Introduction
High tsunami waves after the Tohoku earthquake on 11 March 2011 damaged the
cooling system of the Fukushima Nuclear Power Plant (FNPP). Due to lack of
electricity, it was not possible to cool nuclear reactors and the fuel
storage pools that caused numerous explosions at the FNPP for details
see. The Fukushima accident was classified at the
maximum level of 7, similar to the Chernobyl accident which happened in 1986
in the former Soviet Union. Radionuclides were released from the FNPP through
two major pathways: direct discharges of radioactive water and atmospheric
deposition onto the North Pacific Ocean. Indirect estimation of that
deposition is in the range 6.4–35 PBq . The total
amount of 137Cs isotope released into the ocean was estimated
to be 3.6±0.7 PBq by the end of May 2011 .
A few special research vessel (R/V) cruises were conducted, just after
the accident and later, to measure radioactivity in sea water, zooplankton,
fish and in other marine organisms. 137Cs and
134Cs isotopes with 30.17 and 2.06 years half-life,
respectively, were detected over a broad area in the western North
Pacific in 2011 and 2012
.
137Cs concentration levels off Japan before the accident were
estimated at the background level to be 1–3 mBq kg-1, while
134Cs was not detectable. Because of a comparatively short
half-life time, any measured concentrations of 134Cs could
only be Fukushima derived.
The studied area is shown in Fig. a. It is known as the
Kuroshio–Oyashio confluence zone or a subarctic frontal area
. The Kuroshio Extension prolongs the Kuroshio Current which
turns to the east at about 35∘ N and flows as a strong
meandering jet constituting a front separating the warm subtropical and cold
subarctic waters. It is a region with one of the most intense air–sea heat
exchange and the highest eddy kinetic-energy level. The Kuroshio–Oyashio
confluence zone is populated with several mesoscale eddies that transfer
heat, salt, nutrients, carbon, pollutants and other tracers across the ocean.
They originate, besides from the Kuroshio Extension, from the Tsugaru Warm
Current, flowing between the Honshu and Hokkaido islands, and from the cold
Oyashio Current flowing out of the Arctic along the Kamchatka Peninsula and
the Kuril Islands (Fig. a). The lifetime of those eddies ranges
from a few weeks to a few years.
(a) The AVISO velocity field in the Kuroshio–Oyashio
confluence zone, averaged from 1993 to 2016. TsS stands for the Tsugaru
Strait. Location of the FNPP is shown by the radioactivity sign. The area
just around the FNPP is shown by the yellow lines. (b) The velocity
field on 24 August, 2011, with the Tohoku (TE) and Hokkaido (HE) eddies
studied in the paper and with tracks of some available drifters (the red
circles) and Argo floats (the green stars) present in the area at that time.
Elliptic and hyperbolic stagnation points with zero mean velocity are
indicated by triangles and crosses, respectively, with upward- and
downward-oriented triangles denoting anticyclones and cyclones,
respectively.
![]()
The standard approach in simulating transport phenomena, such as propagation of oil after the explosion
at the Blue Horizon mobile drilling rig in the Gulf of Mexico in April 2010 and propagation of
radioactive isotopes after the accident at the FNPP, is to run global or regional numerical models
of circulation to simulate propagation of pollutants and try to forecast their trajectories.
The outcomes provide “spaghetti-like” plots of individual trajectories which are hard to interpret.
Moreover, as the majority of real trajectories in a chaotic environment are very sensitive to small
and inevitable variations in initial conditions, they are practically unpredictable even over
a comparatively short time.
A specific Lagrangian approach, based on dynamical systems theory, has been
developed in recent decades with the aim of finding more or less robust
material structures in chaotic flows governing mixing and transport of
Lagrangian particles and creating transport barriers preventing propagation
of a contaminant across them for reviews
see. Identification of such structures
in the ocean would help to predict, for short and medium-length periods of
time, where a contaminant will move even without a precise solution of the
Navier–Stokes equations. This approach has been successfully used in
simulating propagation of oil in the Gulf of Mexico
and propagation of
Fukushima-derived radionuclides in the Pacific ocean
.
The present authors have developed a set of Lagrangian tools for tracking the origin, history and fate
of water masses advected by analytic, altimetric and numerical velocity fields generated by eddy-resolved
regional circulation models .
Each elementary
volume of water can be attributed to physico-chemical properties (temperature, salinity, density, radioactivity,
etc.) which characterize this volume as it moves.
