Spatiotemporal Evolution and Impact Mechanisms of Areca Palm Plantations in China (1987–2022)

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

Please find the attached file.

Comments for author File: Comments.pdf

Author Response

Dear Reviewer 1,

Thank you very much for your time and effort reviewing this manuscript. We have responded to the comments below. We believe the edits have resulted in an improved manuscript and our response to all of the comments is provided below. In addition, we attach great importance to the advice given by reviewer and feel guilty about the mistakes in previous articles! We have marked your reply with a yellow background color.

Yours sincerely,

(on behalf of all co-authors)

1 .Reviewer: I suggest to change the title! Specially “First-ever understanding…”.

Authors: We followed the opinions of the reviewer. The original title " First-Ever understanding the spatiotemporal evolution and impact mechanisms of areca palm plantations in China from 1987 to 2022" was changed to " Spatiotemporal Evolution and Impact Mechanisms of Areca Palm Plantations in China (1987–2022)". (See Page 1, Line 2-3). The contents are as follows:

… (This part is the paragraph of the text)

Spatiotemporal Evolution and Impact Mechanisms of Areca Palm Plantations in China.

2. Reviewer: Use full phare at the first time in the paper, then abbreviation can be used in the paper.

Authors: Your question is very helpful to us, which is also our concern. We have revised the paper to ensure that the full term is used the first time it appears, followed by its abbreviation in subsequent mentions. (See Page 12, Line 387-390). The contents are as follows:

… (This part is the paragraph of the text)

including Baisha, Baoting, Changjiang, Chengmai, Danzhou, Dingan, Dongfan, Haikou, Ledong Lingshui, Lingao, Qionghai, Qiongzhong, Sanya, Tunchang, Wanning, Wenchang, and Wuzhishan, as shown in (Fig. 6).

.3. Reviewer: Revise paper carefully; some parts need re-writing.

Authors: We appreciate your feedback and have carefully revised the paper as suggested. We have re-written the discussion sections you highlighted to improve clarity and address any issues. We believe these revisions enhance the overall quality of the manuscript. (See Page 19-21, Line 499-596). The contents are as follows:

… (This part is the paragraph of the text)

4. Discussion

4.1. Historical distribution of the areca palm

Throughout the areca palm mapping process, three independent survey datasets were used to map its historical distributions. High-resolution images were processed using the Res50_U-net model to map the areca palm distribution for 2022 on Hainan Island, achieving an overall accuracy of 90%. This accuracy coincides with that of Luo’s machine learning-based areca palm classification method which achieved as accuracy of 89% [58]. However, the areca palm mapping in this study was oriented to a wider range of scales. Previous research has not evaluated the spatial distribution of the areca palm on such a large scale. This study has thus demonstrated the feasibility of utilizing deep learning for large-scale areca palm classifications using high-resolution imagery. Building upon this, the study has expanded a 1 km buffer zone to generate a potential areca palm distribution area, which was used to create a historical map of the distribu-tions on Hainan Island. Using Landsat imagery and a Random Forest algorithm with constraints on the potential distribution area, the overall accuracy for 2022 was 67%. These results indicate that the model has an adequate ability to identify areca palm dis-tributions. Historical areca palm distribution data was also examined and found to aligns well with the expected growth patterns during cultivation. This study has thus successfully depicted the historical evolution of areca palm cultivation, producing, for the first time, providing a large-scale, high-resolution spatial distribution map over an extended period. This study has also introduced a novel approach for classifying dis-persed and patchy crops. However, it should be noted that the addition of the 1 km potential distribution range will lead to some areca palm not included in the study area, resulting in omission, which is also an aspect that needs to be improved in the future, how to obtain a more refined potential distribution of betel nut through more effective means. In addition, the accuracy of the areca palm distribution data before 2022 cannot be guaranteed due to the lack of field data classification accuracy, which means that more continuous field data will be needed in the future.

4.2. Spatiotemporal evolution of areca palm plantation on Hainan Island

From 1987 to 2022, areca palm plantations on Hainan Island exhibited a relatively clustered pattern. This clustering is influenced by both the actual cultivation practices and the limitations of the areca palm's potential distribution range. While some overes-timation of clustering might occur due to these distribution range limitations, the pattern still reflects the true extent of areca palm plantations. This study identified two distinct areca palm plantation patterns on Hainan Island. The first is a high-value clus-ter in the southeastern region, which has shown an expansion trend over time. This pattern is most prominent in cities like Wanning and Qionghai. Wanning is renowned as the areca palm hub of China, while Qionghai has seen significant growth in areca palm cultivation in recent years, making it a key driver of local economic development. The high-value clustering enhances scale and efficiency, thus promoting the growth of the areca palm industry.

Conversely, a low-value clustering pattern is found in the southwestern and northeastern regions of Hainan Island. This pattern, particularly evident in areas like Haikou, Danzhou, and Changjiang, has not shown any signs of expansion. These re-gions do not rely heavily on the areca palm industry, which is not a major economic pillar for them. These two spatial distribution patterns highlight the significant differ-ences in the spatial distribution and plantation practices of areca palms on Hainan Is-land, creating a clear east-west divide. The differences are influenced by varying levels of governmental attention, policies, and economic incentives directed towards the areca palm industry. However, the overall trend for the areca palm industry is one of rapid expansion in plantation areas, with rising prices playing a significant role in this growth. This has, however, led to rigid cultivation practices, limited the integration of technology, and fostered exploitative management. Therefore, it is necessary for the government to regulate the expansion of the areca palm industry and prevent farmers from blindly following cultivation trends. For instance, oil palm plantations in Malaysia pose a threat to peatlands, which are home to a rich diversity of flora and fauna [59]. Instead, there should be a focus on promoting the integrated application of plantation technologies to accelerate the upgrading of the areca palm industry, realize the three-dimensional comprehensive development and utilization of areca palm, and im-prove the output value of the unit area as well as the benefits of the vacated land for the development of the understory economy.

4.3. Mechanisms influencing the evolution of Areca palm plantation

From 1987 to 2022, the areca palm cultivation area on Hainan Island expanded dramatically, increasing from 8,064 ha to 193,328 ha, over 20 times. This growth in areca palm plantations strongly correlates with Hainan Island’s GDP, underscoring the crop's significant economic role. However, this correlation varies across regions. For example, Sanya shows no significant correlation, as its GDP is primarily driven by tourism rather than agriculture, and it cultivates a diverse array of crops like mangoes, coconuts, and rice alongside areca palms. Similarly, Danzhou, with rubber as its primary crop, does not show a significant correlation due to low rubber prices, despite a growing interest in switching to areca palms. In contrast, regions like Wanning and Qionghai exhibit strong correlations, reflecting the flourishing areca palm-related industries, especially in the past five years.

