ECOMIP: The EarthCARE-ORCESTRA Model Intercomparison Project

Roh, Woosub; Satoh, Masaki; Fiévet, Romain; Fielding, Mark D.; Hogan, Robin J.; Stevens, Bjorn; Kubota, Takuji; Okamoto, Hajime

doi:10.5194/egusphere-2026-3606

Preprints

https://doi.org/10.5194/egusphere-2026-3606

Preprints

16 Sep 2026

| 16 Sep 2026

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

ECOMIP: The EarthCARE-ORCESTRA Model Intercomparison Project

Woosub Roh, Masaki Satoh, Romain Fiévet, Mark D. Fielding, Robin J. Hogan, Bjorn Stevens, Takuji Kubota, and Hajime Okamoto

Abstract. This study proposes the EarthCARE-ORCESTRA Model Intercomparison Project (ECOMIP), a framework for evaluating various range of atmospheric models, including global or regional km-scale models using observations from the EarthCARE satellite and the ORCESTRA (Organized Convection and EarthCARE Studies over the Tropical Atlantic) field campaign. The goal of ECOMIP is not to rank model performance but to provide a clear path to improve models by considering how new satellite observations, combined with in situ measurements and satellite simulators, can constrain cloud microphysics, precipitation, and convective dynamics. This paper presents the protocol for Phase 1 of the ECOMIP framework, consisting of 2-day simulations designed for comparison with EarthCARE and ORCESTRA observations. As a first examination of the model intercomparison, NICAM, ICON, and IFS simulations are evaluated with EarthCARE observations during the satellite overpass over the ORCESTRA domain on 3 September 2024. Along the EarthCARE track, the occurrence of a mesoscale convective system and Southern Hemisphere cirrus clouds is captured in both simulations. The three models exhibit different convective organization characteristics. NICAM and IFS tend to produce larger and more coherent convective systems, whereas ICON shows more scattered convective cores. Although conclusions from a single EarthCARE frame cannot be generalized, these results provide an example of model-dependent characteristics and demonstrate the potential of the proposed protocol for improving our understanding through the analysis of a larger number of EarthCARE cases and datasets in future studies. We also present the comprehensive framework of ECOMIP after Phase 1, including statistical free-run simulations and observationally constrained parameter-sensitivity experiments.

Received: 19 Jun 2026 – Discussion started: 16 Sep 2026

Competing interests: At least one of the (co-)authors serves as topic editor for the special issue to which this paper belongs.

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

Woosub Roh, Masaki Satoh, Romain Fiévet, Mark D. Fielding, Robin J. Hogan, Bjorn Stevens, Takuji Kubota, and Hajime Okamoto

Status: open (until 13 Nov 2026)

Post a comment Subscribe to comment alert

RC1:
'Comment on egusphere-2026-3606', Christian Jakob, 20 Oct 2026 reply
The paper is an amalgamation of two very distinct parts. In the first part (Sections 1-3) it provides some motivation for a model intercomparison project using EarthCARE and ORCESTRA observations to evaluate km-scale models at both global and regional scale. The second part (Section 4) provides a case study of the use of (mostly) EarthCARE observations to demonstrate the feasibility and potential issues that need further exploration when using these observations in evaluating multiple models at once.

I find the first part very weak and not publishable in its current form. It is simply a proposal and I feel just publishing proposals is not warranted, which the authors might even agree with having added the very nice second part. The structure of the first part is also very confusing and not very well thought-out. An example is the fact that Section 3 is called Experimental Design, but so is Subsection 3.2. I will provide more detailed comments below.

The second half of the paper is very nice and is actually publishable more or less as is. What is missing is stronger link between the two parts and a clearer motivation for the intercomparison, which currently still sounds much as: "we have models, we have data, let's compare them", when there are really exciting opportunities emerging from part 2.

My main suggestion is to completely restructure the paper and let the intercomparison idea more naturally emerge from the results of the feasibility study (Section 4). This will make the paper much more readable and acceptable for publication. Please see my suggestions how this might work below. I am naturally open to other ideas, but the current certainly does not work.

