Introduction

Despite their remarkable ability to adapt to all kinds of environments, the European eel, Anguilla anguilla, population has been in decline, at least since the 1960s (ICES, 2019; Dekker & Beaulaton, 2016). Recruitment to freshwater habitats decreased by more than 90% in the early 1980s and since 2008 the European eel has been listed as Critically Endangered on the International Union for Conservation of Nature (IUCN) red list (Jacoby & Gollock, 2014). Causes of the decline are related to habitat loss, overfishing, climate change, pollution, parasites and diseases (Aschonitis et al., 2017; Drouineau et al., 2018).
The European eel is semelparous and panmictic (Als et al., 2011). It spawns in the Sargasso Sea, but is distributed across Europe, from northern Norway to northern Africa and far into the Mediterranean (Schmidt & Regan, 1923; Dekker, 2003a). Larvae drifting with the Gulf Stream metamorphose into glass eels when they reach the continental shelf. These glass eels colonize coastal and freshwater habitats where they spend their growth phase until they start to mature into silver eels which will migrate back to the Sargasso Sea for spawning (Bertin, 1956; Tesch, 2003).
The status of the stock is primarily assessed through time series of recruiting glass eels (or elvers) to freshwater at different monitoring stations across Europe (ICES, 2000). A severe reduction in glass eel recruitment, more marked in the northern part of the distribution area, became apparent in the early 1980s (Moriarty, 1986,1990; Dekker, 2003b, 2004; ICES, 2016; Bornarel et al., 2017). However, signs of decrease in the standing stock (yellow stage) date from the 1960s (Aalto et al., 2016; Dekker, 2003c).
At some of the monitoring stations, like in the river Imsa, Norway, both upstream ascending elvers and downstream migrating silver eels are trapped and counted (Sandlund et al., 2017). The Imsa reference time-series was started in the 1970s, and the age distribution and the population dynamics in this catchment were especially studied in the 1980s and 1990s, yielding fundamental knowledge on the ecology of European eels (Vøllestad et al., 1986; Vøllestad & Jonsson, 1986, 1988; Vøllestad, Jonsson, Hvidsten, & Næsje, 1994). Following the awareness of the European eel population crash, countries across Europe developed management plans in accordance with a European Union Regulation (EC 1100/2007). Although a non-EU country, Norway followed in 2010, with a ban on eel fishing and the Imsa time-series became even more relevant to monitor the local sub-stock (Poole et al. 2018).
Knowledge on age structure is essential to assess the status of a population. Determination of age in fish is challenging, especially in long-lived species such as the European eel. For this species, otoliths (or earstones) are prepared so that annual rings (annuli), marking periods of fast and slow growth, become visible and can be counted to give an age estimate (Moriarty, 1973, 1983, Svedäng, Wickström, Reizenstein, Holmgren, & Florenius, 1998). Four main methods have been used for preparing otoliths: 1) Grinding and polishing, 2) slicing, 3) burning and cracking, and 4) clearing of whole otoliths in ethanol (in toto method). It has been debated what is the better method (Vøllestad & Næsje, 1988; ICES, 2009). Different methods yield different estimates (Moriarty & Steinmetz 1979). The suitability of each method depends on the age and growth of individuals (Vøllestad, Lecomte-Finiger, & Steinmetz, 1988). For example, burning and cracking is more suitable for slow-growing eels, because of the shape of their otolith and the numerous but short growth increments (Vøllestad & Næsje, 1988). The “in toto” method (clearing whole otoliths) is fast and inexpensive, but best suited for young eels (Vøllestad and Næsje, 1988). During recent years, a manual regarding the best practice for ageing eels has been developed (ICES, 2009, 2011). However, methods are still debated due to a general lack of validation and of different growth patterns in such a widespread species.
Once annuli are revealed, interpretation remains challenging. For example, traumatic events, such as high temperatures in the summer, diseases, or stress can cause supernumerary check or “false” checks (Domingos, Costa, & Costa, 2006; Graynoth, 1999; ICES, 2009; Tzeng, Wu, & Wickström, 1994; Svedäng et al., 1998). In addition, observations can vary between readers, over time and between laboratories.
Age estimates have been validated in some cases using chemical marking of otoliths (Chrisnall & Kalish 1993; Dekker, 1986; Oliveira, 1996; Svedäng et al., 1998), external color marking (Chisnall & Kalish 1993; Poole & Reynolds 1996) or indirect methods using individual mark-recapture techniques (Poole & Reynolds 1996; Beentjes & Jellyman, 2015), or by introducing eels in a pristine waterbody (Vøllestad & Næsje 1988; Wickström, Westin, Clevestam, 1996; ICES, 2009). However, validations for the European eel have been done mostly on young individuals with a maximum age of 14 years and unfortunately, there are still too few examples of validation of ageing methods in eels.
In the earlier years (1980-90s), the age of silver eels from the river Imsa was determined using the in toto method (IT), and it was later suspected that these ages were underestimated. Since the otoliths had been maintained, it was decided to reanalyze them using the consensus method: grinding and polishing method (GP) and to compare both age estimates. Additionally, new otolith samples were collected in the 2010s and treated in the same way (GP method) to investigate possible changes in their age distribution of silver eels since the decline of the population (Poole et al., 2018). Finally, using this new dataset, the possible relationship between the number of ascending recruits and the number of descending silver eels was examined to reanalyze previous models (Vøllestad & Jonsson 1988) established for the Imsa eel stock during the period 1975-1987.