In addition, each water parcel can be attributed to other types of diagnostics which
are exclusively a
function of its trajectory. We call them “Lagrangian indicators”.
They are, for example, distance traveled by a fluid particle for some
period of time; absolute, zonal and meridional displacements of particles from their original
positions; the number of their cyclonic and anticyclonic rotations;
time of residence of fluid particles inside a given area; exit time out off that area;
and the number of times particles visited different places in a studied region.
The Lagrangian indicators contain information about the origin, history and
fate of the corresponding water masses and allow the identification of water
masses that move coherently, either by propagating together or by rotating
together. Even if adjacent waters are indistinguishable, say, by temperature
(e.g., the satellite SST images indicate no thermal front), the corresponding
water masses could still be distinguishable by, for example, their origin,
traveling history and other factors. The Lagrangian indicators are computed
by integrating advection equations (Eq. ) for a large number of
synthetic particles forward and backward in time. When integrating
Eq. () forward in time, one computes particle trajectories to know
the fate of the corresponding particles, and when integrating
Eq. () backward in time, one could know where the particles came
from and the history of their travel.
The purpose of this paper is threefold. Firstly, we develop a Lagrangian methodology in order to track
and document the origin and history of water masses constituting prominent mesoscale features. It allows
the distinction of water masses inside mesoscale eddies originating from the main currents in the
Kuroshio–Oyashio confluence zone.
Secondly, we apply that methodology in order to identify and track the mesoscale eddies, advected by
the altimetric AVISO velocity field, with a risk of being contaminated by Fukushima-derived radionuclides.
Finally, the simulation results are compared qualitatively with in situ sampling of those eddies
in the R/V cruises. The location and form of the simulated eddies are verified, when possible, by tracks
of surface drifters and diving Argo floats available at the sites aoml.noaa.gov/phod/dac and
www.argo.net, respectively.
Data and methodology
All the simulation results are based on integrating equations of motion for
a large number of synthetic particles (tracers) advected by the AVISO velocity
field.
dλdt=u(λ,φ,t),dφdt=v(λ,φ,t),
where u and v are angular zonal and meridional velocities, and φ and λ are latitude
and longitude, respectively. The altimetry-based velocities were obtained from the AVISO database
(aviso.altimetry.fr) archived daily on a 1/4∘×1/4∘ grid. The velocity
field was interpolated using a bicubical spatial interpolation and third-order Lagrangian
polynomials in time. In integrating Eq. () we used a fourth-order Runge–Kutta
scheme with an integration step of 0.001 days.
The velocity field is from altimetry data, which provide the geostrophical component of the real
near-surface velocities valid at the mesoscale. In order to display the enormous amount of
information, we plot maps
of specific Lagrangian indicators versus particle's initial positions.
The region under study is seeded with a large number of Lagrangian particles whose trajectories
are computed for a given period of time. The results obtained are processed to get a data file
with the field of a specific Lagrangian indicator in this area. Finally, its values are coded
by color and represented as a map in geographic coordinates.
It is informative also to identify “instantaneous” stagnation elliptic and hyperbolic points on the
Lagrangian maps. We mark them by triangles and crosses, respectively. They are points with zero
velocity which are computed daily with the AVISO velocity field. The elliptic points are called
stable and the hyperbolic ones are unstable. Their local stability properties are characterized
by a standard method calculating eigenvalues of the Jacobian matrix of the velocity field.
The elliptic points, situated mainly in the centers of eddies, are those points around which
the motion is stable and circular. Upward (downward) orientation of one of the triangle's top on
the maps means anticyclonic (cyclonic) rotations of water around them. The hyperbolic points,
situated mainly between and around eddies, have stable manifolds along which water parcels
converge to such a point and unstable manifolds along which they diverge. The stagnation
points are moving Eulerian features and may undergo bifurcations in the course of time.
In spite of nonstationarity of the velocity field, some of them may exist for weeks and
much more. The hyperbolic points and their attracting and repelling manifolds were
recently identified with the help of drifter's tracks in the Gulf of La Spezia in the
northwestern Mediterranean Sea , in the
Gulf of Lion , in the Gulf of Mexico
and in the northwestern Pacific .
