Contributors: Oleksandr Bobryshev (DWD)
Issued by: Oleksandr Bobryshev (DWD)
Date: 03/08/2023
Ref: C3S2_D312a_Lot1.2.3.1-v2.2_202308_PQAR_ECVSurfaceRadiationBudget_v1.2
Official reference number service contract: 2021/C3S2_312a_Lot1_DWD/SC1
History of modifications
List of datasets covered by this document
Related documents
Acronyms
List of Figures
List of Tables
General definitions
Table 1: Summary of variables and definitions
Variables | Abbreviation | Definition |
|---|---|---|
Surface Incoming Shortwave radiation | SIS | The total incoming shortwave (solar) flux, measured at the Earth’s surface. |
Surface Reflected Shortwave radiation | SRS | The total upwelling shortwave flux, measured at the Earth’s surface. |
Surface Net Shortwave radiation | SNS | The net downwelling shortwave flux, measured at the Earth’s surface (equal to SIS – SRS). |
Surface Downwelling Longwave radiation | SDL | The total downwelling thermal (longwave) flux, measured at the Earth’s surface. |
Surface Outgoing Longwave radiation | SOL | The total upwelling thermal flux, measured at the Earth’s surface. |
Surface Net Longwave radiation | SNL | The net downwelling thermal flux, measured at the Earth’s surface (equal to SDL-SOL). |
Surface Radiation Budget | SRB | The total net downwelling radiative flux, measured at the Earth’s surface (equal to (SIS+SDL) – (SRS+SOL)). |
Table 2: Definition of satellite data provessing levels
Processing level | Definition |
|---|---|
Level-1 | The full-resolution geolocated radiometric measurements (for each view and each channel), rebinned onto a regular spatial grid. |
Level-2 | Derived geophysical variables at full input data resolution and location as L1 source data (‘swath data’) |
Level-3 | Geophysical variables mapped on uniform grid, derived from multiple satellites/radiometers, averaged over a specific time, such as monthly and daily means. |
Table 3: Definition of special terms
Term | Definition |
|---|---|
Brokered product | The C3S Climate Data Store (CDS) provides both datasets produced within the C3S and so-called brokered products. The latter are existing products (data) produced under an independent programme or project which are made available through the CDS. |
TCDR | It is a consistently-processed time series of a geophysical variable of sufficient length and quality. |
ICDR | An Interim Climate Data Record (ICDR) denotes an extension of TCDR, processed with a processing system as consistent as possible to the generation of TCDR. |
User requirements | Depending on the different user needs, different product requirements may be applied and they are used to evaluate validation results. This document uses three accuracy categories: Optimal: Ideal accuracy which meets requirements for global and regional climate analysis. Target: Accuracy that meets requirements for global and regional climate modelling. Threshold: Minimum accuracy that meets requirements for operational climate monitoring and climate services |
Scope of the document
This document is the Product Quality Assessment Report (PQAR) for CLARA Product family CDR. This document provides validation results for the Climate Data Record (CDR) of the Surface Radiation Budget.
In the scope of the Copernicus Climate Change Service (C3S), two Surface Radiation Budget datasets are combined into one Climate Data Record (CDR) called “CLARA product family”.
The first dataset, EUMETSAT’s CM SAF CLARA-A2.1, is the core of this CDR. The second dataset, generated specifically within the C3S project, includes extra data products that are not included in the CLARA-A2.1 dataset. The core data products are brokered from CM SAF. The extra net fluxes are not included in the brokered CLARA-A2.1 dataset and are complimentary data provided for the convenience of users. Their format is as close as possible to the CLARA-A2.1 and the datasets are meant to be used together. Both datasets are frequently updated with Interim Climate Data Records (ICDRs) or simply extensions, generated using the same software and algorithms to cover more recent periods. The ICDR part for all products covers period from 2019 to 06/2021.
This separation into two datasets is necessary to keep the origin of the data, e.g. licence affiliations: “EUMETSAT’s CM SAF” and “Copernicus”, clear for individual products (Table 4).
Table 4: Licence overview of the CLARA product family Surface Radiation Products, including products SIS, SDL, SOL, SRS and Net Fluxes. Products marked in blue are taken from the CLARA-A2.1 TCDR and the remaining products are Extra Data Products produced in the C3S project.
