PS-InSAR Based Monitoring of Land Subsidence by Groundwater Extraction for Lahore Metropolitan City, Pakistan

Round 1

Reviewer 1 Report

The paper is very interesting and well written and can be accepted for publication.

Author Response

The paper is very interesting and well written and can be accepted for publication.

Reviewer 2 Report

The resubmitted version of the paper "PS-InSAR Based Monitoring of Land Subsidence by Groundwater Extraction for Lahore Metropolitan City, Pakistan" by Muhammad Afaq Hussain and collaborators (manuscript no. remotesensing-1866487, formerly: remotesensing-1787618) contains changes that follow my comments (most of them; some are consistently ignored - I can live with that, but I still believe the presentation of the results would have benefited from those changes). Therefore now I'm leaning towards accepting the manuscript for publication. The minor issues remaining are listed below.

(1)
Abstract, lines 22-23:
Concerning numbers, subsidence -109.76 mm and uplift 21.32 mm: this is the 3rd time I ask the authors not to give so many decimal digits because the measurements are clearly not that accurate! Now it is stated in the paper that the measurement errors are estimated as 2-3 mm (see line 72). It is therefore absolutely meaningless to give numbers that suggest 0.01 mm (!) level of accuracy. This is misleading. Please round the numbers to the first *significant* digit, in this case to -110 mm and +21 mm. Also, do it so throughout the entire paper when the decimal digits are physically meaningless.

(2)
line 187:
change SARPROZ to Sarproz (to be consistent with the rest of the manuscript)

Author Response

"Please see the attachment."

Author Response File: Author Response.docx

This manuscript is a resubmission of an earlier submission. The following is a list of the peer review reports and author responses from that submission.

Round 1

Reviewer 1 Report

Dear Authors, the paper is interesting and well written.

Only minor comments from my side.

Introduction: this section is confusing and should be written in a more precise and direct way. There are couples a references that are put without a specific order.

Big attention should be put instead to the existing literature on the assessment of the effect of the landslides on the urban context (e.g. see the following ones) and in particular on the structural damages:

D. Infante, D. Di Martire, P. Confuorto, S. Tessitore, R. Tòmas R, D. Calcaterra, M. Ramondini, Assessment of building behavior in slow-moving landslide-affected areas through DInSAR data and structural analysis. Engineering Structures, 199, 109638, 2019.

M. Del Soldato, L. Solari, F. Poggi, F. Raspini, R. Tomás, R. Fanti, N. Casagli, Landslide-Induced Damage Probability Estimation Coupling InSAR and Field Survey Data by Fragility Curves. Remote Sensing, 11(12), 1486, 2019.

A. Miano, A. Mele, D. Calcaterra, D. Di Martire, D. Infante, A. Prota, M. Ramondini, The use of satellite data to support the structural health monitoring in areas affected by slow-moving landslides: a potential application to reinforced concrete buildings. Structural Health Monitoring, 1475921720983232, 2021.

Talledo, D. A., Miano, A., Bonano, M., Di Carlo, F., Lanari, R., Manunta, M., ... & Stella, A. (2022). Satellite radar interferometry: Potential and limitations for structural assessment and monitoring. Journal of Building Engineering, 46, 103756.

Peduto D, Pisciotta G, Nicodemo G, Arena L, Ferlisi S, Gullà G, Reale D. A procedure for the analysis of building vulnerability to slow-moving landslides. In: 1st IMEKO International Workshop on Metrology for Geotechnics, Benevento, Italy, March 17–18, 2016.

Tofani V, Raspini F, Catani F, et al. Persistent scatterer interferometry (PSI) technique for landslide characterization and monitoring. Remote Sens 2013; 5: 1045–1065. 

Methodology: there is no clear information on the accuracy and precision of the measurements. Please comment on the possible errors.

Author Response

Please see the attachment.

Author Response File: Author Response.docx

Reviewer 2 Report

The paper "PS-InSAR Based Monitoring Land Subsidence by Groundwater Extraction for Lahore Metropolitan City, Pakistan" by Muhammad Afaq Hussain and collaborators (manuscript no. remotesensing-1787618) presents a PS-InSAR analysis of 2 years of Sentinel-1 SAR data (2020-2021) to show that the city of Lahore is characterized by an overall land subsidence, of which the amount varies spatially, the highest subsidence rates of ~100 mm/yr being measured in the city center area.