The Lagrangian maps show evolution of the Tohoku eddy (TE) from after the
accident to the days of its sampling and the origin of waters in its core and
at the periphery. The red, black and blue colors specify the tracers which
came for 2 years in the past to their places on the maps from the Kuroshio,
Oyashio and Tsushima currents, respectively, more exactly, from the
corresponding line segments shown in Fig. a. The yellow color
marks the Lagrangian particles coming from the area around the FNPP in
Fig. a (shown in Fig. a by the dashed line), after
the day of the accident on 11 March 2011. The TE was sampled on 10
and 11 June 2011 by along the transect
35.5– 38∘ N, 144∘ E shown in (c) and at
the end of July 2011 by along the transect
35∘ N–41∘ N, 144∘ E shown in (d).
The locations of stations with surface seawater samples (collected by and
) with measured radiocesium
concentrations at the background level are indicated by the green diamonds.
Stations where the concentrations were measured to be much higher are
marked by the magenta diamonds.
![]()
The altimetry-based Lagrangian maps allow accurate identification and tracking of mesoscale eddies and
document their transformation due to interactions with currents and other eddies. Inspecting
daily-computed Lagrangian maps for a long period of time (up to 2 years in this paper) and
computing stagnation elliptic points daily, one can track the origin and fate of water masses
within a given eddy if it is sufficiently
large and long lived (i.e., more than a week). For this purpose Lagrangian diagnostics are more
appropriate than commonly used Eulerian techniques, because Lagrangian maps are imprints of
the history of water masses involved in the vortex motion,
whereas vorticity, Okubo–Weiss parameter and similar indicators are only
instantaneous snapshots (see , and , for comparison).
Being motivated by the problem of identification of Fukushima-contaminated waters in the core
and at the periphery of persistent mesoscale eddies in the area, we develop in this paper a
specific Lagrangian technique designed to distinguish water masses of a different origin
inside the eddies with a risk of being contaminated. With this aim we specify, besides
Fukushima-derived waters, water masses originated from the main currents in the
Kuroshio–Oyashio confluence zone. The integration was performed backward in time.
We removed from consideration all the particles entered into any AVISO grid cell with two
or more corners touching the land in order to avoid artifacts
due to the inaccuracy of the altimetry-based velocity field near the coast.
In what follows, we define the “yellow” waters on the maps as those which
have a large risk of being contaminated because they came after the accident
from the area just around the FNPP, enclosed by the yellow straight lines in
Fig. a, for the period from the day of the accident,
March 11, 2011, to May 18, 2011, when direct releases of radioactive isotopes
to the ocean and atmosphere stopped. The “red” waters are salty and warm
Kuroshio waters. To be more exact, they came from the red zonal line
(34.5∘ N, 139–144∘ E) in Fig. a,
crossing the Kuroshio main jet. The “black” waters came from the warm
Tsushima Current flowing via the Tsugaru Strait out off the Japan Sea and
across that strait (the black line with 40–43∘ N,
141.55∘ E). The “blue” waters are fresher and colder waters
originating from the Oyashio Current and crossing the blue zonal line
(48∘ N, 153–159∘ E) shown in Fig. a.
The “white” waters on the Lagrangian maps have not been specified as originating from one of the segments mentioned above. They could reach their
places on the maps from anywhere besides those segments.
We are interested in advective transport for a comparatively long period of time, up
to 2 years. It is hardly possible to adequately simulate motion of a specified passive
particle in a chaotic flow, but it is possible to reproduce transport of a statistically
significant number of particles. Our results are based not on simulation of individual
trajectories but on statistics for 490 000 Lagrangian particles. We cannot, of course,
guarantee that we compute “true” trajectories for individual particles. The description
of the general pattern of transport for half a million particles is much more robust. However,
we do not try to quantitatively simulate the concentration of radionuclides or estimate the
content of water masses of different origin inside the studied eddies.
Results
A few mesoscale eddies were present in the studied area on the day of the accident. The cyclonic
eddies with the centers, marked by the downward-oriented triangles on the Lagrangian maps,
prevailed in the
area to the north of the Subarctic Front, the boundary between the subarctic (blue) and
subtropical (red) waters in Fig. . The anticyclonic eddies with the centers,
marked by the upward-oriented triangles, prevailed to the south of the front.
The large anticyclonic Tohoku eddy (TE,) with the center at around
39∘ N, 144∘ E in March 2011, was sampled after the
accident in the two R/V cruises in June and July 2011
, showing large concentrations of 137Cs and
134Cs. The anticyclonic Hokkaido eddy (HE), genetically
connected with the TE, originated in the middle of May 2011 with the center at
around 40∘ N, 145∘ E. After that it captured some
contaminated water from the TE. It was sampled at the end of July 2011
.
