Records of ocean sunfish Mola mola along the Norwegian coast spanning two centuries, 1801–2015

Introduction

Ocean sunfishes (genus Mola Koelreuter 1766, Tetraodontiformes, Molidae), including the widespread ocean sunfish Mola mola (L. 1758), are shrouded in mystery, both in their ecology and taxonomy (Yoshita et al., 2009; Pope et al., 2010; Yamanoue et al., 2010). Over the past two decades several records of occurrence, distribution and strandings of M. mola from all over the world have been published (Al-Baz et al., 1999; Sims & Southall, 2002; Houghton et al., 2006; Dulčić et al., 2007; Fulling et al., 2007; Dube et al., 2013; Jawad, 2013; Bas et al., 2015). Apart from one Icelandic study (Palsson & Astthorsson, 2017) none of these were from cold waters such as those in the northern part of the North-East Atlantic Ocean. More data on the species' morphology and regarding correspondence between morphology and genetic identity (Sawai et al., 2017) are needed.

Although basically a tropical and subtropical species, M. mola, being a deep diver, has a broad thermal tolerance. Its lower tolerance limit to sea surface temperature (SST) has been estimated as 8–11° C (Sims et al., 2009; Dewar et al., 2010; Sousa et al., 2016a). Its preferred range of SST seems to be roughly 10–24° C and its optimum range around 15–20° C. It may dive to 800 m depth, but spends over 60% of its time in the upper 50 m of the sea (Potter & Howell, 2010). In one study, M. mola was active by day only, diving to depths of 100–200 m where it fed mainly on siphonophores (Nakamura et al., 2015). Radio-tracked specimens returned several times each day to warmer surface waters, where they regained a higher body temperature. Close to the surface, M. mola often lies passively on its side, perhaps to aid buoyancy as it lacks a swim bladder, or to increase the warming effect of the sun (Nakamura et al., 2015). On the other hand, Potter et al. (2011) found no relationship between time spent in colder waters (<10° C) and time spent near the surface. Another function of basking behaviour may be to allow birds to remove parasites, several incidences of which have been witnessed (Sims et al., 2009; Abe & Sekiguchi, 2012). Time spent at or near the surface will greatly influence detectability at sea.

As mainland Norway spans >13° latitude (from Mandal at 58° N to Kinnarodden at 71° N), it serves as a good, wide latitude case study of the occurrence and abundance of M. mola in the northern hemisphere. The Norwegian coast is mainly affected by the North Atlantic Drift, transporting warm water from the North Atlantic Ocean into the Nordic Seas and by the Norwegian Coastal Current (Blindheim, 2004; Sætre, 2007). The Norwegian Coastal Current, which flows northward within the top 50–100 m of the sea and near the coast, shows stronger seasonal and annual temperature fluctuations than the North Atlantic Drift (Sætre, 2007). The Institute of Marine Research (IMR) in Norway has operated several fixed stations along the coast, which measured SST and salinity (Aure & Strand, 2001). As an illustration, monthly averages during 1936–2000 for their station at Stad (62° 12′ N, 05° 05′ E) were 13·7° C with a salinity of 31·2 for August (the warmest month) and 4·7° C and a salinity of 33·1 for March (the coldest month). There are no specific upwelling areas along the Norwegian coast that can be expected to work as hotspots for M. mola, but large concentrations of jellyfish may occur during the summer both on the coast and in the fjords.

The aim of this study was to investigate recent and historical records of M. mola along the Norwegian coast, collected from all possible sources. The aim was to amalgamate as many recordings as possible. Specimens and records from natural history museums are considered the most reliable material. Most records of M. mola along the Norwegian coast have only been published in Norwegian, in Norwegian newspapers and magazines and hence have been largely inaccessible to non-Norwegian readers. The advent of the internet and social media facilitated distribution of more recent observations, but also challenged their quality in terms of possible multiple records of the same specimen.

