Fog and low clouds in the Namib Desert may be more resilient than previously thought

Mass, Alexandre; Cermak, Jan; Knutti, Reto; Merrifield, Anna L.; Andersen, Hendrik

doi:10.5194/egusphere-2026-3242

Preprints

https://doi.org/10.5194/egusphere-2026-3242

Preprints

25 Jun 2026

| 25 Jun 2026

Fog and low clouds in the Namib Desert may be more resilient than previously thought

Alexandre Mass, Jan Cermak, Reto Knutti, Anna L. Merrifield, and Hendrik Andersen

Abstract. Fog and low clouds (FLCs) are essential moisture sources for Namib desert ecosystems. However, their response to climate change remains uncertain because fog processes are not resolved in climate models. Here, we apply a cloud-controlling factor framework in which FLC anomalies are expressed as a linear function of large-scale meteorological drivers, including estimated inversion strength (EIS), relative humidity at 700 hPa (R700), sea surface temperature (SST), and the eastward and northward components of 10 m wind (U10, V10). Sensitivities of FLCs to these drivers are quantified using a statistical model. By applying these sensitivities to projections of the corresponding predictors from CMIP6, we produce the first observationally constrained projections of Namib FLC occurrence. Projected trends remain uncertain and scenario-dependent; however, a robust physical signal emerges. Changes in FLCs are governed by competing influences: SST increase over the southeast Atlantic region reduces FLCs, while increased lower-tropospheric stability as well as circulation changes enhance them. Overall, these results suggest that Namib FLCs may be more resilient to climate change than previously assumed, raising the question of whether similar compensating mechanisms operate in other eastern-ocean boundary-layer upwelling systems, such as those of the Atacama Desert and California.

Received: 04 Jun 2026 – Discussion started: 25 Jun 2026

Competing interests: Hendrik Andersen is guest editor for the inter-journal (ACP/AMT/AR/ESSD) Special Issue “Aerosol, fog, climate, and biogeochemistry in southern Africa”. The remaining authors declare that they have no conflicts of interest.

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.

Download & links

Alexandre Mass, Jan Cermak, Reto Knutti, Anna L. Merrifield, and Hendrik Andersen

Status: closed (peer review stopped)

RC1:
'Comment on egusphere-2026-3242', Anonymous Referee #1, 25 Jul 2026
This paper presents the CMIP6-projected changes of fog and low clouds (FLCs) in the Namib Desert by applying a linear cloud-controlling factor (CCF) framework regressed from satellite observations and atmosphere reanalysis. It also discusses the contributions from different thermodynamic and circulation factors to the FLC changes and their uncertainties.
Major concerns:
This paper is largely built upon Mass et al. (2026) which is a preprint demonstrating the CCF framework for Namib Desert FLCs and discussing the importance of different CCFs from historical data. Mass et al. (2026) is essential for this paper because future FLC projections are directly computed from CCF sensitivities estimated during the historical period. Given a specific reanalysis dataset, these sensitivities are fixed and thus the link between future FLC projections and future CCF changes is quite straightforward. As a result, I feel the scientific contents presented in this paper is insufficient as an independent ACP paper, especially considering the substantial uncertainty in FLC projections. I would recommend either combining these two manuscripts into a more comprehensive paper or extending this paper with more analysis and more insights into the future FLC changes. Indeed, many of my following comments are addressed in Mass et al. (2026), suggesting that they may fit better into a single paper.

I appreciate the discussion of the regional differences in FLC changes in the introduction and the analyses. However, this difference is mainly described as coastal versus inland regions in the introduction while it is demonstrated as central (CN) versus Angolan Namib (AN) in the analyses. I feel that the CN vs. AN setup needs better justifications. I also love to see more careful discussion of the regional differences in Section 2.2, for example, by comparing the two columns in Figure 2 in more details.

Minor comments:
Fogs may also be important for radiation (Anber et al., 2015, https://doi.org/10.1073/pnas.1505077112).

The title and the abstract say “more resilient than previously thought” or “more resilient to climate change than previously assumed”, but the comparison between the results in this paper and referenced previous results is not clearly stated.

In Figure 1, is total uncertainty represented by one standard deviation? It could be stated more clearly.

Line 84 mentions that GCM uncertainty remains relatively constant with time. This could be better illustrated with absolute values in addition to the fractional contributions in Figure A2.