Field investigations revealed various areca palm cultivation types in Hainan, in-cluding plantations near farmlands, within natural tropical rainforests, in backyards, on flatlands, and in hilly and mountainous areas. The dispersed and fragmented nature of these plantations, sometimes encroaching on Hainan Tropical Rainforest National Park, raises environmental concerns, including biodiversity loss and reduced carbon seques-tration capacity. Additionally, the expansion of areca palm plantations at the expense of traditional agricultural land poses risks to ecological balance, as noted in the work of Takeuchi, where monoculture and forest destruction can lead to ecosystem degrada-tion and rural population outflow [60]. The areca plam trade often results in a series of associated infrastructures, including trading markets and other facilities, and these de-velopments are frequent in urban areas. Cities and towns have developed transporta-tion systems and convenient trading environments. The development of the areca palm is also a microcosm of economic evolution, as the focus shifts from rural to urban areas. Indeed, the areca palm is a high-value cash crop that can be lucrative for farmers; however, achieving large-scale planting and efficient management while ensuring eco-logical sustainability and a harmonious relationship between humans and nature are significant challenges for Hainan Island and the rest of the world. Srinivasan’s work provides an example with India’s oil palm. However, the current expansion of  oil palm cultivation in India comes at the cost of biodiversity-rich landscapes. But their model suggests that, on a national scale, India seems to have viable options to meet its projected palm oil demand without compromising its biodiversity or food security [61]. Can the planting of Areca palm in Hainan meet the needs of the forest through macro policy regulation and control under the consideration of the existing planting concen-tration. On a more precise spatial scale, areca palm cultivation needs to take into ac-count local climatic conditions, biodiversity, local agricultural input ratios, and trade-offs between economy and social security. China's policy decisions on areca palm have largely mitigated the current set of problems facing China's tropical rainforests.

4. Reviewer: In the abstract, the proposed method, its accuracy and data aren’t introduced!

 Authors: We appreciate you pointing out the problems in the abstract. We have revised the abstract to include a concise introduction of the proposed method, its accuracy, and the data used in our study. These additions should provide a more comprehensive overview of our research. (See Page 1, Line 19-21). The contents are as follows:

… (This part is the paragraph of the text)

Using Landsat and Google Earth imagery combined with machine learning, the geographical distribution of areca palm was mapped at a 30-meter resolution from 1987 to 2022, achieving an overall classification accuracy of 0.67 in 2022.

5. Reviewer: L 84: ref [30] doesn’t use Landsat imagery!

Authors: Your comments would be very helpful to us. We removed the article and re-added crop studies with similar questions using Landsat data. (See Page 2, Line 95-96). The contents are as follows:

… (This part is the paragraph of the text)

Asgarian, A.; Soffianian, A.; Pourmanafi, S. Crop type mapping in a highly fragmented and heterogeneous agricultural landscape: A case of central Iran using multi-temporal Landsat 8 imagery. Computers and Electronics in Agriculture 2016, 127, 531-540, doi:https://doi.org/10.1016/j.compag.2016.07.019.

 6. Reviewer: Which image resolution is appropriate for scattered planting pattern?

Authors: We appreciate your understanding of these practical considerations. We're sorry that we can't clearly explain why we use data in this way, and we've added a section to explain this. (See Page 2-3, Line 95-107). The contents are as follows:

… (This part is the paragraph of the text)

Landsat data generally has poorer recognition ability for crops with scattered planting patterns due to its resolution limitations[35]. However, if the potential range and geographic distribution information of the crops can be obtained, it is still possible to achieve good classification results by leveraging this prior knowledge. This is because known prior knowledge: such as slope, distribution area, potential range of distribution that can help guide and refine the classification algorithm to more accurately identify target areas, compensating for the limitations of the satellite data resolution [36]. The potential range limitation of the distribution of classification results is an effective method of spatial constraint. In other crop mapping, one category of existing land classification data is often used as the potential distribution range of the crop. However, the actual planting situation of areca palm is very complicated, and the existing land classification data does not meet the requirements as its potential distribution range

Jin, Y.; Guo, J.; Ye, H.; Zhao, J.; Huang, W.; Cui, B. Extraction of Arecanut Planting Distribution Based on the Feature Space Optimization of PlanetScope Imagery. Agriculture 2021, 11, doi:10.3390/agriculture11040371.

Gumma, M.K.; Nelson, A.; Thenkabail, P.S.; Singh, A.N. Mapping rice areas of South Asia using MODIS multitemporal data. Journal of Applied Remote Sensing 2011, 5, doi:10.1117/1.3619838

 

7. Reviewer: Literature review is weak. More crop mapping study based on deep learning should be added, especially for palm plantations.

Authors: Thanks for your advice. We have added the relevant study in literature review. (See Page 2, Line 77-87). The contents are as follows:

… (This part is the paragraph of the text)

For instance, Freudenberg et al. proposed a neural network of the U-Net to detect oil and coconut palms on high resolution satellite images with accuracies between 89% and 92% [26]. Li used a number of manually interpreted samples to train and optimized the convolutional neural network (CNN) , and can be dectected 96% of the oil palm trees using the QuickBird images compared with the manually interpreted ground truth [27]. Cheng combined PALSAR-2 with maximum likelihood classifier (MlC) to map the oil palm with 1 km meter resolution in Malaysia, which is the closest to the official MPOB inventories (~8.87% overestimation)[28]. Lee tested two classification al-gorithms Classification and Regression Trees (CART) and Random Forests (RF) and evaluated various band combinations for extracting oil palm in India using Landsat 8 imagery [29]

Freudenberg, M.; Nölke, N.; Agostini, A.; Urban, K.; Wörgötter, F.; Kleinn, C. Large Scale Palm Tree Detection in High Resolution Satellite Images Using U-Net. Remote Sensing 2019, 11, 312

Li, W.; Fu, H.; Yu, L.; Cracknell, A. Deep Learning Based Oil Palm Tree Detection and Counting for High-Resolution Remote Sensing Images. Remote Sensing 2017, 9, 22

Cheng, Y.; Yu, L.; Xu, Y.; Lu, H.; Cracknell, A.P.; Kanniah, K.; Gong, P. Mapping oil palm extent in Malaysia using ALOS-2 PALSAR-2 data. International Journal of Remote Sensing 2018, 39, 432-452, doi:10.1080/01431161.2017.1387309.

Lee, J.S.H.; Wich, S.; Widayati, A.; Koh, L.P. Detecting industrial oil palm plantations on Landsat images with Google Earth Engine. Remote Sensing Applications: Society and Environment 2016, 4, 219-224, doi:https://doi.org/10.1016/j.rsase.2016.11.003

 

8. Reviewer: Move Figure 1 to second section. Authors have not described the figure!

Authors: We understand the reviewer's concern about the placement of Figure 1. However, we believe that keeping Figure 1 in its current position in the manuscript enhances the flow and context for the readers. In the revised manuscript, we have provided a more detailed description of Figure 1 in the introduction section to ensure that its relevance and significance are clearly communicated. The updated description highlights the key elements of the figure and their relationship to the study's objectives. (See Page 3, Line 125-129). The contents are as follows:

… (This part is the paragraph of the text)

This study has aimed to collaboratively map the spatial distribution of the areca palm in the Hainan region of China across a large-scale and long-term with high- and medium-resolution imagery, as well as machine learning and deep-learning techniques, and has also investigated the spatiotemporal dynamics of the areca palm and the in-fluencing mechanisms (Fig. 1).

 

9. Reviewer: L 132: just high-resolution images are used? Merge sections 2.1.2 and 2.1.4 to describe datasets.

Authors: Thank you for your feedback. We have merged sections 2.1.2 and 2.1.4 to create a more streamlined and comprehensive description of the datasets. (See Page 5, Line 158-183). The contents are as follows:

 … (This part is the paragraph of the text)

2.1.2. Image data sources

Digital orthophotos were sourced from QuickBird-2, Geoeye-1, WorldView-2, and WorldView-3, which contain only RGB three-channel information. They were download from the Google Historical Imagery server, with a resolution of 0.59 m, and ranging from November 2021 to February 2023. For the selection of digital orthophotos, images with minimal cloud cover, consistent overall color tones, and clear imaging were selected. For ease of data storage, they were cropped into 256 × 256-pixel tiles, and there was total of 2,610,000 tiles.

