The MRV system of greenhouse gaseous emission

in the world high-carbon reservoirs

Hironori AraiMasayoshi ManoNaoko SaitohKazuyuki Inubushi

Dr. Takeuchi Lab. IIS Tokyo Univ. xx

Development of MRV system of Methane emissions from

Rice paddies in the Mekong delta

Hironori Arai1,3), Wataru Takeuchi1),Kei Oyoshi2), Lam Dao Nguyen4),Towa Tachibana5), Ryuta Uozumi,

Koji Terasaki3), Takemasa Miyoshi3),Hisashi Yashiro3), Kazuyuki Inubushi5)

Modified from Yasuoka 2015

Cycle from Observation to Countermeasure

To target To solution To target

Observ.Survey

ModellingSimulation

Forcast・eval.Solutiondesign

optimization

management

Evaluation

DATAinformation

Estimation

Strategy Policy

EffectObservation of the effect

GOSAT K-computer,Fugaku

Each country must submit INDC (Intended Nationally Determined Contributions) to UNFCCC before 2020

3

Yan et al.2009 Xuan and Matsui, 1998

alluvial soil

1.1 Mha, 28%

acid sulfate soil

1.1 Mha, 28%potentially acid sulfate

0.5 Mha, 13%

431st / Oct / 2011

Rice farmers participatory field observation

6

Characteristics of the Mekong delta

Arai et al., 2018

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100% そのまま圃場等に放置

家畜飼料

その他作物の栽培

キノコ栽培（自身が栽

培）

知人に譲渡

売却（売却先では主にキ

ノコ栽培に用いられる）

焼却Winter-Spring

*n=50Straw use in TL2, Can Tho

(ha ha-1)

Spring-Summer

Summer-Autumn Burned

Sold to mushroom farmers

Free transfer to mushroom farmers

Compost for Mushroom cultivation

Compost forvegetable cropping

Feedstuffs forlivestocks

Left on soil

n=50

Greenhouse gas emission derived from rice straw use

8

0

1

2

3

4

04

69

21

38

18

42

30

27

63

22

36

84

14

46

0

CO flux

CO2 flux

Times after ignition (minutes)

Em

issi

on r

ate

(gC

/ h)

CO

CO2

Straw 40kg Moisture 20%

・CH4・N2O・VOC・O2・LEL

40kg 10kg Inlet

Outlet

Greenhouse gas emission derived from straw burning - Comparison among different straw size and moisture -

0

4

8

12

16

20

土壌還元 焼却 持ち出し

0

30

60

90

120

- CF AWD CF AWD CF AWD

0

700

1400

2100

2800

3500

土壌還元 焼却 持ち出し

■Conventional ■ Straw-use■AWD

- Reduction of irrigation rate & GHGs (2012-2016)- Increase of rice grains and its quality

GH

Gs

em

issi

on

(t- C

O2

eq.h

a-1

year

-1)

60%reduction

30%reduction

Irrigation r

ate

(mm

ha-

1ye

ar-1

)

Gra

in y

ield

(t- 1

4%

mo

istu

re h

a-1

Year

-1)

C/N

ratio o

f gra

ins

0

7

14

21

28

3510%increase

ImprovedRice quality

Strawincorporation

Strawburning

StrawTaken out

Strawincorporation

Strawburning

StrawTaken out

Strawincorporation

Strawburning

StrawTaken out

Strawincorporation

Strawburning

StrawTaken out

Philippines CF

Philippines AWD

Arai et al., 2018

PALSAR-2 Lv.1.1

quadruple6m

Volumediffusion

Coherency matrix [T3] generation

Speckle filtering

Lee-refined (7×7)

Field water level & cropping calendar

CH4 flux data & Soil map

Sampling (25-30 pixel）from each ROI

・Statistical analysis

CH4 emission estimation

model

Inundated / non-

inundated paddies

characterization

Freeman-Durden

decomposition

Single scattering

Doublebounce

Field data collection

Hierarchical bayesian modelling

MODIS (MOD13Q1)Normalized Vegetation/Water index,

Days after sowing, 1km

AMSR-218.7 & 23.8 GHz (V), 10km

Normalized frequencyIndex

Local Maximum Fitting-Kalman filter

Validation&

Data Fusion

HH, HV σ0Enhanced Lee filter (3×3)

PALSAR-2 Lv.1.1

SCAN-SAR25-100m

Multi-looking

Co-registration

De-Grandimulti-temporalspeckle filtering

Geocoding,Radiometric calibration

HH, HV σ0

Local Incidence

angles

Inundated / non-

inundated paddies

classification

Flow chart

CH4 emission map

ValidationGOSAT

(Retrieval)10km

NICAM-TM(LETKF)

Field observation & flux modeling

Remote sensing on field water & vegetation

Validation & integration

with daily EO data

Verification with GOSAT

Straw incorporation timeand amount

Water regime prior to rice cultivation

Water regime duringrice cultivation

A. ①

②

B.

