24 Oct 2025
| 24 Oct 2025
Status: this preprint is open for discussion and under review for The Cryosphere (TC).
Brief communication: Evaluation of the ESA CCI+ ESMR v1.1 sea-ice concentration product
Abstract. I evaluated a novel NIMBUS-5 Electrically Scanning Microwave Radiometer (ESMR) sea-ice concentration (SIC) data product. I manually classified 50 Landsat-1 Multispectral Scanner (MSS) images obtained in the Northern Hemisphere during 1974 into open water and ice. I mapped these onto the ESMR product’s grid (25 km resolution) and computed Landsat-1 SIC. The resulting ~3300 grid cells, covering mostly compact sea ice, have a mean difference (median), standard deviation, and linear correlation coefficient of -1.4 % (0.0 %), 6.0 %, and ~0.9, respectively. This suggests using this novel ESMR SIC data product as an extension of existing SIC climate data records back in time.
Publisher's note: Copernicus Publications remains neutral with regard to jurisdictional claims made in the text, published maps, institutional affiliations, or any other geographical representation in this paper. While Copernicus Publications makes every effort to include appropriate place names, the final responsibility lies with the authors. Views expressed in the text are those of the authors and do not necessarily reflect the views of the publisher.
Review of “Evaluation of the ESA CCI+ ESMR v1.1 sea ice concentration product” by Stefan Kern.
Evaluation of sea ice concentration (SIC) is very important for product development and assessment of product quality. There are a number of good reasons to extend the SIC data records back in time, before 1978, and evaluation of these records using independent SIC estimates is particularly important because there is little overlap between SIC products in the 1970s. Stefan has been developing a valuable tool for evaluating SIC products using high resolution optical satellite imagery from Landsat and this has been used for more contemporary SIC records and published (Kern et al. 2022). Here the NIMBUS 5 ESMR v1.1 SIC has been evaluated. I have some suggestions which I think could improve the MS and then some specific comments below.
50 Landsat-1 images from the Arctic have been manually classified. The images are mostly from March and April 1974 and obviously the ESMR SIC quality changes seasonally. Could the scenes in Fig. 2c be color coded with time of year so that the reader could get an idea of the seasonal (and geographical) distribution?
One of the updates to ESMR SIC v1.0 to v1.1 was a treatment of the ice types in the SIC processing. The v1.0 SIC was underestimation multiyear ice SIC and overestimating first-year ice SIC. Evaluating this update is important. However, most of the Landsat scenes are over first-year ice and therefore a full evaluation is not possible. Please include a discussion of this limitation in the MS.
Overlap between the contemporary SIC records and NIMBUS 5 ESMR is not possible. However, the Kern et al. (2022) article presents a similar evaluation of Landsat imagery and a number of SIC products. The coverage in Hudson Bay and Baffin Bay is comparable to the selection of Landsat in the ESMR evaluation. How does the ESMR evaluation compare with the Kern et al. (2022) evaluation in these two regions?
Specific comments:
In the abstract: Please include the time of year for Landsat scenes and which ice type is covered by the Landsat scenes.
L 14 - 15: include some references to the statement.
L 15: suggest using “changes” instead of “developments”
L 16: add “satellite” before “climate”
L 17: delete “first data of the”, there was a SMMR on Seasat.
L 18: add “on NIMBUS 7” after “(SMMR)” and “NIMBUS 7” before “SMMR”
L 18: the sentence starting with “Prior…” rewrite “Prior to the NIMBUS 7 SMMR sensor there were other satellites carrying microwave radiometers, e.g. the NIMBUS 5….”
L 31: after “brightness temperatures” add “due to instrument and geophysical noise…”
L33: before “EASE” add “a predefined 25 x 25 km2 EASE grid”
L37: In Kolbe et al (2024) a threshold of 30% was selected because of the open water noise level. Why 10%? When 15% is more common for contemporary records.
L40: It is worth mentioning that the coverage in v1.1 is much better than in v1.0 (the list of missing files in v1.1 is much shorter in v1.1).
L43: what is meant with “reasonable”?
L45: It is not so clear that it is the SIC thresholds that are compared. Please reformulate.
L51: Is this criterion comparable to the evaluation criterion in Kern et al. (2022)? And if not, could this be aligned?
L85: Is it not better consistently to use 4, 5 and 7 and not just sometimes?
L95: add “v9.1” after “SNAP” and delete next sentence.
L105: Would it not be better to remap the Landsat data onto the Pol. Ster. Grid when evaluating the NSIDC ESMR SIC? So that this process is comparable to remapping the Landsat data to the EASE2 grid when evaluating the ESA ESMER SIC.
L131: are these regional differences comparable to Kern et al. (2022)?
L152: add “NIMBUS 7” before “SMMR”
L157: The reasoning about the Landsat SIC systematic uncertainty could perhaps be used for quantifying the Landsat SIC uncertainty including the pixel resolution and SIC dependent uncertainty, albedo variations, instrument noise, atmospheric noise?
L185: last sentence: “will reveal the full bias”?