the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 22 May 2024
The impact of synoptic storm likelihood on European subseasonal forecast uncertainty and their modulation by the stratosphere
Abstract. Weather forecasts at subseasonal-to-seasonal (S2S) timescales have little or no forecast skill in the troposphere: individual ensemble members are uncorrelated and span a range of atmospheric evolutions that are possible for the given set of external forcings. The uncertainty of such a probabilistic forecast is then determined by this range of possible evolutions – often quantified in terms of ensemble spread. Various dynamical processes can affect the ensemble spread within a given region, including extreme events simulated in individual members. For surface pressure or geopotential height forecasts over Europe, such extremes are mainly comprised of synoptic storms propagating along the North Atlantic storm track. We use ECMWF re-forecasts from the S2S database to investigate the connection between different storm characteristics and ensemble spread in more detail. We find that the presence of storms in individual ensemble members at S2S time scales contributes about 20 % to the total geopotential height forecast uncertainty over Northern Europe. Furthermore, certain atmospheric conditions associated with substantial anomalies in the North Atlantic storm track show reduced geopotential height ensemble spread over Northern Europe. For example, during periods with a weak stratospheric polar vortex, the genesis frequency of Euro-Atlantic storms is reduced and their tracks move equatorwards. As a result, we find weaker storm magnitudes and lower storm counts, and hence anomalously low subseasonal ensemble spread, over Northern Europe.
-
Notice on discussion status
The requested preprint has a corresponding peer-reviewed final revised paper. You are encouraged to refer to the final revised version.
-
Preprint
(3888 KB)
-
The requested preprint has a corresponding peer-reviewed final revised paper. You are encouraged to refer to the final revised version.
- Preprint
(3888 KB) - Metadata XML
- BibTeX
- EndNote
- Final revised paper
Journal article(s) based on this preprint
Interactive discussion
Status: closed
-
RC1: 'Comment on egusphere-2024-1423', Anonymous Referee #1, 09 Jul 2024
GENERAL COMMENT
This manuscript analyses the relation between strong storms and forecast uncertainty, demonstrating how the prediction of storms, associated with strong geopotential height anomalies, affects the spread of ensemble forecasts in the Euro-Atlantic sector. Even though the ECMWF hindcast ensemble is small, as argued by the authors the results should naturally extend to larger ensembles.
This outcome is applied to the case of weak and strong stratospheric vortex state. The weaker frequency and depth of strong storms in the Euro-Atlantic sector (and Northern Europe) during weak stratospheric vortex conditions likely explains the decrease in uncertainty that follows these events.
Although the topic is original and interesting, the manuscript still requires substantial improvements in terms of linearity of the discussion and, ultimately, readability. Therefore, I encourage the authors to work specially on the structure of the text and on the precision of the langage, specially in the Introduction and Discussion sections. Other methodological and technical comments are included below. Once these issues have been addressed, I am willing to recommend publication of this work in Weather and Climate Dynamics.
SPECIFIC COMMENTS
Introduction and Discussion :
- From reading these sections, I was expecting the whole work to be focused on SPV impacts on uncertainty. The main results, however, are those relating the presence of strong storms and forecast uncertainty. Could you work on these in order to make the focus of your work clearer to the reader ?
- Both are rather dishomogeneous and difficult to follow. Connections between the different paragraphs and topics are often lacking.
Figure 1 : for an equal comparison between the ensemble spread with and without storms the size of the ensemble must be consistent. For example, for the "with storms" case you can compute a line (or better, a distribution of lines) where the ensemble includes storm cases, plus a number of random members necessary to equal the size of the "without storm" ensemble.
Figure 5 : see comment to Figure 1.
Figures 6a and 7b : The fact that storm strength is high for small Pmin is confusing. Please address this issue as you think best.
TECHNICAL CORRECTIONSLine 19 : "prescribed" or "initialised" in the case of S2S ?
Lines 21-23 : Sentence is clear but quite convoluted. Please improve.
