the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 15 Apr 2024
Estimating Bowen Index in urban environment based on Landsat 8/9 imagery
Abstract. This study explores the Olomouc region’s thermal dynamics and energy balance in northeastern Czech Republic during the spring and summer of 2022. Using Landsat 8 and Landsat 9 imagery, meteorological data and the S-SEBI model to determine Bowen Index (BI). Analysis of Solar Net Radiation and Sensible and Latent Heat Fluxes contributes to our understanding of energy exchanges at the surface. The S-SEBI algorithm helps determine Bowen Index, highlighting patterns across land cover classes. Urban and artificial areas consistently exhibit higher Bowen Index values, suggesting a higher sensible heat flux. Areas with vegetation showcase lower BI values, indicating the dominance of latent heat flux and the cooling effect of vegetation. Temporal trends in BI vary during specific sensing days. Despite challenges associated with atmospheric variability, our integration of meteorological data and remote sensing techniques contributes to understanding the local climate. Validation with EUMETSAT evaluation data indicates a consistent relationship with some biases, implying potential reliability of our approach and the derived heat fluxes and Bowen Index values.
- Preprint
(8636 KB) - Metadata XML
- BibTeX
- EndNote
Status: final response (author comments only)
-
CC1: 'Comment on hess-2024-85', X. W. Shen, 30 Apr 2024
In this paper, titled “Estimating Bowen Index in urban environment based on Landsat 8/9 imagery” (hess-2024-85), the authors discussed the use of satellite data (i.e. Landsat 8/9) in estimating BI over an urban area, especially during the spring and summer of 2022. Specifically, in this study, the authors used remotely sensed images with the S-SEBI model to estimate and monitor the temporal trends in BI over the Olomouc region in northeastern Czech Republic. In the reviewer’s opinion, this paper (as current version) did not make important contribution to this field (i.e. in local climate), or did not show clearly readers their contribution, although general findings were obtained.
Major problems:
- For processing data. LST and albedo are critical inputs for heat flux estimation, by using the S-SEBI model. However, the reviewer did not agree with the authors in estimations for “3.2.1 Land Surface Temperature” and “3.2.2 Albedo”. In particular, Eq. (1) is not suitable for LST estimation, in which the atmospheric corrections are not considered. Furthermore, the constants might not be applicable for the TISR of Landsat 8/Landsat 9. For example, the center or effective wavelength is varied among sensors, according to the reviewer’s previous investigations. To estimate emissivity over urban area is another challenge for LST retrieval using Landsat observations. Similarly, Eq. (2) for albedo estimation might be not applicable exactly for the OLI on Landsat 8/Landsat 9. If it is used, the authors should discuss possible or potential uncertainties. Consequently, other errors in inputs and the possible uncertainties should also be considered, especially for the simplification of S-SEBI in “3.2.4 Determining Heat Fluxes using S-SEBI”.
By the way, many procedures for processing Landsat data could not be applicable. The authors should reference the data provider (https://www.usgs.gov/landsat-missions/landsat-science-products), and obtain newly reprocessed data records.
- For the results. Based on the knowledge of the reviewer, in terms of urban climate, findings as shown in this manuscript are not profound. The authors need to show readers and reviewer the contributions to this field, as well as the significant or visible findings compared against other previous investigations. Actually, according to the estimated results, obvious differences between heat flux estimation and EUMETSAT were observed (Tables 6-8).
- The major concern is that what important information can readers obtain from this investigation. The authors should demonstrate and make proved explanation.
Other problems:
- In “Introduction” (about lines15-20), the authors mentioned the searching results considering “Bowen Index” and “Remote Sensing” within different databases. However, details on these results are not shown (e.g., the time range of publications).
- Figure 1 is not suitable and informative. Please give a refined picture containing geoinformation about the area. If it is possible. The Figure 1 and Figure 2 should be combined into one graph.
- As mentioned in p.3, the meteorological data downloaded from CHMI were used. Please show the location of these meteorological stations.
- Several words (or numbers) make confusion, such as “(6)” and “7” in line 122 (see “4 Results”).
Wish the comments above are valuable.
Citation: https://doi.org/10.5194/hess-2024-85-CC1 -
AC2: 'Reply on CC1', Tereza Pohanková, 22 May 2024
Dear Commenter,
Thank you for your thorough review of our paper. Your comments raise several important points that will strengthen our manuscript.
We appreciate your emphasis on clearly articulating the specific contributions to urban climate research. We will revisit the manuscript to ensure the significance, particularly in comparison to previous investigations. Finally, your recommendations regarding figure clarity, meteorological station location details, and clearer wording are valuable.
Thank you again for your insightful comments. We believe by considering these points, we can significantly enhance the quality of our research.
