Paper revised version

[daavid00]

This PR addresses the paper review in https://github.com/OPM/pyopmspe11/pull/84/files and #89.

@totto82 and I we thank both reviewers @gassmoeller and @MatthewFlamm for their valuable comments.

We believe we have addressed all of them, which has significantly improved the quality of the paper, many thanks again.

Please let us know if this PR can be merged, or if additional changes are requiered.

[gassmoeller]

Yup, much better. I have some suggestions for the wording and you should check your paper one more time for typos. However, afterwards this is ready from my point of view.

[gassmoeller]

Suggested change

	Based on the acquired knowledge by developing the aforementioned tools, as well as from using/contributing to other open-source projects, then we have developed and made open the `pyopmspe11` tool which facilitates reproducible solutions to the SPE11 benchmark, which focus on GCS at different scales [@Nordbotten:2024].
	Based on the acquired knowledge by developing the aforementioned tools, as well as from using/contributing to other open-source projects, we have developed and made open the `pyopmspe11` tool. `pyopmspe11` facilitates reproducible solutions to the SPE11 benchmark, which focuses on GCS at different scales [@Nordbotten:2024].

[gassmoeller]

Suggested change

A previous benchmark study for GCS can be found in @Class:2009. One key differece of the SPE11 benchmark from the previous one is that no grids were given in the description, i.e., one of the main task for the participants was to create suitable grids (e.g., structured grids such as Cartesian or unscrtuctured grids such as corner-point grids). The participants were encouraged to share data (e.g., input decks, code, submitted results), with the opportunity to store the data for open access. This is where developing tools that made all steps reprodubible (i.e., preprocessing and postprocessing) become handy, and for this benchmark study, one available tool is `pyopmspe11`.

A previous benchmark study for GCS can be found in @Class:2009. One key differece of the SPE11 benchmark from the previous one is that no grids were given in the description, i.e., one of the main task for the participants was to create suitable grids (e.g., structured grids such as Cartesian or unstructured grids such as corner-point grids). The participants were encouraged to share data (e.g., input decks, code, submitted results), with the opportunity to store the data for open access. This is where developing tools that made all steps reproducible (i.e., preprocessing and postprocessing) become handy, and for this benchmark study, one available tool is `pyopmspe11`.

[gassmoeller]

Suggested change

compatible not only with OPM Flow but also with other simulators. Here, we mention two existing widely-use visualization/postrocessing software for OPM Flow: [ParaView](https://www.paraview.org) and [ResInsight](https://resinsight.org). While these tools are very useful, to the authors knolwedge, there is no existing functionality in these tools to handle all the necessary postporcessing to generate all data reporting as required in the SPE11 benchmark study. For example, to write the csv maps of the quantities from a simulation grid to a given reporting grid, where both grids do not overlap, or to compute the convective mixing as defined in Eq. (17) in [@Nordbotten:2024].

compatible not only with OPM Flow but also with other simulators. Here, we mention two existing widely-use visualization/postprocessing software for OPM Flow: [ParaView](https://www.paraview.org) and [ResInsight](https://resinsight.org). While these tools are very useful, to the authors knowledge, there is no existing functionality in these tools to generate all postprocessed data as required in the SPE11 benchmark study. For example, to write the csv maps of the quantities from a simulation grid to a given reporting grid, where both grids do not overlap, or to compute the convective mixing as defined in Eq. (17) in [@Nordbotten:2024].

Not sure the last sentence is quite correct (you can do a lot of interpolation to other specified grids and/or complex calculations (even embedded python functions) inside Paraview. However, I of course see the benefit of having all that already readily created in a specialized tool, so feel free to keep the sentence.

[gassmoeller]

Suggested change

methods efficiently interpolates quantities over time and map non-overlapping cell values (both intensive and extensive quantities) between the simulation grid and the reporting grid. The Python package Scipy [@Virtanen:2020], specifically the interp1d Class, is used for the time interpolation. The Python package Shapely [@Gillies:2024], speficically the Polygon Class, is the base for the developed methods in `pyopmspe11` to handle the mapping from the simulation grid to the reporting grid.

methods efficiently interpolates quantities over time and map non-overlapping cell values (both intensive and extensive quantities) between the simulation grid and the reporting grid. The Python package Scipy [@Virtanen:2020], specifically the interp1d Class, is used for the time interpolation. The Python package Shapely [@Gillies:2024], specifically the Polygon Class, is the base for the developed methods in `pyopmspe11` to handle the mapping from the simulation grid to the reporting grid.

[daavid00]

Many thanks @gassmoeller for your comments and corrections. We will wait for the comments from @MatthewFlamm before merging this 🙂.