Biodiversity Data Journal :
Data Paper (Biosciences)
|
Corresponding author: S. Jannicke Moe (jmo@niva.no)
Academic editor: Yasen Mutafchiev
Received: 17 Mar 2020 | Accepted: 15 May 2020 | Published: 28 May 2020
© 2020 S. Jannicke Moe, Chloé Nater, Atle Rustadbakken, L. Asbjørn Vøllestad, Espen Lund, Tore Qvenild, Ola Hegge, Per Aass
This is an open access article distributed under the terms of the Creative Commons Attribution License (CC BY 4.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Citation:
Moe SJ, Nater CR, Rustadbakken A, Vøllestad LA, Lund E, Qvenild T, Hegge O, Aass P (2020) Long-term mark-recapture and growth data for large-sized migratory brown trout (Salmo trutta) from Lake Mjøsa, Norway. Biodiversity Data Journal 8: e52157. https://doi.org/10.3897/BDJ.8.e52157
|
|
Long-term data from marked animals provide a wealth of opportunities for studies with high relevance to both basic ecological understanding and successful management in a changing world. The key strength of such data is that they allow us to quantify individual variation in vital rates (e.g. survival, growth, reproduction) and then link it mechanistically to dynamics at the population level. However, maintaining the collection of individual-based data over long time periods comes with large logistic efforts and costs and studies spanning over decades are therefore rare. This is the case particularly for migratory aquatic species, many of which are in decline despite their high ecological, cultural and economical value.
This paper describes two unique publicly available time series of individual-based data originating from a 51-year mark-recapture study of a land-locked population of large-sized migratory brown trout (Salmo trutta) in Norway: the Hunder trout. In the period 1966-2015, nearly 14,000 adult Hunder trout have been captured and individually marked during their spawning migration from Lake Mjøsa to the river Gubrandsdalslågen. Almost a third of those individuals were later recaptured alive during a later spawning run and/or captured by fishermen and reported dead or alive. This has resulted in the first data series: a mark-recapture-recovery dataset spanning half a century and more than 18,000 capture records. The second data series consists of additional data on juvenile and adult growth and life-history schedules from half of the marked individuals, obtained by means of scale-sample analysis. The two datasets offer a rare long-term perspective on individuals and population dynamics and provide unique opportunities to gain insights into questions surrounding management, conservation and restoration of migratory salmonid populations and freshwater ecosystems.
mark-recapture, sclerochronology, individual-based, long-term, Salmo trutta, survival, growth, life-history, hydropower dam, river regulation, migration, freshwater
Important processes in ecology and evolution of vertebrates occur over the course of multiple years and often decades. Many areas of ecological and evolutionary research – including most of studies with the goals of improving species management and conservation – therefore rely on the availability of data spanning long time periods. Long-term ecological data on animal populations may be collected either at the population level (e.g. count or occupancy surveys) or by following individuals with uniquely identifiable marks throughout their lives. Individual-based mark-recapture and life-history data resulting from the latter provide a wealth of opportunities for studies that are impossible with only population-level data, as they not only allow linking population dynamics to vital rates (e.g. survival, growth, reproduction), but also enable the study of individual differences in those vital rates (
Recent declines of freshwater species abundance is more severe than species declines on land or in the ocean, according to the latest Living Planet report (
This paper describes two unique individual-based datasets that have resulted from a long-term study on a land-locked population of large-sized, piscivorous brown trout (Salmo trutta) in Norway, commonly referred to as Hunder trout (Fig.
We here describe a long-term study of the Hunder trout and provide documentation for the resulting (capture-)mark-recapture-recovery (CMRR) data and growth data with the aim of facilitating re-use of this unique data resource. We include important information on biology, habitat, study design and sampling protocols alongside detailed metadata for both datasets and notes on their explicit and implicit connections and potential uses.
The Hunder trout population
The Hunder trout is a migratory brown trout inhabiting Lake Mjøsa and its main inlet River Gudbrandsdalslågen (also called Lågen) in southeast Norway. The population is famous for its large body size (with some individuals measuring > 1 m and weighing > 10 kg) and its life history and spawning biology have been well studied (
Anthropogenic impacts: harvest, hydropower and mitigation measures
The Hunder trout has a long history of being impacted by human activity both directly through harvest and indirectly through habitat alteration.
