CO2M science studies
Science studies focusing on the Copernicus CO2 Monitoring Mission (CO2M).
30 May 2022
26 March 2020
It will provide support on topics including methods for retrieval of the CO2 dry air mole fraction (XCO2), scientific algorithm data flows and functional requirements, and calibration, product validation and monitoring requirements.
At its core, three state-of-the-art retrieval methods are being further developed for the mission:
In pursuit of the highest quality retrievals, particular aims of the study are to take advantage of measurements from all instruments currently proposed for the CO2M platform and to support the identification of synergies between the three algorithms (referred to as ‘companion algorithms’).
Specific objectives include:
The scientific service support will focus on the retrieval of air-column weighted CO2 column concentrations (XCO2) and other climate relevant parameters (e.g. methane and Solar Induced Fluorescence (SIF)) using the following three methods for greenhouse gas retrievals from shortwave infrared, high spectral resolution spectrometers.
The fast atmospheric trace gas retrieval (FOCAL) algorithm has been initially developed for OCO-2 XCO2 retrieval, where XCO2 is the column-average dry-air mole fraction of atmospheric CO2.
FOCAL includes a radiative transfer model, which has been developed to approximate light scattering effects by multiple scattering at an optically thin scattering layer. This reduces the computational costs by several orders of magnitude, which is important especially for future sensors such as CO2M.
FOCAL's radiative transfer model is utilised to simulate the radiance in all three OCO-2 spectral bands, allowing the simultaneous retrieval of CO2, H2O, and SIF. Recently, FOCAL has been modified to also retrieve several other parameters (e.g., CH4 and N2O).
The FOCAL OCO-2 XCO2 product is used by the EU H2020 projects CHE and VERIFY, and algorithm improvements are carried out in the framework of the GHG-CCI+ project of ESA's Climate Change Initiative (CCI). The OCO-2 FOCAL XCO2 product is also used to contribute to the multi-sensor/multi-algorithm merged XCO2 products generated for the Copernicus Climate Change Service (C3S).
FOCAL's OCO-2 XCO2 accuracy and precision has been determined by comparison with collocated ground based TCCON (Total Column Carbon Observing Network) FTS (Fourier Transform Spectrometer) measurements. Regional-scale (station-to-station) biases amount to about 0.6 ppm. The single measurement precision (standard deviation of the difference to TCCON) is about 1.5 ppm. Mainly due to overly conservative cloud filtering using the MODIS MYD35 cloud mask data, probably too few soundings are currently processed and investigations are ongoing to enhance the data yield.
FOCAL is also used for various CO2M related assessments carried out in ESA studies related to CO2M error analysis, performance assessments and requirements consolidation, and for ESA's CO2M End-to-End simulator. Currently FOCAL is also used for the retrieval of XCO2 and other parameters from GOSAT and GOSAT-2.
CO2 retrievals from OCO-2 data via FOCAL v08 are shown in Figure 1.
The University of Leicester Full Physics (UoL-FP) retrieval algorithm is a sophisticated algorithm for the retrieval of greenhouse gas columns, based on Bayesian optimal estimation method. The retrieval employs a forward model that combines a multiple scattering radiative transfer solver with a fast 2-orders-of-scattering vector code to allow for polarisation. To accelerate the radiative transfer component of the retrieval algorithm, an accurate principal component analysis (PCA) method is adopted. Aerosols are described with two aerosol types representing a large and small mode of aerosols, which are created using scene-dependent aerosol information from the Copernicus Atmosphere Monitoring Service (CAMS).
UoL-FP has its heritage in the NASA OCO mission. It is used to:
UoL-FP is further developed to increase its retrieval speed, as required for a mission like CO2M, and to allow the use of aerosol information from the co-located MAP instrument. This development will be guided by intensive simulations and by application to current GHG satellites OCO-2/-3 and GOSAT-2 with its aerosol senor CAI-2.
The CO2M mission is required to measure XCO2 with high accuracy and precision. One of the main challenges of XCO2 satellite remote sensing is the presence of atmospheric aerosols, which are known to modify the light path of the backscattered sunlight and, in turn, introduce errors in the retrieved XCO2. To mitigate this, the MAP instrument is part of the CO2M mission, together with the CO2I spectrometer, measuring reflected radiances in the near and shortwave infrared.
The mission requires cutting-edge remote sensing approaches to provide XCO2 with the required accuracy, making synergistic use of the mission payload. At SRON, the RemoTAP algorithm is currently under development, accounting for three different potential processing lines of the CO2M ground segment: (1) XCO2 from only CO2I spectrometer measurements, (2) aerosol properties from the MAP in support of light path simulation of the CO2 retrieval, and (3) XCO2 from simultaneous MAP and CO2I observations.
Based on preliminary prototype developments, the numerical stability and efficiency of RemoTAP will be improved as part of this study. The software is based on algorithm heritage from RemoTeC, an operational algorithm used to infer CO2 and CH4 column abundances from spectrometer measurements of e.g. GOSAT, GOSAT-2, OCO-2, Sentinel-5 (and its precursor mission S5-P), and the operational algorithm used to infer aerosol properties from the SPEXone multiangle polarimeter of NASA’s PACE mission. The advantage of a joint MAP and CO2I data retrieval in terms of XCO2 data accuracy is illustrated in Fig. 3.
Given the strategic nature of the European integrated system, and its contribution towards assessing the effectiveness of CO2 emission strategies, the service is driven by a stringent set of requirements for the geophysical products, with rigorous targets for accuracy and uncertainty. Within this context, this activity aims to further consolidate the performance and accuracy of the algorithms to date and assess their feasibility as ‘candidate algorithms’ for the operational CO2M level-1b/c to level-2 processor.
All data flows available from CO2M level-1b/c (including MAP and CLIM data and visible channels) will be taken into account, as well as the required and available static and dynamic auxiliary information (including spectroscopy and numerical weather prediction (NWP) and climate modelling (CM)). Opportunities for improvements will also be explored through the use of companion retrieval results and assessed with respect to overall mission requirements.
The proposed algorithm improvements will be demonstrated and their performance evaluated, with both synthetic retrievals and, where possible, real satellite data (e.g. OCO-2/3 and/or GOSAT). This will include a special focus on improvements in the error analysis, and reduction of the overall uncertainties, also through use of the output of the companion algorithm results.
Under this activity the service will address the requirements definition and specification of all aspects of science data flows for the candidate algorithms, as needed for the future operational processor implementation in the PDGS. This includes the flow of ancillary static, dynamic and third-party data, the specification of all data-sets, interface analysis and the specification of the processing needs of the scientific methods applied, with the interaction between the companion level-2 algorithms within the functional processing as a particular focus.
The service also contributes to EUMETSAT’s definition of requirements for continuous instrument and product cal/val, and to the continuous CO2M level-2 product quality monitoring and product evolution, thus informing the specific methods to be implemented operationally and the objectives and procedures forming the CO2M cal/val and monitoring plan.
The service will establish, and maintain, a description of methods for algorithm validation with respect to accuracy, as well as possible systematic and structural random errors, following QA4EO guidelines.
It will also identify, describe and maintain validation data sets, including in situ observations and/or comparable products from other missions. It will also maintain a functional description of the data flows required for continuous calibration, validation and monitoring, including a description of all required (user) interfaces and a proposal for a set of reporting tasks. The development of a basic description of continuous monitoring methods to evaluate data-quality online is also included.