Assessing the Dynamics of Plant Species Invasion in Eastern-Mediterranean Coastal Dunes Using Cellular Automata Modeling and Satellite Time-Series Analyses
Round 1
Reviewer 1 Report
See attached.
Comments for author File: 
Comments.pdf
Author Response
Thank you very much for your suggestions. We did our best to include all your comments. We think the current version is much improved and appreciate very much your efforts in improving our manuscript.
Author Response File: Author Response.pdf
Reviewer 2 Report
Review manuscript remotesensing-1573944
Title: be more explicit on the geographic location of the study, e.g., Mediterranean. This infers that your study applies universally to all coastal dune systems which is not necessary nor is it likely defendable; An integrative approach using RS tools is a bit passive.
Abstract- remove first two sentences. These are introductory; include a brief statement describing the the coastal dune preserve and the apparent urban interface. Its clear this is a major landscape feature dictating spread. How did this coincide with urban sprawl?; How the forecast over the next 50 years was derived is unclear and unconvincing. Is there a CI of this projection? This is a fragmented ecosystem that is highly vulnerable to imposing anthropogenic factors that are likely more important than global change as mentioned. How does this influence species invasion? It appears most of the population expansion is occurring outside of the dune system. How does this impact the dune system. You might refer to the heuristic invasion curve where it appears management is effectively relegated to asset protection (i.e., perimeter exclusion)
- Introduction- Need to better develop the ecological profile of this Mediterranean dune system. remove 4th pp “increased presence…local biodiversity”; reword passive sentence “address the evaluation”; create new pp at “In this study…” and offer a little more practical understanding of CA works and is a good fit for monitoring species invasions; reword passive sentence “…we focused on evaluating…”; replace actual with observed; 1990-2100???; landscape to mean the coastal dune
- Materials and methods- 2.1 What are the plant species that make up the native community in the coastal dune system?; Fig 1 what is the color gradient in fig. 1? I assume elevation which seems in congruent to the study. Recreate map depicting urban, agriculture and natural areas according to the area designations acquired by mapi.gov. Add training and validations points for each species. Where do they match within the same pixel? What is the dash around West Bank represent? Replace dash with hatch referring to the research area; 2.2 Provide some invasion history on the species of interest. When were they introduced to Israel? What stress conditions are being referred to: drought, temperature, salinity, disease, etc..?; 2.3 In what form were spatial data provided by INPA?; 2.4 make mention that level 2 is atmospherically corrected in a way consistent with corresponding literature; What were the criteria for image selection? What was the time frequency over this 30-yr period?; what were some of the basic steps to performing image registration; 2.5 Is bare soil equal to sand dune? how was NDVI calculated, i.e., equation and bands used; provide more details on NDVI thresholds used to mask areas of interest.; rewrite pp on classification: do points fall within pixels? Is it one pixel per point? How was 70/30 distributed? What is a rejection fraction? How did you account for mixed pixels, i.e., assumption is the pixel represents a monospecific sand of the species of interest?; 2.6 It is a very big assumption that other trees and shrubs of the area won’t change or increase. Probably unrealistic.; these three scenarios between A. saligna and H. subaxillaris don’t comprehend. What is the point of this? Is this to be interpreted as a model of interspecific competition; It’s not clear what the expansion simulation produces. Some brief points to guide the reader would help with the methods description. What is the evidence of salt spray limiting revegetation and what are those limits; 2.7 Kappa stat is usually complemented with total, producer and user accuracies. Why not here?; 2.8 What are the classes that are being referred to?
- Results- 3.1 remove first sentence restating the M&M; overall accuracy or total accuracy needs to be added to the M&M; Are these accuracies derived from validation data?; Fig. 2 remove the first panel that is already displayed in Fig. 1. What is the grey classification? Are those Urban/agricultural spaces? It is well over 90% of the surface are within the area of interest. Where are the dunes?; replace a,b,c with years; Why are the panels cut off? They should display the entire AOI as depicted in Fig. 1. Add panels for 1990 and 1995; Table 1 the heading is undescriptive and needs to be written with more details; what are the samples of? Validation points? Add user and producer accuracies.; Where are data from the 1990-1995? Why aren’t species separated in Fig. 2?; Table 2 how was mixed land cover calculated. There are no methods provided; Is vacant area the same as coastal dune?; Is vacant area percent relative to total area?; vacant area reducing over time due to urban sprawl?; Fig. 3 Still not clear how mixed pixels were derived.; the legend can be simplified with three colors and solid and dash lines for class and sim based lines with descriptions in the heading; Fig. 4 I don’t understand the point of these scenarios in panel A? Please explain or consider removing; what is an increment? Replace 10-years period with decadal; None of these scenarios account for urban and agricultural sprawl; 3.2 scenario 1 and 2 these aren’t really hypotheticals in that these aren’t being tested.; Fig.5 make same changes as Fig. 2; 3.3 This section is a broad speculation with models that don’t appear to be very well developed. There needs to be strong consideration to remove this section of the paper.
