Submit to this Journal Review for this Journal Propose a Special Issue

Article Menu

Share Help Cite Discuss in SciProfiles

Open AccessArticle

Peer-Review Record

Estimation of Shortwave Solar Radiation on Clear-Sky Days for a Valley Glacier with Sentinel-2 Time Series

Remote Sens. 2020, 12(6), 927; https://doi.org/10.3390/rs12060927

by Yanli Zhang^1,*

, Xiang Qin², Xin Li^3,4, Jun Zhao¹ and Yushuo Liu²

Reviewer 1: Anonymous

Reviewer 2: Anonymous

Reviewer 3: Anonymous

Remote Sens. 2020, 12(6), 927; https://doi.org/10.3390/rs12060927

Submission received: 4 January 2020 / Revised: 11 March 2020 / Accepted: 11 March 2020 / Published: 13 March 2020

(This article belongs to the Special Issue Remote Sensing of Climatic and Environmental Changes over the Antarctic, Arctic, and the Qinghai-Tibet Plateau)

Round 1

Reviewer 1 Report

The paper proposes the investigation of Shortwave Solar Radiation on Clear-3 Sky Days for a Valley Glacier with Sentinel-2 Time.

Unfortunately, at this stage, I cannot figure out what is the goal of this research and what is intellectual merit and what is the novelty of the paper.

“reliable estimates of the shortwave solar radiation” how to understand this sentence?

The study area

To test the new approach, you need more than one site for investigation.

Data sources

The authors use three different data sources to evaluate different quantities. Not clear what is data management and data validity.

Methods

“This paper adopted the ….” Why and what reason to adopt methods, where is your contribution to science.

What is a purpose of formula 1? I think page 6 is not useful at all.

Concern about data using for validation I think this is too little.

English has to be improved as well as all abbreviations and acronyms must be disclosed.

Overall, the paper must be improved by showing novelty at each step of the research.

Author Response

Reviewer 1’s Comments and Our Responses

We sincerely thank you for your insightful comments. We have provided point-by-point responses below. The comments have been copied below and are followed by an explanation of how they were addressed in the resubmitted manuscript. Your original comments appear in black; our responses follow in blue.

The modifications are marked with red text in the paper.

The paper proposes the investigation of Shortwave Solar Radiation on Clear-Sky Days for a Valley Glacier with Sentinel-2 Time. Unfortunately, at this stage, I cannot figure out what is the goal of this research and what is intellectual merit and what is the novelty of the paper.

Reply: We apologize for your doubts due to the inaccurate description in our manuscript. However, your doubts gave us the opportunity to carefully amend this study. We have revised it from two aspects: one is to adjust the organizational structure of the introduction section (paragraphs 2, 3, 4, and 5 in the original manuscript); the other is to add or modify descriptive statements about the goal of the research and the novelty of this paper, which are described as follows:

Lines 20-22 in the abstract section:

Based on a high-resolution (12 m) DEM, in this paper, the newly launched Sentinel-2 satellites, rather than MODIS and TM, were used to provide input data for our published mountain radiation scheme in a valley glacier.

Lines 106-113 in the introduction section:

Compared with MODIS and Landsat TM, the use of Sentinel-2 (S2) satellite data to estimate valley glacier DSSR has four advantages: (1) S2 can provide both atmospheric and surface information, thus reducing the errors caused by different satellite data; (2) S2 data can improve the DSSR spatiotemporal resolution and provide reliable information for understanding the melting of glaciers; (3) S2 data can almost completely avoid the saturation defect of the snow signal from optical satellites because of its high radiation resolution; (4) the data processing software Sen2Cor for S2 provided by the official website can use S2 itself for atmospheric and surface signals to retrieve glacier albedo with the aid of digital elevation model (DEM).

Lines 115-118 in the introduction section:

The purpose of this research is to introduce images of the newly launched advanced S2 satellite into the previously published mountain radiation scheme, replacing MODIS atmospheric products and Landsat TM imagery, to obtain the DSSR received by a valley glacier surface with high temporal and spatial resolution.

Lines 165-174 in the methods section:

Based on the DEM, the Li mountainous spectral radiation scheme fully considers the effects of terrain factors, such as the slope, aspect, sky view factor, and obstruction coefficient, on solar radiation and calculates the surrounding-reflected radiation by combining Landsat TM images. The Yang model proposes a general formula for estimating atmospheric broadband transmittance. Nevertheless, our published DSSR estimation method can accurately calculate the DSSR received by the slope surface by introducing MODIS aerosol optical depth (AOD) and PW products and Landsat TM images as input parameters. However, in this study, the MODIS atmospheric products and Landsat imagery were replaced by S2 products in estimating the DSSR of a valley glacier for the following three reasons.