Landsat Surface Reflectance (SR) data, with a spatial resolution of 30 m, were downloaded from the Google Earth Engine (GEE) platform. The Landsat 5 SR product have been processed using Landsat ecosystem disturbance adaptive processing system (LEDAPS) [46] on GEE platform. The steps included Geometric Correction, Radiometric Correction, Atmospheric Correction and topographic correction. The Landsat 8 SR product are created with the Land Surface Reflectance Code (LaSRC) [47]. The image selection criteria included cloud cover of less than 20% and high imaging quality. Cloud removal was performed using the QA-PIXEL band. To ensure consistency in the extraction of bands across each of the years for which the Landsat acquisition data was obtained, six common bands were shared by the Landsat satellite, namely, blue, green, near-infrared (NIR), short-wave infrared 1 (SWIR1), and short-wave infrared 2 (SWIR2), and they were extracted and composited using an annual median synthesis method. The study area encompasses a large area, and in practical operations, the results of the annual compositing often suffer from data gaps due to cloud cover. To address this problem, a data supplementation approach was adopted by selecting images from adjacent years with low cloud cover for the median composition. This ultimately helped to reduce data loss. The information of Image composition time, satellite sensor and selected bands used in this study show in Table 2.

10. Reviewer: Image with high and medium resolution which are used in this study are applied with different spectral properties?

Authors: Thank you for your insightful question. In our study, the high-resolution images are based on RGB three-channel information (See Page 5, Line 159-160). For the medium-resolution images, we utilized the six spectral bands that are common to both Landsat 5 and Landsat 8: Blue, Green, Red, Near-Infrared (NIR), Shortwave Infrared 1 (SWIR1), and Shortwave Infrared 2 (SWIR2) (See Page 5, Line 173-177). We appreciate your attention to this detail and hope this clarifies our approach.

 … (This part is the paragraph of the text)

Digital orthophotos were sourced from QuickBird-2, Geoeye-1, WorldView-2, and WorldView-3, which contain only RGB three-channel information.

To ensure consistency in the extraction of bands across each of the years for which the Landsat acquisition data was obtained, six common bands were shared by the Landsat satellite, namely, blue, green, near-infrared (NIR), short-wave infrared 1 (SWIR1), and short-wave infrared 2 (SWIR2), and they were extracted and composited using an an-nual median synthesis method.

11. Reviewer: Why Res50_U-net is applied? Explain the advantage of this network.

Authors: ResNet is a widely used deep learning feature extraction module, where deeper networks theoretically offer stronger learning capabilities but generally require more memory. Given these considerations, we chose ResNet50 due to its balance between performance and memory usage (See Page 7, Line 243-246). Additionally, U-net is a highly reliable network architecture for land cover classification. Combining ResNet with U-net to form Res50_U-net enhances performance, leveraging the strengths of both architectures for improved results. We have included this explanation in the relevant sections to help readers understand the advantages of this network configuration (See Page 8, Line 261-265). The contents are as follows:

… (This part is the paragraph of the text)

Depth is generally considered a crucial factor for improving the neural network per-formance, where deeper networks theoretically offer stronger learning capabilities but generally require more memory. Given these considerations, chose ResNet50 due to its balance between performance and memory usage.

U-Net model surpasses conventional image classification methods in accurately seg-menting different crop types [54] Combining ResNet 50 with U-net to form Res50_U-net enhances performance, leveraging the strengths of both architectures to ensure areca palm classification results. The structure of the Res50_U-net model is illustrated in (Fig. 5).

12. Reviewer: Move Line 243 to 253 to the third section.

Authors: Thank you for your suggestion. We have moved the content from Line 243 to 253 to the third section as recommended. 

 13. Reviewer: Why two different methods (DNN and RF) are applied for high and medium resolution image classification?

Authors: Thank you for your question. We applied two different methods, DNN and RF, based on the characteristics and data requirements of high and medium-resolution images. DNN requires a large amount of training data, which high-resolution images can provide, making it well-suited for such data and resulting in higher accuracy. On the other hand, while RF can be applied to high-resolution images, it has limitations. Specifically, RF requires a considerable amount of manual effort to accurately classify non-target land cover types, which becomes increasingly complex with high-resolution data. Additionally, the accuracy of RF tends to be lower than that of DNN when dealing with the detailed information provided by high-resolution imagery. For medium-resolution images, where the available data may not be sufficient to train a DNN effectively, RF is more suitable, as it can work effectively with smaller datasets and is more practical for medium-resolution imagery.

 14. Reviewer: How do you split the ground truth into the training, validation and test data? Define their percentage!

Authors: We apologize for the oversight. In the data collection process, both the first and second steps focused on gathering ground truth data. The first set of 500 areca palm sample plots was used for training the deep learning model, with the data split into 70% for training and 30% for validation. The second set of 500 areca palm sample plots was reserved exclusively for testing the model's classification performance. We have made revisions in the relevant sections to ensure that readers can better understand this process. (See Page 6, Line 194-197). The contents are as follows:

 … (This part is the paragraph of the text)

The first set of 500 areca palm sample plots was used for training the deep learning model, with the data split into 70% for training and 30% for validation. The second set of 500 areca palm sample plots was reserved exclusively for testing the model's classification performance.

15. Reviewer: Merge tables 5 and 7; remove tables 6 and 8. Extract more criteria from confusion matrix.

Authors: We are grateful for your suggestions. We have merged Tables 5 and 7 as recommended and removed Tables 6 and 8. Additionally, we have extracted the F1-Score from the confusion matrix to enhance the analysis (See Page 13, Line 412). and provided the corresponding formula. F1-Score analysis has been included in the evaluation of the RF results derived from the confusion matrix (See Page 14, Line 432), and we have also incorporated F1-Score analysis in all relevant sections of the manuscript where accuracy is discussed. the contents are as follows:

… (This part is the paragraph of the text)

 Table 5. Classification result of Res50_U-net.

Internal evaluation	Areca	Non-Areca
Precision	0.85	0.98
Recall	0.82	0.97
Dice	0.81	0.97
External evaluation
User’s Accuracy	0.87	0.93
Producer’s Accuracy	0.92	0.88
F1-Score	0.89	0.90
Overall Accuracy	0.90
Kappa coefficient	0.80

 

Table 6. Classification result of random forest.

	User’s Accuracy	Producer’s Accuracy	F1-Score	Overall Accuracy	Kappa Coefficient
Areca	0.68	0.65	0.66	0.67	0.34
Non-Areca	0.66	0.69	0.67

 16. Reviewer: In order to prove the efficiency of the proposed methods, compare proposed U-net with other DNN methods.

Authors: Your suggestion reminds us that we will correct it. We have added a reference to crop classification study that compares U-net with other classification algorithms, which supports the effectiveness and feasibility of using U-net in our research. (See Page 7, Line 261-265) The contents are as follows:

… (This part is the paragraph of the text)

Ayushi; Buttar, P.K. Satellite Imagery Analysis for Crop Type Segmentation Using U-Net Architecture. Procedia Computer Science 2024, 235, 3418-3427, doi:https://doi.org/10.1016/j.procs.2024.04.322.

