the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
Cloudiness retrieved from All-Sky camera and MSG satellite over Reunion Island and Antananarivo Madagascar
Abstract. To gain a deeper understanding of cloud variability over the Southwest Indian Ocean (SWIO) region, various measurement techniques can be used. We have used ground-based observations using all-sky cameras from the UV-Indien network provided by the Reuniwatt company, which capture sky images in the visible spectrum. Two algorithms, namely Elifan and Reuniwatt, were applied to analyze the camera images. Despite differences in the methodologies employed by each algorithm, we have found strong agreement between them, with a Bias of -5.48 %, a Root Mean Square Error (RMSE) of 6.48 %, and a correlation coefficient (r) of 0.99 for Saint-Denis. Ground-based measurements over the SWIO are insufficient, and the vast ocean coverage in this region makes spatial observation important for gathering information. Here we have used cloud products from the Meteosat Second Generation (MSG) satellites. The quality of the classification algorithm employed for camera and satellite plays a significant role in cloud analysis. Comparing camera and satellite observations is essential to ensure the complementarity of each measurement. We observed good consistency between the ground-based camera and satellite measurements (Bias=2.64 %, RMSE=21.43 %, and r=0.87) with Elifan, and (Bias=6.79 %, RMSE=25.70 %, and r=0.82) with Reuniwatt for the Saint-Denis site; and (Bias=6.48 %, RMSE=28.63 %, and r=0.78) for Antananarivo. In Antananarivo, during the dry season, heavy cloud cover (~50 %) is observed in the morning, gradually dissipating as the day progresses. Conversely, in the wet season, cloud cover varies between approximately 30 % and 60 % from December to April, with weaker cloudiness around noon in October and November. As for Saint-Denis, the morning skies are generally clear but become increasingly overcast throughout the day, reaching up to 80 % cloud cover during the wet season and 60 % during the dry season.
- Preprint
(2426 KB) - Metadata XML
- BibTeX
- EndNote
Status: final response (author comments only)
-
RC1: 'Comment on egusphere-2024-1827', Anonymous Referee #1, 18 Mar 2025
The comment was uploaded in the form of a supplement: https://egusphere.copernicus.org/preprints/2024/egusphere-2024-1827/egusphere-2024-1827-RC1-supplement.pdf
-
RC2: 'Comment on egusphere-2024-1827', Anonymous Referee #2, 22 Mar 2025
Manuscript “Cloudiness retrieved from All-Sky camera and MSG satellite over Reunion Island and Antananarivo Madagascar”, Rivonirina et al., egusphere-2024-1827
general comments
This paper compares different datasets from satellite and all-sky cameras on the capability to observe cloud fraction.
It compares different data sources, but remains very general in introduction and state of the art, describes the used data only roughly, and does not discuss the restrictions of the chosen methods sufficiently.
It should be better elaborated what the new data or new research results are compared to existing knowledge.
Restrictions of the underlying data should be better assessed and their impact quantified.
Overall, the study is very general, seems to be of local interest mainly, and does not clearly add new knowledge to the understanding of cloud climatologies beyond what is known from general meteorology and existing satellite or model-based climatologies.specific comments
- The abstract is very unclear, it provides numbers on bias etc. but it is not clear for which geophysical parameter, in general the abstract does not tell which parameter is analysed and seems to repeat twice statistics for Saint-Denis? At the end the reader may guess that it is about cloud coverage? Later in line 56 it is called cloud fraction?
- The commercial algorithm of Reuniwatt is not further specified by the coauthors originating partly from the company Reuniwatt. This is a drawback in a scientific paper which only can be accepted if the method itself is not relevant and the other results of the paper are standing on its own.
- The Elifan algorithm on the other hand is further detailed, but I miss a clear description why it is better/different from other existing cloud masking methods for all-sky cameras.
- I’m missing a specific review of the state of the art in cloud masking of all-sky cameras in the introduction section. Why is the presented work going beyond the state of the art? What is new? Why is the comparison of the two camera algorithm of interest for the reader?
- Also, the introduction is very broad and not specific enough to describe the research question of this paper clearly. It should be rewritten with focus on the study details and not refer generally to textbook knowledge.
- It should be better justified why the assessment of the two locations is of special interest to the broader scientific community. What is known on cloudiness in the area from other studies? How does the work fit to the existing state of the art? Just the aim to understand the cloud properties of two locations operated by the author team is of minor interest to the scientific community.
- The very general, again textbook like satellite description could be replaced by a much better description on the details of cloud mask retrieval and how the assumptions and restrictions of the CLM algorithm may affect your study. What is known from other studies on the CLM product accuracy?
- How is cloud fraction inside a satellite pixel handled? The authors seem to assume that the CLM product provides only 0 and 100% cloud fraction inside a pixel? Is this realistic? What is the impact of this assumption? What happens to inside-a-pixel cloud fraction and what happens to sub-visible clouds which should be visible in the all-sky cameras, but not in the satellite images?
- The reprojection of satellite-based cloud masks by a 3x3 window is assumed to be the same area as the field of view of all-sky cameras. The actual field of view of the all-sky camera depends very much on visibility and the type and height of clouds in the field of view. It can range from very local to hundreds of km. This is not taken into account. This is a major drawback of the study.
- What is the additional value of using all-sky cameras for cloud fraction quantification if the satellite data is already available as open data? If you compare only against satellite data, what is the added value of an all-sky camera? The research question of the study is not well discussed and analysed. The relevance of the work should be better elaborated.
- Why do you introduce cloud type? It is unclear how this is used in the study.
technical corrections
- What is meant by the term ‘nebulosity’?
- What is the difference between cloud fraction, cloud coverage, cloud fraction index? Please better define the terminology used and be more specific.
- I’m not going into more technical details of the paper as it first requires a significant rewriting of the overall study design.
Citation: https://doi.org/10.5194/egusphere-2024-1827-RC2 - The abstract is very unclear, it provides numbers on bias etc. but it is not clear for which geophysical parameter, in general the abstract does not tell which parameter is analysed and seems to repeat twice statistics for Saint-Denis? At the end the reader may guess that it is about cloud coverage? Later in line 56 it is called cloud fraction?
Viewed
HTML | XML | Total | BibTeX | EndNote | |
---|---|---|---|---|---|
265 | 87 | 26 | 378 | 18 | 18 |
- HTML: 265
- PDF: 87
- XML: 26
- Total: 378
- BibTeX: 18
- EndNote: 18
Viewed (geographical distribution)
Country | # | Views | % |
---|---|---|---|
United States of America | 1 | 125 | 33 |
Madagascar | 2 | 38 | 10 |
China | 3 | 30 | 8 |
France | 4 | 26 | 7 |
Germany | 5 | 23 | 6 |
Total: | 0 |
HTML: | 0 |
PDF: | 0 |
XML: | 0 |
- 1
- 125