the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
Brief Communication: Effects of different saturation vapor pressure calculations on simulated surface-subsurface hydrothermal regimes at a permafrost field site
Xiang Huang
Charles J. Abolt
Katrina E. Bennett
Abstract. Air saturation vapor pressure (SVP) can be calculated using different formulas, with or without over-ice correction. These different approaches result in variability that affects the simulation of surface-subsurface thermal-hydrological processes in cold regions; however, this topic has not been well documented to date. In this study, we compared the relative humidity (RH) downloaded and calculated from four data sources in Alaska based on five commonly used SVP formulas. RH, along with other meteorological indicators, was used to drive physically-based land surface models. Results show that RH is a sensitive parameter, and its biases from SVP with or without over-ice correction meaningfully impact model-based predictions of snow depth, sublimation, soil temperature, and active layer thickness.
- Preprint
(1508 KB) -
Supplement
(1758 KB) - BibTeX
- EndNote
Xiang Huang et al.
Status: closed
-
RC1: 'Comment on tc-2023-8', Anonymous Referee #1, 05 Mar 2023
The authors of this paper have tested the sensitivity of a land surface model to different uses and formulations of saturation vapor pressure equations. The title of this work is appropriate for the work contained. The results are important to users of these models, and similar or equivalent models in meteorology. More much work is required to clarify their point, but I think the authors are more than capable of it. If this were not submitted as a brief communication it could be accepted with major revisions. Even so, there are important clarifications to be made in language (e.g. SVP over liquid water or ice), experimental design (e.g. comparison to snow depth without snow density information), and statement of the problem (this is a common issue in meteorology). Specific comments are below. I wish the authors luck in revising this work for resubmission.
1. The problem addressed by the authors is overstated. The choice of saturation vapor pressure formula is an important one, but one that many already take seriously. I agree that authors in our disciplines ought to list the vapor pressure formula in their published work, especially if results might be sensitive to the choice. However, this information is almost always available upon further requests for information from authors about their models, or investigation into open source model code itself.
2. Nevertheless, the authors here have clearly demonstrated a sensitivity of their model to choice of vapor pressure formulation. Some important questions remain about what exactly was done as the language used here is ambiguous. In this work, there is confusion about what the 'over-ice' correction is. It seems that the 'over-ice' correction is really just the SVP over a plane surface of ice, not a correction of an existing formula of SVP over ice. In revising this work, it is critical that the authors use more standard language about saturation vapor pressure over 'liquid water' or 'ice'.
3. In any case, there are implicit procedural assumptions that the authors make here in their sensitivity test that should further be clarified. It seems that this work is not just about a choice of which formula to use, but also when their is saturation with respect to liquid water or saturation with respect to ice in the air and soil/snow. The default assumption in this work, outline in the formulas in Section 2, is that saturation with respect to liquid water does not exist below 0oC, after which all vapor pressures are saturated with respect to ice. This may be true in interstitial pore spaces of soil and snow, but it is not true for the near-surface atmosphere. Air can be saturated and even super-saturated with respect to liquid water below 0oC. In clean Arctic air, liquid water fogs can exist down below -30oC. I wonder how your narrative, or even the results, would change in testing this assumption explicitly, rather than implicitly when suggesting that some authors are not using the appropriate formulas below 0oC? How often this sort of RH occurs can be found in the BEO or similar data sets. It could be that the differences between obs from BEO and the polygon center/trough are the result of assumptions about liquid/ice saturation threshold.
4. I recommend that this work be rejected from the Brief Communications. The results are interesting and important, but require reframing and an expansion to separate assumptions around the use of which formula, as well as when to choose liquid vs ice saturation. I recommend the authors focus their future work on Arctic permafrost hydrology, not the broad misuse of vapor formulations (this problem is not as rampant as is implied here) . I suggest using multiple scenarios to test assumptions around temperature thresholds of liquid vs ice saturation in the near-surface atmosphere, as well as which formula is most appropriate. Better curation and assessment of the observations is also important here. The observations of snow depth that are used as ground truth here require more processing. Specifically in this work, we are left wondering which observations to trust in 2013. But, critically if snow depth is the ground truth, then we need to know the snow density, also, if we are truly tracking RH and latent heat impacts on the snow surface as the result of RH formulations. Finally, I would wish that the authors make clear, strong, and evidence-based recommendations on which formulas are best.
