Life history parameters, and in particular reproductive traits such as spawning cycle, size and age at maturity, fecundity and spawning potential are the basis for assessing the productivity and resilience of fish stocks, making them essential information for stock assessment and management (Dominguez-Petit et al. 2008). The estimation of the size at which 50% of individuals are mature (L₅₀) can be used to define the minimum legal landing size for exploited stocks. However, in most stock assessments the causes of L₅₀ variability are not analysed, though there are intrinsic and extrinsic factors affecting maturity that may act at a temporal scale (Lorenzen and Camp 2019). Size and age at maturity are especially sensitive to population abundance (Dominguez-Petit et al. 2008, Lorenzen and Camp 2019) owing to density-dependent effects. When stock size decreases, individuals tend to mature at large sizes and early ages because of compensatory growth (Ali et al. 2003). However, some environmental and anthropogenic drivers, including high rates of fishing mortality, may also impact on size at maturity (Pepin 2015). Moreover, when the effects of these drivers are intense and/or sustained over a long period of time, they could lead to evolutionary changes because of the selection of certain genotypes (Hollins et al. 2018). Subsequently, understanding mechanisms driving maturity variability within species could provide preliminary evidence of the species’s adaptability to climate change (Dalgleish et al. 2010), habitat degradation (Öckinger et al. 2010) and overfishing (Hobday et al. 2011).

Barot et al. (2004) identified a downward shift in size at maturity in the Georges Bank Atlantic cod (Gadus morhua) that supports the hypothesis that an evolutionary change, likely caused by high fishing mortality, is partially responsible for the observed decline in age and size at maturity in these cod stocks (Olsen et al. 2005). Regardless of the causes of variation in maturation, size and age at maturity can have significant effects on the reproductive performance of the stock by regulating its reproductive potential. For example, fecundity and the quality and viability of eggs and larvae are directly related to the size of spawning females (Trippel et al. 1997, Marteinsdottir and Begg 2002). The extent to which these effects may extend over several generations and affect the adaptive capacity of populations is not yet well understood, although understanding these processes is critical to comprehending stock dynamics.

European hake (Merluccius merluccius) is a resource of great commercial importance in Atlantic Iberian waters. This species is assessed by the International Council for the Exploration of the Sea (ICES) in two units: the northern and the southern stocks (ICES 2021). The southern stock, which is the object of this study, is distributed in the Atlantic Iberian waters that correspond to ICES divisions 27.8.c and 27.9.a. Though this stock has been subjected to a recovery plan since 2006 and the multiannual management plan for Western Waters, fishing mortality is still above that corresponding to the maximum sustainable yield (F_msy) (ICES 2021). For this stock, the main indicator of stock status is the spawning stock biomass, which is calculated as numbers, mean weight and proportion of mature fish. The proportion of mature fish was established for the entire study period (1982-2019) by estimating annual maturity ogives for both sexes combined based on macroscopic observations of hake specimens, although a time-constant female-only maturity ogive was used from 2020 onwards owing to the change of the assessment model. The impact of using different reproductive potential metrics (combined vs only female) was evaluated by Cerviño et al. (2013) for this stock, showing that, although the absolute values change, their impact on the management recommendations is slight because the reference points also change in the same direction.

Several studies have analysed the size at maturity of this stock (Dominguez-Petit et al. 2008, Murua 2010), but generally only for females and never for the two sexes separated. The objective of this study was to provide information on size at maturity for the period 1982-2019 in order to corroborate the decreasing trend of the L₅₀ in this stock for each sex and to determine which factors explain this trend. Understanding which factors influence the size at maturity of the southern European hake stock can contribute to better assessment and management. To this end, we first estimated the size at maturity of the European hake for each sex in Atlantic Iberian waters from 1982 to 2019. Second, the variability of the L₅₀ for each sex was modelled with generalized additive models, considering as explanatory variables environmental factors (Atlantic Multidecadal Oscillation, North Atlantic Oscillation and sea surface temperature), and biological variables (biomass, spawning biomass at length and relative factor condition) to test density-dependent effects and the year to assess the interannual variations.

Understanding biological processes underlying southern hake stock dynamics in Atlantic Iberian waters can help to improve their assessment and management. The overall aim of this study was to assess changes in the size at maturity (L₅₀) of both sexes in the period 1982-2019 and determine the potential drivers of its variation.

