Fig 1.
Observed glass eel catches from 1958 to 2018 in Japan (a), since 2000, with the years used in the simulations shown in red and Kuroshio large meander years with red circles (b), and monthly glass eel catches/input (showing percentage of annual catch) in Japan from 2016 to 2018 (c). Simulated monthly v-larvae arrival (%) to southern Japan (d).
Fig 2.
Mean ocean circulation during the 2018 eel catch season (November 2017 to May 2018).
Color shading is sea surface height (m), and vectors are 50−250 m mean ocean currents. Ky, S, T, and Ka show the Kyushu, Shikoku, Tokai, and Kanto regions of Japan, respectively. The green box shows the spawning area where v-larvae were released. Magenta box south of Kyushu shows the release region for examining the effect of the Kuroshio large meander with different swimming behaviors. The white line shows where the ocean current was stronger than 0.4 m/s during the 2010 Kuroshio non-large meander eel catch season.
Table 1.
List of experiments and the corresponding oceanic conditions used in each simulation for the main experiments on the effect of the presence of the Kuroshio large meander and swimming sensitivity experiments.
For the meander experiments, release dates were 2–4 days before the new moon of each month during the spawning period in the selected years, and for the sensitivity cases, v-larvae were released in the Kuroshio south of Japan during the peak recruitment season. The numbers listed in the parenthesis of the North Equatorial Current (NEC) region are the mean NEC bifurcation latitudes. Both Kuroshio large meander (LM) and no large meander years (NLM) were used in the simulations.
Fig 3.
Visitation frequency distributions of v-larvae at the end of January during the early recruitment period in Kuroshio large meander years (a, 2005, b, 2018) and normal years (c, 2016, d, 2010).
The unit of visitation frequency was normalized by the total number of released v-larvae.
Fig 4.
Visitation frequency distributions of v-larvae at the end of April during the late recruitment period in Kuroshio large meander years (a, 2005, b, 2018) and normal years (c, 2016, d, 2010).
The unit of visitation frequency was normalized by the total number of released v-larvae.
Table 2.
Percentage of successful recruitment to various regions of southern Japan in the swimming sensitivity experiment during the 2 large meander (LM; 2005, 2018) and non-large meander (NLM; 2010, 2016) years.
Fail accounted for those v-larvae that did not approach the Japanese shore.
Fig 5.
Visitation frequency of v-larvae released within the Kuroshio south of Japan in the large meander years (2005 and 2018, left) and non-large meander years (2010 and 2016, right) with different swimming strategies at the end of the recruitment season: Along current (a, b), northwestward swimming (c, d), 24-hour northwestward swimming (e, f), and half speed northwestward swimming (g, h).
Fig 6.
The averaged time taken (days) from the spawning area to each particular grid location within the simulation region.
Day 0 to 180 corresponds to the early migration period (p1), day 180 to 270 to the early recruitment period (p2), and after day 270 is the late recruitment period (p3).
Fig 7.
The simulated 50−250 m depth-averaged current speeds (cm/s) and eddy presence (percentage of eddies and average days of retention within the green box) at the corresponding periods of v-larvae passing through.
Period 1 (p1) is early migration from May to October, period 2 (p2) is early recruitment from November to coming January, and period 3 (p3) is late recruitment from February to April. Magenta lines indicate the sections used for calculating mean velocity. N, W, and E next the numbers indicate the current flow directions of northward, westward, and eastward, respectively.
Fig 8.
The 2018 Kuroshio geostrophic velocity anomaly (cm/s) south of Japan derived from tide gauge stations Oodori (ODO) and Naze (NAZ) (a). Plots of the sea surface height (b−e, color shading shows height in meters) and 50−250 m averaged ocean currents (vectors, m/s) in the late recruitment season (period 3, February to April). Red shading and clockwise circulation often indicate the presence of warm eddies, and blue and anti-clockwise circulation corresponds to cold eddies.
Fig 9.
Schematic diagram illustrating the ocean conditions and v-larvae distributions for the 2018 large meander (a,b) and 2016 non-large meander (c,d) years during the early migration and recruitment season (a, c), and late recruitment period (b, d).
Yellow circles represent eddies, blue dots represent v-larvae, the width of the red shading arrows reflects the strength of the Kuroshio (KUR), and green circle marks the spawning area in the North Equatorial Current (NEC). The dotted line enables comparison of the relative position of the NEC.