the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 19 May 2026
Evaluation of cloud base height provided by ceilometers and a proposal for a visibility-based quantitative definition for aviation
Abstract. Ceilometers are established sensors for deriving cloud base height (CBH) from a laser backscatter signal. However, various types of ceilometers provide different CBHs for the same cloud situation due to the application of manufacturer-specific, proprietary algorithms and, above all, the lack of a generally valid quantitative definition. This is particularly critical for air traffic control in low cloud situations. In the framework of the AutoMETAR project by Deutscher Wetterdienst (DWD) and in collaboration with the Universities of Leicester and Hamburg, the multiphase "Ceilometer campaign Hansestadt Hamburg" (CircaHH) was carried out. Across three measurement campaigns conducted between September 2016 and May 2019, primarily during the fall-spring seasons when low clouds are more prevalent, photographs of the 300 m tall "Hamburg Weather Mast" were taken for subsequent image analysis. Since the mast is increasingly obscured from the top downwards if the CBH descends, the contrast ratio of its alternating red and white segments was used to calculate the extinction coefficient profile. Several independent methods were analyzed for their suitability and accuracy to derive CBH from these extinction profiles. Our results suggest that the slant optical range (SOR) using a threshold value of 1000 m is an appropriate quantitative CBH definition for aviation, providing pilots with more useful information for an oblique downward or upward view. This SOR definition was also applied to extinction coefficient profiles derived from backscatter signals of different ceilometers using our own implementation of the Klett inversion method (KIM). This cross-check reveals that CBH differences due to various ceilometer backscatter input data for the KIM are significantly smaller than CBH differences associated with different ceilometer algorithms. Despite the complex experimental setup, our image analysis can serve as a new reference method for evaluating vertical profiles of the ceilometer backscatter signal for low clouds and can be used as a quality check for CBHs from any ceilometer firmware. Our KIM, combined with the SOR definition, has the potential to improve automated airport weather reports for CBH and cloud amount like routine Meteorological Aerodrome Report (METAR) and local routine report (MET REPORT).
- Preprint
(4716 KB) - Metadata XML
- BibTeX
- EndNote
Status: final response (author comments only)
- RC1: 'Comment on egusphere-2026-1696', Anonymous Referee #1, 04 Jun 2026
-
RC2: 'Comment on egusphere-2026-1696', Anonymous Referee #2, 28 Jul 2026
General review
--------------The paper describes an assessment of ceilometer cloud base height estimates provided by firmware with a calculation from extinction profiles stemming from a camera in the visible range and from the ceilometer backscatter profiles using different cloud base definitions. The topic is absolutely relevant and the main idea of the experiment makes sense. However the presentation of the manuscript and the derivation of the methods need very major reworking to make the manuscript publishable. The authors should not hesitate to re-write larger parts of the paper.
Specific comments
-----------------Please make an assessment on the differences you expect due to the wavelength difference from using ceilometers in the infrared on the one hand and the blue channel of the camera on the other hand to infer your cloud bases according to the definitions. Also, you mention that PollyXT does not have a extinction profiles delivered at 1064 nm and is thus not suited for validation. However it seems to provide extinction at 532 nm what would be even closer to the ceilometer wavelengths. But the overlap might possibly be an issue for low clouds. Please comment on this and re-formulate your introduction.
Some sources of the (empirical) methods applied appear unclear, e.g. Eq. 3. I think it is important to have a sound reasoning for your method transmitted to the reader which is not the case here. In the end this is a key of your estimation of the extinction from camera, so we need to know what is going on here.
In addition the description of the method for the camera estimates and its parameters needs improvement to make it more understandable. Especially the method B for the calculation of the extinction profile appears questionnable or at least hard to understand. I would encourage the authors to come up with a more comprehensive description and derivation of the method.
Between section 3 and 4 you should logically also introduce the Klett inversion. Possibly the image analysis and the Klett inversion would be best suited as subsections of a new section 3 entitled something like "determination of the extinction profile". I see you have it in the appendix, but at least a small introduction and reference to it in the main text would be necessary if you don't want to move everything up. Also make sure that the CBH definitions given in the current section 4 can be applied to ceilometer extinction profiles and camera extinction profiles and give a more detailed description on the reasoning of them.
A lot of editorial work on the manuscript is required, here some examples. At some occasions, the manuscript has a tendency to describe what follows later in the manuscript (e.g. l. 94 - 104, l. 112-116, ...) or to repeat what has been said before instead of just making a reference to the subsequent paragraph. Moreover some information which I consider belonging to the introduction follow in other sections (e.g. l. 126-128). In addition some detail information are not relevant or out of context and other phrases could be completely omitted (e.g. l. 128-130). Similarly don't overly repeat what you already stated earlier (e.g. l. 131-132). This can result in a somewhat fuzzy reading experience. Please do quite some re-structuring, also on points which I haven't mentioned (there are too many).
Review of specific parts of the manuscript
------------------------------------------- l. 3-4: formatting error with spacing and right margin
- l. 6-8: too much details for an abstract in this sentence
- l. 31: I suggest to remove "However" or replace it by a more appropriate formulation.
- l. 60: "\tau is the optical depth" -> "\tau is the optical depth between the instrument and range r"
- l. 60: I would move the sentence "The different methods..." to the next paragraph
- l. 68-69: why capitalizing the cloud types?
- l. 68-69: "to several 100 m or even a few 1 km for Cumulonimbus". Please rephrase and clarify
- l. 113: suggestion "... which is located 178m from the mast and 2m less in the third CircaHH phase as explained in the caption to Table 1." --> "... which is located 176-178 m from the mast (see Table 1)."
