the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
19 Jan 2021
19 Jan 2021
Testing the altitude attribution and vertical resolution of AirCore measurements with a new spiking method
- Institute for Atmospheric and Environmental Sciences, University of Frankfurt, Frankfurt, 60438, Germany
- Institute for Atmospheric and Environmental Sciences, University of Frankfurt, Frankfurt, 60438, Germany
Abstract. AirCores have been increasingly used to capture vertical profiles of trace gases reaching from the ground up to about 30 km, in order to validate remote sensing instruments and to investigate transport processes in the stratosphere. When deployed to a weather balloon, accurately attributing the trace gas measurements to the sampling altitudes is non-trivial especially in the stratosphere. In this paper we present the CO-spiking experiment, which can be deployed to any AirCore on any platform in order to evaluate different computational altitude attribution processes and to experimentally derive the vertical resolution of the profile by injecting small volumes of signal gas at predefined GPS-altitudes during sampling. We performed two CO-spiking flights with an AirCore from the Goethe-University of Frankfurt (GUF) deployed to a weather balloon in Traînou, France in June 2019. The altitude retrieval based on an instantaneous pressure equilibrium assumption slightly overestimates the sampling altitudes, especially at the top of the profiles. For these two flights our altitude attribution is accurate within 250 m below 20 km. Above 20 km the bias becomes larger and reaches up to 1.2 km at 27 km altitude. Differences in descent velocities are uncovered to have a major impact on the altitude attribution bias. We identified the time lag between the theoretically attributed altitude and the actual CO-spike release altitude to be a possible empirical correction parameter for our AirCore altitude retrieval across different flights. Regarding the corrected profiles, the altitude attribution is accurate within 120 m throughout the profile. Further investigations are needed in order to test for the scope of validity of this correction parameter regarding different ambient conditions and maximum flight altitudes. We derive the vertical resolution from the CO-spikes of both flights and compare it to the modelled vertical resolution. The modelled vertical resolution is better than the experimentally derived resolution throughout the profile, albeit agrees within 220 m. All our findings derived from the two CO-spiking flights are strictly bound to the GUF AirCore dimensions. The newly introduced CO-spiking experiment can be used to test different combinations of AirCore configurations and platforms in future studies.
Thomas Wagenhäuser et al.
Status: closed
-
RC1: 'Comment on amt-2020-461', P. P. Tans, 05 Feb 2021
This paper is a very useful contribution to the literature on AirCore. The spiking idea is new, and the experiment has been carried out carefully, so that the results are solid.
I have a few minor comments:
line 81: The internal diameter is what matters, so please provide that. I assume that the predicted results use the ID.
line 137: The fill gas at the closed end of the tube will not be distributed as a Gaussian. It has to be asymmetric because the end is closed off. What it looks like depends on how much fill gas is left, but close to the end the spatial derivative of each gas fraction has to go to zero. If there is a lot of fill gas left, occupying several diffusion length scales, the fill gas fraction must approach 1 at the closed end. When the fill gas enters the analyzer the transition ought to be rapid, unless the tubing toward the analyzer and the analyzer cell itself cause a lot of mixing. Some years ago I analyzed experiments with "plug" transitions, sudden mole fraction shifts inserted very close to the Picarro. In the hypothetical case that the cell would be perfectly mixed all the time, the insertion of a plug should produce a negative exponential approach toward the new steady state. If there is plug flow within the cell, so that the rapid transition is mostly preserved, the approach to the new state should be linear. It turns out that the actual transition was in between these two cases. The "response function" of each analyzer will depend on pressure, cell volume, and shape of the cell. This subject comes back on line 156.
line 193: I suppose the signal gas mixture is CO-in-natural air. If a larger spike is inserted, you don't want to alter the main gases of interest.
line 234: typo - should be June 17
line 245: "resulting calculated vertical profile" (this is for clarity)
line 350: I would be very surprised if junctions between sections could cause much additional mixing. The flow conditions in the tube are extremely far away from any kind of turbulence. My suggestion is to look further into the analyzer contribution to mixing.
line 387: I suggest replacing "proof," with "prove"
- AC1: 'Reply on RC1', Thomas Wagenhäuser, 26 Mar 2021
-
RC2: 'Comment on amt-2020-461', William Thomas Sturges, 17 Feb 2021
This is a clever and useful idea to improve data retrieval from AirCore sampling of the atmosphere, and will be of interest to the (currently small) community of practitioners. AMT is an entirely suitable medium for this report. This appears to have been a very carefully conducted piece of work. One would always like to see more data collected under a variety of conditions (only two flights in this case), but the results appear robust.
