New 30&thinsp;m resolution Hong Kong climate, vegetation, and topography rasters indicate greater spatial variation than global grids within an urban mosaic

Morgan, Brett; Guénard, Benoit

doi:https://doi.org/10.5194/essd-11-1083-2019

Articles | Volume 11, issue 3

https://doi.org/10.5194/essd-11-1083-2019

© Author(s) 2019. This work is distributed under
the Creative Commons Attribution 4.0 License.

https://doi.org/10.5194/essd-11-1083-2019

© Author(s) 2019. This work is distributed under
the Creative Commons Attribution 4.0 License.

Articles | Volume 11, issue 3

Peer-reviewed comment

|

22 Jul 2019

Peer-reviewed comment |

| 22 Jul 2019

New 30 m resolution Hong Kong climate, vegetation, and topography rasters indicate greater spatial variation than global grids within an urban mosaic

Brett Morgan and Benoit Guénard

Download

Final revised paper (published on 22 Jul 2019)
Supplement to the final revised paper
Preprint (discussion started on 21 Dec 2018)

Interactive discussion

Status: closed

AC: Author comment | RC: Referee comment | SC: Short comment | EC: Editor comment

- Printer-friendly version

- Supplement

RC1: '20', Anonymous Referee #1, 12 Jan 2019
- AC1: 'Response to Anonymous Referee #1', Brett Morgan, 26 Mar 2019
RC2: 'Review of Morgan and Guenard: "Local models reveal greater spatial variation than global grids in an urban mosaic: Hong Kong climate, vegetation, and topography rasters"', Anonymous Referee #2, 22 Jan 2019
- AC2: 'Response to Anonymous Referee #2', Brett Morgan, 26 Mar 2019
RC3: 'Review of Morgan and Guenard', Anonymous Referee #3, 23 Jan 2019
- AC3: 'Response to Anonymous Referee #3', Brett Morgan, 26 Mar 2019
RC4: 'Review for the manuscript “Local models reveal greater spatial variation than global grids in an urban mosaic: Hong Kong climate, vegetation, and topography” by Morgan and Guenard.', Anonymous Referee #4, 25 Jan 2019
- AC4: 'Response to Anonymous Referee #4', Brett Morgan, 26 Mar 2019
AC1: 'Response to Anonymous Referee #1', Brett Morgan, 26 Mar 2019

Peer-review completion

AR: Author's response | RR: Referee report | ED: Editor decision

AR by Brett Morgan on behalf of the Authors (13 Apr 2019) Author's response

ED: Referee Nomination & Report Request started (02 May 2019) by Giulio G.R. Iovine

RR by Anonymous Referee #4 (15 May 2019)

RR by Anonymous Referee #5 (21 May 2019)

Suggestions for revision or reasons for rejection

This study presents a set of interpolated bioclimatic variables (plus other variables) specific for Hong Kong, a locale marked by dramatic ecological gradients and interanual climatic variation. The novelty of the study lies in its regional specificity and use of high-resolution DEM and satellite data to make predictions on a 30 m scale. Furthermore, the study is well executed and well written. Somewhat unsurprisingly the high-resolution dataset for this locale was different from the global dataset, with greater spatial variability and more intense extremes. Unfortunately it was not actually demonstrated that the higher resolution data produced improved species distribution models, although the introduction seemed to be setting this up and this was implied throughout (I sense a companion piece may be forthcoming). Furthermore the first half of the results/discussion section is essentially a summary of the climate and topography of HK -- OK, but what new information do the rasters produce compared to climate station summaries? The overall result is that the sections feel a little disjointed, but that may certainly be acceptable for a data paper in this journal.

Major Points:

Errors for each climate variable in units of observation (degrees, mm, etc) should be presented in a table. This would allow for some level of comparison with other datasets (although authors note how this is not truly possible). For instance, I would expect that overall error in temperature (which is fairly easy to predict) would not differ much from the global dataset (a couple degrees), but other variables such as precipitation might be more accurate.

The authors should at least consider the critique offered by Hijmans (2012) and others about the use of randomly-sampled cross-validation groups in spatial applications. Namely, control and validation points located near each other by chance may artificially reduce global error in a way that is not reflective of model performance. It wasn't clear if validation groups were spatially stratified, but this should be mentioned.

Hijmans, R. J. Cross-validation of species distribution models: removing spatial sorting bias and calibration with a null model. Ecology 93, 679–688 (2012).

