21 st Asia-Pacific Advanced Network Meetings 21 st Asia-Pacific - - PowerPoint PPT Presentation

Monitoring for electric power output Monitoring for electric power output and fossil CO and fossil CO2 emissions emissions by means of DMSP/OLS nighttime imagery by means of DMSP/OLS nighttime imagery Monitoring for electric power output Monitoring for electric power output and fossil CO and fossil CO2 emissions emissions by means of DMSP/OLS nighttime imagery by means of DMSP/OLS nighttime imagery

Ma Masa sana nao HARA

• HARA*1, Hiroshi Yagi*

, Hiroshi Yagi*1, Hus , Husiletu tu*2, Fumihiko , Fumihiko Nishio Nishio *2 *1 *1 VTI Research Institute, VisionTech Inc.(VTI) *2 *2 Center for Environmental Remote Sensing, Chiba University Ma Masa sana nao HARA

• HARA

Ma Masa sana nao HARA

• HARA*

*1

1,

, , Hiroshi Yagi , Hiroshi Yagi* *1

1, Hus

, Husiletu tu* *2

2, Fumihiko

, Fumihiko Nishio Nishio * *2

2

*1 *1 *1 *1 VTI Research Institute, VisionTech Inc.(VTI) *2 *2 *2 *2 Center for Environmental Remote Sensing, Chiba University January 22, 2006

21st Asia-Pacific Advanced Network Meetings 21st Asia-Pacific Advanced Network Meetings

• Most electric power is

Most electric power is Most electric power is Most electric power is generated by the fossil fuel. generated by the fossil fuel. generated by the fossil fuel. generated by the fossil fuel.

• Greenhouse gas is one of big issue in the 21st century.

Greenhouse gas is one of big issue in the 21st century.

• Greenhouse gas is one of big issue in the 21st century.

Greenhouse gas is one of big issue in the 21st century. Greenhouse gas is one of big issue in the 21st century. Greenhouse gas is one of big issue in the 21st century.

• In particularly, CO

In particularly, CO In particularly, CO In particularly, CO2

2 emission is increasing because of

emission is increasing because of emission is increasing because of emission is increasing because of hu human activities. man activities. hu human activities. man activities.

• Human activities always ne

Human activities always need e ed energ ergy. Human activities always ne Human activities always need e ed energ ergy.

• Typically

Typically Typically Typically electric power electric power electric power electric power is one is one is one is one

• f
• f
• f
• f most important energy for

most important energy for most important energy for most important energy for hu human activities man activities hu human activities man activities. . Back Ground Back Ground Back Ground Back Ground

1.3 1.3 1.7 1.7 15.5 15.5 4.4 4.4 3.4 3.4 41.5 41.5 3.3 3.3 3.2 3.2 20.4 20.4 5.1 5.1 Hydraulic Hydraulic and other and other energy energy 14.7 14.7 9.5 9.5 7.6 7.6 42.1 42.1 13.0 13.0 0.9 0.9 16.0 16.0 5.6 5.6 0.4 0.4 9.1 9.1 Atomic Atomic energy energy 8.8 8.8 37.6 37.6 29.7 29.7 13.7 13.7 21.1 21.1 4.4 4.4 12.4 12.4 51.9 51.9 2.5 2.5 23.7 23.7 Natural Natural gas gas 53.6 53.6 35.7 35.7 35.1 35.1 33.9 33.9 38.7 38.7 20.4 20.4 50.5 50.5 21.2 21.2 19.4 19.4 38.6 38.6 Oil Oil 21.7 21.7 15.5 15.5 12.1 12.1 5.8 5.8 23.7 23.7 32.9 32.9 17.9 17.9 18.0 18.0 57.4 57.4 23.6 23.6 Coal Coal Energy Energy sources sources (%) (%) 193.6 193.6 232.6 232.6 251.0 251.0 257.1 257.1 339.6 339.6 501.9 501.9 524.7 524.7 614.0 614.0 1,142.4 1,142.4 2,299.7 2,299.7 Total Total

Korea Korea U.K U.K Canada Canada France France Germany Germany India India Japan Japan Russia Russia China China USA USA

