Seasonal and Daily Movement Patterns by Wels Catfish (Silurus glanis) at the Northern Fringe of Its Distribution Range

Abstract

Fish behavior often varies across a species’ distribution range. Documenting how behaviors vary at fringes in comparison to core habitats is key to understanding the impact of environmental variation and the evolution of local adaptations. Here, we studied the behavior of Wels catfish (Silurus glanis) in Lake Möckeln, Sweden, which represent a European northern fringe population. Adult individuals (101–195 cm, N = 55) were caught and externally marked with data storage tags (DSTs). Fifteen DSTs were recovered a year after tagging, of which 11 tags contained long-term high-resolution behavioral data on the use of vertical (depth) and thermal habitats. This showed that the catfish already became active in late winter (<5 °C) and displayed nocturnal activity primarily during summer and late autumn. The latter included a transition from the bottom to the surface layer at dusk, continuous and high activity close to the surface during the night, and then descent back to deeper water at dawn. During the daytime, the catfish were mainly inactive in the bottom layer. These behaviors contrast with what is documented in conspecifics from the core distribution area, perhaps reflecting adaptive strategies to cope with lower temperatures and shorter summers.

1. Introduction

Fish in lakes and rivers at high latitudes must cope not only with short summers but also with low winter temperatures, low light levels, and the formation of ice. Several species have adapted to these conditions, although some are regarded as warm-water fishes and are limited in foraging activity, growth, and reproduction by the low temperatures [1,2]. One such species is the Wels catfish (Silurus glanis, hereafter “catfish”), which is a relic from warm postglacial times in temperate ecosystems and naturally inhabits only three water systems in northern Europe, namely those in southern Sweden. Because the core distribution area for catfish is found in eastern, central, and southern Europe, these Swedish populations have been isolated for a long time [3]. Whether these isolated populations have adapted to the low temperatures or how their activity is affected by these cold conditions remain unknown.

Movement behavior allows fish to adjust to environmental changes and disturbances, as well as intra- and interspecific competition [4,5]. This includes fundamental abilities such as feeding, breeding, and predator avoidance [6], which govern fitness [7]. During climate change, movement can also be of vital importance for coping with unfavorable temperatures. The movement behavior of fish is consequently adapted to patterns of environmental variation at different temporal scales (i.e., seasonal or daily) that optimize growth, reproductive output, and survival [7,8,9,10]. The most prominent temporal pattern is the earth’s 24 h rotation around its own axis making up day (light) and night (dark). Many biological and environmental patterns are linked to this diel rhythm, such as temperature, light, food availability, and migration. To be able to anticipate and prepare for rhythmic events, organisms are controlled by a biological clock to “be on time”.

Biological clocks are controlled and adjusted by different factors. The most important factor is considered to be photoperiod [8,11], but temperatures [12] and access to food also help synchronize biological rhythms in both fishes and mammals [13,14,15]. The diel activity rhythm (diurnal, crepuscular, or nocturnal) in animals is species-specific but may also differ within species due to age, size, or social status [9,10]. Further, activity rhythms are not fixed but plastic, and individuals may change their rhythms over seasons, with food availability, habitats, or other environmental parameters [8,13,14,16].

In fish species like catfish, particularly for native Swedish populations, there are knowledge gaps regarding rhythms in movement and behavioral patterns on yearly and daily bases. Catfish are the largest freshwater fish in Europe. The species is well-documented to reach lengths of 2.7 m and weights of 130 kg in Europe [17], and rumored to reach very large sizes (length of 5 m). Catfish are apex predators with a varied diet and ability to adapt to new food resources [2,18,19,20,21,22]. There are several studies investigating various spatiotemporal behavioral patterns of catfish in Europe. Juvenile catfish have been shown to prefer nocturnal feeding in laboratory experiments, especially when in groups [13,23]. In the wild, results are varied, but a nocturnal predilection has been shown that varies over seasons and locations. For instance, juvenile catfish exhibited strict nocturnal patterns during spring, summer, and autumn in a Czech lake [24]. Another study of both juvenile and adult catfish found diurnal activity during winter and spring, all-hours activity during summer, and then nocturnal activity during autumn [25]. A recent radiotelemetry study in River Po, Italy also indicated increased nocturnal and movement activity during spring and summer compared to the rest of the year [26,27,28]. However, tracking energy usage patterns showed no consistent, detectable relation to the diurnal cycle [29]. In the Římov Reservoir, catfish were observed shifting from deeper waters during periods with low temperatures to shallower waters in warmer seasons. Additionally, they displayed activity peaks during both cold and warm seasons [28]. Disregarding the time of day, several studies agree that catfish movements and activity increase with increasing temperature [24,25,30].