The manuscript starts with a fair introduction about the importance of land subsidence measurements. Then it describes the study area and the analysis technique. Then the results are presented, followed by a discussion of the possible causes of the land motion. Apart from some considerable weaknesses in presenting the results, my main concern with this study is the interpretation. It is largely based on qualitative assessments and the "belief" (see line 272) of the authors that the explanation is groundwater extraction. While this may well be the truth, there is little scientific evidence in this area to support the conclusion. The only apparently supporting information is the map of tube well locations (Fig. 12; by the way, the red dots are barely visible in the map). However, there is no data about the actual water production levels/capacities which might be vastly different for different wells. On the other hand, data about precipitation are presented (Fig. 14) but their relation to subsidence rates is unclear since the latter seem to be quite constant.

An immediate question from the potential reader is why only 2 years are covered in the analysis? (More exactly, overlapping ~1.5 yr with ascending and descending track data, respectively.) The Sentinel-1 mission lifetime is longer. This limited scope has at least to be explained in the paper, and perspectives given for the time extension of the PS-InSAR study.

I do not recommend publication of this manuscript until a major revision is made. My comments and suggestions are given below in details.

Title:

add the word "of":
PS-InSAR Based Monitoring [of] Land Subsidence by Groundwater Extraction for Lahore Metropolitan City, Pakistan

Abstract:

line 20:
Strictly speaking, the PS-InSAR technique is not based on radar "images" but the detection of phases. So it would be more appropriate to use the term "Sentinel-1 data" instead of "Sentinel-1 images" (also elsewhere throughout the text)

line 22-23:
"high subsidence rate of -109.76 mm" ... "low subsidence rate of 21.32 mm"
I have a couple of problems here:
(1) Do you mean subsidence in both cases? If so, why the signs are different (- and +)?
(2) The unit of mm is a distance (amount of subsidence), not a subsidence rate which should be measured in mm/yr. Actually, this confusion is also present elsewhere in the paper. Look at e.g. Fig. 9. The time span in the horizontal axis is ~2 yr. The vertical axis is subsidence (in mm) starting with 0 at t=0. The subsidence rate (in mm/yr) is given in the top right corner of the graphs. For example, look at point (a): how can ~100 mm subsidence over a 2-yr period result in a subsidence rate of ~100 mm/yr? The same question applies to all other permanent scatterers.
(3) Every kind of physical measurements should have an uncertainty. I wonder what can be the error of the subsidence rate estimation? It is not discussed anywhere in the paper (it must have been!). However, giving the subsidence rate values with a precision of 0.01 mm/yr (i.e. 10 micrometers/yr) is obviously unphysical since the technique is not capable of providing that level of accuracy. Please, give only decimal digits that are significant.

Paper body:

line 81-83:
The sentence starting with "Because..." is unclear, please rephrase.

line 97:
abbreviations, like LOS here, should be explained at their first occurrence: line of sight (LOS)

Sect. 2.3 (data analysis)

It is not clear whether any of the permanent scatterers located within the study area were regarded as a "stable" reference. If not, how can be the subsidence rates determined, where every PS may be moving?

line 125:
[A]laska [S]atellite [F]acility

line 153-154:
Sentinel-1 SAR satellites do not make "photos" or "photographs", please rephrase.

line 164:
SAPROZ or Saproz? Spell the name consistently throughout the paper.

Sect. 3 (results)

Fig. 9
See my earlier concerns about incompatible subsidence numbers (vertical axis) versus subsidence rate (top right corner). Also, it is unclear what the continuous curves mean in Fig. 9. The best possible time sampling with Sentinel-1 is 6 days, so we must see discrete measurement points. In addition, I presume the subsidence rates are obtained by linear regression (i.e. fitting a line to the discrete measurement points). Ideally, that line should also be indicated along with the individual measurement points. The linear regression should also provide an uncertainty of the slope (the subsidence rate) originating from the scatter of the measurement points.

Fig. 11
On the horizontal axis, apparently the serial number of PS points is given. As I guess these are not equidistant in reality, it would be much more useful to plot the actual physical distance along the N-S and E-W profiles.

around line 250:
Subsidence (in mm) and subsidence rate (mmy/yr) is mixed here, too.

