Space-Time Evolutions of Land Subsidence in the Choushui River Alluvial Fan (Taiwan) from Multiple-Sensor Observations

Round 1

Reviewer 1 Report

This is an interesting work in line with the aims of Remote Sensing. The Ms is well organized, clearly and concisely written. Specifically, the introduction is clear and provides a comprehensive review of the pertinent literature. Data and outcomes are original and well presented. The conclusions are supported by data included in the manuscript.

I consider this work a valuable contribution for the scientific community, and highly recommend it for publication after addressing some minor issues.

Fig. 1: check the word ‘leveling line’ positioned outside the legend, just at the top-right side of the figure.

Line 132-133: do the additional leveling surveys refer to the same reference benchmark?

Line 150: show the position of KMNM, or at least provide coordinates.

Line 268: the following locations do not appear in the map: Erlin, Xihu, Xizhou…..etc. I suggest to include them at least in fig. 1.

Line 260-264: I suggest to explain better how the cumulative land subsidence over 1993-2019 was computed. For example: (a) each dataset (e.g., ERS, ENVISAT, etc.) was preliminary gridded on the same grid cell size (100 m) and then all grids were summed considering the years span of each dataset…..

In addition, which dataset was considered for filling the gaps of data (i.e. 2011-2014) and for the superpositon of ENVI and ALOS in the period 2006-2008)?

Fig 5: I suggest to move the names of the Satellites to the top of the SAR displacement maps (below the interpolated leveling maps), or at least inside the SAR maps.

Fig. 5: how were interpolated the ground movements obtained by leveling?

Fig 5 caption: I suggest “Figure 5. Interpolated ???  vertical velocities from precise leveling (upper row) and average SBAS InSAR measurements ……

Line 316: Dacheng, Xigang, etc are difficult to find in the maps.

Line 382: Although explained in this work, the names 'primary' and 'secondary' compaction could be confused with the similar one used in soil mechanics that are different meaning. I suggest to change them, if possible, especially in fig 8.

Fig. 6a caption and right axis: ………… average??? groundwater …….

Fig. 7: the names of the locations are not shown in the map. I suggest to insert them or at least some letters for reference, perhaps in fig 4 or in a new position map inset in fig 7, e.g., shallow and deep compaction.

Fig. 7 caption: Xizhou, Huwei , etc., see the previous  comment.

Line 405: Add a couple of lines for explaining how fig 10 maps were obtained (e.g., by interpolation of … etc..)

Figure 10. Average??? Deep compactions at intervals of 5 mm/yr …

Author Response

Please see the attachment.

Author Response File: Author Response.docx

Reviewer 2 Report

Dear authors, thank you for the paper. The paper discusses monitoring method for ground subsidence, using InSAR and CGPS, levelling results are used for validation.

I have just minor comments to the paper.

• The title and abstract do not contain the information that it was done in Taiwan, and, in addition, it mentions only "multiple sensor" and the methods are not listed there.  The information is only in the keywords.
• page 5, line 173: "In order to overcome the limitations of InSAR technology" - please be more concrete about the limitations
• page 5,  line 207: it is not right that orbits create systematic errors in the InSAR displacement rates (or, it has to be discussed in which sense they are systematic). However, comparing the results "before and after calibration" (and I cannot see any ramps), it seems to me that the difference may simply be caused by inappropriate InSAR reference: StaMPS references the results to the average of all the points, so in your area which is largely unstable, it needs to be re-referenced (simply adding/subtracting a constant, it would not be necessary to adjust to the plane). It would be interesting to include information about the number of CGPS stations and the tilts finally calculated.
• page 2, line 62: "which are using based on temporal"
• page 2, line 66: "were consistent the difference"
• page 5, line 192: "interferograms was generated" -> were
• page 6, line 227: "the horizontal motions of CRAF" - this sentence lacks a verb
• page 6, line 227: the rates you show here are not objectively "relatively small", so please disclose to what (vertical) rates you compare them.
• page 6, line 238: "figure 2 show" -> shows
• page 6, line 244: "root means squared" -> mean
• page 10, line 311: "this difference may be attributed" etc. I am afraid that displacement rates about 200 mm/yr are above the InSAR limit for ERS-1/2 because of sparse temporal sampling.

Author Response

Please see the attachment.

Author Response File: Author Response.docx

Reviewer 3 Report

This manuscript is well written. The authors integrated multi-observational datasets to track long-term land subsidence over time and space. The results from this study would be important for groundwater management in the study area.

I have several minor comments:

• This study used leveling measurements, which are essentially different from GPS data, as well as the InSAR data/measurements. Leveling data are physical measurements----orthometric heights; GPS data are geometric measurements----ellipsoidal heights. Leveling researchers may argue the difference or sameness between the two datasets. Anyway, it is necessary to address the conceptional difference. I recommend a reference for this topic:

Wang, G., and T. Soler. 2014. “Measuring land subsidence using GPS: Ellipsoid height vs. orthometric height.” J. Surv. Eng. 141: 05014004. https://0-doi-org.brum.beds.ac.uk/10.1061/(ASCE)SU.1943-5428.0000137.

• Lines 140-150---GPS data

The description for GPS data processing is very brief. It is OK.  You may consider to add one figure showing the vertical displacement time series of KMNM (your reference) and one or two your stations in CRAF. KMNM (2012-2020) is far away (approximately  xx km) from your study area. Add one sentence to explain why you select KMNM as your reference and briefly assess the long-term stability of KMNM. Reference is always a complex concept in regional subsidence monitoring.  A stable regional reference frame would do a better job than a single reference station for your future subsidence investigations. I read through this article for several times, but can not realize the contribution of GPS to your final products. Continuous GPS measurements are important for assessing your leveling and InSAR measurements.  

• Fig. 4 is great. You may consider to add several plots  showing the integrated subsidence time series at critical sites?
• Lines 308—315. I feel that this paragraph needs to be rewritten. The large difference between the leveling measurement (200 mm/yr, 2993-2009) and the InSAR measurement (110 mm/yr, 1993-2003) needs to be addressed very carefully.

“However, in areas of expected subsidence (based on groundwater extraction and leveling measurements) the subsidence patterns from InSAR are similar to those from leveling”. This statement is not precise. Be honest, I can not realize the “similar pattern” between leveling and INSAR images. In order to highlight the “similar patterns”, you may need to redraw Fig. 5. For example, add contour lines in InSAR images; only show the values between -60 to 15 mm/year.

• Line 354--- However, cur-353 rent data show that land subsidence in Dacheng is still ongoing but is decreasing. Should say “subsidence rate is decreasing?”  I found that in several places, subsidence should be replaced with “subsidence rate”. Please check!
• Line 422----“deep he compaction”??
• Lines 434-436--- The overall groundwater levels reducing in Xizhou, 434 Huwei, Tuku, and Yuanchang rose since this measure (de-commissioning wells) took ef-435 fect in 2011. Difficult to understand this sentence.

Author Response

Please see the attachment.

Author Response File: Author Response.docx