【 【Source Source】 】 IEA, Energy Balance of OECD Countries(2002) IEA, Energy Balance of OECD Countries(2002) IEA, Energy Balance of NON IEA, Energy Balance of NON-

• OECD Countries(2002)

OECD Countries(2002) units units： ：million tons converted into oil million tons converted into oil

Consumption com Consumption composition in

• sition in the

the top 10 countries top 10 countries

• f primary e
• f primary energy supply

ergy supply（2000 2000） Consumption com Consumption composition in

• sition in the

the top 10 countries top 10 countries

• f primary e
• f primary energy supply

ergy supply（2000 2000）

Source : Ministry of Environment, 2001 Source : Ministry of Environment, 2001 Electric Power Electric Power plant plant 37.9% 37.9% Transportation Transportation ( (Ship, car, plane, etc) Ship, car, plane, etc) 21.4% 21.4% Industry Industry 21.1% 21.1% Other business Other business 7.9% 7.9% Industrial process Industrial process 4.2% 4.2% Waste material Waste material 2.0% 2.0% Others 0.1% Others 0.1% Home Home 5.4% 5.4%

CO2 emission in Japan

The total amount

• f year 2001

1.4 billion-tons

CO2 emission in Japan

The total amount

• f year 2001

1.4 billion 1.4 billion-

• tons

tons

Structure of CO Structure of CO2 emission sources in Japan emission sources in Japan Structure of CO Structure of CO2 emission sources in Japan emission sources in Japan

Objective Objective Objective Objective

• Estimate electric power consumption and monitoring

Estimate electric power consumption and monitoring Estimate electric power consumption and monitoring Estimate electric power consumption and monitoring CO CO CO CO2

2 emission from the electric power consumption

emission from the electric power consumption emission from the electric power consumption emission from the electric power consumption by using satellite remote sensing. by using satellite remote sensing. by using satellite remote sensing. by using satellite remote sensing.

・ ・DMSP(Defe efense Meteorological Satellite P se Meteorological Satellite Program) /

• gram) / O

OLS(O S(Optical tical DMSP(Defe efense Meteorological Satellite P se Meteorological Satellite Program) /

• gram) / O

OLS(O S(Optical tical linescan linescan linescan linescan System) data were used. System) data were used. System) data were used. System) data were used. ･ ･A whole year of 1999 data (365 scenes) were used. A whole year of 1999 data (365 scenes) were used. A whole year of 1999 data (365 scenes) were used. A whole year of 1999 data (365 scenes) were used. ・ ・1 km spatial resolution data 1 km spatial resolution data generated by NGDC in U generated by NGDC in U.S.A. S.A. 1 km spatial resolution data 1 km spatial resolution data generated by NGDC in U.S.A. generated by NGDC in U.S.A.

Data Data Data Data

DMSP (Defense Meteorological Satellite Program) specification DMSP (Defense Meteorological Satellite Program) specification Orbit Sun-synchronous near polar orbit Altitude 830km Revisit 101min Sensor OLS （Operational Linescan System） SSM/I （Microwave Imager ) SSM/T （Atmospheric Temperature Profiler） SSM/T2（ Atmospheric Water Vapor Profiler ） SSJ/4 (Precipitating Electron and Ion Spectrometer ) SSIES (Ion Scintillation Monitor ) Visible (Night) 0.47－0.95μm 0.55km 2.7km 3000km 6bit Thermal-IR 10.0－13.4μm 0.55km 2.7km 3000km 8bit OLS Sensor specification Band Spectral range Spatial resolution Swath width Radiometric resolution Fine Smooth Visible 0.40－1.10μm 0.55km 2.7km 3000km 6bit

Description of the satellite and the s Description of the satellite and the sensor nsor Description of the satellite and the s Description of the satellite and the sensor nsor

韓国 北朝鮮 ロ シ ア 中国

Study area Study area Study area Study area

Methodology Methodology Methodology Methodology Methodology Methodology

• Remove cloud from a time series of DMSP/OLS imagery

Remove cloud from a time series of DMSP/OLS imagery by the by the 10-days Maximum 10-days Maximum Value Composite method. Value Composite method.

