This section collects the atmospheric datasets on which the Complete Data Fusion (CDF) algorithm has been applied and verified during the EMM project. Each “tested dataset” is a retrieval product (instrument + provider + processing chain) for which we have checked whether the file content and format comply with the requirements of the CDF algorithm, and we have run the verification tests described in the dedicated sections of this website.
For each dataset we report the FAIR references (persistent identifiers, access URLs, documentation, citation), the general characteristics of the product, the results of the completeness and auto-consistency tests, and the mapping between the quantities required by the CDF algorithm and the actual variable names inside the data files.
These are data quality tests. The auto-consistency tests documented here are not merely a verification of algorithmic compatibility. A dataset that passes the CDF auto-consistency test demonstrates that the error covariance matrices and averaging kernels stored in the product file are internally consistent and mathematically well-formed — in other words, that the characterisation of the retrieval uncertainty is reliable. A failure indicates that the uncertainty information in the file is inconsistent or incomplete, which would compromise any data fusion application regardless of the algorithm used. The test results should therefore be interpreted as indicators of data quality, not just as a check of CDF readiness.
How to read this index. The Master overview table lists all five datasets; for IASI/EUMETSAT the six constituents are shown as separate rows. Separate columns give CDF(2022) and CDF(2015) auto-consistency test outcomes. The Constituent × Instrument cross-reference table shows which datasets cover each atmospheric constituent. The Test summary boxes collect the overall results at a glance.
Master overview
| Dataset | Provider | Instrument / Platform | Constituents | Format | CDF(2022) | CDF(2015) |
|---|---|---|---|---|---|---|
| GOME-2 / AC-SAF | AC SAF (EUMETSAT) | GOME-2 / Metop-A,B,C | O3 | HDF5 | ✓ passes | ✓ passes |
| IASI / EUMETSAT | EUMETSAT | IASI / Metop-A,B,C | T (Temperature) | netCDF / EPS binary | ✓ passes GS basis expansion required† |
✓ passes GS basis expansion required† |
| H2O (Water vapour) | ✓ passes GS basis expansion required† |
✓ passes GS basis expansion required; profile space: 10.88% |
||||
| O3 (EPS netCDF) | ~ partial PC space confirmed; profile space partial |
~ partial PC space confirmed; profile space: 6.44% |
||||
| O3 (FORLI native) | ✓ passes | ~ partial max |Δx| = 3.71% |
||||
| CO | ✓ passes | ~ partial max |Δx| = 0.89% |
||||
| HNO3 | ✓ passes | ~ partial max |Δx| = 3.29% |
||||
| IASI / AERIS | AERIS (IASI Portal) | IASI / Metop-A,B,C | O3 (FORLI) | netCDF | ✓ passes legacy & CDR formats |
✗ fails legacy: max |Δx| = 3523%; CDR: max |Δx| = 243% |
| MIPAS / IFAC | IFAC-CNR | MIPAS / Envisat | O3 | netCDF | ✓ passes | ✓ passes |
| OMPS / NASA | NASA (Earthdata) | OMPS / Suomi-NPP | O3 | HDF5 | ~ partial passes Variant A (AK from file); fails Variant B (AK from Jacobian) |
✗ fails both variants (Var. A: max |Δx| = 209%) |
† For T and H2O, the covariance \mathbf{S}_a is rank-deficient in profile space (PC-based retrieval). A Gram–Schmidt orthonormal basis expansion is required to make the inversion well-defined. The expansion procedure differs between the two constituents (see the dedicated page for details).
Legend — auto-consistency test outcome: ✓ passes = max(|Δx|) and |ΔDOF| both negligible (errors indicate no issues with the uncertainty characterisation) · ~ partial = results not fully satisfactory: errors not negligible in some configurations, test confirmed only in a subset of spaces (e.g. PC space only), or uncertainty about the correct file reading · ✗ fails = errors clearly not negligible, indicating problems with the uncertainty characterisation in the product file.
Constituent × Instrument cross-reference
Find which datasets cover a given atmospheric constituent. Click the column header to open the corresponding page. A ✓ indicates that CDF(2022) passes for this constituent; ~ indicates results are not fully satisfactory (partial, uncertain, or PC space only); — indicates the constituent is not provided by this dataset.
| Constituent | GOME-2 AC-SAF |
IASI EUMETSAT |
IASI AERIS |
MIPAS IFAC |
OMPS NASA |
|---|---|---|---|---|---|
| O3 (profile / total column) | ✓ | ✓ | ✓ | ✓ | ~ |
| O3 (EPS, PC-based) | — | ~ | — | — | — |
| T (Temperature) | — | ✓ | — | — | — |
| H2O (Water vapour) | — | ✓ | — | — | — |
| CO | — | ✓ | — | — | — |
| HNO3 | — | ✓ | — | — | — |
Legend: ✓ CDF(2022) passes · ~ CDF(2022) passes for some variants only · — constituent not provided by this dataset.
Test summary
Related pages
- CDF Prerequisites — the six requirements that input products must satisfy for CDF application
- CDF Tests — description of the completeness and auto-consistency tests applied to each dataset
- CDF Algorithm — overview of the Complete Data Fusion method
- Datasets catalogue — the general project catalogue of available datasets