the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
Comparing thaw probing, electrical resistivity tomography, and airborne lidar to quantify lateral and vertical thaw in rapidly degrading boreal permafrost
Abstract. Permafrost thaw across earth’s high latitudes is leading to dramatic changes in vegetation and hydrology. We undertook a two-decade long study on the Tanana Flats near Fairbanks, Alaska to measure permafrost thaw and associated ground surface subsidence via field-based and remote-sensing techniques. Our study focused on four transects that included an unburned area and three fire scars (1988, 2001, and 2010). Three types of permafrost quantification were used. First, repeat measurements of ground-surface elevation and depth to the top of near-surface permafrost were made between 1999 and 2020. Widespread near-surface permafrost degradation was evident between 2004 and 2020 with top-down thaw of near surface permafrost doubling from 18 % to 36 % over the study period. Multi-year frost and repeat thin permafrost, two types of permafrost aggradation, were almost completely absent by 2020. Second, we calculated rates of top-down versus lateral thaw using airborne lidar measurements collected in 2014 and 2020. Lateral thaw of tabular shaped permafrost boundaries and development of unfrozen zones between the bottom of the seasonally frozen layer and the top of near-surface permafrost (taliks) were evident. Third, repeated electrical resistivity tomography measurements in 2012 and 2020 supported surface-based thaw observations and allowed subsurface mapping of permafrost morphologies up to 20 m deep. The study identified strengths and limitations of the three methods we used to quantify permafrost thaw degradation. Future applications of these methods should apply geospatial analyses to identify variables relating surface and subsurface conditions to project finer scale field-based spatial assessments across broader regions.
- Preprint
(1961 KB) - Metadata XML
-
Supplement
(107 KB) - BibTeX
- EndNote
Status: open (until 07 May 2025)
-
RC1: 'Comment on egusphere-2024-3997', Di Wang, 31 Mar 2025
reply
This is interesting and timely work. Some comments:
- The description of permafrost degradation is quite general, please provide some data (temperature increasing rate, extreme climate event frequency and severity) to support it.
- The literature review is more like a background description. Please explicitly highlight the novelty and necessity of combining thaw probing, ERT, and airborne lidar.
- It is not clear why these three technologies were selected. Please revise it.
- The data collection periods are different among these three technologies. How do you compare them properly?
- Line 490: The limitations of each technology were discussed in detail; excellent work. However, the application of emerging ML is vague.
- Conclusion: Among these technologies mentioned, what is the most promising one?
- The findings were based on the condition in Alaska. If such knowledge is being transferred to other cold regions, what is the most important issue for localization?
Citation: https://doi.org/10.5194/egusphere-2024-3997-RC1 -
RC2: 'Comment on egusphere-2024-3997', Anonymous Referee #2, 17 Apr 2025
reply
GENERAL COMMENTS
The authors use repeat probing, electrical resistivity tomography, and LiDAR surveys to assess lateral and vertical thaw of permafrost across one undisturbed and three fire disturbed transects in the Tanana Flats near Fairbanks, Alaska from 2012 to 2020. The authors use the frost probing results to classify sections of each of the study transects into three permafrost aggradation conditions or five permafrost degradation stages, and they find that the degraded proportion increases through time. The greatest degradation was found to occur at the site that was disturbed by fire most recently.
Overall, the manuscript is interesting as it uses three important permafrost monitoring/measurement methods in the assessment of permafrost thaw at different stages following fire disturbance. However, I believe that the manuscript requires major revisions prior to further consideration for publication in The Cryosphere.
- Each of the methods that is used is a recognized method for monitoring or measuring permafrost change through time. However, the authors do not currently provide any information on accuracies, precisions, and/or errors associated with each of these methods. This information needs to be included for the reader to be able to properly assess the significance of thaw/change. For example, the authors use thresholds of 0.3 m for assessing subsidence using LiDAR and of 0.15 m for assessing settlement from ground surface measurements, but there are no descriptions or explanations on the errors of the methods.
- The authors apply annual frost probing, repeat ERT from 2012 and 2020, and repeat LiDAR from 2014 and 2020.
- Frost probing is a standard field method for confirming permafrost presence and for measuring the depth of seasonal thaw on the date of the measurement. The authors are strongly recommended to provide the exact dates of frost probe measurements and to provide contextual information that could support comparisons between measurements from year to year. For example, a sum of the total number of thawing degree days per year up to the measurement date for frost probing, ERT surveys, and LiDAR acquisitions could demonstrate that field measurements were conducted at similar points in the thaw season throughout the study period and that no bias was introduced by the field visit/measurement date.