Author Response File: Author Response.pdf

Reviewer 2 Report

Comments and Suggestions for Authors

The authors have brought out a very pertinent issue with scientific facts. The article is well written, however the following points may be incorporated to improve the article

a) The map of the study area will be useful for the readers

b) In the ground survey section (2.1.3) authors may add more details about the sampling strategy. How the survey points are chosen/decided, tble may be presented to show how many points are chosen for survey in each city?, How many points are surveyed vs area of the of the site/location, why 1500 points are chosen, how 100 points are arrived at? why 500 points are used for classification performance.

c) Table 2 is a generalized information readily available to all!Authous may add little more information of the radiometric calibration, atmospheric corrections and geometric corrections. why its is need and what approach has been adopted.

d) section 2.2.1 on mapping with deep learning- authors may explain what it means ' high resolution image is richer and greater data magnitude'? line 197 on data redundancy needs to be discussed.

e) line no. 410 what it means by 'area size'? is not clear

f) line no. 442 ' precipitation changes are relatively complex' is not clear to the reader, needs to be discussed properly.

g) the discussion is very long. There is scope to improve the section as per the results

h) in section 2.1.3 ' a tea of 20 individuals' is not necessary to be mentioned

Comments on the Quality of English Language

English language corrections are suggested in few sections:

line no.141'was organized' may be checked

line no. 199 'by which to automatically' needs correction

Similar correction may be checked in the article carefully.

 

 

Author Response

Dear Reviewer 2,

Thank you very much for your time and effort reviewing this manuscript. We have responded to the comments below. We believe the edits have resulted in an improved manuscript and our response to all of the comments is provided below. In addition, we attach great importance to the advice given by reviewer and feel guilty about the mistakes in previous articles! We have marked your reply with a green background color.

Yours sincerely,

(on behalf of all co-authors)

1. Reviewer: The map of the study area will be useful for the readers.

Authors: Your suggestion is very good. We have added the map of the study area which will help readers to understand our research more clearly. (See Page 4, Line 155-156). The contents are as follows:

… (This part is the paragraph of the text)

Figure 2. Overall view of the study area. (a) Administrative Map of China. (b) Remote Sensing images of Hainan. (c) Areca palm on Google Images. (d) Mobile Phone photograph of Areca palm;

2. Reviewer: In the ground survey section (2.1.3) authors may add more details about the sampling strategy. How the survey points are chosen/decided, table may be presented to show how many points are chosen for survey in each city? How many points are surveyed vs area of the of the site/location, why 1500 points are chosen, how 100 points are arrived at? why 500 points are used for classification performance.

Authors: Thank you for your valuable feedback. We have revised Section 2.1.3 to include more details about our sampling strategy (See Page 5-6, Line 187-200). A new table has been added (Table 3. The Survey information of Areca Palm Plot.) to show the distribution of survey points across different cities. This table includes the number of points and average area surveyed in each city (See Page 6, Line 204). In addition, we have produced a survey distribution map of areca palm sample plots to make it more convenient for readers to understand the survey situation (See Page 6, Line 203). The contents are as follows:

… (This part is the paragraph of the text)

Figure 3. Areca palm sample plots distribution in Hainan Province.

From March 2023 to August 2023, our team conducted surveys at the areca palm plantations in each county and city within Hainan Province. Each number used a handheld GPS (Garmin GPSMAP 63csx) in conjunction with Google Earth (Google Inc., Santa Clara County, CA, USA) to record and store data detailing the distribution of the areca palm samples. To ensure spatial representativeness, we maintained a distance of approximately 1 kilometer between two areca palm sample. Ultimately, data for 1500 areca palm sample plots (Table 3). Our research divided the areca nut survey plots into three equal parts in order to analyze and construct a more reasonable areca palm classification model. The first set of 500 areca palm sample plots was used for training the deep learning model, with the data split into 70% for training and 30% for validation. The second set of 500 areca palm sample plots was reserved exclusively for testing the model's classification performance. The third set comprised 500 areca palm sample plots and 500 other land cover sample plots, which were used to evaluate the practical accuracy of the Random Forest model that was developed. The distribution of different Areca palm sample plots shown in Fig. 3.Table 3. The Survey information of Areca Palm Plot.

Cities/Counties	Areca palm plot	Average Area (ha)
BS	29	0.43
BT	112	0.37
CJ	1	0.17
CM	111	0.59
DZ	5	0.32
DA	149	0.34
DF	3	0.43
HK	90	0.35
LD	48	0.24
LS	12	0.16
LG	33	0.36
QH	263	0.52
QZ	174	0.36
SY	37	0.25
TC	150	0.41
WN	168	0.44
WC	78	0.31
WZS	37	0.19

 

3. Reviewer: Table 2 is a generalized information readily available to all!Authous may add little more information of the radiometric calibration, atmospheric corrections and geometric corrections. why its is need and what approach has been adopted.

Authors: We apologize for the oversight in not including detailed information regarding radiometric calibration, atmospheric corrections, and geometric corrections in the initial version of Table 2. We have now added this crucial information to the table to enhance clarity and provide a more comprehensive understanding. These corrections are essential for ensuring the accuracy and reliability of remote sensing data, as they address sensor-specific biases, atmospheric distortions, and geometric inconsistencies. (See Page 5, Line 166-171) The contents are as follows:

… (This part is the paragraph of the text)

Landsat Surface Reflectance (SR) data, with a spatial resolution of 30 m, were downloaded from the Google Earth Engine (GEE) platform. The Landsat 5 SR product have been processed using Landsat ecosystem disturbance adaptive processing system (LEDAPS) [46] on GEE platform. The steps included Geometric Correction, Radiometric Correction, Atmospheric Correction and topographic correction. The Landsat 8 SR product are created with the Land Surface Reflectance Code (LaSRC) [47]

Schmidt, G.; Jenkerson, C.B.; Masek, J.; Vermote, E.; Gao, F. Landsat ecosystem disturbance adaptive processing system (LEDAPS) algorithm description; 2013-1057; Reston, VA, 2013; p. 27.

Vermote, E.; Roger, J.C.; Franch, B.; Skakun, S. LaSRC (Land Surface Reflectance Code): Overview, application and validation using MODIS, VIIRS, LANDSAT and Sentinel 2 data's. In Proceedings of the IGARSS 2018 - 2018 IEEE International Geoscience and Remote Sensing Symposium, 22-27 July 2018, 2018; pp. 8173-8176.

4. Reviewer: section 2.2.1 on mapping with deep learning- authors may explain what it means ' high resolution image is richer and greater data magnitude'? line 197 on data redundancy needs to be discussed.

Authors: Thank you for your feedback. We have clarified the phrase 'high-resolution image is richer and greater in data magnitude.' By this, we meant that high-resolution images provide more detailed information, capturing finer spatial details, which we refer to as 'richer.' Additionally, 'greater data magnitude' refers to the larger volume of data in these images due to the increased number of pixels, offering a higher level of granularity. This abundance of detailed data enhances the effectiveness of deep learning models in image analysis. Regarding the mention of data redundancy, we recognize that it is primarily a concern in hyperspectral data. Its inclusion here was an oversight, and we have revised the text accordingly to correct this. (See Page 7, Line 232-237) The contents are as follows:

… (This part is the paragraph of the text)

Compared to the data obtained from medium- and high-resolution images, the data obtained from high-resolution images is richer and of a greater data magnitude. However, high-resolution images required greater computational power to process. Convolutional Neural Networks (CNNs) provide an effective approach for automatically extracting features from raw images using a mathematical operation known as convolution, making them suitable for processing large datasets.

5. Reviewer: line no. 410 what it means by 'area size'? is not clear.

Authors: Thank you for your suggestion, and we apologize for the lack of clarity in the 'area size' definition. We have changed it to the cultivation area. (See Page 15, Line 439). The contents are as follows:

… (This part is the paragraph of the text)

Areca palm is distributed in all cities and counties on Hainan Island, and based on cultivation area, it can be roughly categorized into three groups:

6. Reviewer: line no. 442 ' precipitation changes are relatively complex' is not clear to the reader, needs to be discussed properly.

