the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 08 Nov 2023
Overview: Cascading spatial, seasonal, and temporal effects of permafrost thaw on streamflow in changing nested Arctic catchments
Abstract. In the Arctic, the thawing of permafrost affects how catchments store and release water. However, the effects of thawing on the hydrological response remain poorly documented. In addition, it remains unclear how the effects of a thawing landscape will propagate through nested catchments. Here we investigate 10 nested catchments within the Yukon basin (Alaska and Canada) to study how permafrost thaw impacts catchments’ streamflow seasonality and storage-discharge relationships, and how these effects cascade through the nested catchments, from headwaters to downstream. Our results indicate that upstream catchments, characterized by continuous permafrost, have stronger streamflow seasonality and that these catchments also exhibit the most nonlinear storage-discharge relationships. Larger catchments downstream sustain year-round streamflow with baseflow continuing during winter. Since the 1950s flow regimes have become increasingly seasonal in the upstream catchments, with an earlier and more abrupt freshet, whereas further downstream flow seasonality has remained stable. Across the Yukon, storage-discharge relationships for 9 out of 10 sub-catchments have become increasingly nonlinear over time, with the biggest change occurring in the largest downstream catchments. In smaller catchments, each season has distinct recession characteristics, but those seasonal differences are not apparent further downstream. Upstream catchments are strongly influenced by localized change, whereas downstream catchments receive the effects of many different localized upstream impacts, making it difficult to detect a singular cause of change. Seasonal and long-term shifts in storage-discharge relationships are typically not accounted for by hydrological models and make accurate streamflow predictions more difficult. These shifts highlight how the changing landscape of the Arctic has far-reaching hydrological consequences.
- Preprint
(3568 KB) - Metadata XML
- BibTeX
- EndNote
Status: closed
-
RC1: 'Comment on egusphere-2023-2391', Anonymous Referee #1, 09 Jan 2024
The comment was uploaded in the form of a supplement: https://egusphere.copernicus.org/preprints/2023/egusphere-2023-2391/egusphere-2023-2391-RC1-supplement.pdf
-
RC2: 'Comment on egusphere-2023-2391', Anonymous Referee #2, 10 Jan 2024
General comments
The manuscript deals with seasonality and flow recession features in the nested Arctic catchments within the Yukon River basin, its major aim is to explore the cascading effect of permafrost thaw on streamflow differentiated by position within the cathchment, i.e., headwater vs lower-lying reaches. The manuscript leaves rather bleak impression and I would not recommend its publication in the present form. A resubmission with re-review is appropriate.
The manuscript starts with an unimpressive and unambitious Introduction. While the main idea of exploring nested sub-catchments is in fact catchy, the Introductions falls short to deliver, revolving around vaguely defined concepts. The description of seasonal effects of climate and permafrost warming (L53-68), as I feel, could be more consistent in a hydrological sense – though the authors are correct to say that the effects are interrelated. At least I see no difference between what the authors put as ‘firstly’, ‘secondly’ or ‘finally’ in their narration, in the sense that in all these cases, the surface-subsurface hydrological connectivity is concerned, and the persistence of residual thaw layers a.k.a. residual thaw zones, residual taliks, non-merging permafrost, supra-permafrost taliks, is implied. Note also that (Lamontagne-Hallé et al., 2018) exactly describes the hydrological role of non-merging permafrost as reproduced in SUTRA model.
Basically, the seasonality in Arctic hydrology response to thawing permafrost is mostly related to the time period when the additional thaw-related hydrological pathways are active, including, e.g., enhanced sub-permafrost groundwater discharge through open taliks and related winter runoff redistribution via icings. As permafrost becomes more discontinuous, the reduction in peak freshet discharge can also be observed as documented by Yang et al. (2015). Also, the thicker active layer in mineral soils can have no immediate effect on hydrologic response owing to their low hydraulic conductivity and high water retention capacity upon full saturation.
In the transition paragraph between active layer hydrology and catchment hydrology, the discussion around climate warming rate as compared to mean annual air temperature is alien to the textflow and should be rewritten and put into context. Eventually, it is clear that for upstream (headwater) catchments in continuous permafrost, most hydrological effects will be related to catchment surface (vegetation, ET, etc) and active layer features, while more large-scale connections are in play when moving downstream toward larger catchments. At L104-107, after having presented the recession flow analysis, I would also consider if the observed changes are in fact more notable and important at larger spatial scales, considering potential multiplicative effects from increased hydrological connectivity. In L111-115, the explications on the importance of spatial scales are redundant. Overall the Introduction section lacks solidity and needs to be reworked in this sense, to better explain the research aims, scope and objectives.
