Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 7-Day Trial for You or Your Team.

Learn More →

Individual variability in the movement ecology of Northern pike Esox lucius in a highly connected wetland system

Individual variability in the movement ecology of Northern pike Esox lucius in a highly connected... Maintaining hydrological connectivity is important for sustaining freshwater fish populations as the high habitat connec- tivity supports large-scale fish movements, enabling individuals to express their natural behaviours and spatial ecology. Northern pike Esox lucius is a freshwater apex predator that requires access to a wide range of functional habitats across its lifecycle, including spatially discrete foraging and spawning areas. Here, pike movement ecology was assessed using acoustic telemetry and stable isotope analysis in the River Bure wetland system, eastern England, comprising of the Bure mainstem, the River Ant and Thurne tributaries, plus laterally connected lentic habitats, and a system of dykes and ditches. Of 44 tagged pike, 30 were tracked for over 100 days, with the majority of detections being in the laterally connected lentic habitats and dykes and ditches, but with similar numbers of pike detected across all macrohabitats. The movement metrics of these pike indicated high individual variability, with total ranges to over 26 km, total movements to over 1182 km and mean daily movements to over 2.9 km. Pike in the Thurne tributary were more vagile than those in the Ant and Bure, and with larger Thurne pike also having relatively high proportions of large-bodied and highly vagile common bream Abramis brama in their diet, suggesting the pike movements were potentially related to bream movements. These results indicate the high individual variability in pike movements, which was facilitated here by their access to a wide range of connected macrohabitats due to high hydrological connectivity. Keyword Pike movement ecology; River connectivity; Acoustic telemetry; Individual variability Introduction trends in fluvial connectivity are contrary to this, with most rivers around the world no longer being free flowing (Grill Longitudinal and lateral hydrological connectivity is impor- et al. 2019; Belletti et al. 2020). This disrupted connectiv- tant for sustaining fish assemblage structure and diversity ity tends to be associated with impacts on fish dispersal, (Amoros et al., 2002; Shao et al. 2019). However, global with barriers located throughout the aquatic system that * Simone Cittadino Environment Agency, Dragonfly House, Norwich NR3 1UB, [email protected] UK 2 South End Farm Cottages, Fishtrack Ltd, Department of Life and Environmental Sciences, Beccles NR34 8TG, UK Bournemouth University, Poole BH12 5BB, UK The Rivers Trust, Rain Charm House, Kyl Cober Park, Stoke Department of Ecology and Vertebrate Zoology, Faculty Climsland, Callington PL17 8PH, UK of Biology and Environmental Protection, University of Lodz, Lodz, Poland River Waveney Trust, Mill Ln, Pulham Market, Diss IP21 4XL, UK Department of Basic Sciences, Faculty of Fisheries, Muğla Sıtkı Koçman University, Menteşe, Muğla, Türkiye Environment Agency, Scarrington Road, Nottingham, United Kingdom Vocational School of Health Services, Eskişehir Osmangazi University, Eskişehir, Türkiye Environment Agency, Inworth Road, Rivers House, Threshelfords Business Park, Feering CO5 9SE, UK Vol.:(0123456789) 105 Page 2 of 13 S. Cittadino et al. can limit the movement of anadromous, catadromous, and being capable of consuming larger-bodied prey (Nilsson and potamodromous fishes (Fullerton et al. 2010). Conversely, Brönmark 2000; Nolan et al. 2019). Pike movements display the maintenance of natural flow regimes and an absence of considerable intra-population variation along a continuum of anthropogenic barriers provides high habitat connectivity, spatial behaviours, where some individuals remain confined facilitates fish dispersal processes, and supports complex to relatively small areas (less than 1 km of river length), trophic dynamics through increasing access of fishes to more while others move repeatedly over several km (Masters et al. diverse foraging habitats that alter inter- and intra-specific 2002). This variation can influence gene flow as highly vag- interactions, and also affects nutrient cycling and productiv - ile individuals move relatively large distances in search of ity (Thieme et al. 2023). In connected systems, the move- suitable spawning habitats and mates (Ouellet-Cauchon et al. ment of organisms across different habitat types allows for 2014). Additionally, it may increase their exposure to height- a more integrated food web where energy and nutrients can ened risks of both intra- and interspecific competition, which flow more freely through various trophic levels (Power and can further impact their population dynamics (Nyqvist et al. Dietrich 2002). 2018). Consequently, some pike populations consist of indi- High habitat connectivity enables individual fish to viduals with annual home ranges of less than 2 km, while express their natural traits and behaviours, which can vary others have home ranges exceeding 14 km (Sandlund et al. considerably between individuals (Amat-Trigo et al. 2024). 2016). While there is usually a positive relationship between This variability is expressed through differences in spatial home range size and body size, this does not necessarily behaviours where, for example, some individuals are rela- show allometric scaling. For example, large pike in a chalk tively sedentary, characterised by having small home ranges, stream in Southern England moved less than smaller individ- while others are more active and have much larger home uals when their body mass was taken into account (Rosten ranges (Gutmann Roberts et  al. 2019; Amat-Trigo et  al. et al. 2016). Longer distance movements of pike tend to be 2024). Individual variability in the trophic niche often mani- for spawning, with populations showing discernible parti- fests as generalist populations being composed of relatively tioning between foraging and spawning areas (Craig 1996), specialised individuals whose niches are small subsets of which often include tributaries, side channels and ditches the population niche (Araújo et al. 2011). Consistency in (Nyqvist et al. 2018; Oele et al. 2019). In the Baltic Sea, individual specialization has been argued to have important resident pike spawn in brackish coastal waters, while there is ecological, evolutionary, and conservation consequences, an anadromous form that spawns in freshwater streams and including reducing levels of intraspecific competition and wetlands (Lappalainen et al. 2008). High habitat connectiv- potentially improving population resilience through the min- ity is thus important in enabling pike to express individual imising competitive pressures (Vander Zanden et al. 2010). specialisation and to maintain their access to suitable spawn- Reproductive activities and trophic interactions within ing areas that then serve as nursery grounds. populations can also profoundly shape the spatial habitat use In this study, the aim was to describe the movements of of fish populations (Speed et al. 2010), such as instigating pike in a protected and highly connected wetland system, the seasonal migratory events for spawning (Winter et al. 2021a) lower River Bure network in eastern England, recognized for and more regular movements for foraging (Nunn et al. 2010). its ecological importance and designated as a Ramsar site. While daily foraging movements can be relatively limited in This aim included assessment of the influence of spawn- distance, the extent of spawning migrations can vary widely, ing, connectivity, individual traits and trophic ecology on ranging from a few kilometres for potamodromous species to the individual variability in movements between individual over 1000 km for diadromous species (Verhelst et al. 2018; pike. The study system was the lower River Bure network van Puijenbroek et al. 2019). However, traits, such as sex, in eastern England, a highly connected, free-flow wetland age, length size and behaviour, can substantially influence system comprising the main River Bure, connected side the distribution, behaviour and movements of individuals channels and stillwater habitats (Fig.  1). The movements (Koed et al. 2006; Nyqvist et al. 2020). The interplay of of pike were measured using acoustic telemetry, a key tool these variables thus has important implications for the spa- for assessing the movements of animals in aquatic environ- tial ecology of individuals, populations and communities. ments (Brownscombe et al. 2022), and complemented with The Northern pike Esox lucius (‘pike’) is a freshwater the ecological application of stable isotope analysis (SIA), apex predator with a Holarctic range, being encountered which provides information on trophic relationships and diet across Europe and North America, and is recognised as a (Layman et al. 2012; Costantini et al. 2018). The objectives keystone piscivore (Craig 2008). Their diet can be highly were to assess the individual variability in the movement plastic, with the consumption of prey that includes mac- metrics of pike over a 3-year period and the influence of roinvertebrates as well as fish (Pedreschi et  al. 2015). body size, spawning season, and individual trophic ecology Pike prey size is typically limited by gape size, and there on this. We posit that larger pike will move greater distances, are ontogenetic shifts in their diet, with larger individuals have larger home ranges, and consume larger-bodied prey Individual variability in the movement ecology of Northern pike Esox lucius in a highly connected… Page 3 of 13 105 Fig. 1 Map illustrating the study area within the Bure system, situ- ing the location of the station for environment variables (HOBO® ated in Broad National Park, with red squares indicating the locations Pendant; model MX2202, Onset Computer Corporation). The red where acoustic receivers were deployed and the red triangle indicat- arrows indicate the general flow compared with smaller pike, and all pike will move greater promotes sustainable land use practices (https://rsis. r amsar. distances during spawning versus non-spawning periods. org/r is/6 8). The system features more than 60 km of barrier- The study then quantifies the use of different macrohabi- free main river, with a complex network of secondary slow- tats by pike and explores how this habitat use varies across flow channels (< 2 m3/s) and numerous small shallow lakes the study area. In completing these objectives, an increased (4 m depth), commonly known as “Broads” (Natural Eng- understanding of pike behaviour and habitat use in lowland land, 2020). As a consequence, the system has high lateral river systems is generated, providing important information and longitudinal connectivity, with the rivers intersected by on the management of wetlands, including the maintenance a dense network of lateral broads and dyke systems (Fig. 1). of hydrological connectivity. The land use of the Broads primarily includes woodland and agriculture, with approximately 40% comprising of wet grassland grazed by livestock. As the lower Bure flows Methods downstream, it transitions into semi-artificial grazing marsh and the water becomes more brackish owing to groundwater Study area interaction with seawater (Winter et al. 2021a). The study area is influenced by tidal activity owing to its proximity to The study area was the northern region of the Broads the sea, lack of barriers and the low-lying, relatively homo- National Park, Eastern England, comprising the lower geneous topography of the landscape. The area is therefore River Bure and its tributaries of the Rivers Ant and Thurne prone to saline intrusion during large spring tides and storm (Fig.  1). This important wetland area has Ramsar status, surges in winter. The salt edge can be pushed more than which provides protection against habitat degradation and 10 km inland during strong saline incursion events. 105 Page 4 of 13 S. Cittadino et al. the main rivers longitudinally (to enable long distance move- Fish capture, acoustic transmitter implantation and receiver network ments to be tracked). Some receivers were also positioned at the mouth of lateral connections (e.g., at the entrance of the Pike were sampled in the three main rivers (Bure, Ant and broads from the main river) and inside channels considered to be important for pike foraging and spawning (based on Thurne; Fig. 1). Rod and line angling was used to capture the fish as it was more effective in capturing large-bodied extant knowledge of the authorship team) (Fig. 1). Receivers were attached to permanent structures, moored on naviga- individuals compared with seine netting and electrofishing compromised by the increasing conductivity downstream. A tion posts or suspended from floating objects, at approxi- mately mid-water depth (1.0–1.5 m) to maximise detection total of 44 pike were captured; 16 were captured in Novem- ber 2017, 27 in January 2018 and 1 additional pike was cap- efficiency during tidal changes in river levels. Detection data were downloaded every 4 months and batteries replaced tured in September 2018 (Table S1). Before surgery, all fish were examined by the operator to assess any signs of stress annually. Movement data of the fish were collected from 2018 to 2020. All receivers remained operable throughout and whether the length of the individual was appropriate for the size of the acoustic tag (all tagged fish were > 500 mm). the study period and prior to any data analyses, the detection data were processed in actel R package (R version 4.2.3) Under general anaesthesia (Tricaine methanesulfonate, −1 MS-222; approximately 0.04 g  l ), the pike were measured to remove any false detections (e.g. transmitter code errors caused by code collisions) (Flávio and Baktoft 2021). (fork length, nearest mm), sexed (external identification; Casselman 1974) and scale samples and pelvic fin biopsy Stable isotope analysis and abiotic data collection taken (equivalent to approximately 1 mg dry weight), with the fin tissue frozen for storage. Each pike was then surgi - The fin biopsies of the pike, common bream and roach were cally implanted with an acoustic transmitter (Vemco V13; 69 kHz; 36 mm length × 13 mm diameter; mass in water: individually used for stable isotope analysis (SIA). After drying at 60 °C for 24–36 h, the samples were ground and 6.0 g; random transmission interval approximately 90 s; estimated battery life: 1200 days; manufacturer: Innovasea, weighed to 1000 µg in tin capsules. These samples were then analysed for δ13C versus VPDB and δ15N versus At. USA). The transmitters were inserted ventrally anterior to the pelvic fins, and incisions were closed using a single Air (%) at the Cornell University Stable Isotope Laboratory, New York, USA, using a Thermo Delta V isotope ratio mass suture (2–0 absorbable monofilament; Ethicon Ltd, USA) and wound sealer (Orabase; ConvaTec Ltd, UK). During spectrometer interfaced with an NC2500 elemental analyser (CE Elantach Inc., USA). As C:N ratios were < 3.5, no lipid surgery, instruments were sterilised between fish though immersion in a 10% iodine based solution (Videne antisep- correction was applied. The same preparation and analy- ses were completed for the macro-invertebrate samples, a tic solution; Ecolab Ltd, UK), followed by rinsing in 0.9% saline solution. During surgery, scales were removed from minimum of four replicates each capture site and date, with amphipods (Gammaridae) the most represented group, with the incision site to facilitate the entry of the scalpel and sutures. Following completion of the surgical procedure, the other samples including killer shrimp Dikerogammarus vil- losus (Table S2, Winter et al. 2021c). In subsequent analy- fish were then transferred to tanks containing river water, held until normal body orientation and