The potential of space borne imagery to The potential of space - PowerPoint PPT Presentation

The potential of space ‐ borne imagery to The potential of space ‐ borne imagery to quantify fossil fuel CO 2 emissions quantify fossil fuel CO 2 emissions Philippe Ciais Philippe Ciais LSCE, France LSCE, France 1

The global carbon budget • Emissions of fossil fuels are estimated from energy use statistics Satellite data have not been used in this estimate After Lequéré et al., 2014

The 2 ° C goal of the Paris Agreem ent requires a reversal of the trend of CO 2 em issions in the short term , and large and sustained reductions in the long term 3

Are w e losing the anchor of the carbon cycle ? Uncertainty of fossil CO 2 em issions has not decreased … Liu et al. based on new coal carbon content measurements from Chinese mines and coal samples Liu et al. Nature, 2015 • This 14% correction of emissions translates into adjusting the global land sink (residual) by 0.4 GtC yr ‐ 1 (~30%) 4

Local scale : tw o em ission m aps for transportation in the London area European data (Stuttgart University) Regional data from UKNAEI Source ‐ B. Thiruchittampalam IER, Stuttgart U. GgC y ‐ 1 per 1 km grid cell 5

Critical uncertainties in CO 2 Uncertain em issions from hotspots inventories No.1 emitting power plant? (postal address) ! So care must be taken... Source ‐ Courtesy T. Oda 6

How m easurem ents of atm ospheric CO 2 concentration can be used to support em ission reduction policies ? 7

Atm ospheric inverse m odeling: from CO 2 concentrations to em issions The atmosphere is a powerful integrator of surface fluxes  Atmospheric inverse modeling was already proven to be effective to quantify regional CO 2 fluxes at global, continental, regional scales 8

Em issions are highly localized – > 2 % of the w orld area contain 9 8 % of em issions  We do not have a dense sampling of the atmosphere in space and in time to elucidate the spatial details of fossil CO 2 emissions beyond continental scales 9

New generation of satellites to m onitor global anthropogenic em issions • Satellites offer an unparalleled global spatial coverage for monitoring greenhouse gas budgets. • Existing satellites are focused on natural fluxes, but imagery and long ‐ term observations provide the opportunity to estimate emissions and their trend . • Significant investments around the world are preparing future space missions to measure emissions for supporting climate policies. 10

Sam pling strategies to m onitor em issions by atm ospheric m easurem ents Sampling over the main Sampling across continental and emission areas : synoptic scales : cities, large power plants fossil CO 2 gradients and proxy tracers Proxy tracers of CO 2 emissions • Synoptic structures contain a mix of fossil and • Needs very high ‐ resolution measurements & natural CO2 transport models • In situ : may not be captured with a sparse • In situ : requires sampling of all large cities network Needs a new generation of satellites with • • Needs a new generation of satellites with imagery capabilities imagery capabilities • Small sources missed. May not give a complete • Emissions signals are rapidly diluted by 11 picture transport

The CO 2 m issions planned in the year of the Paris Agreem ent GEOCARB No European mission focused on anthropogenic CO 2 emissions 12

Tow ards a European operational m onitoring system of anthropogenic CO 2 em issions 2015 2017 • Policy relevant need for an observing system to • Operational System components monitor anthropogenic emissions • Detection of emissions from space Attributes of space infrastructure ( imagery of • • Modeling requirements for column CO2 concentration ) anthropogenic emissions and natural Attributes of complementary in ‐ situ infrastructure • fluxes Part of the Copernicus Programme Part of a coordinated international carbon observing system 13

Needed Capabilities Needed attributes of space observations of column CO 2 for emission monitoring  Dense sampling (imagery) : images of CO 2 plumes produced by emitting areas  High spatial resolution : capture emission hotspots and avoid clouds, pixel size < 3 km  High accuracy : resolve the small atmospheric gradients, individual precision ≈ 1 ppm  Global coverage European mission SCIAMACHY OCO ‐ 2 Microcarb GOSAT GOSAT for monitoring CO 2 (stopped) GOSAT (sounder) (sounder) (sounder) emissions (imager) CO 2 and CH 4 CO 2 CO 2 CO 2 and CH 4 Sampling sparse Resolution too low Sampling too sparse Imager : Sampling dense Accuracy good Accuracy too low Accuracy moderate Accuracy good 14