Authors: Thank you for your valuable feedback. We have revised the manuscript to include a more detailed discussion on precipitation changes. (See Page 17, Line 470-430). The contents are as follows:

… (This part is the paragraph of the text)

At the same time, temperature changes on Hainan Island are relatively stable, while the precipitation showed fluctuating annual precipitation, with increases from 1,109 mm in 1987 to peaks like 1,917 mm in 2012, followed by decreases, reaching 1,590 mm in 2022.

7. Reviewer: the discussion is very long. There is scope to improve the section as per the results

Authors: Thank you for your feedback regarding the length of the discussion section. We acknowledge that the section was lengthy, which was an oversight on our part. However, due to the extensive research content, a detailed discussion was necessary to adequately cover all aspects. We have now revised the section, adding new content to improve its logical flow and coherence. (See Page 19, Line 499-596). The contents are as follows:

… (This part is the paragraph of the text)

4. Discussion

4.1. Historical distribution of the areca palm

Throughout the areca palm mapping process, three independent survey datasets were used to map its historical distributions. High-resolution images were processed using the Res50_U-net model to map the areca palm distribution for 2022 on Hainan Island, achieving an overall accuracy of 90%. This accuracy coincides with that of Luo’s machine learning-based areca palm classification method which achieved as accuracy of 89% [58]. However, the areca palm mapping in this study was oriented to a wider range of scales. Previous research has not evaluated the spatial distribution of the areca palm on such a large scale. This study has thus demonstrated the feasibility of utilizing deep learning for large-scale areca palm classifications using high-resolution imagery. Building upon this, the study has expanded a 1 km buffer zone to generate a potential areca palm distribution area, which was used to create a historical map of the distribu-tions on Hainan Island. Using Landsat imagery and a Random Forest algorithm with constraints on the potential distribution area, the overall accuracy for 2022 was 67%. These results indicate that the model has an adequate ability to identify areca palm dis-tributions. Historical areca palm distribution data was also examined and found to aligns well with the expected growth patterns during cultivation. This study has thus successfully depicted the historical evolution of areca palm cultivation, producing, for the first time, providing a large-scale, high-resolution spatial distribution map over an extended period. This study has also introduced a novel approach for classifying dis-persed and patchy crops. However, it should be noted that the addition of the 1 km potential distribution range will lead to some areca palm not included in the study area, resulting in omission, which is also an aspect that needs to be improved in the future, how to obtain a more refined potential distribution of betel nut through more effective means. In addition, the accuracy of the areca palm distribution data before 2022 cannot be guaranteed due to the lack of field data classification accuracy, which means that more continuous field data will be needed in the future.

4.2. Spatiotemporal evolution of areca palm plantation on Hainan Island

From 1987 to 2022, areca palm plantations on Hainan Island exhibited a relatively clustered pattern. This clustering is influenced by both the actual cultivation practices and the limitations of the areca palm's potential distribution range. While some overes-timation of clustering might occur due to these distribution range limitations, the pattern still reflects the true extent of areca palm plantations. This study identified two distinct areca palm plantation patterns on Hainan Island. The first is a high-value clus-ter in the southeastern region, which has shown an expansion trend over time. This pattern is most prominent in cities like Wanning and Qionghai. Wanning is renowned as the areca palm hub of China, while Qionghai has seen significant growth in areca palm cultivation in recent years, making it a key driver of local economic development. The high-value clustering enhances scale and efficiency, thus promoting the growth of the areca palm industry.

Conversely, a low-value clustering pattern is found in the southwestern and northeastern regions of Hainan Island. This pattern, particularly evident in areas like Haikou, Danzhou, and Changjiang, has not shown any signs of expansion. These re-gions do not rely heavily on the areca palm industry, which is not a major economic pillar for them. These two spatial distribution patterns highlight the significant differ-ences in the spatial distribution and plantation practices of areca palms on Hainan Is-land, creating a clear east-west divide. The differences are influenced by varying levels of governmental attention, policies, and economic incentives directed towards the areca palm industry. However, the overall trend for the areca palm industry is one of rapid expansion in plantation areas, with rising prices playing a significant role in this growth. This has, however, led to rigid cultivation practices, limited the integration of technology, and fostered exploitative management. Therefore, it is necessary for the government to regulate the expansion of the areca palm industry and prevent farmers from blindly following cultivation trends. For instance, oil palm plantations in Malaysia pose a threat to peatlands, which are home to a rich diversity of flora and fauna [59]. Instead, there should be a focus on promoting the integrated application of plantation technologies to accelerate the upgrading of the areca palm industry, realize the three-dimensional comprehensive development and utilization of areca palm, and im-prove the output value of the unit area as well as the benefits of the vacated land for the development of the understory economy.

4.3. Mechanisms influencing the evolution of Areca palm plantation

From 1987 to 2022, the areca palm cultivation area on Hainan Island expanded dramatically, increasing from 8,064 ha to 193,328 ha, over 20 times. This growth in areca palm plantations strongly correlates with Hainan Island’s GDP, underscoring the crop's significant economic role. However, this correlation varies across regions. For example, Sanya shows no significant correlation, as its GDP is primarily driven by tourism rather than agriculture, and it cultivates a diverse array of crops like mangoes, coconuts, and rice alongside areca palms. Similarly, Danzhou, with rubber as its primary crop, does not show a significant correlation due to low rubber prices, despite a growing interest in switching to areca palms. In contrast, regions like Wanning and Qionghai exhibit strong correlations, reflecting the flourishing areca palm-related industries, especially in the past five years.

Field investigations revealed various areca palm cultivation types in Hainan, in-cluding plantations near farmlands, within natural tropical rainforests, in backyards, on flatlands, and in hilly and mountainous areas. The dispersed and fragmented nature of these plantations, sometimes encroaching on Hainan Tropical Rainforest National Park, raises environmental concerns, including biodiversity loss and reduced carbon seques-tration capacity. Additionally, the expansion of areca palm plantations at the expense of traditional agricultural land poses risks to ecological balance, as noted in the work of Takeuchi, where monoculture and forest destruction can lead to ecosystem degrada-tion and rural population outflow [60]. The areca plam trade often results in a series of associated infrastructures, including trading markets and other facilities, and these de-velopments are frequent in urban areas. Cities and towns have developed transporta-tion systems and convenient trading environments. The development of the areca palm is also a microcosm of economic evolution, as the focus shifts from rural to urban areas. Indeed, the areca palm is a high-value cash crop that can be lucrative for farmers; however, achieving large-scale planting and efficient management while ensuring eco-logical sustainability and a harmonious relationship between humans and nature are significant challenges for Hainan Island and the rest of the world. Srinivasan’s work provides an example with India’s oil palm. However, the current expansion of  oil palm cultivation in India comes at the cost of biodiversity-rich landscapes. But their model suggests that, on a national scale, India seems to have viable options to meet its projected palm oil demand without compromising its biodiversity or food security [61]. Can the planting of Areca palm in Hainan meet the needs of the forest through macro policy regulation and control under the consideration of the existing planting concen-tration. On a more precise spatial scale, areca palm cultivation needs to take into ac-count local climatic conditions, biodiversity, local agricultural input ratios, and trade-offs between economy and social security. China's policy decisions on areca palm have largely mitigated the current set of problems facing China's tropical rainforests.