Lines 519-528 in the conclusions section:

“reliable estimates of the shortwave solar radiation” how to understand this sentence?

Reply: Thanks for question. This sentence may not clearly describe the purpose of this article, we have modified it (lines 115-118).

The study area

To test the new approach, you need more than one site for investigation.

Reply: You are an expert in model research and estimation. For a model method, if we can conduct experiments on different glaciers, it will be beneficial to compare and improve the estimation robust results. That is what we want to achieve. However, due to the limitation of glacier observation conditions, there are fewer large-scale glaciers with radiation observation instruments in the Qilian Mountains.

Thank you for your constructive suggestion. In the future, we hope that we can cooperate with other glacier observatory researchers, thus can improve the accuracy of DSSR received by valley glacier surface.

Data sources

The authors use three different data sources to evaluate different quantities. Not clear what is data management and data validity.

Reply: Thank you for your suggestion. There are three different data sources in this article. Since DEM datasets and Sentinel-2 A/B Level-1C products are all from their corresponding official websites, data management and data validity are widely known. However, the pyranometer datasets observed from Laohugou Glacier No. 12 need to be supplemented with an introduction to data processing and quality evaluation in lines 145-151.

Furthermore, we have added a reference, which had described in detail the methods of solar radiation data processing for pyranometer observations of Laohugou Glacier.

AWS1 and AWS2 were equipped with Kipp and Zonen CNR1 four-component net radiometers and were connected to a data collector (CR1000, Campbell Scientific Inc., USA) with a low temperature resistance (–55. °C), which acquired observations every 30 min by taking the average every 10 seconds. Notably, due to snow cover and rime formation on the pyranometer sensor and the low solar altitude angle, the values of DSSR measurements can be distorted. The detailed data processing and quality control methods of field observations are provided in the literature [37].

Sun W , Qin X , Du W , et al. Ablation modeling and surface energy budget in the ablation zone of Laohugou glacier No. 12, western Qilian mountains, China. Annals of Glaciology, 2014, 55(66):111-120.
Methods

“This paper adopted the ….” Why and what reason to adopt methods, where is your contribution to science.

Reply: Thank you for reminding. Based on Li mountainous spectral radiation scheme and the Yang radiation model, we had developed an integrated algorithm, which can estimate the spatial distribution of solar radiation components illuminated on a sloped grid based on DEM, MODIS and Landsat TM images. To more clearly present the advantages of our published mountain radiation scheme, more explanation has been added in lines 165-173 of the revisions:

6. What is a purpose of formula 1? I think page 6 is not useful at all.

Reply: In mountainous areas, the global radiation DSSR received on the slope surface includes four components: direct, diffuse, isotropic, anisotropic, and reflected radiation, which are affected by different terrain factors. Therefore, in order to accurately estimate solar radiation DSSR in the estimation model, these four components of each pixel are estimated separately.

In addition, the input parameters of the mountain radiation estimation model are critical to the calculation results of DSSR. The purpose of page 6 of the original article is to introduce how to retrieve the atmospheric and surface albedo input parameters from Sentinel-2 products.

The reason for your misunderstanding is that we did not elaborate on this content, so we have modified the article, such as:

In line 188, “Different solar radiation components are affected by different terrain factors.”
In line 198, “there are four input parameters for the instantaneous DSSR estimation model”
In lines 205-207, “Therefore, extracting the atmospheric factors and glacier albedo from S2 satellite data is the key to DSSR estimation in a valley glacier. The following subsection will introduce in detail how to retrieve the PW and aerosol concentrations and the ice/snow albedo from S2A/B imagery.”

Concern about data using for validation I think this is too little.

Reply: In this paper, the spatiotemporal DSSR variations on the glacier of a mass-balance year are calculated using 62 typical Sentinel-2 scenes (less cloud cover) during from September 2017 to August 2018. However, as above mentioned, due to the limitation of mountain glacier observation conditions, it is extremely difficult to obtain glacial surface observation data, and it is not possible to obtain all field observations matching satellite observations.

There are two reasons: One is the failure of the instrument. Because observation data cannot be obtained by wireless transmission in Laohugou Glacier No. 12, people must go to the glacier regularly to collect data. Therefore, the problem of instrument failure is not easy to find, and data loss usually occurs in unobservable seasons such as winter. Second, due to the computational complexity of solar radiation in mountainous areas, we only estimated solar radiation for clear-sky conditions. However, clear-sky days are very scarce in mountainous regions, which greatly reduces the validation data.

