Estimating the In‐use Steel Stock of 

Ci il
Civil Engineering and Building in China 
i i d ildi i Chi
via Nighttime Light
via Nighttime Light
Hsu (David)1; ;
Feng Chi, Hsu (David)
Feng‐Chi,
Christopher D. Elvidge 2; 
Yasunari Matsuno 1
1 Department of Materials Engineering, Graduate School of Engineering, The University of Tokyo
p g g, g g, y y
2 National Geophysical Data Center, National Oceanic and Atmospheric Administration

APAN31, Hong Kong, 2011.2.21
 Background
• Rapid growth in construction sector in China
– Fast urbanization, Population growth
• Steel consumption in China (2004) Hu et al.
Total:  286 Mt 
Construction sector : 145 Mt ( 50%) 
• In
In‐use
use steel stock in China   
steel stock in China (2005) 
(2005) Hatayama et al.
Hatayama et al.

Total:  2334Mt 
Construction sector： 1456Mt (
Construction sector： 1456Mt ( 65%) 
65%)
• Steel used in construction sector in China needs more 
d t il d t d
detailed study
– Material Flow Analysis (MFA) 2
 M t i l Fl A l i (MFA)
Material Flow Analysis (MFA) 
• Account
Account for flows of materials in defined system.
for flows of materials in defined system
• System can be geophysical boundaries and/or 
product life cycle.
• This study focus on the in
This study focus on the in‐use
use stock in China
stock in China

Import / Export
Import / Export

Produce Transport Use Dispose

Stock

Environment
3
 Issues in MFA
Issues in MFA
• Building stocks and their material contents in China 
are estimated by MFA approaches.
– Estimate housing stock using dynamic MFA modeling for 
1900‐2100.  Hu et al.
– Iron and steel stock of the buildings based on estimated 
housing stock. Hu et al.
– MFA model to account for the material and energy flows in 
MFA d l f h i l d fl i
construction sector. Yang et al.
– Estimation of steel stock regarding multiple end uses in 
Estimation of steel stock regarding multiple end uses in
different countries. Hatayama et al.; Kakiuchi et al.
– Etc.
Etc
4
 I
Issues and Challenges
d Ch ll
• Mostly country level studies.
l l l d
• No comprehensive study on sub‐national level 
p y
(prefectures).
• Lack of sub‐national level statistical data.
L k f b ti l l l t ti ti l d t
• Methodology capable without detailed 
gy p
statistical data?
• Satellite images →
Satellite images → Proxy to unknown quantities 
Proxy to unknown quantities
 Nighttime Light
Nighttime Light
• Anthropological lighting that was observed by 
satellite.
• Radiance positively correlated to human 
activities.
activities
– GDP, Population, Energy consumption, copper and 
steel stock.

6
NGDC, NOAA
 P i
Previous Study ‐
St d Outline
O tli
• Civil engineering/building steel stock 
Civil engineering/building steel stock Nighttime light
• Prefectures in Japan
• Land cover considered  →
L d id d → Correlation refined
C l i fi d
Hsu et al. (2010)

Land Cover 
• Indicate
Indicate coverage properties on 
coverage properties on
earth surface
• ISCGM
• 1km grid
ISCGM • 20 categories
20 categories
7
 P i
Previous Study ‐
St d Result
R lt
• Civil engineering steel stock → Total nighttime light.
Civil engineering steel stock → Total nighttime light
• Building steel stock   →  Urban nighttime light.
• Land cover is effective in refining the correlation
Building steel stock ֞
g Building steel stock ֞
g
Total Nighttime Light Urban Nighttime Light
60 60
06 t)

Building SSteel Stock (106 t)
50 50
Building SSteel Stock (10

40 40

30 30
20 20
10 y = 36.786x 10
y = 91.342x
0 R² = 0.5458
0
R² = 0.8227
0 50 100 150
Total Nighttime Light (104 w/cm2/sr) 0 10 20 30 40 50 60
4 2
Urban Nighttime Light (10 w/cm /sr)
8
 Objective
• Use nighttime light image to estimate the civil 
g g g
engineering/building steel stock for each 
prefecture in China.
– Civil engineering/building steel stock 
Civil engineering/building steel stock Nighttime 
Nighttime
light
– Asia Pacific countries
A i P ifi ti
– Estimate civil engineering/building steel stock in each 
Chinese prefecture.