Citation: https://doi.org/10.5194/tc-2023-8-RC1 -
RC2: 'Comment on tc-2023-8', Anonymous Referee #2, 21 Mar 2023
Recommendation: Reject, but encourage resubmission.
General comment:
The authors drive a land-surface model with different formulae for saturation vapour pressure (SVP) and compare the model outputs. The results are obscure because the figures are difficult to comprehend. In fact, I was unable to review the paper in detail because I could not make sense of the figures, but I do have some suggestions, as follows.
Major comments:
(1) Key references are not cited. My understanding is that the accepted formula for SVP over ice is that of Marti and Mauersberger (1993), which is not cited. And the authoritative review by Murphy and Koop (2005), for both ice and supercooled water, is also not cited.
(2) The authors use peculiar language, as on line 130, where they write that “some studies suggest” distinguishing SVP over liquid from SVP over ice. But of course these two vapour pressures must be distinguished, and much has been written about exactly how they differ and why, for example in the review by Murphy and Koop. It belittles ice to say that ice is just liquid water with an optional “correction”, as on line 154. The authors conclude their paper on line 238 with a recommendation that SVP should be “corrected over the ice surface”, by which they mean that the SVP difference between liquid and ice should be accounted for.
(3) Line 90. The Goff-Gratch formula is presented here without discussion, and without saying which version of Goff-Gratch is used in this work. When the authors describe it as “internationally accepted”, they probably mean the version presented in WMO’s Document 49. But Murphy and Koop (2005) pointed out three typographical errors in WMO’s equation (in comparison to the equation in Goff’s 1957 paper).
(4) Figure 1. There are two vertical axes: RH on the left, air temperature on the right. Which of the plots are temperature, and which are RH?
The legend symbols are indistinct, so I can’t relate them to the different colours of the dots.
The hundreds (thousands?) of dots are overlapping, thus signifying nothing.
Are we meant to compare Figures 1a and 1b? This is not easy. Perhaps the thing to do is to form monthly averages, then plot differences of the monthly averages from a reference case.
(5) Figure 2. I have difficulty distinguishing the three separate blue lines.
The legend indicates that ten different quantities are plotted; this is far too many for one graph. I can’t distinguish them all; when I count them I get less than 10.
(6) Figure S1. As on the figures in the main paper, far too much data are presented, overlapping so as to obscure any interpretation. Especially in frame (b), it is impossible to tell the level of agreement among the three sets of closely-packed dots. Here again, monthly averages could be differenced to show the bias, instead of plotting the daily data.
Minor comments:
Line 52. Define CRN.
Line 54. Define GLDAS.
Line 104. “polynomial fits”. Equation 5 is not a polynomial.
Lines 114-115. “over the water surface or without over-ice correction”. What do you mean here by “or”?
Citation: https://doi.org/10.5194/tc-2023-8-RC2 -
CC1: 'Comment on tc-2023-8', Lance McCormick, 03 Apr 2023
I take great interest in reading this manuscript examining the impacts of the variations of water vapor (content) in the air on the subsurface hydrothermal and surface snowpacks. Humidity is a prevalent and essential meteorological indicator and is widely used/monitored indoors and outdoors. Observations of humidity always show a high spatiotemporal variation, but their corrections rarely receive attention. According to the uncertainties of humidity inputs in Figures 2 and 3, it’s surprising to me has up to 20~30 percent variations on the calculated snow depth and temperature values. I noted substantial contents were presented in the Supplement and consider this short paper will attract large readerships, which generally uplift the average quality of TC’s brief communication from my observations. Several places need to be corrected before being accepted.
i) The figure is hard to read, and I suggest showing weekly or monthly averaged values instead of daily changes. Also, there have typos in “Rletive Humidity”.
ii) Also hard to identify Figure 2, and I would add another sub-figure showing zoom-in features.
