the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 23 Jun 2026
Cloud occurrence, properties, and ice crystal effective dimension parameterization from nine years of far-infrared observations on the Antarctic Plateau
Abstract. Clouds over the Antarctic Plateau exert a strong influence on the regional radiation budget, yet observations and modelling of their properties remain scarce. Here, nine years (2012–2020) of ground-based high-resolution spectral radiance measurements from the REFIR-PAD spectroradiometer at Concordia Station (Dome C, Antarctica) are analyzed in synergy with co-located lidar observations. A machine-learning Cloud Identification and Classification (CIC) algorithm is applied to discriminate clear sky, ice cloud, and mixed-phase cloud conditions, enabling the construction of a long-term cloud climatology. Cloud optical and microphysical properties are subsequently retrieved using a simultaneous atmospheric and cloud retrieval framework, for cases with reliable cloud boundaries and cloud base heights above 500 m. Results confirm that cloud occurrence over Dome C is dominated by optically thin ice clouds, with approximately 95 % of cases exhibiting optical depths below 1. Median optical depth ranges from 0.11 in summer to 0.32 in winter. The median temperature of the ice layers is approximately 237 K. Mixed-phase clouds are rare and mainly confined to the austral summer, but exhibit larger optical depths (median 1.7) and warmer temperatures (approximately 246 K). Based on the retrieved dataset, a new parameterization of ice crystal effective dimension is derived. Compared with commonly used parameterizations developed for tropical and midlatitude conditions, the proposed scheme predicts systematically smaller particle sizes, highlighting the inadequacy of existing formulations for the Antarctic environments. These results provide new observational constraints on Antarctic cloud microphysics and support improved cloud representation in climate and numerical weather prediction models.
- Preprint
(6041 KB) - Metadata XML
- BibTeX
- EndNote
Status: open (until 13 Sep 2026)
- RC1: 'Comment on egusphere-2026-2806', Anonymous Referee #1, 17 Aug 2026 reply
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 52 | 29 | 7 | 88 | 6 | 9 |
- HTML: 52
- PDF: 29
- XML: 7
- Total: 88
- BibTeX: 6
- EndNote: 9
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
Review of "Cloud occurrence, properties, and ice crystal effective dimension parameterization from nine years of far-infrared observations on the Antarctic Plateau", Fabbri et al, submitted to ACP.
This manuscript uses retrievals from the REFIR-PAD data collected at Dome C in Antarctica to develop an ice cloud particle size parameterization. The data used in the analyses are unique, and are an important description of cloud properties in this poorly understood, but climactically important region. I think the topic is quite relevant to ACP, and the work is an important application of the Dome C dataset.
There are a number of issues with the presentation that I think should be addressed before the paper could be accepted. I think these are all minor issues in that I do not think they require much detailed work to address.
-- There is a large amount of overlap with the Donat 2025 (D2025) manuscript submitted to AMT. For example, Figure 4 appears be the same data as Table 1 in D2025; Figure 5 is the same as Figure 4(b) and (c) in D2025.
-- In addition, I also think that the manuscript is rather long, and there are some figures that could be omitted. This, I would suggest substantially reducing section 3, and simply referring to D2025 when warranted.
-- Finally, I also think the primary new work in this paper is the analysis of the cloud properties themselves, and the new parameterization, not the occurrence statistics (as that was done in D2025). Thus, I think the paper can be shortened and focused on the novel aspects.
-- To address the above three points, I would suggest removing Figures 4 and 5 as these duplicate information in with D2025. I would also suggest removing Figure 3, as it is not central to the paper. This would also mean substantial text could be removed from Section 3.
-- I generally agree with the methodology in the retrieval setup, but there are several very important simplifying assumptions. One key assumption is that the cloud is a homogeneous layer of solid single ice crystals. It would be valuable to know how sensitive the results are to this assumption, for example repeat the retrievals with a different habit assumption and quantify the difference in retrieved cloud properties (similar to what was done in Maestri 2019, doi:10.1029/2018JD029205). If the authors argue that is out of scope for this paper, that is fine, but I think these assumptions need to be emphasized more clearly. Right now, those important assumptions are described in the methodology (lines 219, 230-234). These should be repeated and emphasized again in the conclusions, probably expanding the paragraph at line 480.
-- One minor methodological issue, in the discussion of the parameterization performance, several statistics are computed to evaluate the 'goodness of fit' (Table 7). One of these statistics is the bias - presumably this is the mean of residuals between the data and fit. Note that the mean residual of a least squares fit is zero by definition. The small number reported for bias, in Table 7, is therefore not meaningful and should be omitted. (The nonzero value is floating point roundoff error propagated through the bias calculation).
-- On data availability: The new and novel data used in this manuscript is not the remote sensing datasets (this is described in several other papers, e.g. D2025 among others), but rather the cloud property retrievals from the REFIR-PAD. My reading of the Copernicus data policy (https://publications.copernicus.org/services/data_policy.html) is that data release is not required, but encouraged. Therefore I would encourage the retrieval data to be shared publicly. The editor could also give their opinion in this case.
Other minor corrections, suggestions, needed clarifications:
Line 45: I think the "broad temperature range" for the ice clouds is incorrect, doesn't Figure 8 show many cases below -40 C? It also seems strange to describe the mixed phase cloud range as "narrower" when it is only 2 C warmer. But, I think the ice cloud range should contain much colder temps.
Line 156: "The criteria adopted for the selection of the spectra consist in the individuation of neatly defined cases observed by the lidar."
This sentence is unclear, can you rewrite?
Figure 1: This would be better represented as a table, as there are not very many values. If the authors insist on keeping this figure, a different colormap is needed, as the Ice Cloud - ThS is not readable (black on dark blue).
Line 288: Is "surface temperature" the surface air temperature, or surface skin temperature?
Figure 7: Please explain the reason for ~25% of the data that lack co-located LIDAR data; is that simply due to instrument outages, or are these cases that the LIDAR fails to detect?
Line 320: I thought that the ice crystal size was assumed to be constant (cloud properties are homogeneous in the layer); so why does the mean value of D_eff need to be computed?
Line 364: How was the D_eff computed in the mixed phase cases? Is it weighted by the optical depths, or something else?
Figure 13: This plot is confusing to me. It could be omitted entirely, as the only new information is that the residuals do not strongly depend on T_cloud or IWC, and I think Figure 12 already shows this - the spread of the distributions does not vary much across the T_cloud in and IWC variables in each plot. If the authors want to keep this figure:
The boxes with the sample counts are hard to read (please choose a lighter background color). The boxes could also be omitted, as this is not new information, as the count histograms are already in Figure 8. The bin widths are chosen at seemingly random values, it would help to bin by regular values (2.5 or 5 K for T_cloud, 0.25 in log IWC) which would make it easier to visually compare to other plots (e.g. Figure 12). The y-ranges should be significantly narrowed to be able to more clearly see the variation that is there (+/-25 um).