Remote Sensing Extraction of Agricultural Land in Shandong Province, China, from 2016 to 2020 Based on Google Earth Engine

Round 1

Reviewer 1 Report

Title: Remote sensing extraction of agricultural land in Shandong Province, China from 2016 to 2020 based on Google Earth Engine

 

Overview: This study use Google Earth Engine and Landsat data (TM, ETM+, OLI- which are not actually mentioned- just the satellites Landsat 5, 7, 8) to map agricultural areas in Shandong Province, China. It could use more details about the study area (what crops are being grown there?) and clearer explanation of the methods. More comments below.

 

Other comments:

Line 29 Change “thema2ic” to “thematic”

Line 43 Should this read “more than ten million hectares”?

Lines 56, 61, 65, 73, 207 have coding errors in the citations

Line 70 Should that word be “researchers”?

Line 76 I would use “supervised” rather than “supervising”

Line 78 What is meant by “7th period”?

Line 95 “e” for the longitude should be “E”

Lines 98-99 are confusing

Line 117 I would put a space between “Landsat” and “5”

Line 125 I would add “of” between “calculation” and “NDVI”

Figure 2 Why are these curves called “exponential”?

So far there is no mention of what specific crops are grown in the region- it would be helpful to include this information.

Table 1 There is likely a wide variation among agricultural land types- how was this handled?

Figure 3 needs a more detailed caption

Section 3.1.1. Wouldn’t keeping all original bands and adding the SVIs lead to a lot of redundancy in the inputs?

Table 2 Use “Description” rather than “Describe”

Line 194 I believe that “hillshade” is the name used in ESRI products and “shaded relief” would be a general term to use. This also seems like it would have redundancies with other DEM data.

Table 3 How were these “optimized features” selected or generated?

Section 4 “Results and Analysis” seems an odd section title. Shouldn’t analysis be in the methods? Or do you mean “Results and Discussion”?

Figure 4 Many of the features seem to have similar importance values. Why 15?

Figure 7 The colors in the legend don’t seem to quite match those of the map.

Line 326 How were the areas reported in the Statistical Handbook obtained?

Line 352 You can also spell it “mosaicking”

Author Response

We gratefully thank editor and all reviewers for their time spend making their constructive remarks and useful suggestions, which has significantly raised the quality of the manuscript and has enables us to improve the manuscript. Each suggested revision and comment, brought forward by the reviewers was accurately incorporated and considered. Below the comments of the reviewers are response point by point and the revisions are indicated. And the full text has been revised carefully for another time.The detailed modification information is in the PDF file we upload to you. Thank you very much for your valuable advice.

Author Response File: Author Response.pdf

Reviewer 2 Report

This manuscript reports on the use of GEE platform for accessing Landsat time series image data as well as machine learning algorithm to process these images to obtain the spatial variation distribution information of agricultural land in Shandong Province from 2016 to 2020.

The significance of this manuscript is in the comparison of three different ML algorithms against an Ensemble Learning. It also reflect on the authors experience regarding data pre-processing and optimization of features used in the classification.

The manuscript, however, does not add noteworthy new knowledge on the topic. As such, in my opinion it is of a limited originality, significance of content, and interest to the readers.

 

Missing References on Lines 56, 61, 65, 66, 73, 207

 

Weak Sentences; re-write:

Line 18-19

Lines 125-129

 

Define abbreviation before 1^st use:

Line 20: GEE

Line 54: API

Line 112: SRTM

Line 162: CART

Line 174: EVI

 

Recommendations/Corrections to text:

Line 29: correct “thema2tic”

Line 42: include rough equivalent in US$.

Line 61: Start new paragraph.

Lines 125-126: include definitions/equations for NDVI, LSWI, NDBI, and NDWI.

Line 126: insert “and” before “NDWI”.

Line 134: vague word “exponential”.

Line 166: Figure 3; add arrows indicating directions for the roles of “Sample Data” and “Topographic Data”.

Line 167: change “route” to “flowchart” or “work flow”.

Line 169: change “have” to “has”.

Line 173: vague word “phase”

Line 202: Table 3, change “Nir” to NIR”.

Line 320: Table 5; replace “Extraction area” by “Area Extracted by Remote Sensing” or similar meaning.

Lines 326-328: repeated from previous section; delete.

Line 375: change “accuracy” to “resolution”.

Author Response

We gratefully thank editor and all reviewers for their time spend making their constructive remarks and useful suggestions, which has significantly raised the quality of the manuscript and has enables us to improve the manuscript. Each suggested revision and comment, brought forward by the reviewers was accurately incorporated and considered. Below the comments of the reviewers are response point by point and the revisions are indicated. And the full text has been revised carefully for another time.The detailed modification information is in the PDF file we upload to you. Thank you very much for your valuable advice.

Author Response File: Author Response.pdf

Reviewer 3 Report

The manuscript is well and detailed written and I do not any further correction. It can be accepted in its current form.