The Authors
Citation: https://doi.org/10.5194/hess-2024-85-AC2
-
RC1: 'Comment on hess-2024-85', Anonymous Referee #1, 01 May 2024
Review:
General summary of review:
The contribution is characterized by poor English, contains incorrect and irrelevant figures and references. Incorrect wording, or naming of definitions and databases occurs, as well as inconsequent reasoning. Furthermore insufficient literature is cited/consulted. At numerous occasions in the text it is unclear what is meant exactly and it is unclear which data is used at which instance in the analysis. At many occasions the text improperly reflects what is shown in the equations, tables and figures and at times derivations are not provided and/or parameters not defined. No critical review of the S-SEBI method is done and reference to other S-SEBI results in urban areas is not provided. No validation versus ground truth is performed, only a coarse comparison versus EUMETSAT SAF output is done which cannot be regarded as a proper validation, also given its coarse resolution.
This review summary is exemplified by several, but not exhaustive, examples below.
Details/Examples:
Line 3: “Bowen Index” is generally known as “Bowen Ratio”; this more appropriate wording should be used.
Line 16: Unclear what is meant by “Stable season”
Line 17 mentions that there are not many studies on Bowen Ration. Whereas “many” is a relative word, the reviewer would think that the literature is swarmed by publications on the Bowen Ratio. The reviewer suspects that the authors may have searched for “Bowen Index”, which is the improper wording for “Bowen Ratio”.
Line 18/19: Database “Word of Science” should probably read as “Web of Science”
Line 19 mentions that the first study on Bowen Ratio is from 1992, whereas the original paper (The Ratio of Heat Losses by Conduction and by Evaporation from any Water Surface, I.S. Bowen, Phys. Rev. 27, 779 – Published 1 June 1926) dates back to 1926.
Line 24-25 states that meteorological methods such as the BREB or Eddy Covariance require a lot of in-situ data, which is a rather strange remark, since these two are measurements.
Line 25 mentions that these measurements are point data, whereas they originate from a certain source area, also known as the footprint area (not a point!) of the observation. In the next line the authors mention that this “enable us for spatial analyzing”, which is completely contradicting.
Line 32 mentions this approach innovative aspect is in its application as a remotely sensed variable, whereas the method the authors use is the S-SEBI approach which is a remote sensing-based approach based particularly on the Bowen Ratio!
Line 39-40: “intricate interplay of environmental factors”; unclear which factors.
Line 43: unclear what is meant with “near-natural”; not defined.
Line 45: Olomouc is reported to have “one hundred citizens”; given tha the city has some 102,000 inhabitant most probably the word “thousand” disappeared.
Line 48: “The climate is affected by its continental conditions”; improper phrasing.
Line 50: “Fig 3” should be replaced by “Fig 2”
Above line 52, Figure 1. As displayed here the location map hardly contains relevant information and is better left out.
Line 56: “Optimal” is not defined
Line 56: Unclear is which criteria are used for selecting the nine days/observations.
Line 57: “L1TP” and “L2TP” are not defined.
Below line 57, Table 1 shows “19th May 2022”, which might be 19th June 2022? When examining figures 6 and 7, the 19th of May is missing, but 19 June is present.
Above line 60, Table 2. Average, minimum and maxim air temperatures are mentioned but it is unclear over which period (hourly, daily, weekly, monthly).
Line 62: “layer of habitats”; unclear what is meant here.
Line 65-66: “habitats smaller than….the satellite pixel”; completely unclear what is meant/done here.
Line 67: Unclear what is meant with the “(2)”.
Line 72: Unclear what is meant with “Water et al”
Line 76-77: The authors explain here first that S-SEBI uses Evaporative Fraction, generally indicated by capital lambda, to determine sensible and latent heat flux, followed by an explanation that they then used sensible and latent heat flux to determine the Bowen Ration. This is a circular reasoning since evaporative fraction, Lambda, is directly related to Bowen Ration, beta, by Beta=(1-Lambda)/Lambda.
Line 82, Eq. 1: LST is calculated incorrectly, using at sensor temperature (so Top Of Atmosphere, TOA) and surface emissivity (so Bottom Of Atmosphere, BOA). No atmospheric correction is performed.
Above line 85, figure 3. The process of Bowen Ratio calculation is incorrectly represented. At several locations in the figures arrows are incorrect; for example “Landsat bands” do not yield “Elevation”.
Line 90, Eq. 2: Surface albedo is calculated, which is used in the S-SEBI method in a combined manner with the temperature (which was in principle determined at TOA); this is an incorrect manner of applying the S-SEBI methodology.
Line 93: Solar net radiation (i.e. net shortwave radiation) is NOT a balance of incoming and outgoing short- and longwave radiation. In addition it is unclear from where (stations, imagery?) this data is obtained.
Line 101-102: Unclear what is meant with “limited pixels”
Line 103: Evaporative fraction is not determined by the three fluxes but only by two; the sensible and latent heat flux.