Year | CDR Type | CLARA Product Family | |||
|---|---|---|---|---|---|
| CLARA-A2.1 | Extra data products | ||||
| SIS | SDL, SOL Longwave fluxes | SRS | Net fluxes | ||
1982 – 2018 | TCDR | CM SAF (CLARA-A2.1) | CM SAF (CLARA-A2.1) | C3S | C3S |
2019 – 06/2021 | ICDR | CM SAF (CLARA-A2.1 ICDR) | C3S | C3S | C3S |
The longwave fluxes are not included in the CM SAF ICDR plans for the current phase. To ensure the dataset integrity and continuity, they are calculated within the C3S for the ICDR part. As such, the longwave fluxes change their licence affiliation, namely they are provided within the C3S project for the ICDR part (2019-06/2021) and are brokered from EUMETSAT’s CM SAF for the TCDR part (1982 to 2018).
Furthermore, the CM SAF CLARA-A2.1 dataset has a temporal coverage of January 1982 to June 2019 (as described in [D2]), but TCDR data are only being brokered to the CDS up to December 2018. Data available from the CDS for January 2019 to June 2021 are brokered/derived from the CLARA A2.1 ICDR1.
In contrast to the original CM SAF CLARA-A2.1 dataset, the brokered service within Copernicus solely provides:
Level-3 data – excluding the level-2b data format.
Data on a global equal angle grid – excluding the polar grid format.
An aggregated version of all satellite data – excluding the provision of the individual satellite datasets.
1 The CM SAF CLARA-A2.1 and CLARA-A2.1 ICDR have a 6 month overlap from 01/2019 to 06/2019. One of the input datasets used in CLARA-A2.1, namely ERA-interim reanalysis data, is only available until August 2019 and dates after that are only covered by the newer ECMWF reanalysis version (ERA-5) and ECMWF IFS model. The CLARA-A2.1 ICDR uses ERA data from January 2019 onwards. The 6-month overlap between CLARA-A2.1 and CLARA-A2.1 ICDR is needed to ensure a sufficient period for data comparison and validation. CM SAF ATBD-ICDR contains more information on the overlap period [D8].
Executive summary
The brokering service of the CLARA-A2.1 data products includes 37 years (1982-2018) of level-3 data (monthly means) on a regular global latitude-longitude grid (with 0.25° x 0.25° resolution), merged from various polar orbiting satellites. It includes three products: the Surface Incoming Shortwave radiation (SIS), the Surface Outgoing Longwave radiation (SOL), and the Surface Downwelling Longwave radiation (SDL). SIS daily averages are also available within C3S. This CDR is brokered from EUMETSAT’s CM SAF. Therefore, this document only refers to the validation results from the original EUMETSAT CM SAF Validation Report [D1, D7]. It describes the validation methodology and the validation results.
The extra data products (produced specifically within the C3S project) are: the Surface Reflected Shortwave radiation (SRS), the Surface Net Shortwave radiation (SNS), the Surface Net Longwave radiation (SNL), and the Surface Radiation Budget (SRB). Validation results for these extra data products are described within this document. Validation results for SDL and SOL ICDR data products are described within this document.
An executive summary of the evaluation of the CLARA-A2.1 and CLARA-based ICDR surface radiation dataset can be found in CM SAF Validation Report, Section 1 [D1, D7].
The predefined requirements for accuracy of the SIS, the SOL and the SDL are given in CM SAF Product Requirement Document (PRD) [D2], Annex A. The achieved accuracies of CLARA-A2.1 correspond to the target accuracy requirements, only SIS daily accuracy corresponds to threshold accuracy. The validation methodology for the extra data products is selected in a way to provide the most conservative accuracy estimation, i.e. by assuming that the uncertainties in the input datasets align perfectly and contribute fully to the overall uncertainty. A summary of accuracy for CLARA-A2.1 and Extra data product is given in Table 5 and Table 6. Achieved accuracy and stability results allow consistent quantification of mean values, anomalies, variability and the Earth energy budget in general.