The Hunder trout has been the target of intense harvest for centuries; see
Two major dams were established in River Gudbrandsdalslågen in the 1960s, the Harpefoss and Hunderfossen dams (Fig.
Map of Lake Mjøsa and the river Gudbransdalslågen, where the Hunder brown trout occurs. The fish ladder and the hatchery are located at the Hunderfossen dam. Sources: The Norwegian Mapping Authority (https://www.kartverket.no) and ESRI (https://www.esri.com). After
Damming also drastically altered the hydrological conditions in the river, leading to a reduction in availability and quality of spawning and nursing habitats in the river, both upstream and downstream of the dam (
Project: Sustainable management of renewable resources in a changing environment: an integrated approach across ecosystems (SUSTAIN). https://www.sustain.uio.no/
Work Package 1: Demographic structure in harvested ecosystems. https://www.sustain.uio.no/research/wp/work-package-1
Lake Mjøsa and River Gudbrandsdalslågen
Lake Mjøsa is a deep fjord lake (max. depth 453 m) situated in southeast Norway (Fig.
In the 1970s and 1980s, eutrophication due to excessive nutrient loads from agriculture, industry and households resulted in poor water quality and harmful algal blooms in Mjøsa. A subsequent large restoration effort (the Lake Mjøsa campaign) resulted in a period of re-oligotrophication and the current good ecological status of the lake (
Research Council of Norway, project SUSTAIN, contract no. 244647/E10. https://www.sustain.uio.no/
The establishment of the fish ladder within the Hunderfossen dam resulted in unique opportunities for monitoring the spawning population of the Hunder trout. Concurrent with the opening of the fish ladder in 1966, a mark-recapture programme was initiated and was run continuously until termination in 2016. Harvest recoveries were still recorded in 2017. This monitoring programme has resulted in two unique long-term data series: a 51-year time series of (capture-)mark-recapture-recovery (CMRR) data and 37 years of scale samples resulting in a 51-year time series of growth data (Fig.
Overview of availability of (a) (capture-)mark-recapture-recovery (CMRR) data and (b) growth data over the course of the study period (1966 - 2017). For CMRR data, counts of trout captures are categorised by capture type (see colour legend). Captures of types "trap dead" and "trap alive" occur in the trap within the Hunderfossen fish ladder; "harvest" and "catch release" are captures by fishers in any location and resulting in death of the fish for the former and alive re-release for the latter; capture type "found dead" is assigned to fish that are discovered already dead in any location except for the trap within the Hunderfossen fish ladder. For more details, see Table S.2 in Suppl. material
In the following, we describe the generation of mark-recapture data from trapping of spawning trout in the Hunderfossen fish ladder, its supplementation with recovery data obtained as reports of marked fish by local fishers and sampling and sclerochronological analysis of scales resulting in data on individual growth and life-history scedules.
All data were originally collected, recorded, maintained and processed by a variety of individuals and institutions, as listed in the Author contributions and Acknowledgements sections.
Marking and recaptures in the fish ladder
All adult trout ascending the Hunderfossen hydropower dam on their spawning migration between 1966 and 2016 were captured in a trap situated in the lower part of the fish ladder. Until 2015, all unmarked trout were individually marked with Carlin tags, consisting of a plastic disc with a unique mark number (alphanumeric code) and a stainless steel thread to attach to the fish (Fig.
Recoveries from fishing reports and others
Marked trout could be recaptured in the fish ladder during later spawning runs, as well as by fishers in any part of the river or lake. Reports of such harvest recoveries are available from 1966 to 2017. When trout were caught by fishers and reported, the amount of data collected varied. As a minimum, fishers reported the mark number of the caught fish, the date of capture and whether the fish was killed or released. Additional information could include capture location, fishing gear, sex, origin, length and weight. Reporting the capture/harvest of a marked trout was done on a voluntary basis and no monetary reward was given in exchange for reports. In combination with declining feedback to fishers from the marking project, this likely led to a gradual decrease in fisher’s reporting rate over the course of the study period (
On rare occasions, marked trout were found dead and reported by people who did not capture them through fishing themselves. This includes, for example, trout that died while attempting dam passage and were found dead in the vicinity of the hydropower station, trout that died naturally and washed up on the shores of the lake or the river and even reports from people which had purchased a marked trout on a fish market. All of these reports were included in the CMRR dataset.
