Hypoxia Controlled by Hydrodynamics Akihide Kasai

Aqua-BioScience Monographs, Vol. 7, No. 4, pp. 117–145 (2014)
www.terrapub.co.jp/onlinemonographs/absm/
Hypoxia Controlled by Hydrodynamics
Akihide Kasai
Field Science Education and Research Center
Kyoto University
Oiwake, Kitashirakawa, Sakyo, Kyoto 606-8502, Japan
e-mail: [email protected]
Abstract
In summer dissolved oxygen is often depleted in the lower and bottom layers in many
coastal basins all over the world. This phenomena is called hypoxia. When the oxygen
consumption exceeds oxygen supply, the water becomes hypoxic. The oxygen is consumed by decomposing organic matter by bacteria (biochemical processes), while the
oxygen is supplied by physical processes such as convection, advection and diffusion.
The primary cause of hypoxia is the consumption of oxygen in the water column, but
physical processes mainly control its generation, distribution and configuration. In addition to the vertical supply of oxygen by mixing, horizontal transport by estuarine circulation plays the major role in the formation of hypoxia in regions of freshwater influence.
As the hypoxic water contains a lot of nutrients, it plays an important role for primary
production, producing middle layer chlorophyll maximum in summer and inducing bloom
of phytoplankton in autumn.
1. General introduction
Oxygen is essential for almost all marine biota, including fishes and invertebrates, to maintain their life.
However, the amount of oxygen diluted in the water is
limited. The saturation rate of oxygen in the water is
only 5.2 mL L –1 (=7.4 mg L–1) under 1 atoms at 20°C.
This concentration is significantly lower than that in
the air (210 mL L–1). Therefore, marine animals develop advanced gills to take in oxygen efficiently from
the seawater. In spite of the advanced organ intrinsic
to marine animals, significant decrease of dissolved
oxygen (DO) in the water damages them. DO concentration sometimes reduces seriously to the level which
has harmful effects on marine animals especially in
summer. This water mass is called hypoxia or hypoxic
water. The water including nearly zero amount of oxygen is called anoxia.
Oxygen depletion exerts a serious impact on marine
ecosystems, although the tolerability of marine animals
is different among the species. For instance, fishes such
as red sea bream and yellowtail are going to die within
a few days by exposure to the water with 3 mg L –1 of
DO (Ishioka 1982; Yamamoto et al. 1990). In general,
oxygen deficiency lower than 4 mg L–1 exerts a baneful
influence upon cultured fish (Inoue 1998). On the contrary, benthic animals tend to be tolerant to low oxy-
© 2014 TERRAPUB, Tokyo. All rights reserved.
doi:10.5047/absm.2014.00704.0117
Received on December 20, 2013
Accepted on
April 15, 2014
Online published on
October 24, 2014
Keywords
• circulation
• dissolved oxygen
• estuarine circulation
• fortnightly shifts
• hydrographic condition
• hypoxia
• tide
gen. Starfish and brittle stars are stressed when DO
drops below 1.5 mg L–1, and are found dead when there
is less than 1 mg L–1 (Simpson and Sharples 2012).
Some bivalves can survive over one week even if DO
concentration is less than 1.5 mg L–1. However, even
those low oxygen tolerant animals decrease their activities in hypoxia, and all marine animals becomes
debilitated and cannot survive in anoxic water.
Animals that have sufficient swimming ability to
control their position vertically within the water column (e.g., fish, demersal invertebrates such as shrimp
and crab) attempt to leave the hypoxic region in the
lower layer. Animals that have less ability to leave the
region of low DO gradually become stressed and, if
concentrations of oxygen drop low enough, weaken and
finally die. It has been reported that hypoxia often
causes problems in many eutrophic estuaries and
coastal areas (e.g. Borsuk et al. 2001; Hagy et al. 2004;
Gilbert et al. 2005). More than 400 hypoxic waters have
been reported in many coastal areas, including Baltic
Sea, Mexico Bay and various estuaries. Ise Bay is
highly eutrophic and is famous for its bottom hypoxia
in summer (Kuno 1996), and the oxygen depletion severely damages the ecosystem and fisheries in the bay
(Hossain and Sekiguchi 1996). In Mikawa Bay, the
number of benthic animals seriously decreases in the
DO concentration under 3 mg L–1 (Suzuki 1998). In
A. Kasai / Aqua-BioSci. Monogr. 7: 117–145, 2014
118
(a)
(b)
3
2
1
(c)
Fig. 1. (a) The position of hypoxic water in the bottom layer in summer, (b) diversity index of macro benthos and (c) the
distribution of log 10(H/U3), where H is the depth of water and U is the amplitude of M2 tidal current (after Yanagi 1990) in the
Seto Inland Sea. Reprinted from Yanagi, Science of tidal front, 169 pp.,  1990, with permission from Koseisha-Koseikaku.
the Seto Inland Sea, there are some regions including
Osaka Bay, Harima-Nada, Hiuchi-Nada, Hiroshima
Bay, Suo-Nada and Beppu Bay, where hypoxia occurs
every summer (Fig. 1a). The diversity of species of
macro benthic animals is low in the hypoxic regions
(Fig. 1b). One of the most wide-spread hypoxia is observed off the coast of Louisiana and Texas, USA
(Rabalais et al. 2002). The survey of the region shows
an area of about 17,000 km2 experiencing hypoxia,
which leads to large changes in bottom water marine
life. This region is called the “dead zone” because of
the failure to catch demersal fish and benthic animals.
Not only with the objective of environments, but also
fisheries are of course damaged by hypoxia, especially
in enclosed euphotic bays and lakes. Demersal fishes,
crabs and shellfishes are rarely observed in the bay head
of Mikawa Bay in summer. Shijimi clam fisheries in
Lake Shinji and Lake Ogawara, both of which are foremost Shijimi fisheries brackish lakes in Japan, are restricted in the narrow coastal areas which are shallower
than 5 m depth, because the deep central areas become
hypoxic in summer and shijimi clams cannot survive
doi:10.5047/absm.2014.00704.0117 © 2014 TERRAPUB, Tokyo. All rights reserved.
A. Kasai / Aqua-BioSci. Monogr. 7: 117–145, 2014
(a)
119
(b)
Mixed
Winter
Mixed
Summer
Fig. 3. Schematic view of estuarine circulation in (a) winter
and (b) summer in a coastal embayment.
Fig. 2. Schematic view of vertical profiles of (a) tempera-
ture and (b) dissolved oxygen concentrations in lakes or
fjords. Sp: spring, Su: summer, F: fall, and W: winter.
there. It is therefore crucially important to clarify the
formation mechanism of hypoxia and to develop a
scheme to reduce it.
2. Formation mechanism of hypoxia
In general, both physical and biochemical processes
control the generation of hypoxia. The oxygen supply
is mainly determined by the physical processes such
as water exchange between the water mass and the surrounding water. DO is usually saturated in the sea surface, because oxygen is supplied into seawater from
the air in the process of turbulence and mixing by
windswell and/or strong currents. As phytoplankton
releases oxygen through photosynthetic activities, the
oxygen in the surface layer sometimes over saturates
in the surface and subsurface layers, especially in
blooming seasons. In the deep layer, on the other hand,
oxygen is not supplied from the air and photosynthesis would be weak because of insufficient irradiation.
Both physical and biological processes therefore cannot supply oxygen in the deep layer.
Oxygen consumption is mainly determined by biochemical processes concerning bacterial activity. Carcass of phytoplankton produced in the euphotic layer
gradually sinks, as well as other organic matter such
as terrestrial matter. Oxygen is consumed as bacteria
decompose organic matter during sinking. Not only
small organic particles, which sink slowly, but also the
larger particles such as carcass and feces of
zooplankton sink faster and accumulates at the bottom.
Oxygen is subsequently consumed when the accumulated organic matter is decomposed and/or consumed
by other biota. The oxygen consumption exceeding the
supply reduces DO concentration in the water, and accordingly leads the water hypoxic.
The vertical supply of oxygen has been mainly focused for the formation and demise of hypoxia in the
conventional theory. In spring and summer, strong irradiation warms surface water and makes the water
density smaller in the surface than in the bottom (stratification, Fig. 2a). The mixing between the oxygen rich
upper water and bottom water is thus limited when the
water is stratified. The bottom water becomes hypoxic,
because oxygen consumption is larger than the supply
(Fig. 2b). In autumn, in contrast, the air cools sea surface, making the surface water denser. This leads to
mixing between the surface and lower layer, supplying oxygen to the lower layer from the surface. The
hypoxia disappears when the whole water column is
well mixed in the late autumn or winter. This process
is often observed in fjords and lakes, where horizontal
currents are weak in the bottom layer.
On the other hand, recent studies have shown that
horizontal currents play an important role in the variation of oxygen concentration in coastal areas
(Takahashi et al. 2000). Within the regions where freshwater flows in, buoyancy input into the bay head is
responsible for producing a physical regime that is radically different from lakes and open ocean (Simpson
1997). It is well known that less saline lighter water
flows seaward in the upper layer, while more saline
heavier water flows landward in the lower layer
(Hansen and Rattray 1966). This flow pattern is called
“estuarine circulation” and often observed in estuaries
and bays, as the water exchange between coastal basins and adjacent ocean has long been a topic of great
interest to oceanographers (e.g., Geyer and Cannon
1982; LeBlond et al. 1991; Allen and Simpson 1998).
If the estuarine circulation is dominant, bottom water
is readily renewed by oceanic water. This indicates that
oxygen would be supplied to the bottom layer and hypoxia hardly happens. The hypoxia is, nonetheless,
ubiquitous in many bays all over the world.
Recent surveys have shown that the circulation does
not always follow the classical estuarine circulation
pattern in regions where the stratified area (wide bay
area) located next to the mixed area (narrow strait).
The flow pattern in those regions has a bimodal character: bottom inflow of the mixed water in the cooling
season, while mid-layer inflow in the heating season
(Kasai et al. 2000, 2002; Takahashi et al. 2000;
Fujiwara et al. 2002). In winter, the temperature is reversed and weak stratification is maintained only by
doi:10.5047/absm.2014.00704.0117 © 2014 TERRAPUB, Tokyo. All rights reserved.
120
A. Kasai / Aqua-BioSci. Monogr. 7: 117–145, 2014
freshwater buoyancy input inside the bay (Fig. 3a). Sea
surface cooling causes the freshwater to sink through
the mixing process, so that the water in the bay is lighter
than that at the strait. This density difference leads to
bottom inflow in the cooling season, following the
conventional estuarine circulation pattern. During summer, on the other hand, the effect of surface heating is
restricted to the upper layer by the strong thermocline
and halocline in the stratified area (inside the bay),
while it extends to the bottom due to the strong tidal
currents in the well-mixed area (strait). This process
causes the density of the bottom bay water to be greater
than that of the strait water. The well-mixed strait water thus has a density equivalent to the middle layer
inside the bay, and then tends to flow not through the
bottom layer but through the middle layer, when it intrudes into the bay in the heating season (Fig. 3b,
Takahashi et al. 2000; Kasai et al. 2002). The bottom
water in the stratified area is therefore a relic spring
water, which temperature is lower than the surrounding water. This process limits oxygen supply to the
relict water, making it hypoxic in summer.
3. Three-dimensional circulation and hypoxia in
a coastal embayment
3-1. Introduction
Ise Bay is one of the major coastal embayments on
the Pacific coast of Japan (Fig. 4). It is ~30 km wide
and 60 km long, and has a mean depth of ~20 m with
the deepest longitudinal depression of about 35 m depth
in the middle. Three large rivers (Ibi, Nagara, Kiso
Rivers) flow into the bay head in the north, while in
the south the bay opens to the Pacific Ocean via the
narrow Irago Strait, which has a width of ~10 km.
Nutrient and organic loading from the rivers results in
serious eutrophication, especially at the bay head. Inside the bay, the buoyancy input by freshwater and
weak tidal currents (~10–1 m s–1) cause the water to
stratify. Based on the stratification-circulation diagram
by Hansen and Rattray (1966), the water condition is
classified as a strongly stratified regime in summer,
and a weakly stratified regime in winter. Both large
freshwater discharge and sea surface heating reinforce
the stratification in summer. On the other hand, stirring by strong tidal currents in the strait (~1 m s–1)
promotes strong vertical mixing of the oceanic heavy
water with relatively lighter water from the bay to produce intermediate density water. The coexistence of
stratified and mixed regions complicates the circulation in this system (Kasai et al. 2002).
The oxygen in the bottom water is often depleted in
the stratified season, although the bay has no sill in its
mouth (Fig. 4). It is reported that shellfish fisheries
are strongly damaged by the hypoxia in summer espe-
Fig. 4. Bathymetry of the Ise Bay. ADCP observations were
carried out along the thick straight lines A, B and C. Solid
circles and a triangle indicate CTD stations. Open circles
are the CTD and oxygen observation points shown in Fig. 5.
Reprinted from Cont. Shelf Res., 22, Kasai et al., Circulation and cold dome in a gulf-type ROFI, 1579–1590,  2002,
with permission from Elsevier.
cially in the western basin of the bay. An example of
the longitudinal distribution of hydrographic and DO
conditions is shown in Fig. 5. A significant thermocline
between 10 and 15m separates the cold water from the
upper layer in the bay. Some of the isotherms (T = 16–
18°C) bend down and create a sharp bottom front at
the bay mouth, while other (T = 20–21°C), along with
isohalines of 30–32, are curved upwards to the surface. In contrast to the stratification inside the bay, the
water column in the Irago Strait is strongly mixed. The
distribution of DO was similar to that of temperature.