Table 5: Summary of the accuracy of the brokered CLARA-A2.1 data products; [D1] Section 1 for SIS, SOL, SDL TCDR; [D7] Section 1 for SIS ICDR. The accuracy estimation of SOL and SDL ICDR extensions is performed within the C3S using the algorithms developed by CM SAF
Product Name | Dataset accuracy [W/m²] | |
|---|---|---|
TCDR | ICDR | |
SIS monthly means | 9.5 | 8.7 |
SIS daily means | 18.6 | 22.1 |
SDL | 8.1 | 7.2 |
SOL | 13.8 | 9.9 |
Table 6: Summary of the accuracy of the extra data products developed within the C3S.
Product Name | Propagated accuracy [W/m²] | |
|---|---|---|
TCDR | ICDR | |
SRS | 7.8 | 6.6 |
SNS | 13.0 | 12.0 |
SNL | 21.9 | 17.1 |
SRB | 34.9 | 29.1 |
1. Product validation methodology
In the scope of the Copernicus Climate Change Service (C3S), two Surface Radiation Budget datasets are combined into one Climate Data Record (CDR) called the “CLARA product family”.
1.1 Validation methodology for SIS, SOL, SDL
CLARA-A2.1 data products include the Surface Incoming Shortwave radiation (SIS), the Surface Outgoing Longwave radiation (SOL), and the Surface Downwelling Longwave radiation (SDL) datasets. These datasets were validated against reference data records from surface measurements obtained by the Baseline Surface Radiation Network (BSRN)2. The reference dataset is described in detail in CM SAF Validation Report [D1], Section 4.
Therefore, this document refers to the original EUMETSAT CM SAF Validation Report [D1, D7]. It describes the validation methodology and the validation results. The accuracy estimation of SOL and SDL ICDR extensions is performed within the C3S using the algorithms developed by CM SAF.
The validation of the SIS ICDR (2019-06/2021) product is described in CM SAF Validation Report ICDR, Section 5.2 [D7]. The validation methodology of SOL and SDL ICDR extensions is performed within the C3S using the algorithms developed by CM SAF.
The validation methodology and definitions of the main metrics are provided in the C3S Product Quality Assurance Document (PQAD) [D4], Section 3, which refers to CM SAF Validation Report [D1], Section 5 and 5.1.
2 Ohmura et al., 1998, Driemel et al., 2018
1.2 Validation methodology for CLARA family extra data products
The performed validation approach for extra data products is based on the method of error propagation. These data products are the Surface Reflected Shortwave radiation (SRS), the Surface Net Shortwave radiation (SNS), the Surface Net Longwave radiation (SNL), and the Surface Radiation Budget (SRB).
The ICDR part data are validated in the same way as the TCDR.
1.2.1 Surface Reflected Shortwave Radiation (SRS)
The accuracy of the SRS is determined based on the accuracy of the Surface Incoming Shortwave Radiation (SIS) from the Surface Radiation Budget brokered from EUMETSAT’s CM SAF CLARA-A2.1 dataset [D1], and the Surface Albedo (SAL), not provided within the C3S.
The accuracy for the SRS is calculated using the fixed accuracy values for the SIS (ΔSIS = 10 W/m2) and the relative error of the SAL (ΔSAL = 25% of the SAL value). These values are the target requirements for accuracy from CM SAF Product Requirement Document [D2]. Resulting global mean accuracy for the SRS is 7.8 W/m2 for TCDR and 6.6 W/m2 for ICDR. Figure 1-1 illustrates the spatial distribution of the propagated SRS accuracies.
Figure 1-1: Spatial distribution of the propagated SRS accuracies for 1982-2018 (left) and 2019-06/2021 (right)
1.2.2 Surface Net Shortwave Radiation (SNS)
The accuracy of the SNS is determined based on the accuracy of the SIS and the SAL datasets. Estimations of the accuracy of the SIS product is given in CM SAF Validation Report [D1] and the accuracy of the SAL product is given in CM SAF Validation Report, Surface Albedo [D5].