The collected documentation was used to examine the distribution and occurrence of M. mola along the Norwegian coast and to investigate a possible association between occurrence and sea temperature. The species has a distinctive morphology and ecology, is easily recognizable and may possibly be an indicator species for sea warming. Records of M. mola in its northern distribution range are of particular interest in the context of a tropical and temperate species reaching colder waters at higher latitudes in the light of climate change and sea warming (Dulčić et al., 2007; Lipej et al., 2007).

Materials and methods

Records of M. mola were collected from various sources; scientific and grey literature, internet, media (newspapers, radio and TV through internet searches), bibliographic databases (including the National Library database for older Norwegian newspaper records back to 1931) and the Norwegian Biodiversity Information Centre (www.artsobservasjoner.no). Some records had previously been summarized (Collett, 1875, 1906; Dons, 1921, 1933; Hognestad & Vader, 1979; Syvertsen, 2002) and notes on individual records were also found (Lilljeborg, 1891; Johansen, 1915; Soot-Ryen & Grønlie, 1925; Hammerstad, 1953; Barth, 1963; Bakke, 1981). The earliest record was from 1801 (Collett, 1875) and the most recent from January 2016. This latter specimen must have been dead for quite some time, since late 2015, and was included as belonging to December 2015 in the analyses. The last web search was carried out 5 January 2016. Additionally, all known M. mola in Norwegian museums were examined and measured by personnel at each institution. Records of M. mola were also made available from the database of fish records (mostly observed from research vessels or fishing boats) kept at IMR. The growth of electronic media, including social networks, in the last decades has probably promoted the reporting of sunfish sightings. Hence, the numbers of recent and old records are most likely not directly comparable. The two world wars in this period probably reduced the number of M. mola reports, perhaps for substantial periods of time. The data used are shown in Table S1 (Supporting information).

Because reports of the M. mola are rare in Norwegian waters and due to its extraordinary and strange appearance, the species easily attracts attention from the public and media. Hence, most specimens are most likely correctly identified. Furthermore, most records in the media, even older ones, were documented by photos. Some of the records lacked an exact date, but nearly all could be assigned to a specific month and year. Some records lacked details about the geographic location, although most could be assigned to an approximate location or a region. Only very few records (c. five) had to be discarded due to insufficient information leaving doubt about species identification. Records were carefully screened to avoid duplicates, but multiple records of the same live individual observed at different sites were probably not entirely avoided (most notably in the Oslofjord in 2014). It is, however, likely that many more M. mola were captured by fishing boats, but discarded without the media being notified.

The museum specimens were measured according to Fig. 1 in Yoshita et al. (2009). Most specimens were stored in ethanol and had initially been fixed and preserved in formalin. Only a few were stored in freezers and measured frozen. Some media records also included a few measurements, such as mass, total length (L_T) and width (D), although some of these variables were only estimated (notably for specimens heavier than 50 kg).

The Norwegian coast was divided into four regions (Fig. 1) representing a South–North gradient; East (East of the southernmost tip of Norway), West (from the southernmost tip North to the Sogn og Fjordane county), Central (Møre og Romsdal, Sør-Trøndelag, Nord-Trøndelag) and North (Nordland, Troms, Finnmark). The reason for selecting these regions was to achieve a North–South gradient, with a reasonable sample size in each group (East can in several respects be considered to be the region furthest South). Generally, leisure activities decrease and commercial fisheries increase northwards along the coast. Data for sea temperatures were from IMR. This time series from the research station at Flødevigen was the longest available (1 m depth, daily average summer temperature for July–September 1925–2015, as shown in Fig. 1 in Aure, 2015).

The principal component-based indices of the North Atlantic Oscillation (NAO) for each month June to September were downloaded from www.climatedataguide.ucar.edu/ climate-data/hurrell-north-atlantic-oscillation-naoindex-pc-based. The NAO index provides a large-scale proxy of weather conditions in the North Atlantic Ocean, including changes in the surface westerly winds across the North Atlantic into Europe. Positive values of NAO are typically associated with stronger-than-average westerlies over the middle latitudes, more intense weather systems over the North Atlantic ocean and wetter–milder weather over western Europe (Hurrell et al., 2017).

