Page 60 System Dynamics '91
Simulation for a Deadly Blow to Tokyo
by the Coming Next Large Barthquake.
by
Yasoi YASUDA
Institute of Socio-Economic Planning ,
University of Tsukuba
1-1-1 , Tennodai , Tsukuba
305., JAPAN
TEL 0298-53-5182
FAX 0298-53-5070
* This is a revised version of my paper which was presented at the
1990 European Conference of System Dynamics, at Universita L. Bocconi,
Milan, Italy, October 1990.
System Dynamics '91 Page 61
1. Introduction
(1)Introduction of Japan and -Tokyo
After the Second World War, Japan has achieved a rapid economic
growth. GNP of Japan becomes the second in the developed industrialized
countries after the USA. Per capita GNP will be the first in the
world pararelled with Switzerland. Following to the high economic gro-
wth , Tokyo, capital of Japan has become a big monster concentrating
many functions. :
At present, Japan has 120,000,000 of population. And in the Tokyo
metropolitan area, there are 30,000,000 of the population, in truth,
25% of the total population of Japan.
Tokyo is one of the most softy city in the world in the field of
crime. Homer. Tokyo is one of the most dangerous. cities in the field
of the earthquake like ITALY.
(2)The purpose of this study
If a large earthquake were to attack Tokyo. in these days, what would
happen?
Perhaps Tokyo will be destroyed by big fires and by big floods. In
Japan , on September 1, 1923, there was a large earthquake , so called
“Kanto Dai Shinsai” ( in Japanese ) , that is ,“ The large Tokyo. disaster
* This disaster meant the collapse of Tokyo, the capital of Japan.
According to a recent theory set forth by the«Institute of Geophysi-
cal Sciences , a large earthquake similar to the “ Kanto Dai Shinsai ”
will likely occur within about sixty nine years of the former large
earthquake.
Why , then, has there not an appropriate policy been accepted for the
next coming big Tokyo disaster ? The major reason is that the scientific
prediction and the evaluation of policy's effectiveness have not been
carried out.
The main purpose of this study is to build a system dynamics model
for simulating impacts on the Tokyo Metropolitan area by the coming
next large earthquake.
Tokyo may well suffur a deadly blow, and the Japanese economy could
be destroyed by the consequences of such a disaster.
Page 62 System Dynamics '91
As a result, international conflict between Japan and USA will dis-
appear.
2.A System Dynamics Model of the
Disaster and Recovery Process
of the Coming Tokyo Earthquake
2.1 The purpose of the model
The components of disaster brought on by the large earthquake are
destruction of buildings and roads by the direct impact of the earth-
qtake , houses and offices burned down by big fires , and subsequent
damage to the social system: for example confusion of economic activi-
ties. We would like to make a comprehensive system simulation model
of the damage and recovery process of Tokyo. The area of study for
this model is the Tokyo Metropolitan area. This area consists of four
prefectures : Tokyo , Kanagawa , Saitama , and Chiba. We calculated 60 days
for a simulation period with one day as a unit interval. We chose a
System Dynamics method for system simulation because we cannot wse past
data for these situations.
2.2 The Basic Structure of the Model
Our model consists of four subsectors : that is population, regional
economy , transportation , and material sectors. The interrelationship of
four sectors is shown in figure 1. The causal feedback loop which indi-
cates the basic structure of the model is shown in figure 2. We note
that transportation capacity plays an important role in the process of
recovery from the disaster.
8. Results of Simulation Experiment
3,1 Simulation Case
System Dynamics '91 Page 63
The scale of damage brought on by the large earthquake is quite de-
pendent on the size of the earthquake. In this model we set up several
simulation cases given the occurrence of a large earthquake whose charac-
teristics are the same as those of Kanto Daishinsai (the former Tokyo
disaster). We build up four simulation cases which are shown. in table
1 according to the season ,time , and weather , and also recovery policy
after the earthquake.
A. Case I (Pessimistic Case)
For this pessimistic case we assume the earthquake takes place at
dinner time under strong wind conditions in the winter season.
There will be many deaths since the earthquake occurs during the
commuting hours.
Thirty percent of the wooden buildings are burnt down. We assume
there is one road obstacle per 1 km distance. The road damage is net-
work type destruction.