Sect. 4 (Discussion)

line 260-264:
The starting sentences:
"This study found that various elements are responsible for ground subsidence in Pakistan’s largest metropolis, Lahore. Contributing factors include groundwater removal to fulfill human demands, natural consolidation of quaternary alluvium, and inappropriate building construction."
These statements are simply not true! What this study found so far is that there is overall land subsidence in the Lahore region over a time period of ~2 yr in 2020-2021. The rest of the discussion is essentially based on assumptions taken from the literature. I'm not saying these assumptions are necessarily wrong. I'm just saying that these should be presented as such, and not as the findings/result of this study. There is no supporting scientific evidence given here on why the groundwater extraction is responsible for the effect.

Fig. 14.
The horizontal axis is time, spanning 1 yr range. The columns indicate average precipitation in 2020 and 2021. On the other hand, the graphs (adopted from Fig. 9) show subsidence, but over ~2 yr time span. How these graphs can be put together with the same horizontal axis scale?

Sect. 5 (Conclusions)
It must clearly distinguish between the actual new scientific results presented in the paper, and the assumptions made in the discussion. I also suggest to give some perspective on the continuation of this research: how can the causal connection between subsidence and groundwater level possibly be verified in a quantitative way? What do we expect on time scales longer than 2 yr? Etc.

Author Response

Please see the attachment.

Author Response File: Author Response.docx

Round 2

Reviewer 2 Report

In the revised version of the paper "PS-InSAR Based Monitoring of Land Subsidence by Groundwater Extraction for Lahore Metropolitan City, Pakistan" by Muhammad Afaq Hussain and collaborators (manuscript no. remotesensing-1787618), the authors fixed many of the issues brought up in my previous report. However, some of them are still there. Maybe I was not clear enough so I repeat them here with additional explanation.

Abstract, line 22-23:

(2) The unit of mm is a distance (amount of subsidence), not a subsidence rate which should be measured in mm/yr. Actually, this confusion is also present elsewhere in the paper. Look at e.g. Fig. 9. The time span in the horizontal axis is ~2 yr. The vertical axis is subsidence (in mm) starting with 0 at t=0. The subsidence rate (in mm/yr) is given in the top right corner of the graphs. For example, look at point (a): how can ~100 mm subsidence over a 2-yr period result in a subsidence rate of ~100 mm/yr? The same question applies to all other permanent scatterers.

Response:
(2) Please check figure 9.

Referee:
Yes, I do see Fig. 9, and the problem persists. On panel (a), the subsidence is about -100 mm, as you can see from the vertical scale. The time span in the horizontal axis is from January 2020 (although the reader can only guess it because the labeling does not start at the beginning) to December 2021. From Jan 2020 to Dec 2021 it is 2 years. So the subsidence rate is about -100 mm/(2 yr) = 50 mm/yr. Do I misunderstand something, or Fig. 9 is wrong? The same applies to all sub-panels.  

(3) Every kind of physical measurements should have an uncertainty. I wonder what can be the error of the subsidence rate estimation? It is not discussed anywhere in the paper (it must have been!). However, giving the subsidence rate values with a precision of 0.01 mm/yr (i.e. 10 micrometers/yr) is obviously unphysical since the technique is not capable of providing that level of accuracy. Please, give only decimal digits that are significant.

Response:
(3) Thanks for pointing out, our study do have uncertainty of 2 – 3 mm, please see line 71 to 72.

Referee:
OK, this is a general number for the PS-InSAR technique. But the statement in lines 71-72 refers to *deformations*, measured in mm. On the other hand, here you give *deformation rates*, in mm/yr. What are the uncertainties of subsidence rates you measure? Are they as small as ~0.01 mm/yr? If not, you should avoid displaying decimal digits of 0.01 mm/yr. In any case, without clearly stating the errors in subsidence rate, the values determined cannot be interpreted correctly. The formal errors on subsidence rates must be given in the paper.

---

Sect. 2.3 (data analysis)

It is not clear whether any of the permanent scatterers located within the study area were regarded as a "stable" reference. If not, how can be the subsidence rates determined, where every PS may be moving?

Response: It is not permanent scatterers; it is persistent scatterers (PS). We choose one reference point (Ground control point (GCP)) in ascending and descending tracks which is stable on ground (please see fig 2). With respect to this reference point, we can analyze the movement of other PS points.