• Extract the nighttime stable light(DN) by the NRF (Noise

Extract the nighttime stable light(DN) by the NRF (Noise Reduction Filter) algorithm. Reduction Filter) algorithm.

• Comp

Compen ensa sate te th the sa e satu tura ration of DN

• n of DN by a sim

by a simple meth le method

• d na

named med Deltaic Model. Deltaic Model.

• Find correlation between DN an

Find correlation between DN and statistical electric power d statistical electric power consumption. consumption.

• Estimate the electric power cons

Estimate the electric power consumption from DN by using the umption from DN by using the correlation. correlation.

• Calculate CO

Calculate CO2 emission level from the es emission level from the estimated po timated power consumptio wer consumption. n.

∑

= +

            +       + + =

N l l l l l t

t M k c t M k c t c c f

1 1 2 2 1

2 cos 2 sin π π

Where; M : Cycle（1 year = 36 for 10-days composite） ｔ : Time period (1Year＝36，2Year＝72，3Year＝108, 4Year＝144） Kt : Frequency（1, 2, 3, 4, 6, 12 month(s)) N : Number of pixel C0，C1，C2l，C2l+1 : Coefficient of each function Where; M : Cycle（1 year = 36 for 10-days composite） ｔ : Time period (1Year＝36，2Year＝72，3Year＝108, 4Year＝144） Kt : Frequency（1, 2, 3, 4, 6, 12 month(s)) N : Number of pixel C0，C1，C2l，C2l+1 : Coefficient of each function

“ “NRF NRF NRF NRF” ” is developed to remove a cloud from a time series compo is developed to remove a cloud from a time series composite ite is developed to remove a cloud is developed to remove a cloud from a from a time series time series compo composite ite images especially designed images especially designed for NDVI and SST products. for NDVI and SST products. images especially designed images especially designed for NDVI and SST products. for NDVI and SST products.

The Algorithm of Noise Reduction Filter The Algorithm of Noise Reduction Filter The Algorithm of Noise Reduction Filter The Algorithm of Noise Reduction Filter （ （NRF NRF NRF NRF） ）

・ ・Finding a cyclic patterns(seaso Finding a cyclic patterns(seasonal changes) and e al changes) and extract harmo tract harmonic ic Finding a cyclic patterns(seaso Finding a cyclic patterns(seasonal changes) and e al changes) and extract harmo tract harmonic ic series from a time series vegetation index(SST). series from a time series vegetation index(SST). series from a time series vegetation index(SST). series from a time series vegetation index(SST). ・ ・Estimating the correct NDVI(SST) Estimating the correct NDVI(SST) of a pixel under the cloud or

• f a pixel under the cloud or

Estimating the correct NDVI(SST) Estimating the correct NDVI(SST) of a pixel under the cloud or

• f a pixel under the cloud or

its shadow by the cyclic patte its shadow by the cyclic patterns and the harmonics series. rns and the harmonics series. its shadow by the cyclic patte its shadow by the cyclic patterns and the harmonics series. rns and the harmonics series.

Before NRF processing After NRF processing P1 P2 P1 P2

Digital Number

P2

50 100 150 200 250 1 11 21 31 41 51 61 71 81 91 101 111 121 131 141

NRF RAW

P1

50 100 150 200 250 1 11 21 31 41 51 61 71 81 91 101 111 121 131 141 Digital Number

NRF RAW

Profile of the observed DN Profile of the observed DN Day 1 Day 1 Day 2 Day 2 night night night Day time Day time Day time Day time Light on Light on Light on Light on Light on Light on Light off Light off Light off Light off Light off Light off 18 0 6 12 18 0 6 12 18 18 0 6 12 18 0 6 12 18 0 6 0 6 → →Time Time → →Time Time DN DN ○ ○ ○ ○ ○ ○

Observation Observation

NRF for DMSP/OLS nightt F for DMSP/OLS nighttime t e time ser me series data set es data set NRF for DMSP/OLS nightt F for DMSP/OLS nighttime t e time ser me series data set es data set NRF for DMSP/OLS nightt F for DMSP/OLS nighttime t e time ser me series data set es data set

