C-2-2 Corroborative calculations of tritium
concentrations in seawater simulated in the
radiological impact assessment using ROMS
NIISOE Tamon
Nuclear Regulation Authority JAPAN
IAEA Regulatory Review Mission on ALPS Treated Water Handling
23 March 20222Abstract
• Radiological impacts of ALPS treated water discharged into the ocean from
the Fukushima Daiichi Nuclear Power Station (FDNPS) depend on
concentrations of the discharged radionuclides in seawater.
• TEPCO conducted dispersion simulations in the radiological impact
assessment (RIA) using a regional ocean modeling framework, ROMS, to
assess discharged tritium concentrations in seawater, which were multiplied
by amount ratios in the source terms to derive concentrations of the other
discharged radionuclides in seawater.
• The NRA conducted dispersion simulations using ROMS and the same source
term in order to see whether tritium concentrations in seawater reported in
the RIA can be replicated.
• The NRA also checked the time trend of the modeled concentrations within
the simulated single year to see whether the first-year simulation can
represent any following year.
• The NRA also assessed tritium concentrations at the lateral boundaries of the
target domain of the simulations to compare them with the background
levels in seawater.3What is ROMS?
• ROMS (https://www.myroms.org/) is an open-source 3D
modeling framework of ocean dynamics developed by a
scientific community organized mainly by UCLA and Rutgers
Univ. It has been broadly applied to various research fields
including dynamics, geology, and biology in the ocean.
• ROMS calculates time evolution of variables in the ocean such
as current, surface elevation, temperature, and salinity. The
modules to calculate advection/diffusion for temperature and
salinity enable dispersion simulations of tracers like tritium.
• The governing equations are the primitive equations which
are common to ocean modeling frameworks.
• Hydrostatic equilibrium is assumed implying vertical scale
much smaller than horizontal scale for motion of seawater.ROMS(Regional Ocean Modeling System)4Methods(1)
• Dispersion simulations like the RIA were conducted applying
ROMS to the generally same domain with horizontal
resolution not of 200m but of 10km, because the
concentrations in seawater used for assessment in the RIA
were those averaged over an area of 10km square.
The target domain
and bathymetry(m)
Model grid cells in the vicinity of the FDNPS
Horizontal resolution of 10km
Target cell
Generally
same as that in
the RIA
(Figure 4-1).
Concentrations
in seawater at
the target cell
were compared
with those
reported in theRIA.290km wide from west to east
480km wide
from south to
north5Methods(2)
• Horizontal resolution:1km for
whole of the domain; 200m for
the vicinity of the FDNPS.
• Vertical structure:30 layers.
• Annual amount of discharge:
2.2E13Bq(constant).
• Target year:2019.
• Input data for ocean:JAMSTEX
reanalysis JCOPE2.
• Input data for atmospheric
forcing:Products made from JMA
Numerical Weather Prediction.
• Tracer as tritium : passive tracer*.
• Horizontal resolution:10km for
whole of the domain.
• Vertical structure:30 layers.
• Annual amount of discharge:
2.2E13Bq(constant).
• Target year:2019.
• Input data for ocean:HYCOM
analysis GOFS3.1.
• Input data for atmospheric
forcing:JMA Japanese 55-year
reanalysis.
• Tracer as tritium : passive tracer*.
Computational conditions
• *Passive tracer is a modeled substance which undergoes no removal processes except
outflow from the target domain.
• It is conservative to estimate concentrations of radionuclides other than tritium based on
those of tritium and the ratios of radioactivity in the source terms neglecting removal
processes such as gravity settling of particulate matters and radioactive decay.
RIA by TEPCO NRA6Results(1)
Annually averaged tritium concentrations in seawater (Bq/L)
Vertical profile
at the target cell
Depth(m)Sea
surface
Seabed
Concentration(Bq/L)
• The annually averaged tritium concentration
for all the model layers was comparable with
the RIA.
• Vertical mixing was not active due to the model
assumption of hydrostatic equilibrium.
• However, discharged ALPS treated water
could activate vertical mixing by promoting
local turbulence.
Similar trends are
found in the cross
sections of
concentration in
Figure 5-3 and 5-
4 of the RIA.
All layers
RIA 5.6E-02
(from Table 5-5 of theRIA)NRA 4.4E-027Results(2)
Time series of tritium concentrations in seawater (Bq/L)
• A monotonous increase was found only in the
first few days.
• It is not clear whether the cause of the gentle
ascending trend in the first 100 days was
accumulation or seasonality.
• After that, there was no ascending trend.
• If the year of 2019 is not an anomaly year, it is
reasonable to judge that the model results for a
single year of 2019 represent a typical state for
whole of the duration of discharge.
Start of the
model run on
Dec. 27, 2018,
with the initial
condition of zero.
End of the model
run on Jan. 01,
2020.
Averaged over
all the layers.8Results(3)
The maximum value of hourly averaged tritium concentrations in
seawater (Bq/L) at the lateral boundaries of the target domain
(For reference)
Background tritium level in seawater
・Reference value of UNSCEAR:1E-1a
・Observed values in Pacific Ocean and
Indian Ocean (1996-1997): 4E-2〜4E-1b
・Observed values in the coastal areas of Aomori
Prefecture in Japan (2001-2005):〜2.4E-1c
Lateral
boundaries
of the
target
domain
aUNSCEAR, 2016. Sources, effects, and risks
of ionizing radiation. Annex C: Biological
effects of selected internal emitters – Tritium.
bKakiuchi et al.,1999. Tritium concentration in
ocean. J. Radioanal. Nucl. Chem., 239 (3),
523-526.
c海洋生物環境研究所, 2017. 平成28年度原子力
施設等防災対策等委託費(海洋環境における放
射能調査及び総合評価)事業調査報告書.
North side:3.4E-4
East side:1.0E-3
South side:1.8E-39Summary
• The annually averaged tritium concentration for all the model layers was
comparable with the RIA.
• Vertical mixing of tritium was not active due to the model assumption of
hydrostatic equilibrium. In fact, however, the discharged ALPS treated
water could activate vertical mixing by promoting local turbulence.
• It is reasonable to judge that the model results for a single year of 2019
represent a typical state for the whole of the duration of discharge.
• The maximum values of hourly averaged tritium concentrations in
seawater at the lateral boundaries of the target domain of the
simulations were smaller than the background levels reported in
literatures by more than an order of magnitude.