It should be mentioned that the data used in this study were limited to the northern part of the southern stock distribution and that, owing to the missing data and the way the data were compiled, these results should be interpreted with caution. However, our results highlight the importance of understanding the factors affecting the balance between the energy invested in reproduction, maintenance and growth and the interannual variability of this trade-off. This is the only way to understand how environmental pressures influence the population dynamics of this species at a regional scale.

The results of this study showed that differentiated long-term patterns were observed for the two sexes, with major long-term declines for male hakes. For males, the size at maturity decreased by 12.9 cm over the study period, from 36.1 cm in 1982 to 23.2 cm, the lowest observed value, in 2019. However, for females two trends in L₅₀ can be distinguished: an increasing trend starting in 1990 and reaching its highest value of 56.2 cm in 1996, followed by a decreasing trend for the rest of the time series, reaching a value of 38.8 cm in 2019.

In relation to Kn, both sexes showed the same structure, that is, a tendency to increase from 1993 onwards, in which year both males and females reached the minimum values of 0.37 and 0.35, respectively. This translates into a decrease in Kn of 39% for males and 68.2% for females with respect to the mean. In contrast, the highest Kn value was 0.87 in 2014 for males and 2.48 in 1998 for females, which is an increase of 42.6% and 125.45% percentage points, respectively. It should be noted that the high Kn value for females in 1998 is due to a bias in the sample, as only 15 mature hakes showed very high weights. This can be explained by the fact that when total weight (heavily influenced by gonad size/maturity stage and stomach size/repletion stage) is used to calculate Kn, these weights can fluctuate markedly because of factors unrelated to the actual physiological condition of the hake specimen.

From the above, it can be deduced that though both sexes follow the same pattern for Kn, male and female maturity responded differently to over time, likely linked to external drivers, including fishing pressure.

The reproductive system of fish reacts to any changes in life conditions, so variations in growth and reproduction dynamics of fish populations substantially affects fish production not only quantitatively but also qualitatively (Godø and Haug 1999, Oven 2004). When fishing pressure removes old and large individuals from the spawning stock, it may favour density-dependent growth, leading to an increase in size at maturity (Olsen et al. 2004); however, if fishing pressure is too intense and maintained for a long time, it can remove the fast-growth or large-size genotypes from the stock, resulting in a decrease of size at maturity and even causing evolutionary changes in the stock (Trippel 1995, Jørgensen et at. 2007, Hidalgo et al. 2012).

In Iberian Atlantic waters, European hake has been exploited beyond safe biological limits since the 1990s, with the largest landings being recorded in this period, but in 2004 it was declared in a critical state because of overfishing (Dominguez-Petit 2007, Korta et al. 2010). Currently, the stock is considered to be in a relatively healthy status but still shows reduced reproductive capacity, according to the annual observed recruitment. It is still intensely exploited (in 2018, F=0.60), exceeding the rate of fishing mortality for the maximum sustainable yield (F_msy= 0.25) but not above the precautionary limits (F_pa=0.75) (ICES 2019). In addition, European hake is the target of mixed fisheries delivered by the industrial and artisanal fleets with different gears: gillnet, longline and different trawl gears (Korta et al. 2015, ICES 2019) targeting medium-large specimens (<40cm). Gillnets and longlines, have accounted for 16% and 13% of catches during the last decade, respectively, targeting hake adults (especially longliners that target large hake). However, Dominguez-Petit (2007) and Korta et. al (2015), reported a decrease in the mean hake size in the landings in the last few decades independently of the gear type used, supporting the perception of overexploitation.

In line with the above, fishing pressure has been widely reported as a selection factor for larger individuals with an effect that is especially problematic when the oldest and largest individuals of the population are removed (Law 2000), because the size/age truncation of the stock leads to a decrease in reproductive potential (Hixon et al. 2014). However, no proxy for fishing effort was considered in our models (for both males and females) owing to the lack of detailed and reliable fleet information (with the same spatial and temporal coverage) that could support this hypothesis. On the contrary, biological (biomass, spawning biomass at length and Kn) and environmental factors (NAO and SST) were considered as drivers of the variations in the decrease of L₅₀ for each sex.

Several studies have addressed the changes in the maturity ogives in different species ( Haug and Tjemsland 1986, Junquera et al. 1999, Marteinsdottir and Begg 2002, Engelhard and Heino 2004). In particular, Dominguez-Petit et al. (2008) studied these changes for European hake, although males were not considered in their study. They observed for the southern stock that the size at maturity increased even when total biomass and spawning biomass continued to decrease, which is contrary to the compensatory theory, probably as a consequence of environmental drivers such as NAO and upwelling. However, their results indicate that the pattern of the size at maturity in the Bay of Biscay (northern stock) was different, as a steady decline was observed throughout the study period.