- Given that the first ERT survey and the first LiDAR acquisition were conducted in different years, it would be very helpful if the authors could provide a description of the climatic context for the 2012 versus 2014 years. The authors should also include a limitations section in the Discussion that addresses overall limitations to their study, including issues in comparing between methods in different years.
- The authors have compiled and collected valuable data that describes permafrost conditions under different disturbance contexts in the Tanana Flats. However, much of this work has already been described in previous publications (e.g., Douglas et al., 2016), and the advancement of this contribution relative to previous contributions appears to be the development of a classification scheme that identifies an increase in degradation through time. The classification scheme has eight total classes, three for aggradation (?) and five for degradation. The main message of this paper appears to be focused on trends in the proportions of these classes at each study site through time. This classification scheme is an interesting approach for evaluating broad patterns of permafrost aggradation versus degradation between sites, but it also introduces several issues.
- First, eight classes is a lot of classes. The authors are encouraged to consider combining some of these classes together and recategorizing them as intact permafrost (UD), newly aggraded permafrost (MF/NPR), vertically degraded permafrost (DI/DPS/DPD), and laterally degraded permafrost (DL). The patterns of decreased intact and newly aggraded permafrost and increased vertically/laterally degraded permafrost at the study sites would still be evident, but having fewer classes would improve readability and overall comprehension.
- Second, the UD class is defined according to a threshold of 0.8 m, which the authors describe as being a typical maximum range of late summer thaw for organic-rich soil. The DI and DPS classes are defined according to a threshold of 1.1 m, and the DPD and DCO are defined according to the limit of the frost probe, which the authors describe as ranging from 2.5 to 4 m. These thresholds are not necessarily applicable to other study regions, nor to other materials within the same study region, and do not take into account settlement of the ground surface. As permafrost thaws, it is possible for ground ice to melt, for the ground surface to settle, and for the overall thaw penetration to increase, but for the frost table to remain similar from year to year. The limit of the frost probe, measuring from 2.5 to 4 m, is also quite a large range, and the authors are encouraged to either be more specific or to use the lower limit of 2.5 m. By combining the DI/DPS/DPD classes together into an overall vertically degraded permafrost class, as recommended above, this would eliminate the use of the 1.1 and 2.5-4 m thresholds. These changes would make the classification scheme more general and potentially more applicable to other permafrost regions.
- The classification scheme, which makes up the main message of the paper, is derived completely from the frost probing results that occurred annually from 2012 to 2020. However, the title, introduction, and discussion suggest that each of the three methods will be separately evaluated for their utility in monitoring or delineating permafrost. At the very end of the discussion, P22 L497-498, the authors describe the importance of “a coupled application of these methods”. The authors are encouraged to follow what they have written here and to actually integrate these methods. Frost probing is an important validation measurement for ERT, as the authors describe in P16 L313-314, and repeat LiDAR acquisitions are very useful for evaluating lateral versus vertical thaw. It would be helpful to present the data in a more integrated way, perhaps by extracting the old thaw/lateral thaw/vertical thaw shown in Figure 7 along the actual frost probing/ERT study transects and comparing those results with the repeat ERT results from Figure 9 and the designation of vertically versus laterally thawed permafrost from the frost probing.
SPECIFIC COMMENTS
TITLE
P1 L1, lidar should be written as LiDAR, as it is actually an acronym for “Light Detection and Ranging”.
ABSTRACT
P1 L10-24, the abstract does not make direct mention of the classification scheme that was developed as part of this study. It does mention the loss of “multi-year frost and repeat thin permafrost”, which are part of the classification scheme, but these are not commonly known terms in permafrost science. Please revise to include more information here, explaining the development of the classification scheme.
P1 L14, here the repeat measurements of ground surface elevation and depth to the top of permafrost are described as having occurred from 1999 to 2020, but later in the main text, the transects are described as having first been established in 2011 and that surface topography and seasonal thaw depth were measured in fall 2012. I would recommend changing the text in the abstract from 1999 to 2020 to 2012 to 2020 to remain consistent with what is actually presented and described.
INTRODUCTION
P2 L41, please remove “in the subsurface”, as permafrost is already below the ground surface.