9
 Flow Chart of Methodology
Flow Chart of Methodology
[Global]
[Global]
Nighttime light
In-use steel stock

[Chinese Prefectures] [Asian Pac. Countries]

Boundaries Boundaries [Asian Pac. Countries]
Building steel stock
[Chinese Prefectures] [Asian Pac. Countries] [Asian Pac. Countries]
Total of nighttime light Total of nighttime light Civil engineering steel stock

Land cover
Gas flare mask
[Chinese Prefectures] [Asian Pac. Countries]
Linear
Urban nighttime light Urban nighttime light
regression
g analysis
y
[Chi
[Chinese P f t
Prefectures]
] [A i Pac.
[Asian P Countries]
C t i ]
Gas flare excluded Gas flare excluded
nighttime light nighttime light

Model off steel stock

and nighttime light
Global datasets
[Chinese Prefectures] Asian Pac. datasets
Building steel stock
[Chinese Prefectures] China datasets
Civil engineering steel stock Analysis & Model 10
 Data Preparation
Data Preparation
Sample countries
• 16 Asia Pacific Countries
— Restriction of data accessibilityy
• Russia, China,  India, Japan, Korea, 
Pakistan, Turkey, Indonesia, Australia, 
Thailand, Malaysia, Taiwan, Philippine,
Thailand, Malaysia, Taiwan, Philippine, 
New Zealand, Bangladesh, Singapore

Nighttime Light Steel Stock
• F162006 • Hatayama et al.
• Radiance Calibrated • In‐use steel stock for multiple 
end uses
– Steel stocks are city concentrated • Data of year 2006 is used.
Data of year 2006 is used.
– Better NTL data for city center
– Avoid saturation 11
 Ancillary Data
Ancillary Data
New
Land Cover (ISCGM) Gas Flare Mask (Elvidge et al.)
• Proved valid to refine  • Extraordinary bright
correlation between in‐use  • Does not necessary 
steel stock and nighttime  correlate to steel stock 
li ht
light • Exclude?

ISCGM Wikipedia 12
Li
Linear Regression analysis
R i l i
y = 15..57 x y = 82.5 x y = 19.93x
R2 = 0.91 R2 = 0.85
R2 = 0.92

y = 15.71 x y = 99.31 x y = 21.92 x

R2 = 0.73 R2 = 0.87
R2 = 0.96

13
 Results
Estimated Civil Engineering/Building Steel Stock in China
Estimated Civil Engineering/Building Steel Stock in China
Prefecture Civ. Eng. Bld. Prefecture Civ. Eng. Bld.
1 Anhui 15700 15500 17 Jilin 16400 21600
2 Beijing 21400 63400 18 Lianoning 28800 47100
3 Chongqing 5600 7000 19 Nei Mongol 19700 20000
Ningxia
4 F ji
Fujian 15400 19300 20 4600 3800
Hui
5 Gansu 8700 9800 21 Qinghai 2400 2700
6 Guangdong 65700 125800 22 Shaanxi 18400 17700
7 Guangxi 11200 10000 23 Shandong 57900 61800
8 Guizhou 5200 2200 24 Shanghai 21400 61600
9 Hainan 3800 3200 25 Shanxi 25000 19300
10 Hebei 35800 30000 26 Sichuan 15700 15200
11 Heilongjiang 32600 41000 27 Tianjin 13500 26700
Xinjiang
12 Henan 35900 36900 28 15000 17600
Uygur
13 Hubei 14000 14400 29 Xizang 800 1100
14 Hunan 10600 10100 30 Yunnan 15800 16400
15 Jiangsu 53500 67500 31 Zhejiang 33000 38400
16 Jiangxi 8700 9100 Unit: kt
14
 Results 
Results
Density Graph of Steel Stock in China
Beijing

Ti ji
Tianjin

Shanghai

Guangzhou 15
 Discussion
• Estimated civil engineering/building steel stock
i d i il i i /b ildi l k
– Beijing (2006) : 84.8 Mt
• Reported total steel stock (Rauch et al.)
– Beijing (2000) : 29.4 Mt
Reason?
• Fast urbanization
F t b i ti
• Transition in building styles
– Steel intensity for residential buildings
– +13% (2004→2010) (Hu et al.)

16
 Conclusion
• C
Confirmed linear correlation between civil 
fi d li l i b i il
engineering/building steel stock and nighttime 
li ht
light.
– Civil engineering → Gas flare excluded nighttime light
– Building → Urban nighttime light
• Civil engineering/building steel stock estimated
– Highest density, largest quantity in few highly 
developed prefectures
– Eastern China, along coast line
• Seek for ground‐truth in China
17