Coupling the atmospheric model Meso-NH-v5.5 with the Monte-Carlo solver of conductive-radiative-convective heat exchanges stardis-v0.11.1 to calculate the surface energy balance of complex geometries

Schoetter, Robert; Caliot, Cyril; Retailleau, Florent; Nagel, Tim; Coustet, Christophe; Eymet, Vincent; Forest, Vincent; Hourdin, Frédéric; Inagaki, Atsushi; Kotthaus, Simone; Masson, Valéry; Morrison, William; Villefranque, Najda

doi:10.5194/egusphere-2026-2061

Preprints

https://doi.org/10.5194/egusphere-2026-2061

Preprints

15 Jun 2026

| 15 Jun 2026

Status: this preprint is open for discussion and under review for Geoscientific Model Development (GMD).

Coupling the atmospheric model Meso-NH-v5.5 with the Monte-Carlo solver of conductive-radiative-convective heat exchanges stardis-v0.11.1 to calculate the surface energy balance of complex geometries

Robert Schoetter, Cyril Caliot, Florent Retailleau, Tim Nagel, Christophe Coustet, Vincent Eymet, Vincent Forest, Frédéric Hourdin, Atsushi Inagaki, Simone Kotthaus, Valéry Masson, William Morrison, and Najda Villefranque

Abstract. The meteorological impact on humans and infrastructure in urban areas is strongly influenced by the air temperature, humidity, wind velocity, and radiative fluxes. These can be strongly heterogeneous in urban environments with complex 3-D building and vegetation geometry. Micrometeorological building-resolving models have been developed to solve the prognostic equations of the atmospheric variables in urban areas at metric resolution. They take into account the radiative exchanges in the 3-D urban geometry and heat conduction in the building and ground surfaces with separate deterministic approaches. An alternative introduced in this study is to solve the 3-D heat conduction, radiation, and convection with the stochastic Monte-Carlo Method (MCM) in a unified algorithm. Such MCM solvers have become available thanks to recent breakthroughs in the Monte-Carlo community, which allow to establish a connection between conductive and radiative random walks. The advantage of the MCM is that the calculation is independent of the complexity of the urban geometry, the calculations can deal with a large range of scales, and are easy to parallelise. In this study, the coupling between the atmospheric model Meso-NH and the Monte-Carlo solver stardis of conductive-radiative-convective heat exchanges is presented. Humid processes cannot yet be considered with the new model. The coupled model is evaluated against observational data for dry heat wave conditions at the impervious Comprehensive Outdoor Scale MOdel (COSMO) urban-like site in the northern outskirts of Tokyo (Japan). Evaluation results show that the model represents well the temporal evolution of the sensible heat flux and the vertical profiles of air temperature in the urban roughness sublayer. Future developments could focus on including urban vegetation and moist processes such as to enable the study of adaptation measures in urban areas.

How to cite. Schoetter, R., Caliot, C., Retailleau, F., Nagel, T., Coustet, C., Eymet, V., Forest, V., Hourdin, F., Inagaki, A., Kotthaus, S., Masson, V., Morrison, W., and Villefranque, N.: Coupling the atmospheric model Meso-NH-v5.5 with the Monte-Carlo solver of conductive-radiative-convective heat exchanges stardis-v0.11.1 to calculate the surface energy balance of complex geometries, EGUsphere [preprint], https://doi.org/10.5194/egusphere-2026-2061, 2026.

Received: 10 Apr 2026 – Discussion started: 15 Jun 2026

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

Status: open (until 26 Aug 2026)

Post a comment Subscribe to comment alert

RC1:
'Comment on egusphere-2026-2061', Anonymous Referee #1, 15 Jul 2026 reply
Summary:
The authors present the coupling of a surface energy balance model (stardis) to an atmospheric model (Meso-NH). They first describe the two models that are coupled and then how Meso-NH forces stardis, including limitations. The authors test their model with observations from an urban-like setup in Japan which is represented by concrete cubes. Then, the authors present different model configurations and how they compare to the “standard” model (TEB) and observations. Various metrics are analysed including temporal evolution of sensible heat flux, temperature profiles and top/wall/ground temperatures. The metrics show that the coupled model can generally represent the spatial and temporal patterns of temperature but still faces issues such as biases in temperature and difficulties in capturing the temperature peak at the cube height. The discussion includes a test of the convergence of the algorithm used in stardis and computational costs. Overall, the manuscript describes the model development of a subgrid model and the evaluation compared to observations and different model configurations. In general, the writing and presentation is clear, but the structure could be improved, see comments below.
General comments:
The manuscript is in some places convoluted, making it difficult to get the main message. For example:
The limitations in section 2.5 list 11 points, but they could be summarized as some have the same cause. Point 2 and 5 refer both to the time step. Point 7 and 11 relate to the fact that only dry and sunny situations were considered. Point 9 and 10 relate to the MC algorithm

Figure 7 seems to have limited added value and could be moved to the appendix. The values look similar to the models on Figure 6 and only show larger variability at noon.