Lines 32-33 : Do you want to say that uncertainties in the position of the jet (and its associated baroclinic zone) lead to a large ensemble spread in the mid latitudes ? Is there a difference in laying the focus on the jet or on the strong pressure gradients ? In my opinion, they just correspond to different aspects of the same weather feature.
Line 34 : Maybe "synoptic-to-planetary scale" ?
Line 36 : Since you are introducing the focus of this work, maybe make new paragraph ?
Line 57 : "NAO-like".
Line 58 : "and OF THE North Atlantic storm track".
Line 59-61 : This sentence is rather vague. Please consider removing if outside of the scope and topic of this paper.
Lines 90-93 : I find these two sentences confusing.
- Why do you separate the 2020-21 hindcast from the hindcasts in previous years ?
- Could you clarify the sentence "Each real-time initialisation provides hindcasts for the previous 20 years". Is this again 10-member hindcasts over the previous 20 years ?
- It would be useful to state explicitely which terminology is used to distinguish the runs initialised in 2000-2019 and the ensemble runs from 2020-2021, if there is a distinction. Also, what does the term "forecasts" in the rest of the text refer to ? Note the inconsistency with the term "hindcasts" used to present the dataset.
Line 123-124 : This sentence is an overstatement in my opinion. Maybe mention "intense" or "strong" storms ?
Line 131 : Specify "Z1000 distribution".
Line 135 : Specify the target domain from the beginning. Please add a motivation for choosing this region.
Line 143 : How is the vicinity of the storm to the target domain defined ?
Line 158 : The case study is for "Northern Europe" right?
Line 161 : You have not yet defined the "Euro-Atlantic sector".
Line 163 : Specify "a single storm CENTRE is therefore...". Further away from the centre, the percentage increase in spread is weaker.
Line 172 : Could you justify why you go back to analysing Northern Europe, since the previous analyses in this subsection referred to the Euro-Atlantic sector ?
Lines 238-240 : A bit out of the blue here. Specify connection with previous or shift to a different paragraph.
Line 244-246 : Acknowledge briefly that reduction in uncertainty does not necessarily mean an improvement in the predictions.
Line 257-258 : It’s a bit strange to consider Rossby waves and storms in opposition, since storms can be seen as Rossby-wave troughs. Please consider this while revising the text.
Line 259 : Correct the syntax of the last part of the sentence.
Line 260- : I believe a correct representation of the climatological storm track is not enough to ensure spread and error to be proportional. For example, incorrect model response to the initialised/prescribed forcing (because of lacking model physics, e.g. parametrisation), or initialisation issues (errors, sampling, ecc) may also result in forecast errors or may affect the spread.
Line 280 : This last sentence is vague and inconclusive. Please specify how storm forecast and risk mitigation can benefit from the outcomes of this work.
Citation: https://doi.org/10.5194/egusphere-2024-1423-RC1 -
AC1: 'Reply on RC1', Philip Rupp, 24 Aug 2024
The comment was uploaded in the form of a supplement: https://egusphere.copernicus.org/preprints/2024/egusphere-2024-1423/egusphere-2024-1423-AC1-supplement.pdf
-
RC2: 'Comment on egusphere-2024-1423', Anonymous Referee #2, 16 Jul 2024
This paper looks at the impact that storms have on ensemble spread of S2S forecasts, focussing on Northern Europe/the North Atlantic region. The results show that the occurrence of strong storms contributes around 20% of the Z1000 spread. The study then moves on to look at the impacts of strong and weak SPV states on storms and ensemble spread. Overall the paper is well written, with clear and useful figures, and a logical structure. I have only a couple of minor comments.
Minor comments:
Fig 1: it is interesting to see that on some days the spread without storms is (albeit slightly) higher than the spread with storms: do you have any ideas why this would happen? I can see that it happens when the Z1000 spread of the active storm members is within the range of other members, but in a dynamical sense how would you interpret this?