In the Methods section, it is totally unclear how the climate data were obtained and treated, and if the further hydrological analysis is sensitive to the propagation of uncertainties in climate records. It is unclear how point-based GHCN-Daily data were used to quantify the catchment-scale processes, and how exactly GHCN-Daily was substituted by ERA5 data. In general, the station- and reanalysis-based data can not be used interchangeably within the single research setting, or otherwise the ERA5 dataset should be evaluated for regional performance and compared to existing station-based observations. If the gridded GHCN-Daily data were used, please note the difference in spatial resolution between the two datasets. The authors should also reflect on a fact that, for the gauging stations presented in Table 1, there is no common period of record when data from all gauges overlap, and on the presence of missing data. Then, while the IPA reference dataset may be the one published by NSIDC, the most recent and comprehensive effort includes the ESA Permafrost CCI Project, v. 3.0 (Obu et al., 2021). Permafrost extent and active layer thickness for 2003-2019 period can be derived from this dataset.
The analysis presented in the Results and discussion section is insufficiently sound and needs to be significantly enhanced in order to be viable. The presented analysis, as it is also noted by the authors, does not account for different periods of record at most gauges. Are these records directly comparable in a manner presented in Figure 5, or its meaning is blurred by this difference? And could this analysis be enhanced to account for such difference? If the answer to the last question is ‘yes’, this should be done. If the answer is ‘no’, this has long-standing consequences for this manuscript. Likewise, in the following Section 3.2, does the analysis include only data from the overlapping period for the two gauges? Also, the temporal limits of seasons need to be explicitly explained.
On Figure 6, is the scale logarithmic at both (a) and (b) figures? With log Q = 10, assuming log base = 10, the Q becomes 10^10 which is weird. Also, does the manuscript discuss the potential effect of flow rate (mm/d) reduction in the downstream direction across the studied sub-catchments? Is this effect relevant to the discussion around the nested nature of the considered sub-catchments? Besides, on Figure 6a, it seems that winter points, if singled out, have rather different regression line direction with negative beta, if my understanding is correct. Is this the case, and if yes, does this have a plausible explanation from the viewpoint of the authors’ hypotheses? Any other gauges where such behavior occurs?
Instead of arrows on Figure 7, can the manuscript be revised in a way that uppermost sub-catchments are considered before the lowermost, i.e., in the Table 1 and in Figure 7? The arrows here only serve to order figures, so do the letters from (a) to (j), so either one is expletive. Also, because the periods are dissimilar, the data are not necessarily directly comparable. I would suggest adding the supporting hydrological information. The average flow rate will be instructive, i.e., to say whether the change in peak PC at Ruby (June) comes from lower mean annual discharge, or higher peak month runoff.
In L350-371, the explanation of the ‘analogue catchment’ follows, which I found excessively detailed. It is a well-known concept in hydrology and across disciplines, and as such it barely merits such an extensive description.
Overall, the manuscript will benefit from strengthening the Introduction, and giving clear and explicit explanations of the Methods used including GHCN-Daily and ERA5 reanalysis data which fall completely out of context. I strongly feel that rethinking the permafrost thaw effects on flow seasonality will provide more insights on the effects presented in the manuscript.
Line-by-line comments:
L85 : I would suggest using ‘permafrost-affected’ rather than ‘permafrost-laden’ catchment, and later in L86, propose the well-accepted ‘state of permafrost’ term.
L123-125 : the figure captions would normally be self-sufficient and mostly understandable without the manuscript text, so at least the caption should include the basin name; also since each larger catchment includes all smaller catchments, these are all correctly called ‘sub-catchments’.
L164 : what is ‘permafrost control on catchment response’ ? What kind of response is implied ?
L185-186 : though, e.g., Lyon et al. (2009) implied that b = 1 in their analysis, they also show that it is somehow not necessarily a linear relationship, with b = 1.16 for the studied catchment, while b = const suggestion still holds. The relative unimportance of the b intercept stems from the (Brutsaert, 2005) analysis and this should be referenced.