swimming behaviour ses, SI data of gammarids and killer shrimp were combined into a single group (‘Amphipods’), as differences in their resumed. After a visual inspection by the operator fish were released back to their approximate area of capture within 2 h values were not significant (t -test, p-value > 0.05). Abiotic data were recorded by a stationary receiver of their tagging. During the same sampling periods, scales and fin biopsies were collected from 83 common bream (HOBO® Pendant; model MX2202, Onset Computer Cor- poration), maintained by the Environment Agency, and (Abramis brama) and 47 roach (Rutilus rutilus) coming from the three different rivers. Additionally, macroinvertebrate located in the main River Bure (52°38′56.5"N 1°34′03.5" E). Water temperature (± 0.5 °C), salinity and conductivity samples were gathered from six distinct sites along the riv- ers that cover the main River Bure, Thurne and Ant, with (us/cm) were recorded every 15 min throughout the entire study period. sampling between 11 September and 3 October 2018 using a sweep net for stable isotope analysis (Table S2, Winter Pike movement metrics et al. 2021a,b,c). All surgical procedures were conducted by the same surgeon and in accordance with UK Home Office Of the 44 pike tagged, 41 were included in the calcula- project license 70/8063 and after ethical review. To track the movements of the tagged fish, a fixed net- tion of movement metrics, with the remaining 3 individu- als excluded owing to their lack of detections (Table S1). work of 44 Vemco VR2W acoustic receivers was deployed in October 2017 and retrieved in summer 2020. These receivers Data were downloaded from receivers using VUE software (version 2.7.0), and subsequent data manipulation was were strategically placed to provide coverage of the length of Individual variability in the movement ecology of Northern pike Esox lucius in a highly connected… Page 5 of 13 105 conducted using the Vtrack R package (Campbell et al. using the adonis2 function implemented in the R package 2012), which enables the detection data of Vemco acoustic vegan (Oksanen 2012). transmitters to be assimilated across all receivers and then analysed appropriately (Udyawer et al. 2018). To calculate Statistical analyses river distances between detection points, the distance matrix was computed using the Field Calculator function in The data collected in the first five days following each tag- QGIS (version 3.28.8). The analyses completed here were ging event were excluded from the sample (in case the fish a series of movement metrics as per Gutmann Roberts et al. were demonstrating unusual post-surgical behaviours). Only (2019) (Table S1). Three indices were calculated to assess pike that were detected for a period of at least 100 days (as fish residency within the receiver array: the residency index, days between the first and the last detection) were included the linearity index and the residence array index (Table S1; (n = 30). This arbitrary threshold was chosen to focus on Craig 2008; Acolas et al. 2017). The residence array index long-term movements and exclude shorter detections that was included to address instances where a tagged fish could bias the results. The three main movement variables might have been within the receiver array but remained of each individual (TR, TM and MDM) were first tested undetected on a given day. The determination of total range at univariate level against total length (α = 0.05) and then (TR), expressed as the river distance (m) between the far- included as response variables in different models and tested thest receivers where an individual was detected, served as against the predictors (and their combinations) of the river of a proxy for home range. Additionally, total movement (TM), capture, spawning or non-spawning period (spawning period defined as the total distance covered by an individual, was covered the pre-, spawning and the post-spawning period: 1 computed by summing individual movements, regardless of December to 30 April; non-spawning period: 1 May to 30 their directionality (i.e. upstream, or downstream). In calcu- November), water temperature, salinity, fish identification lating mean daily movement (MDM), to avoid overestimates, (transmitter code), fork length and the proportion of com- only those occasions when the time between one observation mon bream in diet (from stable isotope analyses; see below). and the next did not exceed 24 h were considered. With the response variables based on fish being detected Kernel density estimation (KDE) was implemented to for at least 100 days, the number of days of detection was delineate the primary areas of pike spatial occupancy within not included as a predictor of any movement variable as the the study area. This method involves computing the den- relationships were all non-significant (Pearson’s correlation sity of point features surrounding each output raster cell. coefficient, all r < 0.33, P > 0.05). Before model fitting, data A continuous, smoothly curved surface is then fitted over exploration was undertaken following the protocol described each point, with the surface exhibiting its highest value at in Ieno & Zurr (2015) involving examination for missing the point’s location. This value gradually decreases as the values, outliers in the response and explanatory variables, distance from the point increases, eventually tapering off homogeneity and zero inflation in the response variable, col- to zero at a distance equal to the specified search radius linearity between explanatory variables, the balance of cat- (Cantrell et al., 2018). KDE was performed by dividing the egorical variables, and the nature of relationships between individuals according to their capture site (Bure, Ant and the response and explanatory variables. Thurne) and by dividing detections according to whether To test differences in movement metrics between pike they were in the spawning period (1st Dec–30th Apr) or spawning and non-spawning periods, all movement metrics non-spawning period (1st May–31st Nov). These analyses were subjected to analysis with ‘spawning’ designated as were all conducted in ArcMap version 10.8.2. and gener- the predictor variable. The best fitting model for the total ated heat maps that displayed the pike’s spatial occupancy range of movement (‘TR’) was a linear mixed-effects model probability. To assess variations in how the pike use the dif- (LMM, ‘lme4’ R package), which was chosen due to the ferent macrohabitats across the study area, the receivers on normally distributed nature of the TR data and to account which they were detected were then categorised according to for individual variability by including ‘FishID’ as a random whether the receiver was located in the main river (‘main’; effect (Table S3). To assess how pike utilized different mac- as the Bure mainstem, and the Ant and Thurne tributaries), rohabitats during spawning versus non-spawning periods, in ‘broads’ (the laterally connected, shallow, lentic habitats) a generalized linear mixed model (GLM) with general- or within the ditch and dyke systems (‘lateral’). For each ized Poisson regression was employed. In this model, the river, the receiver detection data by macrohabitats were response variable was the number of detections per receiver. summed and converted to proportions (%). To see the dif- The predictors included the macrohabitats of the main riv- ferences in number of detections between habitats and wild ers (‘main’, comprising the Bure mainstem, Ant and Thurne populations, and isotopic signatures between the sites, we tributaries) and the laterally connected habitats (‘lateral’, used a permutational univariate analysis of variance (PER- which included the broads, ditches, and dyke systems). For ANOVA), with Euclidean distance and 9999 permutations the analysis, the macrohabitats ‘broad’ and ‘lateral’ were 105 Page 6 of 13 S. Cittadino et al. combined as this enabled evaluation of the overall use of in the individual diet of pike. The proportion of bream in secondary connections versus the main river channels. As pike diet was entered as the response variable in a GLMM the total number of detections per receiver was not necessar- to investigate potential associations with pike fork length. ily related to the number of fish detected (e.g. one individual The predictor ‘river’ was also incorporated into the model to could have been detected multiple times) the number of pike account for variations among the three primary rivers under detected at each receiver in each period was included as a study. This approach aimed to elucidate potential correla- random effect. In addition, two other random effects were tions between the dietary habits of pike and their size while included in the model: the area of river the receiver was considering differences within the study area. located (‘River’; as Bure, Ant and Thurne) and the antenna receiver identification number (‘ReceiverID’). The relationship between pike length and movement pat- Results terns was analysed using a generalized linear mixed model (GLMM) with a gamma distribution. In this model, total Fish length and detections movement (TM) was the response variable and fork length was the fixed factor. The GLMM was chosen to address The mean fork length of the 41 pike analysed (31 female, the high variance and potential overdispersion in the TM 9 male, 1 undetermined) was 772 ± 122  mm (range data, providing a more accurate assessment of movement 583–1143 mm), with females significantly larger than males patterns. TM was used as other studies have observed that (t-test, t = −5.111, p < 0.01) (Table S1). Detections of the larger pike tend to move significantly more than small pike tagged pike occurred across 933 days, with almost 3 million (day of detection and TM are autocorrelated) (Vehanen et al. of individual detections, which were detected on individual 2006; Craig 2008; Sandlund et al. 2016). An additional inter- receivers between 1117 and 521,771 times. Tracking periods action term between pike length and river of capture (as were highly variable between individuals (3–931 days), with Bure, Thurne or Ant) was incorporated to accommodate individuals being detected between 2 and 172 days of detec- potential disparities in movement tendencies across die ff rent tion, with 30 fish having more than 100 days between their river locations (Table S3). To then test the effect of abiotic first and last detection (Table S1). variables on pike movement, a generalized additive model (GAM) was constructed using the 'mgcv' R package; a Movement metrics and macrohabitat use GAM was used owing to its ability to capture the nonlinear effects of temperature and salinity. Mean daily movement Movement metrics varied between individuals (total range: (MDM) was the response variable, given its sensitivity to 1487–26,099  m; total movement 8032–1,182,287  m; fluctuations in environmental conditions, water temperature mean daily movement 46–2948  m; Table  S1). For fish and salinity (daily mean) were included as predictor vari- detected for at least 100 days, Thurne pike had the larg- ables. Since the metrics are temporally repeated (across the est total ranges [mean ± 95% confidence interval (CI): season) to address individual variation, the fish unique iden- 16,076 ± 5782 m], with Ant pike having higher total ranges tification code (FishID) was integrated as a random effect (8141 ± 4951 m) than those in the Bure (5743 ± 1992 m). within the model structure (Table S3), with the number of This was also reflected in the patterns of mean daily distance knots set to 5 for all parameters to prevent overfitting. The moved (Thurne: 1524 ± 383 m; Ant: 1261 ± 658 m; Bure: model fitting process aimed to identify smooth trends that 717 ± 403 m). At a univariate level and across all fish, the optimally captured the data while balancing goodness of fit movement metrics of total range, total distance moved and and smoothness, with the final, best-fitting model identified mean daily distance moved were not significantly related to by backward selection using AIC (ΔAIC ≤ 2). fish length (linear regression, P > 0.05 in all cases, Fig. S1). The heat maps of pike spatial occupancy indicated most Analyses of stable isotope data fish remained in the river where they were initially caught in both spawning and non-spawning periods (Fig. 2, Fig. To predict the dietary contribution of the putative prey spe- S2). Individuals captured in the main River Bure were never cies (common bream, roach, and macroinvertebrates) to detected in the other tributaries, while some individuals pike diet, stable isotope mixing models were applied using from the Ant and the Thurne were detected in the main Bure the R package simmr (Parnell and Inger, 2016; Flávio and in both the spawning and non-spawning seasons (Fig. 2, Fig. Baktoft 2021). To achieve this, we adopted a prior-free S2). There were substantially more detections of pike on approach and applied trophic discrimination factors (TDFs) receivers located in the laterally connected macrohabitats suggested by Post (2002), specifically 1.0 for δ C and 3.4 than the main river (PERANOVA: p = 0.043), but the num- for δ N. Results of the mixing models are presented as a bers of pike detected across these macrohabitats were similar posteriori distribution for the proportion of each prey item (PERANOVA: p = 0.9) (Table S4). The GLM assessing the Individual variability in the movement ecology of Northern pike Esox lucius in a highly connected… Page 7 of 13 105 Fig. 2 Heat maps of pike occupancy during the spawn- ing (SW) and non-spawning period (NSW) in the river Bure system according to the loca- tion of tagging (Bure, Ant and Thurne), where the probability of occupancy ranges from absence (blue) through to low (green), medium (yellow) and high (red). Approximate release locations are noted with the black lines number of receiver detections across these macrohabitats (Table 1B). Mean daily distances moved were significantly revealed the influence of spawning season was not signifi- and positively influenced by water temperature (GAM, cant (P = 0.82), but the macrohabitat predictors were signifi- P = 0.03; Table 1C), with the fish identification interaction cant (‘Lateral’, P = 0.01; ‘Main’, P = 0.003), indicating the term also being significant in the model, emphasising the use of these different macrohabitats was important all year individual variability in the dataset. round (Table S5). Testing the relationships of the movement metrics in Pike diet composition by river multivariate models revealed spawning season had a sig- nificant and positive effect on total range (LMM, P < 0.01; Mean δ C of all pike was −27.55 ± 1.54% (−28.75% to Table 2A), with the total range being on average 36% greater −24.83%) and mean δ N was 19.45 ± 1.05% (17.86% to in the spawning season than on the non-spawning season. 