J, Yli-Rosti J, Stark M, Rothäusler E, Jormalainen V. 2016. Nutrient enrichment overwhelms top-down control in algal communities around cormorant colonies. Journal of Experimental Marine Biology and Ecology 476: 31–40.
Gagnon K, Yli-Rosti J, Jormalainen V. 2015. Cormorant-induced shifts in littoral communities.
Marine Ecology Progress Series 541: 15–30.
Gremillet D. 1997. Catch per unit effort, foraging efficiency, and parental investment in breeding great cormorants (Phalacrocorax carbo carbo). ICES Journal of Marine Science 54: 635–644.
Grémillet D, Argentin G, Schulte B, Culik BM. 1998. Flexible foraging techniques in breeding Cormorants Phalacrocorax carbo and Shags Phalacrocorax aristotelis: benthic or pelagic feeding? Ibis 140: 113–119.
Grémillet D, Kuntz G, Delbart F, Mellet M, Kato A, Robin J-P, Chaillon P-E, Gendner J-P, Lorentsen S-H, Maho YL. 2004. Linking the foraging performance of a marine predator to local prey abundance. Functional Ecology 18: 793–801.
Grémillet D, Schmid D, Culik B. 1995. Energy requirements of breeding great cormorants Phalacrocorax carbo sinensis. Marine Ecology Progress Series 121: 1–9.
Hansson S, Bergström U, Bonsdorff E, Härkönen T, Jepsen N, Kautsky L, Lundström K, Lunneryd S-G, Ovegård M, Salmi J, Sendek D, Vetemaa M. 2018. Competition for the fish – fish extraction from the Baltic Sea by humans, aquatic mammals, and birds. ICES Journal of Marine Science 75: 999–1008.
Havs och Vatten myndigheten. 2015a. Provfiske med kustöversiktsnät,nätlänkar och ryssjor på kustnära grunt vatten. 48.
Havs och Vatten myndigheten. 2015b. Provfiske i Östersjöns kustområden - djupstratifierat provfiske med Nordiska kustöversiktsnät. 46.
Hermann C, Gregnballe T, Larsson K, Ojaste I, Lilleleht V. 2018. Population Development of Baltic Bird Species: Great Cormorant (Phalacrocorax carbo sinensis). Helcom Indicator Fact Sheets, pp. 1–12. HELCOM,
Honda T, Yamabata N, Iijima H, Uchida K. 2019. Sensitization to human decreases human- wildlife conflict: empirical and simulation study. European Journal of Wildlife Research 65: 1–10.
Inoue‐Murayama M, Ueda Y, Yamashita T, Nishida‐Umehara C, Matsuda Y, Masegi T, Ito S.
2002. Molecular sexing of Japanese cormorants used for traditional fishing on the Nagara River in Gifu City. Animal Science Journal 73: 417–420.
Johnson JH, Ross RM, McCullough RD, Mathers A. 2010. Diet shift of double-crested cormorants in eastern Lake Ontario associated with the expansion of the invasive round goby. Journal of Great Lakes Research 36: 242–247.
Karlson AML, Almqvist G, Skóra KE, Appelberg M. 2007. Indications of competition between non-indigenous round goby and native flounder in the Baltic Sea. ICES Journal of Marine Science 64: 479–486.
Kassambara A. 2021. rstatix: Pipe-Friendly Framework for Basic Statistical Testsrstatix.pdf.
Klimaszyk P, Rzymski P. 2016. The complexity of ecological impacts induced by great cormorants. Hydrobiologia 771: 13–30.
Lamperti AM, French AR, Dierenfeld ES, Fogiel MK, Whitney KD, Stauffer DJ, Holbrook KM, Hardesty BD, Clark CJ, Poulsen JR, Wang BC, Smith TB, Parker VT. 2014. Diet selection is related to breeding status in two frugivorous hornbill species of Central Africa. Journal of Tropical Ecology 30: 273–290.
