Forecasting Pacific saury (Cololabis saira) fishing grounds off Japan using a migration model driven by an ocean circulation model
Introduction
Pacific saury Cololabis saira is a small fish that is widely distributed throughout subtropical to subarctic regions of the North Pacific (Hubbs and Wisner, 1980). This species, caught primarily by stick-held dip nets, is of commercial importance to Japan, Russia, Korea, Taiwan, and China. Whereas both Japan and Russia fish for this species mainly within their exclusive economic zones (EEZs), other countries exploit it within the North Pacific high seas (Huang et al., 2007; Tseng et al., 2014).
The life span of Pacific saury is short, approximately 2 years (Suyama et al., 2006), with fish of 1 year age being the main fishing target. Its total catch has varied greatly, from 18.1 × 104 tons in 1998 to 63.1 × 104 tons in 2014 (http://www.fao.org/fishery/species/3001/en). Until 2000, Japan caught more than 70% of the global catch, but after 2010, with catch in other countries increasing, Japan's proportional take of the total catch decreased to less than 50%.
During their short life cycle Pacific saury migrate over a wide area (Fig. 1). Their primary spawning ground is in the Kuroshio region in the winter (Watanabe and Lo, 1989). From here hatched larvae and juveniles are transported east along the Kuroshio Extension (Oozeki et al., 2015), before migrating north from subtropical regions into subarctic waters from May to July for feeding (Fukushima, 1979). After sufficient growth, they commence a return migration towards subtropical spawning grounds in August or September (Ito et al., 2004; Tseng et al., 2014), during which time the Japanese Pacific saury fishery starts.
Early fishing grounds for Japanese vessels occur mainly along the Kuril Islands archipelago or around the southeast coast of Hokkaido, depending on the distribution of the Oyashio Current (Fig. 1. Fukushima, 1979; Yasuda and Kitagawa, 1996). It is known that the Oyashio forms two tongue-shaped intrusions, the nearshore Oyashio First (O1) and the offshore Oyashio Second (O2) intrusion (Sekine, 1988). Because fishing grounds around O1 are closer to Japanese fishing ports than those of O2, fishing there reduces vessel fuel costs and leads to the landing of fresher fish.
Interannual geographical shifts in the offshore Oyashio front correspond with those of Pacific saury fishing grounds (Yasuda and Watanabe, 1994). Previous studies have suggested that the spatial and temporal distributions and migration routes of Pacific saury are linked to environmental conditions (Fukushima, 1979; Tian et al., 2003; Watanabe et al., 2006; Yasuda and Watanabe 1994; Yasuda and Kitagawa, 1996; Tseng et al., 2014). Accordingly, the areas where early fishing grounds form off Japan vary from year to year. The primary factor affecting the horizontal distribution of these fish is considered to be temperature (Watanabe et al., 1999, 2006; Tian et al., 2003; Tseng et al., 2014), with 15 °C regarded as preferred (Huang et al., 2007). In addition, food density is also important for growth and migration (Ito et al., 2004; Oozeki et al., 2004) and likely influences the horizontal distribution of these fish.
Since 2003 the Tohoku National Fisheries Research Institute (TNFRI) has conducted stock assessment surveys using surface trawls from June to July (pre-fishing season) between 143°E and 165°W (TNFRI, 2018). Pre-fishing season nearshore Pacific saury abundance is low because fish are mainly distributed east of 150°E (Miyamoto et al., 2019). Fish later migrate west to recruit into Japanese nearshore waters during the main September–November fishing season. A similar western movement is found in Taiwanese fishery data (Tseng et al., 2014). Because Taiwanese fishing vessels mainly operate in the high seas, post-August fishing grounds tend to occur along the Russian EEZ boundary (from 42.4°N, 151°E to 41.5°N, 165°E, black line in Fig. 1). Therefore, migration characteristics discerned from territorial fisheries statistics represent an incomplete picture of the entire life history of this species. TNFRI stock-assessment surveys also reveal fish abundance in waters west of 162°E to have decreased drastically after 2010, which resulted in poor early fishing season catches and delays in formation of early nearshore fishing grounds in Japan (TNFRI, 2018).