21.69%), with differences between sexes being not sig- 13 15 The fish identification interaction term was also significant, nificant (t-test: δ C, t = −0.26, p = 0.80; δ N, t = −0.89, indicating that although movements were greater in spawn- P = 0.39). Across all pike, there was a significantly 13 2 ing periods, this varied between individuals. The GLMs higher value of δ C as pike length increased (R = 0.16; indicated pike length had a significant and positive effect F = 6.41, P = 0.02), but with no relationship between 1,34 15 2 on total movements (P = 0.03), where the interaction term pike length and δN (R = 0.03; F = 0.90, P = 0.35) 1,34 between length and capture location was significant for (Fig. S3). The common bream analysed for their stable the River Thurne (P = 0.02), indicating larger pike in this isotope ratios were relatively large putative prey items ver- tributary moved more than those from the Bure and Ant sus roach (mean lengths ± 95% CI: bream 388 ± 13 mm, 105 Page 8 of 13 S. Cittadino et al. Table 1 Results of statistical (A) Total range models on pike movements. (LM) TR ~ SW + R + (1 | FishID) (A) Linear model assessing Predictors Estimates 95% CI P the relationship between total range (TR) and the predictors (Intercept) 5547.59 1710.23–9384.94 0.006 spawning season (SW) and R [BURE] -2800.53 -7602.50–2001.44 0.246 river of capture (R), accounting for individual variability with R [THURNE] 5244.21 −342.87 to 10,831.29 0.065 the unique identification code SW [S] 2227.64 765.54–3689.74 0.004 (FishID) included as a random Random Effects effect. (B) Generalized linear σ 6,159,254.14 mixed model (GLMM) testing the influence of total length τ   24,905,982.87 00 Transmitter (TL), river of capture (R) and (B) Total movement spawning season (SW) on total (GLMM) TM ~ TL*R + SW movement (TM), with an added Predictors Estimates 95% CI P interaction term (TR*SW) (Intercept) 915.67 32.91–25,477.95 < 0.001 to explore potential joint effects. (C) General additive TL 1.01 1.00–1.01 0.003 model (GAM) evaluating the R [BURE] 0.24 0.00–29.08 0.562 impact of temperature (T) and R [THURNE] 76.74 1.27–4624.76 0.038 salinity (SAL) on the mean TL × R [BURE] 1 0.99–1.01 0.833 daily movement (MDM) of pike. Significant P values are TL × R [THURNE] 0.99 0.99–1.00 0.022 indicated in bold (C) Mean daily movement (GAM) MDM ~ s(T) + s(SAL) + SW + s(FishID, bs = ‘re’) Predictors Estimates 95% CI P (Intercept) 834.05 491.33–1415.83 < 0.001 SW [S] 1.98 1.38–2.84 < 0.001 Smooth term (T) 0.034 Smooth term (SAL) 0.193 Smooth term (FishID) < 0.001 range 212–502 mm, n = 83); roach: 131 ± 11  mm, range Discussion 78–229  mm, n = 78). The common bream showed also a significantly higher δ C ratio compared to roach Across the 30 pike used in movement analyses, there was (mean −28.78 ± 0.34 versus −29.92 ± 0.52%; t = 3.58, high individual variability, with the pike moving total dis- P < 0.01), but with no significant difference in δ N tances from 18 to 1182 km, with their mean daily move- ( me an 1 7. 05 ± 0. 3 2 ver su s 17 .2 9 ± 0. 57 %; t = − 0 .72 , ments being between 340 m and 3 km, with total ranges of P = 0.48). The isotopic signatures of Bream and Roach 6–26 km. The individuals that moved more tended to be showed no significant differences between the Bure and larger-bodied and were predicted to consume larger prey, Thurne sites (PERANOVA: P = 0.9 and P = 0.18, respec- with all movements tending to be greater during spawning tively). However, both species exhibited distinct isotopic versus non-spawning periods, with these results generally ratios in Thurne and Bure when compared with those in consistent with the predictions. The individual variability Ant (PERANOVA: P = 0.0015). in pike movements and spatial behaviour observed here The diet composition predictions from stable isotope is also evident in other pike populations. For example, in mixing models suggested that common bream was an the River Frome, southern England, pike revealed a con- important prey item for pike from the River Bure and tinuum of spatial behaviours, where some individuals were Thurne (Fig.  3), where there was a significant rela- always recorded in the same river reach of < 1 km length tionship between increasing pike length and increased while others moved repeatedly over several km (Masters predicted bream dietary contribution (GLM, P < 0.02; et al. 2002). In a Norwegian connected reservoir and river Table  2). Conversely, pike in the River Ant were pre- system, most pike had annual home ranges of less than dicted to have relatively high dietary contributions of 2 km, but with some individuals undertaking migrations of macroinvertebrates. Individual variability in the movement ecology of Northern pike Esox lucius in a highly connected… Page 9 of 13 105 Fig. 3 Predicted proportions (from stable isotope mixing models) of roach, common bream (‘Bream’) and macroinvertebrates (‘Macro’) in the diet of pike in the samples sites of a River Ant, b River Bure and c River Thurne. Whiskers display 95% confidence intervals over 14 km (Sandlund et al. 2016). Indeed, individual vari- ability in movements are also being apparent in a broad spectrum of species across aquatic and terrestrial environ- Table 2 Results of the general linear mixed model (GLMM) assess- ments (Shaw 2020). In entirety, these results emphasise ing the relationship between the proportion of bream in pike diet the importance of maintaining high habitat connectivity in (‘Bream’) and two predictor variables: pike total length (TL) and the wetland systems to enable pike and other fishes to access river of capture (R). Significant P values are indicated in bold a range of functional habitats over relatively large spatial (GLMM) Bream ~ TL + R Total length versus bream proportions areas. Although some of the individual pike in this study were Predictors Estimates 95% CI P highly vagile, there was minimal mixing of pike between the (Intercept) 0.27 0.21–0.34 < 0.001 three rivers. Pike tagged in the Bure were never detected in TL 1 1.00–1.00 0.048 the Ant or Thurne, while pike tagged in the Ant and Thurne Site [Bure] 1.19 1.09–1.31 < 0.001 were only rarely detected in the main Bure. This might Site [Thurne] 1.08 0.98–1.19 0.12 suggest a metapopulation structure, with fish in the Bure, 105 Page 10 of 13 S. Cittadino et al. Thurne and Ant comprising three distinct subpopulations to more permanent waters as the waters recede, although with limited dispersal among them. This partitioning of pike some individuals will move earlier to either avoid cannibal- into local groups with little ecological connectivity between ism (Cucherousset et al. 2009; Nilsson et al. 2014) or owing them was also detected by Lukyanova et al. (2024) in both to competitive displacement (Nyqvist et al. 2020). In the freshwater and brackish environments. Moreover, strong pat- Bure wetland, however, spawning period had no influence terns of site fidelity in pike are often detected in telemetry on the number of receiver detections in the main river ver- studies, where movements away from core areas are limited, sus the laterally connected macro-habitats, with these lateral especially outside of the reproductive period (Miller et al. habitats important all year round. The fish assemblage of the 2001; Kobler et al. 2008). For example, in the River Yser, ditch and dyke systems of the study area have previously Belgium, individually tagged pike demonstrated preferences been shown to be dominated by small numbers of pike, with for specific regions in the river in which they were detected their presence even recorded in the absence of prey fish spe- most frequently (Pauwels et al. 2014) and in a German lake, cies (other than occasional European eel Anguillia anguilla) translocated pike in the summer all returned to their main (Townsend and Peirson 1988). Townsend and Peirson (1988) activity centre within 6 days, suggesting strong fidelity to suggested that pike in these dyke systems would thus have these areas (Kobler et al. 2008). diets heavily reliant on non-fish prey, including from ter - The variability in movement metrics between individual restrial sources. pike was owing to some individuals moving between the The suggestion that even adult pike diet can rely on non- different macrohabitats present in each river, with most fish prey by Townsend and Peirson (1988) has been sup- detections occurring on receivers located in the laterally ported by other studies where individual pike have been connected waters rather than the main river, but with similar shown to be highly invertivorous (Beaudoin et al. 1999; Ven- fish numbers detected in all macrohabitats. This variability turelli and Tonn 2005) including in pike of up to 60 cm fork could have been influenced by variations in receiver detec- length (Pedreschi et al. 2015). This trophic flexibility was tion efficiency between the different macrohabitats. Indeed, also evident here, where stable isotope mixing models pre- receiver efficiencies can be influenced by many biotic and dicted that the diet of River Ant pike had relatively high pro- abiotic factors, including temperature, precipitation, extent portions of macroinvertebrates when compared with those of vegetation and suspended sediments (Winter et al. 2021b). in the Bure and Thurne. In contrast, Bure and Thurne pike However, the receiver array used here was considered as were predicted as having diets where common bream was an highly reliable during the study period, with detection range important – and large – prey item, with this dietary impor- very rarely falling below the width of the river (Winter et al. tance increasing as pike length increased (with common 2021b). Thus, the observed differences in pike detections bream having higher values of δ C compared with other and movements were considered as valid rather than an aquatic resources, as well as being an abundant fish species artefact of detection inefficiencies of the acoustic receivers. (Winter et al. 2021a,c)). Ontogenetic dietary switches are Moreover, high intra-population variability in movement common in pike through their development of larger gape ecology is a feature of many fish populations. In European sizes with increased body size, which facilitates their capture barbel, the majority of individuals have relatively limited and handling of larger prey items (Bry et al. 1995; Nilsson total ranges (< 10 km) but with a low proportion of individu- and Brönmark 2000). In the lower River Severn, western als having much larger total ranges (Gutmann Roberts et al Britain, the stable isotope ecology of pike of over 65 cm 2019; Amat Trigo et al., 2024). Juvenile Atlantic salmon indicated that relatively large fish (> 30 cm) were important Salmo salar predominantly have low mobility, adopting sed- prey items, but with these prey items generally absent in entary behaviour on most days, but with some individuals the diet of smaller pike (Nolan et al. 2019). However, the occasionally exhibiting bursts of high mobility, either by putative prey samples collected for SIA were limited here frequently moving within a confined area or moving between to aquatic resources, with terrestrial resources not being pools (Roy et al., 2013). considered as important to collect at the time of sampling. Highly connected systems are recognised as being impor- Accordingly, while we have confidence that the larger pike tant for providing the key functional habitats needed for all were consuming larger common bream as the reason for aspects of the pike lifecycle (Adolfsson 2020). Pike gener- their higher values of δ C (due to their high abundance in ally use wetland areas and back channels as spawning and the system), we cannot discount that it could also related nursery areas, which provide good habitats for larval pro- to some consumption of terrestrial prey items of relatively duction and recruitment compared with alternative habitats high δ C. (Nilsson, Engstedt and Larsson, 2014). The use of these The multivariate analyses indicated that the relationship wetland areas by pike is, however, often limited to spring between pike body length and total distance moved was sig- and early summer owing to their propensity of these areas nificant and positive, with individuals in the Thurne also to dry up later in summer, resulting in juveniles emigrating moving more than those in the Bure and Ant. In the River Individual variability in the movement ecology of Northern pike Esox lucius in a highly connected… Page 11 of 13 105 Frome, southern England, larger pike utilized 60% less of likely to be initiated by increasing water temperatures, the river length, relative to their body size, compared with with these warmer temperatures also playing a pivotal smaller individuals. This reduced movement is thought to role in the spawning movements of other pike popula- result from larger pike expanding their home ranges, which tions (Ovidio and Philippart, 2003). Although we argue increases spatial overlap between individuals, likely owing our results provide important insights into the dynamics to high levels of intra-specic fi competition, with this posited of pike movements in wetland systems, we acknowledge as owing to the pike increasing their home range sizes – and there are some limitations that necessitate some cautious thus the spatial overlap between individuals – owing to rela- interpretations. While the sample size provided valuable tively high intra-specific competition (Jetz et al. 2004, Ros- data, the relatively small number of analysed fish and the ten et al. 2016). Such patterns were, however, not evident in reliance on a single temperature monitoring station may our results, where the relationship between pike total range have constrained our ability to fully capture the variability and body size was not significant. Instead, we posit that our in pike behaviour and the influence of temperature on this. larger pike generally moved more than smaller pike – espe- Additionally, the deployment of acoustic receivers, while cially in the River Thurne – as a response to movements of generally efficient in detection, could have introduced their prey species. This hypothesis is based on our results variability in detection numbers across different habitats that indicated these larger pike not only moved significantly (given some areas had a relatively low receiver density), further than smaller pike but were also predicted to have potentially influencing the spatial patterns in detections. diets comprising relatively large proportions of the highly The absence of terrestrial prey resources in our analysis vagile and abundant common bream. This hypothesis was, also leaves room for further exploration of the trophic however, unable to be tested here owing to the use of low- influences on pike movement. Consequently, while our resolution acoustic telemetry, which was unable to accu- study contributes to the understanding of pike ecology, it rately quantify the movement relationships of the tagged highlights the need for more extensive research to build pike with the movements of shoals of common bream. on these findings and address the inherent challenges in Indeed, a limitation of low resolution telemetry is that while studying complex aquatic systems. it enables measurement of fish movements, it provides negli- In summary, there was high individual variability in the gible information on the actual activity of those fish, coupled movements of pike in this wetland system, with larger pike with it being unable to detect fish out of range of receiv - moving more that potentially facilitated their ability to pre- ers, which is an issue in a species, such as pike that spend date upon the highly vagile, abundant and large-bodied com- relatively long periods without moving far (Crossin et al. mon bream. Although some pike were highly vagile, their 2017; Jacoby and Piper 2023). Nevertheless, we speculate movements were generally between different macrohabitats that predator–prey dynamics could have played a significant in each river, rather than being long distance movements role in driving the movements of the larger tagged pike here, between rivers. If the connectivity of these habitats were where they actively sought optimal foraging areas where compromised owing to infrastructure developments, such large-bodied profitable prey, such as common bream, were as barriers, or declines in habitat quality, such as decreasing more abundant (Goldbogen et al. 2015; Florko et al. 2023). water levels and eutrophication (Ventura et al. 2023), pike The study pike significantly increased their total ranges might be compelled to cover greater distances in search of during the spawning period. Pike typically demonstrate suitable functional habitats. This enforced dispersal could a discernible partitioning between foraging and spawn- have a dual effect: on one hand, it might promote genetic ing areas, with fidelity expressed for specific spawning flow between distinct pike metapopulations, potentially grounds year after year (Craig 1996), and their spawn- influencing genetic diversity (Ouellet-Cauchon et al. 2014), ing migrations can be extensive, with some exceptional but could also result to increased energy expenditure, which cases reaching up to 78 km (Carbine and Applegate, 1948 might reduce overall fitness. This heightened energy demand in Craig 1996), and with Baltic pike demonstrating ana- could compromise the ability of individuals to effectively dromy during spawning (Lappalainen et al. 2008). Thus, compete for spatial and food resources, potentially leading to the increased total range of Bure pike during spawning increased risks of both interspecific and intraspecific compe- was likely owing to them making movements to specific tition (Bonte et al. 2012). Such extended movements might spawning areas, although this could not be tested within also expose fish to areas with varying resource availability our telemetry data owing to its low resolution that did not (Cooke et al. 2022). Accordingly, these results highlight that allow spawning areas to be reliably detected. Although preserving habitat connectivity is crucial for maintaining there were limitations owing to having a single tracking sustainable pike populations, as the connectivity ensures station for environmental variables and therefore possi- individuals can access diverse and profitable functional bility of temperature fluctuations throughout the system, habitats throughout the year and express their full range of our analyses suggested these spawning movements were movement behaviours. 