Thanks for the opportunity to read your work. The manuscript is set to describe a new R package to calculate metrics of integrity and stability in ecosystems. The paper is clear and methods well explained with some simulation results coming from LPJmL models. 

The paper, however, falls short in introducing the software, its requirements and documentation for a larger adoption by the community. If the central goal of the paper is to introduce the `biosphericmetrics` package, more attention should be given to the software documentation. For example, the package does not contain vignettes or demos that help the user understand what are the main functionalities. Since the package is meant to be used as companion of LPJmL model output, the authors should make available a small datasets that the users can test to understand the usage of functions (the equivalent of mtcars in other packages, not the 3GB in Zenodo). The user should be able to install the package, read the help of functions, and be able to run a demo with some results. I understand the output can be large files, but you can get inspiration of other packages such as stars or terra in making available small spatial data cubes (NetCDF files) for testing.

The GitHub repository has some useful information on for example what functions to use when working on small datasets, sending jobs to servers, or plotting. However, the package documentation is not stand alone. For example, it seems one needs json files to use `calc_biocol()`, but that dependency is not documented on the help files of the function. Most functions on the package do not have examples and are not working out of the box, many assumes the user has access to Fabian’s folder structure. Thus I can install the package, but as a reviewer I cannot test if it works. The package also lacks testing and better documentation on scalability.

I strongly recommend to check the standards for software development in specialised journals such as Journal of Open Source Software, and the Journal of Statistical Software (JStatSoft). Both of them recommend the workflow and standards of the rOpenSci consortia: https://stats-devguide.ropensci.org

At minimum the package should contain a vignette explaining its functionality and a demo data that users can use to test its functionality.

Best regards,

Juan

## Specific comments

Line Comment

190 Unfinished sentence

This manuscript presents software that computes two measures of human impact on the terrestrial system, i.e. the human colonization of the biosphere and the risk of ecosystem destabilization.

To my opinion, this is a relevant manuscript, however, I must admit that I am not specialized on this topic.  I am in favor in adding the nitrogen components into the assessment.

The preprint is very well written and well-documented, with nice figures and maps. 

I only have few minor/textual comments.

• L4: “2” should be “two”; A number smaller or equal to ten should be written fully, if the number is not accompanied with physical units that refer to processes. Check also the reminder of the manuscript (i.e. L65 4/four components; L174 7/seven).
• L5: I should start with (i), and (ii) on line 6;
• L8-10: this sentence is to large and should be split in order to improve readability; the word “extraction” here should be explained here since I did not understand what the authors mean at this stage of the manuscript; Later on, it becomes clearer;
• L19 and elsewhere. References in the text: I am not familiar with the official formats of this journal, but to my opinion it is more logical to range the citations chronologically first and then alphabetically. After all, one refers first to the paper that came first;
• L51: Should be, “HANPP, however, cannot be …”;
• Add reference for Eq. 5:
• Eqs 13-17 are referred from Ostberg et al. (2018)? If not, add reference;
• L190: insert whit space between 1000 and m:
• L223-225: format “th” as superscript;
• Fig 2: Use capital M for “map”;
• L245: Better write: “Values of gi are high in regions …”;
• L272-275: It would be better to use (i), (ii), (iii), when one lists up items;
• L280: Better “Figure 5k,m”;
• L316-318: a list so use (i) and (ii);