Proposed revised paper structure

Introduction

motivate the bringing together of EarthCARE, ORCHESTRA and models for all the obvious reasons you already mention

highlight some of the challenges, such as the need for satellite simulators, likely inconsistencies in the information each model provides, challenges in running km-scale models for a long time AND storing the information needed

motivate a feasibility study with a potential intercomparison to follow

Feasibility study

present what is now Section 4 as a first attempt to see whether and how this all might work

highlight successes and likely challenges

if ORCESTRA is to be a major feature, add some more material on comparing data from it to models

Proposal for an intercomparison

write a very short section on taking the feasibility study forward into the ECOMIP project

describe the protocols in general, but put all the details in an appendix to have them recorded without disrupting the flow of the story (see also my more detailed points below)

introduce the idea that regional km-scale models could also participate and spell out how

An important aside: You make throw-away comments that low-resolution models could participate too, but you never say how. There are massive challenges for this to happen. How do we turn prametrised cloud microphysical information into the very fine-scale information needed? We do have of course some long tradition in satellite simulators in "traditional" GCMs (which you should expand on in the introduction), but this is quite a stretch for those. If you cannot provide more detail on how this might all work, I suggest to just leave this off for now and explore it at a later time.

The title would have to be changed reflect these changes. It could be something like: "Evaluating global storm-resolving models with EarthCARE/ORCESTRA observations - Feasibility and a future model intercomparison project (ECOMIP)". Or something similar

Some more detailed comments

Introduction

There is a rich history in using satellite simulators in model evaluation that dates back to the late 1990s and early 2000s. Please include some reference to that.

More broadly, there is a rich history in model to satellite to field experiment "intercomparisons", both for process-based CSRMs (for instance in the GEWEX Cloud System Study - Randall, D., Krueger, S., Bretherton, C., Curry, J., Duynkerke, P., Moncrieff, M., et al. (2003). Confronting Models with Data: The GEWEX Cloud Systems Study. Bulletin of the American Meteorological Society, 84(4), 455–469. https://doi.org/10.1175/bams-84-4-455) and in global models through Cloud Feedback Model Intercomparison Project (CFMIP) and others. I fell it would be appropriate to bring this out here even without going into great detail.

I am not sure where this comment best fits, so I put it here. One thing missing from both the evaluation plans and the feasibility study in Section 4 is the critical link to the dynamical state of the atmosphere. How to cloud properties vary with "large-scale" dynamical state? The simplest example could be ascent vs subsidence, but we probably can do better than that. There is ample literature that has attempted this in the past and it would be good to see that connection made here too. Please add at least a short paragraph on what you envisage to happen in that area.

Section 2

This section does not need to be there. It partially repeats material from the introduction (e.g., lines 163 to 170) or previews what is coming in Section 3, without spelling it out. I suggest to remove it.

Section 3

This section is very confusing. It jumps from models to observations, broad statements to fine details (in the Tables) without providing much clarity. This is not helped by the fact that it comes before we see the nice results of Section 4.

I suggest to move this entire discussion to after the feasibility study (see suggested structure above) and completely rewrite it.

I suggest to move all the fine detail (e.g., all the tables) into an appendix and refer to that as needed to make potential participants aware of what would be possible and required from them.

I suggest to add an explicit discussion on the potential participation of regional km-scale models, as there are many of them and as this would increase the community that could benefit from this by a large number of groups.

Section 4

This section is overall very nice and can remain largely as is.

I suggest to refrain from statements that models do something "reasonably well". For example, in the description of Figure 4 in lines 334-335 you say the rainfall structures in the tropics and mid-latitudes are "resonably well" reproduced. Without wanting to open a long discussion on what "reasonably well" even means, I can see very large differences between the models and in comparison to the observations. There can be many reasons for this and the differences are actually the intersting part. So why brush over them with such a vague statement?

I suggest to introduce the satellite simulator concepts and ideas when you start the comparisons to EarthCARE - currently Section 4.2.

I feel there needs to be one or two more examples on how you will use ORCESTRA, as currently this part is very weak.

I hope you find these comments helpful and I am looking forward to reading the revised manuscript.