Scale sampling and sclerochronological analysis
During the period 1966-2015, 4-6 scales were obtained from a large number of trout captured in the fish ladder. The scales were sampled from a standardised location above the lateral line between the dorsal and the adipose fin using small forceps. For a subset of individuals that ascended the fish ladder between 1966 and 2003, the sampled scales have been analysed using schlerochronological methods (
Illustration of a fish scale and interpretation of growth, development and spawning events. The green line represents the time line from hatching (centre) to the current age (the edge). Red lines indicate winters (marked "V"), where the growth is slower (denser sclerite pattern). Smoltification ("S") after the 4th winter is indicated by subsequent rapid growth. Spawning runs ("G") are followed by very slow growth during the summer proper to and during the spawning run.
The spatial extent of the river and lake system the Hunder trout inhabits spans the latitudes 60.40° - 61.22°N and longitudes 10.43° - 11.29°E (GCS, WGS84) (Fig.
60.40 and 61.22 Latitude; 10.43 and 11.29 Longitude.
The data are from a single population of large-sized migratory brown trout (Salmo trutta), also referred to as Hunder trout, inhabiting Lake Mjøsa and River Gudbrandsdalslågen.
Rank | Scientific Name | Common Name |
---|---|---|
species | Salmo trutta | Brown trout |
The (capture-)mark-recapture-recovery (CMRR) data contains captures from all years in the period 1966-2017. As individual marking was discontinued after the 2015 sampling season, no new individuals were marked in 2016 (recaptures and recoveries only) and only harvest reports were recorded for 2017.
Scales were sampled in the period 1966-2015. Scales sampled during the years 1966-2003 have been analysed so far. Individual growth data is therefore available from 1952 (due to back-calculation) to 2003.
The data described here has been made publicly available.
We do note, however, that successful re-use of long-term individual-based data requires good understanding of both data stucture and the biological system, as well as an overview over previous work using the data, whether that work has been published in a scientific journal or not (see
For the above reason, we encourage parties interested in re-using the here-described data to contact the authors CRN, SJM, AR or LAV for a discussion of research ideas and potential opportunities for collaboration.
The mark-recapture-recovery dataset (CMRR) contains altogether 18,488 capture records from years 1966-2017. Of these, 13,975 are individual marking events of mature trout in the ladder. The remaining 4,513 capture records are recaptures by different means (as defined by CaptureType): 2,106 recaptures in the trap (376 of which resulted in death); 2,322 harvest reports (1,944 in the lake, 358 in the river, 20 in other or unknown locations); 72 reports of fish found dead; and nine reports of catch-and-release fishing. An overview of all recorded captures across years is given in Fig.
Ratios of (a) origin (stocked vs. wild) and (b) sex (male vs. female) of trout captured alive in the fish ladder (black) and scales sampled (grey) over the course of the study period. Ratios for scales samples in 2003 are not depicted, since only a single scale is available for this year. The dotted blue line marks the line of equal ratios (50%). Note that y-axis scaling differs for panels a) and b).
Annual distributions of individual body lengths measured upon capture in the fish ladder. The box plot shows the median (horizontal black bar), 25% and 75% quantiles (range of the box), largest and smallest values within 1.5 times the interquartile range (whiskers = vertical lines) and outliers (dots beyond whiskers).