P2 L43, Holloway et al. 2020 does not describe repeat geophysical measurements in lowland sites, but is rather a review paper on post-fire permafrost and ecosystem response.
P2 L44, consider changing “the greatest means” to “an effective means”.
P2 L52-53, consider changing “in an area of degrading ice-rich lowland permafrost” to “in a lowland area containing ice-rich permafrost”.
STUDY SITE AND METHODS
P2 L59, consider renaming to “Study area and methods” or “Study sites and methods”, as there is more than one study site.
P3 L64-65, more information is required on the permafrost conditions at the study sites. Given how much prior work has been conducted here at these study sites, it would be helpful if the authors could provide more information on the permafrost conditions, either here in the study area section or in the beginning of the results section. More specific information on permafrost thickness, permafrost temperatures, active layer thicknesses, surficial materials, depth to bedrock, etc. would all be beneficial.
P3 Figure 1, consider including the location of Fairbanks on the inset map, as the study region is described in the text as being south of Fairbanks, but the location of Fairbanks is not provided.
P3 Figure 1, consider including some site photos.
P3 Figure 1, please provide the source for the fire perimeters in the figure caption.
P3 Figure 1, please review the accessibility of this figure and the other figures by running them through a colour blindness simulator.
P4 L76, is there a particular station within the Tanana Flats at which these values are measured? Please provide the corresponding period for the mean annual air temperature and mean summer and winter monthly air temperatures. For example, do these values correspond to the 1991-2020 climate normal?
P4 L76-81, please consider providing additional context related to climatic conditions during the study period. A plot showing mean annual air temperatures and total annual precipitation from 2012 to 2020 would help to contextualize the study.
P4 L83, what is meant by “pure or mixed white spruce”?
P4 L85, please provide an estimate of the fire interval here, rather than saying “regular intervals”.
P4 L95, is “TF50” supposed to be “TF88”?
P4 L87-95, please provide a rationale for the extension of the lengths of TF01, TF10, and T1 from 2011 to 2012.
P4 L100, there is no mention of the exact timing or dates of the site visits that were completed each year from 2012 to 2020. This information is very important, as biases can be introduced if site visits occurred later in the thaw season as the study period progressed. Please provide a table of the dates of the site visits and additional contextual information, such as the sum of thawing degree days leading up to the site visit date for each year. This can be included as supplementary information if needed.
P4 L99-102, what was used to measure surface topography and elevation? For example, was the elevation estimated from a handheld GPS and the topography then estimated using an Abney level or clinometer? Or was a differential GPS used, as indicated by Figure 2? Please provide a description of the measurement method, along with mention of the accuracy and precision of the measurement method.
P4-5 L103-118, consider sharing the classification scheme in the form of a table or a schematic.
P4-5 L103-118, the three “permafrost conditions” are later described in the caption for Figure 4 as “aggradation stages” and are earlier described in the abstract as “types of permafrost aggradation”, but these are not really aggradation processes. Please revise.
P4 L106-107, please provide a reference for the threshold of 0.8 m as the typical maximum range of late summer thaw for organic-rich soil.
P5 L131-132, please provide a reference for this statement on the dipole-dipole array.
P6 L166, please provide a brief description and a reference for this statement on salt and pepper effects.
RESULTS
P7 L173, here there are six degradation stages that are mentioned, but the earlier classification scheme only describes five stages? Please revise and correct.
P7 L173-187, this section repeats the descriptions for the different classifications, which were first described on P5. The descriptions for the classifications here on P7 are also not consistent with those on P5. For example, here, DI is described as regions where thaw depths are >100 cm and < 120 cm or had increased by 30 cm over time. On P5, DI is described as regions where thaw depths increase to 1.1 m. Please ensure that the classes are consistently described and applied, and please revise to remove repetition between sections.
P8-9 Figures 2-3, the figures are very comprehensive; however, they are also a bit confusing as they provide so much information. It would be best to perhaps show only the relevant information in 2012 and 2020, as annual information is provided in Figure 4. Please also consider combining Figures 3 and 4 into one large figure with all four sites together and re-organizing the figures to present burned site results in chronological order of fire disturbance: TF88, TF01, TF10. This should help with interpretation of inter-site comparisons/results.
P13 L263-267, what is the accuracy of the elevation measurement method? Are these elevation difference thresholds of 0.15 m beyond the error of the instrument/method? Please clarify.
