the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 01 Dec 2025
Effects of Model Grid Spacing for Warm Conveyor Belt (WCB) Moisture Transport into the Upper Troposphere and Lower Stratosphere (UTLS) – Part II: Eulerian Perspective
Abstract. Warm conveyor belts (WCBs) are important features of extratropical cyclones that transport water vapor and hydrometeors into the upper troposphere and lower stratosphere (UTLS), influencing Earth's radiative budget. Previous studies have demonstrated that the horizontal grid spacing of numerical weather prediction (NWP) models influences modeled WCB properties such as ascent rates and diabatic heating. This two-part study investigates how model grid spacing affects the transport of moisture. We analyze two ICON model simulations of one North Atlantic WCB case study: a convection-parameterizing run at ~ 13 km and a convection-permitting run at ~ 3.5 km approximate grid spacing. Here, present the Eulerian perspective to complement the Lagrangian perspective from the first part of this study. We determine that the convection-parameterizing simulation produces a more humid UTLS in the WCB outflow. This results from (i) larger and fewer ice crystals, slowing the depletion of supersaturation, and (ii) the convection parameterization scheme, which injects excess vapor into the UTLS, when compared to the convection permitting simulation. The convection-permitting simulation experiences larger vertical velocities, which allows for the formation of thicker clouds with more graupel. Cloud-top temperatures are similar, yet the convection permitting simulation produces more outgoing long-wave radiation, which we can attribute to differences in UTLS vapor. Our findings indicate that convection-parameterizing simulations likely misrepresent moisture and hydrometeor transport by WCBs. This has implications for how global climate models simulate the radiative impact of WCBs and their potential influence on upper-level flow.
- Preprint
(10809 KB) - Metadata XML
- BibTeX
- EndNote
Status: final response (author comments only)
- RC1: 'Comment on egusphere-2025-5631', Anonymous Referee #1, 08 Jan 2026
-
RC2: 'Comment on egusphere-2025-5631', Anonymous Referee #2, 07 Feb 2026
I thank the authors for a well-written and interesting manuscript addressing the impact of model grid spacing on WCB moisture transport using an Eulerian perspective. The study is timely, the figures are of high quality, and the topic is highly relevant for the ACP readership. My comments in the attached document are intended to help strengthen the clarity, robustness, and interpretability of the results. In particular, I raise some substantive concerns regarding the dependence on a companion paper still under review, the description of the experimental setup, and the attribution of the results to convection versus resolution effects. I therefore recommend major revision and hope that the detailed comments provided in the attachment will be helpful in improving the manuscript.
-
RC3: 'Comment on egusphere-2025-5631', Anonymous Referee #3, 17 Feb 2026
General Comments
This is the second part of a two-part paper looking into moisture transport in a warm-conveyor-belt (WCB) case study. This paper in particular presents results from an Eulerian point of view. The paper’s topic is within ACP’s scope. The study consists of the analysis of two simulations of the same case at two (three) different resolutions. One simulation (at 13 km grid spacing) includes a convection parameterisation scheme; the second simulates convection explicitly. The study also describes a new method to determine the location of a WCB from an Eulerian point of view. While this method makes sense in general, I point out some aspects that require clarification. The analysis is generally sound. However, I am not fully convinced by the last part of the analysis, as I detailed below in my Specific Comments section. In my view the manuscript also requires some work to tidy it up in terms of statistical analysis as well as organisation and formatting that although minor when taken all together might represent more substantial changes. Therefore I cannot recommend the publication of the paper in ACP in its current form.
Specific Comments
L86-90: Please expand on the discussion on the influence of the convection parameterisation scheme on the trajectories in the low resolution simulation. Why is the vertical motion of the trajectories not being influenced by the parameterisation of convection? Do you simply mean that the scheme does not output a momentum tendency? In my view the vertical motion is still affected, albeit indirectly, by the scheme by modifying the atmosphere’s thermal structure. Then, I don’t understand your next sentence: “In particular, the latter means we are only able…”. Isn’t it contradictory to say that a WCB trajectory would be able to see the impact of convective moisture transport if the trajectory passes over a grid-point where the convection scheme is triggered when it was just stated that the vertical motion of the trajectories are not influenced by the convection parameterisation? I think this paragraph needs rewriting to make it clearer.
L166-170: I really don’t understand step 2. First, you discard all grid points with q_v less than a custom threshold, but then you say that this allows the mask to ‘crawl’ forward. How can the mask ‘crawl’ forward (‘grow’?) by removing grid points? I would expect the exact opposite effect! Then, why do you need to iterate? Why don’t you simply fix the threshold, or does this threshold depend on the iteration? Please, expand your explanation on this step.
Section 5: I’m not fully convinced about the analysis presented in Section 5. I agree there is a link between OLR and the amount of moisture that is placed at UTLS by each simulation. However, I’m not quite sure that this can be uncovered by analysing CTT and OLR separately. Even though this is not stated anywhere in the text, I understand that OLR is output by the model. What about CTT? Is it not possible to output CTT? I agree that the method to find CTT is sensible, but it will still introduce biases when compared to the model’s thermodynamics. I’m very surprised about the functional relationship between OLR and CTT (Fig 18). I would have expected a more monotonic relationship. After all, OLR is used to estimate cloud height.
The analysis will benefit also from my recommendation to conduct significance tests on all your statistical results. Several arguments seem to rely on differences between distributions or means but it is difficult to gauge how important these differences are without a statement on significance.
I’m not sure it is correct to talk about interactions between q_v and OLR at pressures that are not the top of the clouds (or the surface in the case of clear-sky conditions). If the interaction occurs at other points, then it is not outgoing longwave radiation.
Technical Comments
L6: Add ‘we’ after ‘Here’.
