the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 27 May 2024
Observing atmospheric rivers using GNSS radio occultation data
Abstract. Atmospheric Rivers (AR) are comparatively narrow regions in the atmosphere that are responsible for most of the horizontal transport of water vapor in the extra tropics, which are responsible for many extreme precipitation events and floodings at mid-latitudes, including Europe and the US. The critical role of ARs in global moisture transport and precipitation dynamics necessitates accurate water vapor measurements for both understanding and forecasting these phenomena. While the integrated water vapor content (IWV) of ARs can be well measured with microwave and infrared sounders, the vertical structure is less well known. In this study, we analysed if specific humidity profiles and IWV values from Global Navigation Satellite System Radio Occultation (GNSS-RO) measurements provide additional information for the study of ARs, in particular regarding their vertical structure. The retrieval of water vapor from GNSS-RO data requires background information, which is usually incorporated by the one-dimensional variational method (1D-Var) that combine observations and background in an optimal manner. We compared data from the COSMIC Data Analysis and Archive Centre (CDAAC), operated by the University Corporation for Atmospheric Research (UCAR) in Boulder, Colorado with data from the Wegener Center for Climate and Global Change (WEGC) at the University of Graz, Austria. We found that retrievals from both centres agree very well in the altitude range, where the 1D-Var weights the observations strongly, even if the employed background profiles are very different. This demonstrates that GNSS-RO data provide indeed additional vertically-resolved information, which was not already contained in the background or in operational analyses. IWV values from CDAAC and WEGC agree generally very well, however, both tend to underestimate the values obtained by Special Sensor Microwave Imager/Sounder (SSMI/S) data, since GNSS-RO profiles not always reach the lowermost part of the atmosphere, leading to a systematic bias in the IWV data, which decreases with better penetration characteristics of the GNSS-RO data. The results suggest that is promising to combine the GNSS-RO data – with very high vertical resolution with SSMI/S data – with high horizontal resolution to get a more compete view of the 3D structure of ARs.
- Preprint
(25167 KB) - Metadata XML
- BibTeX
- EndNote
Status: final response (author comments only)
-
RC1: 'Comment on amt-2024-81', Anonymous Referee #1, 21 Jun 2024
As a limb sounding remote sensing system, radio occultation observations have been used to characterize atmosphere with high vertical resolution for more than two decades. The basic measurements RO provides is refractivity for each layer, but the subsequent temperature water vapor retrieval relies on the auxiliary dataset (a-priori) using variational method. In this manuscript the authors vertically integrate the RO water vapor retrieval to acquire integrate water vapor content (IWV), and compare the its values between different centers (UCAR and WEGC) and collocated passive microwave sensor observations (SSMI/S). The comparisons in this manuscript locate at numbers of selected atmospheric river (AR) regions, where the moisture structure plays a critical role in characterizing the phenomenon.
The authors concluded that the RO water vapor retrieval using different backgrounds does not significantly change the results, which suggests the current RO observation does preserve the information that’s not subject to the a-priori. In addition, the general agreements between RO and SSMI/S (except slight low bias cost by RO penetration) also validates the reliability of RO water vapor retrieval and implies possible combination of both observations. Based on its detailed comparison design, convincing results, and importance to the field I recommend this article to be published after minor revision.
General comments:
- Introduction: is there any existing literatures showing high vertical resolution moisture retrieval, instead of just IWV, can significantly improve the AR forecast? If so, it will be more convincing to include them.
- I agree that combining the MWR and RO can better resolve the water vapor in both dimensions and compensate the missing data gap of RO at lower troposphere, but it is not directly related to the work presented in the manuscript. It will be good to strengthen the link between these two and state why this is a good idea based on the results you get. For example, “the consistent RO and SSMI/S IWV values shown in this manuscript demonstrate the possibility of combining these two observations within the current variational method framework”.
- It's better to point out which physical observation is used in 1DVar for each center: refractivity, bending angle, or phase delay?
- It seems the sensitivity to the vapor a-priori is somewhat mixed – Fig. 4(a) is small, Fig. 6(a) is large. Some lines stress the insensitivity of the retrieval from various backgrounds (L21, L352, L726), others saying the opposite (L623, L656). I suggest having a consistent, clear view on this to avoid confusing the readers. Besides, RO data could be biased (especially in the lower troposphere as shown in Fig. 15 and 16 at 0-2 km) too. This factor should also be included in the discussion of background/data impacts on 1DVar.
Minor comments:
- Line 284: By this definition how do we interpret RAER when it is 100% (above 10km)? The RO observation uncertainty is extremely high in this altitude range and the retrieval mainly follows the background? It will be helpful to explain further on what RAER represents.
- Line 314: Should be Fig. 3(d) instead of Fig. 3(e) ?!
- Line 459: Interesting results. Both centers should use the same data - is there a BA or N retrieval difference between the centers?
- Line 472: period between “stable” and “Here”.
- Line 546: What is the anticipated IWV considering the missing lowest 200m of RO data? What could be the possible cause? This is maybe an open question but I suggest to discuss it if it is being mentioned.
