4.1 Fishing-induced life-history variation
Overall, this study provides evidence of serious fishing-induced life-history variation in L. crocea population and represent a glimpse of fishery collapse. The observed life-history parameters show that body size of L. crocea have on average, decreased during the last five decades (table 1), whereas we would expect an overall decrease of body size if the heavily fishing was the dominated driver of body size changes and STE. China’s marine resources have suffered serious overfishing issues, characterized by the progressive depletion of economic stocks (Cao et al. 2017). Specifically, indiscriminate fishing (~50% trawl with 54 mm mesh size) in China can cause both age or size truncation of target populations, leading to increasing instability and vulnerability of the population dynamic (Kirby 2004, Gaichas et al. 2014, Kuparinen et al. 2016). Hence, the extremely high fishing mortality and exploitation rate have caused systematic decline in landing catch. Previous studies in the 1970s and 1980s revealed that the main catch of L. crocea consisted of two or three years old (400 – 800g) individuals (Yu and Lin 1980), while 95% of catches in this study were individuals aged zero or one years old. Another major finding was that ~70% of potential catch area have disappeared (Supporting information Fig. S2), with the highest disappearance rate in offshore areas (122°E – 125°E), which follows the life-history variation of L. crocea during the last five decades.
Previous studies have identified continuous fishing pressures and subsequent fishing-induced life-history variation is likely to have negative effects on overall population structure and their distribution patterns in several ways. First, continuous fishing pressure can erode fish biomass by substantially decreasing the proportion of large, old individuals. Still, STE-caused decline in maturity size and increase in mortality, may reduce reproductive output or the size of the breeding population (Johnston and Temple 2002, Morita et al. 2005, McMahan et al. 2020), and subsequent cause irreversible changes such as recruitment bottleneck, even extinction, to some populations, such as the collapse of Eastern blue groper, gemfish and blue-eye trevalla populations, following intense commercial fishing in Australia (Last et al. 2011). Second, a decrease in the population size can also reduce variation (causing inbreeding and genetic abnormalities) and, therefore, restrict the potential for adaption (Langangen et al. 2019). Thirdly, fishing-induced life-history variation may constrain species distribution because the migratory ability of a species is strongly dependent on dispersal characteristics, such as morphological traits, like body-size (Hsieh et al. 2009). Consequently, constriction of geographic distribution, associated with a decline in body size, may reduce the ability to respond to climatic stress, by limiting movement (Reusch et al. 2005).
4.2 Climate-induced overwintering habitat degradation may intensify the effect of overfishing
The best model (next best model: ∆AIC=2, for other models, Supporting information Table S2) to explain catch patterns under unexploited status over time included SST and depth. The result of SI suggested that the optimal overwintering temperature range (18.2 – 20.5℃) and depth (36 – 72m) of L. crocea mirror previous lab-based observation of optimal growth temperature (17℃– 24℃) and empirical observation of optimal overwintering depth (50 – 60 m) (Xu and Chen 2011, Liu 2013). It is worth noting that SST in the mid-southern ECS have decreased by an average of 1℃ between the 1980s and 2010s with annual decrease SST rate – 0.028℃/year (Fig. 4). HSI variation results in the last five decades suggests that cooling trend of SST in the ECS has significantly reduced the proportion of optimal and common overwintering habitats for L. crocea (Fig. 5B).
Consequently, in response to the SST decrease in winter, migratory species, like L. crocea are expected to respond in two ways as follows. Generally, marine organisms respond to climate change through shifts in distribution (Guisan and Thuiller 2005). For instances, in the North Sea, both exploited and un-exploited fish species have shifted to higher latitudes and deeper water between 1977 and 2001, in response to rising sea temperatures (Perry et al. 2005); in the Eastern Tropical Pacific, demersal species were projected to move into shallow water by the mid-21st century, in response to high greenhouse gas emissions (Representative Concentration Pathways, RCP8.5) and strong migration (RCP2.6) scenarios (Clarke et al. 2020). Alternatively, it is also commonly observed that species stay in poor habitat against climate change but suffered climate-induced life-history variation. Particularly, it occurs when marine fishes are living in temperature outside their physiology optima: this results in reduced aerobic scope, which negatively affects their growth and reproduction (Pearson and Dawson 2003, Pörtner and Knust 2007, Toresen et al. 2019). Hence, species like L. crocea must either migrate to remain within suitable habitat or suffer the consequences (Bell et al. 2015).
Interesting, L. crocea overwintering distribution pattern did not shift alongside a decrease in winter SST, which is a good indicator that temperature per se did not explain the overall shift of L. croceadistribution. Such absence of a clear systematic impact of temperature may be due to the life-history parameters degradation, which could constrain the hedging capacity against climate change (Thorson et al. 2017). It is not uncommon that that heavily exploited species that do not ‘fill’ their suitable habitat against climate change background. For example, STE-caused a change in the length-age structure is the main diver of inter-annual shifts in summer flounder distribution, while temperature had little influence on the change in distribution (Bell et al. 2015). More broadly, the ‘match/mismatch hypotheses’ may explain the combined effects of heavily fishing and climate change on decrease of overall population (Cushing 1990, Edwards and Richardson 2004). Based on our study, we suggest that fishing-induced life-history variation leading to the ‘mismatch’ of L. crocea optimal overwintering habitat. Firstly, our study demonstrated that L. crocea have both life-history degradation and size truncation compared with 1980s, with significantly smaller body size and advanced maturation (Fig. 2B). This truncation in overall size-structure can significantly affect swimming ability, such as reducing the sustained swimming time and average swimming speed, namely size dependent, consequently reducing the distribution range of L. crocea (Jorgensen et al. 2008, Opdal and Jørgensen 2015, 2016). Ultimately, given the climate-induced changes in overwintering habitat suitability that occurs in the mid-southern ECS, the fishing-induced life-history population variation that constrains dispersal capability could pose a significant ‘mismatch’ of optimal overwintering habitat to L. crocea -like migratory species (Fig. 6). Such applicability of ‘mismatch hypotheses’ to the specific long-lived migratory fish exposed to fishing and climate change had rarely been demonstrated.