8. Reviewer: in section 2.1.3 ' a team of 20 individuals' is not necessary to be mentioned

Authors: We followed the opinions of the reviewer. We deleted ' a team of 20 individuals' and corrected the phrase to make it clearer and grammatically (See Page 5, Line 187-188). The contents are as follows:

… (This part is the paragraph of the text)

From March 2023 to August 2023, our team conducted surveys at the areca palm plantations in each county and city within Hainan Province.

9. Reviewer: English language corrections are suggested in few sections:

line no.141'was organized' may be checked

line no. 199 'by which to automatically' needs correction

Similar correction may be checked in the article carefully.

Authors: Your comments would be very helpful to us. We have carefully corrected the English language throughout the article. We removed the 'was organized' and reorganize the language. (See Page 5, Line 187-188). We deleted the awkward "by which to automatically" and streamlines the sentence for clarity. (See Page 7, Line 234-237). The contents are as follows:

  … (This part is the paragraph of the text)

From March 2023 to August 2023, our team conducted surveys at the areca palm plantations in each county and city within Hainan Province.

Convolutional Neural Networks (CNNs) provide an effective approach for automatically extracting features from raw images using a mathematical operation known as convolution, making them suitable for processing large datasets

Author Response File: Author Response.pdf

Reviewer 3 Report

Comments and Suggestions for Authors

1.       Introduction line 38-39, you also can give example of Oil palm which is the main cash crop in SE Asia.

2.       Line 125-126, “Adundant” typo should be “abundant”.

 

3.       Line 354 “Literation” shouldn’t it be “iteration”?

Author Response

Dear Reviewer 3,

Thank you very much for your time and effort reviewing this manuscript. We have responded to the comments below. We believe the edits have resulted in an improved manuscript and our response to all of the comments is provided below. In addition, we attach great importance to the advice given by reviewer and feel guilty about the mistakes in previous articles! We have marked your reply with a blue background color. In addition, we have changed the font of the added references to red.

Yours sincerely,

(on behalf of all co-authors)

 1. Reviewer: Introduction line 38-39, you also can give example of Oil palm which is the main cash crop in SE Asia.

Authors: Your suggestion is enlightening to us. We added a reference about Oil palm which is the main cash crop in SE Asia. (See Page 1, Line 32-46). The contents are as follows:

… (This part is the paragraph of the text)

Qaim, M.; Sibhatu, K.T.; Siregar, H.; Grass, I. Environmental, Economic, and Social Consequences of the Oil Palm Boom. Annual Review of Resource Economics 2020, 12, 321-344,

2. Reviewer: Line 125-126, “Adundant” typo should be “abundant”

Authors: Your suggestion reminds us that we will correct it. We have modified the typo error. (See Page 4, Line 146-147). The contents are as follows:

… (This part is the paragraph of the text)

Abundant sunlight and favorable temperatures increase the photosynthetic potential, making the area suitable for areca palm cultivation.

 3. Reviewer: Line 354 “Literation” shouldn’t it be “iteration”?

Authors: We appreciate your advice. We feel guilty about our mistakes. We have revised the spelling mistakes. (See Page 12, Line 392). The contents are as follows:

… (This part is the paragraph of the text)

The loss function stabilized at 0.1 after 20,000 iterations.

Author Response File: Author Response.pdf

Reviewer 4 Report

Comments and Suggestions for Authors

The manuscript under review demonstrates a significant contribution to the understanding of the spatiotemporal evolution and impact mechanisms of areca palm plantations in China, utilizing advanced remote sensing techniques and deep learning models. However, several areas could be refined to enhance the clarity and impact of the research.

The title is engaging and captures the scope of the study effectively, but it could be more concise. A suggested revision would be “Spatiotemporal Evolution and Impact Mechanisms of Areca Palm Plantations in China (1987–2022),” which maintains the core message while being more succinct.

In terms of authorship, the affiliations listed are comprehensive but display inconsistencies in punctuation and capitalization. For example, "Hainan University" is correctly capitalized, while "Sanya institute of Breeding and Multiplication" is not. These minor discrepancies should be corrected to ensure professional consistency.

The abstract offers a solid summary of the research, though some sentences are overly complex, making the information difficult to follow. For example, the sentence describing the rapid increase in plantation area from 8,064 hectares in 1987 to 193,328 hectares in 2022 could be simplified for clarity. Additionally, while the correlations mentioned are informative, they lack statistical support. Including statistical tests, such as p-values, would bolster the scientific rigor of the abstract.

The introduction is well-constructed, but a few typographical and structural issues need attention. For instance, "cash crop" should be pluralized to "cash crops," and there is a typographical error in the phrase “In recent years., the area used globally.” Furthermore, while the mention of spiritual attributes and social values is intriguing, this point could be more closely connected to the main subject of the study to avoid feeling tangential. Linking these elements to areca palm plantations more directly would create a more cohesive argument.

The materials and methods section is thorough but lacks some details concerning the verification of datasets. It would be beneficial to describe any quality control measures that were implemented. The model descriptions, particularly regarding ResNet and Random Forest, are rigorous, but the rationale behind choosing specific parameters, such as the use of “50 decision trees,” could be further elaborated. Moreover, the resolution of the imagery used during different time periods should be emphasized to account for potential variability in accuracy when mapping historical versus recent plantation areas.

The results section provides a comprehensive presentation of the findings, particularly in terms of model performance and historical trends. However, some areas, such as the spatial analysis involving Global Moran's Index, are dense and could benefit from simplification to aid reader comprehension. Additionally, while the figures are generally well-designed, they could be improved in terms of clarity, specifically with regard to axis labels and color scales. Adding subtitles to the figures could further enhance their accessibility.

The discussion section effectively interprets the results, but it would benefit from a more critical analysis of the study's limitations, particularly with respect to potential biases in historical data extrapolation using the 1 km buffer. The discussion of economic and environmental trade-offs is strong, yet additional references to similar case studies outside of China would provide a broader context for the findings. Additionally, the mechanisms through which socioeconomic factors influence areca palm plantation growth should be explored in more depth, including consideration of policy factors.

The conclusion effectively summarizes the key findings but could be more explicit about the broader implications of the research, particularly in terms of informing sustainable agricultural policy in China and beyond. The statement regarding the feasibility of utilizing deep learning could be strengthened by suggesting potential next steps for improving or scaling the models.

The references are extensive and support the claims made in the manuscript. However, incorporating a few more recent studies on the intersection of remote sensing and machine learning in agriculture could help reflect the latest advances in the field.

Overall, the language of the manuscript is professional, though it requires revisions for clarity and conciseness. Several grammatical errors and instances of awkward phrasing need to be addressed. Additionally, consistent formatting of figures, tables, and units is necessary for a more polished presentation.

In summary, the manuscript is rigorous and well-researched, making a valuable contribution to the field. However, refinements in the presentation of results and discussion, particularly in simplifying complex sections and addressing limitations, would enhance the overall impact of the research.

The manuscript is suitable for acceptance with minor revisions.

 

Comments on the Quality of English Language

 Overall, the language of the manuscript is professional, though it requires revisions for clarity and conciseness. Several grammatical errors and instances of awkward phrasing need to be addressed. 

Author Response

Dear Reviewer 4,

Thank you very much for your time and effort reviewing this manuscript. We have responded to the comments below. We believe the edits have resulted in an improved manuscript and our response to all of the comments is provided below. In addition, we attach great importance to the advice given by reviewer and feel guilty about the mistakes in previous articles! We have marked your reply with a pink background color. In addition, we have changed the font of the added references to red.