The anticyclonic Tsugaru eddy (TsE) was genetically connected with the HE. It
originated in the beginning of February 2012 with the center at around
41.9∘ N, 148∘ E and captured some contaminated water
from the HE. The TsE was sampled in the R/V Professor Gagarinskiy cruise on 5 July, 2012, and found to have concentrations of
137Cs and 134Cs over the background level at
the surface and at intermediate depths . All these eddies will
be studied in this section from the Lagrangian point of view in order to
simulate and track by which transport pathways they could have gained water
masses from the Fukushima area or from other origins and to compare
qualitatively the simulation results with in situ measurements.
The Tohoku eddy
We tracked with daily-computed Lagrangian maps the birth, metamorphoses and
decay of the mesoscale anticyclonic TE. It originated in the middle of May 2010
with the elliptic point at around 38∘ N, 144∘ E at that
time as the result of interaction of a warm anticyclonic Kuroshio ring with a
cyclone with mixed Kuroshio and Oyashio core waters. It has interacted with
other eddies almost for a year, with multiple splitting and merging in the
area to the east off the Honshu Island. Just after the accident, it began to
gain yellow water from the area around the FNPP with a high risk of
contamination. That eddy is clearly seen in an earlier simulation just after the
accident in Fig. 3b by and on the Lagrangian map in
Fig. a as a red patch labeled as TE with the center at
39∘ N, 144∘ E on 26 March 2011.
The maps in Fig. and in the subsequent figures were
computed, as was explained in Sect. 2. The red color in the core of the TE means
that its core water was of subtropical origin. More precisely, the red
tracers were advected for 2 years from the red line segment in
Fig. a to the current place on the map. In March 2011 yellow
water, coming from the area around the FNPP with a comparatively high risk of being contaminated, wrapped round the TE. A thin streamer of Tsugaru black
water, coming from the black line segment in Fig. a, wrapped a
periphery of the TE at the end of March. Yellow waters propagated
gradually to the east and south due to a system of currents wrapping around
the eddies present in the area. The straight zonal boundary along
36.5∘ N and meridional boundary along 144∘ E,
separating water masses of different origin in Fig. a on 26 March 2011, are just fragments of the boundary in Fig. a restricting
the area around the FNPP. These boundaries separate the yellow tracers
which were present within the area from those which have not yet managed
to penetrate inside the area for 15 days after the accident.
In April and May 2011 the TE had a sandwich-like structure, with the red
subtropical core belted with a narrow streamer of Fukushima yellow waters
which, in turn, was encircled by a red streamer of Kuroshio subtropical water
(Fig. b). A new eddy configuration appeared at the end of May in
Fig. b, with the TE interacting with a blue cyclone with the
center at 39.9∘ N, 144.7∘ E and a newborn yellow
anticyclone which we call the Hokkaido eddy with the center at
40.4∘ N, 145.5∘ E. The core of that cyclone consisted
of a blue subarctic Oyashio water with low risk of being contaminated, but
the HE core water came from the area around the FNPP with a high risk of being
contaminated.
In the course of time the TE moved gradually to the south. Its periphery was sampled at the beginning of June by , and the whole
eddy was crossed at the end of July 2011 by .
Fukushima-derived cesium isotopes were measured on 10 and 11 June during
the R/V Ka'imikai-o-Kanaloa cruise along the
144∘ E meridional transect where the cesium concentrations were found to be in the range from the background level,
C137= 1.4–3.6 mBq kg-1 (stations 13 and 14), to a high level up
to C137=173.6±9.9 mBq kg-1 (station 10). The ratio
134Cs/137Cs was close to 1.
For ease of comparison, we mark, with the green diamonds in Fig. c, the locations
of stations 13 and 14 with collected surface seawater samples by in which
the cesium concentrations were measured to be at the background level (≲3.6 mBq kg-1).
The stations 10,
11 and 12, where the concentrations were found to be much larger, are indicated by the
magenta diamonds. Our simulation in Fig. c shows that stations 13 and 14 on the
days of sampling were located in red and white waters with a low risk
of containing Fukushima-derived radionuclides.
