Prototyping a Methodology for Long-Term (1680–2100) Historical-to-Future Landscape Modeling for the Conterminous United States
Round 1
Reviewer 1 Report
The paper deals with challenging issues of land use modelling. The land use issues are topical today and the subject of the paper is actual. However, text of the paper is complicated and it is difficult to read it. The general picture about FORE-SCE is not presented it makes difficult to read the text. Obviously it is not problem for the readers who are familiar with the FORE-SCE methodology. I would expect to get some general information about this method and the workflow chart might be useful. It would give more holistic picture about the carried out research.
It seems that too much material is put into one paper. Lot of the space is devoted to the data management issues in the paper. It is stated that the aim of the study was “… to produce a long-term (1680 through 2100) 107 LULC dataset…” Nevertheless, the paper deals also with backcasting and forecasting. I would propose to divide the material of that paper into two (or even three) parts and compose separated papers.
The figures are too complex. For example, Figure 2 might be divided into two separated drawings. Parts a, b and c of that Figure will be one drawing and parts c and d respectively second drawing. The similar division of some other figures is also recommended. The smaller figures will avoid the situation that the figure is on the separated pages.
The discussion part of the paper is proportionally very short and the conclusions are missing at all.
Author Response
Response to Reviewer 1 Comments
Point 1: text of the paper is complicated and it is difficult to read it. The general picture about FORE-SCE is not presented it makes difficult to read the text. Obviously it is not problem for the readers who are familiar with the FORE-SCE methodology. I would expect to get some general information about this method and the workflow chart might be useful. It would give more holistic picture about the carried out research
Response 1: The FORE-SCE model’s underlying methodologies have been detailed in prior publications, which we reference. Given the manuscript’s length we felt it prudent to focus on methodologies of new model capabilities.
Point 2: It seems that too much material is put into one paper. Lot of the space is devoted to the data management issues in the paper. It is stated that the aim of the study was ‘… to produce a long-term (1680 through 2100) 107 LULC dataset…’ Nevertheless, the paper deals also with backcasting and forecasting. I would propose to divide the material of that paper into two (or even three) parts and compose separated papers.
Response 2: The long-term landscape modeling to which the author refers is comprised of backcasting and forecasting. We believe this is clear throughout the manuscript (e.g. lines 99, 113, 124, 913) and contend that to fully characterize the scope of the FORE-SCE model all components presented in the manuscript should be examined together.
Point 3: The figures are too complex. For example, Figure 2 might be divided into two separated drawings. Parts a, b and c of that Figure will be one drawing and parts c and d respectively second drawing. The similar division of some other figures is also recommended. The smaller figures will avoid the situation that the figure is on the separated pages.
Response 3: It is unclear how the reviewer’s suggestions with regards to Figure 2 would improve clarity.
Point 4: The discussion part of the paper is proportionally very short and the conclusions are missing at all.
Response 4: Per Land journal guidelines a conclusion is not mandatory, rather, it is recommended when the discussion is long or complex. Our discussion was designed to serve as both discussion and conclusion, with a concluding paragraph on next steps for our work. We feel our discussion is sufficient and does not require additional summary, particularly when reviewers commented on the long length of the manuscript. Based on the current condition of the concluding discussion section and concerns about overall manuscript length, we decided against lengthening the manuscript with an additional section.
Reviewer 2 Report
Authors present a very interesting study regarding land use / land cover historical and future scenarios. It is well elaborated and explained.
However, some improvements can be made before publishing it:
First of all, the title talks about the “conterminous United States” but it is based in the Delaware River Basin.
Introduction section: A wider introduction about other LULC changes and LULC scenarios carried out, worldwide and in the US is needed. More information regarding FORE-SCE should be given. For example, who has developed it? Explain why it was being developed from 1936 to 2100, but for the US Great Plain was modelled from 2012 to 2100. Figures should be placed after they are named (figure 1).
Methods: Authors can introduce a table showing the historical data collected to carry out the historical scenarios and the future scenarios. That way, we can see it all the information at a glance. As authors make use of an extensive number of previously cartography made by different institutions or researchers, is there any information regarding to accuracies?
Results: It would be nice to see a calibration performance (spatial and statistical), for example, simulating land use change to year 2018, starting from a historical modeled map. In figure 4, why modeled quantities are showed as aggregated land covers if they are not aggregated in the modelling procedure?
Discussion: It should be improved. Compare your results with other LULC scenarios carried out worldwide and specifically in the US. Which LULC are modeled better and which worse and why?
There is no conclusion. At least a short paragraph is needed.
Author Response
Response to Reviewer 2 Comments
Point 1: the title talks about the “conterminous United States” but it is based in the Delaware River Basin
Response 1: The title refers to the prototyping of a methodology with the goal of using it to model land cover for the conterminous US. The prototype, or test case, is the Delaware River Basin. The intention is to scale the prototype to CONUS level modeling. We believe this is clear in the introduction (line 108), materials and methods (line 130), and discussion (lines 913, 933).