In the economic sector we assume that the petrochemistry combinate
belt suffers deadly destruction , and thus does not recover during the
sixty day period. None of the petroleum tanks can be employed because
of the destruction and fires in the belt.
The main parts of steam power generationg stations are damaged, and
as a result the supply of electric power is curtailed.
B. Case I (Optimistic Case)
For this optimistic case we suppose that the large earthquake ocurrs
at midnight at a time of weak winds , during the summer season.
We assume both of mortality of the refuges and loss of buildings due
to fire at the low level. The occurrence of road obstacles are few,
and road damage is single-type destruction.
C. Case M (Intermediate Case)
For this intermediate case we assume the middle level between case |
and I. Hence , each parameter is set up at an intermediate level of
Page 64 System Dynamics '91
the two cases.
D.Case W (Standard Case)
For this standard case we suppose that each parameter is chosen as
the most feasible .situation to be encountered in reality. This case
is the most probable case in our model.
3.2 Results of Simulation © Experiments
For each case we carried out a computer simulation’, and obtained
certain interesting results. We want to explain the simulation results
comparing the several cases.
First , simulated regional population levels are shown in -figure 3,
The minimum of the population occurrs with standard case ¥, and the
passage of time when the minimum takes place’ is earlier than in the
pessimistic case I. In the pessimistic case | , the. transportation multi-
plier becomes very small , hence the population who can not find refuge
elsewhere’ stay in the region.
The changing levels of food stocks are shown in: figure 4. For the
pessimistic case | and the standard case. , the food stocks are ex-
hausted 10 and 40 days. after the earthquake , respectively. This is be-
cause of population staying in the region, very low transportation ca-
pacity.
The number of roads interruped is shown in figure 5. For the opti-
mistic case [ and the intermediate case , over 50% recovery is project-
ed within 20 days. However for the pessimistic case | and the standard
case ¥ , the recovery pace is slightly slower. This reason is the short-
ness of the behavior according to the decrease of the population.
Changes in fuel stocks are shown in figure 6. For the standard case
WY, the exhaustion of stocks appears earlier time than in the pessimis-
tic case |.
Transportation capacity changes are shown in figure 7. For the pessi-
mistic case | and the standard case ¥ , share declines in transportation
capacity appear 50 and 30 days after the earthquake , respectively , brought
on by lack of fuel.
System Dynamics '91 Page 65
4. Concluding Remarks and Discuussion
We succeeded in formulating the simulation model for a deadly blow
to Tokyo by the next large earthquake. System Dynamics method is useful
rather than Econometric method , because this problem is future oriented
and complex. e
We had an interesting result using simulation experiments by System
Dynamics Model. A complex relationship among subsystems of the socio-
economic system, has been analyzed by the feedback loop of the System
Dynamics Model.
We found: out that in this complex model the bottle neck element
plays an important role in’ the behavior of the socio-economic system,
In this case, transportation capacity is a bottle neck factor and be-
havior of each subsystem is influenced by the transportation capacity.
Moreover , we have to develop our model including other sectors such
as a communication system, land use and environmental systems.
