Thanks for your  valuable comments concerning our manuscript entitled “Retrieval of dominant methane (CH4) emission sources, the first high resolution (1-2m) dataset of storage tanks of China in 2000-2021” (essd-2024-10). These comments are very helpful in revising and improving our paper, as well as guiding our future research. We have studied the comments carefully and tried our best to revise the manuscript.

This manuscript generates a storage tank dataset (STD) by implementing a deep learning model with manual refinement based on high spatial resolution images (1-2m) from the GaoFen-1, GaoFen-2, GaoFen-6, and Ziyuan-3 satellites over some region in China in 2021. Since oil gas infrastructure data is barely publicly available in China, this dataset could give researchers a tool to estimate or scale up methane emissions from relevant industry sectors, and thus it is of value to be published in this journal.

The reviewer has several comments which may help improving the quality of the manuscript:

Q1. Targeted tanks: The manuscripts screens tanks under two criteria: footprint bigger than 500 m2; located within the built area and bare land in the LULC product of the Esri Land Cover in 2021.While the objective of this study is to retrieve dominant methane emission sources, the reviewer is wondering how many of tanks retrieved are methane emission sources, and how many tanks which are methane emissions sources, are neglected. For example, oil gas batteries are major methane emission sources, and located in the rural agriculture area or desert. Will these batteries be identified under the current algorithm. What is the rational for 500 m2 threshold? Are there tanks who have footprints smaller than 500 m2?

To make this clear, the reviewer would suggest a table summarizing the key features of tanks, including the following aspects: Usage/function; typical size (water volume, m3); typical footprint (m2); typical location (urban, rural?); industry sector.

Answer: Thanks very much for your suggestion. The storage tanks are extracted from remote sensing images, according to their spectral, textual, and morphological features based on deep learning frameworks as illustrated in our manuscript. It is very difficult to tell the exact usage of each storage tank only according to the appearance in the image, since storage tanks may take similar appearances with different functions.

Moreover, it is very difficult to estimate typical size of each storage tank because of lack of height. The typical footprint in the unit of m² has been provided in the attribute table of our dataset. The storage tanks located in rural agriculture area or desert are not considered in our proposed dataset, because storage tanks of crude oil or other petroleum, and industrial materials, such as alcohols, gases, or liquids, are among the most significant sources of emitting CH₄ (Im et al., 2022; Johnson et al., 2022), and they are predominantly situated in urban area or development zones in China due to pipeline transport logistics. Additionally, the high cost and limited access to high-resolution spatial imagery, coupled with the extensive coverage required for rural and desert areas, make it impractical to include these tanks in our study. Therefore, we did not distinguish the typical location of the storage tanks in our dataset since they are all located in urban area.

In terms of the threshold of 500m², it is determined based on the spatial resolution of remote sensing images we collected to construct the dataset. The storage tanks with 125 pixels or more in remote sensing images can be distinguished more accurately from the complicated background objects. There are tanks with footprints smaller than 500m², but they are more difficult to be extracted accurately due to the small size. Moreover, since large capacity storage tanks are known to release significant levels of CH₄, our proposed inventory focuses on storage tanks with an area of no less than 500 m². The corresponding sentence has been revised to be ‘Given that large capacity storage tanks are known to release significant levels of CH₄, resulting in climate warming, the proposed inventory focuses on storage tanks with an area of no less than 500 m²’ in Line 400-402 of the revised manuscript.

Q2. Dataset accuracy verification. The reviewer did not see any verification of the dataset in the manuscript. Have the researchers verified their findings, for example, checking the accuracy of the dataset for one city/region? How many true positive, false positive? How many tanks are missed? Without such verification process, this is not a complete stand-alone paper.

Answer: Actually, each storage tank of our proposed dataset has been manually validated and refined by six experienced experts through visual interpretation based on our collected high spatial resolution images and field survey to facilitate the corresponding construction year. Therefore, we did not conduct dataset evaluation specifically because each storage tank has been verified and refined. To make the manuscript easier to understand, the corresponding sentence has been modified to be ‘To facilitate the dating of each storage tank's construction year, the reconstructed circle for each extracted storage tank has been manually validated and refined by six experienced experts through visual interpretation based on our collected high spatial resolution images and field survey.’ in Line 369-372 of the revised manuscript.

Q3. Construction year assignment. The researchers identified the construction year of tanks by the high-resolution historical images available on Google Earth. Can the researchers also identify how many tanks disappear during the period of study?

Answer: The storage tanks extracted in this study are those existing in the high spatial resolution images we collected in year of 2021, and the corresponding construction year is determined for each extracted storage tank referring to the historical images available on Google Earth. Due to the large scale of our study area and the limited availability of high spatial resolution images, the storage tanks are extracted based on mono-temporal images, rather than time-series images. Therefore, it is difficult to identify the disappeared storage tanks during the period of study.

Minor Observations:

Q1. Line 53, “ … with 85 times more climate warming potency than CO2”: please specify the time length of 85 times.

Answer: Thanks very much for your suggestion. We have specified the time length in Line 52-54 of the revised manuscript as ‘Meanwhile, CH4 is more effective in trapping heat, with 85 times more climate warming potency than CO2 for a decade or two (Stocker, 2014).’.

Q2. Line 108, “size > 500 m2”. Tank size is usually characterized by its water volume. Suggest changing tank size to tank footprint.

Answer: Thanks very much for your generous suggestion. We have changed tank size to tank footprint throughout the revised manuscript.

Q3. Line 114. “cities”: city in English usually refers to urban area. Here it looks like that the city covers both the urban area, as well as the rural counties. Would suggest to change the “city” to “region”, and explain the true meaning of it.

Answer: Your suggestion is much appreciated. The word region is better, but it is difficult to describe the administrative boundary of our study area. Therefore, we have modified city to city region throughout the revised manuscript. Hopefully, it is clearer this time.

Q4. Line 168, digital elevation model (DEM), and Figure 1: Is DEM a model or elevation? In figure 1. DEM High 4708, Low -91. What’s the unit for DEM? Please clarify, and add unit in Figure 1 (m?)

Answer: The unit for DEM is meter. Thanks very much for your suggestion. We have added the unit in the caption of Figure 1 as ‘Figure 1. Study area demonstration with digital elevation (in the unit of meter) from the Shuttle Radar Topography Mission (SRTM) product.’ in Line 178-179 of the revised manuscript.

Q5. Line 204 “Given that storage tanks are constructed mainly in residential areas”: Please elaborate why. To the reviewer’s knowledge, most oil gas storage tanks should be outside of the residential areas.

Answer: The storage tanks are mostly constructed in urban area because of the high expense of pipeline transportation. Moreover, according to our field survey, the storage tanks, especially storing oil and gas, are mostly located in industrial area of urban area. To make the manuscript easier to understand, the corresponding sentence has been modified to be ‘Given that storage tanks are constructed mainly in urban area due to the high expense of transportation of pipelines’ in Line 205-206 of the revised manuscript.

Thanks for your  valuable comments concerning our manuscript entitled “Retrieval of dominant methane (CH4) emission sources, the first high resolution (1-2m) dataset of storage tanks of China in 2000-2021” (essd-2024-10). These comments are very helpful in revising and improving our paper, as well as guiding our future research. We have studied the comments carefully and tried our best to revise the manuscript. 

Please see the detailed responses in the attchment.