Point 2: More information regarding FORE-SCE should be given. For example, who has developed it? Explain why it was being developed from 1936 to 2100, but for the US Great Plain was modelled from 2012 to 2100.
Response 2: Clarification regarding FORE-SCE capacity vs application was added to the section 2 (beginning line 114).
Point 3: Figures should be placed after they are named (figure 1).
Response 3: Figure 1 has been moved to follow its reference.
Point 4: Authors can introduce a table showing the historical data collected to carry out the historical scenarios and the future scenarios. That way, we can see it all the information at a glance. As authors make use of an extensive number of previously cartography made by different institutions or researchers, is there any information regarding to accuracies?
Response 4: Table A.1, which includes availability of historical accuracy assessments, has been added to appendix material.
Point 5: It would be nice to see a calibration performance (spatial and statistical), for example, simulating land use change to year 2018, starting from a historical modeled map.
Response 5: Section 3.5, assessing the FORE-SCE model’s ability to simulate land cover change between RS-based reference datasets has been added.
Point 6: In figure 4, why modeled quantities are showed as aggregated land covers if they are not aggregated in the modelling procedure?
Response 6: Plotting every land cover category creates an overly complicated visual. We aggregate here to enhance visualization of trends for multiple scenarios. Language was added to clarify this in section 3.1 (line 614).
Aggregation of categories is also necessary for cross-model comparisons introduced in section 3.7. We elect to maintain this approach for both simplicity and consistency.
Point 7: A wider introduction about other LULC changes and LULC scenarios carried out, worldwide and in the US is needed. Compare your results with other LULC scenarios carried out worldwide and specifically in the US. Which LULC are modeled better and which worse and why?
Response 7: Judging which scenarios are best is largely dependent on stakeholders and application. It is not an objective of the work described in the manuscript. Able to ingest any scenario, FORE-SCE is ultimately scenario agnostic and can map any scenario at high spatial and thematic resolutions. The goal of FORE-SCE is not to produce the “best” scenario but to produce the most realistic mapping possible for any useful scenario. Language to this affect has been added in section 2.3 (starting line 197).
We have also added a section (3.7) comparing the scenarios modeled by FORE-SCE in the DRB to harmonized global integrated assessment model scenarios for the same region and time period.
Point 8: There is no conclusion. At least a short paragraph is needed.
Response 8: Per Land journal guidelines a conclusion is not mandatory, rather, it is recommended when the discussion is long or complex. Our discussion was designed to serve as both discussion and conclusion, with a concluding paragraph on next steps for our work. We feel our discussion is sufficient and does not require additional summary, particularly when reviewers commented on the long length of the manuscript. Based on the current condition of the concluding discussion section and concerns about overall manuscript length, we decided against lengthening the manuscript with an additional section.
Reviewer 3 Report
Dear Authors,
This is a very good paper which I support for publication as is with minor corrections.
I found the production of such long time series of land cover very interesting and well carried out. I do not have any remarks to improve the paper except for a few minor remarks in the end of this letter.
Some small remarks/ corrections:
- Line 55, remove full stop before the references;
- There is a problem with figure 9, 9b is in a different page;
- There is a problem with figure 10 which is clipped;
- The discussion section should establish some comparison with previous studies (currently has only one reference) so the reader understands better how this paper enhanced the field.
- Recommendation (optional), It would be better to shorten the paper. It is quite long making more difficult to pass the message. Maybe some figures could be put in a annex (e.g. Fig. 6)
I am looking forward to see the whole US tested with this methodology.
Good luck with the paper!
Author Response
Response to Reviewer 3 Comments
Points 1-3: Line 55, remove full stop before the references;
There is a problem with figure 9, 9b is in a different page;
There is a problem with figure 10 which is clipped
Responses 1-3: These edits have been completed.
Point 4: The discussion section should establish some comparison with previous studies (currently has only one reference) so the reader understands better how this paper enhanced the field.
Response 4: Section 3.7, Cross-model Scenario Comparison, has been added to address this.
Point 5: Recommendation (optional), It would be better to shorten the paper. It is quite long making more difficult to pass the message. Maybe some figures could be put in a annex (e.g. Fig. 6)
Response 5: Appendix A has been added and populated with tables considered supplementary.
Round 2
Reviewer 1 Report
The following imperfection should be corrected.
The following subheadings are the last rows on the page:
- 2.6.1. Scenario-based Demand (line 511);
- 3.7 Cross-model Scenario Comparison (line 897),
- References (line 980).
The caption of the Figure 5 is not on the same page of the Figure itself bur on the next page (lines 660 and 661).
Figure 3, Figure 7 and Figure 10 is divided between two pages. It is very recommended to re-arrange the text so that the figures will be entirely on the one page. The other option is to resize the figures.
Author Response
Spacing has been formatted to comply with Reviewer 1's suggestions.