Transport and mixing at and around stations 10, 11 and 12 with high measured
values of the cesium concentrations were governed
mainly by the interaction of the TE with the yellow mesoscale cyclone
with the center at 37.2∘ N, 142.8∘ E. This cyclone
formed in the area in April and captured yellow waters with a high risk
of contamination. Unfortunately, it has not been sampled in the
R/V Ka'imikai-o-Kanaloa cruise. The surface seawater samples at
stations 10, 11 and 12 were collected on the days of sampling at the
eastern periphery of that cyclone and at the southern periphery of the TE
with the yellow streamer there. Station 10, with the highest measured
level of the 137Cs concentration of C137=173.6±9.9 mBq kg-1, was located at 38∘ N, 144∘ E
inside the wide streamer of yellow water around the TE. Stations 11 and
12, with C137=103.7±5.9 mBq kg-1 and C137=93.6±4.9 mBq kg-1, respectively, were located within the narrow
streamers with yellow simulated water in Fig. c intermitted
with narrow streamers of red water. So, we estimate the likelihood of finding
Fukushima-derived radionuclides there (the magenta diamonds) to be much
higher than at stations 13 and 14 (the green diamonds), and it is confirmed by
a qualitative comparison with measured data.
A specific configuration of mesoscale eddies occurred in the area to the
northeast of the FNPP at the end of July 2011, the days of sampling by
along the 144∘ E meridian from
35 to 41∘ N during the R/V Kaiun maru cruise.
That transect is shown in Fig. d. It crosses the TE and the
cyclone with blue Oyashio water, which is genetically linked to the
blue cyclone at 39.9∘ N, 144.7∘ E in
Fig. b. The transect also partly crosses the periphery of the
anticyclonic HE. The measured 137Cs concentrations in
surface seawater samples at the stations C43–C55 were found to be in
the range from the background level, 1.9±0.4 mBq kg-1 (station
C52), to a much higher level of 153±6.8 mBq kg-1 (station C47).
The colored tracking maps in Fig. 5 by show where the
simulated tracers of that transect were moving from 11 March to 10 April,
2011, being advected by the AVISO velocity field.
The risk of radioactive contamination of the markers placed at
36–36.5∘ N was estimated by , to be small,
because they were advected mainly by the Kuroshio Current from the
southwest to the east (the corresponding concentrations were measured by
, to be 2–5 mBq kg-1). The present simulation in
Fig. d also shows that stations C51, 52 and 53 (the green
diamonds), with the measured cesium concentrations at the background level on
the days of sampling by were located in the red
waters (stations C51 and C53) advected by the main Kuroshio jet from the
southwest and in the white waters (station C52) between the TE and the
jet. Therefore, we estimate the likelihood of finding Fukushima-derived radionuclides
there to be comparatively low.
The transect 36.5–38∘ N in Fig. d (the red one in
Fig. 5 by ) crossed the TE. The 137Cs concentrations
at the stations C49 and C50 of that transect were measured to be 36±3.3 and 50±3.6 mBq kg-1 . Comparing those
results with simulated ones, we note the presence of yellow water in the
TE core at the locations of those stations. Surface samples at station C48
(38.5∘ N) were measured to contain the 137Cs
concentration to be at the background level 2.7±0.6 mBq kg-1
. The corresponding green diamond is located in our
simulation in the area with red and white waters.
(a)–(b) The Lagrangian maps show evolution of
the Hokkaido eddy (HE) after the FNPP accident to the days of its sampling
and the origin of waters in its core and at the periphery. (c)–(d) A fragment of the track of the drifter no. 39123 is indicated by
the full circles for 3 days before the day indicated with the size of
circles increasing in time. Tracks of three Argo floats are shown by the
stars. The largest star corresponds to the day indicated and the other ones each
show float positions 7 days before and after that date.
![]()
Inspecting the Lagrangian maps on the days between 6 June and 28 July (not
shown), we have found that the yellow cyclone with the center at
37.2∘ N, 142.8∘ E in Fig. c collapsed at the
end of June. Its yellow core water with a high risk of being contaminated
was wrapped around the neighbor anticyclone TE in the form of a wide
yellow streamer visible in Fig. d. The highest concentration,
C137=153±6.8 mBq kg-1, was measured by
at station C47 (39∘ N), situated in the area of that streamer.
Stations C46 (39.5∘ N) with C137=83±5.0 mBq kg-1 is
situated in the close proximity to a yellow streamer sandwiched between
white and black waters.