105 Page 12 of 13 S. Cittadino et al. Supplementary Information The online version contains supplemen- Bonte D et al (2012) Costs of dispersal. Biol Rev. https:// doi. org/ 10. tary material available at https://doi. or g/10. 1007/ s00027- 024- 01124-4 .1111/j. 1469- 185X. 2011. 00201.x Brownscombe JW et al (2022) Application of telemetry and stable Acknowledgements We thank the anglers for their assistance in pike isotope analyses to inform the resource ecology and management sampling, especially the Pike Anglers Club of Great Britain (PAC) and of a marine fish. J Applied Ecol. https:// doi. org/ 10. 1111/ 1365- Norwich and District Pike Club (NDPC). We also thank the Environ-2664. 14123 ment Agency for their logistical help in pike tagging Bry C et al (1995) Early life characteristics of pike, Esox lucius, in rear- ing ponds: temporal survival pattern and ontogenetic diet shifts. Author Contribution J.R.B., E.W., A.H., S.L., R.W. and J.L. conceived J Fish Biol. https:// doi. org/ 10. 1111/j. 1095- 8649. 1995. tb059 49.x the study; J.R.B., E.W., A.H. and S.L. captured and tagged the fish; Campbell HA et al (2012) V-Track: software for analysing and visualis- S.C., E.W., A.S.T. and S.A. analysed the data; S.C., J.R.B. and A.S.T. ing animal movement from acoustic telemetry detections. Marine wrote the manuscript, and all authors edited the manuscript, and all Freshwater Res. https:// doi. org/ 10. 1071/ MF121 94 authors agree to its submission. Cantrell, D.L. et al. (2018) ‘The Use of Kernel Density Estimation With a Bio-Physical Model Provides a Method to Quantify Con- Funding Bournemouth University,Environment Agency,TÜBİTAK nectivity Among Salmon Farms: Spatial Planning and Manage- BİDEB,2219 Program, EU LIFE + Nature and Biodiversity Pro- ment With Epidemiological Relevance , Frontiers in Veterinary gramme, LIFE14NAT/UK/000054 Science. https://www.frontiersin.org/articles/https:// doi. org/ 10. 3389/ fvets. 2018. 00269. Data availability Raw data used in the manuscript that are not already Casselman JM (1974) ‘External Sex Determination of Northern Pike, provided are available from the corresponding author upon reasonable Esox lucius Linnaeus. Transact American Fisheries Society. request.https://doi. or g/10. 1577/ 1548- 8659(1974) 103% 3c343: ESDONP% 3e2.0. CO;2 Cooke SJ et al (2022) ‘The movement ecology of fishes. J Fish Biol. Declarations https:// doi. org/ 10. 1111/ jfb. 15153 Costantini ML et al (2018) ‘The role of alien fish (the centrarchid Conflict of interests The authors declare no competing interests. Micropterus salmoides) in lake food webs highlighted by stable isotope analysis ,. Freshwater Biol. https:// doi. org/ 10. 1111/ fwb. Open Access This article is licensed under a Creative Commons Attri- bution 4.0 International License, which permits use, sharing, adapta- Craig JF (1996) Population dynamics, predation and role in the tion, distribution and reproduction in any medium or format, as long community. In: Craig JF (ed) Pike: Biology and exploitation. as you give appropriate credit to the original author(s) and the source, Dordrecht, Springer, Netherlands provide a link to the Creative Commons licence, and indicate if changes Craig JF (2008) ‘A short review of pike ecology. Hydrobiologia. https:// were made. The images or other third party material in this article are doi. org/ 10. 1007/ s10750- 007- 9262-3 included in the article’s Creative Commons licence, unless indicated Crossin GT et al (2017) Acoustic telemetry and fisheries management. otherwise in a credit line to the material. If material is not included in Ecol Applicat. https:// doi. org/ 10. 1002/ eap. 1533 the article’s Creative Commons licence and your intended use is not Cucherousset J et al (2009) Spatial behaviour of young-of-the-year permitted by statutory regulation or exceeds the permitted use, you will northern pike (Esox lucius L) in a temporarily flooded nursery need to obtain permission directly from the copyright holder. To view a area. Ecol Freshwater Fish. https:// doi. org/ 10. 1111/j. 1600- 0633. copy of this licence, visit http://creativecommons.org/licenses/by/4.0/. 2008. 00349.x Flávio H, Baktoft H (2021) actel: Standardised analysis of acoustic telemetry data from animals moving through receiver arrays. Met Ecol Evol. https:// doi. org/ 10. 1111/ 2041- 210X. 13503 References Florko KRN et al (2023) The dynamic interaction between predator and prey drives mesopredator movement and foraging ecology. Bio. https:// doi. org/ 10. 1101/ 2023. 04. 27. 538582 Acolas ML, Le Pichon C, Rochard E (2017) ‘Spring habitat use by Fullerton AH et al (2010) ‘Hydrological connectivity for riverine fish: stocked one year old European sturgeon Acipenser sturio in the measurement challenges and research opportunities. Freshwater freshwater-oligohaline area of the Gironde estuary. Estuarine Biol. https:// doi. org/ 10. 1111/j. 1365- 2427. 2010. 02448.x Coastal Shelf Sci. https:// doi. org/ 10. 1016/j. ecss. 2017. 06. 029 Goldbogen JA et al (2015) Prey density and distribution drive the three- Adolfsson, O. (2020) Consequences on population dynamics following dimensional foraging strategies of the largest filter feeder. Funct regained connectivity in pike (Esox lucius) spawning location. : Ecol. https:// doi. org/ 10. 1111/ 1365- 2435. 12395 https:// urn. kb. se/ resol ve? urn= urn: nbn: se: lnu: diva- 104213. Grill G et al (2019) Mapping the world s free-flowing rivers. Nature. Amat-Trigo F et al (2024) Variability in the summer movements, habi- https:// doi. org/ 10. 1038/ s41586- 019- 1111-9 tat use and thermal biology of two fish species in a temperate Gutmann Roberts C, Hindes AM, Britton JR (2019) ‘Factors influenc- river. Aquatic Sci. https:// doi. org/ 10. 1007/ s00027- 024- 01073-y ing individual movements and behaviours of invasive European Amoros C, Bornette G (2002) Connectivity and biocomplexity in barbel Barbus barbus in a regulated river. Hydrobiologia. https:// waterbodies of riverine floodplains. Freshwater Biol. https:// doi. doi. org/ 10. 1007/ s10750- 018- 3864-9 org/ 10. 1046/j. 1365- 2427. 2002. 00905.x Home | Ramsar Sites Information Service. https:// rsis. ramsar. org/. Araújo MS, Bolnick DI, Layman CA (2011) The ecological causes of Ieno, E.N. ‘A beginner s guide to data exploration and visualisation individual specialisation. Ecol Letters. https:// doi. org/ 10. 1111/j. with R. : https:// cir. nii. ac. jp/ crid/ 11308 48327 91313 9328. 1461- 0248. 2011. 01662.x Jacoby DMP, Piper AT (2023) What acoustic telemetry can and can- Beaudoin CP et al (1999) Individual specialization and trophic adapt- not tell us about fish biology. J Fish Biol. https://doi. or g/10. 1111/ ability of northern pike (Esox lucius): an isotope and dietary jfb. 15588 analysis. Oecologia. https:// doi. org/ 10. 1007/ s0044 20050 871 Kobler A, Klefoth T, Arlinghaus R (2008) ‘Site fidelity and seasonal Belletti B et al (2020) More than one million barriers fragment Europe changes in activity centre size of female pike Esox lucius in a s rivers. Nature. https:// doi. org/ 10. 1038/ s41586- 020- 3005-2 Individual variability in the movement ecology of Northern pike Esox lucius in a highly connected… Page 13 of 13 105 small lake. J Fish Biol. https://doi. or g/10. 1111/j. 1095- 8649. 2008. Roy L et al (2012) Individual variability in the movement behaviour of 01952.x juvenile Atlantic salmon. Canadian J Fish Aquatic Sci [preprint]. Koed A et al (2006) Annual movement of adult pike (Esox lucius L.) https:// doi. org/ 10. 1139/ cjfas- 2012- 0234 in a lowland river. Ecol Freshwater Fish. https://doi. or g/10. 1111/j. Sandlund OT, Museth J, Øistad S (2016) Migration, growth patterns, 1600- 0633. 2006. 00136.x and diet of pike (Esox lucius) in a river reservoir and its inflowing Lappalainen A et al (2008) Reproduction of pike (Esox lucius) in reed river. Fisher Res. https:// doi. org/ 10. 1016/j. fishr es. 2015. 08. 010 belt shores of the SW coast of Finland, Baltic Sea: a new survey Shao X et al (2019) River network connectivity and fish diversity. Sci approach: Boreal Environment Research. Boreal Environ Res Total Environ. https:// doi. org/ 10. 1016/j. scito tenv. 2019. 06. 340 13(4):370–380 Shaw AK (2020) Causes and consequences of individual varia- Layman CA et al (2012) Applying stable isotopes to examine food-web tion in animal movement. Move Ecol. https:// doi. org/ 10. 1186/ structure: an overview of analytical tools. Biol Rev. https:// doi. s40462- 020- 0197-x org/ 10. 1111/j. 1469- 185X. 2011. 00208.x Speed CW et al (2010) Complexities of coastal shark movements and Masters, J.E.G. et al. (2002) ‘Habitat utilisation by pike Esox lucius their implications for management. Marine Ecol Progres. https:// L. during winter floods in a southern English chalk river , in E.B. doi. org/ 10. 3354/ meps0 8581 Thorstad, I.A. Fleming, and T.F. Næsje (eds) Aquatic Teleme- The Scaling of Animal Space Use | Science. https://www.science.org/ try: Proceedings of the Fourth Conference on Fish Telemetry in doi/full/https:// doi. org/ 10. 1126/ scien ce. 11021 38. Europe. Dordrecht, Springer, Netherlands. Thieme M et al (2023) Measures to safeguard and restore river con- Miller LM, Kallemeyn L, Senanan W (2001) Spawning-site and natal- nectivity. Environment Rev [preprint]. https:// doi. org/ 10. 1139/ site fidelity by northern pike in a large lake: mark-recapture and er- 2023- 0019 genetic evidence. Transaction American Fisher Societ. https:// Townsend CR, Peirson G (1988) Fish community structure in lowland doi.or g/10. 1577/ 1548- 8659(2001) 130% 3c0307: SS ANSF%3e2.0. drainage channels. J Fish Biol. https:// doi. org/ 10. 1111/j. 1095- CO;28649. 1988. tb053 62.x Nilsson PA, Brönmark C (2000) ‘Prey vulnerability to a gape-size lim- Udyawer V et al (2018) A standardised framework for analysing animal ited predator: behavioural and morphological impacts on northern detections from automated tracking arrays. Animal Biotelemet. pike piscivory. Oikos. https:// doi. org/ 10. 1034/j. 1600- 0706. 2000. https:// doi. org/ 10. 1186/ s40317- 018- 0162-2 880310.x van Puijenbroek PJTM et al (2019) Species and river specific effects of Nilsson J, Engstedt O, Larsson P (2014) Wetlands for northern pike river fragmentation on European anadromous fish species. River (Esox lucius L) recruitment in the Baltic Sea. Hydrobiologia. Res Application. https:// doi. org/ 10. 1002/ rra. 3386 https:// doi. org/ 10. 1007/ s10750- 013- 1656-9 Vander Zanden HB et al (2010) Individual specialists in a generalist Nolan ET, Gutmann Roberts C, Britton JR (2019) ‘Predicting the con- population: results from a long-term stable isotope series. Biol tributions of novel marine prey resources from angling and ana- Letter. https:// doi. org/ 10. 1098/ rsbl. 2010. 0124 dromy to the diet of a freshwater apex predator. Freshwater Biol. Vehanen T et al (2006) Patterns of movement of adult northern pike https:// doi. org/ 10. 1111/ fwb. 13326 (Esox lucius L) in a regulated river. Ecol Freshwater Fish. https:// Nunn AD et al (2010) ‘Seasonal and diel patterns in the migrations of doi. org/ 10. 1111/j. 1600- 0633. 2006. 00151.x fishes between a river and a floodplain tributary. Ecol Freshwater Ventura M et al (2023) When climate change and overexploitation meet Fish. https:// doi. org/ 10. 1111/j. 1600- 0633. 2009. 00399.x in volcanic lakes: the lesson from lake bracciano, rome’s strategic Nyqvist MJ et al (2018) Relationships between individual movement, reservoir. Water. https:// doi. org/ 10. 3390/ w1510 1959 trophic position and growth of juvenile pike (Esox lucius). Eco Venturelli PA, Tonn WM (2005) Invertivory by northern pike (Esox Freshwater Fish. https:// doi. org/ 10. 1111/ eff. 12355 lucius) structures communities of littoral macroinvertebrates in Nyqvist MJ et al (2020) Dispersal strategies of juvenile pike (Esox small boreal lakes. J North American Benthol Soci. https://do i. lucius L.): Influences and consequences for body size, somatic org/ 10. 1899/ 04- 128.1 growth and trophic position. Ecol Freshwater Fish. https:// doi. Verhelst P et  al (2018) Downstream migration of European eel org/ 10. 1111/ eff. 12521 (Anguilla anguilla L.) in an anthropogenically regulated fresh- Oele DL et al (2019) Growth and recruitment dynamics of young-of- water system: Implications for management. Fisheries Res. https:// year northern pike: Implications for habitat conservation and man-doi. org/ 10. 1016/j. fishr es. 2017. 10. 018 agement. Ecol Freshwater Fish. https://doi. or g/10. 1111/ eff. 12453 Winter ER et al (2021a) Acoustic telemetry reveals strong spatial pref- Oksanen, J. (2012) ‘Constrained Ordination: Tutorial with R and vegan. erences and mixing during successive spawning periods in a par- Ouellet-Cauchon G et al (2014) Landscape variability explains spatial tially migratory common bream population. Aquatic Sci. https:// pattern of population structure of northern pike (Esox lucius) in doi. org/ 10. 1007/ s00027- 021- 00804-9 a large fluvial system. Ecol Evolut. https:// doi. org/ 10. 1002/ ece3. Winter ER et al (2021b) ‘Detection range and efficiency of acoustic 1121 telemetry receivers in a connected wetland system. Hydrobiologia. Ovidio M, Philippart JC (2005) Long range seasonal movements of https:// doi. org/ 10. 1007/ s10750- 021- 04556-3 northern pike (Esox lucius L.) in the barbel zone of the River Winter ER et al (2021c) ‘Dual-isotope isoscapes for predicting the Ourthe. River Meuse basin, Belgium scale of fish movements in lowland rivers. Ecosphere. https://doi. Pauwels IS et al (2014) ‘Movement patterns of adult pike (Esox lucius org/ 10. 1002/ ecs2. 3456 L.) in a Belgian lowland river. Ecol Freshwater Fish. https:// doi. Winter ER et  al (2021d) ‘Movements of common bream Abramis org/ 10. 1111/ eff. 12090 brama in a highly connected, lowland wetland reveal sub-popu- Pedreschi D et al (2015) ‘Trophic flexibility and opportunism in pike lations with diverse migration strategies. Freshwater Biol. https:// Esox lucius. J Fish Biol. https:// doi. org/ 10. 1111/ jfb. 12755doi. org/ 10. 1111/ fwb. 13726 Power ME, Dietrich WE (2002) Food webs in river networks. Ecolog Res. https:// doi. org/ 10. 1046/j. 1440- 1703. 2002. 00503.x Publisher’s Note Springer Nature remains neutral with regard to Rosten CM, Gozlan RE, Lucas MC (2016) Allometric scaling of jurisdictional claims in published maps and institutional affiliations. intraspecific space use. Biol Letters. https://doi. or g/10. 1098/ rsbl. 2015. 0673 http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Aquatic Sciences Springer Journals

