Kyuji WATANABE

1
A study on dynamics and management for North
Japan Sea sandfish, Arctoscopus japonicus, stock

• Kyuji WATANABE

Tokyo University of Marine Science and
Technology
2
Depletion and recovery of the North Japan Sea
sandfish stock
Fishing moratorium from 1992 to 1994 in fisheries
of Akita
Fishermen of other prefectures (Aomori, Yamagata
and Niigata) harvesting this stock did not
participate in the moratorium and TAC system.
Total allowable catch (TAC) system from 1995
3



Motivation
Akita Prefecture
Korea
Catch(104ton)
Mechanism of dynamics in North Japan Sea stock
and Korean Peninsula stock of sandfish
Management policy to exploit sustainable the
sandfish stocks
4
dredge net (Oct. - Nov.)
5
(No Transcript)
6
Presentation Outline
1. Biology and fisheries of sandfish
2. Data used in the present study
3. Mechanism of the stock-recruitment
relationship of North Japan Sea stock
4. Dynamics and catch forecasting of the Korean
Peninsula stock
5. Synchronous fluctuations in two stocks
6. Different fluctuations in two stock from the
1980s to 1990s
7. Management policy of sandfish ( North Japan
Sea stock )
7
1 Biology (stocks)
8
Biology of North Japan Sea stock
Distribution area
Aomori Pref.
Akita Pref.
Yamagata Pref.
Niigata Pref.
Spawning ground
9
Juvenile inhabits in the depth of 10 m in April.
10

• Young fish migrate in the depth of 300 m from June

11
Sandfish fisheries in Akita Prefecture
Offshore fisheries
Danish seine small-scale Danish seine boats
Coastal fisheries
set and gill nets
12
Fisheries

• Spawning grounds
• Set net and gill net

Offshore areas Danish seine and trawl boats
13
Catches
3
Catch(104ton)
14
Catches at stock
Catch(104ton)
North Japan Sea
Korean Peninsula
15
2. Data
North Japan Sea
Korean Peninsula
Catch
Catch

• Frequency distribution of BL

Mean weight
Eastern waters
Station 1 off Akita
0-300m
0-300m
16
3. Mechanism of the stock-recruitment
relationship of North Japan Sea stock
17

• Materials

(1). Water temperature data in station 1
0-50m, 75-150m, 200-300m in each month 1965-1991
(2). Frequency distribution of BL at sex
(3). Monthly catches in the stocks
(4). Mean weight
Catch number at age
Age composition
VPA
Natural mortality coefficient M (Pauly, 1980)
Fishing mortality coefficient at terminal age FT
18
Recruitment (Rt)
Stock number at age 1 from 1966 to 1991
Egg production (St )
Escapement stock number at age
S
egg production at age
2
19
Stock number estimated by VPA
Stock number(106)
20
Egg production-recruitment relationship
r0.67, plt0.01
Recruitment (t)(106)
Egg production (t-2)(108)
21
Life history of sandfish
Year
t
t
-2
t
-1
recruitment (age 1)
Spawning season(Dec.)
Period from Hatching to recruiting
recruiting
22
Relationships between R and water temperature
(correlation coefficient)
R
ln(R)
positive
Plt0.05
positive
Plt0.01
23
Recruitment models
( St )
1.
( Tn )
2.
( St , Tn )
3.
4.
(Ricker model)
5.
( St , Tn ?density-dependent mortality)
Rt recruitment? St egg production? M1, t-t
watertemperature-dependent mortality Tn water
temperature index
24
Evaluation of models
W1 Spring (Apr.), W2 Summer (Aug., Sept.)? W3
Autumn(Oct.), W4 Winter (Nov., Dec.)
25
(No Transcript)
26
(No Transcript)
27
10
Spring
6
2
Water temperature(?)
16
Winter
12
6
M1, t-tWater temperature-dependent mortality
4
1970
1980
1990
Year
28
Observed value and reproduced value
Observed value
Reproduced value
Recruitment(106)
Year
RtSt-2exp-(-0.24W1t 0.33W4t 9.4)-2.010-12
St-2
M1, t-t Water temperature-dependent mortality
29
After 1980
Recruitment (t)(106)
Egg production (t-2)(108)
Egg production (t-2)(108)
30
Summary
Fluctuation in recruitment is affected by factors
as follows
1. Egg production
2. Ocean environmental conditions (Spring and
winter)
3. Egg production-dependent mortality
31
4. Dynamics and catch forecasting of the Korean
Peninsula stock
Catch(104ton)
Year
32
Materials
Catches of Korea, Shimane, Tottori and Hyogo
from 1960 to 1999 Water temperature data from
1960 to1999(KODC) area(A, B, C, D, E, F,
G) depth(0-50, 100, 150, 200, 300m)
month(Feb., Apr., Jun., Aug., Oct. and Dec.)
33
Area
A
B
C
Korea
D
F
G
E
Japan
34
Total catch of Korean Peninsula stock
Y(t)KSTH
K catch of Korea
S catch of Shimane Prefecture, Japan
T catch of Tottori Prefecture, Japan
H catch of Hyogo Prefecture, Japan
35