- l. 115: temporal resolution is confusing here. The shutter of the camera was not open for 1 minute, I guess. Moreover I think this information might not be needed here as you describe it in more detail at line 134-136. You can also consider to move the latter one upward, but don't duplicate.
- Eq. 3.: I would like a more explicit reference than just "Wikipedia" from where this empirical formula is coming and also an explanation on the reason for using this mapping function rather than another.
- Eq. 4: Could you consider using another symbol than C to avoid confusion with lidar constant?
- l. 155-157: Please add a formula for this normalisation, it is needed for the good understanding of this statement and section.
- l. 152-169: I struggle to understand the difference between C_n and C and then their difference to C(h) and C_0(h). I was under the impression that C_0 must be the contrast value from the lowest 2 segments, but then how about C_0(h). Similarly, what is the difference between h and H. Please clarify the variables used in this section.
- Eq. 6: I would keep it general with just Eq. 5 here and then move Eq. 6 to approach A) at line 172.
- l. 181 and Eq. 9: Please explain the reasoning of setting V=d and of using Eq. 9. For me it is unclear why you use Eq. 9 just to set V=d afterwards. And you should also state what d is. Moreover, I don't understand why you come up with an equation from Weitkamp 2005 instead of just solving Eq. 5 with your linear sigma assumption. With this you should also come up with Eq. 10 without detour over external references and a reference to a subsequent equation and an obscure V=d setting.
- From Eq. 10 -> Eq. 11. Please state that you are setting s=d. Actually why do you have s and d as distinct variables?
- l. 186-187: This sentence does not tell anything to the reader at the current stage. Probably best to remove it.
- Fig. 3: I was puzzled that you aligned the subfigures from bottom to top. Why breaking the convention of going from top to bottom?
- Fig. 3 c): I still don't understand how you calculated the inherent contrast and what you are using it for. Please make this clearer in the manuscript.
- Fig. 3 f): These are vertical, not horizontal dashed lines. But it might be a good idea to have a horizontal dashed line denoting the 1000 m threshold
- Fig 3 f): At the time you introduce Fig. 3 in the manuscript you have not defined MOR or SOR, so you should introduce the figure after this definition.
- l. 199: I would not call this section "Attempts" but rather just speak of different cloud base definitions"
- Section 4: This whole section deserves re-writing. You should guide the reader on the reasoning behind the different definitions. You do this somehow (but not well enough) a bit later but it should be done at the time you introduce them instead of throwing all the definitions in first.
l. 222-223: "The CBH results..." -> This sentence does not have anything to do here.
l. 224: but Eq. 13 does so as well...
- Figure 4: You state that SOR-based CBH lead to lower values, but the drawing shows them being larger than MOR. It is also in contrast to what you state in the main text. Moreover you have 2 CBH SOR which I assume are for the two different extinction profiles in the right panel, but you should mention this, the reader isn't assumed to guess.
- Eq. 14: Please explain this equation, i.e. the geometry, where the 3 in the nominator comes from, what is the idea behind, etc. This also applies to Eq. 12 and 13, but here the lack of explanation is most striking.
- l. 215: What is the value of SOR_thr, 5%? Please re-formulate.
- l. 241-142: The first 2 sentences belong to the figure caption, not the main text.
- l. 284-285: omit this question falling out of the blue and which seems not to be absolutely needed
- l. 288-289: I don't understand how you associate the MORs at the different altitude levels with the plausibility of the increasing CBH. Please explain.
- l. 296-303: What is the added value of this reasoning in addition to Figure 7 what you introduce below?
- l. 307: What senor problems and of what sensor, PWD22 or camera? Please be more specific and don't throw in such sentences which the reader cannot really understand.
- Section 5.3.: I would not call this a thorough error estimation. Please provide a more in-depth analysis
- l. 335: Why do you consider that the Klett inversion is not directly applicable to operational use? Please specify.
- l. 346: smaller or larger?
- Fig. 3 e): Can you attribute this behavior to an artefact described in Kotthaus et al. 2016?
- l. 374: when making such a statement, it would be needed that you specify which firmware algorithm uses which definition. Moreover, I would be a bit afraid to give reasons for the differences while the firmware algorithms used by the manufacturers are mostly black box.
- l. 378-379: This sentence is not placed well in the Conclusions- Appendix A: The main equations of your Klett inversion must be outlined here. As current, this section is not followable by the reader with reasonable effort. Moreover, I have trouble to understand why you use LD40 parameters where e.g. for pulse per power, diameter, etc. Such basic parameters should also be available for the other ceilometers.
Moreover it appears unclear to me what is your reference region for the backward Klett algorithm. I guess the molecular signal is not accessible with ceilometers in the conditions of interest to you, hence, also from your statement at l. 424-428, I assume you use the known lidar ratio in the cloud as reference region. But I don't want to speculate, can you please be more explicit in the description of the method. I will be happy to read this section again after revision.
- Appendix B: This would be better placed in a supplement than the appendix, I guess.Citation: https://doi.org/10.5194/egusphere-2026-1696-RC2
Data sets
Subset of Ceilometer campaign Hansestadt Hamburg (CircaHH) data Daniel Klaus and Ulrich Görsdorf https://doi.org/10.5281/zenodo.20134337
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 243 | 95 | 15 | 353 | 21 | 11 |
- HTML: 243
- PDF: 95
- XML: 15
- Total: 353
- BibTeX: 21
- EndNote: 11
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
Please note comments in the attached file:
Feet (instead of, or at least as well as!) should be used since this is standard throughout the global aviation community
All Charts, all values quoted by and to ATC etc.
Presentation of Figure 5 (etc.) needs to be improved to understand the important contents