Länsstyrelsen, Gotlands Län. 2019. Fiske på Gotland. Bestämmelser för fiske inom Gotlands län. Länsstyrelsen i Gotlands län, Visby.
Länstyrelse Gotlands Län. 2020. Skyddsjakt efter skarv på Fårö. online January 21, 2020:
https://www.lansstyrelsen.se/gotland/om-oss/nyheter-och-press/nyheter---
gotland/2020-01-21-skyddsjakt-efter-skarv-pa-faro.html. Accessed November 30, 2020.
Larsson A. 2017. A diet study of post-breeding Great cormorants (Phalacrocorax carbo sinensis) on Gotland. SLU
Lehikoinen A. 2005. Prey-switching and Diet of the Great Cormorant During the Breeding Season in the Gulf of Finland. Waterbirds 28: 511–515.
Lehikoinen A, Heikinheimo O, Lappalainen A. Temporal changes in the diet of great cormorant (Phalacrocorax carbo sinensis) on the southern coast of Finland — comparison with available fish data. 16: 10.
Lehikoinen A, Heikinheimo O, Lappalainen A. 2011. Temporal changes in the diet of great cormorant (Phalacrocorax carbo sinensis) on the southern coast of Finland — comparison with available fish data. 16: 10.
Leopold MF, van Damme CJG, Philippart CJM, Winter CJN. 2001. Otoliths of North Sea Fish.
online 2001: https://otoliths-
northsea.linnaeus.naturalis.nl/linnaeus_ng/app/views/introduction/topic.php?id=3327
&epi=87. Accessed November 30, 2020.
Martucci O, Pietrelli L, Consiglio C. 1993. Fish otoliths as indicators of the cormorant Phalacrocorax carbo diet (Aves, Pelecaniformes). Bolletino di zoologia 60: 393–396.
Marzano M, Carss DN, Cheyne I. 2013a. Managing European cormorant-fisheries conflicts:
problems, practicalities and policy. Fisheries Management and Ecology 20: 401–413.
Marzano M, Carss DN, Cheyne I. 2013b. Managing European cormorant‐fisheries conflicts:
problems, practicalities and policy. 13.
Mecenero S, Krakstad J-O, Roux J-P, Underhill LG. 2007. Can seal diet predict future catches of commercial prey? Final Report of the BCLME (Benguela Current Large Marine Ecosystem) Project on Top Predators as Biological Indicators of Ecosystem Change in the BCLME. Avian Demography Unit, Cape Town.
MoKay Helen V, Robinson K, Carss D, Parrott D. 2003. The limitations of pellet analysis in the study of cormorant Phalacrocorax spp. diet. Vogelwelt 124: 227–236.
Moorhouse-Gann RJ, Kean EF, Parry G, Valladares S, Chadwick EA. 2020. Dietary complexity and hidden costs of prey switching in a generalist top predator. 10: 6395–6408.
Mutze G. 2017. Continental-scale analysis of feral cat diet in Australia, prey-switching and the risk: benefit of rabbit control. Journal of Biogeography 44: 1679–1681.
Nadjafzadeh M, Voigt CC, Krone O. 2015. Spatial, seasonal and individual variation in the diet of White‐tailed Eagles Haliaeetus albicilla assessed using stable isotope ratios. 15.
Naturvårdsverket. 2013. Nationell förvaltningsplan för skarv 2014. 67.
Ojaveer H, Galil B, Lehtiniemi M, Christoffersen M, Clink S, Florin A-B, Gruszka P, Puntila R, Behrens J. 2015. Twenty five years of invasion: management of the round goby Neogobius melanostomus in the Baltic Sea. Management of Biological Invasions 6:
329–339.