• Assumptions

(1) Y(t) is proportional to stock size in t
(2) Y(t) is proportional to spawning abundance in
t
(3) Y(t) mainly consists of ages 1 and 2
(Frequency distribution of BL)
Modeling
36
Time lag -1
Time lag -2
negative
Plt0.01
positive
Plt0.01
positive
negative
Plt0.05
Plt0.05
37
Multiple regression analysis
Predictors
SSB Spawning abundance index
W1 water temperature index in spring (Apr.) W2
water temperature index in summer (Jun., Aug.)
W3 water temperature index in Autumn (Oct.)
W4 surface water temperature index in winter
(Dec., Feb. ) W5 bottom water temperature index
in winter (Dec., Feb. )
Results
ln (Yt) 1.2ln(SSBt-t)-5.0ln(W1t-t)-0.8ln(W4t-t)3
.8ln(W5t-t)14
coefficient of determination adjusted for the
number of freedom r20.77
38
Observed and reproduced values






r20.77
39
Evaluation of forecasting by extrapolation
Catch(104ton)
Index of fitting by extrapolation IFE the
square of the correlation coefficient
IFE 0.62
40
Summary
Catch fluctuation in Korean Peninsula stock is
affected by factors as follows
1. Spawner abundance
2. Ocean environmental conditions (Spring and
winter)
These factors coincide with those for North Japan
Sea stock
41
5. Synchronous fluctuations in two stocks
42
Mechanism of stock fluctuations
North Japan Sea
Korean Peninsula
Bottom waters
Bottom waters
Winter
Spring
Bottom
Bottom
Spawner abundance
Spawner
Egg production
Redpositive
Bluenegative
43
Relationships between water temperature and mixed
layer depth (MLD)
In winter and spring , vertical mixing is active
in the Japan Sea
MLDthe depth (m) between the sea surface and a
layer of water where the difference in water
temperature between the sea surface and this
layer is 1º (Hamawa and Hoshino, 1988 Kim and
Isoda, 1998)


44
Water temperature (?)
45
Relationships between water temperature and MLD
in the offshore off Akita
46
Relationships between water temperature and MLD
in the offshore off Korean Peninsula
Dec.
Feb.
Apr.
positive
Plt0.05
positive
Plt0.01
47

MLD
Water temperature
48
Water temperature in winter at 200-300 m
Water temperature (?)
Regime shift
49
Mechanism of stock fluctuations
Large-scale change of ocean environments in
winter in the whole Japan Sea
MLD in winter ?
Synchronous fluctuations in the two stocks
Catch(104ton)
50
6. Different fluctuations in the two stocks from
the 1980s to 1990s
Catch(104ton)
Korean Peninsula
North Japan Sea
51
Stock number (106)
Exploitation rate
Offshore fisheries of Akita
Coastal fisheries
Offshore fisheries of Akita
52
Why different were fluctuations in the two stock?
Difference in fishing intensity

• North Japan Sea

Overfishing by the offshore fisheries from the
1980s
Korean Peninsula
The fishing intensity would be lower compared to
that of the North Japan Sea stock, because
sandfish is not main target in the Korean waters
53
7. Management policy of sandfish ( North Japan
Sea stock )
(1) Evaluation of fishing moratorium and
management policies
(2) Comparisons of management policies
54
(1) Evaluation of fishing moratorium and
management policies