There is a strong DO front at the bay mouth and the
DO level in the cold dome is substantially reduced with
a minimum <10% of saturation. This strong depletion
of DO within the dome indicates that the bottom water
was isolated from the surrounding water masses, in a
similar way to the domes on the shelf regions or fjords.
These hydrographic and DO features resemble the schematic views shown in Fig. 3b.
However, the formation mechanism of hypoxic water, namely the cold dome, in Ise Bay is different from
that on the shelf or fjords because we are dealing with
a system in which buoyancy induced by freshwater
input is an important part of the problem. In the con-
doi:10.5047/absm.2014.00704.0117 © 2014 TERRAPUB, Tokyo. All rights reserved.
A. Kasai / Aqua-BioSci. Monogr. 7: 117–145, 2014
121
drowned river valley estuaries and bays have complicated topography, which changes the circulation pattern (Wong 1994; Wong and Münchow 1995; ValleLevinson and Lwiza 1995), so that the current structure in bays is essentially three-dimensional. Figure 5
indeed indicates that the prominent bottom front would
prevent the intrusion of the oceanic water into the lower
layer, and that circulation in Ise Bay would be different from the classical estuarine circulation. The movement of the intruding oceanic water is thus essentially
important for the formation of the cold dome.
In this section, therefore, we show the mechanism
responsible for the formation of the cold dome and
hypoxia. Focusing on the cold dome and the intruding
oceanic water, detailed observational results of threedimensional temperature, salinity and velocity fields
in Ise Bay are presented. Based on the observational
results, the principles which govern the formation of
hypoxia are demonstrated using a numerical model
applied to an idealized condition. The key point for
understanding the mechanism is the existence of a region of strong vertical mixing which maintains a well
mixed condition next to a stratified bay.
3-2. Cold hypoxic dome in Ise Bay
Fig. 5. Longitudinal distributions of temperature (°C), sa-
linity, density (sigma-t unit) and dissolved oxygen concentration (%), observed on 17 June 1997. Reprinted from Cont.
Shelf Res., 22, Kasai et al., Circulation and cold dome in a
gulf-type ROFI, 1579–1590,  2002, with permission from
Elsevier.
ventional theory on regions of freshwater influences,
two-dimensional (vertical-longitudinal) estuarine circulation develops with upper layer freshwater flowing
seaward while lower layer salty water flows landward.
If this estuarine circulation were dominant, the inflow
water would replace the bottom water so that the hypoxia would not persist. The dynamics are further complicated by the influence of the Earth’s rotation if the
horizontal scale of a bay is large compared with the
internal Rossby radius of deformation (Kasai et al.
2000). Ise Bay, for example, has a width of ~30 km
comparing with 2–7 km of the radius. In addition, many
The detailed observations were carried out on August 28–30, 1995, along three east-west lines (Fig. 4).
Observed temperature, salinity, density and residual
current distributions across the section are shown in
Fig. 6. The panels show views looking up-estuary, with
east being to the right. In the figure, the distributions
indicate that all transects were strongly stratified. Temperature predominantly controls density in the lower
layer, although both salinity and temperature contribute to the density structure in the upper layer. Figure 7
shows the T–S diagram, which supports the dependence of density on temperature in the lower layer, as
the salinity changed little under 20 m depth. Since the
lower layer is substantially important for the cold dome,
temperature structure rather than salinity is mainly focused on hereafter. In the all sections, the thermocline
was around 10 m deep west of the section, while it
bifurcated in the eastern part of the bay (Fig. 6). The
upper half of the thermocline (T > 23°C) was still lying horizontally in the eastern part, but the lower half
(T = 20–21°C) bent down and generated a prominent
bottom front. Under the thermocline there was a cold
(T < 20°C) water mass, which extended from the middle to the western lower (deeper than 15 m) basin in
each line.
Vertical profiles of temperature in the eastern and
western part of the line C and at the Irago Strait are
presented in Fig. 8. In the western area of the bay (C5),
a sharp thermocline between 8 m and 13 m depth separated the upper and lower layers. The temperature dif-
doi:10.5047/absm.2014.00704.0117 © 2014 TERRAPUB, Tokyo. All rights reserved.
A. Kasai / Aqua-BioSci. Monogr. 7: 117–145, 2014
122
Temperature (oC)
Salinity
A12
0
0
28
28
31
32
Depth (m)
24
33
20
22
20
20
18
Line A
Line A
40
40
B17
Depth (m)
0
26
0
24
32
31
22
33
20
20
20
18
34
Line B
Line B
40
40
C5
0
C13
0
26
31
32
32
Depth (m)
33
24
22
20
20
Line C
40
0
10
20
20
30
Line C
40
Distance (km)
40
0
10
20
30
40
Distance (km)
Fig. 6. Transversal distributions of temperature, salinity, density and north (Line A) or northwest (Lines B and C) components
of residual currents along the three lines. Positive and negative values of residual currents indicate inflow and outflow respectively, in the velocity panels. Reprinted from Cont. Shelf Res., 22, Kasai et al., Circulation and cold dome in a gulf-type ROFI,
1579–1590,  2002, with permission from Elsevier.
ference reached 1.5°C m –1. A weaker but prominent
thermocline (∆T ~ 0.7°C m–1) was also found at the
same depth in the eastern part (C13), although the
western upper (lower) layer was much warmer (colder)
than the eastern. In both areas, the upper and lower
layers were relatively uniform. On the other hand, no
explicit thermocline was observed at the Irago Strait,
indicating the water column was mixed. The temperature at the Irago Strait, especially deeper than 13 m,
was similar to that at C13, although the thermocline
was indistinct compared with that at C13. This feature
is also revealed in the T–S diagram (Fig. 7). Temperature and salinity of the water shallower than 20 m at
the Irago Strait were between 21°C and 24°C and between 32.5 and 33.8, respectively. In the eastern part
of each line (A12, B17 and C13), the water located
between the upper and lower thermocline had the same
character as the Irago Strait water, indicating that the
shallow water at the Irago Strait intruded into the eastern middle layer of the bay. The deeper Irago water
belonged to another group, which was more saline (S
> 33.8) and dense (σ t > 23.6) water. Notice that the
bottom water from the middle to the western basin (C5)
had an explicitly different character; the temperature
was lower than 20°C, which was explicitly colder than
the eastern middle layer (C13). In addition, it is indicated that this cold water mass extended from lines A
to C at least, as the bottom waters in the all lines occupy the same position in the diagram (Fig. 7).
In Fig. 6, the temperature from 20°C to 24°C, which
was the same as that at the Irago Strait, and inflowing
areas faster than 10 cm s–1 are shaded in the temperature and velocity panels, respectively. The area between
20°C and 24°C was thick in the eastern part. As the
bifurcation of the thermocline, it corresponded to the
inflowing area. In line C, the strongest middle layer
inflow was ~17 cm s–1 just above the tilting bottom
front, while near the seabed the flow was ~4 cm s–1 to
the southeast. The difference ∆u ~ 21 cm s–1 is approximately consistent with the thermal wind relation, as
doi:10.5047/absm.2014.00704.0117 © 2014 TERRAPUB, Tokyo. All rights reserved.
A. Kasai / Aqua-BioSci. Monogr. 7: 117–145, 2014
123
Residual current (cm s-1 )
Sigma-T
A12
Depth (m)
0
0
19
20
21
22
0
10
23
20
20
24
Line A
Line A
40
40
B17
0
21
20
0
21
10
Depth (m)
22
23
20
20
0
24
Line B
Line B
40
40
C5
0
Depth (m)
20
19
C13
21
20
0
21
0
10
-10
22
0
0
23
20
20
24
Line C
40
0
Line C
10
20
30
40
0
40
Distance (km)
10
20
30
40
Distance (km)
Fig. 6. (continued).
∆ρ ~ 0.4 kg cm –3 in a distance of 5 km with a depth of
20 m would lead ∆u ~ 22 cm s–1. The inflowing velocities in the other two lines were also observed above
the front and consistent with the relation. This indicates that the well-mixed strait water intruded along
the eastern coast as the geostrophic flow. In contrast,
the flow speed was weak in the western lower layer. A
prominent bottom front separated the eastern brisk intrusion water from the western stagnant cold dome.
This flow pattern, namely the inflow in the eastern side
of the bay with the maximum in the middle layer, is
considerably different from the conventional two-layer
estuarine circulation. Since the low temperature
(<20°C) was observed nowhere else in the bay in this
season (Figs. 6, 7) but existed in spring, the water of
the cold dome would be a relict of cold spring water,
which would be trapped beneath the thermocline.
3-3. Numerical model
From the detailed observations there are strong indications that the strait water intrudes along the eastern coast and relict water in the western lower layer
forms the cold dome. The formation of the cold dome
by this process can be demonstrated using a suitable
model with two different density structures; a stratified and a mixed region (Kasai et al. 2002).
The model is applied to a rectangular bay of 38 km
width and 64 km length, corresponding to Ise Bay (Fig.
9). To reproduce circulation and density structures in
a gulf next to a well-mixed area in summer situation,
the water is stratified with low vertical eddy viscosity
and diffusivity (=10–5 m2 s –1) in the bay (x < 50 km),
while the water is homogeneous with high viscosity
and diffusivity (=10–3 m2 s–1) in the mixed region (x >
50 km), which corresponds to the Irago Strait. The
water in the bay is stratified in the initial stage but no
freshwater input is induced, because the movement of
surface layer fresh water is dispensable. Typical temperature and salinity values in summer, with reference
to those shown in Fig. 5, were chosen as the initial
values; surface temperature and salinity inside the bay
are 22°C and 30 and those at the bottom are 15°C and
33.5, respectively. Initial temperature and salinity in
the mixed region are 18.0°C and 31.7, which are the
same as those in the middle layer of the bay. Both tem-
doi:10.5047/absm.2014.00704.0117 © 2014 TERRAPUB, Tokyo. All rights reserved.
A. Kasai / Aqua-BioSci. Monogr. 7: 117–145, 2014
124
28
20
21
z
22
Temperature ( C)
0 - 10 m
y
10 - 20 m
26
++
+
20 - 30 m
23
++
o
x
+
Stratified area
Small Kz & Az
30 m -
24
+
+
22
+
+
+
+
20
24
++
+++
+
+
++
+
+
31
++
+
32
33
A12
B17
C5
C13
Irago
Mixed area
Large Kz & Az
Bay
Strait
34
Salinity (psu)
Fig. 7. Temperature-salinity diagram at A12, B17, C5, C13
and the Irago Strait. Larger symbols indicate deeper areas.
See Figs. 3 and 6 for the observational positions. Reprinted
from Cont. Shelf Res., 22, Kasai et al., Circulation and cold
dome in a gulf-type ROFI, 1579–1590,  2002, with permission from Elsevier.
Fig. 9. A schematic view of the model basin. Reprinted from
Cont. Shelf Res., 22, Kasai et al., Circulation and cold dome
in a gulf-type ROFI, 1579–1590,  2002, with permission
from Elsevier.
Fig. 8. Vertical profiles of temperature at C5, C13 and the
Irago Strait. See Figs. 3 and 6 for the observational positions. Reprinted from Cont. Shelf Res., 22, Kasai et al., Circulation and cold dome in a gulf-type ROFI, 1579–1590, 
2002, with permission from Elsevier.
perature and salinity are uniform in the ydirection. Velocity is set to be zero in the whole region.
Figure 10 shows the model results indicating the
horizontal distribution of temperature and velocity in
the middle layer (23 m depth) after 25 days, when the
field reaches nearly steady state. The mixed water flows
into the stratified region in the center (y = 14–24 km)
and is forced towards the right wall facing landward
as it intrudes into the bay. This inflow is ~15 cm s–1
and parallel to the isotherms (namely isohalines and
Fig. 10. Simulated horizontal distributions of temperature
and velocity at 23 m depth after 25 days. Reprinted from
Cont. Shelf Res., 22, Kasai et al., Circulation and cold dome
in a gulf-type ROFI, 1579–1590,  2002, with permission
from Elsevier.
isopycnals), suggesting that it is in the geostrophic
balance. At the warmer side of the front (x = 30–50
km, y > 20 km) temperature is ~17.5°C, which is close
to that in the mixed region (T = 17.5–18°C). On the
other side of the front, the water is cold with a minimum of 16°C. This contrast of temperature is consist-
doi:10.5047/absm.2014.00704.0117 © 2014 TERRAPUB, Tokyo. All rights reserved.
A. Kasai / Aqua-BioSci. Monogr. 7: 117–145, 2014
125
Cold Dome
Fig. 12. A schematic view of the intrusion process and the
formation of the cold and hypoxic dome. Reprinted from
Cont. Shelf Res., 22, Kasai et al., Circulation and cold dome
in a gulf-type ROFI, 1579–1590,  2002, with permission
from Elsevier.