Most Pacific saury fishermen concern themselves with the time and location of early fishery grounds, because initial seasonal products trade at higher prices. To this end, predictions as to when and where early fishing grounds will form are commercially important, and are provided at https://www.jfa.maff.go.jp/j/press/sigen/190731.html. Forecasts have historically been made qualitatively based on empirical knowledge—when pre-fishing season Pacific saury occurred nearer to the Japanese coast, Pacific saury reached the main Japanese fishing area earlier (Suyama et al., 2012), and vice versa, with fishing grounds formed along the Oyashio (Fukushima, 1979; Yasuda and Watanabe, 1994; Yasuda and Kitagawa, 1996).
Until ocean circulation models were developed, the future Oyashio distribution was determined using a persistence forecast method based on observed hydrographic conditions. That is, the observed Oyashio distribution was assumed to be maintained for 1–2 months. This resulted in the forecast tending to fail when the Oyashio distribution changed abruptly. Since 2012 the Oyashio distribution has been forecast using FRA-ROMS, a Regional Ocean Modeling System (ROMS) developed by the Japan Fisheries Research and Education Agency (FRA) (Kuroda et al., 2017). By using the forecast based on physical processes, the forecast accuracy of the Oyashio distribution has improved. Forecasts of the timing of the Pacific saury fishing ground formation in nearshore Japanese waters had been conducted empirically based on the pre-fishing fish distribution until Miyamoto et al. (2019) proposed a new method based on the radius of otolith annual rings. Using this method, we became able to forecast migration dates. However, this method enable us to determine migration dates regardless of oceanic conditions that vary every year. The ability to forecast both the time and location of early fishing grounds taking oceanographic conditions into consideration would represent a considerable improvement on existing forecast methods.
Recent numerical models have simulated the migration of various fish species, such as sardine and herring (e.g., Megrey et al., 2007; Okunishi et al., 2009; Moriarty et al., 2016). Pacific saury transport processes during their early life stages have also been examined with particle tracking models using observed current velocity fields (Iwahashi et al., 2006; Oozeki et al., 2015). For adult Pacific saury, mechanisms controlling growth were examined using a bioenergetics model NEMURO.FISH (North Pacific Ecosystem Model for Understanding Regional Oceanography for Including Saury and Herring) (Ito et al., 2007; Kishi et al., 2007; Mukai et al., 2007), accounting for growth of Pacific saury throughout their life history (Ito et al., 2004), for which the influence of ocean warming on fish migration patterns is discussed (Ito et al., 2013). In this model, Pacific saury migrate among three meridionally distributed oceanographic domains, in which the temperature of each is assumed to be constant spatially but to change over time, with spatial differences in temperature considered to represent differences among domains. This model does not take zonal temperature distributions into consideration, and Pacific saury migrate south during their spawning migration. Models including the western migration of fish in two-dimensional horizontal temperature and current velocity fields would improve forecasts of early fishing ground times and locations.
To develop a migration model, knowledge about migration characteristics of a target fish is necessary. For Pacific saury, while approximate migration patterns can be estimated (Fig. 1), detailed trajectories of migration routes, and the environments experienced during migration are unknown. The reason is that tag and recapture methods (e.g., Schwarz et al., 1993; Block et al., 2005) and biologging technologies (e.g., Cooke et al., 2016) are difficult to apply to Pacific saury because of their small size and delicate nature. For this reason, migration model parameters must be estimated from limited data relating to Pacific saury distributions and be incorporated into migration model properly. Therefore, the modelling of migration characteristics is an important key to reproduce a plausible migration route. Because our goal is to improve the accuracy of forecast times and locations of the early Pacific saury fishing grounds, modelling specialized for forecast migration from pre-fishing to early fishing seasons would be effective.
We endeavor to develop a migration model for Pacific saury from pre-fishing to early fishing seasons using pre-fishing season Pacific saury distributions, fishery data, and oceanic conditions (temperature and velocity fields) obtained from an ocean circulation model (FRA-ROMS). For the initial distribution of Pacific saury we use data on the distribution of age-1 fish obtained from pre-fishing season stock assessment surveys. We determined migration characteristics from analysis of actual fishing grounds. For oceanic conditions, we incorporate two-dimensional temperature and velocity fields calculated by FRA-ROMS. Then, using reanalysis data and 2-month forecasts from FRA-ROMS, in July 2018, after completing stock assessment surveys, we forecast Pacific saury distributions 2 months in advance to predict the location of early fishing grounds from August to September. Our forecast fishing grounds were validated using fishery data on the actual fishing grounds for that year.
