EarthCARE level-2 demonstration products from simulated scenes
Bibliographic record
Abstract
<strong>Overview</strong> The EarthCARE satellite combines four instruments, a Cloud Profiling Radar (CPR), an Atmospheric Lidar (ATLID), a Multispectral Imager (MSI) and a Broadband Radiometer (BBR), from which many products will be generated on the properties of clouds, aerosols, precipitation and radiation. The dataset in this repository consists of test products generated from simulated 3D scenes produced by the Canadian Global Environmental Multiscale (GEM) model. This includes both Level 1 (L1) "input" products containing simulated satellite measurements, and Level 2 (L2) "output" products produced by running the various European retrieval algorithms on the inputs. The dataset has been produced as part of the European Space Agency funded "CARDINAL" project involving numerous European and Canadian scientists, and there are several versions representing the evolution of the algorithms during preparation for the launch of EarthCARE. The scenes and algorithms are discussed in detail in a Special Issue of the journal Atmospheric Modelling Techniques (AMT), so the overview here is limited to a summary of what is contained in this dataset. <strong>Directory structure</strong> The three top-level directories inside the zip file are for the three simulated scenes (except for the C-APC directory, described in a separate section below), each of which represent a single 6000-km long EarthCARE granule: Halifax: A swath over the Atlantic Ocean from the Caribbean to the Labrador Sea, passing close to Halifax in Nova Scotia Baja: A swath passing over the Rocky Mountains and the Baja California peninsula in Mexico Hawaii: A swath over the Pacific Ocean passing close to Hawaii Each of these directories contains three further directories: input: simulated level-1 instrument data output: level-2 meteorological products generated from the input data logs: text files logging the progress of the algorithms as they generated the level-2 data The input and output directories contain subdirectories for each product, and these each contain two files: A NetCDF4/HDF5 file with the suffix "h5" containing the retrieved data An XML file with the suffix "HDR" containing a description of the variables in the file (reproducing metadata already in the NetCDF4/HDF5 file) <strong>Input products (L1)</strong> The level-1 products are named <em>Y-ZZZ</em> where <em>Y</em> indicates the source of the data (A=ATLID, C=CPR, M=MSI, B=BBR and X=auxiliary) and <em>ZZZ</em> is a shortened version of the product name: A-NOM: Nominal ATLID measurements C-NOM: Nominal CPR measurements M-RGR: Regridded MSI measurements B-NOM: Nominal BBR measurements averaged to various scales B-SNG: Single-pixel BBR measurements X-JSG: Definition of the Joint Standard Grid on to which several of the observations are interpolated X-MET: Meteorological data from ECMWF <strong>Single-instrument output products (L2a)</strong> The naming convention is the same as the L1 data products. A-AER: ATLID aerosol profiles A-ALD: ATLID aerosol layer descriptor A-CTH: ATLID cloud top height A-EBD: ATLID extinction, backscatter and depolarization A-FM: ATLID feature mask A-ICE: ATLID ice cloud properties A-TC: ATLID target classification C-CD: CPR corrected Doppler velocity C-CLD: CPR cloud properties C-FMR: CPR feature mask and reflectivity C-TC: CPR target classification M-AOT: MSI aerosol optical thickness M-CM: MSI cloud mask M-COP: MSI cloud optical and physical properties <strong>Multi-instrument output products (L2b)</strong> The level-2b output products are named <em>YY-ZZZ </em>where YY is 2-4 character code conveying which of the four instruments were used and <em>ZZZ</em> is the shortened version of the product name.<em> </em> AC-TC: ATLID-CPR target classification AM-ACD: ATLID-MSI aerosol column descriptor AM-CTH: ATLID-MSI cloud top height BM-RAD: BBR-MSI broadband radiances (unfiltered) ACM-3D: ATLID-CPR-MSI constructed 3D scene ACM-CAP: ATLID-CPR-MSI synergistic retrieval of cloud, aerosol and precipitation ACM-COM: ATLID-CPR-MSI composite of single-instrument cloud and aerosol retrievals ACM-RT: ATLID-CPR-MSI radiative fluxes and heating rates computed on the retrievals BMA-FLX: BBR-MSI-ATLID broadband fluxes ACMB-DF: ATLID-CPR-MSI-BBR difference between radiances and fluxes computed from retrievals (ACM-RT) and measurements (BM-RAD, BMA-FLX) <strong>CPR Antenna Pointing Correction (C-APC) product</strong> From version 10.01 of the dataset, an additional top-level "C-APC" directory contains test data for the CPR Antenna Pointing Correction product, which will be generated via statistical analysis of a larger sample of C-NOM data. The generated information about mis-pointing of the radar antenna will then be used to improve interpretation of subsequent radar Doppler observations. There are two subdirectories: GEM_scenes: contains a file used in subsequent radar processing of the three GEM scenes, although in these scenes it has been assumed that there is no mis-pointing to correct, indicated by the file containing missing data and zeros. Antenna_mispointing_example: contains input and output files (in further subdirectories) illustrating what the data would look like with mis-pointing present. It was generated by artificially modifying the Doppler veclocities in the original Baja, Halifax and Hawaii scenes, concatenating them to generate a full orbit, and then running the C-APC algorithm to characterize the mis-pointing. <strong>Filename format</strong> The data filenames and directories have the following name format: ECA_EXAA_<em>YYY_ZZZ_LL_OBSERVATION-TIME_GENERATION-TIME_VVVVVV</em> where: "ECA" indicates EarthCARE. EXAA indicates the file class: ESA, Latency N/A, Simulator, Baseline N/A. <em>YYY</em> is a three-character code indicating the instrumental source of the data. The 1-4 character codes defined in the sections above are converted to 3 character codes as follows: A -> ATL, C -> CPR, M -> MSI, B -> BBR, X -> AUX, AM -> AM_, AC -> AC_, BM -> BM_, ACMB -> ALL. <em>ZZZ</em> is a three-character code abbreviating the full name of the product, as defined in the sections above. <em>LL</em> represents the level of the data from 1B, 1C, 1D, 2A and 2B. <em>OBSERVATION-TIME</em> is a code representing the date and time the observations were taken in the form <em>yyyymmdd</em>T<em>hhmmss</em>Z. Note that for the present datasets the dates are fictional future dates. <em>GENERATION-TIME</em> is a code of the same form representing the time that the processor was run to produce the data product. <em>VVVVVV</em> indicates the orbit number and frame letter. Further information may be obtained from the papers in the special issue of AMT. The description here may be expanded in future. Contacts: Gerd-Jan van Zadelhoff <gerd-jan.van.zadelhoff@knmi.nl> and Robin Hogan <r.j.hogan@ecmwf.int>
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame distilled prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. Learned from the 10,348 direct Codex labels and 10,348 direct Gemma labels. Candidate is the union of thresholded teacher heads; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels or direct frontier model labels.
Codex and Gemma teacher scores by category
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.000 | 0.000 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.000 | 0.001 |
| Science and technology studies | 0.003 | 0.000 |
| Scholarly communication | 0.001 | 0.000 |
| Open science | 0.001 | 0.000 |
| Research integrity | 0.000 | 0.001 |
| Insufficient payload (model declined to judge) | 0.275 | 0.009 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; both teacher heads agree on what is shown here.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".