However, it is worth mentioning that the radiation field observations are all obtained in the summer season during the field investigation period when the glacier changes very intensively.

English has to be improved as well as all abbreviations and acronyms must be disclosed.

Reply: Thank you for pointing out deficiencies in English language usage in our manuscript. We have rewritten some sentence, and the paper has been sent to AJE, a professional language editing service. The paper was edited again after the revisions were completed.

Thank you for your suggestion. We have added a table of abbreviations, which records all abbreviations and acronyms. And the details are provided below (lines 33-34):

Abbreviations
DSSR	downward surface shortwave radiation (W m^-2)
MODIS	Moderate Resolution Imaging Spectroradiometer
TM	Landsat Thematic Mapper
ESA	European Space Agency
S2 A/B	Sentinel-2A and Sentinel-2B
MSI	Multispectral Imager
DLR	German Aerospace Center
WMO	World Meteorological Organization
UTM/WGS84	Universal Transverse Mercator/World Geodetic System
AWS	automatic weather station
S2	Sentinel-2
VNIR	visible and near-infrared
SWIR	shortwave infrared
AOD	aerosol optical depth
PW	precipitable water(cm)
TOA	top of the atmosphere
BOA	bottom of atmosphere
BRDF	bidirectional reflectance distribution function
GIPP	ground image processing parameter
DDV	dense dark vegetation
APDA	atmospheric pre-corrected differential absorption
MBE	mean bias error
MBE%	MBE percentage
RMSD	root-mean-square difference
RMSD %	RMSD percentage
	surface global irradiance (W m^-2)
	direct irradiance (W m^-2)
	diffuse irradiance (W m^-2)
	isotropic diffuse irradiance (W m^-2)
	anisotropic diffuse irradiance (W m^-2)
	surrounding-reflected irradiance (W m^-2)
	cosine of solar illumination angle on a sloped grid
	obstruction coefficient
	sky view factor
	topographic configuration factor

Overall, the paper must be improved by showing novelty at each step of the research.

Reply: Thank you for giving us the opportunity to comment on the novelty of this paper. We have modified almost every part of the article, including the abstract, introduction, methods, and conclusion (the revisions are marked with red text in the paper); we hope that you find this article acceptable in its revised state.

Author Response File: Author Response.docx

Reviewer 2 Report

The main objective of the paper seems to have been achieved, that is, to obtain reliable estimates of the shortwave solar radiation received by the surfaces of mountain glaciers by entering S2 data instead of MODIS atmospheric products. It is an important advance in the field.

The article is very well organized and methodologically well presented.

Although specific references are provided, I suggest a more detailed explanation of geometric problems with regard to terrain modeling, solar zenith distance, view-factors, etc.

Author Response

Reviewer 2’s Comments and Our Responses

Thank you very much for your evaluation. Thank you for your constructive suggestions. The comments have been copied below and are followed by an explanation of how they were addressed in the resubmitted manuscript. The original reviewers’ comments appear in black; our responses follow in blue. The modifications are marked with red text in the paper.

The main objective of the paper seems to have been achieved, that is, to obtain reliable estimates of the shortwave solar radiation received by the surfaces of mountain glaciers by entering S2 data instead of MODIS atmospheric products. It is an important advance in the field. The article is very well organized and methodologically well presented.

Reply: Thank you very much for your approval.

Although specific references are provided, I suggest a more detailed explanation of geometric problems with regard to terrain modeling, solar zenith distance, view-factors, etc.

Reply: Thank you for your constructive suggestions. We have made appropriate revisions based on your opinions; we hope that you find this article acceptable in its revised state.

We have added a more detail explanation of the estimation modeling and terrain factors in the introduction section and methods section. For example,

In lines 58-62:

the cosine of the solar illumination angle (cos i_s) on a sloped grid affected by the terrain slope and aspect, solar zenith and azimuth angles; the obstruction coefficient (V_s), which indicates whether the surface is obscured by the surrounding terrain; the sky view factor (V_iso), which describes the area ratio of the sky dome seen by the slope surface; and the topographic configuration factor (F_ij), which describes the energy ratio of the surrounding visible pixels to the object pixel.

In lines 165-173:

Furthermore, we have added a table of abbreviations, which records all terrain factors and corresponding explanation, shown as below (lines 33-34):

Author Response File: Author Response.docx

Reviewer 3 Report

Level of English language usage: Authors must check the manuscript from native English speaker.

Abstract: Poorly written with no significant novelty or analysis.