Fig 4 a and b: I didn’t entirely understand what is plotted here or follow the interpretation about this - Please could you try to clarify more in the text.
Typos:
Fig 4 caption: anomay – should be anomaly
Fig 5 caption: Evolution of …
Citation: https://doi.org/10.5194/egusphere-2024-1423-RC2 -
AC2: 'Reply on RC2', Philip Rupp, 24 Aug 2024
The comment was uploaded in the form of a supplement: https://egusphere.copernicus.org/preprints/2024/egusphere-2024-1423/egusphere-2024-1423-AC2-supplement.pdf
-
AC2: 'Reply on RC2', Philip Rupp, 24 Aug 2024
Interactive discussion
Status: closed
-
RC1: 'Comment on egusphere-2024-1423', Anonymous Referee #1, 09 Jul 2024
GENERAL COMMENT
This manuscript analyses the relation between strong storms and forecast uncertainty, demonstrating how the prediction of storms, associated with strong geopotential height anomalies, affects the spread of ensemble forecasts in the Euro-Atlantic sector. Even though the ECMWF hindcast ensemble is small, as argued by the authors the results should naturally extend to larger ensembles.
This outcome is applied to the case of weak and strong stratospheric vortex state. The weaker frequency and depth of strong storms in the Euro-Atlantic sector (and Northern Europe) during weak stratospheric vortex conditions likely explains the decrease in uncertainty that follows these events.
Although the topic is original and interesting, the manuscript still requires substantial improvements in terms of linearity of the discussion and, ultimately, readability. Therefore, I encourage the authors to work specially on the structure of the text and on the precision of the langage, specially in the Introduction and Discussion sections. Other methodological and technical comments are included below. Once these issues have been addressed, I am willing to recommend publication of this work in Weather and Climate Dynamics.
SPECIFIC COMMENTS
Introduction and Discussion :
- From reading these sections, I was expecting the whole work to be focused on SPV impacts on uncertainty. The main results, however, are those relating the presence of strong storms and forecast uncertainty. Could you work on these in order to make the focus of your work clearer to the reader ?
- Both are rather dishomogeneous and difficult to follow. Connections between the different paragraphs and topics are often lacking.
Figure 1 : for an equal comparison between the ensemble spread with and without storms the size of the ensemble must be consistent. For example, for the "with storms" case you can compute a line (or better, a distribution of lines) where the ensemble includes storm cases, plus a number of random members necessary to equal the size of the "without storm" ensemble.
Figure 5 : see comment to Figure 1.
Figures 6a and 7b : The fact that storm strength is high for small Pmin is confusing. Please address this issue as you think best.
TECHNICAL CORRECTIONSLine 19 : "prescribed" or "initialised" in the case of S2S ?
Lines 21-23 : Sentence is clear but quite convoluted. Please improve.
Lines 32-33 : Do you want to say that uncertainties in the position of the jet (and its associated baroclinic zone) lead to a large ensemble spread in the mid latitudes ? Is there a difference in laying the focus on the jet or on the strong pressure gradients ? In my opinion, they just correspond to different aspects of the same weather feature.
Line 34 : Maybe "synoptic-to-planetary scale" ?
Line 36 : Since you are introducing the focus of this work, maybe make new paragraph ?
Line 57 : "NAO-like".
Line 58 : "and OF THE North Atlantic storm track".
Line 59-61 : This sentence is rather vague. Please consider removing if outside of the scope and topic of this paper.
Lines 90-93 : I find these two sentences confusing.
- Why do you separate the 2020-21 hindcast from the hindcasts in previous years ?
- Could you clarify the sentence "Each real-time initialisation provides hindcasts for the previous 20 years". Is this again 10-member hindcasts over the previous 20 years ?
- It would be useful to state explicitely which terminology is used to distinguish the runs initialised in 2000-2019 and the ensemble runs from 2020-2021, if there is a distinction. Also, what does the term "forecasts" in the rest of the text refer to ? Note the inconsistency with the term "hindcasts" used to present the dataset.