L14-15: The quoted implication seems limited in scope. Are there any broader implications, e.g. considering the Earth’s planetary energy budget?
L40: Reference formatting: replace semi-colon by ‘and’ in between Oertel et al. (2025) and Schwenk et al. (2025).
L90: Add ‘to’ between ‘is’ and ‘expand’.
L121: ‘Neighboring grids interact through a two-way coupling’. But this is only for R03B08 and R03B09, or is R03B07 also included? A table with the features of the simulations or the domains would be useful to explain how the simulations fit together. For example, I’m not clear on whether the 6.5km grid covers the whole area inside the red rectangle in Fig 1b or only the area between the red and the black rectangles.
L127-129: The sentence ‘In contrast to the nested simulation…’ is a repetition of what was stated just before it and therefore can be deleted.
L132-133: Format: Put the references to Miltenberger et al. and Oertel et al. between brackets.
L138: I don’t understand what is meant by ‘within a visually defined region at a specific time.’ When is this time? And how was the region visually defined?
Figure 1 caption: The caption refers to nested domains 2 and 3, but these have not been defined anywhere in the text. Instead the designations RXXBYY have been used. Decide on one nomenclature and use it consistently throughout the manuscript.
L175: TF2 and TF3 are first mentioned here without explaining what they are. I know after reading the manuscript that they are going to be defined later, but I recommend to re-organise the manuscript so that it follows a logical order.
L176: ‘The Area and volume of the masks…’: Is the area referred to here the horizontal area? Is the volume achieved by integrating over pressure levels?
L184: Delete ‘that across’.
L190-191: The statement ‘If there are too few trajectories at a given pressure level, regions of the WCB could still be excluded, even with dilation.’ contradicts the one in lines 180-181: ‘In theory, one trajectory per pressure level could suffice to fill the WCB area at that pressure level through dilation’. Perhaps it would be enough to explain why or how the conditions of the theory are not satisfied.
L211: You defined three time frames, but only gave any details of two of them. What is the third TF? I’m not sure ‘time frame’ is a very fortunate name. ‘Frame’ is an instant in time, rather than a period or a stage.
L214: It should read ‘both simulations’.
L215: From the times defining the TFs they overlap. Is this intended or is there a typo somewhere?
L223: ‘More chaotic’ (also in L463). ‘Chaos’ has a more or less precise definition involving, e.g. sensitivity to initial conditions. Is this what you mean here? Otherwise, it’s probably best to avoid the term.
L223: ‘2017-09-23’ I believe ACP has a specific format for times and dates. Please refer to the journal’s guidelines.
L232: Define all symbols such as N_i, q_i and r_i when they first appear in the text.
L234: The figures are discussed in a different order to that in which they appear in the manuscript. Re-order them.
L235-236: How realistic are the peaks at low values of r_i? The one at 250 hPa (Fig. 9b) gives the impression to be related to some sort of threshold as values suddenly drop from high to zero.
L260: Rewrite by substituting ‘they’ for ‘both simulations’.
L272: ‘patchiness’ could be replaced or perhaps described further as containing small-scale structure.
L276: Remove ‘This is an interesting result for those who study supersaturated regions in UTLS (and for instance their fractal characteristics).’ Or expand on how this is interesting and add references.
L282: tau_{sat,ice} has not been defined.
Figure 5 (and all histograms): How have the distributions been normalised?
L296: ‘As these areas account for a substantial proportion of the data points’ Where is this shown or how do we know this?
L296: Avoid ‘it is obvious’. At least it is not so obvious to me. In fact, why then does the red distribution not dominate?
L327-328 (and all comparisons between histograms and statistics): Include some statistical significance analysis in all comparisons between histograms and statistics (mean, median). Otherwise all assertions about which one is larger become very subjective. They may appear different but if that difference is not significant then the interpretation becomes more complex.
L368: The figure for N_i is included in the main manuscript but that for q_i is only in the appendix. What criteria was taken to decide what figures to include in the main text and what to relegate to the appendix?
L368: Can the differences in histograms between simulations be an artefact of the mask construction method? It looks like the trajectories cover more fully the Eulerian mask in the nested simulation.
L371: ‘not observed’ Do you mean ‘not included’ or some other term? You are not talking about observations here.
L410-411: Is there a functional relationship between OLR and CTT? If so, what is it?
L435: It should read ‘chosen’.
Figure A2 caption: Do you really mean heatmaps or simply maps (horizontal cross-sections)?
Citation: https://doi.org/10.5194/egusphere-2025-5631-RC3
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 249 | 134 | 22 | 405 | 35 | 26 |
- HTML: 249
- PDF: 134
- XML: 22
- Total: 405
- BibTeX: 35
- EndNote: 26
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
In this manuscript, Schwenk and Miltenberger present a North Atlantic warm conveyor belt (WCB) case study comparing moisture transport, hydrometeor properties, and radiative effects between convection-permitting and convection-parameterizing simulations. A key contribution is a novel algorithm that combines Lagrangian trajectory information with moisture fields to construct three-dimensional WCB masks, which are then used for an Eulerian analysis of WCB outflow regions.
The studies findings -- more moisture, fewer but larger hydrometeors, and reduced outgoing longwave radiation in the WCB outflow of the convection-parameterizing simulation -- are highly relevant for understanding WCB-related biases in upper-level dynamics and radiative budgets in climate models.
I particularly commend the authors on their innovative approach to create WCB masks and the generally clear and well structured presentation.
The effort to diagnose and interpret the processes underlying the simulated differences is especially valuable, as it points toward potential pathways for improving convection parameterizations.
I look forward to seeing the study published after one major issue is addressed. I lay out this concern, along with additional general and my specific comments on individual sections and figures in the attached document.