Introduction: Detailed but repetitive. Authors may consider making a table showing results from past studies and highlighting their limitations. At the moment, there are paragraphs after paragraphs.

Study area: Figure 1 is not appropriate. Impossible to imagine the location on a broader perspective. The study area is not described well, authors must explain the topography, weather, precipitation and other details about the study area.

Methods: Described well but there is no novelty.

Results and discussion: Very well explained.

Suggestion: This study is limited by two facts, which are my main criticism; (1) this study is carried out on only one glacier, which does not provide any robust results (2) methods used in the paper are not novel (3) authors have not designed new method but they only changed the sensor.

Author Response

Reviewer 3’s Comments and Our Responses

We sincerely thank your detailed, valuable, and insightful comments. The comments have been copied below and are followed by an explanation of how they were addressed in the resubmitted manuscript. Your original comments appear in black; our responses follow in blue. The modifications are marked with red text in the paper.

Level of English language usage: Authors must check the manuscript from native English speaker.

Abstract: Poorly written with no significant novelty or analysis.

Reply: Thank you for your instruction. We have modified the abstract section (lines 18-30) to illustrate the purpose and novelty of thesis, and supplement the results analysis.

Furthermore, an estimation model based on optical satellite images must simultaneously consider terrain and atmospheric effects and the transient nature of ice/snow albedo. Based on a high-resolution (12 m) DEM, in this paper, the newly launched Sentinel-2 satellites, rather than MODIS and TM, were used to provide input data for our published mountain radiation scheme in a valley glacier. Considering Laohugou Glacier No. 12 as the study area, 62 typical Sentinel-2 scenes were selected, and spatiotemporal DSSR variations on the glacier surface were obtained with a 10 m spatial resolution during a mass-balance year from September 2017 to August 2018. Ground-based measurements on 52 clear-sky days were used for validation, and the mean bias error (MBE=-16.0 W/m²) and root-mean-square difference (RMSD=73.6 W/m²) were relatively low. The results confirm that DSSR is affected mainly by the solar zenith angle and atmospheric attenuation in flat areas of valley glaciers, while in areas with complex terrain, the DSSR received by the glacier surface is affected primarily by the terrain and ice/snow albedo, which exhibits very high spatial heterogeneity.

Introduction: Detailed but repetitive. Authors may consider making a table showing results from past studies and highlighting their limitations. At the moment, there are paragraphs after paragraphs.

Reply: We are very impressed by your detailed review. Your suggestion may come from our confusing descriptions in the introduction section. Because there are already 5 tables in this article, we have revised it from two aspects: one is to adjust the organizational structure of the introduction section (paragraphs 2, 3, 4, and 5 in the original manuscript); the other is to add or modify descriptive statements about the goal of the research, which are described as follows:

In lines 81-84:

Satellite remote sensing has become an effective means for achieving glacial monitoring without the need to visit the site [28-29]. It is challenging to obtain field measurements of glacial DSSR because glaciers are poorly accessible due to their adverse natural conditions, and the observation instruments used to monitor glacial melting and movement are unstable. Two methods based on remote sensing data are used to estimate DSSR:

In lines 106-114:

Compared with MODIS and Landsat TM, the use of Sentinel-2 (S2) satellite data to estimate valley glacier DSSR has four advantages: (1) S2 can provide both atmospheric and surface information, thus reducing the errors caused by different satellite data; (2) S2 data can improve the DSSR spatiotemporal resolution and provide reliable information for understanding the melting of glaciers; (3) S2 data can almost completely avoid the saturation defect of the snow signal from optical satellites because of its high radiation resolution; (4) the data processing software Sen2Cor for S2 provided by the official website can use S2 itself for atmospheric and surface signals to retrieve glacier albedo with the aid of digital elevation model (DEM). Therefore, S2 has become a significant source of information for glacier research [29, 33-34].

In lines 115-118:

The purpose of this research is to introduce images of the newly launched advanced S2 satellite into the previously published mountain radiation scheme, replacing MODIS atmospheric products and Landsat TM imagery, to obtain the DSSR received by a valley glacier surface with high temporal and spatial resolution.

Study area: Figure 1 is not appropriate. Impossible to imagine the location on a broader perspective. The study area is not described well, authors must explain the topography, weather, precipitation and other details about the study area.

Reply: Thank you for your advice. We have redrawn Figure 1 by taking the Tibetan Plateau as a reference area in the revised paper, which can imagine the location on a broader perspective.

According to your suggestion, we have rewritten the description of the study area (lines 127-132), and added one reference.