For male hake, both environmental and biological factors failed to explain the trend towards decreasing L₅₀for our entire study period. This may well be explained by the fact that fishing pressure has eliminated larger individuals (mostly females), which could lead to a clear alteration of growth rates, causing profound changes in L₅₀, either because fishing intensity has caused a regime shift, which has caused the main drivers of maturation/growth to change, or even because there has been a regime shift in the fishery itself, so fishing pressure has lost influence relative to other drivers. Considering the above, fishing pressure and behaviour of the trawl fleet could explain the decline, but further work is required because males and females do not show the same response.

In the case of female hake, the reduction in stock biomass, usually as a consequence of high fishing exploitation, leads to a reduction in size at maturity. The results showed a negative relationship between biomass and size at maturity for female hake, which is to be expected in a density-dependent relationship. The higher the biomass, the greater the intraspecific competition, the lower the growth and the lower the L₅₀. However, the results showed a positive relationship between biomass at spawning and L₅₀, which is contrary to the compensatory theory. The lack of compensatory response is known as depensation. According to Liermann and Hilborn (2001), there are four mechanisms of depensation: i) reduced probability of fertilization, ii) altered group dynamics, iii) environmental influences and iv) predator saturation. These four mechanisms are responsible for the lack of a compensatory response, although population density and intraspecific competition decrease.

Major observed declines in L₅₀ are usually explained by unfavourable environmental conditions that reduce growth or by high fishing pressure (or a combination of both) as a mechanism to maximize fitness (Albo-Puigserver et al. 2021). L₅₀ is expected to be more variable than other traits, particularly in heterogeneous environments (Hidalgo et al. 2014).

The only significant environmental variable included in our model for European hake females was the NAO index, which showed a positive relationship with the L₅₀. The NAO is the most important source of variability in the North Atlantic region. The NAO index acts as an integrator of several environmental factors that can have a synergistic effect on fisheries dynamics. NAO impacts on air temperature, wind and precipitation regime and on large water mass distribution and flow, with the subsequent influence that all these factors have on marine ecosystems (Greene and Pershing 2000 , Köster et al. 2005). Effects of the NAO ripple influence life history traits and population dynamics and do so across trophic levels from primary production to predators (Ottersen et al. 2001). In fact, many examples show that NAO drives the life history and dynamics of many fish populations (Ottersen et al. 2001, Hjermann et al. 2004, Sullivan and Cowen 2005) and in particular of the European hake stocks (Meiners-Mandujano 2007, Dominguez-Petit et al. 2008, Goikoetxea and Irigoien 2013). Similarly, Drinkwater (2005) showed that NAO explains approximately 50% of the variability in the growth increase of northern cod between ages three and five in Newfoundland, which is undoubtedly a determinant of L₅₀, as these are the ages at which cod mature in this area. For female hake something similar may occur: positive NAO values imply higher growth, which implies higher L₅₀. The NAO index may influence the maturation process of hake indirectly through alterations in ecosystem composition and resource availability (Köster et al. 2005, Kell et al. 2005).

The size at maturity of the southern European hake stock has declined in both sexes in the last few decades. The present results show that the maturation of southern hake is influenced by a combination of demographic, anthropogenic and environmental factors, leading to a deeper understanding of the factors explaining the annual decrease in L₅₀ for both sexes, which was the objective of this study.

However, determining which variables explain the decreasing trends in L₅₀for both sexes requires applying more complex models than an overall linear trend to all the data, because there are much more complex relationships between L₅₀and the factors that may explain the variability in L₅₀.

Total biomass, spawning biomass at length and the NAO partially explain the decline in the L₅₀ of females, but for males no relevant factor was found among the analysed ones that could explain this drastic decline. Overall, life history traits that rapidly respond to external changes, such as L₅₀, might be good indicators to anticipate further declines in populations and require close monitoring and evaluation. That is why this study reiterates the need for a posteriori studies to confirm the hypothesis of fishing pressure as a maturity driver in order to have a better understanding of the underlying dynamics of this stock. Additionally, the impact of the reduction in the L₅₀ in males on the reproductive potential of the stock should be also investigated as paternal effects have proved to be more important than previously thought in stock dynamics (Dominguez-Petit et al. 2022).