③

CROPflood

CROPNon-flood

CROPNon-flood

>30 days

180 days

IPCC guideline (Tier1)[Emission factor × Scaling factor in IPCC guideline]

Sow

ing

day

s (d

ays

afte

r 2

00

1/0

1/0

8)

The ad

jacent fallo

w’s len

gth (d

ays)Cropping calendar evaluation with

MODIS-NDVI (LMF-KF) for GCOM-C

Samples of paddies

Paddy A (x:184, y:229)Double cropping → Triple cropping

Paddy B (x:185, y:220):Triple cropping

Arai et al., 2018

DAS←days after sowing

β ←acid sulfate・coastal sandy soil coefficient

CH4 emission on a specific date= γ * carbon_management / non-inundated_fallow / inundated_fallow * water_management *α*β

carbon_management (Michaelis-Menten KINETICS)=[exp(-DAS*δ)-exp(-DAS*(δ+ω))+κ]

non-inundated_fallow (OXYDATION CAPACITY)

= [1+EXP(-1* ζ *(DAS – ι* days of nonflooding days of the former fallow))]

inundated_fallow= EXP(ε * days of flooding days of the former fallow)

water_management= EXP(η * inundated days during the last 10days)

γ, η, δ, ε , ω , ζ , ι, κ←constant (>0)

α ←straw incorporation coefficient

Semi-empirical daily CH4 flux (mg C m-2 hr-1) Model

0 days non-flooding fallow 5 days non-flooding fallow

30 days non-flooding fallow 60 days non-flooding fallow

Ob

serv

ed 100

100101

10

1

0.1

0.01

0.001

0.01

Estimated

Da

ily C

H4

fluxe

s(k

g C

km

-2h

-1)

0

50

100

0 20 40 60 80100Days after sowing

CH

4fl

ux

Days after sowingCH

4fl

ux

Arai et al., 2018

ALOS-2/PALSAR-2– Lband-Synthetic Aperture Radar –

Xuan and Matsui, 199830paddies × 1village

5paddies × 4villages

Alluvial soils

1.1 Mha, 28%

Acid sulfate soils

1.1 Mha, 28%potential acid sulfate soils

0.5 Mha, 13%

Strip map

SCAN SAR

Coherency matrix [T3] generation

Polarimetric decomposition

Speckle filteringLEE refined

(7×7)

Sampling (25-30pixel）from each ROI

&Statistical analysis

Modified from Avtar et al. 2012

Freeman-Durden

Cloud-Pottier

PALSAR-2 Lv.1.1(quad. CEOS)

23 scenes

PALSAR-2 Lv.1.1(SCANSAR CEOS)

105 scenes

Multilooking

Co-registration

De Grandimulti-temporal

filtering

Geocoding&

Radiometriccalibration

HH HVIncidence

angle

Rice paddy masking&

Statistical analysis

Classification of inundated paddies and non-inundated paddieswhich is covered by rice plants

Inundated(fallow)

Inundated(cropping)

Non-inundated (cropping)

Non-inundated (fallow)

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1

Single scattering(%)1009080706050403020100

10

20

30

40

50

60

70

80

90

100

100

90

80

70

60

50

40

30

20

10

0

Dry fallow (+rice stumps)

Fallow after plowing or flooding fallow

0-20 days after sowing

21-40 days after sowing

Inundated Non-inundated

41-60 days after sowing

61-100 days after sowing

-Freeman-Durden decomposition-

0

Full-polarimetry (3m)SCAN-SAR (25m)

Fallow paddies

0-20 days after sowing

21-40 days after sowing

41-60 days after sowing

61-100 days after sowing

Inundated Non-inundated

HV

(d

B)

HH (dB)

-40

-30

-20

-10

-40 -30 -20 -10 0

(a) Quadruple-polarimetry

Arai et al., 2018

HV

+H

H t

hre

shold

(d

B*1000)

HH

thre

shold

(d

B*1000)

cosine(incidence angle)

cosine(incidence angle)

HH threshold (dB) = 0.550*HV +12.9*cosine(IA) -11.2

y = 242983x2 - 360742x + 77749R² = 0.7936

-60000

-57000

-54000

-51000

-48000

0.6 0.7 0.8 0.9 1

R² = 0.9321

-20000

-15000

-10000

-5000

0

0.6 0.7 0.8 0.9 1

HV -25dB

Arai et al., 2018

SCAN-SAR (25m)

Truth (Unknown)

Prediction

Analy-sis

Anal.

