Response to reviewer comment RC1:

(Comment / Response)

Summary: The manuscript presents software to access data sets that may be interesting for researchers of the paleoclimate of the past centuries. This data set comprises three ‘reconstructions’ of seasonal climate fields based on climate simulations and an off-line assimilation scheme that merged the output of these simulations with natural proxies and long instrumental records.

Recommendation: I have some recommendations to improve the clarity of the text. In some instances, the text is indeed rather unclear, for instance, in the description of the differences between the three different data sets. To really understand these differences, I needed to look up the Valler et al. (2024) paper, and I think that this manuscript should provide enough clear information for the interested reader without the need to look up the original publications. Other than these recommendations, I think that the manuscript and the software are a relevant contribution to palaeoclimate research, and it will facilitate the use of these data sets by other groups.    

• Thankyou for your comments and suggestions. We agree with your general point that the difference between the datasets was not clear in the original manuscript. We will rewrite this section to hopefully make this much more transparent. Please find below our responses to your individual points:

1) Title: I found the title too ‘literary’. This title would be fine for an internet site or a press release, but not really for a research paper. In its present form, it is not informative, and it should include specifications of the time scale, type of tool, and spatial extent of the data. I suggest including ModE-RA Climate Reanalysis, webtool, global, past centuries, and seasonal in the title and keeping the title technical. 

• We understand your argument for a more technical title, but since we are trying to attract as broad a readership as possible, we would like to retain the first part of our title – particularly since we feel this more attractive for historians (who may not even know what a reanalysis is until they read the paper). We do however agree that it would be useful to give more detailed specifications in the title, and will update it accordingly.

2) 'ClimeApp allows access to the ModE-Sim climate simulation, which is the basis of ModE-RA before assimilating early instrumental, documentary and proxy data. '

Actually, ClimApp allows access to all three data sets. This sentence may confuse the reader.

• Agreed. This will be updated to make it clearer.

3) 'ModE-Sim is a climate model experiment '

ModE-Sim is not really a climate model experiment, and this terminology can be confusing for tan average paleoclimate reader - Please, keep the expected reader in mind (!). ModE-Sim is an ensemble of global climate simulations driven by external forcings. 

• Accepted. We will use this recommendation.

4) Originally designed to form the physical basis for ModE-RA, 

This sentence may be unclear to the average reader. Observations also form ‘a physical basis’, so it can be argued that MedE_Sim and the observations both are the physical basis for ModE-RA

• Accepted. We will rephrase it to say that it only forms part of the physical basis.

5) 'The ModE-Sim ensemble mean used in ClimeApp represents the average over a set of climate states (the "ensemble members") that the model assumes to be realistic given the external forcings and boundary conditions.'

Consider a clearer version of this sentence, for instance: Each member in the ModE-Sim ensemble represents a possible climate state that is compatible  (from the model’s perspective) with the external forcing). The ensemble mean is the average over all ensemble members.

• Accepted. We will use your recommendation.

6) 'Averaging reduces temporal variability in the ensemble mean, compared with observations, but retains and highlights signals caused by variations in the forcings and boundary conditions, e.g. the climate's reaction to a volcanic eruption. '

Averaging over the ensemble members also reduces the spatial variability, not only the temporal variability.  The original sentence is, in my opinion, correct, but it may mislead the reader. Also, consider replacing boundary conditions by specifying SSt and sea-ice. This will help the average paleoclimatologist.

• Accepted. We will rewrite this to simplify it and include both spatial and temporal variability to hopefully make it clearer.

7) ‘ ModE-RA it can also help climatologists identify how observations affect the final reanalysis.’

This sentence, and actually the description of ModE-Clim is rather cryptic. 

• This use of ModE-Clim is demonstrated in the case study, so to aid the reader we will refer to the case study in this section. We do agree that the description of ModE-Clim is not entirely clear though, so we will partly rewrite this section to make it more understandable. 

8) ‘with observations increasing exponentially through time. Starting from a few thousand natural proxies and historical documents in the 15th century, by the late 19th century approximately 100000 mostly instrumental measurements are assimilated each year.’

Exponentially ? I do not tink this is the case. Probably, the authors mean increasing very rapidly - until they reach saturation. 

• Correct, the number of sources actually peaks around the 1930s since after this the reanalysis can rely on the instrumental records. We will correct this.

9) ‘ this allows accurate reconstruction of the autumn, winter and spring seasons, in addition to the widespread tree-ring based summer reconstructions.’

In principle, the setup allows for a seasonal reconstruction. Whether or not the reconstruction is accurate is another matter.

• Depending on the time period and region, ModE-RA will be more or less accurate. As long as there are early instrumental observations nearby, it will be most accurate. In times of historical and phenological information ModE-RA will be better than previous reconstructions that had only summer information. Without any data around, ModE-RA is simply a forced model simulation. Hence, we agree that “accurate” is probably too strong of a statement and we will remove it from the text.