Statistical methods include Pearson′s correlation (r), Pearson's χ² and non-parametric generalized linear model (GLMZ) with Wald χ² and Akaike′s information criterion (AIC). Significance level was set to P < 0·05.

Results

A total of 216 records of M. mola along the Norwegian coast were found (Fig. 2 and Fig. S1, Supporting information). With the exception of 1985, 0–5 records were made annually prior to 2000. All years in which more than five fish were recorded, excluding 1985, were in or after 2000 (7 years), with a maximum of 23 in 2014 (Fig. S1, Supporting information). Most records were from July–December (92·4%), but in the past decade some were also from January–April (total n = 197 fish with known month; Fig. 2). Most (20·8%) were from November, whereas none were from May. During the period 1879–2015, there were 54 years with no records of M. mola. The distributions of these were: 1879–1899 = 11 years, 1900–1929 = 16 years, 1930–1959 = 16 years, 1960–1999 = 10 years, 2000–2015 = 1 year.

Mola mola were reported along most of the Norwegian coast with the northernmost occurrence just North of Alta, Finnmark (c. 70° 44′ N, from December 1881; Fig. 2). Two simultaneous records were reported a little further North at Hammerfest, but this report was only a headline containing no further information and was discarded. Most records were from East Norway (Fig. 3), the most densely populated region of the country. Relatively many records were from North Norway, of which the majority were from Nordland. In East and West Norway, many were observations of live fish, whereas in northern Norway most were caught in fishing gear (Fig. 3). Decade and month were the factors affecting regional differences most (Table I), with a trend towards observations later in the year in the North (West–East–Central–North). There was no significant correlation between year and latitude (r = −0·13, P > 0·05, n = 200), i.e. no trend of more fish being found in the North in recent years.

A high percentage of the samples (43·5%) were individuals accidentally caught in fishing gear, 26·8% were found stranded and 29·7% were observed still alive at sea (n = 138). Recording method by month was tested for the months July to December. Fish caught in fishing gear occurred mostly in August–November, stranded fish were found mostly in November–December, whereas most observations of live fish were made in August (χ² = 45·9, d.f. = 10, P < 0·001, n = 124 cases). This suggests a trend of decreasing numbers of observations of live fish and increasing numbers of strandings (dead specimens) during the autumn. All M. mola recorded in January–March were found at latitudes South of 62° N (n = 13).

Most specimens whose size was recorded were small, <30 kg and <70 cm L_T (Fig. 4 and Table II). Only 10 were heavier than 100 kg and these were all recorded before 1960. The masses of the three largest individuals were estimated at 500 kg (two specimens) and 400 kg. Maximum L_T was 280 cm and maximum D 312 cm. Fish measured in museums (presumably the most accurate measurements) averaged 65 cm L_T and 27 kg (Table II). None of the overall size dimensions in Table II (L_T, D and mass) were correlated with latitude. Total length was significantly affected by decade and type of record, mass by decade only (Table I). Average size decreased after 1960 and specimens caught in fishing gear were smaller than those observed alive (when most sizes recorded were merely estimates).

The yearly numbers of M. mola were weakly correlated with corresponding annual mean sea temperatures at Flødevigen research station during July–September (r = 0·20, P < 0·05, n = 95 years). The correlation is no longer significant when 2014 (a deviant outlier) is excluded (r = 0·11, P > 0·05, n = 94). Including years from 1970 onwards, the correlation coefficient increases slightly, but becomes statistically insignificant due to a smaller sample size (r = 0·26, P > 0·05, n = 46). The frequency of M. mola was negatively correlated with the NAO index for August only, (r = −0·19, P < 0·05, n = 117) and not quite if 2014 was excluded (r = −0·17, P > 0·05, n = 116). For June, July and September: P > 0·05. This may reflect the fact that only 4 years had a positive NAO and 12 had a negative NAO for August 2000–2015.