My only comment would be that, for the sake of anyone not working directly on AirCores, this would benefit from having a little more explanatory text. E.g. under Section 2.1, a clearer summary of these steps that does not require reference to Engel et al. would make reading/understanding much easier.
A few minor points:
L.15 add “positive” to “bias”
L.16 “shown” not “uncovered”
L.17 “to be represented by possible empirical”
L.19 is it +/- 120 m or +120 offset?
L.50 “needs to be attributed to positional data” – doesn’t it just need altitudinal data? Lat/Long you’d get from GPS, wouldn’t you?
L.90 what is the push gas made of?
Fig. 1 is quite tough to follow unless you have a little more background
L.102 what does “PG resp. a calibration standard” mean? I didn’t understand this.
L104 I am not clear about “only tubing involved at the start of the AirCore measurement is coloured”; what is meant by “involved” – it's all involved isn’t it?
L.117 not clear what “starting point in the analysis” refers to.
L.129 how high is high CO?
L.131 maybe explain how “Cal gas is used to distinguish between PG and FG”?
L.189 what does “fastening valve” mean? I’ve not heard of this before (shutoff valve?).
Fig. 8 It took a while for me to realise that the steps in the curve related to the three diameters of tubing - maybe point this out from the start?
- AC2: 'Reply on RC2', Thomas Wagenhäuser, 26 Mar 2021
-
RC3: 'Comment on amt-2020-461', Anna Karion, 22 Feb 2021
Review of "Testing the altitude attribution and vertical resolution of AirCore
measurements with a new spiking method", by Wagenhauser et al.The manuscript presents a well-written and clear description of a method that can be used to assess and correct the altitude assignments of AirCore measurements and also their vertical resolution. This relatively simple empirical method for correcting for a pressure drop inside the AirCore tubing during descent seems to work very well in the two flights shown. As the authors mention, the extent of applicability of a given linear fit might need to be investigated more in future work, especially around issues such as individual dryer packing. This data can also provide a check of any future more theoretically-based estimation of the pressure drop. It is a valuable study to be included in the literature on AirCore balloon deployments, as altitude assignment is one of the uncertainty components of these measurements.
I only have very minor grammatical comments or clarifications, below.
L188, perhaps I missed this earlier but what is ml_n ? (and again elsehwere including L326, after 1.4 liters (ln?))?L197, rather than "bar" perhaps SI units would be used here (editors can comment on journal policy) (same comment, line 332 using "atm".)
L213, approximately should be spelled out here and elsewhere I believe (editors can comment on that)
L 229 Typo, June 18 is used twice, should be June 17
L243 should be the analyzer's (apostrophe added)
Fig 4 and Fig 5, one is labeled GPS Altitude and one geometric - are these the same thing? (i.e. both based on the GPS reading?). (and same question for other figures - perhaps they should all be made with consistent labeling).
L286: "between the both" should be "between the two"
L335 and on: It would help the reader if discussion of Figure 8 could mention the sharp changes in the modeled resolution occur at the junctions between different diameter parts of the Aircore.
- AC3: 'Reply on RC3', Thomas Wagenhäuser, 26 Mar 2021
Status: closed
-
RC1: 'Comment on amt-2020-461', P. P. Tans, 05 Feb 2021
This paper is a very useful contribution to the literature on AirCore. The spiking idea is new, and the experiment has been carried out carefully, so that the results are solid.
I have a few minor comments:
line 81: The internal diameter is what matters, so please provide that. I assume that the predicted results use the ID.
line 137: The fill gas at the closed end of the tube will not be distributed as a Gaussian. It has to be asymmetric because the end is closed off. What it looks like depends on how much fill gas is left, but close to the end the spatial derivative of each gas fraction has to go to zero. If there is a lot of fill gas left, occupying several diffusion length scales, the fill gas fraction must approach 1 at the closed end. When the fill gas enters the analyzer the transition ought to be rapid, unless the tubing toward the analyzer and the analyzer cell itself cause a lot of mixing. Some years ago I analyzed experiments with "plug" transitions, sudden mole fraction shifts inserted very close to the Picarro. In the hypothetical case that the cell would be perfectly mixed all the time, the insertion of a plug should produce a negative exponential approach toward the new steady state. If there is plug flow within the cell, so that the rapid transition is mostly preserved, the approach to the new state should be linear. It turns out that the actual transition was in between these two cases. The "response function" of each analyzer will depend on pressure, cell volume, and shape of the cell. This subject comes back on line 156.