Few previous studies of catfish ecology have examined movement patterns in native wild populations using a high frequency of measurements throughout the year. The aim of this study was to investigate yearly and daily activity patterns in catfish in Lake Möckeln, which is a northern fringe habitat in Sweden where catfish have been isolated from the main distribution area since the warmer postglacial era 9500–8000 years ago. This isolation has led to distinct genetic differences for catfish in Lake Möckeln compared to populations in southern and central Europe [31,32]. Here, we aimed to increase our understanding of whether this northern fringe population of catfish displays different seasonal and daily activity patterns compared to those documented in the core distribution area, thus reflecting latitudinal variation in temperature and light conditions. We hypothesized that we would see strict nocturnal activity and minimal diurnal activity over the seasons. Secondly, we expected to see long periods of low activity during the months with low temperatures (November–April), and activity peaking during summer (June–July) when the temperature was highest. Particular interest centered around the onset of activity in spring and the activity patterns during catfish spawning season.

2. Materials and Methods

2.1. Study Area

This study was conducted in Lake Möckeln (56°39′48.9″ N 14°8′57.7″ E), situated in the upper part of the Helge River system in southern Sweden. Helge River harbors one of three native populations of catfish in Sweden, with a population size of 720 ± 80 mature individuals [33]. Lake Möckeln has an area of 46 km² and a mean depth of 3 m with a maximal depth of 12 m. The lake is a mesotrophic (Tot-P 23 ug/L [34]), brown water (130 mg Pt/L [34]) lake with an annual mean temperature of 9.4 °C [34] (1982–2020). Other fish predators present in the lake are pike (Esox lucius), perch (Perca fluviatilis), and pikeperch (Sander lucioperca), whereas the main prey species are roach (Rutilus rutilus), bream (Abramis brama), zope (Abramis ballerus), and silverbream (Blicca bjoerkna).

2.2. Fishing and Tagging

Fishing was conducted with longlines in the deepest part of the lake (12 m depth), with the objective of capturing and equipping adult catfish (>100 cm) with data storage tags (DST G5, and G5 PDST, Cefas, Suffolk, UK) to record the depth (every minute) and temperature (every 5 min). Longlines consisted of a floating mainline (400 m) with monofilament (1.2 mm) leaders (1–1.5 m) attached every 10–20 m [35]. Leaders were fitted with single treble hooks and baited with native cyprinids. After capture, a DST was attached to the catfish using a braided fishing line (0.35 mm) tied around the base of the first fin ray of the pectoral fin. During August 2018, catfish (N = 20, size range 105–195 cm) were captured and fitted with G5 PDSTs (pop-ups) that were programmed to release after one year. Additionally, another 35 catfish were tagged in August 2019 (size range 101–167 cm) with a combination of both a G5 DST and a radio transmitter (transmitter F1580, ATS, Isanti, MN, USA) to facilitate recovery of the tags. In addition, all individuals were measured to the nearest cm (total length) and weighed to the closest 0.1 kg (Berkley 50lbs/22 kg, Columbia, SC, USA or Steinberg Systems SBS-KW-300/100-O, Berlin, Germany). A passive integrated transponder (PIT, Biomark, Boise, Idaho, USA, 23 mm HDX) was also injected into the pelvic girdle, or the abdominal cavity, to allow identification of recaptures regardless of whether the DST stayed attached.

Several searches for detached DSTs were conducted in Lake Möckeln and adjacent streams (2019, 2020), with the aid of a radio receiver (ATS, Isanti, MN, USA, R410) and antenna (ATS, Isanti, MN, USA, 5 element Yagi). Data from all recovered DSTs were initially checked by plotting both depth and temperature against time to determine when the DST had detached from the catfish. The DST was determined to be free-floating when the depth was constant (at the water surface) and showed no further amplitude changes, and the temperature showed indications of being affected by direct sunlight through sharp peaks during the daytime.