P14 Figure 7, please show the location of each of the transects within their respective 300 by 500 m LiDAR study areas.
P15 L290, please provide additional information on the overall thickness of permafrost along each of the study transects. The authors do state in the study area section that permafrost in this overall region can measure up to 50 m in thickness, though it would be helpful to know how thick the permafrost is at each study site. For example, do the ERT surveys show permafrost thicknesses extending beyond 20 m?
P15 L299-300, this statement that the resistivity values decreased between 2012 and 2014 should be removed if the authors do not present the tomograms from 2014 in this paper.
P16 L313-314, great, frost probing is an important measurement to collect to support interpretations of permafrost presence from ERT!
DISCUSSION
P17 L325, change “Interior Alaska permafrost” to “Permafrost in interior Alaska”.
P17 L333, this part of the discussion mentions that the study sites consist of mixed forest, birch forest, or black spruce woodland; however, the study area section does not mention black spruce and rather describes the region as containing deciduous Alaska paper birch and aspen mixed with pure or mixed white spruce. Please revise and correct.
P17 L333, “These are the most common landforms above boreal discontinuous permafrost” – what does this mean? That the permafrost plateaus mentioned in the previous sentence are the most common landforms? Please clarify.
P17 L337, what is “braided ice”? I am not familiar with this term as a cryostructure, and it does not seem to be a term that is commonly used outside of this author team (see Brown et al., 2015; Douglas et al., 2016). Please clarify or describe what braided ice is.
P17 L352-354, many years of personal experience with frost probing in various materials would lead me to disagree with these statements. I would say that frost probing does not provide precise measurements, as the depth to the frost table may vary according to the person performing the probing, to the material, and to the timing of the measurement from year to year. I also do not think that it is possible to detect precise changes in soil texture with depth, and if this is the method that is used to determine stratigraphy, then soil pits need to be dug to verify these interpretations. The discussions of the limitations of probing on P18 L370-373 also contradict this description of probing as a “precise” method that can “detect marked changes in soil texture”.
P19 L391, how was the threshold of 0.3 m for significant permafrost thaw determined? What is the error of the LiDAR method? Please include this information.
P19 L395, these are important limitations to the LiDAR method. Please consider providing context for precipitation in the two years of acquisition (2014, 2020).
P19 L399-403, these sentences are quite repetitive and a bit contradictory, as L399 states that repeat ERT can delineate permafrost boundaries, but L403 states that ERT is less precise in delineating boundaries.
P19 L404, replace “it’s” with “its”.
P19 L403-405, these sentences are quite repetitive as well, mentioning the importance of boreholes in both sentences. Please revise.
P20 L442-447, this information would actually fit best in the study area section, as it provides helpful context on recent climate history for this region. Please consider moving this.
P21 L458-460, this information on sediments at the study sites would also fit well in the study area section. Please consider moving this.
P21 L476-477, please provide a reference for this statement that describes a positive feedback that facilitates further lateral thaw.
CONCLUSION
P22 L494, what data was provided from 2004? The methods section describes the survey transects as being established in 2011/2012, not 2004? Please revise and correct.
P22 L502, which one field site is this referring to? Please provide the name of the site that is experiencing substantial vertical thaw.
P22 L502-504, given the timing of the field studies relative to the timing of the fire disturbances, it is a bit difficult to really determine whether the thaw that has occurred at TF88, TF01, and T1 is associated with the press disturbance of climate warming rather than their respective pulse disturbances. Please consider rewording.
P22 L510-512, agreed that it is very challenging to monitor permafrost! But integrating these three methods is a great way to start.
Citation: https://doi.org/10.5194/egusphere-2024-3997-RC2
Viewed
HTML | XML | Total | Supplement | BibTeX | EndNote | |
---|---|---|---|---|---|---|
130 | 40 | 6 | 176 | 23 | 5 | 5 |
- HTML: 130
- PDF: 40
- XML: 6
- Total: 176
- Supplement: 23
- BibTeX: 5
- EndNote: 5
Viewed (geographical distribution)
Country | # | Views | % |
---|---|---|---|
United States of America | 1 | 63 | 40 |
Canada | 2 | 13 | 8 |
Germany | 3 | 12 | 7 |
China | 4 | 6 | 3 |
United Kingdom | 5 | 6 | 3 |
Total: | 0 |
HTML: | 0 |
PDF: | 0 |
XML: | 0 |
- 1
- 63