line 193: I suppose the signal gas mixture is CO-in-natural air. If a larger spike is inserted, you don't want to alter the main gases of interest.
line 234: typo - should be June 17
line 245: "resulting calculated vertical profile" (this is for clarity)
line 350: I would be very surprised if junctions between sections could cause much additional mixing. The flow conditions in the tube are extremely far away from any kind of turbulence. My suggestion is to look further into the analyzer contribution to mixing.
line 387: I suggest replacing "proof," with "prove"
- AC1: 'Reply on RC1', Thomas Wagenhäuser, 26 Mar 2021
-
RC2: 'Comment on amt-2020-461', William Thomas Sturges, 17 Feb 2021
This is a clever and useful idea to improve data retrieval from AirCore sampling of the atmosphere, and will be of interest to the (currently small) community of practitioners. AMT is an entirely suitable medium for this report. This appears to have been a very carefully conducted piece of work. One would always like to see more data collected under a variety of conditions (only two flights in this case), but the results appear robust.
My only comment would be that, for the sake of anyone not working directly on AirCores, this would benefit from having a little more explanatory text. E.g. under Section 2.1, a clearer summary of these steps that does not require reference to Engel et al. would make reading/understanding much easier.
A few minor points:
L.15 add “positive” to “bias”
L.16 “shown” not “uncovered”
L.17 “to be represented by possible empirical”
L.19 is it +/- 120 m or +120 offset?
L.50 “needs to be attributed to positional data” – doesn’t it just need altitudinal data? Lat/Long you’d get from GPS, wouldn’t you?
L.90 what is the push gas made of?
Fig. 1 is quite tough to follow unless you have a little more background
L.102 what does “PG resp. a calibration standard” mean? I didn’t understand this.
L104 I am not clear about “only tubing involved at the start of the AirCore measurement is coloured”; what is meant by “involved” – it's all involved isn’t it?
L.117 not clear what “starting point in the analysis” refers to.
L.129 how high is high CO?
L.131 maybe explain how “Cal gas is used to distinguish between PG and FG”?
L.189 what does “fastening valve” mean? I’ve not heard of this before (shutoff valve?).
Fig. 8 It took a while for me to realise that the steps in the curve related to the three diameters of tubing - maybe point this out from the start?
- AC2: 'Reply on RC2', Thomas Wagenhäuser, 26 Mar 2021
-
RC3: 'Comment on amt-2020-461', Anna Karion, 22 Feb 2021
Review of "Testing the altitude attribution and vertical resolution of AirCore
measurements with a new spiking method", by Wagenhauser et al.The manuscript presents a well-written and clear description of a method that can be used to assess and correct the altitude assignments of AirCore measurements and also their vertical resolution. This relatively simple empirical method for correcting for a pressure drop inside the AirCore tubing during descent seems to work very well in the two flights shown. As the authors mention, the extent of applicability of a given linear fit might need to be investigated more in future work, especially around issues such as individual dryer packing. This data can also provide a check of any future more theoretically-based estimation of the pressure drop. It is a valuable study to be included in the literature on AirCore balloon deployments, as altitude assignment is one of the uncertainty components of these measurements.
I only have very minor grammatical comments or clarifications, below.
L188, perhaps I missed this earlier but what is ml_n ? (and again elsehwere including L326, after 1.4 liters (ln?))?L197, rather than "bar" perhaps SI units would be used here (editors can comment on journal policy) (same comment, line 332 using "atm".)
L213, approximately should be spelled out here and elsewhere I believe (editors can comment on that)
L 229 Typo, June 18 is used twice, should be June 17
L243 should be the analyzer's (apostrophe added)
Fig 4 and Fig 5, one is labeled GPS Altitude and one geometric - are these the same thing? (i.e. both based on the GPS reading?). (and same question for other figures - perhaps they should all be made with consistent labeling).
L286: "between the both" should be "between the two"
L335 and on: It would help the reader if discussion of Figure 8 could mention the sharp changes in the modeled resolution occur at the junctions between different diameter parts of the Aircore.
- AC3: 'Reply on RC3', Thomas Wagenhäuser, 26 Mar 2021
Thomas Wagenhäuser et al.
Thomas Wagenhäuser et al.
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 322 | 110 | 15 | 447 | 3 | 3 |
- HTML: 322
- PDF: 110
- XML: 15
- Total: 447
- BibTeX: 3
- EndNote: 3
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1