Body length changes for Atlantic salmon (Salmo salar) over five decades exhibit weak spatial synchrony over a broad latitudinal gradient.

Understanding the factors that drive spatial synchrony among populations or species is important for management and recovery of populations. The range-wide declines in Atlantic salmon (Salmo salar) populations may be the result of broad-scale changes in the marine environment. Salmon undergo rapid growth in the ocean; therefore changing marine conditions may affect body size and fecundity estimates used to evaluate whether stock reference points are met. Using a dataset that spanned five decades, 172,268 individuals, and 19 rivers throughout Eastern Canada, we investigated the occurrence of spatial synchrony in changes in the body size of returning wild adult Atlantic salmon. Body size was then related to conditions in the marine environment (i.e., climate indices, thermal habitat availability, food availability, density-dependence, and fisheries exploitation rates) that may act on all populations during the ocean feeding phase of their life cycle. Body size increased during the 1980s and 1990s for salmon that returned to rivers after one (1SW) or two winters at sea (2SW); however, significant changes were only observed for 1SW and/or 2SW in some mid-latitude and northern rivers (10/13 rivers with 10 of more years of data during these decades) and not in southern rivers (0/2), suggesting weak spatial synchrony across Eastern Canada. For 1SW salmon in nine rivers, body size was longer when fisheries exploitation rates were lower. For 2SW salmon, body size was longer when suitable thermal habitat was more abundant (significant for 3/8 rivers) and the Atlantic Multidecadal Oscillation was higher (i.e., warmer sea surface temperatures; significant for 4/8 rivers). Overall, the weak spatial synchrony and variable effects of covariates on body size across rivers suggest that changes in Atlantic salmon body size may not be solely driven by shared conditions in the marine environment. Regardless, body size changes may have consequences for population management and recovery through the relationship between size and fecundity.

Coherent long-term body-size responses across all Northwest Atlantic herring populations to warming and environmental change despite contrasting harvest and ecological factors.

Body size is a key component of individual fitness and an important factor in the structure and functioning of populations and ecosystems. Disentangling the effects of environmental change, harvest and intra- and inter-specific trophic effects on body size remains challenging for populations in the wild. Herring in the Northwest Atlantic provide a strong basis for evaluating hypotheses related to these drivers given that they have experienced significant warming and harvest over the past century, while also having been exposed to a wide range of other selective constraints across their range. Using data on mean length-at-age 4 for the sixteen principal populations over a period of 53 cohorts (1962-2014), we fitted a series of empirical models for temporal and between-population variation in the response to changes in sea surface temperature. We find evidence for a unified cross-population response in the form of a parabolic function according to which populations in naturally warmer environments have responded more negatively to increasing temperature compared with those in colder locations. Temporal variation in residuals from this function was highly coherent among populations, further suggesting a common response to a large-scale environmental driver. The synchrony observed in this study system, despite strong differences in harvest and ecological histories among populations and over time, clearly indicates a dominant role of environmental change on size-at-age in wild populations, in contrast to commonly reported effects of fishing. This finding has important implications for the management of fisheries as it indicates that a key trait associated with population productivity may be under considerably less short-term management control than currently assumed. Our study, overall, illustrates the need for a comparative approach within species for inferences concerning the many possible effects on body size of natural and anthropogenic drivers in the wild.

Declines in body size of sockeye salmon associated with increased competition in the ocean.

Declining body sizes have been documented for several species of Pacific salmon; however, whether size declines are caused mainly by ocean warming or other ecological factors, and whether they result primarily from trends in age at maturation or changing growth rates remain poorly understood. We quantified changes in mean body size and contributions from shifting size-at-age and age structure of mature sockeye salmon returning to Bristol Bay, Alaska, over the past 60 years. Mean length declined by 3%, corresponding to a 10% decline in mean body mass, since the early 1960s, though much of this decline occurred since the early 2000s. Changes in size-at-age were the dominant cause of body size declines and were more consistent than trends in age structure among the major rivers that flow into Bristol Bay. Annual variation in size-at-age was largely explained by competition among Bristol Bay sockeye salmon and interspecific competition with other salmon in the North Pacific Ocean. Warm winters were associated with better growth of sockeye salmon, whereas warm summers were associated with reduced growth. Our findings point to competition at sea as the main driver of sockeye salmon size declines, and emphasize the trade-off between fish abundance and body size.

Spatial and ontogenetic variations in sardine feeding conditions in the Bay of Biscay through fatty acid composition.