The accuracy for the SRS is calculated using the fixed accuracy values for the SIS (ΔSIS = 10 W/m2) and the relative error of the SAL (ΔSAL = 25% of the SAL value). These values are the target requirements for accuracy from CM SAF Product Requirement Document [D2]. Resulting global mean accuracy for the SNS is 13.0 W/m2 for TCDR and 12.0 W/m2 for ICDR. Figure 1-2 illustrates the spatial distribution of the propagated SNS accuracies.
Figure 1-2: Spatial distribution of the propagated SNS accuracies for 1982-2018 (left) and 2019-06/2021 (right). There are gaps in the Arctic and Antarctic areas due to the difficulties in distinguishing between clouds and snow covered surfaces.
1.2.3 Surface Net Longwave Radiation (SNL)
The accuracy of the SNL is defined by the accuracy of the Surface Downwelling Longwave Radiation (SDL) and the Surface Outgoing Longwave Radiation (SOL), both datasets are the Surface Radiation Budget brokered from EUMETSAT’s CM SAF CLARA-A2.1 [D1].
Estimations of the accuracy of the SDL and the SOL products are given in CM SAF Validation Report [D1].
The accuracy for the SNL is calculated using the fixed accuracy values for the SDL (ΔSDL = 8.13 W/m2) and the SOL (ΔSOL = 13.77 W/m2). Resulting global mean accuracy for the SNL is 21.9 W/m2. Global mean accuracy for the SNL ICDR is 17.1 W/m2.
1.2.4 Surface Radiation Budget (SRB)
The accuracy of the SRB is defined by the accuracy of the SNS and the SNL datasets.
The accuracy for the SRB is calculated using the fixed accuracy values for the SNS (ΔSNS = 13.0 W/m2 ) and the SNL (ΔSNL = 21.90 W/m2 ). Resulting global mean accuracy for the SRB is 34.9 W/m2. Global mean accuracy for the SNL ICDR is 29.1 W/m2.
2. Validation results
2.1 Validation results for SIS, SOL, SDL
The TCDR (1982-2018) validation results are fully described in CM SAF Validation Report [D1], Section 5. The validation results for the SIS are provided in CM SAF Validation Report [D1], Section 5.2. The validation results for the SOL are provided in CM SAF Validation Report [D1], Section 5.3. The validation results for the SDL are provided in CM SAF Validation Report [D1], Section 5.4. Considerations for climate applications are provided in CM SAF Data Set Description [D3], Section 6. Conclusions for the validation results are provided in CM SAF Validation Report [D1], Section 6. Results are summarised in Table 2-2 in Section 2.2 below.
The validation results for SIS ICDR monthly and daily products are fully described in CM SAF Validation Report ICDR [D7], Section 5.2.
The validation results of SOL and SDL ICDR extensions is performed within the C3S using the algorithms developed by CM SAF. The summary of validation results for the monthly ICDR SDL and ICDR SOL datasets are shown in Table 2-1.
At the time of writing (May 2022) only 34 out of 60 BSRN stations have submitted data for years 2019-06/2021. 12 stations had less than 4 months worth of data and they were discarded. A full list of BSRN stations used in the validation and station-wise results are presented in Table 7 and Table 8, which can be found in the Appendix of this document. The outgoing longwave measurements are not carried out at all BSRN stations. Only 9 stations have submitted data for the years 2019-2021.
Table 2-1: Summary of validation results for SDL and SOL ICDR datasets (2019-06/2021)
Dataset | Number of months analysed | BSRB Stations | Abs.bias, W/m2 | Std.Dev, W/m2 |
SDL, ICDR | 465 | 21 | 7.2 | 9.32 |
SOL, ICDR | 253 | 9 | 9.9 | 13.45 |
2.2 Validation results for CLARA family extra data products (SRS, SNS, SNL, SRB)
The validation methodology for the extra data products is selected in a way to provide the most conservative accuracy estimation, i.e. by assuming that the uncertainties in the input datasets align perfectly and contribute fully to the overall uncertainty. Table 2-3 below provides a summary of the accuracies for the SRS, the SNS, the SNL, and the SRB.
