the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
10 Feb 2021
10 Feb 2021
Invited perspectives: The ECMWF strategy 2021–2030 Challenges in the area of natural hazards
- European Centre for Medium-range Weather Forecasts, Reading/Bologna/Bonn, UK/Italy/Germany
- European Centre for Medium-range Weather Forecasts, Reading/Bologna/Bonn, UK/Italy/Germany
Abstract. The European Centre for Medium-Range Weather Forecasts mission is to deliver high quality global medium‐range numerical weather predictions and monitoring of the Earth system to its Member States. The modelling and forecasting of natural hazards are an important part of this mission. The new strategy identified several goals: producing world leading weather and earth-system science, cutting edge technology and computational science, high quality products fit for purpose, efficient and easy access to products, and aiming to inspire and hire the best experts. Challenges for the future include integration of innovative observations into the earth system, a realistic representations of water, energy, and carbon cycles, coupling and initialisation of all earth system components, adequate representation of uncertainties and advances in software engineering. Progress in all these areas will need enhanced collaboration with Member States and partners across Europe and beyond.
Florian Pappenberger et al.
Status: final response (author comments only)
-
CC1: 'Comment on nhess-2021-33', Ricardo Tavares da Costa, 11 Feb 2021
Dear authors,
first of all thank you for your paper, I like it. Congratulations for the vision of bringing code to data, instead of data to code and for open the possibility for AI. Here are my notes, hopefully helpful ones:
corrections:
- Paragraph 35, line 4"form cars" > from cars;
- Paragraph 45, line 1 "significant research challenge" > a significant research challenge;
- Paragraph 45, line 1 to 2 "This is closely linked to the challenge of how" > This is closely linked to questions such how to";
- Paragraph 45, line 3 to 4 "Initialising the snowpack on a global scale which is important for flood forecasts is an example of such a challenge." > maybe just "initialising the snowpack on a global scale which is important for flood forecasts"
suggetion:
- Paragraph 40, line 2 "does not conserve mass" > the mass in the system does not remain constant (principle of mass conservation);
- Paragraph 45, line 3 to 4 "Initialising the snowpack on a global scale which is important for flood forecasts is an example of such a challenge." > maybe snow calculation module?
- Paragraph 50, line 3 to 4 "which will allow domain experts (natural hazard scientists)" > maybe just natural hazard scientists
- Paragraph 50, lines 4 to 5 "GPU based architectures are particularly suited to be used by artificial intelligence and machine learning" > well also FGPAs and ARM-based systems this will require you to repharse the subsequent statements about GPUs...
- Paragraph 50, line 3 to 5 and Paragraph 60, line 1 "Other novel ways need to be found to address the need of increasingly compute and storage hungry forecasts and simulations and should therefore also include computing solutions beyond supercomputers" > maybe just forecast that are increasingly demanding in terms of computing and storage... computing solutions beyond supercomputers? OpenIFS@home example not familiar and I did not clearly understand your explanation.
- Paragraph 70, Line 2 "provide maximum benefits to user" > maybe just provide maximum benefits
- Conclusion is ok, but I would be more bold and incisive in this section.
rephrase:
- Paragraph 40, lines 2 to 4 "It is also essential to capture and represent in our predictions as many of tahe model and other uncertainties as necessary: for example, we know that many parts of the Earth system are inadequately observed (Beven et al 2020). However, finite computing and requirements to produce timely forecasts will only allow a limited number of ensemble members to represent these uncertainties." > "model and other uncertainties", maybe just uncertainties? > how does "many parts of the Earth system are inadequately observed", better link to the previous statement about uncertainties > "finite computing and requirements", maybe just requirements (or even resources)? > "ensemble members", maybe just complete set of forecasts;
- Paragraph 45, line 5 and Paragraph 50, lines 1 to 2 "The compute and storage power of novel HPC architectures is required to improve the representation of processes and 50 uncertainties, however existing large code bases to model and forecast natural hazards are ill equipped to scale for such novel architectures." > maybe like this: Novel HPC architectures, both for computing and storage, are required to... at the same time, existing codes and algorithms need to be adpated so that they can take advantage of these new resources (speedup and scaleup of parallel processing);
- Paragraph 80, lines 3 to 5 and Paragraph 85, lines 1 to 2 "Computer, whilst the other tasks are more suited to be executed in a cloud environment, although these different technologies will converge in the future. Research into the efficient managing and orchestrating of workflows spanning these different compute environments will crucially improve the overall performance of a forecast and model change and thus allow a larger proportion of limited human and compute resources to be used to improve natural hazard forecasts and models." > too long, please rephrase
-
AC1: 'Reply on CC1', Florian Pappenberger, 11 Feb 2021
Thank you very much for reading it and all the excellent corrections, suggestions and rephrasing - we will include them all as it will greatly improve the manuscript. You are particular rigth regarding your comment on GPUs (should have been phrased more generic!). We will need some time to think how to me more bold in the conclusion :) !