Column label | Column description |
---|---|
MarkNo | The alphanumeric code of an individual’s tag. All information on individual fish is linked to the MarkNo. |
CaptureNo | The running number of the captures of an individual ordered by capture date, including the marking event as CaptureNo = 1. |
CaptureNoMax | The total number of captures recorded for an individual. |
MarkDate | The exact date of the marking event, if this information is available. |
MarkYear | The year of the marking event. |
CaptureDate | The exact date of the capture event, if this information is available. |
CaptureDay | The day of the capture event |
CaptureMonth | The month of the capture event. |
CaptureYear | The year of the capture event. |
CaptureArea | Area of the capture event. All captures in the trap at the Hunderfossen dam (including marking events) are denoted as “river trap”. The reported locations of harvests (captures by fishers) are aggregated into five areas: "lake" (between Minnesund and Lillehammer), "river below" (between Lillehammer and Hunderfossen), "river above" (between Hunderfossen and Harpefossen), “river” (north of Lillehammer, but unclear if above or below Hunderfossen) and "far away" (elsewhere). |
CaptureType | Classification of the type of capture events based on information on capture location and alive/dead state prior to and after capture. The distinguished capture types are defined as follows: “trap alive” = captured in the fish ladder during a spawning run and released alive, “trap dead” = captured in the fish ladder during a spawning run but died prior to/shortly after release (or found dead in the immediate vicinity of the fish ladder), “harvest” = captured by a fisher and killed, “catch release” = captured by a fisher and released alive and “found dead” = found dead in any location other than the fish ladder. |
CaptureEquipment | The gear involved in the capture summarised into five standardised categories (aggregated from more detailed reported information of equipment). For harvest captures, the fishing gear has been categorised as either "net" or "rod"; for other capture types, the possible values are "trap", "dead in trap" or "found dead". |
LengthAtCapture | The measured length of the fish at the capture (unit mm). |
WeightAtCapture | The measured weight of the fish at the capture (unit g). |
Sex | Sex of the individual, categorised as female, male or NA. Determined based on secondary sex characteristics. |
Origin | Origin of the individual categorised as wild, stocked or NA. Stocked individuals are first-generation offspring of wild fish that have been reared in a hatchery near the Hunderfossen and are recognisable by their clipped adipose fin. |
ScaleData | Information on whether growth and life-history data (from at least one scale sample) is available for this individual in the growth dataset. |
The growth dataset contains altogether 41,605 size records back-calculated from 7047 scale samples collected between 1966 and 2003. The scales originate from 6,875 marked individuals, which are also present in the mark-recapture-recovery (CMRR) data. For the majority of individuals, growth data is available from a single scale sample in the dataset. Some 161, 10 and 1 individual(s) have growth data from two, three and four scales collected at different capture events, respectively. Backcalculated sizes are divided into the juvenile period of life spent in the river and the adult period of life spent predominantly in the lake and are supplemented with information on key life history events (smolting, maturation, spawning status). An overview of all sampled scales across years is given in Fig.
Column label | Column description |
---|---|
MarkNo | The alphanumeric code of an individual’s tag. Refers to the same tag as used in the mark-recapture-recovery (CMRR) dataset. |
ScaleNo | The running number of the scale readings from the individual ordered by sampling (capture) date. |
ScaleNoMax | The total number of scales sampled and analysed for the individual. |
Period | Information on individual's life-history stage at a given age. "river" refers to the juvenile period of life spent in the river (prior to smolting), "lake" refers to the adult period of life spent predominantly in the lake (after smolting). Data from the "river" period are mostly unavailable for stocked fish as they spend the juvenile period in the hatchery. Data from the "river" period may also be missing for wild-born individuals when sampled scales were replacement scales or when growth increments for the early period were otherwise unreadable. |
AgeTotal | Total age in of the individual based on the number of growth increments on the scale. Equals the sum of AgeRiver and AgeLake and is, therefore, only available when information on river growth is available. |
AgeRiver | Age of the individual while in the juvenile period of life in the river. Equals AgeTotal until smolting. AgeRiver = 1 corresponds to having spent 1 year in the river since hatching. |
AgeLake | Age of the individual while in the adult period of life in the lake, independent of the number of years spent in the river or in the hatchery as a juvenile previously. AgeLake = 1 corresponds to having spent 1 year in the lake since smolting (migration from river to lake for wild fish, release from the hatchery for stocked fish). Normally available also when information on river growth is missing. |
Length | Backcalculated spring length at a given age (AgeRiver and/or AgeLake). Backcalculated assuming isometry of scale and fish growth and using the Lea-Dahl method (Bagenal 1978). |