iii) I checked your “code and data availability” but failed to access the link the authors provided. I suggest adding the critical equations of the ATS simulator because it is not a widely used software. Please carefully check them.Citation: https://doi.org/10.5194/tc-2023-8-CC1
Status: closed
-
RC1: 'Comment on tc-2023-8', Anonymous Referee #1, 05 Mar 2023
The authors of this paper have tested the sensitivity of a land surface model to different uses and formulations of saturation vapor pressure equations. The title of this work is appropriate for the work contained. The results are important to users of these models, and similar or equivalent models in meteorology. More much work is required to clarify their point, but I think the authors are more than capable of it. If this were not submitted as a brief communication it could be accepted with major revisions. Even so, there are important clarifications to be made in language (e.g. SVP over liquid water or ice), experimental design (e.g. comparison to snow depth without snow density information), and statement of the problem (this is a common issue in meteorology). Specific comments are below. I wish the authors luck in revising this work for resubmission.
1. The problem addressed by the authors is overstated. The choice of saturation vapor pressure formula is an important one, but one that many already take seriously. I agree that authors in our disciplines ought to list the vapor pressure formula in their published work, especially if results might be sensitive to the choice. However, this information is almost always available upon further requests for information from authors about their models, or investigation into open source model code itself.
2. Nevertheless, the authors here have clearly demonstrated a sensitivity of their model to choice of vapor pressure formulation. Some important questions remain about what exactly was done as the language used here is ambiguous. In this work, there is confusion about what the 'over-ice' correction is. It seems that the 'over-ice' correction is really just the SVP over a plane surface of ice, not a correction of an existing formula of SVP over ice. In revising this work, it is critical that the authors use more standard language about saturation vapor pressure over 'liquid water' or 'ice'.
3. In any case, there are implicit procedural assumptions that the authors make here in their sensitivity test that should further be clarified. It seems that this work is not just about a choice of which formula to use, but also when their is saturation with respect to liquid water or saturation with respect to ice in the air and soil/snow. The default assumption in this work, outline in the formulas in Section 2, is that saturation with respect to liquid water does not exist below 0oC, after which all vapor pressures are saturated with respect to ice. This may be true in interstitial pore spaces of soil and snow, but it is not true for the near-surface atmosphere. Air can be saturated and even super-saturated with respect to liquid water below 0oC. In clean Arctic air, liquid water fogs can exist down below -30oC. I wonder how your narrative, or even the results, would change in testing this assumption explicitly, rather than implicitly when suggesting that some authors are not using the appropriate formulas below 0oC? How often this sort of RH occurs can be found in the BEO or similar data sets. It could be that the differences between obs from BEO and the polygon center/trough are the result of assumptions about liquid/ice saturation threshold.
4. I recommend that this work be rejected from the Brief Communications. The results are interesting and important, but require reframing and an expansion to separate assumptions around the use of which formula, as well as when to choose liquid vs ice saturation. I recommend the authors focus their future work on Arctic permafrost hydrology, not the broad misuse of vapor formulations (this problem is not as rampant as is implied here) . I suggest using multiple scenarios to test assumptions around temperature thresholds of liquid vs ice saturation in the near-surface atmosphere, as well as which formula is most appropriate. Better curation and assessment of the observations is also important here. The observations of snow depth that are used as ground truth here require more processing. Specifically in this work, we are left wondering which observations to trust in 2013. But, critically if snow depth is the ground truth, then we need to know the snow density, also, if we are truly tracking RH and latent heat impacts on the snow surface as the result of RH formulations. Finally, I would wish that the authors make clear, strong, and evidence-based recommendations on which formulas are best.
Citation: https://doi.org/10.5194/tc-2023-8-RC1 -
RC2: 'Comment on tc-2023-8', Anonymous Referee #2, 21 Mar 2023
Recommendation: Reject, but encourage resubmission.
General comment:
The authors drive a land-surface model with different formulae for saturation vapour pressure (SVP) and compare the model outputs. The results are obscure because the figures are difficult to comprehend. In fact, I was unable to review the paper in detail because I could not make sense of the figures, but I do have some suggestions, as follows.