2.3. Estimates of Daylength

Seasonal patterns of daylength in the study area were based on photo-active radiation (PAR) data downloaded from the STRÅNG archive (Swedish Meteorological and Hydrological Institute, SMHI, Norrköping, Sweden), covering the period 1 August 2018–1 September 2020 (https://strang.smhi.se, accessed on 5 October 2022). STRÅNG provides hourly estimates of several radiation parameters each day of the year. Here, we defined PAR > 0 as day and PAR = 0 as night, resulting in a daylength varying between 7 and 17 h throughout the year.

2.4. Data Handling and Statistical Analysis

To estimate catfish activity, we utilized information about vertical movement (i.e., vertical activity) by calculating the delta values of depth (m) between each timestamp (every minute). To assess seasonal and circadian patterns in activity, we calculated the mean delta values per month (from m/min, Figure 1A) and hour of each day (i.e., 00-23, henceforth TOD), respectively. The calculations were carried out for each catfish.

To visually explore seasonal patterns in daily activity (i.e., circadian rhythm), we plotted mean activity per hour of the day by calculating a monthly mean activity per TOD (Figure 2). Visual inspections of plots indicated that activity generally peaked at night (Figure 2). Consequently, we hypothesized that the circadian rhythm could be described with a cosine function peaking in activity at night. The cosine function to model circadian activity patterns was specified as follows: circadian rhythm = cos(2πt/τ), where t is the time variable (TOD) and τ is the period of a cycle (i.e., 24 h). To investigate whether this variable could describe the hourly activity patterns (calculated for each individual and day), we performed a generalized linear mixed model (GLMM) with gamma distribution and a log-link function. Circadian rhythm and month were included as fixed explanatory variables with an interaction between them, and an individual was included as a random factor. This was performed with glmer in the lme4 package (v1.1-30) [36]. For this analysis, we added a constant (1) to the response variable because observations of zero activity cannot be used with gamma distribution [37].

Due to a significant interaction effect between the circadian rhythm and month, we analyzed each month separately using paired t-tests to evaluate if there were any significant differences in activity between hours categorized as night or day. Due to multiple testing, p-values were adjusted with the Bonferroni method in the p.adjust function in R. All data handling, statistical analysis, and graphics were performed with R (v. 4.2) [38] and RStudio (v2022.07.2) [39] using packages like dplyr (v1.0.10), ggplot2 (v3.3.6), ggpubr (v0.4.0), and lubridate (v1.8.0) [40,41,42,43]. Data and R-scripts are available in the Supplementary Materials (https://doi.org/10.5061/dryad.nk98sf82f).

3. Results

During 2018, catfish were tagged (N = 20) with a G5 PDST, and two of those loggers were recovered in 2019, which contained data for 260 days/individual. During the second tagging event in 2019, G5 DSTs with an attached radio transmitter were attached to catfish (n = 35), and 13 were recovered in 2020. In total, tags from 15 individuals (length 101 to 155 cm; seven males, six females, and two unidentified) were recovered containing temperature and depth data from 22 to 364 days. The recovery of pop-up DSTs was 10% after the first year, and the recovery of tags increased to 37% the second year after adding a radio transmitter. Of the 15 tags that were reclaimed, 11 tags represented data from a time span longer than 300 days while four DSTs contained less than one month of data.

The DST data showed marked variations in activity of the catfish during a full year. The mean vertical activity (delta values for depth changes) per individual and month, together with the ambient temperature experienced by the catfish, brought forth yearly activity patterns (Figure 1). High activity (0.198 to 0.162 m/min averaged per month) was recorded during the summer from July to August when water temperatures were about 20 °C. High activity (0.167 m/min) was maintained throughout September although the water temperatures decreased to 15 °C. The water temperatures (falling from 10 to 5 °C) and activity then decreased from October to November, from 0.140 to 0.087 m/min per month. During midwinter, in December and January, the catfish were inactive (0.050–0.041 m/min). Activity then continuously increased from February to April, starting at a water temperature <5 °C, from 0.083 to 0.108 m/min per month. In May, activity decreased to 0.085 m/min with increasing water temperatures but increased again to 0.138 m/min during June.