10) ' The current resolution for ModE-RA is 1.875° (longitude) by 1.865° (latitude)'

spatial horizontal resolution

• Accepted. We will use your correction. 

11) ‘ This means that in ModE-RAclim, the externally forced signal in the model simulations is removed from the ensemble and only added back if it appears in the observations. ‘

As noted before, I found the description of Mod-E-RAclim rather confusing, and I needed to go back to the original paper by Valler et al. to really understand the difference. If I am not mistaken, the difference between ModE-RA  and ModE-RAClim is the construction of the prior. For ModE-RA, the prior is constructed from the time-aligned ensemble members of  Mod-Sim, i.g. the prior for the year 1800 is constructed from all simulated states for that particular year. For ModE-RAClim the prior is constructed from temporally non-aligned simulated states, e.g. the prior for 1800 includes all years of the ensemble ModE-Sim, regardless of the simulated year.  Is my interpretation correct? If so, please spare a few lines to describe it more clearly. If not, please consider describing the ModE-RAClim in a much more detailed manner. 

In my interpretation, the model error-covariance (spread) for ModE-RAClim is generally larger than for Mod-RA. For this reason, the impact of assimilating observations in ModE-RAClim is stronger. Please confirm if this is correct. 

• Yes, your interpretation is correct. The forced signal is removed because we take random years and ensemble members as a prior before starting the assimilation. If there is, for example, a strong volcanic eruption leading to a cooling, the data set will only show it if the assimilated observations show it. If the eruption causes, e.g., an El Nino in the model but we have no observations around the Pacific, the El Nino climate anomalies will only appear in ModE-RA due to the response of the model but not in ModE-RAclim.
• We agree that the description of ModE-Clim is currently not clear enough, so we will partly rewrite this section to make it more understandable. 

Response to reviewer comment RC2:

(Comment / Response)

Dear authors of the manuscript titled "ClimeApp: Opening Doors to the Past Global Climate New Data Processing Tool for the ModE-RA Climate Reanalysis", the development of ClimeApp is a significant advancement in making paleoclimate reanalysis data more accessible. The application adeptly bridges climatology with humanities and other non-climatological disciplines, fulfilling a crucial interdisciplinary need. The manuscript commendably elucidates the technical underpinnings and functionalities of ClimeApp, offering a detailed exposition of features such as anomaly mapping, compositing, and statistical analyses. Additionally, the use of the Tambora eruption case study effectively showcases ClimeApp’s utility in deriving new insights from historical climate events, thereby demonstrating its practical application and value.

• Thank you very much for your comments and your review.

However, the manuscript could be enhanced in the following ways:

Comparative Analysis: A comparative study between ClimeApp and other existing tools in paleoclimate research would enrich the manuscript. Such an analysis should highlight ClimeApp’s unique features and advancements, further substantiating its contribution to the field.

• Accepted. We will briefly highlight the functions and features of ClimeApp in comparison with other paleoclimate research tools and databases, such as Climate Explorer, Tambora.org and Euro-Climhist so users can easily identify for which circumstances it is best to use ClimeApp and differentiate between them.

Methodological Detailing: The manuscript could benefit from more detailed explanations in certain sections to enhance reader understanding. Specifically, a more comprehensive description of how ClimeApp differentiates between external forcing and internal variability through its statistical or computational methods would be beneficial.

• Accepted. As our paper aims for a broad interdisciplinary audience we will update the data section to make the differentiation between external forcing and internal variability more clear.

User Experience Documentation: While the paper provides an in-depth description of ClimeApp's interface functionalities, it lacks empirical data from user feedback or usability studies. Including findings from beta testing or initial user interactions would lend credence to the claims regarding the app's user-friendliness and effectiveness.

• The app did undergo extensive beta-testing both in development and at workshops and conferences. Initial user interactions and feedback were recorded and many suggested changes implemented. However we felt that detailing this would be of limited use to the audience of this paper, hence it was not included. If required, a summary could be included as an appendix, but again we feel this would be of limited benefit to the readership since the aim is to explain the functionality of the current version of ClimeApp not the development of the app.
• We feel that by including a link to the app, readers will be able to verify for themselves the user-friendliness and effectiveness of the app.

Future Development Roadmap: The discussion regarding future enhancements and the expansion potential of ClimeApp is intriguing yet lacks specificity. Detailing forthcoming features, enhancements, and a clear development roadmap, particularly concerning scalability issues like handling larger datasets or increased user traffic, would provide a clearer picture of the app’s growth prospects.

• Accepted. With regard to the future development of ClimeApp, we will clearly depict the roadmap, particularly concerning the priority and order of upcoming features and enhancements. We are currently applying for funding to upgrade of the processing power available for ClimeApp which, combined with the strategic addition of pre-processed data, should allow it to cope with simultaneous user traffic and larger datasets.