Yours sincerely,

(on behalf of all co-authors)

 1. Reviewer: The title is engaging and captures the scope of the study effectively, but it could be more concise. A suggested revision would be “Spatiotemporal Evolution and Impact Mechanisms of Areca Palm Plantations in China (1987–2022),” which maintains the core message while being more succinct. In terms of authorship, the affiliations listed are comprehensive but display inconsistencies in punctuation and capitalization. For example, "Hainan University" is correctly capitalized, while "Sanya institute of Breeding and Multiplication" is not. These minor discrepancies should be corrected to ensure professional consistency.

Authors: We followed the opinions of the reviewer. The original title " First-Ever understanding the spatiotemporal evolution and impact mechanisms of areca palm plantations in China from 1987 to 2022" was changed to " Spatiotemporal Evolution and Impact Mechanisms of Areca Palm Plantations in China (1987–2022)". (See Page 1, Line 2-3). Thank you for your thorough review of our manuscript. We have carefully reviewed and corrected the inconsistencies in punctuation and capitalization among the affiliations, as you pointed out. "Sanya institute of Breeding and Multiplication" has been corrected to "Sanya Institute of Breeding and Multiplication," ensuring consistency across all affiliations. We appreciate your attention to detail, which has helped improve the quality and professionalism of our manuscript. (See Page 1, Line 6). The contents are as follows:

… (This part is the paragraph of the text)

Spatiotemporal Evolution and Impact Mechanisms of Areca Palm Plantations in China.

 School of Breeding and Multiplication (Sanya Institute of Breeding and Multiplication), Hainan University,

2. Reviewer: The abstract offers a solid summary of the research, though some sentences are overly complex, making the information difficult to follow. For example, the sentence describing the rapid increase in plantation area from 8,064 hectares in 1987 to 193,328 hectares in 2022 could be simplified for clarity. Additionally, while the correlations mentioned are informative, they lack statistical support. Including statistical tests, such as p-values, would bolster the scientific rigor of the abstract.

Authors: Thank you for your insightful comments on our abstract. We have revised the sentence describing the increase in plantation area to make it more concise and easier to follow. The original sentences " Results revealed a rapid increase in plantation area, soaring from 8,064 hectares in 1987 to 193,328 hectares in 2022." was changed to " Plantation area rapidly expanded from 8,064 hectares in 1987 to 193,328 hectares in 2022.". (See Page 1, Line 21-22). Additionally, we appreciate your suggestion to include statistical tests to support the correlations mentioned. We have now incorporated the relevant p-values to enhance the scientific rigor of the abstract. (See Page 1, Line 24-27). The contents are as follows:

… (This part is the paragraph of the text)

Plantation area rapidly expanded from 8,064 hectares in 1987 to 193,328 hectares in 2022.

Moreover, the plantation area exhibited a significant positive correlation with both GDP (r = 0.98, p < 0.001) and total population (r = 0.92, p < 0.01), while negatively correlating with rural population (r = -0.76, p < 0.05).

 

3. Reviewer: The introduction is well-constructed, but a few typographical and structural issues need attention. For instance, "cash crop" should be pluralized to "cash crops," and there is a typographical error in the phrase “In recent years., the area used globally.” Furthermore, while the mention of spiritual attributes and social values is intriguing, this point could be more closely connected to the main subject of the study to avoid feeling tangential. Linking these elements to areca palm plantations more directly would create a more cohesive argument.

Authors: Thank you for your valuable feedback. We have addressed the typographical and structural issues in the Introduction, including correcting "cash crop" to "cash crops" and the typographical error in the phrase “In recent years., the area used globally.” (See Page 1, Line 35-36). We have revised the discussion on the spiritual attributes and social values to better align with the main subject of the study (See Page 2, Line 59-61). We have also enriched the Introduction by adding relevant content on deep learning and machine learning applications in the study of palm-type crops. The contents are as follows:

… (This part is the paragraph of the text)

In recent years, the global area dedicated to cash crops cultivation has been steadily increasing

.Hainan betel nuts have been sold directly after ripe picking; or simple preliminary processing, and then resold to enterprises in Hunan and Guangxi for further processing. Therefore, it is necessary to explore the mechanisms of areca palm impact.

Freudenberg, M.; Nölke, N.; Agostini, A.; Urban, K.; Wörgötter, F.; Kleinn, C. Large Scale Palm Tree Detection in High Resolution Satellite Images Using U-Net. Remote Sensing 2019, 11, 312

Li, W.; Fu, H.; Yu, L.; Cracknell, A. Deep Learning Based Oil Palm Tree Detection and Counting for High-Resolution Remote Sensing Images. Remote Sensing 2017, 9, 22

Cheng, Y.; Yu, L.; Xu, Y.; Lu, H.; Cracknell, A.P.; Kanniah, K.; Gong, P. Mapping oil palm extent in Malaysia using ALOS-2 PALSAR-2 data. International Journal of Remote Sensing 2018, 39, 432-452, doi:10.1080/01431161.2017.1387309.

Lee, J.S.H.; Wich, S.; Widayati, A.; Koh, L.P. Detecting industrial oil palm plantations on Landsat images with Google Earth Engine. Remote Sensing Applications: Society and Environment 2016, 4, 219-224, doi:https://doi.org/10.1016/j.rsase.2016.11.003

 

5. Reviewer: The materials and methods section is thorough but lacks some details concerning the verification of datasets. It would be beneficial to describe any quality control measures that were implemented. The model descriptions, particularly regarding ResNet and Random Forest, are rigorous, but the rationale behind choosing specific parameters, such as the use of “50 decision trees,” could be further elaborated. Moreover, the resolution of the imagery used during different time periods should be emphasized to account for potential variability in accuracy when mapping historical versus recent plantation areas.

Authors: We appreciate the reviewer’s feedback regarding the verification of datasets. To ensure the quality and reliability of our datasets, we added the relevant contents in materials and methods section (See Page 5, 2.1.3. Ground survey data sources). The contents are as follows:

.… (This part is the paragraph of the text)

From March 2023 to August 2023, our team conducted surveys at the areca palm plantations in each county and city within Hainan Province. Each number used a handheld GPS (Garmin GPSMAP 63csx) in conjunction with Google Earth (Google Inc., Santa Clara County, CA, USA) to record and store data detailing the distribution of the areca palm samples. To ensure spatial representativeness, we maintained a distance of approximately 1 kilometer between two areca palm sample. Ultimately, data for 1500 areca palm sample plots (Table 3). Our research divided the areca nut survey plots into three equal parts in order to analyze and construct a more reasonable areca palm clas-ification model. The first set of 500 areca palm sample plots was used for training the deep learning model, with the data split into 70% for training and 30% for validation. The second set of 500 areca palm sample plots was reserved exclusively for testing the model's classification performance. The third set comprised 500 areca palm sample plots and 500 other land cover sample plots, which were used to evaluate the practical accuracy of the Random Forest model that was developed. The distribution of different Areca palm sample plots shown in Fig. 3.

5. Reviewer: The results section provides a comprehensive presentation of the findings, particularly in terms of model performance and historical trends. However, some areas, such as the spatial analysis involving Global Moran's Index, are dense and could benefit from simplification to aid reader comprehension. Additionally, while the figures are generally well-designed, they could be improved in terms of clarity, specifically with regard to axis labels and color scales. Adding subtitles to the figures could further enhance their accessibility.