Individual variability in the movement ecology of Northern pike Esox lucius in a highly connected wetland system

Download PDF
13 pages

Loading next page...
 
/lp/springer-journals/individual-variability-in-the-movement-ecology-of-northern-pike-esox-0U3aKB6B30

References (64)

Publisher
Springer Journals
Copyright
Copyright © The Author(s) 2024
ISSN
1015-1621
eISSN
1420-9055
DOI
10.1007/s00027-024-01124-4
Publisher site
See Article on Publisher Site

Abstract

Maintaining hydrological connectivity is important for sustaining freshwater fish populations as the high habitat connec- tivity supports large-scale fish movements, enabling individuals to express their natural behaviours and spatial ecology. Northern pike Esox lucius is a freshwater apex predator that requires access to a wide range of functional habitats across its lifecycle, including spatially discrete foraging and spawning areas. Here, pike movement ecology was assessed using acoustic telemetry and stable isotope analysis in the River Bure wetland system, eastern England, comprising of the Bure mainstem, the River Ant and Thurne tributaries, plus laterally connected lentic habitats, and a system of dykes and ditches. Of 44 tagged pike, 30 were tracked for over 100 days, with the majority of detections being in the laterally connected lentic habitats and dykes and ditches, but with similar numbers of pike detected across all macrohabitats. The movement metrics of these pike indicated high individual variability, with total ranges to over 26 km, total movements to over 1182 km and mean daily movements to over 2.9 km. Pike in the Thurne tributary were more vagile than those in the Ant and Bure, and with larger Thurne pike also having relatively high proportions of large-bodied and highly vagile common bream Abramis brama in their diet, suggesting the pike movements were potentially related to bream movements. These results indicate the high individual variability in pike movements, which was facilitated here by their access to a wide range of connected macrohabitats due to high hydrological connectivity. Keyword Pike movement ecology; River connectivity; Acoustic telemetry; Individual variability Introduction trends in fluvial connectivity are contrary to this, with most rivers around the world no longer being free flowing (Grill Longitudinal and lateral hydrological connectivity is impor- et al. 2019; Belletti et al. 2020). This disrupted connectiv- tant for sustaining fish assemblage structure and diversity ity tends to be associated with impacts on fish dispersal, (Amoros et al., 2002; Shao et al. 2019). However, global with barriers located throughout the aquatic system that * Simone Cittadino Environment Agency, Dragonfly House, Norwich NR3 1UB, [email protected] UK 2 South End Farm Cottages, Fishtrack Ltd, Department of Life and Environmental Sciences, Beccles NR34 8TG, UK Bournemouth University, Poole BH12 5BB, UK The Rivers Trust, Rain Charm House, Kyl Cober Park, Stoke Department of Ecology and Vertebrate Zoology, Faculty Climsland, Callington PL17 8PH, UK of Biology and Environmental Protection, University of Lodz, Lodz, Poland River Waveney Trust, Mill Ln, Pulham Market, Diss IP21 4XL, UK Department of Basic Sciences, Faculty of Fisheries, Muğla Sıtkı Koçman University, Menteşe, Muğla, Türkiye Environment Agency, Scarrington Road, Nottingham, United Kingdom Vocational School of Health Services, Eskişehir Osmangazi University, Eskişehir, Türkiye Environment Agency, Inworth Road, Rivers House, Threshelfords Business Park, Feering CO5 9SE, UK Vol.:(0123456789) 105 Page 2 of 13 S. Cittadino et al. can limit the movement of anadromous, catadromous, and being capable of consuming larger-bodied prey (Nilsson and potamodromous fishes (Fullerton et al. 2010). Conversely, Brönmark 2000; Nolan et al. 2019). Pike movements display the maintenance of natural flow regimes and an absence of considerable intra-population variation along a continuum of anthropogenic barriers provides high habitat connectivity, spatial behaviours, where some individuals remain confined facilitates fish dispersal processes, and supports complex to relatively small areas (less than 1 km of river length), trophic dynamics through increasing access of fishes to more while others move repeatedly over several km (Masters et al. diverse foraging habitats that alter inter- and intra-specific 2002). This variation can influence gene flow as highly vag- interactions, and also affects nutrient cycling and productiv - ile individuals move relatively large distances in search of ity (Thieme et al. 2023). In connected systems, the move- suitable spawning habitats and mates (Ouellet-Cauchon et al. ment of organisms across different habitat types allows for 2014). Additionally, it may increase their exposure to height- a more integrated food web where energy and nutrients can ened risks of both intra- and interspecific competition, which flow more freely through various trophic levels (Power and can further impact their population dynamics (Nyqvist et al. Dietrich 2002). 2018). Consequently, some pike populations consist of indi- High habitat connectivity enables individual fish to viduals with annual home ranges of less than 2 km, while express their natural traits and behaviours, which can vary others have home ranges exceeding 14 km (Sandlund et al. considerably between individuals (Amat-Trigo et al. 2024). 2016). While there is usually a positive relationship between This variability is expressed through differences in spatial home range size and body size, this does not necessarily behaviours where, for example, some individuals are rela- show allometric scaling. For example, large pike in a chalk tively sedentary, characterised by having small home ranges, stream in Southern England moved less than smaller individ- while others are more active and have much larger home uals when their body mass was taken into account (Rosten ranges (Gutmann Roberts et  al. 2019; Amat-Trigo et  al. et al. 2016). Longer distance movements of pike tend to be 2024). Individual variability in the trophic niche often mani- for spawning, with populations showing discernible parti- fests as generalist populations being composed of relatively tioning between foraging and spawning areas (Craig 1996), specialised individuals whose niches are small subsets of which often include tributaries, side channels and ditches the population niche (Araújo et al. 2011). Consistency in (Nyqvist et al. 2018; Oele et al. 2019). In the Baltic Sea, individual specialization has been argued to have important resident pike spawn in brackish coastal waters, while there is ecological, evolutionary, and conservation consequences, an anadromous form that spawns in freshwater streams and including reducing levels of intraspecific competition and wetlands (Lappalainen et al. 2008). High habitat connectiv- potentially improving population resilience through the min- ity is thus important in enabling pike to express individual imising competitive pressures (Vander Zanden et al. 2010). specialisation and to maintain their access to suitable spawn- Reproductive activities and trophic interactions within ing areas that then serve as nursery grounds. populations can also profoundly shape the spatial habitat use In this study, the aim was to describe the movements of of fish populations (Speed et al. 2010), such as instigating pike in a protected and highly connected wetland system, the seasonal migratory events for spawning (Winter et al. 2021a) lower River Bure network in eastern England, recognized for and more regular movements for foraging (Nunn et al. 2010). its ecological importance and designated as a Ramsar site. While daily foraging movements can be relatively limited in This aim included assessment of the influence of spawn- distance, the extent of spawning migrations can vary widely, ing, connectivity, individual traits and trophic ecology on ranging from a few kilometres for potamodromous species to the individual variability in movements between individual over 1000 km for diadromous species (Verhelst et al. 2018; pike. The study system was the lower River Bure network van Puijenbroek et al. 2019). However, traits, such as sex, in eastern England, a highly connected, free-flow wetland age, length size and behaviour, can substantially influence system comprising the main River Bure, connected side the distribution, behaviour and movements of individuals channels and stillwater habitats (Fig.  1). The movements (Koed et al. 2006; Nyqvist et al. 2020). The interplay of of pike were measured using acoustic telemetry, a key tool these variables thus has important implications for the spa- for assessing the movements of animals in aquatic environ- tial ecology of individuals, populations and communities. ments (Brownscombe et al. 2022), and complemented with The Northern pike Esox lucius (‘pike’) is a freshwater the ecological application of stable isotope analysis (SIA), apex predator with a Holarctic range, being encountered which provides information on trophic relationships and diet across Europe and North America, and is recognised as a (Layman et al. 2012; Costantini et al. 2018). The objectives keystone piscivore (Craig 2008). Their diet can be highly were to assess the individual variability in the movement plastic, with the consumption of prey that includes mac- metrics of pike over a 3-year period and the influence of roinvertebrates as well as fish (Pedreschi et  al. 2015). body size, spawning season, and individual trophic ecology Pike prey size is typically limited by gape size, and there on this. We posit that larger pike will move greater distances, are ontogenetic shifts in their diet, with larger individuals have larger home ranges, and consume larger-bodied prey Individual variability in the movement ecology of Northern pike Esox lucius in a highly connected… Page 3 of 13 105 Fig. 1 Map illustrating the study area within the Bure system, situ- ing the location of the station for environment variables (HOBO® ated in Broad National Park, with red squares indicating the locations Pendant; model MX2202, Onset Computer Corporation). The red where acoustic receivers were deployed and the red triangle indicat- arrows indicate the general flow compared with smaller pike, and all pike will move greater promotes sustainable land use practices (https://rsis. r amsar. distances during spawning versus non-spawning periods. org/r is/6 8). The system features more than 60 km of barrier- The study then quantifies the use of different macrohabi- free main river, with a complex network of secondary slow- tats by pike and explores how this habitat use varies across flow channels (< 2 m3/s) and numerous small shallow lakes the study area. In completing these objectives, an increased (4 m depth), commonly known as “Broads” (Natural Eng- understanding of pike behaviour and habitat use in lowland land, 2020). As a consequence, the system has high lateral river systems is generated, providing important information and longitudinal connectivity, with the rivers intersected by on the management of wetlands, including the maintenance a dense network of lateral broads and dyke systems (Fig. 1). of hydrological connectivity. The land use of the Broads primarily includes woodland and agriculture, with approximately 40% comprising of wet grassland grazed by livestock. As the lower Bure flows Methods downstream, it transitions into semi-artificial grazing marsh and the water becomes more brackish owing to groundwater Study area interaction with seawater (Winter et al. 2021a). The study area is influenced by tidal activity owing to its proximity to The study area was the northern region of the Broads the sea, lack of barriers and the low-lying, relatively homo- National Park, Eastern England, comprising the lower geneous topography of the landscape. The area is therefore River Bure and its tributaries of the Rivers Ant and Thurne prone to saline intrusion during large spring tides and storm (Fig.  1). This important wetland area has Ramsar status, surges in winter. The salt edge can be pushed more than which provides protection against habitat degradation and 10 km inland during strong saline incursion events. 105 Page 4 of 13 S. Cittadino et al. the main rivers longitudinally (to enable long distance move- Fish capture, acoustic transmitter implantation and receiver network ments to be tracked). Some receivers were also positioned at the mouth of lateral connections (e.g., at the entrance of the Pike were sampled in the three main rivers (Bure, Ant and broads from the main river) and inside channels considered to be important for pike foraging and spawning (based on Thurne; Fig. 1). Rod and line angling was used to capture the fish as it was more effective in capturing large-bodied extant knowledge of the authorship team) (Fig. 1). Receivers were attached to permanent structures, moored on naviga- individuals compared with seine netting and electrofishing compromised by the increasing conductivity downstream. A tion posts or suspended from floating objects, at approxi- mately mid-water depth (1.0–1.5 m) to maximise detection total of 44 pike were captured; 16 were captured in Novem- ber 2017, 27 in January 2018 and 1 additional pike was cap- efficiency during tidal changes in river levels. Detection data were downloaded every 4 months and batteries replaced tured in September 2018 (Table S1). Before surgery, all fish were examined by the operator to assess any signs of stress annually. Movement data of the fish were collected from 2018 to 2020. All receivers remained operable throughout and whether the length of the individual was appropriate for the size of the acoustic tag (all tagged fish were > 500 mm). the study period and prior to any data analyses, the detection data were processed in actel R package (R version 4.2.3) Under general anaesthesia (Tricaine methanesulfonate, −1 MS-222; approximately 0.04 g  l ), the pike were measured to remove any false detections (e.g. transmitter code errors caused by code collisions) (Flávio and Baktoft 2021). (fork length, nearest mm), sexed (external identification; Casselman 1974) and scale samples and pelvic fin biopsy Stable isotope analysis and abiotic data collection taken (equivalent to approximately 1 mg dry weight), with the fin tissue frozen for storage. Each pike was then surgi - The fin biopsies of the pike, common bream and roach were cally implanted with an acoustic transmitter (Vemco V13; 69 kHz; 36 mm length × 13 mm diameter; mass in water: individually used for stable isotope analysis (SIA). After drying at 60 °C for 24–36 h, the samples were ground and 6.0 g; random transmission interval approximately 90 s; estimated battery life: 1200 days; manufacturer: Innovasea, weighed to 1000 µg in tin capsules. These samples were then analysed for δ13C versus VPDB and δ15N versus At. USA). The transmitters were inserted ventrally anterior to the pelvic fins, and incisions were closed using a single Air (%) at the Cornell University Stable Isotope Laboratory, New York, USA, using a Thermo Delta V isotope ratio mass suture (2–0 absorbable monofilament; Ethicon Ltd, USA) and wound sealer (Orabase; ConvaTec Ltd, UK). During spectrometer interfaced with an NC2500 elemental analyser (CE Elantach Inc., USA). As C:N ratios were < 3.5, no lipid surgery, instruments were sterilised between fish though immersion in a 10% iodine based solution (Videne antisep- correction was applied. The same preparation and analy- ses were completed for the macro-invertebrate samples, a tic solution; Ecolab Ltd, UK), followed by rinsing in 0.9% saline solution. During surgery, scales were removed from minimum of four replicates each capture site and date, with amphipods (Gammaridae) the most represented group, with the incision site to facilitate the entry of the scalpel and sutures. Following completion of the surgical procedure, the other samples including killer shrimp Dikerogammarus vil- losus (Table S2, Winter et al. 2021c). In subsequent analy- fish were then