- Discussion- 4.1 You need to do a more convincing analysis of this work that acknowledged the limitations of how pixels are classified. Need citations to support claims of herbaceous species more adapted to these conditions30 m resolution is considered to be medium resolution. Misidentification would also come from pixels mixed with urban hardscape and agricultural bareground.; 4.2 What do you mean by strong? How is that measured? There are no native flora listed or described in this manuscript. It appears greater impacts are derived from urban and agricultural expansion. This needs to be incorporated into he paper in order to understand available suitable habitat. 4.3 first mention of when species were introduced needs to go in the introduction. What is the heterogeneity – diversity hypothesis? Remove or rewrite the last sentence with a statement supported by empirical evidence. There is a complete lack information on the ecology of the the coastal dune system in this region to support a need to protect biodiversity.
- Conclusion- OLI is Landsat also
Author Response
We appreciate very much your comments and suggestions. We did our best to include your suggestions. Please see the attachment with the response to your suggestions.
Author Response File: Author Response.pdf
Reviewer 3 Report
Dear colleagues,
the paper is interesting and sound. I think it needs more work prior to publication. I have some comments :
1) authors must add hypotheses at the end of the introduction section.
2) authors must add something about the novelty of the research in the introduction section. Why is it interesting? and for who?
3) First section of the method: authors must provide more details on soil and vegetation cover? what kind of trees? shrubs? etc. Then, add a description of the disturbance dynamics and vegetation succession? is it fire-driven? Since when and for what reason, the invasive plants are in the country? please provide more details about autoecology of the species. I encourage the authors to add two pictures of the ecosystem (A. saligna and H. subaxillaris). for readers around the world, it will be useful.
4) Section 2.3: how exactly the spatial data had been collected? absence and presence? cover? and when? I understand that there was a field campain in 2019 but how about the spatial data? what's the gap?
5) section 2.4: why resampled at 15m? Landsat TM5 and 8 are 30m? it increases errors? why not keep 30m?
6) section 2.5: please provide more details (table for example) about the areas that cannot support the vegetation.
7) section 2.5: why extracting 'sparse/medium vegetation' and 'high'. the plant occurrence is 0 or near 0?
8) section 2.5: classification algorithms are more robust in PCI Geomatica, Envi or some open-source software. Please, justify the use of Arc Gis. Authors must also justify the use of Max likelihood classifier instead of other methods. For example, according to the pictures of the ecosystems that I see on Google Images, why not using spectral unmixing? It will give much more accurate results, for something big as A. saligna.
9) section 2.6 must be supported by references. The assumptions are sound but they need support.
10) Authors are sure that is a good idea to have 2020 as baseline? Considering climate change, extreme weather throughout the world, etc., as the gathered data since 1990, my suggestion is to take 2000 instead of 2020. The average change of 20% per 100 years is absurd as everywhere, the rate is accelerating really fast since 1990.
11) Section 2.6 third paragraph. I don't live near the sea. What's the maximum distance for salt sprays? On the shore, for sure, it has an impact. But how far? Also, what about stand disturbances? fire, humain, agriculture, deforestation? etc.
12) Section 2.6 fifth paragraph. Please justify the use of Kappa. See: Pontius Jr, R. G., & Millones, M. (2011). Death to Kappa: birth of quantity disagreement and allocation disagreement for accuracy assessment. International Journal of Remote Sensing, 32(15), 4407-4429.
13) where are the models? readers must see the tested models? authors show the results on figure 4 but we don't know anything about the models.
14) Figure 2: there are too many colors in contrast with the legend. Gray? yellow? green? what are they?
15) Table 1: why the empty row for H. subaxillaris? no data? not present in the study area or not observed? they have data in 2000 shown in Table2 and Figure 3.
16) P8 last paragraph. Please provide a new table showing the models, the AIC and the delta AIC.
17) Figure 4: the forecast is biased as they used the 2020 as baseline. Global vegetation models shown that the rate of change is exponential since 1990 with a peak around 2018-2021. How can we trust figure 4.
18) Figure 5: again, what are the color green, orange, yellow and gray?
19) Figure 6: please remove the 'both species' on the graph. I don't see the relevance.
20) P12 first paragraph: why not provide a table of omissions/commissions? it will give confidence to the readers.
21) The discussion section is redundant as many statements are already provided in the results section. e.g. beginning of second paragraph of the discussion or section 4.2 first two paragraphs.
22) P12 L 405: ability for rapid development? what is that?
23) P12 L 406: unique phenology. For exotic species, it's not unique but adaptable (optimal) phenology to their new environment. Most of the time, their phenology don't match with native species. Their flowering is earlier or during a period where the native species are dormant.
24) Section 4.2. As I understand, both plant species are pioneer. So, they will not saturate the ecosystem on the long term. Native plant will begin to establish? when? how about the life-span of other pioneer species? in areas infested with A. saligna and H. subaxillaris, do authors observed other plant species? what about their dynamics? They must add text about the natural dynamics and succession.
25) L437. ok, but what about the other species, the native ones? what about the disturbances? what about climate change? warmer and dryer climate = rapid expansion or it may slows?
26) Conclusion: action plan of invasive plant expansion? such as? if the land managers want to eradicate the alien plants? how do they proceed? manual removal, burning, steam, other plant species (more agressive), chemicals?
27) table S3: first column is weird.
Author Response
We appreciate very much your excellent comments and suggestions. We did our best to include them in the revised version. Please see the attached file with our response to your suggestion.
Author Response File: Author Response.pdf