Major comments:
(1) Key references are not cited. My understanding is that the accepted formula for SVP over ice is that of Marti and Mauersberger (1993), which is not cited. And the authoritative review by Murphy and Koop (2005), for both ice and supercooled water, is also not cited.
(2) The authors use peculiar language, as on line 130, where they write that “some studies suggest” distinguishing SVP over liquid from SVP over ice. But of course these two vapour pressures must be distinguished, and much has been written about exactly how they differ and why, for example in the review by Murphy and Koop. It belittles ice to say that ice is just liquid water with an optional “correction”, as on line 154. The authors conclude their paper on line 238 with a recommendation that SVP should be “corrected over the ice surface”, by which they mean that the SVP difference between liquid and ice should be accounted for.
(3) Line 90. The Goff-Gratch formula is presented here without discussion, and without saying which version of Goff-Gratch is used in this work. When the authors describe it as “internationally accepted”, they probably mean the version presented in WMO’s Document 49. But Murphy and Koop (2005) pointed out three typographical errors in WMO’s equation (in comparison to the equation in Goff’s 1957 paper).
(4) Figure 1. There are two vertical axes: RH on the left, air temperature on the right. Which of the plots are temperature, and which are RH?
The legend symbols are indistinct, so I can’t relate them to the different colours of the dots.
The hundreds (thousands?) of dots are overlapping, thus signifying nothing.
Are we meant to compare Figures 1a and 1b? This is not easy. Perhaps the thing to do is to form monthly averages, then plot differences of the monthly averages from a reference case.
(5) Figure 2. I have difficulty distinguishing the three separate blue lines.
The legend indicates that ten different quantities are plotted; this is far too many for one graph. I can’t distinguish them all; when I count them I get less than 10.
(6) Figure S1. As on the figures in the main paper, far too much data are presented, overlapping so as to obscure any interpretation. Especially in frame (b), it is impossible to tell the level of agreement among the three sets of closely-packed dots. Here again, monthly averages could be differenced to show the bias, instead of plotting the daily data.
Minor comments:
Line 52. Define CRN.
Line 54. Define GLDAS.
Line 104. “polynomial fits”. Equation 5 is not a polynomial.
Lines 114-115. “over the water surface or without over-ice correction”. What do you mean here by “or”?
Citation: https://doi.org/10.5194/tc-2023-8-RC2 -
CC1: 'Comment on tc-2023-8', Lance McCormick, 03 Apr 2023
I take great interest in reading this manuscript examining the impacts of the variations of water vapor (content) in the air on the subsurface hydrothermal and surface snowpacks. Humidity is a prevalent and essential meteorological indicator and is widely used/monitored indoors and outdoors. Observations of humidity always show a high spatiotemporal variation, but their corrections rarely receive attention. According to the uncertainties of humidity inputs in Figures 2 and 3, it’s surprising to me has up to 20~30 percent variations on the calculated snow depth and temperature values. I noted substantial contents were presented in the Supplement and consider this short paper will attract large readerships, which generally uplift the average quality of TC’s brief communication from my observations. Several places need to be corrected before being accepted.
i) The figure is hard to read, and I suggest showing weekly or monthly averaged values instead of daily changes. Also, there have typos in “Rletive Humidity”.
ii) Also hard to identify Figure 2, and I would add another sub-figure showing zoom-in features.
iii) I checked your “code and data availability” but failed to access the link the authors provided. I suggest adding the critical equations of the ATS simulator because it is not a widely used software. Please carefully check them.Citation: https://doi.org/10.5194/tc-2023-8-CC1
Xiang Huang et al.
Supplement
Xiang Huang et al.
Viewed
| HTML | XML | Total | Supplement | BibTeX | EndNote | |
|---|---|---|---|---|---|---|
| 245 | 92 | 17 | 354 | 36 | 5 | 4 |
- HTML: 245
- PDF: 92
- XML: 17
- Total: 354
- Supplement: 36
- BibTeX: 5
- EndNote: 4
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1