SmoltingStatus | Binary variable indicating whether this individual has smolted (1) or not (0). The first occurrence of 1 marks the year of smolting. Based on the transition from slow river growth (dense increments on the scale) to fast lake growth (more widely-spaced increments). |
MaturationStatus | Binary variable indicating whether this individual has reached sexual maturity and has had at least one spawning run (1) or not (0). The first occurence of 1 marks the year of maturation and first spawning run. Based on spawning marks on the scale and capture in the fish ladder for the sampling year. |
SpawnStatus | Binary variable indicating whether this individual had a spawning run in autumn of this year (1) or not (0). Based on spawning marks on the scale and capture in the fish ladder for the sampling year. |
Year | Year of backcalculated spring length based on AgeTotal or AgeLake (if AgeTotal unavailable) and CaptureYear. |
Sex | Sex of the individual, categorised as female, male or NA. Assessed based on secondary sex characteristics. |
Origin | Origin of the individual categorised as wild, stocked or NA. Stocked individuals are first-generation offspring of wild fish that have been reared in a hatchery near the Hunderfossen and are recognisable by their clipped adipose fin. |
HatchYear | Individual's backcalculated year of hatching (spring of year when AgeTotal = 0). Only available if AgeTotal is available. |
AgeAtSmolting | Individual's backcalculated age in spring of the smolting year. Equals last AgeRiver. Only available if data on river (juvenile) growth and AgeTotal are available, thus unavailable for the majority of stocked fish. |
LengthAtSmolting | Individual's backcalculated length in spring of the smolting year. Equals backcalculated length at the last AgeRiver when river growth is available. |
AgeAtMaturation | Individual's backcalculated age in spring prior to the first spawning run (first spawning mark on this scale). Only available if AgeTotal is available. |
LengthAtMaturation | Individual's backcalculated length in spring prior to the first spawning run (first spawning mark on this scale). |
SpawnCount | The total number of spawning runs this individual has had up to and including the spawning run during which the scale sample was taken. Based on spawning marks visible on the scale plus the spawning run during which the scale was sampled. |
CaptureYear | The year of the capture event during which the scale sample was obtained. Equivalent to the same field in the CMRR dataset. |
CaptureNo | The running number of the capture event during which the scale sample was obtained. Equivalent to the same field in the CMRR dataset. |
The SUSTAIN trout database is a compilation of all available data on individuals, captures and scale-based information (growth and life history schedules) available for the Hunder trout, from which the two datasets, described in this paper, have been derived. For further use of the Hunder trout data, we recommend using the published datasets. To better explain the context and the data processing underlying the published dataset, we provide a more thorough description of this database and its structure in Suppl. material
The above described CMRR and growth data are linked in the relational database as described in Suppl. material
An individual's body length measured at capture in the fish ladder (LengthAtCapture) during the spawning season is used as the starting point for the backcalculation of size using a scale collected during the first capture. Since increments on scales are used to backcalculate spring size, however, this is done under the assumption that a trout's length measured during the spawning run in autumn is roughly equal to its spring length earlier in the same year. This is not unreasonable, since growth in spawning years is greatly reduced relative to non-spawning years and close to negligible for larger individuals (e.g.
Scale samples for some individuals have been obtained not (only) during the first capture in the fish ladder, but (also) during later recaptures in the fish ladder. For such individuals, information on body length in certain years may thus be available from both scale backcalculations and actual growth measurements during capture in the fish ladder. By assuming no growth during the winter (a reasonable assumption for brown trout,
Correlation of backcalculated length from scales in spring (year t) and measured length in the fish ladder in autumn of the previous year (t-1). The blue dots represent data points (N = 201), the dashed line marks the 1-to-1 perfect correlation, and the red line is the fit of a linear regression including its confidence interval (intercept = 10.726, slope = 0.974, adj. R2 = 0.875).
Besides the two here-described long-term individual-based datasets, additional data and samples have been collected on the Hunder trout in connection with the stocking programme and/or as part of shorter-term projects. Some of these data/samples may be made available on request by their respective owners and we briefly mention some of them here due to their potential value for integration with the here-described long-term data in future studies.
Additional scale samples
Scale samples were collected continually for the period 2004-2015 under the same protocol as described in this article. These additional scale samples have not been analysed using sclerochronological methods yet. These are stored by the County Governor of Innlandet and the Lands museum (https://randsfjordmuseet.no/lands-museum). These scale samples may be used for sclerochronological - and potentially also genetic analysis - in the future (in the context of genetic analyses, stored samples for the period 1966-2003 are also available).