Line 123-124 : This sentence is an overstatement in my opinion. Maybe mention "intense" or "strong" storms ?
Line 131 : Specify "Z1000 distribution".
Line 135 : Specify the target domain from the beginning. Please add a motivation for choosing this region.
Line 143 : How is the vicinity of the storm to the target domain defined ?
Line 158 : The case study is for "Northern Europe" right?
Line 161 : You have not yet defined the "Euro-Atlantic sector".
Line 163 : Specify "a single storm CENTRE is therefore...". Further away from the centre, the percentage increase in spread is weaker.
Line 172 : Could you justify why you go back to analysing Northern Europe, since the previous analyses in this subsection referred to the Euro-Atlantic sector ?
Lines 238-240 : A bit out of the blue here. Specify connection with previous or shift to a different paragraph.
Line 244-246 : Acknowledge briefly that reduction in uncertainty does not necessarily mean an improvement in the predictions.
Line 257-258 : It’s a bit strange to consider Rossby waves and storms in opposition, since storms can be seen as Rossby-wave troughs. Please consider this while revising the text.
Line 259 : Correct the syntax of the last part of the sentence.
Line 260- : I believe a correct representation of the climatological storm track is not enough to ensure spread and error to be proportional. For example, incorrect model response to the initialised/prescribed forcing (because of lacking model physics, e.g. parametrisation), or initialisation issues (errors, sampling, ecc) may also result in forecast errors or may affect the spread.
Line 280 : This last sentence is vague and inconclusive. Please specify how storm forecast and risk mitigation can benefit from the outcomes of this work.
Citation: https://doi.org/10.5194/egusphere-2024-1423-RC1 -
AC1: 'Reply on RC1', Philip Rupp, 24 Aug 2024
The comment was uploaded in the form of a supplement: https://egusphere.copernicus.org/preprints/2024/egusphere-2024-1423/egusphere-2024-1423-AC1-supplement.pdf
-
RC2: 'Comment on egusphere-2024-1423', Anonymous Referee #2, 16 Jul 2024
This paper looks at the impact that storms have on ensemble spread of S2S forecasts, focussing on Northern Europe/the North Atlantic region. The results show that the occurrence of strong storms contributes around 20% of the Z1000 spread. The study then moves on to look at the impacts of strong and weak SPV states on storms and ensemble spread. Overall the paper is well written, with clear and useful figures, and a logical structure. I have only a couple of minor comments.
Minor comments:
Fig 1: it is interesting to see that on some days the spread without storms is (albeit slightly) higher than the spread with storms: do you have any ideas why this would happen? I can see that it happens when the Z1000 spread of the active storm members is within the range of other members, but in a dynamical sense how would you interpret this?
Fig 4 a and b: I didn’t entirely understand what is plotted here or follow the interpretation about this - Please could you try to clarify more in the text.
Typos:
Fig 4 caption: anomay – should be anomaly
Fig 5 caption: Evolution of …
Citation: https://doi.org/10.5194/egusphere-2024-1423-RC2 -
AC2: 'Reply on RC2', Philip Rupp, 24 Aug 2024
The comment was uploaded in the form of a supplement: https://egusphere.copernicus.org/preprints/2024/egusphere-2024-1423/egusphere-2024-1423-AC2-supplement.pdf
-
AC2: 'Reply on RC2', Philip Rupp, 24 Aug 2024
Peer review completion
Journal article(s) based on this preprint
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 346 | 83 | 27 | 456 | 20 | 17 |
- HTML: 346
- PDF: 83
- XML: 27
- Total: 456
- BibTeX: 20
- EndNote: 17
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
Philip Rupp
Jonas Spaeth
Hilla Afargan-Gerstman
Dominik Büeler
Michael Sprenger
Thomas Birner
The requested preprint has a corresponding peer-reviewed final revised paper. You are encouraged to refer to the final revised version.
- Preprint
(3888 KB) - Metadata XML