A comparatively high concentration, C137=65±4.3 mBq kg-1, was
measured by at station C45 (40∘ N) during
the days of sampling in the core of the blue cyclone with the center at
39.7∘ N, 144.2∘ E (Fig. d). Our simulation
shows that it was formed mainly by Oyashio blue waters (with a low
risk of being contaminated by Fukushima-derived radionuclides) and partly by
white waters.
The Lagrangian maps in the study area in the first half of 2012.
(a) The locations of stations in the beginning of February with
surface seawater samples (collected by ) with measured
radiocesium concentrations at the background level (the green diamonds) and
with higher concentration levels (the magenta diamonds). (b–d) The
Lagrangian maps show evolution of the Tsugaru eddy (TsE), which originated on
4 February 2012 (a) after splitting of the HE and was sampled by
at station 84 on 5 July 2012 and shown to have increased radiocesium
concentrations (the magenta diamond in d).
![]()
When comparing simulation results in Fig. d with the measurements by ,
we have found that the simulation is consistent with samplings at stations C48, 51, 52 and 53 in the sense
that the cesium concentrations were measured to be at the background level in those places on
the maps where there is no signs of yellow water with a high risk of containing Fukushima-derived
radionuclides. Our simulation is also consistent, at least quantitatively, with samplings at stations
C47, 49 and 50 with high measured levels of the cesium concentrations because the yellow
water is present there in our simulation.
However, there is an inconsistency of simulation with samplings at stations C45 and C46, where there
are practically no yellow tracers but rather only blue and white ones. The reasons for this inconsistency
might be different. In this paper we track only those tracers which originated from the blue,
red and black segments as well as the yellow rectangular around the FNPP shown in Fig. a. So we
did not specify the origin of white waters. They could reach their places on the maps from anywhere
besides those segments and the area around the FNPP. They could in principle contain Fukushima-derived
radionuclides that were deposited at the sea surface from the atmosphere after the accident and then
advected by eddies and currents in the area. Moreover, they could be those tracers which
were located inside AVISO grid cells near the coast around the FNPP just after the accident and were then
advected outside. We removed from consideration all the tracers entered into any AVISO
grid cell with two or more corners touching the land because of inaccuracy of the altimetry-based
velocity field there and in order to avoid artifacts.
Thus, the white streamers inside the core and at the periphery of the blue
cyclone with the center at 39.7∘ N, 144.2∘ E (nearby
stations C45 and C46 with high measured concentrations of cesium by
) could, in principle, contain contaminated water.
However, it has not been proved in our simulation due to the above-mentioned reasons.
The Hokkaido eddy
Now we consider the anticyclonic HE. It originated in the middle of May (see
the yellow patch in Fig. b with the center at 40.3∘ N,
145.5∘ E), being genetically linked to the TE. During May, the TE
gradually lost a Fukushima yellow water from its periphery to form the
core of the HE. Fig. a shows the HE with a yellow core surrounded
by modified subtropical red water which, in turn, is surrounded by
Tsugaru black water.
The sampling of that eddy and its periphery by along the
144∘ E meridian at the end of July showed comparatively high
concentrations of C137=60±4.0 and 71±4.6 mBq kg-1 at
stations C44 (40.5∘ N) and C43 (41∘ N), respectively.
Station C43 was located inside the anticyclone HE filled mainly by yellow
waters, and we estimate the likelihood of finding Fukushima-derived radionuclides
there to be large. Station C44 was located at the southern periphery of the
anticyclone HE at the boundary between white and blue waters but in
close proximity to a yellow streamer.
The location of the HE on 24 August 2011 is shown in the AVISO velocity field
in Fig. b. To verify the simulated locations of the HE and its
form, we plot in Fig. c and d fragments of the tracks of a
drifter and three Argo floats captured by that eddy in September 2011. A
fragment of the track of the drifter no. 39123 is shown by the red circles
with the size increasing in time for 3 days before the dates indicated in
Figs. c and d and decreasing for 3 days after those dates, i.e.,
the largest circle corresponds to the drifter position at the indicated date.
It was launched after the accident on 18 July 2011 at the point
45.588∘ N, 151.583∘ E in the Oyashio Current, advected
by the current to the south and eventually captured by the HE moving around
clockwise. Fragments of the clockwise tracks of the three Argo floats are
shown by stars in Fig. c and d for 7 days before and 7 days after the indicated dates. The float no. 5902092 was released long
before the accident on September 9, 2008 at the point 32.699∘ N,
145.668∘ E to the south of the Kuroshio Extension jet and was
able to cross the jet and go far north. The float no. 2901019 was
released before the accident on 19 April 2010 at the point
41.723∘ N, 146.606∘ E. The float no. 2901048 was
released just after the accident on 10 April 2011 at the point
37.469∘ N, 141.403∘ E nearby the FNPP.
