atmospheric monitoring of co2 emissions: an innovative ... wu... · atmospheric monitoring of co2...
Post on 23-Jul-2020
10 Views
Preview:
TRANSCRIPT
MRV full day workshopBonn, June 3, 2014
Atmospheric monitoring of CO2 emissions:an innovative solution for cities
Lin Wu, Grégoire Broquet, Felix Vogel, Philippe Ciais,Valentin Bellassen, François-Marie Breon, Irène Xueref-Remy and collaborators
LSCE – CEA/CNRS/UVSQ, Gif-sur-Yvette, FranceChaire BridGES ( UVSQ / CEA / CNRS / Thales Alenia Space / Veolia )
CDC-Climat
MRV GHG emissions under existing and developing carbon pricing mechanismsJune 3, 2014
1
MRV full day workshopBonn, June 3, 2014
Duren & Miller, Nature 2012
Dominant city CO2 emissionsIPCC working group 3: recommendation/challenges for city CO2 monitoring
MRV full day workshopBonn, June 3, 2014
Urbanisation decreases per capita emissions in developed countries
City CO2 emissions: challenge or chance?
MRV full day workshopBonn, June 3, 2014 4
Paris Climate Plan-30% GHG 2004-20-75% GHG 2004-50
Los Angeles:GreenLA & ClimateLA-35% GHG 1990-2030
Many others (US-EPA & C40)Secotral climate plan(energy, transport, building)
Cities in actions: political plans
MRV full day workshopBonn, June 3, 2014
Cities in actions
5
IPCC WG3: lack of comparable/reference methods/data for local/urban carbon accounting
• Vulantary reporting Networks & associations
C40 network, Carbon Disclosure Project
MRV full day workshopBonn, June 3, 2014
City CO2 MRV: down to sectors
6
MRV full day workshopBonn, June 3, 2014
MRV Sector: Regional & City Territorial Inventories
7
� Uncertainties (input data & emis. Factors)� Accuracy (uncertainties)
� Not obligatory , diverse voluntary tools� Needed standardized/harmonized protocal
Ian Cochran
MRV full day workshopBonn, June 3, 2014
City Fossil fuel CO 2 emission inventories
EDGAR10 AIRPARIF05 AIRPARIF08 IER05 Max - Min
Resolution 0.1°x 0.1° 1km x 1km 1km x 1km 1min x 1min
Annual budget of Ile-de-France (TgC)
Residential 6.75 5.65 7.80 5.36 31%
Road 8.50 3.63 3.37 6.03 60%
Industry 5.19 3.02 3.09 4.61 42%
Total 24.65 12.34 14.26 16.39 50%
50% difference at 25km when disaggregating into regions (Duren and Miller 2012)
TYPE 1 Existing inventories
• Research based tools (EDGAR, IER, PKU): top-
down, global coverage; uncertainty can be
high
• Simple budget methods (low price, no
assessment of the uncertainty)
• High quality inventories by local agencies (e.g.
AIRPARIF) ~20% monthly uncertainty
TYPE 2 Potential inventories if
incentive in the future
• MRV practice (e.g.CITEPA); verified
national uncertainty ~5%; following
IPCC guideline
MRV full day workshopBonn, June 3, 2014
Innovative approach: atmospheric inversion (constraint from conc. observations)
9
Bottom-up
Top-down
CityNetworkPlatform
Algorithm
• Observations linked with area emissions by numerical models
• Have been used for 15 years in scientific community to estimate natural fluxes at large scales
• Emergent regional/local applications: ability to estimate anthropogenic fluxes
MRV full day workshopBonn, June 3, 2014
Cities with monitoring/modeling/inversion projects
10
CO2MegacitiesParis London LA IndianapolisSalt lake city Houston Shenzhen …
Cities in actions: scientific attempts
MRV full day workshopBonn, June 3, 2014
In-situ network of 5 stations for Paris CO 2 monitoring
� High-precision, multi-species observations� Picarro(45k€ )+ add cost: 1 M€ / ~10 Picarro
Model G1302 (CO2/CO/H2O)
CO2 CO
Accuracy < 0.13 ppm <11 ppb
Precision < 0.1 ppm
(0.38 ppm for EIF)
<10.2 ppb
Irène Xueref-Remy et al, AGU 2013 & article in prep.
Fossil Fuel (FF) + Bio = Total
Prior CO 2 emission Posterior CO 2 emission
Data selection
Clipping out many data,but still have considerable uncertainty reductions
• Afternoon data• Wind > 3ms-1
• Obs gradient but not obs
Inversion results, Breon et al. 2014
MRV full day workshopBonn, June 3, 2014
Extending network with cost effective sensors?