Measuring CO 2 from Space • Record spectra of CO 2 Retrieve variations in the Validate measurements to ensure column averaged CO 2 dry air and O 2 absorption in X CO2 accuracy of 1 ppm (0.25%) mole fraction, X CO2 over the reflected sunlight sunlit hemisphere Initial Generate Surf/Atm Synthetic State Spectrum GOSAT and OCO ‐ 2 Instrument New Model State (inc. Difference X CO2 ) Spectra Tower Aircraft Inverse Model Flask FTS X CO2 15

What are the capabilities of a constellation of CO 2 imagers (LEO) to quantify fossil CO 2 emissions ? 16

Single Single sa satellit llite S7 S7 ‐ 1: 1: 1 da day an anim imatio ion (350 (350 km km) • MODIS Terra 1km x 1km MOD35 L2 cloud mask as baseline for the orbit • MODIS provides (nadir centered) LON, LAT, SAZ, CFC, Time (+60min=11:30) • Computed from modified time (and LON, LAT): SUZ, AZI, GLI • All other quantities read from geophys. data bases according to LON, LAT, Time 17

First step : clum ps of em itting pixels that can create plum es detected from space A Clump : “ a cluster of emitting pixels whose CO 2 emissions can be detected from space “ • Principle : adjacent high emitting pixels are grouped together • The plume of a clump can be detected even if plumes of component pixels may not • Difficult problem : cannot use simply administrative boundaries (e.g. cities) because of hot ‐ spots near urban areas and complex patterns of urban emissions (multiple centers) Aggregation : the method retained is based on detection thresholds and an inverse distance attraction of pixels emissions Threshold: 1140 tonC a ‐ 1 64% of global total 73% of global total Threshold: 515 tonC a ‐ 1 18 Blue background: Urban area from Natural Earth project database

Clum ps distribution in Northern populated regions Europe China USA 19

Second step : a global 1 km inversion system Observation at: day 44: 1145 Green: center not in the window; response ignored Red: center in the window; 500 km response computed day 44: 1150 Orange: center in the window; response computed “Inversion window” Black dots: Emission ‐ weighted center of a clump Dark blue: Track of a time window that is targeted day 44: 1155 Light blue: Pre ‐ and post ‐ overpasses Strategy : global coverage of all emitting clumps, Bayesian inversion 100 billions of plumes forming H 100 billions of plumes forming H Preserve high spatial and temporal resolution attributes of emissions Numerically optimized Numerically optimized Allows imagers LEO and GEO OSSE Full OSSE takes ≈ 1 week CPU Full OSSE takes ≈ 1 week CPU 20 Simple transport model (Gaussian)

How m any days can w e constrain em issions of a city from space? Three cities in France Dijon Paris Marseille 0.8 MtCO 2 30 MtCO 2 8 MtCO 2 Good Bad Number of days with an Paris Marseille Dijon uncertainty reduction better than: 115 days 150 30 20% 70 105 10 50% 80% 35 10 1 21

Num ber of ‘good days’ w ith an uncertainty reduction of at least 5 0 % with one space ‐ borne imager More than 30 days You want more than 12 ‘good days’ 21 clumps 39 clumps 28% of national 32% of national emission emission … and more than 100 days More than 50 days 3 clumps 19 clumps 4% of national 26% of national emission emission 22

Up ‐ scaling to annual emission budgets Importance of temporal error correlations strong correlation intermediate correlation intermediate correlation strong correlation If we know emission from one day, do we know emission from another day? To address this question, systematic tests were performed 23

Up ‐ scaling to annual emission budgets Importance of temporal error correlations strong correlation intermediate correlation intermediate correlation strong correlation 24

Co Constella llatio tion C4: C4: 1 da day an anim imatio ion (35 (350 km km) Also test GEO configurations 25

Can proxy tracers help to constrain fossil fuel CO 2 emissions? 26

Separation of the fossil fuel CO 2 em issions using tracers Proxy tracers are co ‐ emitted with fossil fuel CO 2 through combustion processes Transportation sector Energy sector Column NO x Energy sector Column CO Konovalov et al. 2016 ACPD

Separation of the fossil fuel CO 2 com ponent using tracers Additional constraints on fossil fuel CO 2 emissions of the EU from satellite measurements of NO x and CO From NO x (OMI) From CO (IASI) Potential of S5, MTG future CO observations Combining NO x and CO Uncertain variations in emission factors Konovalov et al. 2016 ACPD 28