Within one paragraph, there are two figures mentioned but not shown (Lines 94 and 97). I feel that having too many “not-shown” figures may weaken the credibility of this paragraph.

Line 111 argues for a positive FLC change due to drier free troposphere. However, Figure 2 displays opposite signs of RH700 contributions in two regions. Therefore, other possible mechanisms such as cloud-top entrainment may need to be discussed as well.

Line 113 states that SST contributes to the thermal contrast between the ocean surface and the overlying air masses. I think it could also be explicitly linked to surface turbulent fluxes in this sentence.

Line 117 uses the term “thermal wind”, but I am not sure if it is an accurate term here. Thermal wind usually means the vertical wind shear due to a thermal gradient, and it is more pronounced in the free troposphere instead of 10-m winds analyzed here.

Line 132 argues that the inter-model spread mostly arise from SST contributions. However, if excluding FGOALS-f3-L as an outlier, then the contributions from EIS can be as large as or greater than SST.
Citation: https://doi.org/10.5194/egusphere-2026-3242-RC1
RC2:
'Comment on egusphere-2026-3242', Anonymous Referee #2, 02 Aug 2026
Mass et al. 2026 (egusphere-2026-3242; M26b) explores the future change of fog and low clouds with the use of cloud controlling factor framework applying to an ensemble of CMIP6 models. By using two sets of FLC sensitivities to meteorological drivers derived from observational reanalyses (MERRA-2 and ERA5) and large-scale meteorology changes from CMIP6 projections, Mass et al. finds future FLC responses are largely modulated between competing SST and circulation changes.
Major concerns:
Methodology discussed in this manuscript is identical to Mass et al. 2026 (egusphere-2026-1208; M26a), where the difference between the two lies in the using reanalysis data (M26a) vs simulated CMIP6 projections (M26b) of cloud controlling factor changes. Mass et al. apply the well-established cloud controlling factor approach to a fog and low cloud composite over a coastal domain. While the extension to fog is of interest, since fog and low clouds cannot be disentangled in the choice of dataset (SEVIRI), this raise the question of whether the derived sensitivities are informing explicitly about fog behavior, or if they are largely reflecting low cloud signal that dominates the composite, in which the results may not represent a substantial departure from existing marine low cloud CCF studies (e.g. Myers et al. 2021). Authors should clarify the degree to which fog is specifically driving the statistical relationships as identified. If composite is dominated by low cloud occurrence, the claim of this work addressing the gap in understanding future fog changes will be harder to sustain, as sensitivities would largely replicate what is known from the marine low cloud literature. More analyses and discussion will be needed to make this an independent paper from that of M26a; or, as suggested by Reviewer 1, can be included in M26a as a single manuscript.

M26b focuses on two fog hostposts in the Namib Desert - the Central Namib (CN) and Angolan Namib (AN). Using two hotspots to support the manuscript title and conclusions about the entire Namib Desert is a bit beyond what the data can reasonably support. It is recommended to provide either more reasoning on how these two prominent fog hotspots can be representative of the entire desert, or additional analysis on the changes of FLC responses under future climate across more hotspots in the Namib Desert.

There is an implicit assumption in the method that the sensitivities of FLC to drivers derived from present climate (reanalyses) will be applicable to the future changed climate. While Klein et al. (2017) discussed this assumption in the cloud controlling factor framework, and Ceppi and Nowack (2021) and Myers et al. (2021) provide supporting evidence from within GCMs, marine low cloud responses can be strongly state-dependent and nonlinear at higher warming levels (e.g., Chammas et al. 2026). Under scenarios like SSP585, the plausibility of the linear stationarity assumption for FLC sensitivities to large-scale meteorology merits more explicit discussion. It is recommended to 1) include this limitation in the main text, as it bears directly on the interpretation of the central results, and 2) provide additional discussions on why the statistical relationship can be expected to hold true under future conditions, or reframe the results more cautiously as conditional on this assumption.

Minor comments:
Additional elaboration on "advective nature" in L32 can be helpful.

Would recommend removing 'the's before CN/AN (short forms of Central and Angolan Namib), but this is at the authors' discretion.

L63: for -> during

L60-64: Terms like 'weak' and 'statistically significant' are used repeatedly within the 4 sentences; slight reorganizing of the paragraph can be helpful in bringing clarity.