Section snippets
Outline of models
Our migration model for Pacific saury uses pre-fishing distributions obtained from stock assessment surveys as the initial distribution of Pacific saury, and two-dimensional temperature and velocity fields calculated by FRA-ROMS for oceanic conditions. As the horizontal resolution of FRA-ROMS is 0.1°N × 0.1°E, the horizontal resolution of our model is the same. Because the eastern boundary of the FRA-ROMS calculation area is 180°, our model was also run up to 180° (although stock assessment
Northern and spawning migration characteristics
The spatial and temporal distributions of fishery grounds in May (circles), June (crosses), and July (triangles) from 2007 to 2015 (JAMARC data) are depicted in Fig. 2a. Fishing grounds are distributed widely and discontinuously between 146°E and 173°E. Because fisheries operated in the high seas, a distribution limited by the boundary with the Russian EEZ is apparent. Fishing vessels could pursue a fish shoal to the EEZ boundary, but once a shoal moved into Russia's EEZ then vessels ceased
Discussion
Our migration model forecasts the distributions of early fishing grounds for Pacific saury in nearshore Japanese waters using two-dimensional temperature and velocity fields derived from FRA-ROMS. Our initial distributions of Pacific saury were based on observed pre-fishing season distributions obtained by TNFRI. Northern and spawning migration characteristics were estimated from analysis of actual fishing grounds. We estimated swimming speed of Pacific saury during the spawning migration based
Conclusion
We developed a migration model for Pacific saury that extends from the pre-fishing to early fishing seasons based on observed distributions of age-1 fish and their migration characteristics. We trialed this model in July 2018 by forecasting locations of early fishing grounds extending into August through to September. Our migration model successfully forecast the locations of early fishing grounds along the Kuril Islands archipelago. The locations of forecast particles that reached an offshore
CRediT author statement
Shigeho Kakehi: Conceptualization, Methodology, Software, Formal analysis, Validation, Writing - Original Draft
Jun-ichi. Abo: Investigation, Data curation
Hiroomi Miyamoto: Investigation
Taiki Fuji: Investigation
Kazuyoshi Watanabe: Data curation
Hideyuki Yamashita: Investigation, Data curation
Satoshi Suyama: Investigation, Data curation, Supervision, Project administration, Funding acquisition
Declaration of Competing Interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
Acknowledgements
We thank the captains and crews of the Hokko-maru, Hokuho-maru, and Oumi-maru for conducting stock assessment surveys. Fishery data from around Japan was provided by the National Saury Stick-held Dip Net Fishery Cooperation. We thank Drs T. Setou and H. Kuroda for providing FRA-ROMS data. We also thank Dr Y. Kurita of Tohoku National Fisheries Research Institute for his useful comments. This research was supported by the Fisheries Agency of Japan. We thank Edanz Group (www.edanzediting.com/ac)
References (36)
- et al.
Remote bioenergetics measurements in wild fish: opportunities and challenges
Compar. Biochem. Physiol. Part A: Mol. Integr. Physiol.
(2016) - et al.
On the interannual variability of the growth of Pacific saury (Cololabis saira): a simple 3-box model using NEMURO. FISH. Ecol
Model
(2007) - et al.
NEMURO-a lower trophic level model for the North Pacific marine ecosystem
Ecol. Model.
(2007) - et al.
A bioenergetics-based population dynamics model of Pacific herring (Clupea harengus pallasi) coupled to a lower trophic level nutrient-phytoplankton-zooplankton model: description, calibration, and sensitivity analysis
Ecol. Model
(2007) - et al.
Predicting the timing of Pacific saury (Cololabis saira) immigration to Japanese fishing grounds: a new approach based on natural tags in otolith annual rings
Fish. Res
(2019) - et al.
The importance of spawning season on the growth of Pacific saury: a model-based study using NEMURO
Fish. Ecol. Model.
(2007) - et al.
A simulation model for Japanese sardine (Sardinops melanostictus) migrations in the western North Pacific
Ecol. Model.
(2009) - et al.
Variability in transport processes of Pacific saury Cololabis saira larvae leading to their broad dispersal: implications for their ecological role in the western North Pacific
Progress in Oceanogr
(2015) - et al.
Magnetic information calibrates celestial cues during migration
Anim. Behav.
(2000) - et al.
Variations in the abundance of Pacific saury (Cololabis saira) from the northwestern Pacific in relation to oceanic-climate changes
Fish. Res
(2003)