Discussion

A remaining question is whether M. mola migration to the Norwegian coast is incidental, accidental or with a purpose. Arrival in the autumn is likely to reduce survival and the likelihood of returning South and may, therefore, appear accidental. A warming of waters along the Norwegian coast could have led to increased immigration by M. mola, but the correlation was barely statistically significant. The negative correlation with the NAO-index supports this conclusion, as the majority of years since 2000 had a negative NAO. A low NAO is associated with suppressed westerlies and a more southerly track of the North Atlantic Drift, which may potentially entice more M. mola to migrate from South to North. The effects in summer are smaller than in other seasons. The weak relationship between M. mola records and sea temperatures is not surprising, considering the nature of the data and the slow nature of changes in sea temperatures. Furthermore, sea temperatures further South, from where the fish emigrated, may have been of greater significance (see below). The SST of the Mediterranean Sea increased by about 1° C in the period 1982–2012 (Shaltout & Omstedt, 2014). The production or differential survival of year classes may also affect the numbers migrating and migration patterns. Mork (2016) concluded, ‘The heat content of Atlantic water in the Norwegian Sea has since 2000 been above the long-term mean’. In 2014, the year with the highest number of recorded M. mola, the summer temperatures in the upper layers along the West coast of Norway were 3° C higher than normal and the summer of 2014 was among the three warmest since recordings at permanent stations started in 1935 (i.e. IMR stations; Aure, 2015).

Upper-layer summer temperatures at Flødevigen research station (East region) have since 1990 been among the highest since recordings began in 1925 (Aure, 2015). This research station was assimilated by IMR in 1974. Its average SSTs were 16·6 ± 1·1° C (n = 93 years) for the months July–August and 2·3 ± 1·5° C (n = 92 years) for February–March. Sea temperatures at set depths were recorded routinely at several IMR stations (Aure & Østensen, 1993). Data from four of these show how SST decreases towards North and is halved on the way from the extreme South (Lista at 58° 01′ N; 08° 53′ E) to the extreme North (Ingøy at 71° 08′ N; 24° 01′ E). At 50 m depth, temperatures at all stations are below 10° C. Temperature differences between stations decrease with depth, with minor differences below 75 m. Below 100 m, winter temperatures equal summer temperatures. SSTs reach their maxima in August, when M. mola start to arrive on the Norwegian coast. From August to December, the °C SST is roughly halved, although the difference decreases further North. From October, all SSTs are below 10° C and hence outside the comfort zone of M. mola.

It is reasonable to imagine that in late summer and early autumn most individuals either turn South (Sims et al., 2009), or get trapped in colder water. If unable to swim South due to northbound currents or emaciation caused by an insufficient food supply, these fish may seek warmer water close to shore and in fjords. Many may end up dead on the shore or are found dying near the shore. In this study, the majority of stranded specimens were recorded in November and December. In Iceland, under similar oceanographic conditions, stranded specimens were found in July to December, although the number of recorded specimens was considerably smaller (n = 31; Palsson & Astthorsson, 2017). The small size of most fish along the Norwegian coast may indicate that small individuals are less capable of overcoming low temperatures than larger specimens. Basking at the surface to warm up might become more frequent as the sea cools, which may explain the relatively large number of observations of live specimens.