Is it looked into why CN-MERRA2 and CN-ERA5 has such drastic difference in magnitude (L66)?

Recommend capitalizing gcm as GCM in σgcm.
Citation: https://doi.org/10.5194/egusphere-2026-3242-RC2

Status: closed (peer review stopped)

RC1:
'Comment on egusphere-2026-3242', Anonymous Referee #1, 25 Jul 2026
This paper presents the CMIP6-projected changes of fog and low clouds (FLCs) in the Namib Desert by applying a linear cloud-controlling factor (CCF) framework regressed from satellite observations and atmosphere reanalysis. It also discusses the contributions from different thermodynamic and circulation factors to the FLC changes and their uncertainties.
Major concerns:
This paper is largely built upon Mass et al. (2026) which is a preprint demonstrating the CCF framework for Namib Desert FLCs and discussing the importance of different CCFs from historical data. Mass et al. (2026) is essential for this paper because future FLC projections are directly computed from CCF sensitivities estimated during the historical period. Given a specific reanalysis dataset, these sensitivities are fixed and thus the link between future FLC projections and future CCF changes is quite straightforward. As a result, I feel the scientific contents presented in this paper is insufficient as an independent ACP paper, especially considering the substantial uncertainty in FLC projections. I would recommend either combining these two manuscripts into a more comprehensive paper or extending this paper with more analysis and more insights into the future FLC changes. Indeed, many of my following comments are addressed in Mass et al. (2026), suggesting that they may fit better into a single paper.

I appreciate the discussion of the regional differences in FLC changes in the introduction and the analyses. However, this difference is mainly described as coastal versus inland regions in the introduction while it is demonstrated as central (CN) versus Angolan Namib (AN) in the analyses. I feel that the CN vs. AN setup needs better justifications. I also love to see more careful discussion of the regional differences in Section 2.2, for example, by comparing the two columns in Figure 2 in more details.

Minor comments:
Fogs may also be important for radiation (Anber et al., 2015, https://doi.org/10.1073/pnas.1505077112).

The title and the abstract say “more resilient than previously thought” or “more resilient to climate change than previously assumed”, but the comparison between the results in this paper and referenced previous results is not clearly stated.

In Figure 1, is total uncertainty represented by one standard deviation? It could be stated more clearly.

Line 84 mentions that GCM uncertainty remains relatively constant with time. This could be better illustrated with absolute values in addition to the fractional contributions in Figure A2.

Within one paragraph, there are two figures mentioned but not shown (Lines 94 and 97). I feel that having too many “not-shown” figures may weaken the credibility of this paragraph.

Line 111 argues for a positive FLC change due to drier free troposphere. However, Figure 2 displays opposite signs of RH700 contributions in two regions. Therefore, other possible mechanisms such as cloud-top entrainment may need to be discussed as well.

Line 113 states that SST contributes to the thermal contrast between the ocean surface and the overlying air masses. I think it could also be explicitly linked to surface turbulent fluxes in this sentence.

Line 117 uses the term “thermal wind”, but I am not sure if it is an accurate term here. Thermal wind usually means the vertical wind shear due to a thermal gradient, and it is more pronounced in the free troposphere instead of 10-m winds analyzed here.

Line 132 argues that the inter-model spread mostly arise from SST contributions. However, if excluding FGOALS-f3-L as an outlier, then the contributions from EIS can be as large as or greater than SST.
Citation: https://doi.org/10.5194/egusphere-2026-3242-RC1
RC2:
'Comment on egusphere-2026-3242', Anonymous Referee #2, 02 Aug 2026
Mass et al. 2026 (egusphere-2026-3242; M26b) explores the future change of fog and low clouds with the use of cloud controlling factor framework applying to an ensemble of CMIP6 models. By using two sets of FLC sensitivities to meteorological drivers derived from observational reanalyses (MERRA-2 and ERA5) and large-scale meteorology changes from CMIP6 projections, Mass et al. finds future FLC responses are largely modulated between competing SST and circulation changes.
Major concerns:
Methodology discussed in this manuscript is identical to Mass et al. 2026 (egusphere-2026-1208; M26a), where the difference between the two lies in the using reanalysis data (M26a) vs simulated CMIP6 projections (M26b) of cloud controlling factor changes. Mass et al. apply the well-established cloud controlling factor approach to a fog and low cloud composite over a coastal domain. While the extension to fog is of interest, since fog and low clouds cannot be disentangled in the choice of dataset (SEVIRI), this raise the question of whether the derived sensitivities are informing explicitly about fog behavior, or if they are largely reflecting low cloud signal that dominates the composite, in which the results may not represent a substantial departure from existing marine low cloud CCF studies (e.g. Myers et al. 2021). Authors should clarify the degree to which fog is specifically driving the statistical relationships as identified. If composite is dominated by low cloud occurrence, the claim of this work addressing the gap in understanding future fog changes will be harder to sustain, as sensitivities would largely replicate what is known from the marine low cloud literature. More analyses and discussion will be needed to make this an independent paper from that of M26a; or, as suggested by Reviewer 1, can be included in M26a as a single manuscript.