Table 2-2: Summary of the accuracy of the brokered CLARA-A2.1 data products, [D1] Section 1 for SIS, SOL, SDL TCDR, [D7] Section 1 for SIS ICDR. The accuracy estimation of SOL and SDL ICDR extensions is performed within the C3S using the algorithms developed by CM SAF
Product Name | Dataset accuracy [W/m2] | |
|---|---|---|
| TCDR | ICDR | |
SIS monthly means | 9.5 | 8.7 |
SIS daily means | 18.6 | 22.1 |
SDL | 8.1 | 7.2 |
SOL | 13.8 | 9.9 |
Table 2-3: Summary of the accuracy of the extra data products
Product Name | Propagated accuracy [W/m2] | |
|---|---|---|
| TCDR | ICDR | |
SRS | 7.8 | 6.6 |
SNS | 13.0 | 12.0 |
SNL | 21.9 | 17.1 |
SRB | 34.9 | 29.1 |
3. Application(s) specific assessments
A general overview of user requirements needed for climate monitoring is provided in CM SAF Product Requirements Document [D2]. There are three accuracy categories in the CM SAF PRD document ([D2], Section 5): threshold, target and optimal accuracies. They are defined keeping in mind different target users: operational climate monitoring, global and regional climate modelling and global and regional climate studies, respectively.
The Table 3-1 summarizes the achieved accuracy and stability for Surface Radiation Budget products, as well as the target requirements for accuracy and stability. The achieved accuracies correspond to the target accuracy requirements, only SIS daily accuracy corresponds to threshold accuracy.
Table 3-1: Summary of achieved accuracy and stability of the brokered CLARA-A2.1 data products, Section 1 [D1 for SIS, SOL, SDL TCDR, D7 for SIS ICDR]. The accuracy estimation of SOL and SDL ICDR extensions is performed within the C3S using the algorithms developed by CM SA
Product Name | Propagated accuracy [W/m2] | Threshold/ Target/ Optimal accuracies [W/m²] | Stability [W/m²/dec] | |
|---|---|---|---|---|
| TCDR | ICDR | |||
SIS monthly means | 9.5 | 8.7 | 15 / 10 / 8 | 2 |
SIS daily means | 18.6 | 22.1 | 30 / 20 / 15 | 2 |
SDL | 8.1 | 7.2 | 15 / 10 / 8 | 3 |
SOL | 13.77 | 9.9 | 15 / 10 / 8 | 3 |
At the time (~2010) when requirements for the CLARA-A2 data record had to be defined in [D2], there was no guidance available for surface radiation products in the available GCOS documents. Instead, requirements had to be set in a dialogue with experts and potential users (e.g., in association with CM SAF User Workshops).
New GCOS requirements for the ECV Surface Radiation Budget are summarized in GCOS-200 [D6], Table 23, page 279 and include requirements for the horizontal resolution, temporal resolution, accuracy and stability. GCOS targets are:
Frequency: Monthly (resolving diurnal cycle)
Resolution: 100 km
Measurement uncertainty: 1 Wm-2 on global mean
Stability: 0.2 Wm-2dec-1
However, these requirements are only valid for the net fluxes (i.e., SNS and SNL) and not for all individual radiation budget components. However, one could claim that individual radiation budget components should consequently be constrained in the same way as net fluxes.
All products in the brokered CLARA-A2.1 dataset fulfil the GCOS requirements regarding the horizontal and temporal resolution. However, the SNS and SNL products do not fulfil the new GCOS requirements on accuracy and stability which are very stringent. Nevertheless, achieved accuracy and stability results allow consistent quantification of mean values, anomalies, variability and the Earth energy budget in general. We point out the existing uncertainties in the methodology of comparison with area-to point measurements (i.e. satellite-area to point-ground-based reference networks) as important reasons for not fulfilling the new requirements.
The compliance with the user requirements (as defined in the CM SAF PRD document ([D2], Section 5) for the SIS, SDL, and SOL is provided in Table 10 above.
References
Driemel, A., et al. (2018). Baseline Surface Radiation Network (BSRN): structure and data description (1992--2017). Earth System Science Data, 10(3), pp. 1491--1501. doi:10.5194/essd-10-1491-2018
Ohmura, A., et al. (1998). Baseline Surface Radiation Network (BSRN/WCRP): New precision radiometry for climate research. Bulletin of the American Meteorological Society, 79(10), pp. 2115-2136
Further references are listed in CM SAF Validation Report [D1], Section 7.