Fig. 11. Cross-sectional view of simulated temperature and
velocity at x = 40 km. Note that arrows directed upward in
the velocity panel represent the component in x direction
(inflow), and directed right represent the component in y
direction. Reprinted from Cont. Shelf Res., 22, Kasai et al.,
Circulation and cold dome in a gulf-type ROFI, 1579–1590,
 2002, with permission from Elsevier.
Density (sigma-T)
26
24
5m
10m
20m
30m
B-1m
Irago
22
20
18
J
F
M
A
M
J
J
A
S
O
N
D
J
Month
Fig. 13. Seasonal changes in density in the middle of Ise
ent with the observational results (Fig. 6). The velocity in this cold water mass is significantly weak (u ~ 1
cms –1) except for the vicinity of the wall, where there
is return flow.
Figure 11 shows the cross-sectional views of temperature and velocity, facing landward. The water is
stratified in the left half of the model basin. The domelike water in the left deep layer is the coldest, while
the surface water is warmest in the section. On the other
hand, the homogeneous water, which has the same temperature (T = 17–18°C) as that in the mixed region,
exists in the right. The velocity panel clearly shows
that inflow area concentrates in the right middle and
upper regions (y = 17–30 km, z > 30 m), while the
outflow is detected in the upper layer near the left wall.
The velocity adjacent to the coldest region is appreciably weak. Both the temperature and velocity sections
resemble observational results (Fig. 6) in combination
of the stagnant cold dome and intruding mixed water.
The model seems effective in producing both temperature distributions and current structures, and satisfactorily demonstrates the key physical principles detected
by the observation.
Bay and at the Irago Strait (20 m depth). “B-1 m” indicates
“1 m above the bottom”. Reprinted from Cont. Shelf Res.,
22, Kasai et al., Circulation and cold dome in a gulf-type
ROFI, 1579–1590,  2002, with permission from Elsevier.
3-4. Cold dome and hypoxia
The key point to control the mechanism is the coexistence of stratified and mixed areas. Figure 12 illustrates a schematic view of the density and flow structure in the bay. In the stratified season, the mixed strait
water does not intrude along the bottom but enters
through the middle layer of the bay. The Earth’s rotation forces the inflow to the right shore side looking
landward in the northern hemisphere. Therefore, the
water outside the intrusion is isolated because of the
insufficient water exchange. The heating effect cannot
reach the bottom water in the stratified area, so that
spring water has been left cold.
The mechanism responsible for the formation of the
seasonal dome is the summer-time production of horizontal density gradients between the stratified and
mixed waters. According to the stratificationcirculation diagram proposed by Hansen and Rattray
doi:10.5047/absm.2014.00704.0117 © 2014 TERRAPUB, Tokyo. All rights reserved.
126
A. Kasai / Aqua-BioSci. Monogr. 7: 117–145, 2014
Fig. 14. Bathymetry of the study area. Hydrographic observations were carried out at the solid circles. Velocities were mea-
sured along the solid line with an ADCP. Water samples for the Oxygen consumption rate experiments were taken from
Hiuchi-nada (Stns. H7, H12 and H15), Harima-nada (Stn. H) and Bisan-seto (Stn. B). BS and SP denote Bisan-seto and the
Shonai Peninsula, respectively. Reprinted from Coastal and Shelf Science, 71, Kasai et al., Flow structure and hypoxia in
Hiuchi-nada, Seto Inland Sea, Japan, 210–217,  2007, with permission from Elsevier.
(1966), Ise Bay is classified in the strongly stratified
regime from April to October, while during the rest of
the period it is in the weakly stratified regime. Figure
13 shows the seasonal change in water density at five
depths in the middle of the bay and at 20 m depth in
the Irago Strait. In autumn and winter (November–
March) temperature is reversed and weak stratification
is maintained by the only freshwater buoyancy input.
Sea surface cooling sinks the freshwater through the
mixing process, so that the water in the bay is lighter
than that in the strait. During the heating season (April–
October), on the other hand, both large freshwater discharge and sea surface heating makes the water stratified inside the bay. The effect of the heating from the
surface extends to the bottom in the well-mixed area
(Irago Strait), while it is restricted to the surface layer
by the strong thermo- and haloclines in the stratified
area (inside the bay). This procedure makes the density of the bottom bay water larger than that of the strait
water. The strait water is, therefore, expected to go into
the bottom layer in winter but into the middle layer in
summer, when it intrudes into the bay (Fig. 3). In the
transition season, the external factors such as freshwater discharge and wind stress change the strength of
the stratification and could lead the alternation of the
intrusion depth (Takahashi et al. 2000).
The cold dome is strongly related to the hypoxia as
shown in Fig. 5. By inhibiting replacement of the water, DO is not supplied to the dome. Therefore, the
oxygen consumption exceeds the supply, leading to
hypoxia in the cold dome. Hypoxia is observed at several points in the Seto Inland Sea every summer (Fig.
1a). The tidal currents are weak and the water tends to
be stratified in those areas (Fig. 1c). However, the water is well mixed in the straits, which are next to the
stratified areas. This indicates the same physical
mechanism controls DO concentration and causes hypoxia in the Seto Inland Sea.
4. Physical vs. biochemical processes controlling
the formation of hypoxia
4-1. Introduction
As shown in the previous sections, physical processes are important for the formation of hypoxia. In
this section, we present experimental results on the
oxygen consumption over the bottom sediments and
hydrographic observations both in hypoxic and nonhypoxic water. It will be clarified going through the
process, that the contribution of physical processes
outweighs that of biochemical processes in the formation of hypoxia in a coastal basin.
Hiuchi-nada is situated in the central part of the Seto
Inland Sea, Japan (Fig. 14). It is well known that in
the eastern part of the Hiuchi-nada the oxygen in the
bottom layer is depleted every summer (Fig. 1a). The
serious hypoxia has often damaged fisheries since
1960s in this area (Ochi and Takeoka 1986). The historical view on the formation mechanism of hypoxia
is described as follows. Several pulp factories were
constructed in 1960s–70s in Kawanoe and Iyomishima,
which are on the southeastern coast of Hiuchi-nada
(Fig. 14). The pulp factories loaded a large amount of
organic and inorganic matter into the southern area of
Hiuchi-nada. In the eastern part of Hiuchi-nada, there
doi:10.5047/absm.2014.00704.0117 © 2014 TERRAPUB, Tokyo. All rights reserved.
DO Concentration (mg/ L)
A. Kasai / Aqua-BioSci. Monogr. 7: 117–145, 2014
6
4
Hiuchi-nada
Harima-nada
Bisan-seto
2
0
0
Fig. 15. Time sequence of minimum DO concentration in
the east part of Hiuchi-nada. Reprinted from Coastal and
Shelf Science, 71, Kasai et al., Flow structure and hypoxia
in Hiuchi-nada, Seto Inland Sea, Japan, 210–217,  2007,
with permission from Elsevier.
is a cyclonic circulation (6th Regional Coast Guard
Headquarters 1973), which transports the
allochthonous and autochthonous organic matter to the
northeast. En route this organic matter sinks and, therefore, accumulates at the bottom. Oxygen is thus consumed during the decomposition of this organic matter at the bottom (Ochi 1992).
For this reason, various measures have been taken
to reduce the organic matter load from the factories
since the early 1970s. These antipollution measures had
an effect on water quality such that chemical oxygen
demand (COD) in the region rapidly decreased in the
late 1970s (Ukita 1998). However, the DO condition
has unsatisfactorily remained. Figure 15 shows yearto-year variation in the minimum DO concentration in
the eastern part of Hiuchi-nada. Unlike the other chemical components, DO concentration in the bottom layer
has stayed at a low level. It has increased gradually in
recent years, but is still lower than 4 mg l–1. This situation indicates that another mechanism could affect the
oxygen depletion in Hiuchi-nada.
Therefore, we measured oxygen consumption rates
at the bottom in the eastern part of Hiuchi-nada. Comparing the results with those in the other regions, where
hypoxia has never been observed, allows the importance of biochemical processes to be evaluated. To estimate the effect of oxygen supply by the physical processes on the hypoxia in the region, extensive CTD and
ADCP observations were also conducted. Water exchange by horizontal advection in the bottom layer was
evaluated based on the observational results. Combining the results of oxygen consumption experiments and
field observations, it will be discussed whether biochemical or physical processes contribute more to the
formation of the hypoxia in Hiuchi-nada.
4-2. Characteristics of Hiuchi-nada
The study area, Hiuchi-nada, is located in the center
of the Seto Inland Sea, Japan (Fig. 14). As is shown in
127
0.05
0.1
0.15
DO Consumption rate (g/m2/h)
Fig. 16. Relation between the DO consumption rates and in
situ DO concentrations at the bottom. Reprinted from Coastal
and Shelf Science, 71, Kasai et al., Flow structure and hypoxia in Hiuchi-nada, Seto Inland Sea, Japan, 210–217, 
2007, with permission from Elsevier.
Fig. 1, the Seto Inland Sea is characterized by a combination of narrow strait and wide basin. In the western part of Hiuchi-nada the geography is very varied
and includes many small islands, while in the eastern
part the coastline is smooth (Fig. 14c). The complicated topography leads to stronger tidal currents over
1 m s–1 in the western part of the Hiuchi-nada and on
the northern side of the Shonai Peninsula (Bisan-seto),
but typical tidal currents are less than 0.1 m s–1 in the
eastern part. The strong tidal currents result in relatively mixed water, whereas the weaker current makes
the water in the east strongly stratified in summer. Hypoxia has often been observed in the eastern lower
layer, but never in the west. No large rivers empty into
Hiuchi-nada, so that temperature predominantly controls density. The fragmentary data set by observations
is suggestive of cyclonic and anti-cyclonic circulation
in the surface eastern and western part of Hiuchi-nada,
respectively (6th Regional Coast Guard Headquarters
1973).
4-3. Oxygen consumption rate experiments
Sediment cores were taken from the surface to a depth
of 20 cm using acrylic core tubes from three stations
in Hiuchi-nada (Fig. 14c). After taking the cores to
the laboratory, the water in the cores was replaced with
the filtered oxygen-saturated water, taking care not to
disturb the surface of the sediment. The top of each
core was completely capped to exclude air bubbles,
and they were incubated in a 20°C water bath in a dark
room. The concentration of dissolved oxygen in the
water was measured with an oxygen electrode inserted
into each core tube. The water in the cores was gently
stirred by the stirrer equipped with the electrode, and
DO data were continuously monitored for more than
14 hours. As DO gradually decreased and converged
to a certain level within 10 hours, DO consumption
rate R(t) was estimated by the decline of DO.
doi:10.5047/absm.2014.00704.0117 © 2014 TERRAPUB, Tokyo. All rights reserved.
128
A. Kasai / Aqua-BioSci. Monogr. 7: 117–145, 2014
Fig. 17. Horizontal distributions of temperature and DO observed at the bottom in 2002. Reprinted from Coastal and Shelf
Science, 71, Kasai et al., Flow structure and hypoxia in Hiuchi-nada, Seto Inland Sea, Japan, 210–217,  2007, with permission from Elsevier.
To compare the results from Hiuchi-nada with the
other oxygen rich areas, the same experiments were
conducted using the waters from Harima-nada and
Bisan-seto, where hypoxia has never been observed
(Fig. 14b).
Measured
DO
concentration
decreased
monotonically in the three samples from Hiuchi-nada.
The DO consumption rates in Hiuchi-nada were from
0.032 to 0.040 g m–2 h–1. Ochi and Takeoka (1986) and
Hoshika et al. (1989) measured in situ DO consumption rates using a bell-jar type chamber in the eastern
part of Hiuchi-nada. Their results were 0.02 and 0.03
g m–2 h–1 on average, respectively, which are slightly
lower but still comparable to ours.
On the other hand, the Bisan-seto samples showed
considerably high consumption rates (>0.05 g m –2
h–1), although the in situ DO concentration is high. Unfinished feeds from adjoining fish farms accumulate
at the bottom in Bisan-seto, where aquaculture is well
developed. This large amount of organic matter would
result in the highest DO consumption. The rates measured in Harima-nada were from 0.023–0.034 g m–2
h–1, which are comparable to those in Hiuchi-nada.
The DO consumption rates should be correlative to
the in situ DO concentration, if the strength of DO
consumption controls generation of hypoxia at the bottom. However, the relationship demonstrates that there
is no significant correlation between them (Fig. 16, r2
< 0.02). This indicates that the main reason for the generation of hypoxia in Hiuchi-nada is not the high DO
consumption rate (not biochemical processes), but
other processes.
4-4. Hydrographic survey
Field surveys consist of hydrographic observations
over a wide range in 2002 and current observations
along a line in 2003. The hydrographic observations
were carried out in the eastern part of Hiuchi-nada on
21 June, 19 July, 16 August, and 19 September, 2002.
Observational stations are shown in Fig. 14c. Vertical
profiles of temperature and conductivity were obtained
using a CTD profiler. DO was measured using an oxygen meter at 5 m depth intervals, and at every 1 m depth
intervals around oxyclines (usually near the bottom).
Currents were measured by a shuttling operation with
doi:10.5047/absm.2014.00704.0117 © 2014 TERRAPUB, Tokyo. All rights reserved.