L191-194 : while the claim is correct, the early and later periods are also not necessarily comparable as they rarely match at different gauges.
L224-225 : it would be more physically and chronologically correct to say that the results present in this manuscript echo the findings from Curran and Biles (2021), and not vice versa.
L209-210 : how these results are affected by 60% of the missing data for this gauge ?
L237-238 : how these lines are relevant to the other contents of this paragraph ? I see no clear relation.
L243-244 : this phrasing undermines the whole utility of the presented analysis and Figure 5.
L279 : ‘deep thickness’ probably needs to be reworded.
Yang et al., 2015, https://doi.org/10.1016/j.quaint.2014.09.023
Douglas et al., 2021, https://doi.org/10.5194/tc-15-3555-2021
Citation: https://doi.org/10.5194/egusphere-2023-2391-RC2
Status: closed
-
RC1: 'Comment on egusphere-2023-2391', Anonymous Referee #1, 09 Jan 2024
The comment was uploaded in the form of a supplement: https://egusphere.copernicus.org/preprints/2023/egusphere-2023-2391/egusphere-2023-2391-RC1-supplement.pdf
-
RC2: 'Comment on egusphere-2023-2391', Anonymous Referee #2, 10 Jan 2024
General comments
The manuscript deals with seasonality and flow recession features in the nested Arctic catchments within the Yukon River basin, its major aim is to explore the cascading effect of permafrost thaw on streamflow differentiated by position within the cathchment, i.e., headwater vs lower-lying reaches. The manuscript leaves rather bleak impression and I would not recommend its publication in the present form. A resubmission with re-review is appropriate.
The manuscript starts with an unimpressive and unambitious Introduction. While the main idea of exploring nested sub-catchments is in fact catchy, the Introductions falls short to deliver, revolving around vaguely defined concepts. The description of seasonal effects of climate and permafrost warming (L53-68), as I feel, could be more consistent in a hydrological sense – though the authors are correct to say that the effects are interrelated. At least I see no difference between what the authors put as ‘firstly’, ‘secondly’ or ‘finally’ in their narration, in the sense that in all these cases, the surface-subsurface hydrological connectivity is concerned, and the persistence of residual thaw layers a.k.a. residual thaw zones, residual taliks, non-merging permafrost, supra-permafrost taliks, is implied. Note also that (Lamontagne-Hallé et al., 2018) exactly describes the hydrological role of non-merging permafrost as reproduced in SUTRA model.
Basically, the seasonality in Arctic hydrology response to thawing permafrost is mostly related to the time period when the additional thaw-related hydrological pathways are active, including, e.g., enhanced sub-permafrost groundwater discharge through open taliks and related winter runoff redistribution via icings. As permafrost becomes more discontinuous, the reduction in peak freshet discharge can also be observed as documented by Yang et al. (2015). Also, the thicker active layer in mineral soils can have no immediate effect on hydrologic response owing to their low hydraulic conductivity and high water retention capacity upon full saturation.
In the transition paragraph between active layer hydrology and catchment hydrology, the discussion around climate warming rate as compared to mean annual air temperature is alien to the textflow and should be rewritten and put into context. Eventually, it is clear that for upstream (headwater) catchments in continuous permafrost, most hydrological effects will be related to catchment surface (vegetation, ET, etc) and active layer features, while more large-scale connections are in play when moving downstream toward larger catchments. At L104-107, after having presented the recession flow analysis, I would also consider if the observed changes are in fact more notable and important at larger spatial scales, considering potential multiplicative effects from increased hydrological connectivity. In L111-115, the explications on the importance of spatial scales are redundant. Overall the Introduction section lacks solidity and needs to be reworked in this sense, to better explain the research aims, scope and objectives.
In the Methods section, it is totally unclear how the climate data were obtained and treated, and if the further hydrological analysis is sensitive to the propagation of uncertainties in climate records. It is unclear how point-based GHCN-Daily data were used to quantify the catchment-scale processes, and how exactly GHCN-Daily was substituted by ERA5 data. In general, the station- and reanalysis-based data can not be used interchangeably within the single research setting, or otherwise the ERA5 dataset should be evaluated for regional performance and compared to existing station-based observations. If the gridded GHCN-Daily data were used, please note the difference in spatial resolution between the two datasets. The authors should also reflect on a fact that, for the gauging stations presented in Table 1, there is no common period of record when data from all gauges overlap, and on the presence of missing data. Then, while the IPA reference dataset may be the one published by NSIDC, the most recent and comprehensive effort includes the ESA Permafrost CCI Project, v. 3.0 (Obu et al., 2021). Permafrost extent and active layer thickness for 2003-2019 period can be derived from this dataset.