-
AC4: 'Additional Reply on CC1', Florian Pappenberger, 06 May 2021
The comment was uploaded in the form of a supplement: https://nhess.copernicus.org/preprints/nhess-2021-33/nhess-2021-33-AC4-supplement.pdf
-
CC2: 'Comment on nhess-2021-33', Chloe Brimicombe, 16 Feb 2021
I like to check whether there is equal mention to the hydro-met hazards in pieces that mention all of them and this actually does a really good job of that.
My main concern around ECMWF's strategy is this in the impact section "if users can maximise the usefulness and accessibility of high-quality products and outputs ... skilful predictions of extreme temperature anomalies and hydrological impact" - from this you appear not to be addressing thermal impact as part of your ongoing strategy? And as you are aware heat is a massive challenge in the area of natural hazards, so what is the motivation for this? Because you state you are addressing all of the other hydro-met hazards impacts?
ECMWF has skillful prediction for extremes in temperature, often moreso than precipitation so why not develop and maintain GloHAS, to forecast heat/cold risk? This would properly inform users of the impact which would both be useful and accessible and also could save thousands of lives each year.
-
AC3: 'Reply on CC2', Florian Pappenberger, 26 Mar 2021
Great points Chloe – Heat stress is an important topic and a huge challenge. However, the more immediate focus at ECMWF is to improve inputs on heat stress indicators such as the UTCI. Extreme temperature anomaly is only one factor in this. It is correct that we do not directly address thermal impact and that is in only addressed indirectly. I feel that improving inputs to heat stress indicators is already an ambitious goal. The sensitivity to the different inputs for heat indicators will be part of resource distribution in efforts.
A Global Heat Awareness System would be great. There are some significant efforts in Copernicus@ECMWF (i.e. ) but more can be done. We will need partners to achieve that.
- AC6: 'Reply on CC2', Florian Pappenberger, 06 May 2021
-
AC3: 'Reply on CC2', Florian Pappenberger, 26 Mar 2021
-
RC1: 'Comment on nhess-2021-33', Anonymous Referee #1, 23 Feb 2021
The manuscript “Invited perspectives: The ECMWF strategy 2021-2030 Challenges in the area of natural hazards” provides a very interesting concise overview of the ECMWF strategy in respect to natural hazards. The description of the envisaged step changes necessary to tackle the identified key challenges for the future are interesting for the international scientific community also beyond Europe. The challenges are quite generic and relevant not only for ECMWF but for many organisations dealing with forecasting weather and resulting natural hazards, as well as with climate reanalyses and projections. The described step changes can inspire scientific studies and advancements in various domains related to natural hazards.
From my point of view, this manuscript is already valuable and well written, however, with a stronger harmonization of terms, an even more clear structure, particularly a closer link between the identified challenges and the envisaged solutions, this manuscript can become significantly more interesting for the scientific community. Thus, I suggest the following:
Abstract: I would find it more convincing if first the challenges are listed, and then afterwards the goals (which are set to tackle these challenges). When reading the goals, I would expect, that you/ECMWF see the additional challenges of designing user specific products and means of communication and decision support under high uncertainties (or similar).
I suggest to group the goals into the three pillars already in the abstract, this would prepare better for what is coming later. Additionally, I suggest to use harmonized terms, not all interchangeably “goals”, “step changes”, “vision”.
I suggest to provide a closer link between the challenges and the envisaged solutions. E.g. in the part 2 science and technology:
Line 43: “parts of the Earth system are inadequately observed (Beven et al 2020).” -> would be good to mention which ones, so that one has at least an idea without needing to read the paper by Beven.
Line 43: “However, finite computing and requirements to produce timely forecasts will only allow a limited number of ensemble members to represent these uncertainties.” -> are there other means of representing uncertainties, considering uncertainties besides ensembles? What it’s the suggestion or idea how to tackle this challenge?