A. Kasai / Aqua-BioSci. Monogr. 7: 117–145, 2014
H6
H5
0
H7
H6
H8 H5
129
H7
H8
24
Depth (m)
32.5
22
10
32.7
20
20
18
(a) Temp.(oC)
(b) Salinity (psu)
30
0
Depth (m)
22
10
6
5
23
20
4
(d) DO (mg/L)
(c) Density ( t)
3
30
0
5
10
Distance (km)
15
0
5
10
Distance (km)
15
Fig. 18. Vertical distributions of (a) temperature, (b) salinity, (c) density and (d) DO along the east-west section on 21 June
2002. The panels show views looking northward with east being to the right. Triangles indicate observational points. Reprinted from Coastal and Shelf Science, 71, Kasai et al., Flow structure and hypoxia in Hiuchi-nada, Seto Inland Sea, Japan,
210–217,  2007, with permission from Elsevier.
an ADCP for about 12 hours on 3 August 2003 along
the east-west line in the eastern part of Hiuchi-nada
(Fig. 14c). The raw data was analyzed by a harmonic
method to obtain tidal residual currents. Temperature
and conductivity were measured once at each station
using the CTD profiler. The same CTD observation was
conducted at a point (H1) on the northern side of the
Shonai Peninsula, as a representative of the mixed region.
The strong hypoxia was detected at the bottom in
the eastern part of Hiuchi-nada in 2002. Figure 17
shows the observed horizontal distributions of temperature and DO concentration at the bottom. From June to
August, cold (and dense) water was distributed along
the depression. Strong temperature fronts (~0.2°C
km–1) laid along the western edge of the cold water.
The shape of the cold water is similar in each observation, although the temperature increases 4–6°C from
June to August. Sea surface cooling destroyed the stratification in September, and thus the cold water disappears and temperature became nearly uniform.
Distributions of DO concentration resemble those of
temperature; the minimum DO area was observed from
the center to east of the observational area and hypoxia
corresponds reasonably well to cold water from June
to August. Strong DO fronts (~0.5 mg l–1 km–1) were
created along the temperature fronts. The DO concentration decreased until August, and then recovered in
September by mixing, in the same way as the cold water
mass.
Vertical sections along the east-west line (Fig. 18)
show the water was strongly stratified in the center and
eastern area (5–18 km from the western end of the observational line). Temperature and density resemble
each other in structure, indicating that temperature predominantly controls density. Salinity difference in the
observational section was moderate comparing with
temperature. A strong thermocline was observed at
around 10 m depth and the temperature difference between the surface and the bottom reached 9°C in June.
However, the water at the most western point (H5) was
relatively well mixed and the temperature difference
was less than 3°C. The bottom front (T = 19–20°C)
separated the mixed water from the eastern cold water
with a temperature 2–3°C lower than the mixed water.
The distribution of DO deeper than 15 m depth was
similar to that of temperature in the vertical section,
as shown in the horizontal distributions. DO concentration was over 6 mg l–1 in the upper layer, and lower
than 5 mg l–1 in the cold dome. On the other hand, the
DO concentration was relatively high (>5.5 mg l–1) in
the western mixed area even at the bottom. Similar
characteristics of temperature and DO were also observed in the north-south section (not shown here).
A scatter diagram of observed temperature and DO
(Fig. 19) shows that the oxygen was saturated and
nearly uniform in the surface layer (between 0 m and
5m). However, under the thermocline (deeper than 10
m), DO concentration decreased with temperature.
There were significant correlations between temperature and DO, since the coefficient of determination (r2)
is over 0.7 in each month except for September. The
doi:10.5047/absm.2014.00704.0117 © 2014 TERRAPUB, Tokyo. All rights reserved.
A. Kasai / Aqua-BioSci. Monogr. 7: 117–145, 2014
DO Concentration (mg/ L)
130
The horizontal scale of the target phenomena is about
20 km. In contrast, the Rossby internal radius of deformation is estimated to be several kilometers from
Fig. 18, which shows the thickness of upper and lower
layers is ~10 m and density difference between the layers is ~1 kg m–3 in summer. Since the former is considerably larger than the latter, the Earth’s rotation is
expected to be important for the dynamics of water
circulation. Therefore the first-order dynamical balance
ought to be geostrophic. The vertical shear in the along
front current is related to the cross-front density gradient through the thermal wind relation:
8
6
<5m
>10m
4
2
16
20
24
o
Temperature ( C)
28
∂v
g ∂ρ
,
=−
ρf ∂x
∂z
Fig. 19. Relation between temperature and DO concentra-
tion observed on 21 June 2002. Reprinted from Coastal and
Shelf Science, 71, Kasai et al., Flow structure and hypoxia
in Hiuchi-nada, Seto Inland Sea, Japan, 210–217,  2007,
with permission from Elsevier.
colder water is considered to be older water (relict
spring water) in the heating season (Hill 1993; Kasai
et al. 2002). Therefore, the older water contains lower
DO in Hiuchi-nada, indicating that the physical processes are important for the formation of hypoxia.
Temperature, DO and residual currents observed on
3 August 2003 are shown in Fig. 20. The pattern of
salinity and density (not shown here) is similar to that
of temperature. The thermocline was 4–7 m deep in
the center, and deeper and gentler in the eastern and
western part of the section. A dome-like cold (T < 22°C)
and saline (S > 32.3) water existed under the
thermocline around the center of the observational section. This dome corresponds to low-oxygen water mass
(D.O. < 6 mg l –1). These characteristics are the same
as those observed in the hydrographic survey in 2002.
Residual currents from ADCP records show cyclonic
circulation over the bottom cold dome, with maxima
of 14 cm s–1 southward and 10 cm s–1 northward in the
western and eastern part, respectively. This result is
consistent with the flow pattern estimated by 6th Regional Coast Guard Headquarters (1973). However, the
speed was considerably small within the cold dome
with a maximum <5 cm s–1. The vertical profiles of
DO and speed at the center of the observational line
corresponded to each other (Fig. 21). In the surface
layer, DO and speed were high, while in the bottom
water, they were significantly lower. Both properties
reduced rapidly in 12–19 m depths under the
thermocline. Oxygen was depleted in the stagnant bottom water, suggesting that horizontal advection would
play an important role for generation of hypoxia. It is
considered that little oxygen-sufficient water would be
supplied to the bottom layer.
where v is the along frontal velocity (northward velocity for Fig. 20), z is taken to be vertically upwards, g
is gravitational acceleration, ρ is density, f is Coriolis
parameter and x is the cross frontal co-ordinate (eastward for Fig. 20). Using this equation, geostrophic velocities were calculated from the observed density profiles relative to an assumed level of no motion at the
deepest common depth between adjacent stations. The
result (Fig. 20e) demonstrates the existence of
geostrophic circulation in the upper layer of the stratified eastern Hiuchi-nada. Associated with the
thermocline slope, northward and southward
geostrophic velocities of 10 cm s–1 are detected on the
eastern and western side of the dome over the bottom
front, respectively. In the cold dome, however, the velocity is appreciably weaker (v ~ 1 cm s–1), implying
insufficient water exchange. Estimates of geostrophic
currents (Fig. 20e) are consistent with the observed
flow pattern (Fig. 20c), indicating that the residual currents are in approximate geostrophic balance.
The water in Bisan-seto is well mixed even in summer, because of the strong tidal currents. In addition,
freshwater supply from rivers reduces its density. In
Fig. 20, the area where the density ranges from 20.3 to
21.5 kg m–3, which is same as that in Bisan-seto (H1),
is shaded. This area is limited to the upper layer along
the ADCP line, and is compatible with the cyclonic
circulation. At the most western point of the east-west
section in Fig. 20 (H5), the water shallower than 15 m
had the same character as that at H1, indicating that
comparatively mixed water at H1 flowed into the western side of the cold water. The mixed water contains
sufficient DO, so that the DO concentration is relatively high at H5. On the other hand, the cold dome
water was explicitly different in character; the temperature was lower than 19°C, which was colder than
the water in Bisan-Seto. These features correspond to
the flow pattern as shown in Figs. 20c, 20d, 21. The
cold bottom water is stagnant and isolated from the
surrounding water. The prominent bottom front and the
doi:10.5047/absm.2014.00704.0117 © 2014 TERRAPUB, Tokyo. All rights reserved.
A. Kasai / Aqua-BioSci. Monogr. 7: 117–145, 2014
Depth (m)
0
0
10
131
8
8
10
24
23
20
20
4
6
22
(b)
(mg/ L)
DODO
(mg/L)
Temperature
(oC)
(a) Temperature
( oC)
30
30
0
0
-10
Depth (m)
0
10
10
5
-5
5
20
Residual
Current
(c) Residual
Current
V-comp.
(cm/ s)
V-comp.
(cm/s)
30
20
Residual
Current
(d) Residual
Current
V Vector
t r
30
0
0
10
Distance (km)
10 cm/s
20
Depth (m)
-10
10
10
-5
5
20
0
Geostrophic
current
(e) Geostrophic
current(cm/s)
(cm/ s)
30
0
10
Distance (km)
20
Fig. 20. Vertical distributions of (a) temperature, (b) DO, (c) northward components of residual currents and (d) vectors of
residual currents observed on 3 August 2003. (e) Vertical distribution of estimated geostrophic current (northward component). The panels show views looking northward, with east being to the right. Triangles indicate observational points. Positive
and negative values indicate northward and southward flow, respectively, in (c) and (e). Arrows directed upward represent
northward currents, and arrows directed to the right represent eastward currents in (d). Shaded area indicates the area of the
temperature from 23.5 to 26.5°C, which was same as that on the northern side of the Shonai Peninsula (H1). Reprinted from
Coastal and Shelf Science, 71, Kasai et al., Flow structure and hypoxia in Hiuchi-nada, Seto Inland Sea, Japan, 210–217, 
2007, with permission from Elsevier.
thermocline separate the western southward flow from
the bottom cold dome.
Baroclinic circulation in coastal areas shares many
aspects of its dynamics with tidal mixing fronts. The
cold domes have been observed under the thermocline
in a weak tidal area adjacent to the mixed area (LeFavre
1986; Svendsen et al. 1991; Hill et al. 1994, 1997;
Kasai et al. 2002). The formation mechanism of the
dome was clarified by Hill (1993) using a simple dynamical model. In a region of weak tidal currents the
water stratifies during the heating season, even though
the surrounding area is well mixed by strong tides. The
water under the thermocline is warmed very slowly by
the weak diffusion of heat. This relict winter water is,
therefore, isolated from surrounding mixed waters by
horizontal bottom fronts, and develops into a cold
dome. The water over the dome circulates anticlockwise above the bottom front, based on the thermal wind
relation as shown in the aforementioned estimates (Hill
et al. 1994, 1997). Isolation of the bottom water from
doi:10.5047/absm.2014.00704.0117 © 2014 TERRAPUB, Tokyo. All rights reserved.
132
A. Kasai / Aqua-BioSci. Monogr. 7: 117–145, 2014
N
20 km
0
Nagoya
o
35 N
K4
30m
20
M5
m
10
m
K6
o
34 30'N
K7
Irago Strait
o
Fig. 21. Vertical profiles of speed of residual currents and
DO at a position 14 km from the west end of the ADCP line.
Reprinted from Coastal and Shelf Science, 71, Kasai et al.,
Flow structure and hypoxia in Hiuchi-nada, Seto Inland Sea,
Japan, 210–217,  2007, with permission from Elsevier.
the surrounding water is essential: in addition to heat,
oxygen is not transported into the cold dome from the
surroundings, and thus the cold dome becomes hypoxic.
4-5. Conclusions
In order to clarify the formation mechanism of the
hypoxic water mass in Hiuchi-nada, oxygen consumption rate experiments and detailed hydrographic observations were conducted. The oxygen consumption
rate experiments showed the rates in Hiuchi-nada are
less than or comparable to other oxygen rich areas,
indicating that high oxygen consumption is not the
main reason for the formation of hypoxia. On the other
hand, the detailed hydrographic observations demonstrated that an isolated dome is created under the
thermocline in the central and eastern part of the observational area. The water in the dome is significantly
colder than the surrounding water and the current speed
is weak, indicating it is the relict spring water and stagnant. Therefore, the main reason of the formation of
the hypoxia in the eastern part of Hiuchi-nada is not
the high oxygen consumption, but rather the insufficient water exchange in the bottom water.
5. Fortnightly shifts of intrusion depth of oceanic
water into a bay and hypoxia
5-1. Introduction
As we saw in earlier sections, the strength of mixing
136 30'E
o
137 E
Fig. 22. Map of study area. Circles and triangles indicate
observation points by Mie Prefectural Science and Technology Promotion Centre and our observation points, respectively. With kind permission from Springer Science + Business Media: <J. Oceanogr., Fortnightly shifts of intrusion
depth of oceanic water into Ise Bay, 60, 2004, 817–824, Kasai
et al., Fig. 1>.
mainly caused by tidal stirring has a key to control the
bimodal character of the estuarine circulation in a bay.
Hypoxia in coastal areas is greatly affected by the intrusion of mixed water (oxygen-rich water). This process controls the transport and dispersal of DO, so an
understanding and elucidation of the scale and frequency of the intrusion is important, especially for
eutrophic basins. In general, inflow events of oceanic
water into a basin occur over a wide range of
timescales, from days to years (Farmer and Freeland
1983). The variation of tidal stirring with spring-neap
cycles is one of the most plausible phenomena that influence the mixing of strait water (Simpson 1997). In
fjord systems, for example, many observations have
shown that inflows to the deep water change in relation to tidal amplitudes (Griffin and LeBlond 1990;
Allen and Simpson 1998). Some studies have demonstrated enhancement of the inflow by insufficient mixing over the sill during neap tides (Geyer and Cannon
1982). A similar phenomenon may occur in bays, because it is common to both bays and fjords that the
water is stratified in the basin while it is mixed in the
mouth.