Minor Points:

abstract (line 3) -- As written, sounds as if the authors are saying that 1km data are too coarse for SDM, which they are not. They are too coarse for precision applications and certain contexts.

page2 paragraph 1 -- a general SDM/review paper citation here would be helpful

Page3 line 7 -- Careful! Many 'flat deserts' have arguably more ecologically important variation at small scales, related to soil properties. I would generalize this to something like 'contexts with more gradual environmental transitions'

Page 4 line 29 -- Just a thought -- I'd be curious how a 5m resolution model would compare to the 30m resolution, and if the pattern of underestimated climatic variation would continue at this scale.

page 5 line 18 -- cite raster package

page 5 line 27 -- how was collinearity tested?

page 6 line 33 -- cite fields package

Page 7 line 17 -- any citations to back up this assertion?

Section 4.3 -- Was NDVI not included in predicting climate? If not, it doesn't really seem relevant for this paper.

Figure 5 -- I find it difficult to derive much meaningful information from this figure. I suggest it could be removed if limited for space.

Figure 6 -- Beautiful figure!

The dataset seems in order and easy to use. Aggregated climate normals from HKO should also be included so this study could be reproducible (if legally acceptable).

Hide

RR by Anonymous Referee #6 (12 Jun 2019)

Suggestions for revision or reasons for rejection

Review of the manuscript “New 30 m resolution Hong Kong climate, vegetation, and topography rasters indicate greater spatial variation than global grids within an urban mosaic” by Morgan and Guénard

General comments

The manuscript “New 30 m resolution Hong Kong climate, vegetation, and topography rasters indicate greater spatial variation than global grids within an urban mosaic” describes a high- to medium-resolution dataset of a large variety of topography, vegetation and climate rasters for the area of Hong Kong. The authors explain well the motivation and the usefulness of such a dataset emphasizing the applicability especially in Species Distribution Modeling. The selection of different variables, their elaboration and their evaluation are described in detail. While I cannot evaluate if the data manipulation was properly designed and following the standard manipulation procedures, the authors make a great effort to describe their executed procedure in detail. The vast and varied dataset along with the manuscript fit well into the scope of the journal “Earth System Science Data” and could be considered for publication after the authors address some of the comments and technical corrections.

The dataset DOI link works seamlessly and the reference to the discussion paper is provided on the dataset landing page. The authors could include a short instruction on how to cite the discussion/final paper as well as the dataset itself (consider some entries on the Pangaea repository (https://www.pangaea.de/) for nice examples). The few randomly selected datasets download and open (in two different GIS programs without any problems. The dataset names correspond to the descriptions in the discussion paper. However, on the Figshare page I was not able to locate the monthly zip files and the “readme” document (with file names, descriptions and summary statistics) that the authors describe in the “Data availability” section. The authors should upload these files on Figshare or modify the manuscript.

Specific comments

Title: For me personally the second part of the title (“indicate greater spatial variation than global grids within an urban mosaic”) is a bit redundant as it is common knowledge that finer resolution documents variation much better than a coarser resolution. In my opinion, the first part of the title perfectly describes the authors contribution and is adequate on its own. That said, I do not insist on changing the title and just provide my opinion.

P1 L2: Maybe “including” would be a more appropriate word than “particularly” in this context.

P4 L8-12: Here you basically reiterate the motivation you already explained in P3 L9-15. I suggest you remove the part on P4 or include some of the text from P4 in P3.

P6 L17-19 and Table 2: The abbreviations of the variables from Table 2 should be specified in the manuscript (for example: “maximum temperature (tmax)”) or/and in the table caption.

P6 L22-23: “When necessary, each predictor was statistically transformed to approach a normal distribution” What was the criteria you used to determine, whether it was necessary to transform a predictor?

P6 L29: The “AIC” abbreviation is not explained.

P7 L25: Maybe provide a reference for the standard equation?

Section 4.2: I have no experience in modeling climate interpolation, so I will not comment on the technical aspects and used variables, which, nevertheless, seem to be sufficiently described in Section 3.2. However, I do have some problems with understanding the climate interpolation modeling results. You describe monthly results of the climate variables, but I do not understand if this means monthly averages for a period of approx. 20 years (e.g. all the Januaries between 1998 and 2017) or monthly averages for every year (e.g. the average of a variable for January 1998). I think you refer to the first case, but in order to make the manuscript clearer, you should emphasize the considered period in parts of the manuscript and in the figures showing the results.

P8 L15: Why “Minimally”? Do you mean that 32,024 was the lowest used number of measurements for one of the variables? I suggest you rephrase the sentence, to make it clearer.