Line 107, Eq. 4: unclear how fractional vegetation cover is determined.
Line 111: Unclear what is meant with “hot (dry) pixel” and with “cold (wet) pixel”. Authors simply assume readers are familiar with this concept, but this needs additional explanation.
Lines 112-113: Reference should be made to the original paper for such a statement.
Line 117, Eq 8: Bowen ratio is defined as H/LE, which is presented as the author’s manner to derive the turbulent fluxes, but this is implicit to the method used (S-SEBI) which utilizes evaporative fraction.
Line 122: Unclear what is meant with “(6) ”.
Line 121-122: Sensible heat flux is NOT indicative of heat conduction NOR of heat storage; sensible heat flux is a turbulent heat flux. These are completely different phenomena!
Line 122: Latent heat flux is not an indicator of the cooling effect of evapotranspiration; it IS evapotranspiration.
Line 131: Unclear what is meant with “at one point”
Above line 133, fig 6. Dates mentioned do not correspond to the dates mentioned in Table 1.
Line 134-135: Unclear what is meant with the sentence “A contrast can ….the study period”.
Above line 137, fig 7. Dates mentioned do not correspond to the dates mentioned in Table 1.
Line 137: “4.1 Bowen Index”: If this is mentioned under a separate section number, the other results should also be included under a section at a similar level.
Line 142: “22 July 2022” is mentioned as a date with specifically high Bowen Ration, whereas other dates show a clearly higher value; why is this done?
Below line 15, table 4: Improper order of dates; probably due to previously mentioned incorrect dates?
Line 147: Unclear what is meant with “overall varying”.
Line 152: “ofd” should probably be removed?
Line 154: Unclear why these two dates are chosen, no reason provided.
Line 156: Reference is made to 11th of May, which is a date not occurring anywhere in the manuscript.
Line 156: correlation of Bowen Ratio is reported to be negative in table 8, whereas there is no negative value in table 8.
Line 158: direction of bias is reported to be “consistent across all three variables (SHF, LHF, BI)”; this is rather obvious since all three in the S-SEBI are interrelated by definition.
Line 165-166: “identification … from vegetation” is improper reasoning.
Line 168-169: “The findings…local climate studies”; there are no other findings than the observation that over urban areas the Bowen Ratio is higher than over natural landcover, which can hardly be seen as “a finding”, this is rather obvious.
Line 176-177: “our integration of meteorological data”; there is no mentioning of meteorological data in the manuscript.
Line 183-187: It is very hard to detect sense in these lines; completely unclear what is meant here.
Line 205: “during” should probably be replaced by “between”
Line 206: It is unclear where there is an “iterative” part in the “method”used, nor is it clear what is meant with “non-destructive solution”.
Citation: https://doi.org/10.5194/hess-2024-85-RC1 -
AC1: 'Reply on RC1', Tereza Pohanková, 03 May 2024
Dear Referee,
I would like to express my sincere gratitude for your time and insightful feedback on the manuscript. Your corrections regarding English grammar have been invaluable, and I appreciate you pointing out areas where I could improve the precision of the grammar and vocabulary. Your suggestions will be taken in account in resubmission and will surely help to increase the quality of the paper. After resolving all issues the paper will be put into proofreading to ensure quality of the language. Many of your comments relate to the vagueness of the text, this issue will be noted and the next version of the article will address this issue with utmost importance. Resubmitted version will include clear marking on the changes made regarding your comments.To address your comment for Line 82, we implemented corrections following the guidelines set by the USGS for Landsat 8 and Landsat 9 Level 1 Terrain-Corrected (L1TP) products. Additionally, we drew upon the methodology presented in the article by Avdan and Jovanovska (2016) [DOI: 10.1155/2016/1480307] to ensure accurate Land Surface Temperature (LST) retrieval.
As for the validation for this small-scale and local study, we acknowledge the limitations of readily available datasets. Unfortunately, using a complex and demanding technology like an eddy covariance station was beyond the scope of this project. We have done our best to utilize the available resources and methodologies to achieve robust results within the constraints of the study.
Thank you again for taking the time to make these corrections, they are greatly appreciated and we will do our best to incorporate them and improve the next version of the articleCitation: https://doi.org/10.5194/hess-2024-85-AC1
-
AC1: 'Reply on RC1', Tereza Pohanková, 03 May 2024
-
RC2: 'Comment on hess-2024-85', Anonymous Referee #2, 10 May 2024
The article tackles the challenges of estimating the Bowen Index in an urban environment based on satellite data and the S-SEBI model. Although the issue is very relevant, the methods and the results are poorly described and some inconsistencies are present. In some aspects, the work appears incomplete and could use some adjustments to convey its message in a solid way.