M26b focuses on two fog hostposts in the Namib Desert - the Central Namib (CN) and Angolan Namib (AN). Using two hotspots to support the manuscript title and conclusions about the entire Namib Desert is a bit beyond what the data can reasonably support. It is recommended to provide either more reasoning on how these two prominent fog hotspots can be representative of the entire desert, or additional analysis on the changes of FLC responses under future climate across more hotspots in the Namib Desert.

There is an implicit assumption in the method that the sensitivities of FLC to drivers derived from present climate (reanalyses) will be applicable to the future changed climate. While Klein et al. (2017) discussed this assumption in the cloud controlling factor framework, and Ceppi and Nowack (2021) and Myers et al. (2021) provide supporting evidence from within GCMs, marine low cloud responses can be strongly state-dependent and nonlinear at higher warming levels (e.g., Chammas et al. 2026). Under scenarios like SSP585, the plausibility of the linear stationarity assumption for FLC sensitivities to large-scale meteorology merits more explicit discussion. It is recommended to 1) include this limitation in the main text, as it bears directly on the interpretation of the central results, and 2) provide additional discussions on why the statistical relationship can be expected to hold true under future conditions, or reframe the results more cautiously as conditional on this assumption.

Minor comments:
Additional elaboration on "advective nature" in L32 can be helpful.

Would recommend removing 'the's before CN/AN (short forms of Central and Angolan Namib), but this is at the authors' discretion.

L63: for -> during

L60-64: Terms like 'weak' and 'statistically significant' are used repeatedly within the 4 sentences; slight reorganizing of the paragraph can be helpful in bringing clarity.

Is it looked into why CN-MERRA2 and CN-ERA5 has such drastic difference in magnitude (L66)?

Recommend capitalizing gcm as GCM in σgcm.
Citation: https://doi.org/10.5194/egusphere-2026-3242-RC2

Alexandre Mass, Jan Cermak, Reto Knutti, Anna L. Merrifield, and Hendrik Andersen

Viewed

Total article views: 277 (including HTML, PDF, and XML)

HTML	PDF	XML	Total	BibTeX	EndNote
161	95	21	277	23	26

HTML: 161
PDF: 95
XML: 21
Total: 277
BibTeX: 23
EndNote: 26

Views and downloads (calculated since 25 Jun 2026)

Month	HTML	PDF	XML	Total
Jun 2026	1	0	1
Jul 2026	75	31	7	113
Aug 2026	44	35	5	84
Sep 2026	41	29	9	79

Cumulative views and downloads (calculated since 25 Jun 2026)

Month	HTML	PDF	XML	Total
Jun 2026	1	0	1
Jul 2026	75	31	7	113
Aug 2026	44	35	5	84
Sep 2026	41	29	9	79

Viewed (geographical distribution)

Total article views: 231 (including HTML, PDF, and XML) Thereof 231 with geography defined and 0 with unknown origin.

Country	#	Views	%

Latest update: 19 Sep 2026

Short summary

Fog is a critical water source for life in the Namib Desert. To assess how fog may change in the future, we combined satellite observations with climate model projections. Warming ocean temperatures reduce fog, but stronger onshore winds and increased atmospheric stability counteract this effect. Overall, Namib fog appears more resilient to climate change than previously thought, a finding that may apply to other fog-prone coasts worldwide.


Total:	0
HTML:	0
PDF:	0
XML:	0