Page 66 System Dynamics ‘91
References
1. Forester, J.W.,1969, Urban Dynamics MIT Press
2.Goodman,M.R., 1974, Stu Notes in System Dynamics Wright-Allen Press
3.Yasuda,Y. and M.Hijikata, 1982, “A Gaming Simulation of the Human Be-
havior of Taking Refuge in an Emergency —A Case Study of the Com-
ing Big Earthquake in the Tokyo Metropolitan Area—,” Modeling and
Simulation, Vol.13, pp. 1147-1154
4.Yasuda,Y. (ed), 1979, A Study of the Protective Systems for Large Dis-
asters, The Foundation of Advanced International Sciences, Japan (in Jap-
anese)
5. Yasuda, Y. and M.Hijikata, 1979, “Simulation of a Disaster Caused by the
Coming Large Earthquake in Tokyo Using a System Dynamics Approach”,
Communications of Operations Research, September 1979, pp. 549-555 (in Jap-
anese)
System Dynamics '91 Page 67
Table-1 Simulation Case
I I sig Vv
Case Pessimistic |Optimistic |Intermediate|Standard
Case Case Case Case
Sector Index (Age)
O~14: 40thou, U5thou. 80thou,
Population|Death Population |15~64:200 1 150 7 200 1/Same as I
during the Refuge|65~ : 201 10 1 154
Rate of Burnt
Down Wooden 30% 10% 20% 1 sw
Buildings
Facilities
Occurrence Rate
of Road Obstacles] 1(part)/kn 0.47 kn 0.7/kn a |
Transpor- |Pattern of the Network De- |One Part De- |Intermediate
tation Damaged Road stroying Type|stroying Type|Type 1 ow
Electric Power | L ~
bees i ister = 0,G°-0.72 0,86
Supply Hultiplier] gly oa a3 7 “8 a at hy Oo 0.57 a 1
Multiplier of | ha 5.8
Materjals Carried] y's-%7-Ut=%% {= eecaes ; or a s 1a
to the District ae
Economy
Rate of Burnt
Down Stocks of 100% 30% 50% rot
Producers’ Fuel
Level of Fuel
Production L
tb
Page 68 System Dynamics '91
(million)
30
population
X casel
O cascl
20 & caselll
© casclV
_10
ae
Tt . . A A A ‘
0 10 20 30 40 50 60 Time
(days elapsed
Figure 3 Tolal population Inthe Region ~ from earthquake)
System Dynamics '91
Page 69
food stock
{kilo ton)
400
X -cascl
” © casell
© caselll
O caselV
200
100
Time
(ays)
Figure4 Food stock
Page 70 System Dynamics '91
15,000
(points)
10,000
$,000
(ays)
Figure Roads Interrupted
System Dynamics '91 Page 71
fuel stock
Ost)
15007
x casel
O casell
10004 © case lll
O cascIV
500-
1 i ——t ———
0 10 20 30 40 50 60 Time
(days)
Figure 6 Fuel stock
Page 72 System Dynamics '91
Transportation
Sector
Population
Sector
OTransportation
Capacity
\ Multiplier
OWorkers in
Constructing
Industry
O Transportation
O Amount of Food Amount of Fuel Capacity
Consumption Consumption Multiplier
© Producers: OTransportation
Workers Capacity
© Standard of Multiplier
Food Stocks O Number of
Standard of Obstacles
; lO Amount of
Production to the road
Fuel Stocks oCollapsed or Traffi
cs
Burnt Down Area ran
——— OAmount of Stocks
of Building
Materials
Economic
Sector
Material
Sector
O Floor Space
O Amount of Burnt
Down Stocks
© Amount of Consumption
of Building Materials
Figure 1 Relationship between Each Sector
System Dynamics '91 Page 73
traffic capacity
multiplice
10
0,8
: oo
06 x casel
O casell
© caselll
O-casc1V
4 4 z 1 1 4. z
0 10 20 30 40 50 60 Time
(ays)
earthquake Figure7 Capacity of traffic
Page 74
System Dynamics '91
Population
Sector OTransportation
Capacity
\ fultiplier
‘OWorkers in
Transportation
Sector
=
Constructing
Industry
O Transportation
© Amount of Food Amount of Fuel Capacity
Consumption Consumption Multiplier
© Producers- OTransportation
Workers Capacity
© Standard of Multiplier
Food Stocks O Number of
Standard of
Production OAmount of ca iol
Fuel Stocks oCollapsed or Traffics
Burnt Down Area
OAmount of Stocks
of Building
Materials
Economic
Sector © Floor Space
O Amount of Burnt
Down Stocks
© Amount of Consumption
of Building Materials
Material
Sector
Figure 1 Relationship between Each Sector
System Dynamics '91 Page 75
+ Labor Force of
Manufacturing
Industry
Labor Force of
+ Capacity of
Constructing se aed +
Roads
Industry
Population of + Ni i eras ae
the Region i ean t Mee t _ Productivity ir
~ fame eps Manufacturing
Industory
S t
Food
Number of Escaping et Transportation
From the Region Capacity
c. - 2: NG
Food Stock Fuel Stocks
+ + =
Consumption
of Fuel
Figure 2 Loop diagram