Our simulation shows that the HE contained, after its formation in the middle
of May 2011, a large amount of yellow water probably contaminated by the
Fukushima-derived radionuclides. This conclusion is supported by an increased
concentration of radiocesium measured in its core at station C43 by
at the end of July 2011. The HE persisted in the area
around 42∘ N, 148∘ E up to the end of January of the
next year. It eventually split on 31 January 2012 into two anticyclones.
The Tsugaru eddy
The anticyclonic TsE originated on 4 February 2012 after decay of the HE (the
yellow patch with the elliptic point at 42∘ N, 145.6∘ E
in Fig. a). The elliptic point at the center of the TsE appeared
at 41.8∘ N, 146.9∘ E. Just after its birth, the HE
begun to transport its yellow water around the TsE with the core
consisted of an Oyashio blue water (Fig. b). The strong
Subarctic Front is visible in Fig. as a contrast boundary
between Oyashio blue water and Fukushima-derived yellow water, with
the Tsugaru black water in between.
Seawater samples for radiocesium measurements in the frontal area were
collected during the R/V Mirai cruise from 31 January to 5 February
2012 along one of the observation lines of the World Ocean Circulation Experiment
(WOCE) in the western Pacific, specifically the WOCE-P10–P10N line
. We impose on the simulated Lagrangian map in
Fig. a locations of stations to the north of the Kuroshio
Extension (> 36∘ N) with measured levels of the cesium
concentrations. As before, the green diamonds mark locations of those
stations, P10–114 (42.17∘ N, 143.8∘ E), P10–112
(41.75∘ N, 144.13∘ E), P10–110 (41.25∘ N,
144.51∘ E), P10–108 (40.76∘ N, 144.88∘ E),
P10–106 (40.08∘ N, 145.37∘ E) and
P10–104 (39.42∘ N, 145.85∘ E), where the cesium
concentrations in surface seawater samples were measured by
to be at the background level.
The stations, P10–102 (38.75∘ N, 146.32∘ E), P10–100
(38.08∘ N, 146.77∘ E), P10–98 (37.42∘ N,
147.2∘ E), P10-96 (36.74∘ N,
147.63∘ E) and P10–94 (36.08∘ N, 148.05∘ E),
where the concentrations were found to be larger (but not exceeding 25.19±1.24 mBq kg-1
for 137Cs), are indicated by the magenta diamonds. It is worth stressing a good qualitative
correspondence with our simulation results 10 months after the accident in the sense that stations
with measured background level are in the area of Oyashio blue waters with low risk of being
contaminated, whereas stations with comparatively high levels of radiocesium concentrations
are in the area of the Fukushima-derived yellow waters with increased risk of contamination.
As to the TsE, it was sampled later, in 5 July 2012, during the cruise of the R/V Professor Gagarinskiy
when it was a comparatively large mesoscale
eddy around 150 km in diameter with the elliptic point at
41.3∘ N, 147.3∘ E consisting of intermittent strips of
blue and yellow waters (Fig. d), which were wrapped
around during its growth from February to July 2012. Station 84 in that
cruise was located near the elliptic point of that eddy (called “G” by
). The concentrations of 137Cs at the surface and at
100 m depth were measured as 11±0.6 and 18±1.3 mBq kg-1, respectively, an order of magnitude larger than the
background level. As to the 134Cs concentration, it was
measured to be smaller, 6.1±0.4 and 10.4±0.7 mBq kg-1, due
to a shorter half-lifetime of that isotope. In fact, it was one of the
highest cesium concentrations measured inside all the eddy features sampled
in the cruise 15 months after the accident.
The maximal concentration of radionuclides was observed, as expected, not at
the surface but within subsurface and intermediate water layers (100–500 m)
in the potential density range of 26.5–26.7 due to a convergence and
subduction of surface water inside anticyclonic eddies. The corresponding
tracking map in Fig. 10c by confirms its genetic link with
the TE, and, therefore, a probability of detecting increased cesium
concentrations was expected to be comparatively large. We were able to track
all the modification of the TsE up to its death on 16 April 2013 in the
area around 40∘ N, 147.5∘ E.