Appendix
Table 7: Station-wise validation results for SDL, 2019-06/2021
Full name | Abbreviation | # of months | Mean BSRN, W/m2 | Mean CLARA, W/m2 | AbsBias | RMSE | Percentage above 15 W/m2 |
Ny Alesund | nya | 30 | 257.0 | 241.7 | 15.31 | 16.11 | 46.7 |
Toravere | tor | 22 | 305.0 | 300.1 | 5.10 | 5.90 | 0.0 |
Lindenberg | lin | 12 | 318.5 | 318.0 | 3.25 | 4.02 | 0.0 |
Cabauw | cab | 30 | 320.1 | 315.1 | 5.01 | 5.63 | 0.0 |
Palaiseu Cedex | pal | 24 | 322.3 | 317.6 | 4.76 | 5.49 | 0.0 |
Sioux Falls | sxf | 11 | 305.5 | 291.9 | 13.61 | 14.05 | 54.5 |
Sapporo | sap | 23 | 314.0 | 307.9 | 6.31 | 8.08 | 4.3 |
Cener | cnr | 30 | 320.1 | 302.6 | 17.48 | 18.19 | 70.0 |
Rock Springs | psu | 16 | 312.5 | 299.7 | 12.74 | 13.30 | 25.0 |
Bondville | bon | 16 | 315.2 | 308.3 | 7.02 | 7.89 | 0.0 |
Langley Research | lrc | 30 | 338.9 | 343.2 | 4.74 | 5.70 | 0.0 |
Billings | bil | 7 | 338.2 | 323.6 | 14.64 | 15.17 | 42.9 |
Great Plains | e13 | 7 | 338.1 | 324.5 | 13.61 | 13.95 | 28.6 |
Tateno | tat | 30 | 337.8 | 340.4 | 2.64 | 3.10 | 0.0 |
Goodwin Creek | gcr | 16 | 347.5 | 348.8 | 4.88 | 6.03 | 6.2 |
Fukuoka | fua | 30 | 347.1 | 345.7 | 1.96 | 2.24 | 0.0 |
Ishigakijima | ish | 30 | 404.9 | 408.9 | 4.15 | 5.14 | 0.0 |
Minamitorishima | mnm | 30 | 402.4 | 403.4 | 1.99 | 2.46 | 0.0 |
Gobabeb | gob | 30 | 342.0 | 331.3 | 10.85 | 11.80 | 20.0 |
Florinopolis | flo | 30 | 385.1 | 376.4 | 8.64 | 8.85 | 0.0 |
De Aar | daa | 11 | 300.4 | 301.2 | 1.60 | 1.90 | 0.0 |
Total | tot | 465 | 335.8 | 330.6 | 7.21 | 9.32 | 12.5 |
Table 8: Station-wise validation results for SOL, 2019-06/2021
Full name | Abbreviation | # of months | Mean BSRN, W/m2 | Mean CLARA, W/m2 | AbsBias | RMSE | Percentage above 15 W/m2 |
Cabauw | cab | 30 | 369.4 | 368.8 | 3.01 | 4.27 | 3.3 |
Concordia Station | dom | 26 | 144.3 | 141.7 | 11.69 | 14.87 | 23.1 |
Gobabeb | gob | 30 | 444.6 | 448.9 | 6.45 | 7.48 | 0.0 |
Georg von Neumayer | gvn | 25 | 244.5 | 258.0 | 13.59 | 15.21 | 40.0 |
Ny Alesund | nya | 30 | 290.4 | 285.2 | 13.95 | 19.33 | 26.7 |
Payerne | pay | 30 | 367.7 | 344.4 | 23.26 | 23.64 | 100.0 |
Syowa | syo | 30 | 265.0 | 264.1 | 3.09 | 3.95 | 0.0 |
Toravere | tor | 22 | 355.9 | 347.8 | 8.08 | 9.50 | 13.6 |
Tateno | tat | 30 | 391.3 | 396.9 | 5.98 | 7.21 | 0.0 |
Total | tot | 253 | 322.3 | 320.3 | 9.86 | 13.45 | 22.9 |