Oksane J, Guillaume Blanchet F, Friendly Michae, Kindt R, Legendre P, McGlinn Dan, Minchin PR, O’Hara RB, Simpson GL, Solymos P, Stevens H, Szoecs E, Wagner H. 2020. vegan:
Community Ecology Package version 2.5-7 from CRAN. online 2020:
https://rdrr.io/cran/vegan/. Accessed May 15, 2021.
Östman Ö, Bergenius M, Boström MK, Lunneryd S-G. 2012. Do cormorant colonies affect local fish communities in the Baltic Sea? 69: 1047–1055.
Östman Ö, Boström MK, Bergström U, Andersson J, Lunneryd S-G. 2013. Estimating Competition between Wildlife and Humans–A Case of Cormorants and Coastal Fisheries in the Baltic Sea. PLoS ONE 8: e83763.
Ovegård M. 2017. The Interactions between Cormorants and Wild Fish Populations. 55.
Paillisson J-M, Carpentier A, Le Gentil J, Marion L. 2004. Space utilization by a cormorant (Phalacrocorax carbo L.) colony in a multi-wetland complex in relation to feeding strategies. Comptes Rendus Biologies 327: 493–500.
Parsons M, Mitchell I, Butler A, Ratcliffe N, Frederiksen M, Foster S, Reid JB. 2008. Seabirds as indicators of the marine environment. ICES Journal of Marine Science 65: 1520–
1526.
Petersson M. 2020. Inventering av vattenvegetation i Ajkesträsk, Gotlands län, 2020.
Länsstyrelsen i Gotlands län, Visby.
Piatt I, Sydeman W. 2007. Seabirds as indicators of marine ecosystems. Marine Ecology Progress Series 352: 199–204.
Piotr K, Tomasz J, Piotr R. 2015. Roosting Colony of Cormorants (Phalacrocorax Carbo Sinensis L.) as a Source of Nutrients for the Lake. Limnological Review 14: 111–119.
Potier S, Carpentier A, Grémillet D, Leroy B, Lescroël A. 2015. Individual repeatability of foraging behaviour in a marine predator, the great cormorant, Phalacrocorax carbo.
Animal Behaviour 103: 83–90.
Puntila R, Strake S, Florin AB, Naddafi R, Lehtiniemi M, Behrens JW, Kotta J, Oesterwind D, Putnis I, Smolinski MS, Wozniczka A. 2018. Abundance and distribution of round goby (Neogobius melanostomus). HELCOM
Pūtys Ž, Zarankaitė J. 2010. Diet of the Great Cormorant ( Phalacrocorax carbo Sinensis ) at the Juodkrantė Colony, Lithuania. Acta Zoologica Lituanica 20: 179–189.
R Core Team. 2020. R: A language and environment for statistical computingullrefman.pdf.
online 2020: https://cran.r-project.org/doc/manuals/r-release/fullrefman.pdf. Accessed May 15, 2021.
Rakauskas V, Pūtys Ž, Dainys J, Lesutienė J, Ložpys L, Arbačiauskas K. 2013. Increasing Population of the Invader Round Goby, Neogobius Melanostomus (Actinopterygii:
Perciformes: Gobiidae), and its Trophic Role in the Curonian Lagoon, Se Baltic Sea.
Acta Ichthyologica Et Piscatoria 43: 95–108.
Rauschmayer F, Behrens V. 2008. Legitimacy of speciesmanagement: the great cormorant in the EU. Legitimacy in European Nature Conservation Policy, pp. 55–74. Springer, Ray WJ, Corkum LD. 2001. Habitat and Site Affinity of the Round Goby. Journal of Great
Lakes Research 27: 329–334.
Robinette DP, Howar J, Claisse JT, Caselle JE. 2018. Can nearshore seabirds detect variability in juvenile fish distribution at scales relevant to managing marine protected areas?
Marine Ecology 39: e12485.
Rönnberg C, Bonsdorff E. 2004. Baltic Sea eutrophication: area-specific ecological consequences. In: Kautsky H, Snoeijs P (ed.). Biology of the Baltic Sea, pp. 227–241.