From detailed hydrographic and ADCP observations,
Kasai et al. (2000) showed that a strong inflow (~20
cm s–1) toward the north existed just over the deepest
depression in Ise Bay. In Section 3, on the other hand,
we saw that the water that has a similar character to
doi:10.5047/absm.2014.00704.0117 © 2014 TERRAPUB, Tokyo. All rights reserved.
A. Kasai / Aqua-BioSci. Monogr. 7: 117–145, 2014
133
Fig. 24. Relationship between tidal range at Nagoya and
density difference (55–10 m depth) at the Irago Strait (St. 7)
from 15 June to 27 November 2000. With kind permission
from Springer Science + Business Media: <J. Oceanogr.,
Fortnightly shifts of intrusion depth of oceanic water into
Ise Bay, 60, 2004, 817–824, Kasai et al., Fig. 4>.
Fig. 23. Vertical profiles of density at the Irago Strait during
the spring tides (15 June and 14 July, 2000) and the neap
tides (11 August and 6 September, 2000). With kind permission from Springer Science + Business Media: <J.
Oceanogr., Fortnightly shifts of intrusion depth of oceanic
water into Ise Bay, 60, 2004, 817–824, Kasai et al., Fig. 3>.
the Irago Strait water flows into the bay through the
middle layer with a maximum speed of 15 cm s–1. These
studies suggest fortnightly shifts of intrusion depth into
Ise Bay, because the former observation was conducted
during the neap tide and the latter during the spring. It
is therefore expected that the strength of tidal stirring
(tidal strength) varies with spring-neap cycles at the
narrow strait, and that the variation could lead to a
switch from the middle layer intrusion to bottom intrusion of the strait water into the basin.
In this section, we present the observational results
in Ise Bay, and try to clarify the time dependence of
the mixing conditions at the Irago Strait and the time
dependence of the depth of the intrusion of the strait
waters, according to the spring-neap cycles of the tidal
strength. The effect of short time change in
hydrographic condition on the distribution of the hypoxic water mass is discussed.
5-2. Observations
Fisheries Research Division of Mie Prefectural Science and Technology Promotion Centre has measured
temperature, salinity, and DO concentration monthly
in Ise Bay. We used the observed results along the longitudinal section (Fig. 22, open circles). Furthermore,
temperature, salinity and DO concentration were observed twice a month from June 2000 to February 2001
(Fig. 22, closed triangles).
Examples of the vertical profile of density at the Irago
Strait (Stn. K7) are shown in Fig. 23. A strong
pycnocline was always observed in the surface layer
(<10 m) in summer. During spring tides (thick lines),
the water below the pycnocline was well mixed. Especially below 20 m, it was nearly homogeneous. On the
other hand, water was weakly stratified even below the
pycnocline during neap tides (thin lines). Density differences between 10 m and 50 m were smaller during
the spring tides (0.54 and 0.48 kg m–3) than during the
neap tides (1.1 and 1.5 kg m–3). It is clear that mixing
at the Irago Strait was strongly affected by the tidal
strength. This relationship is more readily apparent in
the scatter plot shown in Fig. 24. There is a strong negative correlation between the tidal range on each observational day and the density difference between 10 m
and 55 m depth at the strait, and the coefficient of determination (r2) reaches 0.61 (P < 0.01, 15 June–27
November 2000). The stronger mixing in the middle
and lower layers of the strait is associated with the
stronger tidal currents.
Three consecutive vertical distributions of temperature, salinity, density and DO concentration along the
longitudinal section are shown in Fig. 25. The water
in the basin was always strongly stratified by both surface heating and river discharge. Well-defined
thermoclines and haloclines around 10 m depth separated the upper and lower layers at all observations.
Below the thermocline, however, the water condition
changed completely at every observation. In the first
spring tide (2 August), a cold (<21°C) and saline (>33)
water pool was observed inside the bay. In the strait,
doi:10.5047/absm.2014.00704.0117 © 2014 TERRAPUB, Tokyo. All rights reserved.
A. Kasai / Aqua-BioSci. Monogr. 7: 117–145, 2014
134
2 Aug. (Spring)
Irago
Bay Head
Temp.
(oC)
Irago
Bay Head
0
22
21
23
20
19 21
40
20
23
20
21 22
40
0
32
33
20
34
33.5
33.5
40
0
32.5
33.5
33
32
32.5
33
40
0
0
23
22
20
23.5
22.5
22
24
23
40
D.O.
(mg/l)
24
20
19
19
22
20
Density
(σt)
Irago
Bay Head
0
20
Salinity
22 Aug. (Spring)
11 Aug. (Neap)
24
23.5 23
24.5
22.5
0
20
40
0
20
4
2
6
4
6
6
2
2
4
40
0
20
40
60
0
Distance (km)
20
40
60
0
20
Distance (km)
40
20
40
60
Distance (km)
Fig. 25. Vertical distributions of temperature, salinity, density and dissolved oxygen along the longitudinal section on 2, 11
and 22 August 2000. Darker areas indicate cold (<21°C), saline (>33.5), dense (>23) and hypoxic water (<4 mg l –1). With kind
permission from Springer Science + Business Media: <J. Oceanogr., Fortnightly shifts of intrusion depth of oceanic water
into Ise Bay, 60, 2004, 817–824, Kasai et al., Fig. 5>.
on the other hand, the water was mixed under the
thermocline and halocline, and temperature and salinity were higher than 21°C and less than 33, respectively. A prominent bottom front created at the bay
mouth (45–55 km from the bay head) separated the
strait water from the dense bottom water pool inside
the bay. The density deeper than 15 m at the Irago Strait
was between 22 and 22.3σt and equivalent to that of
the middle layer, but explicitly less than the bottom
layer inside the bay, indicating that the strait water did
not intrude through the bottom, but through the middle layer. In contrast, during neap tide (11 August),
the densities deeper than 15 m at the Irago Strait and
inside the bay were similar to each other. Cold (<20°C)
and saline (>33.5) water was distributed widely from
the bay mouth to the middle of the bay with a depth of
20 m in the lower layer. The oceanic water apparently
intruded through the bottom layer. It appears from the
salinity distribution that the oceanic water reached 30
km from the bay head. The bottom cold water in the
north (15–30 km) would be the water that was widely
distributed in the basin on 2 August (20–40 km) and
be forced to the bay head by the oceanic water, because their temperature (18–19°C), salinity (33.3–33.6)
and DO (<2 mg l–1) were similar. The weaker tidal
current insufficiently mixed the strait water, as shown
in Fig. 23, and thus the bottom water at the bay mouth
became heavier than that inside the bay. It is considered that the density difference at the bay mouth leads
to the downward shift of the intrusion depth of strait
water. The hydrographic condition returned to its initial state during the second spring tide (22 August).
Water at the strait was well mixed again. Saline (>34)
oceanic water disappeared from the observational section and the bottom front with a density difference of
more than 0.3 × 10–3 kg m–4 was recreated at the bay
mouth (50–65 km). The densest water did not exist in
the strait but at the bottom inside the bay, indicating
doi:10.5047/absm.2014.00704.0117 © 2014 TERRAPUB, Tokyo. All rights reserved.
A. Kasai / Aqua-BioSci. Monogr. 7: 117–145, 2014
Sigma-T
30
0
22
Sigma-T
24
20
22
24
Pycnocline
Top
Center
25
Bottom
Average
40
Maximum
20
8m
Stn. K4
32
30
60
34
Irago Strait
(stn. K6 or M5)
Fig. 27. Definition of the intrusion depths of the strait wa-
ter. Intrusion depths are found by comparing the density in
the middle and mouth of the bay. Incoming water is assumed
to flow in the depth of equal density. Top, bottom, and centre of the intrusion depths in the middle of the bay (Stn. K4)
were defined as the depths that had the same densities as
those beneath the pycnocline, at the bottom and their average in the Irago Strait, respectively. With kind permission
from Springer Science + Business Media: <J. Oceanogr.,
Fortnightly shifts of intrusion depth of oceanic water into
Ise Bay, 60, 2004, 817–824, Kasai et al., Fig. 2>.
Salinity
30
22 Aug.
20
25
12m
20
Middle of the
Bay (stn. K4)
25
35m
Stn. K7
20m
25
o
20
20
o
Temperature ( C)
11 Aug.
Temperature ( C)
135
30
32
34
Salinity
into the basin in the middle and bottom layers during
spring and neap tides, respectively.
5-3. Estimates of intrusion depth
Fig. 26. Temperature-salinity diagrams at Stns. K4 and K7
on (a) 11 August and (b) 22 August 2000. Larger (smaller)
marks indicate deeper (shallower) depths. Marks are plotted
every 1 m depth. With kind permission from Springer Science + Business Media: <J. Oceanogr., Fortnightly shifts of
intrusion depth of oceanic water into Ise Bay, 60, 2004, 817–
824, Kasai et al., Fig. 6>.
that strait water would not intrude through the bottom
layer. Modulated by the tidal strength, the hydrographic
condition changed in only 10 days.
This feature is also revealed in the temperaturesalinity diagram (Fig. 26). Since the strait water was
vertically homogeneous on 22 August (spring tide), the
marks (triangles) under the thermocline gather in a
small area in the figure, compared to the rather scattered marks on 11 August (neap tide) when the water
was weakly stratified. On the other hand, the marks
for the central basin (circles) spread in a long line on
both dates. The water under the thermocline in the strait
had a similar character to that in the depth range from
8 m depth to the bottom and from 12 m to 20 m depth
in the central basin on 11 and 22 August, respectively.
The bottom bay water on 22 August was explicitly more
saline and colder than the strait water. The features of
this T–S diagram tell us that the strait water intrudes
The tidal range was used to evaluate the strength of
mechanical stirring by tidal currents, because a larger
tidal range naturally leads to stronger tidal currents.
The tidal range is defined as the difference between
the predicted higher high water and the lower low water in a day. Each constituent tide that is recorded in
“Tidal constituents of the coast of Japan” (Japan Coast
Guard 1983) is used for the prediction of tide.
Estimates of the depth of inflow were made by comparing the density of the strait water to that at the centre of the bay. It is reasonable to assume that the incoming water inflows to the depth of equal density
(Allen and Simpson 1998; Takahashi et al. 2000). From
their analysis of the driving force of density current,
Fujiwara and Yamada (2002) showed that the oceanic
water flows into the Tokyo Bay through the layer where
the inflow water has same density as the ambient water. They explained that the driving force has a maximum at a depth where there is no longitudinal density
gradient between the mouth and the inside of the bay.
As is shown in Section 3, temperature and salinity in
the inflow area are equivalent to those in the Irago Strait
under the thermocline. In addition, the intrusion process of the Irago water through the middle layer, which
has same density as the Irago water, is successfully
demonstrated by a numerical model in Section 3. There-
doi:10.5047/absm.2014.00704.0117 © 2014 TERRAPUB, Tokyo. All rights reserved.
A. Kasai / Aqua-BioSci. Monogr. 7: 117–145, 2014
136
0
Depth (m)
(a)
10
20
30
40
1 Jun.
1 Aug.
200
Tidal Range (cm)
300
250
1 Oct.
Day
1 Dec.
1 Feb.
200
(b)
200
150
Fig. 29. Relationship between tidal range and intrusion depth.
100
50
1 Jun.
200
1 Aug.
1 Oct.
Day
1 Dec.
1 Feb.
200
Fig. 28. Time changes in (a) the intrusion depth and (b) the
tidal range at Nagoya. Open and closed circles indicate middle layer and bottom intrusion, respectively. Error bars in
the upper panel indicate the top and bottom depth of the intrusion layer. With kind permission from Springer Science
+ Business Media: <J. Oceanogr., Fortnightly shifts of intrusion depth of oceanic water into Ise Bay, 60, 2004, 817–
824, Kasai et al., Fig. 7>.
fore, the intrusion depth was defined as follows (Fig.
27). First, the densities at three depths (beneath the
pycnocline, at the bottom, and their average as the representative densities of the intrusion water) were selected in the Irago Strait (Stns. K6 or M5) on each observation. Secondly, the top, bottom, and the centre of
the intrusion depths in the middle of the bay (Stn. K4)
were defined as the depths that had the same densities
as those beneath the pycnocline, at the bottom, and their
average in the Irago Strait, respectively. The strait water
was assumed to intrude through the bottom layer when
the density below the pycnocline in the strait was larger
than that at the bottom in the middle of the bay.
Figure 28 shows time changes in the intrusion depth
and the tidal range. In winter (from late November to
February), the strait water was heavier than the bay
water and then always intruded through the bottom
layer. The strait water tended to intrude through a shallower layer in summer than in winter. In summer, however, the intrusion depth shifts frequently in relation
to the tidal strength. Bottom intrusion, which is shown
as open circles in the figure, occurred only in the neap
tides. The strait water intrudes into the deeper layer
during the neap tides than the spring tides in summer.