• Age-structured model

Na, t Stock number at age a in year t TNt
Total stock size in t TCa, t Catch (individual)
at a in t M Natural mortality coefficient
(0.14/year)
55
S-R model including the water temperatures
Rt Recruitment in year t St Egg production in
t M1, t Natural mortality coefficient
determined by the water temperatures
56
Simulations
1980 -1982
1985 -1987
1992 -1994
No restriction
Scenario 1
(4 pref.)
Moratorium
Scenario 2
(only Akita)
Moratorium
Scenario 3
(4 pref.)
E

2/3
Scenario 4
(4 pref.)
E

2/3
Scenario 5
(4 pref.)
Low
Water temperature
High
E Mean exploitation rate between 1965 and 1979.
57
Scenario 4
Scenario 5
Low
High
Year
The water temperature in winter
58
Results
1995
Scenario 1
0.5
(No restriction)
Scenario 2
1
(Actual policy)
Scenario 3
3
(4 pref.)
12
Scenario 4
85-87
Scenario 5
18
80-82
The moratorium was finished
59
(2) Comparisons of management policies
1) Monte-Carlo simulations
Using the above models, we conducted this
simulations for the period (1966 - 1999) and
repeated 100 times.
2) Forecasted stock size ( )
e Measurement error
Case 1. Here we can accurately forecast the stock
size using the model including the water
temperatures, e LN (1, 0.4).
Case 2. Here we cannot accurately forecast the
stock size using the model such as Ricker model,
e LN (1, 0.8)
60
3) Management policies
I) Constant exploitation rate policies
(Et 0.3, 0.4, 0.5 and 0.6)
Catch quota calculated in year t
Forecasted stock number in year t
II) Feedback control policy (Tanaka, 1980
Sakuramoto and Tanaka, 1989)
61
4) Actual catch Qt

• e LN (1, 0.2)

Actual catch number at age
62
5) Evaluations the performance of of management
policies
I Mean stock biomass Bmean (ton) between 1966
and 1999 in 100 times.
II Mean catch Cmean (ton) between 1966 and 1999
in 100 times
III The number of extinction EX. EX is the number
when the stock biomass reaches 0 in the
simulation (100 times).
63
Results
Case 2
Case 1
e LN (1, 0.8)
e LN (1, 0.4)
Bmean (103t)
Cmean (103t)
EX
E0.3
E0.4
E0.5
E0.6
Feedback
Feedback
An actual mean E 0.6 (1980-1991)
64
Evaluation of fishing moratorium
1) Implementation of the fishing moratorium
conducted in Akita Prefecture had been important
for the increase of the stock size.
2) If more prefectures (including Aomori, Akita,
Yamagata and Niigata) had implemented the
moratorium, a greater increases of the stock
biomass would have been realized.
3) The fisheries in Akita Prefecture could have
been avoided the 3-year fishing moratorium if the
restriction had been conducted early in the 1980s.
65
Proposals for new management policy
1) The stock forecast using the ocean
environmental conditions can decrease the
measurement error of stock size. This art is
useful to avoid the stock collapse.
2) if the exploitation rates Et could be tightly
controlled, then a constant exploitation rate
policy of Et 0.4 would be a better policy.
3) if Et can not be controlled at a constant
level, the stock should be managed by
implementing a tentative reduction of
exploitation rates, i.e., a 30 reduction during
the 3 years when stock size was very low. In
particular, when the ocean environmental
conditions are good, this approach works
effectively.
66
Summary and conclusion

• (1) Fluctuations in the two sandfish stock is
  influenced by the common factors

Ocean environments (spring and winter) Spawner
abundance
(2) Large-scale ocean environments in winter in
the whole Japan Sea
MLD

Primary production and zooplankton ?
(3) Recruitment model of the North Japan Sea stock
67
(4) Main factor of decline in North Japan Sea
stock from the mid 1970s is
not overfishing ocean environments
Main factor of stock collapse from the late 1970s
to the 1990s is
Overfishing
Korean Peninsula
Norh Japan Sea
Catch(104ton)
(5) Timely management policies adapting to the
stock size
68
END

• Thank you