- Line 578: What is the reason of different numbers? Is it because the cut-off height very different between UCAR and WEGC?
- Line 624: “This emphasizes the importance of the choice of background profiles and how they can influence the retrieved IWV values”: Not sure if this is a correct statement. Based on the results it seems ERA has a wet bias, and CDAAC retrieval bring it back to truth so they are drier than ERA. When ECH is used without the wet bias, CDAAC retrieval matches them better. So no matter which background profile is used the retrieved IWV is statistically not sensitive to it, is it correct? If so the background profiles should influence the IWV bias compared to the background, rather than the retrieved IWV. I recommend clarifying the sentence.
- Line 625: Figure 13 instead 11?!
- Line 628: There is actually no CDAAC and WEGC IWV data comparison in the section 4.3.2.. And in section 4.2 this specific case (Iceland -UK 2009) is not shown. I suggested to add this case in 4.2 to validate the statement.
- Line 655: It will be interesting to bring up the COSMIC-2 dry bias issue (Line 457) here again since it is the same case. In Fig. 14 C2 does not show an obvious bias w.r.t the background for both CDAAC and WEGC. However in Fig.7 C2 from CDAAC is obviously dryer than the one from WEGC. Does it indicate that the ECMWF-b is dryer than corresponding ECH profiles for C2?
- Line 670: It will be clearer to define the RSHD with an equation.
- Line 731: should this sentence be in the same paragraph above?
Citation: https://doi.org/10.5194/amt-2024-81-RC1 -
RC2: 'Comment on amt-2024-81', Anonymous Referee #2, 10 Sep 2024
Review on “Observing atmospheric rivers using GNSS radio occultation data”, by Bahareh Rahimi and Ulrich Foelsche
General comments.
In this paper, the authors analyze specific humidity profiles and IWV values from RO data measurements for the study of atmosperic rivers (ARs), focusing on their vertical structure. They consider data obtained at CDAAC and at WEGC, concluding that GNSS-RO data provide indeed additional vertically-resolved information, which was not already contained in the background or in operational analyses. IWV values from CDAAC and WEGC tend to underestimate the SSMI/S data, since GNSS-RO profiles not always reach the lowermost part of the atmosphere. The authors suggest that is promising to combine the GNSS-RO data (high vertical resolution) with SSMI/S data (high horizontal resolution) to get a more compete view of the 3D structure of ARs.
This is an interesting intercomparison between different techniques with data processed at different centres. Some recommendations are made, mainly for obtaining specific humidity profiles in the lower troposphere and in relation to the different calculations of tangent point trajectories and reference points. The authors conclude that their results contribute to the understanding of atmospheric moisture profiles and set a direction for future research. It is not clear to me what that specific direction is for a more accurate and complete understanding of ARs.
In an AR the moisture is transported along narrow corridors, often driven by large-scale weather patterns like cyclones. The process typically starts when warm ocean waters evaporate and the resulting moisture is lifted into the atmosphere. When these rivers encounter land or mountains, the moisture condenses, leading to heavy rainfall or snowfall. Therefore, it would be very illustrative to indicate the synoptic conditions corresponding to each case study.
On the other hand, we know that mesoscale models, like WRF, have been instrumental in advancing our understanding of ARs, particularly with regard to their dynamics. There are numerous studies on ARs using mesoscale models. These are capable of simulating meteorological phenomena on scales ranging from a few km to hundreds of km and are particularly valuable for understanding the detailed structure and dynamics of ARs. In particular, they are able to resolve important features of ARs, such as their interaction with topography, the development of precipitation bands and the processes leading to extreme precipitation and flooding. I would then suggest that the authors indicate to what extent the results presented in these case studies help to understand, through the combination of GNSS-RO data with SSMI/S, data, the structure and dynamics of ARs and the processes leading to extreme rainfall and flooding.
The results obtained in this paper seem to be useful in a more general context and not particularly pertinent to RAs. In summary, what can be concluded about the generation and evolution of RAs from RO data in addition to satellite data that cannot be inferred or forecasted from mesoscale models?
Scientific significance: Good. Scientific quality: Good. Presentation quality: Good.
The paper address relevant scientific questions within the scope of AMT. It presents novel concepts. Additional conclusions could have been reached, specifically to the knowledge of the dynamics of the atmospheric rivers. The scientific methods and assumptions are clearly outlined. The description of experiments and calculations are sufficiently complete and precise to allow their reproduction. The authors give proper credit to related work and clearly indicate their own contribution. Perhaps the seminar paper by Bevis et al (JGR, 1992) could have been included too. The title should include both complement techniques, not only RO data. The abstract provide a concise summary. The overall presentation is well structured. I am not native so I cannot comment about the English language. The mathematical formulae, symbols, abbreviations, and units are correctly defined and used. The references are appropriate.
Line 314 > Figure 3(d).
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 260 | 73 | 37 | 370 | 13 | 17 |
- HTML: 260
- PDF: 73
- XML: 37
- Total: 370
- BibTeX: 13
- EndNote: 17
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1