1st Frequency image 3rd Frequency image 4 Frequency Image 2nd Frequency image

Frequency Component Images Frequency Component Images Frequency Component Images Frequency Component Images Frequency Component Images Frequency Component Images

Direct current component Direct current component

A NRF processed image A NRF processed image

Comparison with a NRF processed data and a Land use map Comparison with a NRF processed data and a Land use map

Land use map(1977, GSI) Urbanization density Land use map(1977, GSI) Urbanization density (m2) (m2)

I I II II III III IV IV V V Low density level Low density level High density level High density level

The composite image of a Land use map And a NRF processed image The composite image of a Land use map And a NRF processed image

#27 Shinseimaru

Gross ton ： １３３ ｔ Ship length ： ４２ ｍ Ship width ： ６．２ｍ Gross ton ： １３３ ｔ Ship length ： ４２ ｍ Ship width ： ６．２ｍ

Number of light bulbs ：１０８ Total Out put Power ：２５０ｋW Number of light bulbs ：１０８ Total Out put Power ：２５０ｋW

Compensation of light intensity saturation

新生丸 8月26日

2 4 6 8 10 12 5 10 15 20 25 30 35 40 45 50 55 60 DN値 度数 度数 累積度数

y = -0.111x + 10

#27 Shinseimaru（Aug26）

新生丸 9月6日

2 4 6 8 10 12 5 10 15 20 25 30 35 40 45 50 55 60 DN値 度数 度数 累積度数

y = -0.190x + 8

#27 Shinseimaru（Sep6）

Deltaic Model Deltaic Model Deltaic Model Deltaic Model Deltaic Model Deltaic Model

The The Deltaic Deltaic model is used for model is used for compensation of Saturated DN compensation of Saturated DN The The Deltaic Deltaic model is used for model is used for compensation of Saturated DN compensation of Saturated DN

The concept of Deltaic Model The concept of Deltaic Model The concept of Deltaic Model The concept of Deltaic Model The concept of Deltaic Model The concept of Deltaic Model

(x,y)= (pixel count DN0～ DNMax,0) (x,y)=(pixel count DN63～ DNMax, 63) Y-axis: DNn X-axis: pixel count DNn ～ DNMax (x,y)=(0 , DNMax)

θ

max

tan DN PNtotal = θ

max

tan DN PNtotal = θ

( )

n n

DN DN PN − =

max

tanθ(

)

n n

DN DN PN − =

max

tanθ

DN = Digital Number of each pixel PN = Number of pixels DN = Digital Number of each pixel PN = Number of pixels

Before Before After After

255 255

Comparison between Deltaic compensation Comparison between Deltaic compensation and non-compensation data and non-compensation data Comparison between Deltaic compensation Comparison between Deltaic compensation and non-compensation data and non-compensation data

Regression analysis are used Regression analysis are used to find correlation between to find correlation between stable nighttime lights and electric po stable nighttime lights and electric power co wer consumption nsumption

• f each prefecture in Japan.
• f each prefecture in Japan.

Regression analysis are used Regression analysis are used to find correlation between to find correlation between stable nighttime lights and electric po stable nighttime lights and electric power co wer consumption nsumption

• f each prefecture in Japan.
• f each prefecture in Japan.

The statistical data of electrical consumption are picked up from the data book of “Summary of electrical supply and demand” issued by Ministry of Economic, Trade and Industry(METI), 2001.

Estimation of electric power consumption from Estimation of electric power consumption from Stable nighttime lights Stable nighttime lights Estimation of electric power consumption from Estimation of electric power consumption from Stable nighttime lights Stable nighttime lights

Hokkaido Hokkaido Tohoku Tokyo Tokyo Hokuriku Hokuriku Chubu Chubu Chubu Kansai Kansai Chugoku Chugoku Shikoku Shikoku Kyushu Kyushu Okinawa Okinawa

Electric power distribution map Electric power distribution map

• f each power supply companies in Japan
• f each power supply companies in Japan

Electric power distribution map Electric power distribution map

• f each power supply companies in Japan
• f each power supply companies in Japan