transferred to tanks containing river water, held until normal body orientation and swimming behaviour ses, SI data of gammarids and killer shrimp were combined into a single group (‘Amphipods’), as differences in their resumed. After a visual inspection by the operator fish were released back to their approximate area of capture within 2 h values were not significant (t -test, p-value > 0.05). Abiotic data were recorded by a stationary receiver of their tagging. During the same sampling periods, scales and fin biopsies were collected from 83 common bream (HOBO® Pendant; model MX2202, Onset Computer Cor- poration), maintained by the Environment Agency, and (Abramis brama) and 47 roach (Rutilus rutilus) coming from the three different rivers. Additionally, macroinvertebrate located in the main River Bure (52°38′56.5"N 1°34′03.5" E). Water temperature (± 0.5 °C), salinity and conductivity samples were gathered from six distinct sites along the riv- ers that cover the main River Bure, Thurne and Ant, with (us/cm) were recorded every 15 min throughout the entire study period. sampling between 11 September and 3 October 2018 using a sweep net for stable isotope analysis (Table S2, Winter Pike movement metrics et al. 2021a,b,c). All surgical procedures were conducted by the same surgeon and in accordance with UK Home Office Of the 44 pike tagged, 41 were included in the calcula- project license 70/8063 and after ethical review. To track the movements of the tagged fish, a fixed net- tion of movement metrics, with the remaining 3 individu- als excluded owing to their lack of detections (Table S1). work of 44 Vemco VR2W acoustic receivers was deployed in October 2017 and retrieved in summer 2020. These receivers Data were downloaded from receivers using VUE software (version 2.7.0), and subsequent data manipulation was were strategically placed to provide coverage of the length of Individual variability in the movement ecology of Northern pike Esox lucius in a highly connected… Page 5 of 13 105 conducted using the Vtrack R package (Campbell et al. using the adonis2 function implemented in the R package 2012), which enables the detection data of Vemco acoustic vegan (Oksanen 2012). transmitters to be assimilated across all receivers and then analysed appropriately (Udyawer et al. 2018). To calculate Statistical analyses river distances between detection points, the distance matrix was computed using the Field Calculator function in The data collected in the first five days following each tag- QGIS (version 3.28.8). The analyses completed here were ging event were excluded from the sample (in case the fish a series of movement metrics as per Gutmann Roberts et al. were demonstrating unusual post-surgical behaviours). Only (2019) (Table S1). Three indices were calculated to assess pike that were detected for a period of at least 100 days (as fish residency within the receiver array: the residency index, days between the first and the last detection) were included the linearity index and the residence array index (Table S1; (n = 30). This arbitrary threshold was chosen to focus on Craig 2008; Acolas et al. 2017). The residence array index long-term movements and exclude shorter detections that was included to address instances where a tagged fish could bias the results. The three main movement variables might have been within the receiver array but remained of each individual (TR, TM and MDM) were first tested undetected on a given day. The determination of total range at univariate level against total length (α = 0.05) and then (TR), expressed as the river distance (m) between the far- included as response variables in different models and tested thest receivers where an individual was detected, served as against the predictors (and their combinations) of the river of a proxy for home range. Additionally, total movement (TM), capture, spawning or non-spawning period (spawning period defined as the total distance covered by an individual, was covered the pre-, spawning and the post-spawning period: 1 computed by summing individual movements, regardless of December to 30 April; non-spawning period: 1 May to 30 their directionality (i.e. upstream, or downstream). In calcu- November), water temperature, salinity, fish identification lating mean daily movement (MDM), to avoid overestimates, (transmitter code), fork length and the proportion of com- only those occasions when the time between one observation mon bream in diet (from stable isotope analyses; see below). and the next did not exceed 24 h were considered. With the response variables based on fish being detected Kernel density estimation (KDE) was implemented to for at least 100 days, the number of days of detection was delineate the primary areas of pike spatial occupancy within not included as a predictor of any movement variable as the the study area. This method involves computing the den- relationships were all non-significant (Pearson’s correlation sity of point features surrounding each output raster cell. coefficient, all r < 0.33, P > 0.05). Before model fitting, data A continuous, smoothly curved surface is then fitted over exploration was undertaken following the protocol described each point, with the surface exhibiting its highest value at in Ieno & Zurr (2015) involving examination for missing the point’s location. This value gradually decreases as the values, outliers in the response and explanatory variables, distance from the point increases, eventually tapering off homogeneity and zero inflation in the response variable, col- to zero at a distance equal to the specified search radius linearity between explanatory variables, the balance of cat- (Cantrell et al., 2018). KDE was performed by dividing the egorical variables, and the nature of relationships between individuals according to their capture site (Bure, Ant and the response and explanatory variables. Thurne) and by dividing detections according to whether To test differences in movement metrics between pike they were in the spawning period (1st Dec–30th Apr) or spawning and non-spawning periods, all movement metrics non-spawning period (1st May–31st Nov). These analyses were subjected to analysis with ‘spawning’ designated as were all conducted in ArcMap version 10.8.2. and gener- the predictor variable. The best fitting model for the total ated heat maps that displayed the pike’s spatial occupancy range of movement (‘TR’) was a linear mixed-effects model probability. To assess variations in how the pike use the dif- (LMM, ‘lme4’ R package), which was chosen due to the ferent macrohabitats across the study area, the receivers on normally distributed nature of the TR data and to account which they were detected were then categorised according to for individual variability by including ‘FishID’ as a random whether the receiver was located in the main river (‘main’; effect (Table S3). To assess how pike utilized different mac- as the Bure mainstem, and the Ant and Thurne tributaries), rohabitats during spawning versus non-spawning periods, in ‘broads’ (the laterally connected, shallow, lentic habitats) a generalized linear mixed model (GLM) with general- or within the ditch and dyke systems (‘lateral’). For each ized Poisson regression was employed. In this model, the river, the receiver detection data by macrohabitats were response variable was the number of detections per receiver. summed and converted to proportions (%). To see the dif- The predictors included the macrohabitats of the main riv- ferences in number of detections between habitats and wild ers (‘main’, comprising the Bure mainstem, Ant and Thurne populations, and isotopic signatures between the sites, we tributaries) and the laterally connected habitats (‘lateral’, used a permutational univariate analysis of variance (PER- which included the broads, ditches, and dyke systems). For ANOVA), with Euclidean distance and 9999 permutations the analysis, the macrohabitats ‘broad’ and ‘lateral’ were 105 Page 6 of 13 S. Cittadino et al. combined as this enabled evaluation of the overall use of in the individual diet of pike. The proportion of bream in secondary connections versus the main river channels. As pike diet was entered as the response variable in a GLMM the total number of detections per receiver was not necessar- to investigate potential associations with pike fork length. ily related to the number of fish detected (e.g. one individual The predictor ‘river’ was also incorporated into the model to could have been detected multiple times) the number of pike account for variations among the three primary rivers under detected at each receiver in each period was included as a study. This approach aimed to elucidate potential correla- random effect. In addition, two other random effects were tions between the dietary habits of pike and their size while included in the model: the area of river the receiver was considering differences within the study area. located (‘River’; as Bure, Ant and Thurne) and the antenna receiver identification number (‘ReceiverID’). The relationship between pike length and movement pat- Results terns was analysed using a generalized linear mixed model (GLMM) with a gamma distribution. In this model, total Fish length and detections movement (TM) was the response variable and fork length was the fixed factor. The GLMM was chosen to address The mean fork length of the 41 pike analysed (31 female, the high variance and potential overdispersion in the TM 9 male, 1 undetermined) was 772 ± 122  mm (range data, providing a more accurate assessment of movement 583–1143 mm), with females significantly larger than males patterns. TM was used as other studies have observed that (t-test, t = −5.111, p < 0.01) (Table S1). Detections of the larger pike tend to move significantly more than small pike tagged pike occurred across 933 days, with almost 3 million (day of detection and TM are autocorrelated) (Vehanen et al. of individual detections, which were detected on individual 2006; Craig 2008; Sandlund et al. 2016). An additional inter- receivers between 1117 and 521,771 times. Tracking periods action term between pike length and river of capture (as were highly variable between individuals (3–931 days), with Bure, Thurne or Ant) was incorporated to accommodate individuals being detected between 2 and 172 days of detec- potential disparities in movement tendencies across die ff rent tion, with 30 fish having more than 100 days between their river locations (Table S3). To then test the effect of abiotic first and last detection (Table S1). variables on pike movement, a generalized additive model (GAM) was constructed using the 'mgcv' R package; a Movement metrics and macrohabitat use GAM was used owing to its ability to capture the nonlinear effects of temperature and salinity. Mean daily movement Movement metrics varied between individuals (total range: (MDM) was the response variable, given its sensitivity to 1487–26,099  m; total movement 8032–1,182,287  m; fluctuations in environmental conditions, water temperature mean daily movement 46–2948  m; Table  S1). For fish and salinity (daily mean) were included as predictor vari- detected for at least 100 days, Thurne pike had the larg- ables. Since the metrics are temporally repeated (across the est total ranges [mean ± 95% confidence interval (CI): season) to address individual variation, the fish unique iden- 16,076 ± 5782 m], with Ant pike having higher total ranges tification code (FishID) was integrated as a random effect (8141 ± 4951 m) than those in the Bure (5743 ± 1992 m). within the model structure (Table S3), with the number of This was also reflected in the patterns of mean daily distance knots set to 5 for all parameters to prevent overfitting. The moved (Thurne: 1524 ± 383 m; Ant: 1261 ± 658 m; Bure: model fitting process aimed to identify smooth trends that 717 ± 403 m). At a univariate level and across all fish, the optimally captured the data while balancing goodness of fit movement metrics of total range, total distance moved and and smoothness, with the final, best-fitting model identified mean daily distance moved were not significantly related to by backward selection using AIC (ΔAIC ≤ 2). fish length (linear regression, P > 0.05 in all cases, Fig. S1). The heat maps of pike spatial occupancy indicated most Analyses of stable isotope data fish remained in the river where they were initially caught in both spawning and non-spawning periods (Fig. 2, Fig. To predict the dietary contribution of the putative prey spe- S2). Individuals captured in the main River Bure were never cies (common bream, roach, and macroinvertebrates) to detected in the other tributaries, while some individuals pike diet, stable isotope mixing models were applied using from the Ant and the Thurne were detected in the main Bure the R package simmr (Parnell and Inger, 2016; Flávio and in both the spawning and non-spawning seasons (Fig. 2, Fig. Baktoft 2021). To achieve this, we adopted a prior-free S2). There were substantially more detections of pike on approach and applied trophic discrimination factors (TDFs) receivers located in the laterally connected macrohabitats suggested by Post (2002), specifically 1.0 for δ C and 3.4 than the main river (PERANOVA: p = 0.043), but the num- for δ N. Results of the mixing models are presented as a bers of pike detected across these macrohabitats were similar posteriori distribution for the proportion of each prey item (PERANOVA: p = 0.9) (Table S4). The GLM assessing the Individual variability in the movement ecology of Northern pike Esox lucius in a highly connected… Page 7 of 13 105 Fig. 2 Heat maps of pike occupancy during the spawn- ing (SW) and non-spawning period (NSW) in the river Bure system according to the loca- tion of tagging (Bure, Ant and Thurne), where the probability of occupancy ranges from absence (blue) through to low (green), medium (yellow) and high (red). Approximate release locations are noted with the black lines number of receiver detections across these macrohabitats (Table 1B). Mean daily distances moved were significantly revealed the influence of spawning season was not signifi- and positively influenced by water temperature (GAM, cant (P = 0.82), but the macrohabitat predictors were signifi- P = 0.03; Table 1C), with the fish identification interaction cant (‘Lateral’, P = 0.01; ‘Main’, P = 0.003), indicating the term also being significant in the model, emphasising the use of these different macrohabitats was important all year individual variability in the dataset. round (Table S5). Testing the relationships of the movement metrics in Pike diet composition by river multivariate models revealed spawning season had a sig- nificant and positive effect on total range (LMM, P < 0.01; Mean δ C of all pike was −27.55 ± 1.54% (−28.75% to Table 2A), with the total range being on average 36% greater −24.83%) and mean δ N was 19.45 ± 1.05% (17.86% to in the spawning season than on the non-spawning season. 