Increased sea temperatures due to climate change may lead to an increased northward migration, perhaps into cooler waters as well (Palsson & Astthorsson, 2017). Increased production of gelatinous zooplankton and jellyfish (Brodeur et al., 1999) or other potential prey of M. mola, might also result in an increase in numbers and an expansion of the species' distribution. M. mola migrates seasonally into higher latitudes (Sims et al., 2009; Pope et al., 2010; Potter & Howell, 2010), but it is not known whether those migrating furthest North survive and are able to return South. Mola mola may follow thermal gradients and move seasonally northward and southward into waters of preferred temperatures (Sousa et al., 2016a). One purpose may be to avoid excessively warm waters (exceeding 25° C) in late spring and summer, or to access upwelling areas and areas of zooplankton blooms that develop as waters warm at higher latitudes, e.g. in the North Sea (Sousa et al., 2016a). In the latter study, no M. mola occupied waters with SST below 9° C, which may explain their southward movement in the autumn. Mola mola is an active, migratory fish able to swim long distances (Cartamil & Lowe, 2004; Potter et al., 2011; Sousa et al., 2016a). Larger specimens migrate further and dive deeper than smaller ones, perhaps because more mass gives a better ability to cope with lower temperatures (Sousa et al., 2016a). Diet differs with size (age), with smaller individuals (<50 cm L_T) feeding on benthic crustaceans found closer to the shore and larger individuals (>200 cm L_T) on pelagic jellyfish (Syväranta et al., 2012; Nakamura & Sato, 2014; Sousa et al., 2016b). Most records along the Norwegian coast were of specimens <100 cm L_T. Collett (1875, 1906) examined the stomach content of some specimens, one of which contained a number of jellyfish, whilst the others were empty or contained seaweed.

In addition to the growth in the traditional mass media, two events may have promoted an increase in records of M. mola along the Norwegian coast: the establishment of natural history museums at the end of the 19th century and the growth of the internet and social media since the late 20th century which facilitated uploading of digital images. The latter factor could explain a part of the increased number of reports since 2000, but probably not all. Most likely, only a small proportion of the M. mola migrating along the coast of Norway is detected and many caught in fishing nets or on lines are probably discarded without any media being notified.

Not many M. mola appear to travel beyond the coast of Nordland at c. 68° N, as only very few were found further North in Troms or Finnmark. The species, however, travel further North in Norway than in Iceland, where only two specimens have been found North of 65° N (Palsson & Astthorsson, 2017). The northernmost record along the Norwegian coast was at about the same latitude as the most northern leatherback turtle Dermochelys coriacea, another jellyfish consumer (Frafjord, 2000). The regional differences in circumstances under which M. mola were found, either caught in fishing gear, observed at sea or found dead on shore, are likely to be related to the density of the human population and the activity of fisheries along the coast. The small size of most specimens and the fact that no large specimen has been found for several decades may seem surprising. Possibly, prior to 1960 larger specimens were more likely to be reported and measured, whereas smaller specimens attracted less attention. In addition, museums rarely have storage means or space sufficient for larger specimens. Exhibits of large fish mostly use replicas or casts and the original carcass may be discarded. Larger specimens would definitely attract attention, but the fact remains: all specimens reported since 1960 have been small in size.

In eastern Irish waters, a density of 0·98 M. mola per 100 km² surveyed was found and only small specimens were observed (Houghton et al., 2006; n = 68), although the authors concluded that the species was more common than previously thought. In the western English Channel, relatively few (n = 15) M. mola were observed, all were quite small (50–70 cm L_T) and most were observed in June and July (Sims & Southall, 2002). These observations support the notion of a regular northward migration of M. mola in the North-East Atlantic Ocean, some of these reach as far North as the Norwegian coast. While the species' occurrence along the Norwegian coast may still be considered a rare event, it does not appear to be unusual. A survey of its occurrence in the North Sea would be very interesting.

The authors would like to thank I. Byrkjedal for data and measurements of M. mola in the collections of Bergen University Museum, R. Wienerroither for data from the Institute of Marine Research and J. Aure, at the same institution, for data on sea temperatures. We are grateful to Å. Brabrand, M. Daverdin and R. Bergersen who assisted recording data from museum specimens at the Natural History Museum (Oslo), the NTNU University Museum (Trondheim) and Tromsø University Museum, respectively. M. Daverdin made the map in Fig. 2. Lastly, we greatly acknowledge the assistance of R. Barrett in improving the manuscript. The comments of two anonymous reviewers also improved the manuscript.