Regression analysis between DN and electric power Regression analysis between DN and electric power consumption of each prefecture in Japan consumption of each prefecture in Japan Regression analysis between DN and electric power Regression analysis between DN and electric power Regression analysis between DN and electric power Regression analysis between DN and electric power consumption of each prefecture in Japan consumption of each prefecture in Japan consumption of each prefecture in Japan consumption of each prefecture in Japan

y = 51070x

• 1E+10

R 2 = 0. 7551 50, 000 100, 000 150, 000 200, 000 250, 000 300, 000 350, 000 1, 500, 000 3, 000, 000 4, 500, 000 6, 000, 000 del tai c データ 発電電力量 (10

6kWh)

東京 関西 東北 北海道

Tokyo Tokyo Kansai Kansai Tohoku Tohoku Hokkaido Hokkaido

Electric power consumption(10 Electric power consumption(106

6kwh)

kwh)

data

The CO The CO2 emission level Computation emission level Computation from electric power consumption from electric power consumption is guided by METI. is guided by METI. The CO The CO The CO The CO2

2 emission level Computation

emission level Computation emission level Computation emission level Computation from electric power consumption from electric power consumption from electric power consumption from electric power consumption is guided by is guided by is guided by is guided by METI METI METI METI. .

Computation of CO Computation of CO2 emission from emission from estimated electrical power consumption estimated electrical power consumption Computation of CO Computation of CO Computation of CO Computation of CO2

2 emission from

emission from emission from emission from estimated electrical power consumption estimated electrical power consumption estimated electrical power consumption estimated electrical power consumption

β ・

e p

C G = β ・

e p

C G =

p

Gp G ： CO2 emission volume (kg） β β ：CO2 emission coefficient( 0.375 at Japan）

e

Ce C：Electric power consumption

Where; Where;

y = 17.77x - 5E+06 R2 = 0.7225 20 40 60 80 100 120 1,500,000 3,000,000 4,500,000 6,000,000 deltaic data 電力消費量に基づくCO

2 排出量 (10 6t)

Tohoku Kansai

CO2 emission from electric consumption(106t) Tokyo Hokkaido

Regression analysis of CO Regression analysis of CO Regression analysis of CO Regression analysis of CO2

2 emission between

emission between emission between emission between DN and statistical electric power consumption DN and statistical electric power consumption DN and statistical electric power consumption DN and statistical electric power consumption

• f each prefecture in Japan
• f each prefecture in Japan
• f each prefecture in Japan
• f each prefecture in Japan

Electric power distribution map of Korea and China Electric power distribution map of Korea and China Electric power distribution map of Korea and China Electric power distribution map of Korea and China

Regression analysis between DN value and electric power Regression analysis between DN value and electric power consumption of each prefec consumption of each prefecture in Korea and China ture in Korea and China Regression analysis between DN value and electric power Regression analysis between DN value and electric power Regression analysis between DN value and electric power Regression analysis between DN value and electric power consumption of each prefec consumption of each prefecture in Korea and China ture in Korea and China consumption of each prefec consumption of each prefecture in Korea and China ture in Korea and China

y = 28987x + 3E+ 09 R

2 =

0.852 10,000 20,000 30,000 40,000 50,000 60,000 70,000 80,000 90,000 500,000 1,000,000 1,500,000 deltaicデ ー タ 発電電力量 (10

6kWh)

そ の 他 ソウ ル

慶尚北道 慶尚南道

Electric power consumption(10 Electric power consumption(106

6kwh)

kwh) Deltaic Data Deltaic Data

Others Others Seoul Seoul

Korea Korea Korea Korea

y = 64924x + 4E+ 09 R

2 =

0.8787 10,000 20,000 30,000 40,000 50,000 60,000 70,000 80,000 90,000 500,000 1,000,000 1,500,000 d eltaic データ 発電電力量 （10

6kWh）

そ の他 青島

遼寧省 浙江省

Electric power consumption(10 Electric power consumption(106

6kwh)

kwh) Deltaic Data Deltaic Data

Others Others Tsigtao Tsigtao

China China China China

Kyongsang-namdo Kyongsang-bukto Liaoning Zhejiang

Korea Korea Korea Korea China China China China

y = 21.245x + 1E+06 R2 = 0.8787 20 40 60 80 100 120 500,000 1,000,000 1,500,000 deltaic データ 電力消費量に基づくCO