A significant correlation (r 2 = 0.74, P < 0.001) was
found between the intrusion depth and the 2-day average of tidal ranges on the day of observation and the
previous day (Fig. 29), showing that intrusion depth is
strongly affected by the tidal strength at the strait.
Data period is from 2 June to 13 September. With kind permission from Springer Science + Business Media: <J.
Oceanogr., Fortnightly shifts of intrusion depth of oceanic
water into Ise Bay, 60, 2004, 817–824, Kasai et al., Fig. 8>.
5-4. Distribution of DO affected by the intrusion
depth
The lowest panels in Fig. 25 show the distributions
of DO concentration. DO is always depleted (<2 mg
l–1) in the bottom water near the bay head (<25 km).
During spring tides, the hypoxic water, which is defined as water with a DO concentration less than 4 mg
l–1, was widely distributed in the lower layer under the
estimated intrusion depths. In addition, the distribution of hypoxia was closely related to the hydrographic
conditions. The contour lines of 4 mg l –1 of DO concentration corresponded well with the 21°C isotherm
in both spring tides. DO concentration was relatively
high in the well-mixed strait water and there was a
strong DO front at the bay mouth, as is shown in the
temperature front. This indicates that the physical processes strongly affect the distribution of hypoxic water.
During the neap tide, on the other hand, the hypoxic
water was forced to the bay head and/or uplifted to the
middle layer by the bottom intrusion of the oxygenrich oceanic water from the strait. The DO pattern was
complicated and fragments of hypoxic water scattered
around the front of oceanic water (25 km and 15–25 m
depth). It appears that the oceanic water intruded landward, mixing with the bay water. The distribution of
hypoxic water changed frequently due to the shifts in
the intrusion of strait water. When waters from the strait
intrude through the middle layer (spring tides) hypoxia
develops in the lower layer below the intrusion depth,
while the bottom intrusion reduces the scale of bottom
hypoxia during neap tides.
5-5. Conclusions
From repeated observations, we have clarified the
shift of intrusion depth of oceanic water into Ise Bay.
doi:10.5047/absm.2014.00704.0117 © 2014 TERRAPUB, Tokyo. All rights reserved.
A. Kasai / Aqua-BioSci. Monogr. 7: 117–145, 2014
137
(a) Spring Tide
(b) Neap Tide
Fig. 30. Schematic views of the intrusion of oceanic water
during (a) spring tides and (b) neap tides, in summer. With
kind permission from Springer Science + Business Media:
<J. Oceanogr., Fortnightly shifts of intrusion depth of oceanic water into Ise Bay, 60, 2004, 817–824, Kasai et al.,
Fig. 9>.
The key point that controls the mechanism is the different mixing condition between the stratified and
mixed area. The conclusions of this section are schematically illustrated in Fig. 30. In the heating season,
the water in the Irago Strait is well mixed during the
spring tides and its density is equivalent to that in the
middle layer inside the bay. The strait water therefore
intrudes through the middle layer and the water under
the intrusion is excluded from water exchange and becomes hypoxic (Fig. 30a). During the neap tides, the
strait water is insufficiently mixed so that the density
in the lower layer at the strait is greater than that inside the bay. The oceanic water therefore intrudes
through the bottom layer and hypoxic water is pushed
to the bay head and/or uplifted (Fig. 30b). The intrusion depth of strait water into the bay shifts frequently
according to the tidal strength. The distribution and
scale of hypoxia changes according to the intrusion
process of oxygen-rich strait water. The hypoxic water can be uplifted and the oxygen minimum can be
observed in the middle layer due to the intrusion of
oceanic water to the bottom according to the shift of
the intrusion depth.
6. Nutrient release from hypoxic water
6-1. Introduction
In coastal environments, the dissolved oxygen concentration changes seasonally in response to stratification of the water column. Hypoxia or anoxia develops in the lower part of the water column in summer
Fig. 31. Time change of temperature in the surface (broken
line) and bottom (solid line) water, DO (µM) and DIN ( µM)
in the bottom water at the central point of Ise Bay from 2002
to 2005. DIN indicates NH4+ (solid line), NO3– + NO2– (broken line) and total concentrations (=DIN, gray line). With
kind permission from Springer Science + Business Media:
<J. Oceanogr., Nitrogen isotopic discrimination by water
column nitrification in a shallow coastal environment, 64,
2008, 39–48, Sugimoto et al., Fig. 2>.
(e.g., Jensen et al. 1990; Kasai et al. 2002; Naqvi et
al. 2006). The changing concentration of DO in the
water just above the sediments alters the depth to which
oxygen penetrates into the sediments, and limits the
NO3– supply from nitrification to denitrification within
the sediment (Kemp et al. 1990; Rysgaard et al. 1994;
Caffrey et al. 2003). Jensen et al. (1990) suggested that
nitrification in the sediments may temporarily cease
after rapid sedimentation events and that in such cases
nutrient rich bottom waters may be the chief NO3–
source for sedimentary denitrification. Therefore, it is
necessary to consider temporal and spatial variations
in these pathways both within the sediments and at the
sediment-water interfaces with changing concentrations of oxygen.
NO 3– concentrations in aquatic systems are affected
by both physical and biogeochemical processes. High
primary production increases the availability of organic
matter in marine environments. Increased organic matter loading enhances O 2 consumption, sedimentary
NH4+ regeneration and consequent NH4+ fluxes from
the sediment into the water column (Jensen et al. 1990).
This NH4+ may be oxidized to NO3– by nitrifying bacteria within the water column. Direct measurements
of nitrification rates by 15N tracer techniques have con-
doi:10.5047/absm.2014.00704.0117 © 2014 TERRAPUB, Tokyo. All rights reserved.
138
A. Kasai / Aqua-BioSci. Monogr. 7: 117–145, 2014
Fig. 32. Vertical profiles of DO concentrations, dissolved inorganic N concentrations (NH 4+ and NO3–) and δ15NNO3 at the
central point of Ise Bay on (a) 12 May 2005 and (b) 1 July 2005. With kind permission from Springer Science + Business
Media: <J. Oceanogr., Nitrogen isotopic discrimination by water column nitrification in a shallow coastal environment, 64,
2008, 39–48, Sugimoto et al., Fig. 3>.
sistently shown that nitrification occurs in the lower
portion of the euphotic zone in coastal environments
(Ward 2005). Even if the nitrification rates in a unit
volume of the water column are lower than those in
the sediments, the total volumetric rates would be significant when integrated over the water column (Pauer
and Auer 2000). In shallow coastal environments, however, the quantitative contribution of each process to
the total NO3– generation is insufficiently understood.
Here we report NO3– isotope data from the water
column in Ise Bay, which is one of the most eutrophic
estuaries in Japan. In spring and summer the water is
strongly stratified as a result of heating and large freshwater run-off (Kasai et al. 2004). Bottom water becomes hypoxic and regenerated dissolved inorganic
nitrogen (DIN) accumulates in the lower layer of the
hypoxic water. In addition, there could be a seasonal
transition in the dominant DIN composition from NH4+
in spring to NO3– in summer, but the quantitative contribution of water-column nitrification to the NO 3–
maximum in summer is little understood. We have thus
investigated the seasonal change in DIN composition
in the central part of the bay. The main objective of
this study is to ascertain the occurrence of nitrification in the water column and to elucidate the effects of
nitrification on the δ15NNO3.
6-2. Nutrient concentrations in hypoxic water
Oxygen concentrations in the bottom water at the
deepest point of the bay (Stn. K4 in Fig. 22) change
seasonally; a hypoxic condition develops with thermal
stratification during summer (Fig. 31). The variation
in the dissolved oxygen concentrations is associated
with seasonal changes in the concentrations of DIN
(=NO3– + NO2– + NH4+) in the bottom water (Fig. 31).
Although DIN concentrations increased with decreasing DO concentrations until June, the loss of DIN was
prominent in summer. The decreasing oxygen concentrations may have induced the temporal variation in
DIN composition. There were lags of few months between the NH4+ and NO3– peaks, the former occurring
in spring, while the latter were found in summer (Fig.
31).
Vertical profiles of NH4+ concentrations in the lower
layer showed different trends from those of NO3– concentrations (Fig. 32). In May, high concentrations of
NH 4+, probably originating from mineralization, were
found in the oxic (DO > 100 µM) lower water column.
NO 3– was vertically homogeneous at low concentrations (2–3 µM) and had δ 15N NO3 values of –6.3 to
–10.4‰ with an average of –8.5 ± 2.0‰ below 10 m
depth. On the other hand, NO 3– concentrations increased significantly in the lower water column with
doi:10.5047/absm.2014.00704.0117 © 2014 TERRAPUB, Tokyo. All rights reserved.
A. Kasai / Aqua-BioSci. Monogr. 7: 117–145, 2014
139
Fig. 33. Seasonal variation in observed and calculated concentrations of Chl-a, NH 4+, and NO3– in the middle (10 m depth) and
bottom waters (34 m depth) at the central point of Ise Bay. Symbols and lines present observed and calculated values, respectively. Reprinted from Coastal and Shelf Science, 86, Sugimoto et al., Modeling phytoplankton production in Ise Bay, Japan:
Use of nitrogen isotopes to identify dissolved inorganic nitrogen sources, 450–466,  2010, with permission from Elsevier.
decreasing NH 4+ and DO concentrations by July. NH4+
fall to nearly zero at all depths. Moreover, δ 15NNO3
increased to 8.4 ± 0.7‰ in the hypoxic water (DO <
100 µM) deeper than 20 m. Although concentrations
of NO 3– and DO decreased simultaneously with depth
in the deepest layer between 30 and 35 m, δ15NNO3 was
almost constant with depth (Fig. 32b). A discrepancy
between stoichiometric variation in NO3– with DO and
isotopic variation in NO3– was also observed in the
deepest layer at the other stations. NO 3– concentrations
therefore showed a strong negative correlation with DO
concentrations in the bottom waters in July (r2 = 0.90).
NO 3– in the water column was significantly 15Ndepleted in May (Fig. 32a). Nitrification has been proposed as the main cause of low δ15N NO3 in several
marine environments (Ostrom et al. 1997; Sutka et al.
2004; Reinhardt et al. 2006). NO3– newly generated
by nitrification is significantly depleted in 15N with
respect to the substrate NH4+, because partial nitrification occurs, with a marked isotope effect. The extremely 15N-depleted NO3– found in Ise Bay in May is
ascribed to newly produced NO3– arising from nitrification. By contrast, δ15NNO3 had increased significantly
within the lower water column by July. The inverse
relationship between the concentrations of NO3– and
NH 4+ at the two observation times, in conjunction with
the decrease in DO concentrations over this period,
indicates that NH4+ was completely oxidized and turned
into NO 3– by July.
6-3. Contribution of nutrients in hypoxic water
to primary production
Nutrients are key elements in phytoplankton production and subsequent ecosystem processes in the sea.
Various nutrient sources, including riverine, oceanic,
and regenerated nitrogen, maintain the phytoplankton
production in coastal ecosystems. Of these major nitrogen sources, riverine input affected by
anthropogenically perturbed nitrogen flux from land
areas to the coast has long been considered primarily
responsible for eutrophication in coastal zones. On the
other hand, we showed in the previous Subsection 6-2
that rich organic matter in the sediments and the water
column in eutrophic coastal ecosystems supply a large
amount of regenerated nutrients, which accumulate in
the oxygen-depleted water mass. Moreover, recent
studies have revealed that a large amount of nitrogen
is supplied from the adjacent marginal sea to the coastal
region. Intrusion of nutrient-rich shelf water would
stimulate phytoplankton production in coastal ecosystems (e.g., Yoder et al. 1981; Sugimoto et al. 2009).
The quantitative characterization of DIN sources to
phytoplankton production is essential for the management of coastal areas. Since these complicated processes are strongly non-linearly coupled, studies on the
relationships between nutrient behavior and
phytoplankton production have usually relied on
physical-ecosystem models that include inorganic and
doi:10.5047/absm.2014.00704.0117 © 2014 TERRAPUB, Tokyo. All rights reserved.
A. Kasai / Aqua-BioSci. Monogr. 7: 117–145, 2014
140
(a)
3.0
Consumption/Supply
DIN flux (tN/day)
80
0
-80
-160
-240
(b)
2.5
2.0
1.5
1.0
0.5
1 2 3 4 5 6 7 8 9 10 11 12
Month
1 2 3 4 5 6 7 8 9 10 11 12
Month
Fig. 34. (a) Seasonal variation in the nitrogen budgets. Positive budgets indicate supply of DIN (=NH4+ + NO3–) from the
rivers (thin dashed line) and ocean (solid bold line), while negative budgets mean consumption of DIN by phytoplankton in
Ise Bay (thin solid line). (b) Seasonal variation in the ratio of DIN consumption by phytoplankton to DIN supply from the
outer regions (rivers and ocean). A linear line of 1.0 indicates equilibrium between external DIN supply and internal DIN
consumption. Reprinted from Coastal and Shelf Science, 86, Sugimoto et al., Modeling phytoplankton production in Ise Bay,
Japan: Use of nitrogen isotopes to identify dissolved inorganic nitrogen sources, 450–466,  2010, with permission from
Elsevier.
organic matter transformation and utilization. We developed a unique three dimensional physicalecosystem model coupled with δ 15N to evaluate the
contribution of the three major nitrogen sources (river,
ocean and regeneration) to phytoplankton production.