Author Response

We gratefully thank editor and all reviewers for their time spend making their constructive remarks and useful suggestions, which has significantly raised the quality of the manuscript and has enables us to improve the manuscript. Each suggested revision and comment, brought forward by the reviewers was accurately incorporated and considered. Below the comments of the reviewers are response point by point and the revisions are indicated. And the full text has been revised carefully for another time.Thank you very much for your valuable advice.

Reviewer 4 Report

The subject is about using existing tools to compare different implemented methods in image classifications so that to extract agricultural land. It is between research and engineering. The work may not be original, though the article is clearly presented and well-written. The analysis and conclusion could be pushed further, in particular with regard to misclassification between items and to the drivers. Some details following the sections are given as follows.

 

Abstract:

-Precise the acronym GEE at first mention

-“thema2tic”?

-Explicit the goal of comparing four classifier algorithms

-“five agricultural land extraction”: indicate this goal of extracting five agricultural types before the description of the results.

 

Introduction: the goals and the state of the art are clearly presented. The novelty of the research should be more explicit, for instance about the lack in the existing approaches and about the objects of research (land use or land cover?).

-Several missing links “[Error! Reference source not found.]. Processing”

-“extraction of cultivated land, crops and land cover”: which other land covers (urban land covers, unbuilt land)?

-Mention of the notion of land cover and then of land use: they are distinct notions that should be detailed and to adequately mentioned. Does “land change” mean land cover or land use changes or mixed items of land use land cover? The targeted item “agricultural land” is a land use.

 

Research area and data preprocessing: acronyms should be explicit. References could be added regarding used algorithms.

-The sample point data are to be specified at first mention (it could be LiDAR points): they are ground-truth data (even if classified manually at distance and not in situ)?

-Landsat 1C for collection 1?

-Reference to add for CFMASK algorithm

-QA to be clarified at first mention

-“The main surface features in the study area can be divided into five categories”: choice for this article, the items could be justified (spatial extent, spatial and semantic consistency between items)

-Sample point data: are they point geometry to be extrapolated to polygon geometry and to be compared to the polygons resulting from classifications?

 

Methods: the technical flow chart for describing the method is useful. Random forest and the other algorithms are clearly explained. Qualification of GEE tools and functions could be useful and added to the article with references in literature (for instance for the algorithms used for topographic features extraction, or index computation).

-CART = Decision tree, not clearly written

-“with the support of GEE platform”: economic model of GEE? (Subscription, trial account for teaching and research...). And which implementation (call a server)?

-“machine learning methods such as”: list the methods used and not “such as” which seems like a list of examples.

-Table 2: solid line to separate the 2 parts of the list

-Specify in part 3.3 about classification methods that the described algorithms are to be compared

 

Results and Analysis: the comparison between algorithms is interesting though it could be clearer, for instance about misclassification between items.

-Figures 5 and 6 are useful to describe the results. In Figure 6, avoid overlapping bars in the graph.

-Specify cartographic accuracy and user accuracy (in comparison with the producer/map-maker accuracy?)

-Same type of errors between the 4 classifiers? Same misclassifications between items of the nomenclature among the classifiers?

-Same questions about land use and land cover in “the change of agricultural land cover”: agricultural land use can change cover (e.g. state of growth of vegetation, change of culture type) though it is not the goal of the work?

-Table 5: underestimation of agricultural land extracted compared to the numbers in Statistical Yearbook, explanations?

-urbanization: definition used in this article, sources of the data?

-“artificial surface consists mainly of buildings”: and of impervious surfaces such as parking lots, roads?

 

Conclusion: the conclusion about the GEE platform should be more supported by the previously presented results. The conclusion about better results by ensemble learning were validated by manual sample data / user accuracy?

Author Response

We gratefully thank editor and all reviewers for their time spend making their constructive remarks and useful suggestions, which has significantly raised the quality of the manuscript and has enables us to improve the manuscript. Each suggested revision and comment, brought forward by the reviewers was accurately incorporated and considered. Below the comments of the reviewers are response point by point and the revisions are indicated. And the full text has been revised carefully for another time.The detailed modification information is in the PDF file we upload to you. Thank you very much for your valuable advice.

Author Response File: Author Response.pdf

Round 2

Reviewer 1 Report

Study area has better details now, but methods could still use clarification in general. And proofreading should be done to find spacing, spelling, grammatical errors.

 

Line 20 Add space before “(Google…”

Lines 87 Change to something like “there is little research on…” (not “few researches”)

Line 97 Add space after “studies.”

Table 1 What are the second numbers in parentheses?

Table 6 Over all should be too words or at least spelled correctly (overall)

Author Response

We gratefully thank editor and all reviewers for their time spend making their constructive remarks and useful suggestions, which has significantly raised the quality of the manuscript and has enables us to improve the manuscript. Each suggested revision and comment, brought forward by the reviewers was accurately incorporated and considered. Below the comments of the reviewers are response point by point and the revisions are indicated. And the full text has been revised carefully for another time.

We have uploaded our modification instructions to you in PDF format. Thanks again for your comments!

Author Response File: Author Response.pdf