Food characteristics are amongst the most influential factors determining the fish life history traits as quantitative and qualitative changes in individuals' diet can lead to a decline in the energy allocated to their growth, and hence influence natural populations' characteristics. The size-at-age and weight of European sardines (Sardina pilchardus) in the Bay of Biscay (BoB) have decreased substantially over the last decade, especially for the youngest age classes, and the factors underlying such changes have not yet been identified. We therefore analysed the fatty acid (FA) composition in the neutral (NL) and polar (PL) lipids in samples collected across the BoB to determine whether the diet of sardines changes with their ages. We found that the total FA contents in both lipid fractions varied mainly with the sampling locations and age. Indeed, sardines aged 1 and 2 years living in South BoB had particularly high contents in FA specific to non-diatom phytoplankton, while older sardines living in the Northern part had higher total FA content and more FA specific to copepods. These differences probably resulted from differences in prey availability and to a lesser extend a change in feeding behaviour with age. The strong dependence of younger sardines' diet to phytoplankton in spring suggests that changes in primary production may explain their decline in size-at-age. Finally, NL clearly reflect finest feeding variations in comparison to PL imprinted by diet variations at longer time scale. Future studies should consider separately NL and PL fractions.

A three-pronged approach that leans on Indigenous knowledge for northern fish monitoring and conservation.

Investigating whether changes within fish populations may result from harvesting requires a comprehensive approach, especially in more data-sparse northern regions. Our study took a three-pronged approach to investigate walleye population change by combining Indigenous knowledge (IK), phenotypic traits, and genomics. We thank Larson et al. (2020) for their critiques of our study; certainly, there are aspects of their critique that are warranted and merit further investigation. However, we argue that their critique is over-stated and misleading, primarily given that (a) one of three prongs of our research, IK, was dismissed in their assessment of our study's conclusions; (b) our Bayesian size-at-age modeling should help to mitigate sample size issues; (c) their re-analysis of our size-at-age data does not actually refute our results; (d) genomic changes that we observed are nascent; (e) the data file that Larson et al. (2020) used for their genomic re-analysis was not correct; and (f) criteria that Larson et al. (2020) use for their genomic re-analysis were not properly justified.

Size reductions and genomic changes within two generations in wild walleye populations: associated with harvest?

The extent and rate of harvest-induced genetic changes in natural populations may impact population productivity, recovery, and persistence. While there is substantial evidence for phenotypic changes in harvested fishes, knowledge of genetic change in the wild remains limited, as phenotypic and genetic data are seldom considered in tandem, and the number of generations needed for genetic changes to occur is not well understood. We quantified changes in size-at-age, sex-specific changes in body size, and genomic metrics in three harvested walleye (Sander vitreus) populations and a fourth reference population with low harvest levels over a 15-year period in Mistassini Lake, Quebec. We also collected Indigenous knowledge (IK) surrounding concerns about these populations over time. Using ~9,000 SNPs, genomic metrics included changes in population structure, neutral genomic diversity, effective population size, and signatures of selection. Indigenous knowledge revealed overall reductions in body size and number of fish caught. Smaller body size, a small reduction in size-at-age, nascent changes to population structure (population differentiation within one river and homogenization between two others), and signatures of selection between historical and contemporary samples reflected coupled phenotypic and genomic change in the three harvested populations in both sexes, while no change occurred in the reference population. Sex-specific analyses revealed differences in both body size and genomic metrics but were inconclusive about whether one sex was disproportionately affected. Although alternative explanations cannot be ruled out, our collective results are consistent with the hypothesis that genetic changes associated with harvesting may arise within 1-2.5 generations in long-lived wild fishes. This study thus demonstrates the need to investigate concerns about harvest-induced evolution quickly once they have been raised.

Spatial, interannual, and generational sources of trait variability in a marine population.

Life-history traits of individuals in marine populations exhibit large sources of variability. In marine fish, variation of individual size at a given age has three main components: (1) spatial, correlated with the location in which individuals are caught, (2) temporal, correlated with the time when individuals are caught, and (3) generational, correlated with the year of birth of the examined individuals. These variations, if present, have practical implications for individual fitness as well as for sampling, survey design, and population assessment. Disentangling these variations and understanding their sources is hard, given the potentially correlated nature of their effects on individual traits. This study examines the size-at-age relationship of the Bering Sea Pacific cod, an economically and ecologically important groundfish. We used extensive records spanning 1994 to 2016 (inclusive) of 25,213 observations of both environmental variables and catch, lengths, and ages. We found that the average size of individuals of the same age could differ up to 7 cm. Notably, we found that the cohort composition of the sampled population explained >75% of the year effect and that individuals caught in the northwest and shallower portion of the sampling area were on average 5 cm smaller than individuals caught in the southern and deeper portion. We further found that northwest movement of young cod (age 1-5) as a result of warming places individuals in areas where we predict them to have smaller size at age. Smaller and less conditioned individuals are less fecund and may not be able to perform long migrations to return to their distant spawning grounds. Both the spatial distribution and water temperature experienced by Pacific cod in the Bering Sea are changing, and this study provides a mechanism for how these changes affect Pacific cod life-history traits and individual fitness.