Figure 9,10, 11: the added value of showing the same plot for three different days is not clear. Similar for Figure 12 and 13.

The description of the coupling can be improved. For an unfamiliar reader, it would be better to first describe the components (Meso-NH-IBM and stardis) and then describe how they are coupled. It is mostly structured like this already, but the section starts with a quick description of the coupling, reference to Fig 1-3 and then describes the components, which is a bit confusing. Additionally, it was not clear on which scales both models operate until section 3. Although the authors mention “metric” scale, it would be helpful to add numbers as the exact resolution was first mentioned on page 14. The meaning of image and ghost points is not entirely clear. The shading in in Figure 3 makes it difficult to understand the path. It is not clear why data from Meso-NH-IBM is saved to disc and then read by stardis, which seems to be a less efficient and rather static coupling. The authors mentioned in the conclusion that it could be kept in memory, which makes it questionable why data must be written to and read from disc. Additionally, it is not clear why the path in stardis is computed backward in time. Although the authors mention Caliot et al. 2026, a quick comment would be helpful.

The authors mention 7 different model configurations and find that basically none of the configurations have a significant effect on the simulation except for the grid resolution. The authors could focus on this parameter and investigate it further.

Specific comments:
The authors justify using Monte Carlo Methods by mentioning “recent breakthrough in the MC community” but don’t provide a reference to the “breakthrough”.

The authors discuss current limitations saying that “some of the most important limitations are mentioned”. This gives the impression, that there are more limitations that are important. If so, the authors should mention the complete list, at least in the appendix.

One of the model configurations is the number of realizations (4REA). This seems to be an important part but is only discussed in the discussion section and not mentioned in the results section. The authors mention “4REAleads to higher precision of T_surf for individual probe points, but does not improve the average H or vertical profiles of θ.” in L611 but don’t show the results. Figure 15 is not clear for readers that are not experts of Monte Carlo methods.

Section 4.1/ Figure 6. The authors discuss the bias of the models compared to observations. Another interesting point to discuss would be “why is the variability of H low during the night and high during the day?”

L503-506: “The observed θ difference between 5 m and just above the ground is about 1 K in the nighttime and 2-3 K between 12:00 and 16:00 when H is highest.” This is not visible in the plot.

L534: “for ε = 0.89 and ν = 0.38, which are the most plausible values of these parameters.” Sensitivity analysis and figure are mentioned later. The authors could add a reference to the figure already here. Additionally, it seems that none of the configurations matches the peaks in Fig. 14 well. This observation is mentioned but not explained.

Section 4.4: “differences are within 2 K” This seems a lot. Is this due to a bias in the camera or stardis? The authors mention dirt on the cubes. But, is this the only reason?

Technical comments:
L56: typo: reveiced -> received

When citing the appendix, it starts with Appendix C instead of A

L197: Equation 8: it is not clear why the others do 4*sigma*T^4-3*sigma*T^4 and then get to equation 9

L203: missing units for temperature

L206: duplicate: equation 10 was already mentioned in line 194

L221: what is capital theta?

L225 exchange coefficient (c_H) is only mentioned in the next section.

Figure 9: It might be helpful to fix the x-axis for all plots so visualize how the temperature profile changes over time.

L531: T_b not explained in the text.

L624: typo: builings -> buildings

Reply
Citation: https://doi.org/10.5194/egusphere-2026-2061-RC1

Model code and software

Source code, numerical simulations, and post processing scripts described in the research article ''Coupling the atmospheric model Meso-NH-v5.5 with the Monte-Carlo solver of conductive-radiative-convective heat exchanges stardis-v0.11.1 to calculate the surface energy balance of complex geometries'', submitted to Geoscientific Model Development R. Schoetter et al. https://doi.org/10.5281/zenodo.19497100

Viewed

Total article views: 79 (including HTML, PDF, and XML)

HTML	PDF	XML	Total	BibTeX	EndNote
47	28	4	79	5	2

HTML: 47
PDF: 28
XML: 4
Total: 79
BibTeX: 5
EndNote: 2

Views and downloads (calculated since 15 Jun 2026)

Month	HTML	PDF	XML	Total
Jul 2026	47	28	4	79

Cumulative views and downloads (calculated since 15 Jun 2026)

Month	HTML	PDF	XML	Total
Jul 2026	47	28	4	79

Viewed (geographical distribution)

Total article views: 55 (including HTML, PDF, and XML) Thereof 55 with geography defined and 0 with unknown origin.

Country	#	Views	%

Latest update: 27 Jul 2026

Short summary

Temperature, humidity, radiation, and wind can have a negative impact on human comfort and infrastructure in urban areas. Building-resolving meteorological models can help to quantify such meteorological impacts and the potential benefits of adaptation measures to climate change. Here, a new unified algorithm to calculate the radiation, heat conduction, and sensible heat exchange in urban areas is coupled with an atmospheric model, and evaluated against experimental data.


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