P8 L20: Why are only 3 of the 10 variables shown in Fig. 5?

P9 L29: Should you even refer to Fig. 1 in this part of the manuscript?

Section 4.2.4: When you compare your models with the WorldClim 2, you do not specify if both models cover the same (or at least a similar) time period from 1998 to 2017. Considering the changing climate, it is important to compare climates in the same time frame. If the models cover different time periods, you can still compare the models, but have to discuss the differences between them in light of the different time frame.

P11 L28-29: The sentence “Models projecting future climate scenarios would enable biodiversity change predictions, with additional variables like cloud cover and solar radiation useful.” is incomprehensible. Probably the comma and “useful” are remnants from a previous version of the manuscript?

References: As a reader I would prefer to have DOI’s (where available) included in the reference list. However, I do not know if DOI inclusion is obligatory in Earth System Science Data.

Table 2: It should be clearly indicated in the table and/or in the table caption, that SD is standard deviation. Additionally, from what was the ratio in this table calculated? From the description in the manuscript I suppose it is the ratio of the standard deviations of the two rasters, however the calculations are off (for example, in the first line: 1/0.5=2 and not 1.9)

Table 2 Caption: The last sentence “Increased standard deviation ranges from 1.4x to 3.4x.” is not very clear, as it is not explained which standard deviation the authors have in mind. Additionally, standard deviation values in Table 2 are much larger than 3.4. If the authors refer to the ratio, they should firstly calculate it again (see previous comment).

Figure S1: I really like this figure as it showcases the whole extent of the work the authors have done. In order to emphasize the several unprecedented datasets that the authors created, they could visually discriminate (by color maybe) between the datasets from other sources and the datasets the authors created themselves (for example: Elevation (30 m) vs. all the Relative Elevations)

Technical corrections

P1 L2: “However, these data, …” instead of “However these data, …”

P3 L11: “… Kong in …” instead of “… Kong, in …”

P5 L18: “… the raster package …” should probably be “… the R raster package …”?

P5 L19: Probably “Secondly, …” instead of “Second, …”

P5 L19-20: The sentence “Second, water proximity … surrounding a given pixel” is a bit difficult to read and understand. As they continue in the manuscript by using the term radius, they could maybe write something in the line of “… as the percent of land surface within a radius of a given pixel.”

P6 L3: Perhaps “… vegetated areas have a value of 0.” Instead of “… vegetated areas are 0.”

P6 L8: Perhaps “… linear regressions …” instead of “… linear regression …”

P6 L32: Probably “Firstly, …” instead of “First, …”

P6 L32: Probably “Secondly, …” instead of “Second, …”

P8 L23: “6° C” instead “6°C”

P8 L24: “> 900 m, < 18° C” instead of “>900 m, <18°C”; “> 24° C” instead of “>24°C”

P9 L4, L6, L10: “> 2500” instead of “>2500”; “< 1600” instead of “<1600”; “52 %” instead of “52%”

P9 L15, L16, L17: “15.5° C” instead of “15.5°C”; “19° C” instead of “19°C”; “90 %” instead of “90%”; “75 %” instead of “75%”

P9 L31: two times “2° C” instead of “2°C”

P10 L 13: “… exception are the … forests at …” instead of “… exception is the … forests, at …”

P10 L20, L26: Probably “Firstly, …” instead of “First, …”; probably “Secondly, …” instead of “Second, …”; “… array of rasters …” instead of “… array rasters …”

P11 L4: “… presented data …” instead of “… data presented …”; probably “Firstly, …” instead of “First, …”

P11 L9, L10: Probably “Secondly, …” instead of “Second, …”; “For example …” instead of “Forexample …”

P12 L5: “… foreseeable …” instead of “… forseeable …”

P12 L10: “Morgan and Guénard …” instead of “Morgan Guénard …”

Fig. 7 caption: “average low temperature” is probably “average temperature”?

Hide

ED: Publish subject to minor revisions (review by editor) (13 Jun 2019) by Giulio G.R. Iovine

AR by Brett Morgan on behalf of the Authors (22 Jun 2019) Author's response Manuscript

ED: Publish as is (02 Jul 2019) by Giulio G.R. Iovine

Short summary

Hong Kong is poised to become a model region for understanding the effects of urbanization, biotic invasions, and protected areas in the tropics. However, until now there have been few suitable GIS layers to address these issues on a landscape scale. This set of 30 m resolution vegetation, topography, and interpolated climate rasters will enable a new generation of spatial studies in Hong Kong. Compared to global datasets, these local models consistently indicate greater climatic heterogeneity.