Referee:
OK, but why don't you describe it in the paper? Please do it, this is relevant information for the reader.

---

Sect. 3 (results)

Fig. 9
See my earlier concerns about incompatible subsidence numbers (vertical axis) versus subsidence rate (top right corner). Also, it is unclear what the continuous curves mean in Fig. 9. The best possible time sampling with Sentinel-1 is 6 days, so we must see discrete measurement points. In addition, I presume the subsidence rates are obtained by linear regression (i.e. fitting a line to the discrete measurement points). Ideally, that line should also be indicated along with the individual measurement points. The linear regression should also provide an uncertainty of the slope (the subsidence rate) originating from the scatter of the measurement points.

Response: We changed it according to your suggestion. Please check fig 9.

Referee:
I check it but unfortunately I see no change: Fig. 9 still displays a continuous curve and not a scatter plot with measurement points. Also the subsidence rates (i.e. amount of subsidence divided by the length of the measurement time interval) seem incorrect - see my first comment above.

---

Fig. 11
On the horizontal axis, apparently the serial number of PS points is given. As I guess these are not equidistant in reality, it would be much more useful to plot the actual physical distance along the N-S and E-W profiles.

Response: We tried to improve the figure. Please check figure 11.

Referee:
Maybe there is a technical problem with the revised pdf version, but I don't see any change in this figure either.

---

Sect. 4 (Discussion)
    
line 260-264:
The starting sentences:
"This study found that various elements are responsible for ground subsidence in Pakistan’s largest metropolis, Lahore. Contributing factors include groundwater removal to fulfill human demands, natural consolidation of quaternary alluvium, and inappropriate building construction."
These statements are simply not true! What this study found so far is that there is overall land subsidence in the Lahore region over a time period of ~2 yr in 2020-2021. The rest of the discussion is essentially based on assumptions taken from the literature. I'm not saying these assumptions are necessarily wrong. I'm just saying that these should be presented as such, and not as the findings/result of this study. There is no supporting scientific evidence given here on why the groundwater extraction is responsible for the effect.

Response: There is scientific evidence available in manuscript related to ground water removal/extraction with references. Please check line number 274 and 281.

Referee:
Indeed, there are references cited in the lines given that report connection between subsidence and groundwater extraction. But those are valid for other areas and not specifically the Lahore metropolitan area which, on the other hand, is the subject of this present study. For the Lahore area, an independent scientific evidence would be e.g. measurement data related to water extraction levels in the study period. Apparently there are no such data available.

Therefore the logic of the discussion should be reversed:
- First of all, give clearly the *original results* of this study, which is the detection of land subsidence in this area using PS-InSAR technique- nothing else.
- Then cite the relevant literature, indicating that *other studies in the literature for other geographic areas* found connection between land subsidence and other factors.
- Finally state that you *assume* this connection also holds for the Lahore area. This of course makes sense. But please do not give the illusion that you measured something which proves this connection here, as this is an assumption only (even though a reasonable one), based on the literature.   

---

Fig. 14.
The horizontal axis is time, spanning 1 yr range. The columns indicate average precipitation in 2020 and 2021. On the other hand, the graphs (adopted from Fig. 9) show subsidence, but over ~2 yr time span. How these graphs can be put together with the same horizontal axis scale?

Response: We tried to improve the figure. Please check figure 14.

Referee:
Thanks for making the time axis compatible for both graphs. However, now the label of the vertical axis "Average Precipitation (mm) 2020-2020" should be corrected as it is no longer the average of 2 years. I suggest "Monthly Precipitation (mm) in 2020-2021".

---

An additional discrepancy I found in the paper: in Introduction, you say that "This study was conducted to record ground subsidence in Lahore, Pakistan, from 10 January 2020 to 30 December 2021 and 2 January 2020 to 22 December 2021, using descending and ascending tracks." On the other hand, in Sect. 2.3, you say that "This research obtained 41 images from the ascending path number 100 (2 January 2020 - 26 April 2021) and 38 from the descending path number 34 (12 October 2020 - 30 December 2021)". These latter are different time ranges, both of them shorter than the ~2 yr mentioned in the introduction. Which one is true? Please make them consistent.

Finally, a very minor typo:

line 85:
angles that range

Author Response

Please see the attachment.

Author Response File: Author Response.docx