The analysis presented in the Results and discussion section is insufficiently sound and needs to be significantly enhanced in order to be viable. The presented analysis, as it is also noted by the authors, does not account for different periods of record at most gauges. Are these records directly comparable in a manner presented in Figure 5, or its meaning is blurred by this difference? And could this analysis be enhanced to account for such difference? If the answer to the last question is ‘yes’, this should be done. If the answer is ‘no’, this has long-standing consequences for this manuscript. Likewise, in the following Section 3.2, does the analysis include only data from the overlapping period for the two gauges? Also, the temporal limits of seasons need to be explicitly explained.
On Figure 6, is the scale logarithmic at both (a) and (b) figures? With log Q = 10, assuming log base = 10, the Q becomes 10^10 which is weird. Also, does the manuscript discuss the potential effect of flow rate (mm/d) reduction in the downstream direction across the studied sub-catchments? Is this effect relevant to the discussion around the nested nature of the considered sub-catchments? Besides, on Figure 6a, it seems that winter points, if singled out, have rather different regression line direction with negative beta, if my understanding is correct. Is this the case, and if yes, does this have a plausible explanation from the viewpoint of the authors’ hypotheses? Any other gauges where such behavior occurs?
Instead of arrows on Figure 7, can the manuscript be revised in a way that uppermost sub-catchments are considered before the lowermost, i.e., in the Table 1 and in Figure 7? The arrows here only serve to order figures, so do the letters from (a) to (j), so either one is expletive. Also, because the periods are dissimilar, the data are not necessarily directly comparable. I would suggest adding the supporting hydrological information. The average flow rate will be instructive, i.e., to say whether the change in peak PC at Ruby (June) comes from lower mean annual discharge, or higher peak month runoff.
In L350-371, the explanation of the ‘analogue catchment’ follows, which I found excessively detailed. It is a well-known concept in hydrology and across disciplines, and as such it barely merits such an extensive description.
Overall, the manuscript will benefit from strengthening the Introduction, and giving clear and explicit explanations of the Methods used including GHCN-Daily and ERA5 reanalysis data which fall completely out of context. I strongly feel that rethinking the permafrost thaw effects on flow seasonality will provide more insights on the effects presented in the manuscript.
Line-by-line comments:
L85 : I would suggest using ‘permafrost-affected’ rather than ‘permafrost-laden’ catchment, and later in L86, propose the well-accepted ‘state of permafrost’ term.
L123-125 : the figure captions would normally be self-sufficient and mostly understandable without the manuscript text, so at least the caption should include the basin name; also since each larger catchment includes all smaller catchments, these are all correctly called ‘sub-catchments’.
L164 : what is ‘permafrost control on catchment response’ ? What kind of response is implied ?
L185-186 : though, e.g., Lyon et al. (2009) implied that b = 1 in their analysis, they also show that it is somehow not necessarily a linear relationship, with b = 1.16 for the studied catchment, while b = const suggestion still holds. The relative unimportance of the b intercept stems from the (Brutsaert, 2005) analysis and this should be referenced.
L191-194 : while the claim is correct, the early and later periods are also not necessarily comparable as they rarely match at different gauges.
L224-225 : it would be more physically and chronologically correct to say that the results present in this manuscript echo the findings from Curran and Biles (2021), and not vice versa.
L209-210 : how these results are affected by 60% of the missing data for this gauge ?
L237-238 : how these lines are relevant to the other contents of this paragraph ? I see no clear relation.
L243-244 : this phrasing undermines the whole utility of the presented analysis and Figure 5.
L279 : ‘deep thickness’ probably needs to be reworded.
Yang et al., 2015, https://doi.org/10.1016/j.quaint.2014.09.023
Douglas et al., 2021, https://doi.org/10.5194/tc-15-3555-2021
Citation: https://doi.org/10.5194/egusphere-2023-2391-RC2
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 218 | 93 | 23 | 334 | 15 | 15 |
- HTML: 218
- PDF: 93
- XML: 23
- Total: 334
- BibTeX: 15
- EndNote: 15
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1