Cost Comments
Sensors / site 2-45 k€
Infrastructure / site
• Container
• Inlet + R-Box + drying
• Calibration
• Others
8-30 k€
10 k€
5 k€
3-5 k€
1 k€
Mininum:
Inlet +
calib
= 8 k€
Annual
• Infrastructure / site
• 1 engineer: network
• 1 engineer: data
• 1 engineer: modeling
0-10 k€
60 k€
60 k€
60 k€
1 Number of sites 70
Cos
t (k€
)
• Cheap sensors• Using existing air quality network
Picarro + maximum cost / site
180
1000Table: detailed cost of monitoring network
MRV full day workshopBonn, June 3, 2014
Objectives and accuracies targets
13
Monte-Carlo simulations of a linear reduction model (5% annual accuracy in a 15-yr horizon)
Inversion objective: 5% annual uncertainty
• Same as (national) inventories for MRV use; thus an alternative/verifying approach• Ability to track graduately the 15-yr trend
• Sectorial verfications
MRV full day workshopBonn, June 3, 2014
All sectors: FF total
Waste
Air, rail, ship
Road transport
Production
Residential
…
Sectorial Inversion of CO 2 fluxes using data from extended networks
Zone1 Zone2 Zone3
Zone4 Zone5
CO2Megaparis
Ext
ende
d
10 sites
30 sites 50 sites
70 sites
MRV full day workshopBonn, June 3, 2014
y
?=x
15
Gaussian unbiased assumption� Obs: � Prior:
� Posterior:
Emis/factorTransport
Obs. Err.Obs
10 sites, number of assimilated gradients ~ 800
Data selectionAfternoon data; Wind > 3ms-1;choose one upwind station tocompute the concentration gradients
Obs uncertainties: transport error + aggreg. error + instru. Error
Control scaling factorsSpatio-temporal disaggregation (diurnal/zone)Monthly prior error budget: ~20%
Sectorial Inversion of CO 2 fluxes using data from extended networks
MRV full day workshopBonn, June 3, 2014 16
• Reduction of uncertainties for total FF with very limited number of stations (~10 sites)
• More sites to reduce uncertainties for spatio-temporally distributed sectorial emissions
Aggregate monthly emissions to annual emissions: 10% 5 %
Sectorial Inversion of CO 2 fluxes using data from extended networks
Public
Industry
Energy
Waste
CommercialResidentialInstitutional
Road
AirlineAgriculture
AIRPARIFInventory
MRV full day workshopBonn, June 3, 2014
Summary
17
• Results to be refined for large networks, for trend detection per sector, and to analyze the ability to solve for the spatial distribution
• Existing inversion framework: ability to monitor the city total emissions with high precision at a reasonable cost
• Near term inversion framework: ability to monitor the city sectorial budgets with high precision at a reasonable cost
� Atmospheric inversion and MRV systems:
• complementary to bottom-up approaches (Bayesian estimation)
• a solution for a fully independent budget verification method
• a robust sectorial monitoring method in a near term
MRV full day workshopBonn, June 3, 2014
The Forest Project
Courtesy of P. Bicheron and F. Maignan
Funded by the Climate KIC
An innovative monitoring service package for tropical forest carbon stocks and fluxes
(trusted, fully integrated, cost-effective) enabling forest carbon projects to achieve
carbon emission reduction and / or stock enhancement
MRV full day workshopBonn, June 3, 2014
� Target :� REDD, Afforestation/Reforestation projects, CDM, voluntary carbon market at national and project (local / regional) levels� prove positive impacts of forest projects linked or not to the carbon economy
Framework
Airbus DS satellites constellation – Spot, Pleiades, TerraSAR,
TanDEM-XBiomass production
based on a simulation of the ORCHIDEE model (gC m-2 yr-1)
� Based on fully optimized and reliable tools:� Observation from space, optical and radar (Astrium)� In situ measurements for calibration (AMAP, ONFI)� Carbon fluxes and stocks modeling (LSCE)� Quality Assurance, calibration, validation, accuracy assessment (NPL)� Cost effective capacities driven by market and end users (CDC-Climat, ONFI)
MRV full day workshopBonn, June 3, 2014
Continuous forest monitoring/mapping
Pleaides product – 50cm. Monitoring of a planting project -Brazil
Land cover classification with TanDEM-X –
Kalimantan, Indonesia
� Land cover / Land use change maps
� Forest base and degradation mapping
� Canopy texture features and forest structure(biomass, height)
MRV full day workshopBonn, June 3, 2014
Integration of satellite datastreamsUsing a process based forest model
� Estimates of sources and sinks due to forest management types, based on different scenarios over a 30 years time horizon (deforestation, degradation, and enhancement)
� Estimates of uncertainties (input data and model parameters)
� Model calibration against plot scale data in selected regions (Guyana, Gabon)
Evolution of circumference distribution (m) for a Tropical broadleaf evergreen grid box
> 22 - 1.81.6 - 1.81.4 - 1.61.2 - 1.41.0- 1.20.8 – 1.00.6 - 0.80.4 - 0.60.2 - 0.40.0 - 0.2
Diameter class
(m)
MRV full day workshopBonn, June 3, 2014 22
Products� Wall to wall quantification of tropical forest carbon stocks and fluxes
� Scale : from project to region and continent
� Accounts for :
� Effects of interannual variability and forest disturbance
� Net carbon balance
� Water and energy fluxes
� Possibility to integrate the calibrated model for future scenarios
Timeline� 2014 Model testing and calibration
� Mid 2015 Site level demonstrator
� End 2015 Prototype of regional applications
top related