Contact: County Governor of Innlandet (fminpost@fylkesmannen.no)
Released stocked trout
Several additonal types of data have been collected in connection with the stocking programme. Numbers, body size distribution (as discrete classes) and location of all hatchery smolt releases have been recorded for most years from 1965 onwards and these data were digitised and aggregated for the period 1965-2018 by Chloé R. Nater and Marlene W. Stubberud. Furthermore, a considerable number of stocked smolt were individually marked upon release as part of experimental studies predominantly between 1970 and 1990. Mark-recapture-recovery data on over 13,000 hatchery-reared smolt (including 10,000+ harvest events and around 200 recaptures in the fish ladder) have been separated out of the here-described CMRR dataset and are available upon request (note, however, that due to this separation happening at an earlier stage, parts of these data have not been harmonised/error-checked completely).
Contact: Chloé R. Nater
Fecundity
A small set of data on female fecundity (number of eggs held by a female of a certain size) has been collected by Chloé R. Nater, Asbjørn Vøllestad and Yngvild Vindenes during the spawning season in 2017 and 2018. Sample size is small (15 females), but the data may be used to infer fecundity under consideration of typical relationships of female size and fecundity in salmonid fishes (see
Contact: Chloé R. Nater
Telemetry
The movement of brown trout in River Gudbrandsdalslågen, especially with regards to hydropower dams, river regulation and habitat fragmentation, has been the subject of a variety of telemetry studies over time (see, for example,
Contact: Refer to corresponding author contact information in the above-cited literature.
Related long-term environmental monitoring data are available for Lake Mjøsa and its catchment, including data on meteorology, hydrology, river and lake water quality and lake biology (see also
Contact: Jannicke Moe
The project SUSTAIN (Research Council of Norway, contract no. 244647/E10) has funded compilation, harmonisation, checking and publication of the SUSTAIN trout data, as well as the writing of this data paper.
Collection and maintenance of the data has been financed, carried out and supported by several institutions and individuals. Opplandskraft DA, who operate the Hunderfossen dam and powerplant, have been responsible for both the stocking programme (since initialisation) and the marking and recapturing of trout in the fish ladder (since 1988). Special thanks go Åse Brenden and Frank Hansen for their outstanding contributions to the operation of the hatchery and stocking programme. The “Merkesentralen” of NINA (Norwegian Institute for Nature Research) has been in charge of managing the system for the reports of tags from outside the fish ladder (especially from harvest). We acknowledge the substantial contribution of the local fishers who have voluntarily reported tags and associated information on their catches for over half a century. The County Governors of Oppland and Hedmark (now merged to "Innlandet") have both played a vital role in combining and maintaining the data from both the marking and recaptures at Hunderfossen and the reported tags. Data maintenance and cleaning has further been supported by The Norwegian Environment Agency. Prof. Thrond O. Haugen (Norwegian University of Life Sciences) has contributed substantially to earlier efforts of partially cleaning and harmonising the data throughout the last 15 years.
SJM and CRN share the first authorship. The authors have contributed to this data paper according to the following CRediT statements. Contributions that mainly predate the SUSTAIN project (2015-2019) are indicated by an asterisk.
Conceptualisation (of data compilation and data paper): LAV, CRN, SJM
Methodology (relational database): SJM
Software (database and supporting algorithms): SJM, CRN
Validation (of compiled and extracted data): CRN, AR
Formal analysis (of extracted data): CRN
Investigation (collection and provision of Hunder trout data): PA*, AR*, OH*, TQ*, LAV
Investigation (sclerochronological analyses of scale samples): PA*, AR*
Resources (materials, samples): PA*, AR*, EL*, OH*, TQ*
Data curation: SJM, CRN, AR*, TQ*, OH*
Writing - original draft: SJM, CRN
Writing - review & editing: CRN, SJM, AR, LAV, EL, OH, TQ, PA
Visualisation: CRN, SJM, EL, AR
Supervision: LAV
Project administration: LAV, SJM, CRN
Funding acquisition: PA* (monitoring programme); LAV, SJM (SUSTAIN project)
The file describes the unpublished relational database (SUSTAIN trout database) from which the two publised datasets are extracted; (1) Hunder trout mark-recapture-recovery dataset and (2) Hunder trout growth dataset.