21.69%), with differences between sexes being not sig- 13 15 The fish identification interaction term was also significant, nificant (t-test: δ C, t = −0.26, p = 0.80; δ N, t = −0.89, indicating that although movements were greater in spawn- P = 0.39). Across all pike, there was a significantly 13 2 ing periods, this varied between individuals. The GLMs higher value of δ C as pike length increased (R = 0.16; indicated pike length had a significant and positive effect F = 6.41, P = 0.02), but with no relationship between 1,34 15 2 on total movements (P = 0.03), where the interaction term pike length and δN (R = 0.03; F = 0.90, P = 0.35) 1,34 between length and capture location was significant for (Fig. S3). The common bream analysed for their stable the River Thurne (P = 0.02), indicating larger pike in this isotope ratios were relatively large putative prey items ver- tributary moved more than those from the Bure and Ant sus roach (mean lengths ± 95% CI: bream 388 ± 13 mm, 105 Page 8 of 13 S. Cittadino et al. Table 1 Results of statistical (A) Total range models on pike movements. (LM) TR ~ SW + R + (1 | FishID) (A) Linear model assessing Predictors Estimates 95% CI P the relationship between total range (TR) and the predictors (Intercept) 5547.59 1710.23–9384.94 0.006 spawning season (SW) and R [BURE] -2800.53 -7602.50–2001.44 0.246 river of capture (R), accounting for individual variability with R [THURNE] 5244.21 −342.87 to 10,831.29 0.065 the unique identification code SW [S] 2227.64 765.54–3689.74 0.004 (FishID) included as a random Random Effects effect. (B) Generalized linear σ 6,159,254.14 mixed model (GLMM) testing the influence of total length τ   24,905,982.87 00 Transmitter (TL), river of capture (R) and (B) Total movement spawning season (SW) on total (GLMM) TM ~ TL*R + SW movement (TM), with an added Predictors Estimates 95% CI P interaction term (TR*SW) (Intercept) 915.67 32.91–25,477.95 < 0.001 to explore potential joint effects. (C) General additive TL 1.01 1.00–1.01 0.003 model (GAM) evaluating the R [BURE] 0.24 0.00–29.08 0.562 impact of temperature (T) and R [THURNE] 76.74 1.27–4624.76 0.038 salinity (SAL) on the mean TL × R [BURE] 1 0.99–1.01 0.833 daily movement (MDM) of pike. Significant P values are TL × R [THURNE] 0.99 0.99–1.00 0.022 indicated in bold (C) Mean daily movement (GAM) MDM ~ s(T) + s(SAL) + SW + s(FishID, bs = ‘re’) Predictors Estimates 95% CI P (Intercept) 834.05 491.33–1415.83 < 0.001 SW [S] 1.98 1.38–2.84 < 0.001 Smooth term (T) 0.034 Smooth term (SAL) 0.193 Smooth term (FishID) < 0.001 range 212–502 mm, n = 83); roach: 131 ± 11  mm, range Discussion 78–229  mm, n = 78). The common bream showed also a significantly higher δ C ratio compared to roach Across the 30 pike used in movement analyses, there was (mean −28.78 ± 0.34 versus −29.92 ± 0.52%; t = 3.58, high individual variability, with the pike moving total dis- P < 0.01), but with no significant difference in δ N tances from 18 to 1182 km, with their mean daily move- ( me an 1 7. 05 ± 0. 3 2 ver su s 17 .2 9 ± 0. 57 %; t = − 0 .72 , ments being between 340 m and 3 km, with total ranges of P = 0.48). The isotopic signatures of Bream and Roach 6–26 km. The individuals that moved more tended to be showed no significant differences between the Bure and larger-bodied and were predicted to consume larger prey, Thurne sites (PERANOVA: P = 0.9 and P = 0.18, respec- with all movements tending to be greater during spawning tively). However, both species exhibited distinct isotopic versus non-spawning periods, with these results generally ratios in Thurne and Bure when compared with those in consistent with the predictions. The individual variability Ant (PERANOVA: P = 0.0015). in pike movements and spatial behaviour observed here The diet composition predictions from stable isotope is also evident in other pike populations. For example, in mixing models suggested that common bream was an the River Frome, southern England, pike revealed a con- important prey item for pike from the River Bure and tinuum of spatial behaviours, where some individuals were Thurne (Fig.  3), where there was a significant rela- always recorded in the same river reach of < 1 km length tionship between increasing pike length and increased while others moved repeatedly over several km (Masters predicted bream dietary contribution (GLM, P < 0.02; et al. 2002). In a Norwegian connected reservoir and river Table  2). Conversely, pike in the River Ant were pre- system, most pike had annual home ranges of less than dicted to have relatively high dietary contributions of 2 km, but with some individuals undertaking migrations of macroinvertebrates. Individual variability in the movement ecology of Northern pike Esox lucius in a highly connected… Page 9 of 13 105 Fig. 3 Predicted proportions (from stable isotope mixing models) of roach, common bream (‘Bream’) and macroinvertebrates (‘Macro’) in the diet of pike in the samples sites of a River Ant, b River Bure and c River Thurne. Whiskers display 95% confidence intervals over 14 km (Sandlund et al. 2016). Indeed, individual vari- ability in movements are also being apparent in a broad spectrum of species across aquatic and terrestrial environ- Table 2 Results of the general linear mixed model (GLMM) assess- ments (Shaw 2020). In entirety, these results emphasise ing the relationship between the proportion of bream in pike diet the importance of maintaining high habitat connectivity in (‘Bream’) and two predictor variables: pike total length (TL) and the wetland systems to enable pike and other fishes to access river of capture (R). Significant P values are indicated in bold a range of functional habitats over relatively large spatial (GLMM) Bream ~ TL + R Total length versus bream proportions areas. Although some of the individual pike in this study were Predictors Estimates 95% CI P highly vagile, there was minimal mixing of pike between the (Intercept) 0.27 0.21–0.34 < 0.001 three rivers. Pike tagged in the Bure were never detected in TL 1 1.00–1.00 0.048 the Ant or Thurne, while pike tagged in the Ant and Thurne Site [Bure] 1.19 1.09–1.31 < 0.001 were only rarely detected in the main Bure. This might Site [Thurne] 1.08 0.98–1.19 0.12 suggest a metapopulation structure, with fish in the Bure, 105 Page 10 of 13 S. Cittadino et al. Thurne and Ant comprising three distinct subpopulations to more permanent waters as the waters recede, although with limited dispersal among them. This partitioning of pike some individuals will move earlier to either avoid cannibal- into local groups with little ecological connectivity between ism (Cucherousset et al. 2009; Nilsson et al. 2014) or owing them was also detected by Lukyanova et al. (2024) in both to competitive displacement (Nyqvist et al. 2020). In the freshwater and brackish environments. Moreover, strong pat- Bure wetland, however, spawning period had no influence terns of site fidelity in pike are often detected in telemetry on the number of receiver detections in the main river ver- studies, where movements away from core areas are limited, sus the laterally connected macro-habitats, with these lateral especially outside of the reproductive period (Miller et al. habitats important all year round. The fish assemblage of the 2001; Kobler et al. 2008). For example, in the River Yser, ditch and dyke systems of the study area have previously Belgium, individually tagged pike demonstrated preferences been shown to be dominated by small numbers of pike, with for specific regions in the river in which they were detected their presence even recorded in the absence of prey fish spe- most frequently (Pauwels et al. 2014) and in a German lake, cies (other than occasional European eel Anguillia anguilla) translocated pike in the summer all returned to their main (Townsend and Peirson 1988). Townsend and Peirson (1988) activity centre within 6 days, suggesting strong fidelity to suggested that pike in these dyke systems would thus have these areas (Kobler et al. 2008). diets heavily reliant on non-fish prey, including from ter - The variability in movement metrics between individual restrial sources. pike was owing to some individuals moving between the The suggestion that even adult pike diet can rely on non- different macrohabitats present in each river, with most fish prey by Townsend and Peirson (1988) has been sup- detections occurring on receivers located in the laterally ported by other studies where individual pike have been connected waters rather than the main river, but with similar shown to be highly invertivorous (Beaudoin et al. 1999; Ven- fish numbers detected in all macrohabitats. This variability turelli and Tonn 2005) including in pike of up to 60 cm fork could have been influenced by variations in receiver detec- length (Pedreschi et al. 2015). This trophic flexibility was tion efficiency between the different macrohabitats. Indeed, also evident here, where stable isotope mixing models pre- receiver efficiencies can be influenced by many biotic and dicted that the diet of River Ant pike had relatively high pro- abiotic factors, including temperature, precipitation, extent portions of macroinvertebrates when compared with those of vegetation and suspended sediments (Winter et al. 2021b). in the Bure and Thurne. In contrast, Bure and Thurne pike However, the receiver array used here was considered as were predicted as having diets where common bream was an highly reliable during the study period, with detection range important – and large – prey item, with this dietary impor- very rarely falling below the width of the river (Winter et al. tance increasing as pike length increased (with common 2021b). Thus, the observed differences in pike detections bream having higher values of δ C compared with other and movements were considered as valid rather than an aquatic resources, as well as being an abundant fish species artefact of detection inefficiencies of the acoustic receivers. (Winter et al. 2021a,c)). Ontogenetic dietary switches are Moreover, high intra-population variability in movement common in pike through their development of larger gape ecology is a feature of many fish populations. In European sizes with increased body size, which facilitates their capture barbel, the majority of individuals have relatively limited and handling of larger prey items (Bry et al. 1995; Nilsson total ranges (< 10 km) but with a low proportion of individu- and Brönmark 2000). In the lower River Severn, western als having much larger total ranges (Gutmann Roberts et al Britain, the stable isotope ecology of pike of over 65 cm 2019; Amat Trigo et al., 2024). Juvenile Atlantic salmon indicated that relatively large fish (> 30 cm) were important Salmo salar predominantly have low mobility, adopting sed- prey items, but with these prey items generally absent in entary behaviour on most days, but with some individuals the diet of smaller pike (Nolan et al. 2019). However, the occasionally exhibiting bursts of high mobility, either by putative prey samples collected for SIA were limited here frequently moving within a confined area or moving between to aquatic resources, with terrestrial resources not being pools (Roy et al., 2013). considered as important to collect at the time of sampling. Highly connected systems are recognised as being impor- Accordingly, while we have confidence that the larger pike tant for providing the key functional habitats needed for all were consuming larger common bream as the reason for aspects of the pike lifecycle (Adolfsson 2020). Pike gener- their higher values of δ C (due to their high abundance in ally use wetland areas and back channels as spawning and the system), we cannot discount that it could also related nursery areas, which provide good habitats for larval pro- to some consumption of terrestrial prey items of relatively duction and recruitment compared with alternative habitats high δ C. (Nilsson, Engstedt and Larsson, 2014). The use of these The multivariate analyses indicated that the relationship wetland areas by pike is, however, often limited to spring between pike body length and total distance moved was sig- and early summer owing to their propensity of these areas nificant and positive, with individuals in the Thurne also to dry up later in summer, resulting in juveniles emigrating moving more than those in the Bure and Ant. In the River Individual variability in the movement ecology of Northern pike Esox lucius in a highly connected… Page 11 of 13 105 Frome, southern England, larger pike utilized 60% less of likely to be initiated by increasing water temperatures, the river length, relative to their body size, compared with with these warmer temperatures also playing a pivotal smaller individuals. This reduced movement is thought to role in the spawning movements of other pike popula- result from larger pike expanding their home ranges, which tions (Ovidio and Philippart, 2003). Although we argue increases spatial overlap between individuals, likely owing our results provide important insights into the dynamics to high levels of intra-specic fi competition, with this posited of pike movements in wetland systems, we acknowledge as owing to the pike increasing their home range sizes – and there are some limitations that necessitate some cautious thus the spatial overlap between individuals – owing to rela- interpretations. While the sample size provided valuable tively high intra-specific competition (Jetz et al. 2004, Ros- data, the relatively small number of analysed fish and the ten et al. 2016). Such patterns were, however, not evident in reliance on a single temperature monitoring station may our results, where the relationship between pike total range have constrained our ability to fully capture the variability and body size was not significant. Instead, we posit that our in pike behaviour and the influence of temperature on this. larger pike generally moved more than smaller pike – espe- Additionally, the deployment of acoustic receivers, while cially in the River Thurne – as a response to movements of generally efficient in detection, could have introduced their prey species. This hypothesis is based on our results variability in detection numbers across different habitats that indicated these larger pike not only moved significantly (given some areas had a relatively low receiver density), further than smaller pike but were also predicted to have potentially influencing the spatial patterns in detections. diets comprising relatively large proportions of the highly The absence of terrestrial prey resources in our analysis vagile and abundant common bream. This hypothesis was, also leaves room for further exploration of the trophic however, unable to be tested here owing to the use of low- influences on pike movement. Consequently, while our resolution acoustic telemetry, which was unable to accu- study contributes to the understanding of pike ecology, it rately quantify the movement relationships of the tagged highlights the need for more extensive research to build pike with the movements of shoals of common bream. on these findings and address the inherent challenges in Indeed, a limitation of low resolution telemetry is that while studying complex aquatic systems. it enables measurement of fish movements, it provides negli- In summary, there was high individual variability in the gible information on the actual activity of those fish, coupled movements of pike in this wetland system, with larger pike with it being unable to detect fish out of range of receiv - moving more that potentially facilitated their ability to pre- ers, which is an issue in a species, such as pike that spend date upon the highly vagile, abundant and large-bodied com- relatively long periods without moving far (Crossin et al. mon bream. Although some pike were highly vagile, their 2017; Jacoby and Piper 2023). Nevertheless, we speculate movements were generally between different macrohabitats that predator–prey dynamics could have played a significant in each river, rather than being long distance movements role in driving the movements of the larger tagged pike here, between rivers. If the connectivity of these habitats were where they actively sought optimal foraging areas where compromised owing to infrastructure developments, such large-bodied profitable prey, such as common bream, were as barriers, or declines in habitat quality, such as decreasing more abundant (Goldbogen et al. 2015; Florko et al. 2023). water levels and eutrophication (Ventura et al. 2023), pike The study pike significantly increased their total ranges might be compelled to cover greater distances in search of during the spawning period. Pike typically demonstrate suitable functional habitats. This enforced dispersal could a discernible partitioning between foraging and spawn- have a dual effect: on one hand, it might promote genetic ing areas, with fidelity expressed for specific spawning flow between distinct pike metapopulations, potentially grounds year after year (Craig 1996), and their spawn- influencing genetic diversity (Ouellet-Cauchon et al. 2014), ing migrations can be extensive, with some exceptional but could also result to increased energy expenditure, which cases reaching up to 78 km (Carbine and Applegate, 1948 might reduce overall fitness. This heightened energy demand in Craig 1996), and with Baltic pike demonstrating ana- could compromise the ability of individuals to effectively dromy during spawning (Lappalainen et al. 2008). Thus, compete for spatial and food resources, potentially leading to the increased total range of Bure pike during spawning increased risks of both interspecific and intraspecific compe- was likely owing to them making movements to specific tition (Bonte et al. 2012). Such extended movements might spawning areas, although this could not be tested within also expose fish to areas with varying resource availability our telemetry data owing to its low resolution that did not (Cooke et al. 2022). Accordingly, these results highlight that allow spawning areas to be reliably detected. Although preserving habitat connectivity is crucial for maintaining there were limitations owing to having a single tracking sustainable pike populations, as the connectivity ensures station for environmental variables and therefore possi- individuals can access diverse and profitable functional bility of temperature fluctuations throughout the system, habitats throughout the year and express their full range of our analyses suggested these spawning movements were movement behaviours. 