The GLMM revealed that catfish daily activity patterns varied across the year (the effect of interaction between the circadian rhythm and month: χ² = 2236.2, df = 11, p < 0.0001; Figure 2). Pairwise comparisons between day and night, repeated for every month, showed that the catfish were significantly more active during the night compared to the day (when the length of day varied between 7 and 17 h) during the months of February–March and July–October (paired t-test, p < 0.05; Figure 3). From April to June, however, no differences in activity between night and day were recorded. In December and January, when the catfish were passive, no diurnal activity occurred. However, as overall activity increased in the following months, a pattern of higher activity during the night and lower during the day emerged. This pattern of high night activity then became more pronounced during July–October (p < 0.01), but the pattern disappeared in November (not significant).

The high nocturnal activity from summer to autumn coincided with a diurnal activity pattern exemplified in Figure 4. During the day, the catfish were passive in the bottom habitat of the deeper part of the lake (Figure 5). At dusk, the catfish started to move and ascended quickly from the bottom (within minutes) to the surface water. During the night, activity was high with movements up and down at a scale of a few meters, whereas at dawn, the catfish returned to the bottom water. There were no indications of long horizontal movements, e.g., to the shallow littoral habitats (a few hundred meters from the deep part of the lake), which would have been recorded on a DST as a continuous decrease in depth over a longer time span.

4. Discussion

4.1. Daily Activity

Catfish in the core distribution area of central and southern Europe are known to display a circadian rhythm, with night activity and day resting being the predominant pattern, although this may vary across seasons and locations [24,25,26,44]. We found a similar behavior in this northern population for parts of the year. In late fall and early winter (November to January), the overall activity of catfish was low with no differences registered between night and day. In February, a significant circadian rhythm could already be distinguished, and the pattern continued throughout March. From April to June, however, no differences in activity between night and day were recorded. The spring period was characterized by increasing water temperatures in the lake and increasing day length (from 13 to 17 h).

A general and daily behavior of adult catfish was evident from July until the end of October: an ascension from the bottom to surface water at dusk, continuous and high activity during night, and descension back to bottom water layers at dawn. Those activity periods were often initiated with an extensive vertical migration of 6–8 m followed by high activity of minor vertical movements in surface waters. During the daytime, the catfish were mainly inactive in the bottom layer. This behavior was repeated over longer time periods, and occasionally interrupted by inactive periods during the night. High activity thus occurred during the night (darkness) for the period July–October. It has been shown experimentally that Wels catfish use the lateral line to detect and pursue prey [45,46]. Swimming by prey creates water movements that remain briefly in the water. These “wakes” are detected by catfish; the prey is then followed and attacked [45,46]. This predation behavior is disturbed by obstacles in the water, e.g., bottom vegetation, stones, or sunken trees. Consequently, catfish most likely hunt and forage in the open water column where prey, such as zooplanktivorous zopes and other cyprinids, can be detected by the lateral line. We, therefore, suggest that the observed movement of adult catfish from the bottom up toward the surface at dusk reflects their main foraging behavior during the warmer season. The fast ascension that was followed by high activity at night in the presumed open water column for several hours, along with the continued high frequency of vertical movement of a lower magnitude, is a behavior that we interpreted as prey search and/or hunting. At dawn, the catfish returned to deeper water near the bottom and remained inactive until the next foraging cycle was initiated the following night. This detailed activity pattern has not been shown for catfish in their central distribution area. During no part of the year were catfish in this northern fringe population more active by day than night; the latter behavior being reported for winter and spring in River Berounka in the Czech Republic [25]. Such differences in behavior may stem from differences in foraging strategies, which could be linked to variations in prey type and behavior. Alternatively, these discrepancies might represent different adaptations to competing needs, like behavioral thermoregulation, foraging, and predator avoidance. The isolated Swedish population under study has long been separated from the core distribution area [3] and is genetically differentiated from other European populations [31,32]. This genetic divergence raises the possibility of adaptations to the colder climate, potentially influencing behaviors and physiological functions.

4.2. Seasonal Activity

The highest activity during the year, and thus the most pronounced foraging, occurred in warm summer months and early autumn (July to October). The warmest period, however, was June when the catfish showed considerably lower activity than in July. Because the lower activity in June was preceded by even lower activity in May, this may indicate a pre-spawning and/or spawning behavior of the catfish [29], which we discuss further below.