Pred.

Estimate actual irrigation practice

Irrigation/field water-level

Model simulation

Our data integration scheme

Simulation scheme with 25m-spatial resolution- Hysteresis of soil matric potential energy-

flood flood

Fie

ld w

ate

r le

vel (

cm)

0-10

-20

10

20

30

40

50

https://slideplayer.com/slide/5038747/

Irrigation, potential energy >> Side flow, ground water flow

Implicit RK4 integration model

Irrigation (init./boundary)

inundated

Not-inundated

Matric-potential at irrigation index (Di) = Σ(soil inundation rate before the irrigation, days after sowing, clay content)・αi

t = days after irrigation

𝑾𝑳 = field water level

Gravitational-potential at irrigation index (G) = field water level after irrigation *β

𝑑𝑾𝑳

𝑑𝑡= 𝜸∗exp δ ∗ 1 − log exp Di ∗ t − G + 2 + exp −Di ∗ t − G ∗ Di ∗ t − G

−δ ∗ exp Di ∗ t − G − exp −Di ∗ t − G ∗ Di ∗ t − G

exp Di ∗ t − G + 2 + exp −Di ∗ t − G

+ Di ∗ 1 − log exp Di ∗ t − G + 2 + exp −Di ∗ t − G + rain-fall

Irrigation functionif 𝑾𝑳 < threshold :

irrigate (i.e., 𝑾𝑳 += X)

Irrigation (init./boundary)

Parameter update by the analysis with EO data

https://slideplayer.com/slide/5038747/

Irrigationthreshold

Irrigationthreshold

constraint

Modelstructure

Estimated (cm)

Ob

serv

ed (

cm)

-20

-15

-10

-5

0

5

10

15

-20 -10 0 10

cal

NICAM-TM(Chem)-LETKFwith AMSU, PREPBUFR and GOSAT/Sentinel-5P

Direct comparison between GOSAT and emission data is meaningless…

→GOSAT data assimilation with NICAM-TM!

Nonhydrostatic ICosahedral

Atmospheric Model-TM(Chem)

Local Ensemble Transform

Kalman Filter

Miyoshi., 2005

Terasaki et al., 2014

Observation operator

development

for GOSAT

Implementation of variable localization scheme in NICAM-TM-LETKF (PREPBUFR&GOSAT)

Back ground covariance matrices

Kang et al., 2012 → Flux parameter estimation!2014051718-1803 Glevel 6, Inflation with RTPS=1

Increment of XCH4 (ppb, 950 hpa) w/ VL

Increment of XCH4 (ppb, 950 hpa) w/o VL

Economic assessment of GHG mitigation under various uncertainties

Kalnay et al. 2017

0

50,000

100,000

150,000

200,000

250,000

300,000温室効果ガス(キノコ栽培)

温室効果ガス(野焼き)

温室効果ガス(水田)

菌種購入費

灌漑費

レベリング

脱穀

種（OM4218）

種（Jasmine 85）

Kali肥料

NPK肥料

DAP肥料

尿素肥料

0

50,000

100,000

150,000

200,000

250,000

300,000

常時湛水

常時湛水

常時湛水

常時湛水

稲わ 稲わ 持出 持出

きのこ

稲わら

米

# State and trends of the carbon market 2012,

# Spot and Secondary offset market 2011, 1,822MtCO2

= 23,250million USD

Arai., 2015

CF

AW

D

AW

D

AW

D

AW

D

CF

CF

CF

GHG from mushroom

GHG from field burning

GHG from paddy soil

Mushroom seeds

Irrigation cost

Land preparation

Grain threshing

Seed (OM4218)

Seed (Jasmine 85)

KCl

NPK

DAP

Urea

Rice

Straw

Mushroom

Expenditure

(JP

Y/y

ear)

Incom

e -

Expenditure

(J

PY/y

ear)

Incorp.Burn sell Mush-room

STRAWMANAGEMENT

Transparent MRV system on baselines/mitigation-effects with

satellite data is the key !