105 Page 12 of 13 S. Cittadino et al. Supplementary Information The online version contains supplemen- Bonte D et al (2012) Costs of dispersal. Biol Rev. https:// doi. org/ 10. tary material available at https://doi. or g/10. 1007/ s00027- 024- 01124-4 .1111/j. 1469- 185X. 2011. 00201.x Brownscombe JW et al (2022) Application of telemetry and stable Acknowledgements We thank the anglers for their assistance in pike isotope analyses to inform the resource ecology and management sampling, especially the Pike Anglers Club of Great Britain (PAC) and of a marine fish. J Applied Ecol. https:// doi. org/ 10. 1111/ 1365- Norwich and District Pike Club (NDPC). We also thank the Environ-2664. 14123 ment Agency for their logistical help in pike tagging Bry C et al (1995) Early life characteristics of pike, Esox lucius, in rear- ing ponds: temporal survival pattern and ontogenetic diet shifts. Author Contribution J.R.B., E.W., A.H., S.L., R.W. and J.L. conceived J Fish Biol. https:// doi. org/ 10. 1111/j. 1095- 8649. 1995. tb059 49.x the study; J.R.B., E.W., A.H. and S.L. captured and tagged the fish; Campbell HA et al (2012) V-Track: software for analysing and visualis- S.C., E.W., A.S.T. and S.A. analysed the data; S.C., J.R.B. and A.S.T. ing animal movement from acoustic telemetry detections. Marine wrote the manuscript, and all authors edited the manuscript, and all Freshwater Res. https:// doi. org/ 10. 1071/ MF121 94 authors agree to its submission. Cantrell, D.L. et al. (2018) ‘The Use of Kernel Density Estimation With a Bio-Physical Model Provides a Method to Quantify Con- Funding Bournemouth University,Environment Agency,TÜBİTAK nectivity Among Salmon Farms: Spatial Planning and Manage- BİDEB,2219 Program, EU LIFE + Nature and Biodiversity Pro- ment With Epidemiological Relevance , Frontiers in Veterinary gramme, LIFE14NAT/UK/000054 Science. https://www.frontiersin.org/articles/https:// doi. org/ 10. 3389/ fvets. 2018. 00269. Data availability Raw data used in the manuscript that are not already Casselman JM (1974) ‘External Sex Determination of Northern Pike, provided are available from the corresponding author upon reasonable Esox lucius Linnaeus. Transact American Fisheries Society. request.https://doi. or g/10. 1577/ 1548- 8659(1974) 103% 3c343: ESDONP% 3e2.0. CO;2 Cooke SJ et al (2022) ‘The movement ecology of fishes. J Fish Biol. Declarations https:// doi. org/ 10. 1111/ jfb. 15153 Costantini ML et al (2018) ‘The role of alien fish (the centrarchid Conflict of interests The authors declare no competing interests. Micropterus salmoides) in lake food webs highlighted by stable isotope analysis ,. Freshwater Biol. https:// doi. org/ 10. 1111/ fwb. Open Access This article is licensed under a Creative Commons Attri- bution 4.0 International License, which permits use, sharing, adapta- Craig JF (1996) Population dynamics, predation and role in the tion, distribution and reproduction in any medium or format, as long community. In: Craig JF (ed) Pike: Biology and exploitation. as you give appropriate credit to the original author(s) and the source, Dordrecht, Springer, Netherlands provide a link to the Creative Commons licence, and indicate if changes Craig JF (2008) ‘A short review of pike ecology. Hydrobiologia. https:// were made. The images or other third party material in this article are doi. org/ 10. 1007/ s10750- 007- 9262-3 included in the article’s Creative Commons licence, unless indicated Crossin GT et al (2017) Acoustic telemetry and fisheries management. otherwise in a credit line to the material. If material is not included in Ecol Applicat. https:// doi. org/ 10. 1002/ eap. 1533 the article’s Creative Commons licence and your intended use is not Cucherousset J et al (2009) Spatial behaviour of young-of-the-year permitted by statutory regulation or exceeds the permitted use, you will northern pike (Esox lucius L) in a temporarily flooded nursery need to obtain permission directly from the copyright holder. To view a area. Ecol Freshwater Fish. https:// doi. org/ 10. 1111/j. 1600- 0633. copy of this licence, visit http://creativecommons.org/licenses/by/4.0/. 2008. 00349.x Flávio H, Baktoft H (2021) actel: Standardised analysis of acoustic telemetry data from animals moving through receiver arrays. Met Ecol Evol. https:// doi. org/ 10. 1111/ 2041- 210X. 13503 References Florko KRN et al (2023) The dynamic interaction between predator and prey drives mesopredator movement and foraging ecology. Bio. https:// doi. org/ 10. 1101/ 2023. 04. 27. 538582 Acolas ML, Le Pichon C, Rochard E (2017) ‘Spring habitat use by Fullerton AH et al (2010) ‘Hydrological connectivity for riverine fish: stocked one year old European sturgeon Acipenser sturio in the measurement challenges and research opportunities. Freshwater freshwater-oligohaline area of the Gironde estuary. Estuarine Biol. https:// doi. org/ 10. 1111/j. 1365- 2427. 2010. 02448.x Coastal Shelf Sci. https:// doi. org/ 10. 1016/j. ecss. 2017. 06. 029 Goldbogen JA et al (2015) Prey density and distribution drive the three- Adolfsson, O. (2020) Consequences on population dynamics following dimensional foraging strategies of the largest filter feeder. Funct regained connectivity in pike (Esox lucius) spawning location. : Ecol. https:// doi. org/ 10. 1111/ 1365- 2435. 12395 https:// urn. kb. se/ resol ve? urn= urn: nbn: se: lnu: diva- 104213. Grill G et al (2019) Mapping the world s free-flowing rivers. Nature. Amat-Trigo F et al (2024) Variability in the summer movements, habi- https:// doi. org/ 10. 1038/ s41586- 019- 1111-9 tat use and thermal biology of two fish species in a temperate Gutmann Roberts C, Hindes AM, Britton JR (2019) ‘Factors influenc- river. Aquatic Sci. https:// doi. org/ 10. 1007/ s00027- 024- 01073-y ing individual movements and behaviours of invasive European Amoros C, Bornette G (2002) Connectivity and biocomplexity in barbel Barbus barbus in a regulated river. Hydrobiologia. https:// waterbodies of riverine floodplains. Freshwater Biol. https:// doi. doi. org/ 10. 1007/ s10750- 018- 3864-9 org/ 10. 1046/j. 1365- 2427. 2002. 00905.x Home | Ramsar Sites Information Service. https:// rsis. ramsar. org/. Araújo MS, Bolnick DI, Layman CA (2011) The ecological causes of Ieno, E.N. ‘A beginner s guide to data exploration and visualisation individual specialisation. Ecol Letters. https:// doi. org/ 10. 1111/j. with R. : https:// cir. nii. ac. jp/ crid/ 11308 48327 91313 9328. 1461- 0248. 2011. 01662.x Jacoby DMP, Piper AT (2023) What acoustic telemetry can and can- Beaudoin CP et al (1999) Individual specialization and trophic adapt- not tell us about fish biology. J Fish Biol. https://doi. or g/10. 1111/ ability of northern pike (Esox lucius): an isotope and dietary jfb. 15588 analysis. Oecologia. https:// doi. org/ 10. 1007/ s0044 20050 871 Kobler A, Klefoth T, Arlinghaus R (2008) ‘Site fidelity and seasonal Belletti B et al (2020) More than one million barriers fragment Europe changes in activity centre size of female pike Esox lucius in a s rivers. Nature. https:// doi. org/ 10. 1038/ s41586- 020- 3005-2 Individual variability in the movement ecology of Northern pike Esox lucius in a highly connected… Page 13 of 13 105 small lake. J Fish Biol. https://doi. or g/10. 1111/j. 1095- 8649. 2008. Roy L et al (2012) Individual variability in the movement behaviour of 01952.x juvenile Atlantic salmon. Canadian J Fish Aquatic Sci [preprint]. Koed A et al (2006) Annual movement of adult pike (Esox lucius L.) https:// doi. org/ 10. 1139/ cjfas- 2012- 0234 in a lowland river. Ecol Freshwater Fish. https://doi. or g/10. 1111/j. Sandlund OT, Museth J, Øistad S (2016) Migration, growth patterns, 1600- 0633. 2006. 00136.x and diet of pike (Esox lucius) in a river reservoir and its inflowing Lappalainen A et al (2008) Reproduction of pike (Esox lucius) in reed river. Fisher Res. https:// doi. org/ 10. 1016/j. fishr es. 2015. 08. 010 belt shores of the SW coast of Finland, Baltic Sea: a new survey Shao X et al (2019) River network connectivity and fish diversity. Sci approach: Boreal Environment Research. Boreal Environ Res Total Environ. https:// doi. org/ 10. 1016/j. scito tenv. 2019. 06. 340 13(4):370–380 Shaw AK (2020) Causes and consequences of individual varia- Layman CA et al (2012) Applying stable isotopes to examine food-web tion in animal movement. Move Ecol. https:// doi. org/ 10. 1186/ structure: an overview of analytical tools. Biol Rev. https:// doi. s40462- 020- 0197-x org/ 10. 1111/j. 1469- 185X. 2011. 00208.x Speed CW et al (2010) Complexities of coastal shark movements and Masters, J.E.G. et al. (2002) ‘Habitat utilisation by pike Esox lucius their implications for management. Marine Ecol Progres. https:// L. during winter floods in a southern English chalk river , in E.B. doi. org/ 10. 3354/ meps0 8581 Thorstad, I.A. Fleming, and T.F. Næsje (eds) Aquatic Teleme- The Scaling of Animal Space Use | Science. https://www.science.org/ try: Proceedings of the Fourth Conference on Fish Telemetry in doi/full/https:// doi. org/ 10. 1126/ scien ce. 11021 38. Europe. Dordrecht, Springer, Netherlands. Thieme M et al (2023) Measures to safeguard and restore river con- Miller LM, Kallemeyn L, Senanan W (2001) Spawning-site and natal- nectivity. Environment Rev [preprint]. https:// doi. org/ 10. 1139/ site fidelity by northern pike in a large lake: mark-recapture and er- 2023- 0019 genetic evidence. Transaction American Fisher Societ. https:// Townsend CR, Peirson G (1988) Fish community structure in lowland doi.or g/10. 1577/ 1548- 8659(2001) 130% 3c0307: SS ANSF%3e2.0. drainage channels. J Fish Biol. https:// doi. org/ 10. 1111/j. 1095- CO;28649. 1988. tb053 62.x Nilsson PA, Brönmark C (2000) ‘Prey vulnerability to a gape-size lim- Udyawer V et al (2018) A standardised framework for analysing animal ited predator: behavioural and morphological impacts on northern detections from automated tracking arrays. Animal Biotelemet. pike piscivory. Oikos. https:// doi. org/ 10. 1034/j. 1600- 0706. 2000. https:// doi. org/ 10. 1186/ s40317- 018- 0162-2 880310.x van Puijenbroek PJTM et al (2019) Species and river specific effects of Nilsson J, Engstedt O, Larsson P (2014) Wetlands for northern pike river fragmentation on European anadromous fish species. River (Esox lucius L) recruitment in the Baltic Sea. Hydrobiologia. Res Application. https:// doi. org/ 10. 1002/ rra. 3386 https:// doi. org/ 10. 1007/ s10750- 013- 1656-9 Vander Zanden HB et al (2010) Individual specialists in a generalist Nolan ET, Gutmann Roberts C, Britton JR (2019) ‘Predicting the con- population: results from a long-term stable isotope series. Biol tributions of novel marine prey resources from angling and ana- Letter. https:// doi. org/ 10. 1098/ rsbl. 2010. 0124 dromy to the diet of a freshwater apex predator. Freshwater Biol. Vehanen T et al (2006) Patterns of movement of adult northern pike https:// doi. org/ 10. 1111/ fwb. 13326 (Esox lucius L) in a regulated river. Ecol Freshwater Fish. https:// Nunn AD et al (2010) ‘Seasonal and diel patterns in the migrations of doi. org/ 10. 1111/j. 1600- 0633. 2006. 00151.x fishes between a river and a floodplain tributary. Ecol Freshwater Ventura M et al (2023) When climate change and overexploitation meet Fish. https:// doi. org/ 10. 1111/j. 1600- 0633. 2009. 00399.x in volcanic lakes: the lesson from lake bracciano, rome’s strategic Nyqvist MJ et al (2018) Relationships between individual movement, reservoir. Water. https:// doi. org/ 10. 3390/ w1510 1959 trophic position and growth of juvenile pike (Esox lucius). Eco Venturelli PA, Tonn WM (2005) Invertivory by northern pike (Esox Freshwater Fish. https:// doi. org/ 10. 1111/ eff. 12355 lucius) structures communities of littoral macroinvertebrates in Nyqvist MJ et al (2020) Dispersal strategies of juvenile pike (Esox small boreal lakes. J North American Benthol Soci. https://do i. lucius L.): Influences and consequences for body size, somatic org/ 10. 1899/ 04- 128.1 growth and trophic position. Ecol Freshwater Fish. https:// doi. Verhelst P et  al (2018) Downstream migration of European eel org/ 10. 1111/ eff. 12521 (Anguilla anguilla L.) in an anthropogenically regulated fresh- Oele DL et al (2019) Growth and recruitment dynamics of young-of- water system: Implications for management. Fisheries Res. https:// year northern pike: Implications for habitat conservation and man-doi. org/ 10. 1016/j. fishr es. 2017. 10. 018 agement. Ecol Freshwater Fish. https://doi. or g/10. 1111/ eff. 12453 Winter ER et al (2021a) Acoustic telemetry reveals strong spatial pref- Oksanen, J. (2012) ‘Constrained Ordination: Tutorial with R and vegan. erences and mixing during successive spawning periods in a par- Ouellet-Cauchon G et al (2014) Landscape variability explains spatial tially migratory common bream population. Aquatic Sci. https:// pattern of population structure of northern pike (Esox lucius) in doi. org/ 10. 1007/ s00027- 021- 00804-9 a large fluvial system. Ecol Evolut. https:// doi. org/ 10. 1002/ ece3. Winter ER et al (2021b) ‘Detection range and efficiency of acoustic 1121 telemetry receivers in a connected wetland system. Hydrobiologia. Ovidio M, Philippart JC (2005) Long range seasonal movements of https:// doi. org/ 10. 1007/ s10750- 021- 04556-3 northern pike (Esox lucius L.) in the barbel zone of the River Winter ER et al (2021c) ‘Dual-isotope isoscapes for predicting the Ourthe. River Meuse basin, Belgium scale of fish movements in lowland rivers. Ecosphere. https://doi. Pauwels IS et al (2014) ‘Movement patterns of adult pike (Esox lucius org/ 10. 1002/ ecs2. 3456 L.) in a Belgian lowland river. Ecol Freshwater Fish. https:// doi. Winter ER et  al (2021d) ‘Movements of common bream Abramis org/ 10. 1111/ eff. 12090 brama in a highly connected, lowland wetland reveal sub-popu- Pedreschi D et al (2015) ‘Trophic flexibility and opportunism in pike lations with diverse migration strategies. Freshwater Biol. https:// Esox lucius. J Fish Biol. https:// doi. org/ 10. 1111/ jfb. 12755doi. org/ 10. 1111/ fwb. 13726 Power ME, Dietrich WE (2002) Food webs in river networks. Ecolog Res. https:// doi. org/ 10. 1046/j. 1440- 1703. 2002. 00503.x Publisher’s Note Springer Nature remains neutral with regard to Rosten CM, Gozlan RE, Lucas MC (2016) Allometric scaling of jurisdictional claims in published maps and institutional affiliations. intraspecific space use. Biol Letters. https://doi. or g/10. 1098/ rsbl. 2015. 0673

Journal

Aquatic Sciences – Springer Journals

Published: Oct 1, 2024

Keywords: Pike movement ecology; River connectivity; Acoustic telemetry; Individual variability

There are no references for this article.