The lowest activity during the year was registered in December and January, which were the cold winter months when ice generally covered the lake, with water temperatures at or below 4 °C. In February, their activity increased despite temperatures still being low (<5 °C), and this pattern continued during March and April at temperatures increasing up to 10 °C. The catfish were active in water temperatures below 5 °C, unlike conspecifics in the central distribution area of Europe, where catfish are reported to be “dormant” and inactive below 8–10 °C [2,47]. The demonstrated low-temperature activity in Swedish catfish may be adaptive for catfish in a northern population. Indeed, the ability to be active in lower temperatures increases the annual time window for foraging and energy gain for spawning and growth. Any trait that contributes to an ability to cope with long winters, with frequent ice cover events during the long lifetime of the catfish [1], will enhance survival and reproductive success. The main environmental variable affecting catfish in these areas—compared to the central distribution region—is the colder climate and thus a considerably lower annual water temperature. It is thus plausible to suggest that catfish in these peripheral, northern regions show adaptation to these harsh conditions, even though the catfish is defined as a warm-water species [2].

In the high water temperatures of July, the catfish once again displayed circadian rhythms with high night activity. This behavior included swimming from bottom to surface waters at dusk, foraging in free water during the night, and returning to bottom waters in the morning. This pattern of diel vertical migration is a known phenomenon among both freshwater and marine species, and is frequently demonstrated by planktivorous fish species [48,49]. It is plausible that the pattern of depth utilization observed here is indicative of catfish tracking the movements of their planktivorous prey. The circadian behavior with high nocturnal activity continued until the end of October and indicates intense foraging, as well as experiencing water temperatures decreasing from 20 to 10 °C. We suggest that the time span from July to October is the main foraging and growth period for catfish during the year. Catfish then became inactive for the coldest midwinter months when water temperatures were below 4 °C.

4.3. Do Early Summer Movement Patterns Coincide with Pre-Spawning Behavior?

The main population of adult catfish spawn in a small, shallow (<2 m), creek (Agunnarydsån, with a mean yearly water discharge of <2 m³/s in the northern part of the lake [33], where the outlet opens into a shallow bay. Pre-spawning behavior includes a migration from central parts of the lake (where tagging took place) to the spawning creek, which is a distance of at least 5 km [33]. This is also supported by tagged-catfish presence in the spawning stream during parts of May–June, where four individuals marked with both a DST and PIT were detected at a PIT station (in operation for another study). When catfish aggregate in these areas before and during spawning, vertical movements are restricted by the shallow water, and their movement activity (as defined) is decreased. Moreover, catfish utilizing this shallow habitat for reproduction is corroborated by the absence of deeper descensions in the DST pressure data and the day–night fluctuation in temperature, which is characteristic of shallow, running water being cooled at night and warmed during the day. The decrease in activity during May–June compared to the following summer months indicate that this time of the year is devoted to pre-spawning and spawning behavior in shallower areas and in the creek. A similar reduction of activity from mid-May to the end of June was also observed by Říha et al., and hypothesized to be linked to spawning [28]. The behavior was not dominated by night activity. The time spent at the spawning grounds varied for the adult catfish, from a few days to months.

5. Conclusions

Our findings suggest that the behavior of adult Wels catfish in this northern fringe population differs from those in the core habitats of eastern, central, and southern Europe. This adds to a previous study finding that these fringe populations differ by growth rate and longevity [1]. These behavioral discrepancies likely arise from distinct thermal conditions and seasonal patterns. Consequently, those may shift as conditions change with global warming. The management of these peripheral populations should prioritize enabling their ability to adapt to changing environmental conditions, which includes ensuring connectivity within and among habitats to facilitate behavioral flexibility for successful feeding and reproduction. Our results suggest that the catfish do not, as would be expected from a warm-water species in a northern peripheral population, seek the warmest habitats of the lake, which most likely would be shallow, vegetated, and sheltered bays. Instead, the catfish displayed distinct behaviors and rhythms, depending on the season, which may reflect adaptations to the colder climate. The activity of the catfish had already started in late winter when the lake is generally frozen, and was so at the time of the study. This was followed by the main foraging period in the summer and early autumn when the catfish utilized the deeper bottom habitat diurnally and the free water toward the surface for nocturnal hunting and foraging. These findings were made possible through the rare use of in situ mark–recapture using data storage tags, which provided high-resolution data on focal catfish behavior and opened up new research to understand its underlying mechanisms and adaptive value.