Laohugou Glacier No. 12 (96.5°N, 39.5°E) is located on the northwest edge of the Tibetan Plateau and faces a northern slope with 9.85 km long, with an area of approximately 20.4 km², which is an important water source for the Shulehe River basin. It is the largest valley glacier located within the Qilian Mountains and is the most typical continental glacier in the region [6, 35]. Its mean annual air temperature is more than 0°C, and major precipitation occurs in summer. In recent years, Laohugou Glacier has shrunk due to increases in the warmth and moisture of the regional climate [36]. The terrain of Laohugou Glacier No. 12 is complicated, especially in the southern accumulation zone, with glacier altitude differences greater than 1200 m (from 4202 m to 5427 m a.s.l.) according to the latest TanDEM-X DEM dataset (spatial resolution of 12 m) provided by the German Aerospace Center (DLR) (Figure 1).

Du W, Qin X, Liu Y and Wang X (2008) Variation of the Laohugou Glacier No.12 in the Qilian Mountains during 1958–2005. J. Glaciol. Geocryol., 30(3), 373–379.

Figure 1. Location and terrain of Laohugou Glacier No. 12.

Methods: Described well but there is no novelty.

Reply: Thank you for giving us the opportunity to comment on the novelty of this paper. We think that your critical objections for this paper may be due to the inaccurate description in our manuscript. We have added descriptive statements about the goal of the research and the novelty of this paper in the methods, introduction and conclusions sections.

Lines 163-174 in the methods section:

This paper adopted the mountain radiative transfer scheme from Zhang et al. [30] that combines the Li Mountain radiation algorithm [23] with the Yang broadband atmospheric attenuation model [16]. Based on the DEM, the Li mountainous spectral radiation scheme fully considers the effects of terrain factors, such as the slope, aspect, sky view factor, and obstruction coefficient, on solar radiation and calculates the surrounding-reflected radiation by combining Landsat TM images. The Yang model proposes a general formula for estimating atmospheric broadband transmittance. Nevertheless, our published DSSR estimation method can accurately calculate the DSSR received by the slope surface by introducing MODIS aerosol optical depth (AOD) and PW products and Landsat TM images as input parameters. However, in this study, the MODIS atmospheric products and Landsat imagery were replaced by S2 products in estimating the DSSR of a valley glacier for the following three reasons.

Lines 106-114 in the introduction section:

Lines 519-528 in the conclusions section:

The main contributions of this article lie in two aspects. One is to introduce the newly launched S2A/B satellite data into a previously published mountain radiation scheme, which is more conducive to the DSSR estimation of mountain glacier surface. The other is to modify the default parameter settings of Sen2Cor toolbox provided by ESA to obtain the improved L2A products than the standard products, thus reliable albedo of ice/snow cover can be retrieved. This study explored the seasonal cycle of solar irradiance in a mass-balance year. The verification results confirm that this mountain radiation scheme with S2 data as the model input parameters can provide reliable solar radiation information for mountain glacier surfaces and is not dependent on ground-based observations; thus, this approach is suitable for estimating the incoming solar irradiance of glaciers in remote and difficult-to-reach areas.

6. Results and discussion: Very well explained.

Reply: Thank you very much for your approval.

Suggestion: This study is limited by two facts, which are my main criticism; (1) this study is carried out on only one glacier, which does not provide any robust results (2) methods used in the paper are not novel (3) authors have not designed new method but they only changed the sensor.

Reply: We apologize for your main criticism. We will explain in detail your three criticisms. We hope we can answer your concerns.

For a model method, if we can conduct experiments on different glaciers, it will be beneficial to compare and improve the estimation robust results. That is what we want to achieve. However, due to the limitation of glacier observation conditions, there are fewer large-scale glaciers with radiation observation instruments in the Qilian Mountains. Thank you for your constructive suggestion. In the future, we hope that we can cooperate with other glacier observatory researchers, thus can improve the accuracy of DSSR received by valley glacier surface.
We have added descriptive statements about the novelty of this paper in the methods, introduction and conclusions sections (above mentioned in Reply 5).
In fact, the main contributions of this article lie in two aspects. The first is to introduce the newly launched S2A/B satellite data into a previously published mountain radiation scheme, which is more conducive to the DSSR estimation of mountain glacier surface. The second is to modify the default parameter settings of Sen2Cor toolbox provided by ESA to obtain the improved L2A products than the standard products, thus reliable albedo of ice/snow cover can be retrieved.

Author Response File: Author Response.docx

Article Menu

Estimation of Shortwave Solar Radiation on Clear-Sky Days for a Valley Glacier with Sentinel-2 Time Series

Further Information

Guidelines

MDPI Initiatives

Follow MDPI