The ecosystem model has five compartments, namely
NH 4 + , NO 3 – , phytoplankton (PHY), zooplankton
(ZOO) and detritus (DET), as the prognostic variables
being concentrations of nitrogen. The evolutions of
these compartments are described with differential
equations composed of biological source and sink
terms, external loading terms, diffusion terms, and
advection terms. Two other parameters, dissolved inorganic phosphorus (PO 43–) and DO, are calculated
using stoichiometric methods in the model. The oxygen consumption by remineralization in the sediments
is proportional to the oxygen consumption rates based
on the observational results. Comparison between
physical parameters such as velocities, temperature and
salinity calculated by this model and observed parameters confirmed that the model can reproduce physical
processes in coastal areas (Kakehi 2006). Details of
the model are described by Sugimoto et al. (2010).
Figure 33 shows the seasonal variations in observed
and calculated concentrations of Chl-a, NH 4 + and
NO 3– concentrations in the middle and bottom layer at
the central part of Ise Bay. Calculated concentrations
of each compartment show seasonal patterns similar
to observed ones. The calculated concentrations of Chla in the middle layer are high in spring (~4.5 µg L–1)
and autumn (~3.5 µg L–1), while low in winter (~1 µg
L–1). Chl-a maximum in spring amply reproduces the
subsurface chlorophyll maximum in the bay. The concentrations of NH 4+ and NO3– and their seasonal vari-
ations in the middle layer are smaller and lower than
those in the bottom layer. The NO3– concentration in
the middle layer decreases to ~1 µM during the spring
bloom. In the bottom layer, the NH4+ concentration
peaks in spring (~7 µM), while NO 3– concentration
peaks in summer (~12 µM). This temporal shift from
NH 4+ to NO3– in the lower layer clearly displays the
effect of nitrification, indicating the accumulation of
regenerated NO3– in the lower layer during the stratified periods. The model can overall reproduce the observed conditions.
Figure 34a shows the seasonal variation in DIN
(=NH4+ + NO3–) fluxes of the new supply from the river
and ocean, and the consumption rates of DIN by
phytoplankton within the bay. The riverine flux was
calculated as a product of the DIN concentration and
the river discharge. The flux of DIN from the outer
ocean to the inner sea was estimated through the transverse section at the open boundary. The oceanic flux
was computed by integration of northward velocities
multiplied by the concentration in each grid. Consumption rates rapidly increase from ~70 to ~200 tN d–1 with
the phytoplankton bloom in spring (Fig. 33). In this
period, the supply of DIN from the shelf increases from
~30 to ~60 tN d–1 because of the strengthened estuarine circulation. High assimilation continues during
summer (~200 tN d–1), peaking in autumn (~220 tN
d–1). However, the DIN supply from the river and shelf
shows decreasing trends in autumn. In winter, consumption rates rapidly decreases to < 80 tN d–1. Figure 34b presents the ratio of DIN consumption by
phytoplankton to the supply from the river and shelf
(consumption/supply ratio). The ratio is higher than
1.0 throughout the year, indicating that the external
doi:10.5047/absm.2014.00704.0117 © 2014 TERRAPUB, Tokyo. All rights reserved.
A. Kasai / Aqua-BioSci. Monogr. 7: 117–145, 2014
141
Summer
Fig. 35. Vertical distributions of simulated temperature, salinity, the uptake ratio of NO3– to total DIN by phytoplankton (f-
ratio), Chl-a concentration, NH 4+ concentration, NO 3– concentration, δ15NPN, δ15NNH4, and δ 15N NO3 along the longitudinal section in summer (1 August). Reprinted from Coastal and Shelf Science, 86, Sugimoto et al., Modeling phytoplankton production in Ise Bay, Japan: Use of nitrogen isotopes to identify dissolved inorganic nitrogen sources, 450–466,  2010, with
permission from Elsevier.
supply of DIN is insufficient and that the phytoplankton
production largely depends on the regenerated DIN
within the bay. The consumption/supply ratio considerably differs in spring and autumn. The lower ratio in
spring (<1.5) suggests a larger contribution of the external DIN supply from the river and ocean, while the
higher ratio in autumn (~3) suggests the dominant contribution of regenerated DIN. In Ise Bay, physical properties change seasonally as was shown in the previous
sections. This seasonal difference strongly influences
the relationship between phytoplankton and DIN
sources.
As an example, model results in summer are shown
in Fig. 35. Water stratification is strengthened by surface heating and high freshwater discharge in summer.
In the bottom layer, the cold water mass (<18°C) is
isolated from the surrounding waters by the horizontal
thermo- and haloclines and the vertical bottom front.
This cold water mass becomes hypoxic (DO < 3 mg
L–1). The Chl-a concentrations show two prominent
maxima; a surface maximum at the bay head and a subsurface maximum at the bay mouth. As both NH4+ and
NO3– are depleted in the upper layer, the δ15NPHY of
the two Chl-a maxima expresses the δ15N values of
assimilated nitrogen, because there is little isotope
fractionation. Near the river mouth, the slightly high
δ15NPHY (~7‰) and low f-ratio (<0.6) indicate uptakes
of riverine NH4+, which has low δ15N values (<20‰)
compared to those of regenerated NH4+ (>40‰) in the
lower layer. Moving from the bay head to the center of
the bay, phytoplankton tends to use the NO3– instead
of NH4+. On the other hand, at the bay mouth, the
phytoplankton forming the subsurface Chl-a maximum
takes up NO3– as the major nitrogen source. However,
the NO3– source (oceanic and/or regenerated) cannot
be determined because both types of NO3– have similar isotope values in summer. The consumption/supply ratio (~2, Fig. 34b) suggests that the contribution
doi:10.5047/absm.2014.00704.0117 © 2014 TERRAPUB, Tokyo. All rights reserved.
A. Kasai / Aqua-BioSci. Monogr. 7: 117–145, 2014
142
Riverine
Loading
54
Zooplankton=159
Phytoplankton=606
2
94
12
93
22
18
12
Boundary
Flux
37
11
67
24
NH4+ =416
NO3 =3313
Detritus=452
7
54
80
5
3
48
55
66
Sediment
Fig. 36. Annual averages of nitrogen fluxes (tN d–1) and standing stocks (tN) in Ise Bay. The magnitude of flux is visually
shown by the thickness of lines. Reprinted from Coastal and Shelf Science, 86, Sugimoto et al., Modeling phytoplankton
production in Ise Bay, Japan: Use of nitrogen isotopes to identify dissolved inorganic nitrogen sources, 450–466,  2010,
with permission from Elsevier.
of regenerated NO3– is comparable to that of external
DIN supplied from the rivers and shelf. On the other
hand, the high (>0.6) f-ratio indicates the weak contribution of NH 4+ to phytoplankton production in the
upper layer. The δ 15N NH4 in the upper layer is considerably lower than that in the lower layer. This indicates that residual NH4+ is left by the nitrification processes in the lower layer, while the new NH4+ is produced by remineralization in the upper layer. Mino et
al. (2002) found that δ15N values of suspended matter
decrease with the uptake of regenerated NH4+ rather
than the uptake of new NO3– supplied from the deep
layer in the Atlantic Ocean. The relatively low δ15NPHY
values (<10‰) associated with lower f-ratios (<0.6)
indicate that phytoplankton do not take up the residual
NH 4+ by nitrification. They rather take up the newly
generated NH4+ by remineralization in the upper layer.
Fluxes of newly regenerated NH4+ by remineralization
(33.5 tN d –1) and excretion (19.8 tN d–1) in the upper
layer (>12 m) was found to be considerably larger than
fluxes of NH4+ supplied from the lower layer (<12 m)
to the upper layer by advection (3.9 tN d–1) and diffusion (12.3 tN d–1).
Figure 36 shows the annual average nitrogen flux
(tN d –1) of each process and of standing stocks (tN) in
Ise Bay. The nitrogen supply of 32 tN d–1 from the river
is composed of 24 tN d–1 as NO3–, 5 tN d–1 of NH 4+,
and 3 tN d–1 as detritus. The net nitrogen flux from the
ocean reaches 29 tN d–1, which is comparable to the
riverine nitrogen flux. Dissolved forms, NO3– (37 tN
d–1) and NH4+ (11 tN d–1), are supplied from the ocean
to the bay, while particulate forms composed of
phytoplankton (12 tN d–1), zooplankton (2 tN d–1), and
detritus (7 tN d–1) flow out from the bay to the ocean.
A total of 77 tN d –1 of DIN is supplied to the inner bay
as new nitrogen.
In contrast, the nitrogen flux for each process within
the bay is considerably larger than the new nitrogen
flux from the outer regions. Phytoplankton takes up
93 tN d–1 of NO3– and 67 tN d–1 of NH 4+. The total
flux of phytoplankton assimilation (160 tN d–1) is more
than twice the total flux from the river and ocean. This
result means that more than half the amount of annual
phytoplankton production is supported by the regenerated DIN. However, high consumption/supply ratios
(>2.0) occur only in periods from September to November (Fig. 34b), suggesting that regenerated DIN,
which is accumulated in the hypoxic water mass during the stratified season, contributes significantly to
annual phytoplankton production. Because a large
amount of phytoplanktonic nitrogen is transferred to
detritus (94 tN d–1) rather than zooplankton (54 tN
d–1), phytoplankton mortality accounts for more than
70% of the total detritus supply, which is one of the
largest processes in Ise Bay. Sedimentation of detritus
is larger than the NH4+ release from the sediments. Regenerated nitrogen from remineralization (54 tN d–1)
and release from sediments (55 tN d–1) also shows large
flux. These two fluxes account for ~70% of the total
NH 4+ supply, suggesting that they are the most important sources for phytoplankton production. Moreover,
zooplankton excretion accounts for 16% of the total
NH 4+ supply. A large supply of NO3– is nitrification
(80 tN d–1), which accounts for ~60% of the total NO3–
supply. The denitrification loss of 48 tN d–1 is twice as
large as the NO3– supply from the river (24 tN d–1).
doi:10.5047/absm.2014.00704.0117 © 2014 TERRAPUB, Tokyo. All rights reserved.
A. Kasai / Aqua-BioSci. Monogr. 7: 117–145, 2014
The calculated denitrification rate is comparable to that
estimated from the Redfield stoichiometry (42 tN d–1;
Kakehi et al. 2005) and larger than the potential
denitrification rate (19–28 tN d–1; Sugawara 2003).
These results indicate that phytoplankton production
in Ise Bay is mainly dominated by the internal cycle
(DIN assimilation by phytoplankton, mortality of
phytoplankton, remineralization of detritus, release
from the sediments, nitrification, and denitrification)
rather than the external supply (fluxes from rivers and
the ocean). The former is accompanied by oxygen consumption, and thus strongly related to the hypoxia in
the bay.
In Ise Bay, the formation of a cold water mass isolated from surrounding waters during the stratified
periods is key to the large contributions of regenerated nitrogen to phytoplankton production, because the
cold water mass covers a large volume in the lower
layer (e.g., Kasai et al. 2002). The oxygen within the
isolated water mass is mainly consumed by
remineralization of organic matter in the waters and
sediments, and thus a large amount of DIN accumulates in the hypoxic water mass during the stratified
periods (Sugimoto et al. 2008). Consequently, the
standing stock of DIN in the lower layer is high (~2,500
tN) in summer, but low in winter (~1,400 tN). The consumption/supply ratio differs considerably per season
(Fig. 34b). The lower ratio in spring (<1.5) means that
the direct DIN supply from the rivers and ocean is more
important than that from regeneration, while the higher
ratio in autumn (~3) means that regenerated DIN is
the major source for phytoplankton production (Fig.
34b).
Although the riverine DIN stimulates phytoplankton
production and forms a Chl-a maximum in the surface
layer at the bay head, phytoplankton production at the
bay mouth is largely controlled by the estuarine circulation. In spring, the intrusion depth of oceanic water
changes from the bottom to the middle layer. δ15N NO3
distributions clearly show that oceanic NO 3– is transported into the euphotic layer by the middle intrusion
and stimulates phytoplankton production at the bay
mouth. Kasai et al. (2007) also pointed out the importance of the middle intrusion of oceanic water for the
subsurface chlorophyll maximum layer. In autumn, the
intrusion depth of oceanic water changes from the middle layer to the bottom layer. Regenerated NO3–, which
is accumulated in the hypoxic water mass, is vertically
supplied into the euphotic layer. This vertical supply
of regenerated nitrogen induces the maximum DIN
consumption rate (~220 tN d–1, Fig. 34a). These results show that seasonal shifts of the intrusion depth
of estuarine circulation induce spatial and temporal differences in DIN behavior and subsequent
phytoplankton production.
143
6-4. Conclusions
An important aspect of the nitrogen cycle in coastal
environments concerns the source of the nitrogen used
in primary production. Phytoplankton production in Ise
Bay, one of the most eutrophic embayments in Japan,
is supported by external nitrogen derived from rivers
and the ocean, and regenerated nitrogen formed in hypoxic water within the bay. We evaluated the contribution of each source of dissolved inorganic nitrogen
(DIN) to phytoplankton production in Ise Bay. The
ecosystem model revealed that DIN consumption by
phytoplankton exceeds the DIN supply from the rivers
and ocean, indicating that a large amount of
phytoplankton production in Ise Bay depends on regenerated DIN within the bay rather than on newly
supplied DIN from hypoxic water. The intrusion depth
of oceanic water changes from the bottom to the middle layer in spring. Oceanic nitrate is transported into
the euphotic layer by the middle-layer intrusion and
stimulates phytoplankton production at the bay mouth.