2 排出量 （106t）

その他 青島 Deltaic Data Deltaic Data CO2 emission from electric consumption(106t)

Others Others Tsigtao Tsigtao

= 0.9371 = 0.9371 y = 9.9308x + 891796 R

2 =

0.852 20 40 60 80 100 120 500,000 1,000,000 1,500,000 deltaic デ ー タ 電力消費量に基づくCO 2 排出量 （10

6ｔ）

そ の 他 ソウ ル Deltaic Data Deltaic Data CO2 emission from electric consumption(106t)

Others Others Seoul Seoul

= 0.8662

Regression analysis of CO Regression analysis of CO Regression analysis of CO Regression analysis of CO2

2 emission between

emission between emission between emission between DN and statistical electric power consumption DN and statistical electric power consumption DN and statistical electric power consumption DN and statistical electric power consumption

• f each prefecture in Korea and China
• f each prefecture in Korea and China
• f each prefecture in Korea and China
• f each prefecture in Korea and China

• The average intensity of artificial stable light can be extracted

The average intensity of artificial stable light can be extracted from the DMSP/OLS from the DMSP/OLS nighttime im nighttime imagery by using NRF method. agery by using NRF method.

• A method to estimate electric

A method to estimate electric power consumption from DMSP/OLS power consumption from DMSP/OLS data is developed data is developed and the co nd the coefficient of determination efficient of determination R2=0.7551 =0.7551 was achieved. was achieved.

• CO

CO2 emission volume is calculat emission volume is calculated by using electric power ed by using electric power co cons nsum umption whic ption which extracted from h extracted from D DMSP/OLS LS s satellite imagery tellite imagery and the and the coefficient of determination R coefficient of determination R2=0.7225 was achieved. =0.7225 was achieved.

• The average intensity of artificial stable light can be extract

The average intensity of artificial stable light can be extract The average intensity of artificial stable light can be extract The average intensity of artificial stable light can be extracted ed ed ed from the DMSP/OLS from the DMSP/OLS from the DMSP/OLS from the DMSP/OLS night nighttime i me imagery by using NRF method. agery by using NRF method. night nighttime i me imagery by using NRF method. agery by using NRF method.

• A method to estimate electric

A method to estimate electric power consumption from DMSP/OLS power consumption from DMSP/OLS A method to estimate electric A method to estimate electric power consumption from DMSP/OLS power consumption from DMSP/OLS data is developed data is developed data is developed data is developed and the coeff and the coefficient of determ cient of determination ination and the coeff and the coefficient of determ cient of determination ination R R2

2=0.7551

=0.7551 =0.7551 =0.7551 was achieved. was achieved. was achieved. was achieved.

• CO

CO CO CO2

2 emission volume is calculat

emission volume is calculated by using electric power ed by using electric power emission volume is calculat emission volume is calculated by using electric power ed by using electric power co cons nsum umption whic ption which extracted from h extracted from D DMSP/OLS LS s satellite imagery tellite imagery co cons nsum umption whic ption which extracted from h extracted from D DMSP/OLS LS s satellite imagery tellite imagery and the and the coefficient of determination R coefficient of determination R and the and the coefficient of determination R coefficient of determination R2

2=0.7225 was achieved

=0.7225 was achieved =0.7225 was achieved =0.7225 was achieved. . . .

Summary Summary Summary Summary Summary Summary

Further Study Further Study Further Study Further Study

・ ・Improve the Improve the Deltaic model to get mo Deltaic model to get more accurate compensation of re accurate compensation of Improve the Improve the Deltaic model to get mo Deltaic model to get more accurate compensation of re accurate compensation of DN. DN. DN. DN. ・ ・Additional e Additional evaluatio aluation of

• f this method will be

this method will be done done to adapt for to adapt for Additional e Additional evaluatio aluation of

• f this method will be

this method will be done done to adapt for to adapt for

• ther countries, especially the country used deferent type o
• ther countries, especially the country used deferent type o
• ther countries, especially the country used deferent type o
• ther countries, especially the country used deferent type of

f f f energy such as atomic energy. energy such as atomic energy. energy such as atomic energy. energy such as atomic energy.