Age determination in echinoderms: first evidence of annual growth rings in holothuroids.

While age is fundamental in animal biology, forming the basis of critical concepts such as life-history strategies, longevity and population structures, measuring this variable in some taxa remains problematic. Such is the case of holothuroid echinoderms, which play key roles in marine benthic communities from the shore to the abyss, and which are extensively fished in many regions across the globe. Here, we present and validate a promising ageing technique using the cold-water species Psolus fabricii. The method involves the extraction of the oldest dermal plates (largest dorsal ossicles) to preserve their original pigments and structure. While plates initially appear to have a uniform texture, polishing and dying them reveals layered ring patterns. A study of laboratory-reared juveniles, from settlement to 40 months of age, confirmed that one layer is added annually, making plates both larger and thicker, and generating successive light and dark rings, the latter representing the transition (overlap) between two layers. Therefore, each pair of rings represents an annual growth band. Size-at-age data obtained using this method revealed that growth of P. fabricii is slow and that wild individuals collected at diving depths had reached an age of several decades.

Growth and otolith morphology vary with alternative reproductive tactics and contaminant exposure in the round goby Neogobius melanostomus.

Round goby Neogobius melanostomus sagittal (saccular) otolith morphology was compared between males of the two alternative reproductive tactics (termed guarder and sneaker males) and between males captured from sites of high or low contamination. Otolith size increased with fish size and also displayed an ontogenetic shift in shape, becoming relatively taller as otoliths grew in size. Despite a considerable overlap in age between males adopting the two reproductive tactics, size-at-age measurements revealed that guarder males are significantly larger than sneakers at any given age and that they invest more into somatic growth than sneaker males. Controlling for body size, sneaker males possessed heavier sagittal otoliths than guarder males. Subtle otolith shape differences were also found between the two male tactics and between sites of high and low contaminant exposure. Sneaker males had relatively shorter otoliths with more pronounced notching than guarder males. Fish captured at sites of high contamination had otoliths showing slower growth rates in relation to body size and their shapes had more pronounced caudal points and ventral protrusions when compared with fish captured at sites of low contamination. The results are discussed in relation to life-history tradeoffs between the male tactics in terms of reproductive and somatic investment as well as the putative metabolic costs of exposure to contaminants. Overall, this study reveals that male alternative reproductive tactics and environmental contaminants can have small, yet measurable, effects on otolith morphology and these factors should be accounted for in future research when possible.

Life-history implications of the allometric scaling of growth.

Several phenomenological descriptions, such as the von Bertalanffy growth model, have been widely used to describe size-at-age and individual growth across a diverse range of organisms. However, for modelling life histories, as opposed to just growth, biologically and mechanistically meaningful growth models, based on allocation decisions, have become increasingly important. This is because fitness is determined by survival and reproduction, which are not addressed directly in phenomenological growth models. To elucidate these considerations, we take as a starting point the biphasic growth model by Quince et al. (2008a, J. Theor. Biol. 254:197) which has the advantage that the underlying allometric scaling of net energy intake can be freely chosen. First, we reformulate this model such that individual size is given in meaningful units of length and weight, facilitating the model׳s interpretation and application. Second, we show that even though different allometric scaling relationships can produce practically identical growth trajectories, the accompanying reproductive investments are highly dependent on the chosen allometric exponent. Third, we demonstrate how this dependence has dramatic consequences for evolutionary predictions, in particular with regard to the age and size at maturation. These findings have considerable practical relevance, because empirically observed allometric exponents are often uncertain and systematically differ from those assumed in current standard growth models.

Shrinking fish: comparisons of prehistoric and contemporary salmonids indicate decreasing size at age across millennia.

A comparison of Upper Palaeolithic and contemporary salmonid vertebrae from the Iberian Peninsula indicates that there has been a significant decrease in the mean body size for a given age among Atlantic salmon and brown trout inhabiting the southernmost range of their endemic distribution. Mean size at age was greater in prehistoric specimens for all age classes during the freshwater phase of their life histories. Fisheries-induced evolution (selection for smaller sizes) is an obvious explanation for the observed reduction in fish body size, but recent changes in the aquatic habitat affecting density-dependent growth cannot be ruled out.