I suggest a thorough revision of the English wording throughout the whole article, as the readability is difficult in many points.
Line 11 and in the entire text: you define the Bowen ratio as BI (which I suspect stands for bowen index). However this bowen index is commonly used as Bowen ratio.
This could also alter the results of lines 17-20, as the bowen ratio has been used in much more researches than what you state.
Introduction: this should be improved as remote sensing based methods for bowen ratio are not discussed
Line26: what do you mean with the eddy covariance method that could work on large scale? The previous lines you state that it works only at point scale. Please rephrase it.
Line 32: I don’t think this is innovative, as the bowen ratio as been computed already using Landsat data (only to cite one paper
Sobrino JA, Souza da Rocha N, Skoković D, Suélen Käfer P, López-Urrea R, Jiménez-Muñoz JC, Alves Rolim SB. Evapotranspiration Estimation with the S-SEBI Method from Landsat 8 Data against Lysimeter Measurements at the Barrax Site, Spain. Remote Sensing. 2021; 13(18):3686. https://doi.org/10.3390/rs13183686
And the S-SEBI model is based on the bowen ratio.
Line 55: Please give a definition for “optimal”
Line 67: not clear what (2) refers to
Line 72: please correct “Water et al”
Lines 74-77: I believe that the applied methodology is not correct. The S-SEBI model is based on the evaporative fraction (which is linked to the Bowen ratio). Then you are using the latent and sensible heat flux to compute the bowen ratio, which are actually derived from the same equations.
Line 82: It seems that eq 1 is not considering the atmospheric correction. Which values have been used for LSE?
Figure.3 is unclear. I also believe some arrows are not correct (e.g landsat bands are not providing elevation. Why the shortwave outoging radiation is used into the incoming shortwave one? I believe is the opposite direction, …).
Line 101-102: please define “limited pixels”
Line 107: please explain how the vegetation cover was derived
Line 112: please cite the original S-SEBI paper.
fig 6. The dates are different from Table.1. The figure is poorly formatted.
Line 155 it is not clear why only to images where compared. All the dates should be included in the analyses.
Line 156: in the tables there are no negative correlations. Please explain.
Line 176: it is not clear from the manuscript where the meteorological data have been integrated into the model.
Line 183-187: These lines need to be rephrased, as there is some confusion. I also do not agree that bowen ratio data computed with different methodologies could not be compared.
Citation: https://doi.org/10.5194/hess-2024-85-RC2 -
AC3: 'Reply on RC2', Tereza Pohanková, 21 Jun 2024
Dear Reviewer,
Thank you very much for your review of our manuscript. We appreciate your comments and suggestions, which have helped us improve the clarity and comprehensiveness of our work. We agree with your suggestion and have revised the manuscript. We have also ensured consistent use of "Bowen ratio" throughout the text. The eddy-covariance section was clarifies to that while eddy covariance provides valuable point-scale data, it is not readily applicable for large-scale estimations due to (but not limited to) the costs.
We acknowledge that the S-SEBI model utilizes the Bowen Ratio concept. However, our aim is to explore the potential of a blended approach, combining S-SEBI with additional data sources, for urban environments. A check was done on references and there applicable it was corrected.
You raise a valid point regarding the potential redundancy in calculating the Bowen Ratio from latent and sensible heat fluxes derived by S-SEBI. We are currently exploring alternative approaches to address this and hopefully incorporate the most suitable method from now on.
We apologize that our explanation regarding thermal corrections on the satellite imagery was unclear. We have applied these corrections following JImenez-Munoz (J. C. Jimenez-Munoz, J. Cristobal, J. A. Sobrino, G. Soria, M. Ninyerola and X. Pons, "Revision of the Single-Channel Algorithm for Land Surface Temperature Retrieval From Landsat Thermal-Infrared Data," in IEEE Transactions on Geoscience and Remote Sensing, vol. 47, no. 1, pp. 339-349, Jan. 2009, doi: 10.1109/TGRS.2008.2007125.) .
"Limited pixels" refer to areas within the image with insufficient data to confidently estimate the wet and dry end members for the S-SEBI model. We will provide a clearer explanation of this concept in the revised manuscript. We recognize the inconsistency between the dates in Figure 6 and Table 1. We will rectify this error and ensure all relevant data is included in the analysis. We apologize that negative correlations were not explicitly mentioned in the tables.
We believe that by addressing your valuable feedback, we can significantly improve the quality and clarity of our manuscript. Thank you again for your time and expertise.
Citation: https://doi.org/10.5194/hess-2024-85-AC3
-
AC3: 'Reply on RC2', Tereza Pohanková, 21 Jun 2024
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 289 | 44 | 24 | 357 | 10 | 14 |
- HTML: 289
- PDF: 44
- XML: 24
- Total: 357
- BibTeX: 10
- EndNote: 14
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1