Disentangling the Effect of Adult Biomass and Temperature on the Recruitment Dynamics of Fishes

Here we analyzed the relation between recruitment dynamic (recruitment or recruitment success) and temperature of 57 commercial fish stocks of the North Atlantic. We showed that, although the effect of temperature was generally significant, spawning biomass is the main factor governing recruitment dynamic. A significant effect of spawning biomass was evident for 67% of all stocks analyzed in this study. For gadoids, the effect of spawning biomass was larger than the effect of temperature (both in terms of number of stocks and proportion of variance explained). For clupeids, spawning biomass was more important than temperature for a higher number of stocks, but the strength of the two effects, when present, was similar. Also, stocks living in colder and warmer areas showed, respectively, a positive and negative response to temperature for both families. Our results highlight that failing to account for spawning biomass effect in climate-recruitment studies could mask the influence of climate variability on recruitment dynamic. In this context, although management of several exploited fish populations cannot be entirely decoupled from the effect of climate on stock reproductive success, it is likely that the observed changes in exploited fish population dynamics are mainly the consequences of an unsustainable human impact and not climate changes. Introduction The physical environment has been long recognized as an essential factor in the regulation of fish population in marine ecosystems (Cushing 222 Cardinale et al.—Effect of Adult Biomass and Temperature 1982, Steele 1985). Among environmental factors, temperature is often the principal variable investigated since it regulates the rate in many ecological and physiological processes (Brett 1979). Therefore, there are strong justifications for studying the link between fish recruitment and temperature, and this information may be crucial for the management of marine resources and to predict the long-term consequence of climate changes on fisheries (Myers 1998). It is now recognized that the twenty-first century will show a significant global warming trend induced by an increase in atmospheric greenhouse effect (Houghton et al. 2001). This will also imply that a global change in the oceanic climate may propagate to the upper level of the food webs (Sarmineto et al. 1998) and consequently affect important marine resources (i.e., Loukos et al. 2003, Drinkwater 2005). Therefore, there is an increasing interest on the relation between climate and fisheries, both at the ecological and economic level (Bopp et al. 2001, Eide and Heen 2002), and evidence suggests that there is a direct link between climate and fish population dynamics (Marshall et al. 2001, Bakun and Broad 2002). In this context, life history traits might constitute a powerful means to unravel the interaction between climate and fish population. Among those, the number of recruits per unit of spawning stock biomass (SSB), i.e., the stock recruitment success (Rs), is one of the most important parameter proxies for population productivity (MacKenzie et al. 2003, Stige et al. 2006). In this study we explored recruitment data in relation to spawning stock biomass and temperature for exploited fish stocks in the North Atlantic—57 selected demersal gadoids (cod, Gadus morhua; haddock, Melanogrammus aeglefinus; saithe, Pollachius virens; and whiting, Merlangius merlangus) and pelagic clupeids (herring, Clupea harengus; and sprat, Sprattus sprattus). Common patterns in life histories and population dynamics are evident only after data from several stocks, species, and areas are combined and standardized in ways that enable fair and unbiased comparisons (Myers 1998, MacKenzie et al. 2003). First, we revised available information to disentangle the effect of spawning stock biomass and temperature on Rs for the selected stocks. Successively, we investigated the potential differences in the strength and sign of these effects between families (gadoids and clupeids) and areas. We deliberately limited the analysis to SSB and temperature, although we were aware that other abiotic factors (i.e., precipitation, wind direction, currents, ice cover, etc.) and biotic factors (i.e., predation, competition, stock structure, food availability, fecundity, condition of the spawners, etc.) play an important role on recruitment processes. 223 Resiliency of Gadid Stocks to Fishing and Climate Change Materials and methods Time series Stock and recruitment data We compiled data of spawning stock biomass (SSB) and number of recruits (R) available at the International Council for the Exploration of the Sea (ICES), www.ices.dk, at the Northwest Atlantic Fisheries Organization (NAFO), www.nafo.ca, and the Ransom Myers Web site, www.mscs.dal.ca/~myers/welcome.html (Table 1). For geographical distribution the reader can refer to www.ices.dk and www.nafo.ca. All estimates are derived from VPA (virtual population analysis) models and thus do not include any assumptions on the form of stock-recruitment curve. Temperature data The temperature data appropriate for comparisons of North Atlantic stocks should fulfill several requirements (Planque and Frédou 1999): they must be available for every stock, be consistent and sufficiently long, and correspond to large oceanic areas to be representative of the change in temperature experienced by the stock over its area of distribution. It is well known (see discussion in Ottersen et al. 1998) that environmental temperature does not strictly correspond to “ambient” temperature, although the population generally experiences higher temperature in warm years and vice versa (Planque and Frédou 1999). At the same time, it is likely that sea surface temperature (SST ) reflects more closely ambient temperature of pelagic fish than demersal species. Anyhow, to accomplish all listed criteria, we used sea surface temperature data from the International Comprehensive Ocean Atmosphere Dataset (ICOADS) available at www.cdc.noaa.gov/coads/ coads_cdc_netcdf.shtml. Monthly average at the spatial resolution of 2 degrees of latitude and 2 degrees of longitude were available for all stocks from 1800. For each stock, we calculated the annual average (SST ) and the average temperature of the time series at the geographical centroid of the stock distribution (SST-A) (available at www.mscs. dal.ca/~myers/welcome.html) (Table 1). The latter was used as a proxy for the mean ambient temperature occurring in the area of distribution of each stock. Statistical analysis To allow for stock comparisons, we chose to estimate R as the number of age 1 individual. For those stocks for which the numbers of age 1 individuals were not available, the logistic equation (Hilborn and Walters 1992) was applied using values of natural and fishing mortality estimated for the adjacent age class. In order to make comparable spring and autumn herring spawners, the number of 1-year-old individuals of 224 Cardinale et al.—Effect of Adult Biomass and Temperature Acronym Fishing area Latitude Longitude Area SST-A Age of recruits (year) Years CA I&II 68 15 NEA 7.6 3 1946-2003 CEB IIId 56 19 NEA 8.8 2 1966-2003 CF Vb 62 –8 NEA 11.0 2 1961-2003 CI Va 63 –21 NEA 8.6 3 1955-2003 CK IIIa 57 12 NEA 9.0 1 1971-2003 CNS IV 55 4 NEA 9.9 1 1963-2003 CWB IIIc&d 55 13.5 NEA 8.8 1 1970-2003 CVIA VIa 58.5 –6 NEA 10.8 1 1966-2003 CVIIA VIIa 54 –5 NEA 10.8