Line 47: “Initialising the snowpack on a global scale which is important for flood forecasts is an example of such a challenge.” -> would be great if first ideas of how to tackle this challenge can be presented.
Part 3 Impacts:
Line 63: “ECMWF will aim to provide detailed Earth system simulations of the past, present and future with a particular focus on extreme events for several weeks ahead” – This sentence is confusing. I guess that you have two separate tasks, one is rather long-term simulations of the earth system in the past and future, e.g. providing climate change projections for the future, also future scenarios of natural hazards, etc. The other one is early warning of natural hazards and for this the aim is to extent the lead time to more than two weeks. In case this is correct, it would be good to write this more clearly in the manuscript, and structure the impacts part accordingly.
In this respect, I also guess, that you have different user groups for these two activities? It would be good if you could write a sentence about who your user groups are (and maybe for which products).
The idea of a user-oriented evaluation seems interesting, it would be interesting to know how this could be realized and how it would influence your work/products. Is it meant to be further developed towards co-design of products together with the users?
Lines 77-86 seem to belong more to the technology part, or it needs to be more focused on the visualization and communication aspect.
Line 85 “model chain” not “model change”
The part 4 on people is limited to the identification of the challenge. Any ideas for the way forward?
-
AC2: 'Request for Clarification by Reviewer - please', Florian Pappenberger, 26 Mar 2021
I will adrress the other comments seperately. However, there is one revieer comment, I am not sure I fully understand:
>>> comment
Line 43: “However, finite computing and requirements to produce timely forecasts will only allow a limited number of ensemble members to represent these uncertainties.” -> are there other means of representing uncertainties, considering uncertainties besides ensembles?
>>>
I am not aware that one could represent uncertainties of a non-linear/chaotic flow dependent system differently than through an emsemble/monte carlo approach. Sure, Post-processing or dressing or similar will also get you a distribution but these are not really the 'same' uncertaintis.
So, I think i am missing the reviewers point (sorry for that) - can you please clarify?
-
RC3: 'Reply on AC2', Anonymous Referee #1, 08 Apr 2021
I was just curious if there is any possibility to tackle this challenge? Given that ensembles are needed to quantify the uncertainties; is it possible to define how many ensemble members are needed to reliably quantify uncertainty? Maybe via post- or pre-processing? Is it possible to further increase computer power, or make models faster? In case the "limited number of ensemble members" is sufficient to represent uncertainties, and there are no plans for improving this situation, then please ignore my comment.
-
RC3: 'Reply on AC2', Anonymous Referee #1, 08 Apr 2021
-
AC2: 'Request for Clarification by Reviewer - please', Florian Pappenberger, 26 Mar 2021
-
RC2: 'Comment on nhess-2021-33', Anonymous Referee #2, 06 Apr 2021
A review of the Invited perspective:
The ECMWF strategy 2021-2030 Challenges in the area of natural hazards
by Florian Pappenberger and co-authors
The perspective provides a clear picture of the ECMWF’s intentions for the next decade. It describes the intended progress from the present practice, describing several key components for a development of capabilities rather than a single step change. These include, in particular, data availability and use, system understanding, computer resources and programming techniques, and human capital.
My impression is that just a few minor improvements of readability of this perspective are possible, which I mentiion below.
Abstract: I am aware that the collaboration with the member states (and further partners) is important, but this aspect is not taken up in the rest of the manuscript. Would part of the strategy, for example, include a development of the collaboration with member states beyond the present practice, for example in terms of computing, use of results, or personnel?
Introduction: As the conclusions mention forecasts of high impact weather at the medium-range time scale of a couple of weeks, I wonder if they it would be worthwhile also to mention them in the introductory section.
Science and Technology: Is the mentioning on remote sensing from satellites and IoT-data, but not of Radar, intentional?
People: While the job market aspect is clearly important, I wonder if cooperation aspects (e.g., with the member states) might be worthwhile to be briefly mentioned.
-
AC5: 'Reply on RC2', Florian Pappenberger, 06 May 2021
The comment was uploaded in the form of a supplement: https://nhess.copernicus.org/preprints/nhess-2021-33/nhess-2021-33-AC5-supplement.pdf
-
AC5: 'Reply on RC2', Florian Pappenberger, 06 May 2021
Florian Pappenberger et al.
Florian Pappenberger et al.
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 597 | 165 | 25 | 787 | 1 | 1 |
- HTML: 597
- PDF: 165
- XML: 25
- Total: 787
- BibTeX: 1
- EndNote: 1
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1