Authors: Thank you for your positive feedback on the results section. We acknowledge your suggestion regarding the density of the spatial analysis involving the Global Moran's Index. We simplified this section to enhance reader comprehension. Thank you for your feedback on the figures. We have carefully reviewed the images and believe that they meet the current standards. However, we appreciate your suggestions for enhancing clarity, and we will take them into consideration for any future revisions. Thank you again for your valuable input on improving the figures. (See Page 16, Line 453-460) .The contents are as follows:

… (This part is the paragraph of the text)

Fig.11 illustrates the Global Moran’s Index for areca palms on Hainan Island from 1987 to 2022. The Global Moran’s Index measures the degree to which similar values are clustered within the study area. Over the period, the index shows an increasing trend, peaking at 0.45 in 1997 and 0.51 in 2012, indicating a significant positive spatial correlation. This suggests that areca palms have become more spatially clustered over time. Fig. 12 presents the Local Moran’s Index, which provides insight into specific clustering and dispersion patterns within the island. This detailed spatial analysis helps to under-stand the distribution and changes in areca palm cultivation from 1987 to 2022.

6. Reviewer: The discussion section effectively interprets the results, but it would benefit from a more critical analysis of the study's limitations, particularly with respect to potential biases in historical data extrapolation using the 1 km buffer. The discussion of economic and environmental trade-offs is strong, yet additional references to similar case studies outside of China would provide a broader context for the findings. Additionally, the mechanisms through which socioeconomic factors influence areca palm plantation growth should be explored in more depth, including consideration of policy factors.

Authors: Your question is very helpful to us, which is also our concern. We have conducted an analysis of the potential biases introduced by the 1 km buffer (See Page 19, Line 519-525) and added references to similar studies from outside of China to support our findings (See Page 20, Line 548-553). Additionally, we have delved deeper into the mechanisms by which socioeconomic factors, including policy influences, affect the growth of areca palm plantations. (See Page 21, Line 586-596) The contents are as follows:

… (This part is the paragraph of the text)

However, it should be noted that the addition of the 1 km potential distribution range will lead to some areca palm not included in the study area, resulting in omission, which is also an aspect that needs to be improved in the future, how to obtain a more refined potential distribution of betel nut through more effective means. In addition, the accuracy of the areca palm distribution data before 2022 cannot be guaranteed due to the lack of field data classification accuracy, which means that more continuous field data will be needed in the future.

This has, however, led to rigid cultivation practices, limited the integration of technol-ogy, and fostered exploitative management. Therefore, it is necessary for the govern-ment to regulate the expansion of the areca palm industry and prevent farmers from blindly following cultivation trends. For instance, oil palm plantations in Malaysia pose a threat to peatlands, which are home to a rich diversity of flora and fauna [59].

Srinivasan’s work provides an example with India’s oil palm. However, the current ex-pansion of oil palm cultivation in India comes at the cost of biodiversity-rich landscapes. But their model suggests that, on a national scale, India seems to have viable options to meet its projected palm oil demand without compromising its biodiversity or food se-curity [61]. Can the planting of Areca palm in Hainan meet the needs of the forest through macro policy regulation and control under the consideration of the existing planting concentration. On a more precise spatial scale, areca palm cultivation needs to take into account local climatic conditions, biodiversity, local agricultural input ratios, and trade-offs between economy and social security. China's policy decisions on areca palm have largely mitigated the current set of problems facing China's tropical rainforests.

7. Reviewer: The conclusion effectively summarizes the key findings but could be more explicit about the broader implications of the research, particularly in terms of informing sustainable agricultural policy in China and beyond. The statement regarding the feasibility of utilizing deep learning could be strengthened by suggesting potential next steps for improving or scaling the models. The references are extensive and support the claims made in the manuscript. However, incorporating a few more recent studies on the intersection of remote sensing and machine learning in agriculture could help reflect the latest advances in the field.

Authors: We sincerely appreciate your thoughtful and constructive feedback on our manuscript. Your insights have been invaluable in helping us enhance the clarity and impact of our work. We are grateful for your recognition of our efforts, and we agree that there are opportunities to further highlight the broader implications of our findings. We cited a few more recent studies on the intersection of remote sensing and machine learning in agriculture. The contents are as follows

… (This part is the paragraph of the text)

This study represents a significant advancement in understanding the spatial dis-tribution and driving forces behind areca palm cultivation on Hainan Island, China, by leveraging the power of satellite remote sensing and deep learning. By producing the first large-scale, long-term spatial distribution analysis at a 30 m resolution, the research has not only achieved high classification accuracy but also provided critical insights in-to the socioeconomic and environmental factors influencing areca palm expansion. The spatiotemporal dynamics revealed in this study underscore the complex interplay be-tween agricultural practices and broader regional development trends. Importantly, the identification of two distinct areca palm plantation patterns offers a new lens through which to assess the sustainability of current practices. These findings highlight the need for future research to delve deeper into the long-term ecological impacts of these plantation patterns and to refine remote sensing methodologies to further enhance accuracy and resolution. Furthermore, this research underscores the broader implications for sustainable agricultural policy, not only in China but also in other regions facing similar challenges. By scaling the models and incorporating more recent advancements in re-mote sensing and machine learning, future studies could expand this work to addition-al countries and regions, providing valuable data to support global agricultural sustainability efforts. Ultimately, this research serves as a crucial reference point for developing policies that promote sustainable agricultural practices, contributing to the global discourse on agricultural and environmental resilience.

Freudenberg, M.; Nölke, N.; Agostini, A.; Urban, K.; Wörgötter, F.; Kleinn, C. Large Scale Palm Tree Detection in High Resolution Satellite Images Using U-Net. Remote Sensing 2019, 11, 312

Li, W.; Fu, H.; Yu, L.; Cracknell, A. Deep Learning Based Oil Palm Tree Detection and Counting for High-Resolution Remote Sensing Images. Remote Sensing 2017, 9, 22

Cheng, Y.; Yu, L.; Xu, Y.; Lu, H.; Cracknell, A.P.; Kanniah, K.; Gong, P. Mapping oil palm extent in Malaysia using ALOS-2 PALSAR-2 data. International Journal of Remote Sensing 2018, 39, 432-452, doi:10.1080/01431161.2017.1387309.

Lee, J.S.H.; Wich, S.; Widayati, A.; Koh, L.P. Detecting industrial oil palm plantations on Landsat images with Google Earth Engine. Remote Sensing Applications: Society and Environment 2016, 4, 219-224, doi:https://doi.org/10.1016/j.rsase.2016.11.003

 

8. Reviewer: Overall, the language of the manuscript is professional, though it requires revisions for clarity and conciseness. Several grammatical errors and instances of awkward phrasing need to be addressed. Additionally, consistent formatting of figures, tables, and units is necessary for a more polished presentation.

Authors: Thank you for your constructive feedback. To ensure the language in our manuscript is clear and highly readable, we have engaged a professional editing service for thorough language polishing. In addition, we have made revisions to address grammatical errors and improve the flow of certain sections. Additionally, we will ensure consistent formatting of figures, tables, and units to enhance the overall presentation of the manuscript.

 9. Reviewer: In summary, the manuscript is rigorous and well-researched, making a valuable contribution to the field. However, refinements in the presentation of results and discussion, particularly in simplifying complex sections and addressing limitations, would enhance the overall impact of the research.

Authors: Thank you for your encouraging feedback and thoughtful suggestions. We are pleased that you find the manuscript rigorous and valuable. We have refined the presentation of the results and discussion by simplifying complex sections and addressing the identified limitations. These improvements have been made to enhance the overall impact and clarity of our research.

Author Response File: Author Response.pdf