The subsurface chlorophyll maximum layer then develops. In autumn, however, the intrusion depth of oceanic water changes from the middle layer to the bottom layer. Regenerated NO3–, which is accumulated in
the hypoxic water mass, is lifted up and supplied to
the euphotic layer. These results imply that
phytoplankton production in Ise Bay is mainly dominated by the internal cycle rather than the external supply.
7. General conclusions
In this article, we reviewed recent studies on the formation mechanism of hypoxic water in the bottom layer
in summer. Although both biological and physical processes control the generation of hypoxia, the physical
processes such as water exchange and convection determine the scale and strength of hypoxia, rather than
the biological processes namely oxygen consumption
(Section 4). The key point for understanding the formation mechanism is the existence of a region of strong
vertical mixing which maintains well mixed condition
next to a stratified area with weak currents (Section
2). The bottom water in the stratified area is not renewed and thus tends to be isolated and hypoxic. Therefore hypoxia occurs in specific places, in which spring
water has been left as a cold dome (Section 3). The
scale of the hypoxia changes according to the relation
between the strengths of stratification and mixing determined by tidal condition and water densities (Section 5). The hypoxia has an enormous effect on the
marine ecosystem through not only the direct negative
impact on the living organisms, but also the nutrient
release and subsequent primary production in coastal
embayment (Section 6).
doi:10.5047/absm.2014.00704.0117 © 2014 TERRAPUB, Tokyo. All rights reserved.
144
A. Kasai / Aqua-BioSci. Monogr. 7: 117–145, 2014
References
6th Regional Coast Guard Headquarters. Public Notice on
the Procedure of the Report related to the Navigation of a
very large vessel, etc. Japan Coast Guard Public Notice
1973; 109.
Allen GL, Simpson JH. Deep water inflows to upper Loch
Linnhe. Est. Coast. Shelf Sci. 1998; 47: 487–498.
Borsuk MR, Stow CA, Luettich RA, Paerl HW, Pinckney
JL. Modelling oxygen dynamics in an intermittently stratified estuary: estimation of process rates using field data.
Est. Coast. Shelf Sci. 2001; 52: 33–49.
Caffrey JM, Harrington N, Solem I, Ward BB.
Biogeochemical processes in a small California estuary.
2. Nitrification activity, community structure and role in
nitrogen budgets. Mar. Ecol. Prog. Ser. 2003; 248: 27–40.
Farmer DM, Freeland HJ. The physical oceanography of
fjords. Prog. Oceanogr. 1983; 12: 147–220.
Fujiwara T, Yamada Y. Inflow of oceanic water into Tokyo
Bay and generation of a subsurface hypoxic water mass.
J. Geophys. Res. 2002; 107: 1–9.
Fujiwara T, Takahashi T, Kasai A, Sugiyama Y, Kuno M.
The role of circulation in the development of hypoxia in
Ise Bay, Japan. Est. Coast. Shelf Sci. 2002; 54: 19–31.
Geyer W, Cannon GA. Sill processes related to deep water
renewal in a fjord. J. Geophys. Res. 1982; 87: 7985–7996.
Gilbert D, Sundby B, Gobeil C, Mucci A, Tremblay GH. A
seventy-two-year record of diminishing deep-water oxygen in the St. Lawrence estuary: The northwest Atlantic
connection. Limnol. Oceanogr. 2005; 50: 1654–1666.
Griffin DA, LeBlond PH. Estuary/ocean exchange controlled by spring-neap tidal mixing. Est. Coast. Shelf Sci.
1990; 30: 275–297.
Hagy JD, Boynton WR, Keefe CW, Wood KV. Hypoxia in
Chesapeake Bay, 1950–2001: Long-term change in relation to nutrient loading and river flow. Estuaries 2004;
27: 634–658.
Hansen DV, Rattray M. New dimensions in estuary classification. Limnol. Oceanogr. 1966; 11, 319–325.
Hill AE. Seasonal gyres in shelf seas. Ann. Geophys. 1993;
11: 1130–1137.
Hill AE, Durazo R, Smeed DA. Observations of a cyclonic
gyre in the western Irish Sea. Cont. Shelf Res. 1994; 14:
479–490.
Hill AE, Brown J, Fernand L. The summer gyre in the western Irish Sea: shelf sea paradigms and management implications. Est. Coast. Shelf Sci. 1997; 44: 83–95.
Hoshika A, Tanimoto T, Kawana K. Oyxgen uptake by
benthic layer in Hiuchi-nada. Water Pol. Bull. 1989; 12:
423–130 (in Japanese).
Hossain M, Sekiguchi H. Influences of oxygen-poor water
on megabenthos communities in Ise Bay, central Japan.
Bull. Jpn. Soc. Fish. Oceanogr. 1996; 60: 237–243.
Inoue H. Water quality. In: Hirano T (ed). Marine Coastal
Environment. 1998; 561–611.
Ishioka H. Changes of haematological constituents of red
sea breams exposed to the ambient oxygen reduction. Bull.
Jpn. Soc. Sci. Fish. 1982; 48: 165–170.
Japan Coast Guard Tidal Constituents of Coast of Japan.
Japan Hydrographic Association, Tokyo. 1983; 172 pp.
Jensen MH, Lomstein E, Sørensen J. Benthic NH4+ and NO3–
flux following sedimentation of a spring phytoplankton
bloom in Aarhus Bight, Denmark. Mar. Ecol. Prog. Ser.
1990; 61: 87–96.
Kakehi S. Seasonal variations in the velocity structure and
nutrient dynamics in eutrophic bays. Kyoto University
Doctoral Thesis 2006; 134 pp.
Kakehi S, Fujiwara T, Yamada H. Seasonal variations in the
nutrients standing mass and nutrient budget of Ise Bay.
Oceanogr. Japan 2005; 14: 527–540.
Kasai A, Hill AE, Fujiwara T, Simpson JH. Effect of the
Earth’s rotation on the circulation in regions of freshwater influence. J. Geophys. Res. 2000; 105(C7), 16961–
16969.
Kasai A, Fujiwara T, Simpson JH, Kakehi S. Circulation and
cold dome in a gulf-type ROFI. Cont. Shelf Res. 2002; 22:
1579–1590.
Kasai A, Fujiwara T, Kimura T, Yamada H. Fortnightly shifts
of intrusion depth of oceanic water into Ise Bay. J.
Oceanogr. 2004; 60: 817–824.
Kasai A, Sugimoto R, Akamine S. Formation mechanism of
subsurface chlorophyll maximum in a coastal embayment.
Umi-to-Sora 2007; 82: 53–60.
Kemp WM, Sampou P, Caffrey J, Mayer M, Henriksen K,
Boynton WR. Ammonium recycling versus denitrification
in Chesapeake Bay sediments. Limnol. Oceanogr. 1990;
35: 1545–1563.
Kuno M. Seasonal variation of sea conditions in Ise Bay.
Bull. Fish. Res. Inst. Mie 1996; 6: 27–46.
LeBlond P, Ma H, Doherty F, Pond S. Deep and intermediate water replacement in the Strait of Georgia. AtmosphereOcean 1991; 29: 288–312.
LeFavre J. Aspects of the biology of frontal systems. Adv.
Mar. Biol. 1986; 23: 164–299.
Mino Y, Saino T, Suzuki K, Maranon E. Isotopic composition of suspended particulate nitrogen (δ15N sus) in surface
waters of the Atlantic Ocean from 50°N to 50°S. Global
Biogeochem. Cycles 2002; 16: 1059; doi:10.1029/2001/
GB001635.
Naqvi SWA, Naik H, Prathihary A, Souza D, Narvekar PV,
Jayakumar DA, Devol AH, Yoshinari T, Saino T. Coastal
versus open-ocean denitrification in the Arabian Sea.
Biogeosciences 2006; 3: 621–633.
Ochi T. Studies on the characteristics of chemical environment and the recycling of nutrient in Hiuchi-nada Sea area.
Memories Fac. Agri., Kagawa Univ. 1992; 58: 88 pp.
Ochi T, Takeoka H. The anoxic water mass in Hiuchi-nada.
Part 1. Distribution of the anoxic water mass. J. Oceanol.
Soc. Japan 1986; 42: 1–11.
Ostrom NE, Macko SA, Deibel D, Thompson RJ. Seasonal
variation in the stable carbon and nitrogen isotope biogeochemistry of a coastal cold ocean environment.
Geochim. Cosmochim. Acta 1997; 61: 2929–2942.
Pauer JJ, Auer MT. Nitrification in the water column and
sediment of a hypereutrophic lake and adjoining river system. Water Res. 2000; 34: 1247–1254.
Rabalais NN, Turner RE, Wiseman WJ. Gulf of Mexico hypoxia, a.k.a. ‘The dead zone’. Annu. Rev. Ecol. Syst. 2002;
33: 235–263.
Reinhardt M, Müller B, Gächter R, Wehrli B. Nitrogen removal in a small constructed wetland: an isotope mass
balance approach. Environ. Sci. Technol. 2006 40: 3313–
3319.
doi:10.5047/absm.2014.00704.0117 © 2014 TERRAPUB, Tokyo. All rights reserved.
A. Kasai / Aqua-BioSci. Monogr. 7: 117–145, 2014
Rysgaard S, Risgaard-Peterson N, Sloth NP, Jensen K,
Nielsen LP. Oxygen regulation of nitrification and
denitrification in sediments. Limnol. Oceanogr. 1994; 39:
1643–1652.
Simpson JH. Physical processes in the ROFI regime. J. Mar.
Syst. 1997; 12: 3–15.
Simpson JH, Sharples J. Introduction to the Physical and
Biological Oceanography of Shelf Seas. Cambridge University Press, Cambridge. 2012; 424 pp.
Sugawara Y. Denitrification rate in the sediments. In: Studies of Ecosystem Recovery for Ise Bay. Mie Pref. Science
and Technology Promotion Center, Japan. 2003; 18–22.
Sugimoto R, Kasai A, Miyajima T, Fujita K. Nitrogen isotopic discrimination by water column nitrification in a
shallow coastal environment. J. Oceanogr. 2008; 64: 39–
48.
Sugimoto R, Kasai A, Miyajima T, Fujita K. Controlling factors of seasonal variation in the nitrogen isotope ratio of
nitrate in a eutrophic coastal environment. Est. Coast. Shelf
Sci. 2009; 85: 231–240.
Sugimoto R, Kasai A, Miyajima T, Fujita K. Modeling
phytoplankton production in Ise Bay, Japan: Use of nitrogen isotopes to identify dissolved inorganic nitrogen
sources. Est. Coast. Shelf Sci. 2010; 86: 450–466.
Sutka RL, Ostrom NE, Ostrom PH, Phanikumar MS. Stable
nitrogen isotope dynamics of dissolved nitrate in a transect
from the North Pacific Subtropical Gyre to the Eastern
Tropical North Pacific. Geochim. Cosmochim. Acta 2004;
68: 517–527.
Suzuki T. Hypoxia. In: Hirano T (ed). Marine Coastal Envi-
145
ronment. 1998; 475–479.
Svendsen E, Satre R, Mork M. Features of the northern North
Sea circulation. Cont. Shelf Res. 1991; 11: 493–508.
Takahashi T, Fujiwara T, Kuno M, Sugiyama Y. Seasonal
variation in intrusion depth of oceanic water and the hypoxia in Ise Bay. Umi no Kenkyu, 2000; 9: 265–271.
Ukita M. Pollution loads into Seto Inland Sea. In: Yanagi, T
(ed). Nature and Environment in Seto Inland Sea. 1998;
178–198.
Valle-Levinson A, Lwiza KM. The effects of channels and
shoals on exchange between the Chesapeake Bay and the
adjacent ocean. J. Geophys. Res. 1995; 100: 18551–18563.
Ward BB. Temporal variability in nitrification rates and related biogeochemical factors in Monterey Bay, California, USA. Mar. Ecol. Prog. Ser. 2005; 292: 97–109.
Wong KC. On the nature of transverse variability in a coastal
plain estuary. J. Geophys. Res. 1994; 99: 14209–14222.
Wong KC, Münchow A. Buoyancy forced interaction between estuary and inner shelf: observation. Cont. Shelf Res.
1995; 15: 59–88.
Yamamoto K, Yamashita H, Wataishi K. Changes of oxygen
consumption in yellowtail saddled weever, scorpion-fish,
Richardson dragonet and tiger puffer under progressive
hypoxia. Aqua. Sci. 1990; 38: 35–39.
Yanagi T. Science of Tidal Front. Koseisha-Koseikaku, Japan. 1990; 169 pp.
Yoder AJ, Atkinson LP, Lee TN, Kim HH, McClain CR. Role
of Gulf Stream frontal eddies in forming phytoplankton
patches on the outer southeastern shelf. Limnol. Oceanogr.
1981; 26: 1103–1110.
doi:10.5047/absm.2014.00704.0117 © 2014 TERRAPUB, Tokyo. All rights reserved.