Springer Netherlands, Dordrecht.
Rosenblatt AE, Nifong JC, Heithaus MR, Mazzotti FJ, Cherkiss MS, Jeffery BM, Elsey RM, Decker RA, Silliman BR, Guillette LJ, Lowers RH, Larson JC. 2015. Factors affecting individual foraging specialization and temporal diet stability across the range of a large
“generalist” apex predator. Oecologia 178: 5–16.
Roth JE, Mills KL, Sydeman WJ. 2007. Chinook salmon (Oncorhynchus tshawytscha) — seabird covariation off central California and possible forecasting applications. 64: 11.
Sapota MR, Skóra KE. 2005. Spread of alien (non-indigenous) fish species Neogobius melanostomus in the Gulf of Gdansk (south Baltic). Biological Invasions 7: 157–164.
Schröder A, Persson L, de Roos AM. 2009. Culling experiments demonstrate size-class specific biomass increases with mortality. Proceedings of the National Academy of Sciences 106: 2671–2676.
Skora KE, Rzeznik J. 2001. Observations on Diet Composition of Neogobius melanostomus Pallas 1811 (Gobiidae, Pisces) in the Gulf of Gdansk (Baltic Sea). J Great Lakes Res 27: 290–299.
Srinivasan S, Chhatre SS, Guardado JO, Park K-C, Parker AR, Rubner MF, McKinley GH, Cohen RE. 2014. Quantification of feather structure, wettability and resistance to liquid penetration. Journal of The Royal Society Interface 11: 20140287.
Terraube J, Guixé D, Arroyo B. 2014. Diet composition and foraging success in generalist predators: Are specialist individuals better foragers? Basic and Applied Ecology 15:
616–624.
Veldkamp R. 1995. DIET OF CORMORANTS Phalacrocorax carbo sinensis AT WANNEPERVEEN, THE NETHERLANDS, WITH SPECIAL REFERENCE TO BREAM Abramis brama. Ardea 13.
Vetemaa M, Eschbaum R, Albert A, Saks L, Verliin A, Jürgens K, Kesler M, Hubel K, Hannesson R, Saat T. 2010. Changes in fish stocks in an Estonian estuary: overfishing by cormorants? ICES Journal of Marine Science 67: 1972–1979.
Waters CN, Zalasiewicz J, Summerhayes C, Barnosky AD, Poirier C, Ga uszka A, Cearreta A, Edgeworth M, Ellis EC, Ellis M, Jeandel C, Leinfelder R, McNeill JR, Richter D d., Steffen W, Syvitski J, Vidas D, Wagreich M, Williams M, Zhisheng A, Grinevald J, Odada E, Oreskes N, Wolfe AP. 2016. The Anthropocene is functionally and stratigraphically distinct from the Holocene. Science 351: aad2622–aad2622.
Wells BK, Santora JA, Henderson MJ, Warzybok P, Jahncke J, Bradley RW, Huff DD, Schroeder ID, Nelson P, Field JC, Ainley DG. 2017. Environmental conditions and prey- switching by a seabird predator impact juvenile salmon survival. Journal of Marine Systems 174: 54–63.
White CR, Day N, Butler PJ, Martin GR. 2007. Vision and Foraging in Cormorants: More like Herons than Hawks? PLoS ONE 2: e639.
Wickham H. 2009. ggplot2. doi 10.1007/978-0-387-98141-3.
Wickham H, Averick M, Bryan J, Chang W, McGowan L, François R, Grolemund G, Hayes A, Henry L, Hester J, Kuhn M, Pedersen T, Miller E, Bache S, Müller K, Ooms J, Robinson D, Seidel D, Spinu V, Takahashi K, Vaughan D, Wilke C, Woo K, Yutani H. 2019.
Welcome to the Tidyverse. Journal of Open Source Software 4: 1686.