Reply
Citation: https://doi.org/10.5194/egusphere-2026-3606-RC1
RC2: 'Comment on egusphere-2026-3606', Anonymous Referee #2, 20 Oct 2026 reply

This manuscript describes the EarthCARE-ORCESTRA Model Intercomparison Project (ECOMIP), a framework for evaluating a range of atmospheric models, including global and regional kilometre-scale models, using observations from the EarthCARE satellite and the ORCESTRA (Organized Convection and EarthCARE Studies over the Tropical Atlantic) field campaign. The manuscript has two main components: Sections 1–3 introduce the experimental protocol, while the second part presents an initial application of the framework to three models, NICAM, ICON and IFS. The overall concept of bringing together EarthCARE and ORCESTRA observations for systematic model evaluation is very valuable. However, in its present form, the manuscript needs further development. In particular, the experimental protocol does not yet appear sufficiently complete or well specified to provide a robust framework for a model intercomparison, while the initial analysis of the three models remains rather superficial. In my view, the experimental protocol is the most important part of the paper and should be strengthened considerably. I therefore recommend major revisions. Please see my specific comments below.
SPECIFIC COMMENTS
Definition of the data request. The data request needs to be defined much more precisely. In particular, the manuscript should specify variable names and definitions, dimensions, file formats, and expected data volumes. These details are important if the protocol is intended to be used consistently by multiple modelling groups. The authors should also consider adopting established CF conventions, as used in other model intercomparison projects.
Radiative fluxes. It is unclear why three-dimensional radiative fluxes are not included in the data request. Cloud-radiation interactions are a key aspect of the EarthCARE mission, so the ability to investigate the relationship between simulated cloud properties and radiative effects would seem to be an important component of ECOMIP. The authors should either include the relevant radiative quantities in the protocol or explain clearly why they have been excluded.
Data-sharing protocol. The manuscript does not appear to describe how model output will be shared between participants. A model intercomparison protocol should provide sufficient information on how the requested data will be stored, transferred or made accessible to the participating groups. A clearer description of the intended data-sharing arrangements would therefore strengthen the experimental design.
Overall maturity of the protocol. More generally, the data request would benefit from further consideration before the protocol is established. The manuscript should demonstrate that the requested variables and associated metadata are sufficient to address the main scientific questions of ECOMIP and to allow meaningful comparisons across models with potentially quite different configurations and resolutions.
Depth of the model intercomparison. The analysis of NICAM, ICON and IFS is currently rather descriptive. It would be considerably more informative if the authors attempted to connect the differences identified between the models with the radiative biases shown in Figures 5 and 6. Establishing such links would move the analysis beyond documenting differences and towards understanding their physical significance.
Use of EarthCARE observations. The analysis presented here considers radar results, but it is not clear why lidar observations are not also included. Given the complementary information provided by radar and lidar measurements, the authors should consider incorporating the lidar observations into the initial intercomparison, or provide a clear explanation for restricting the analysis to radar.

Reply

Citation: https://doi.org/10.5194/egusphere-2026-3606-RC2

Woosub Roh, Masaki Satoh, Romain Fiévet, Mark D. Fielding, Robin J. Hogan, Bjorn Stevens, Takuji Kubota, and Hajime Okamoto

Viewed

Total article views: 626 (including HTML, PDF, and XML)

HTML	PDF	XML	Total	BibTeX	EndNote
416	154	56	626	60	58

HTML: 416
PDF: 154
XML: 56
Total: 626
BibTeX: 60
EndNote: 58

Views and downloads (calculated since 16 Sep 2026)

Month	HTML	PDF	XML	Total
Sep 2026	333	96	30	459
Oct 2026	83	58	26	167

Cumulative views and downloads (calculated since 16 Sep 2026)

Month	HTML	PDF	XML	Total
Sep 2026	333	96	30	459
Oct 2026	83	58	26	167

Viewed (geographical distribution)

Total article views: 626 (including HTML, PDF, and XML) Thereof 626 with geography defined and 0 with unknown origin.

Country	#	Views	%

Latest update: 26 Oct 2026

Short summary

Clouds and rainfall are still difficult to represent accurately in weather and climate models. We introduce a new project called the EarthCARE–ORCESTRA Model Intercomparison Project (ECOMIP), which uses observations from the new Earth Cloud, Aerosol and Radiation Explorer (EarthCARE) satellite and a tropical field campaign to compare and improve atmospheric models. Initial results show that different models can reproduce the same storm system but differ in its size and structure.


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