Space Weather July 25, 2023

Project Leads: Mathew Owens (University of Reading), Luke Barnard (University of Reading) Full questionnaire: .pdf

Space Weather

We asked 144 experts in space weather about the risks of geomagnetic storms and other key topics in their field.

Summary

Over half of experts (51%) believed it is possible we could experience geomagnetic storms that are moderately or significantly larger than the largest storms over the past 200 years, including the 1859 Carrington Event.
On average, participants thought there was a 33% chance that within the next 10 years a space weather event could cause unplanned regional power outages, although there was a very broad range of opinions.
If an extreme geomagnetic storm occurred, operators at power companies would need advance notice to take appropriate action. 40% of participants thought that it was “highly unlikely” or “somewhat unlikely” that space weather forecasts would be “sufficiently accurate” for these end users to take effective action.
To improve our forecasts, experts emphasized the need for investment in the near-sun heliosphere and near-earth heliosphere compared to other physical domains, with 47% of experts ranking one of those two domains as having the highest return on investment.

Question 1

If the Carrington event were to occur today without any warning, what do you think the impact would be? (Select as many as apply)

Results

Participant	Response	Confidence
Edmund Henley Met Office (views my own)	Regional power outages Widespread radio communication issues Widespread loss of internet Space hardware damage/failure Regional power outages Widespread radio communication issues Widespread loss of internet Space hardware damage/failure
Health risk: assuming not a significant uptick in radiation-related health risk (e.g. cancer). Ignoring “health risk = death” from any radiation-related degradation of avionics / HF impacts on aviation. See RAEng report on Extreme space weather.
Anonymous	Regional power outages Widespread radio communication issues Space hardware damage/failure Regional power outages Widespread radio communication issues Space hardware damage/failure

Ryan McGranaghan NASA Jet Propulsion Laboratory	Regional power outages Widespread radio communication issues Space hardware damage/failure Regional power outages Widespread radio communication issues Space hardware damage/failure

Anonymous	Regional power outages Widespread radio communication issues Space hardware damage/failure Significant health risk to airline crew/passengers Regional power outages Widespread radio communication issues Space hardware damage/failure Significant health risk to airline crew/passengers

Dr Mario M. Bisi UKRI STFC RAL Space	Regional power outages Widespread radio communication issues Space hardware damage/failure Significant health risk to airline crew/passengers Regional power outages Widespread radio communication issues Space hardware damage/failure Significant health risk to airline crew/passengers
Ultimately, this all depends on the timing of the events (any associated SEPs, radio blackout(s), geomagnetic storm(s), etc...) with regards to the time of the year and the Earth's rotation (what countries are on the Sun side and night side, etc...).
Anonymous	Regional power outages Widespread radio communication issues Space hardware damage/failure Regional power outages Widespread radio communication issues Space hardware damage/failure

Ian Mann University of Alberta	Regional power outages Widespread radio communication issues Space hardware damage/failure Regional power outages Widespread radio communication issues Space hardware damage/failure

Anonymous	Regional power outages Widespread radio communication issues Widespread loss of internet Space hardware damage/failure Significant health risk to airline crew/passengers Regional power outages Widespread radio communication issues Widespread loss of internet Space hardware damage/failure Significant health risk to airline crew/passengers

Anonymous	Regional power outages Widespread radio communication issues Space hardware damage/failure Regional power outages Widespread radio communication issues Space hardware damage/failure
Experience indicates that extreme geospace reactions are the result of a confluence of factors (i.e. timing, preconditioning, etc) and not dependent on the transient properties as much
Anonymous	Regional power outages Widespread radio communication issues Space hardware damage/failure Regional power outages Widespread radio communication issues Space hardware damage/failure

Anonymous	Regional power outages Widespread radio communication issues Widespread loss of internet Space hardware damage/failure Regional power outages Widespread radio communication issues Widespread loss of internet Space hardware damage/failure
Power outages would probably be regional, but occurring on several continents. Assessing satellite impacts is especially tricky, since the space-age technology has experienced only storms much smaller than the Carrington event.
Anonymous	Regional power outages Widespread radio communication issues Space hardware damage/failure Significant health risk to airline crew/passengers Regional power outages Widespread radio communication issues Space hardware damage/failure Significant health risk to airline crew/passengers

Anonymous	Regional power outages Widespread radio communication issues Space hardware damage/failure Regional power outages Widespread radio communication issues Space hardware damage/failure

Tamas Gombosi University of Michigan	Regional power outages Widespread radio communication issues Widespread loss of internet Space hardware damage/failure Significant health risk to airline crew/passengers Regional power outages Widespread radio communication issues Widespread loss of internet Space hardware damage/failure Significant health risk to airline crew/passengers

Gabor Toth University of Michigan	Regional power outages Continental power outages Widespread radio communication issues Widespread loss of internet Space hardware damage/failure Regional power outages Continental power outages Widespread radio communication issues Widespread loss of internet Space hardware damage/failure

Harlan Spence University of New Hampshire	Regional power outages Continental power outages Widespread radio communication issues Widespread loss of internet Space hardware damage/failure Significant health risk to airline crew/passengers Regional power outages Continental power outages Widespread radio communication issues Widespread loss of internet Space hardware damage/failure Significant health risk to airline crew/passengers
All or most of these are legitimate risks, plus more not included in this short list!
Anonymous	Regional power outages Widespread radio communication issues Space hardware damage/failure Regional power outages Widespread radio communication issues Space hardware damage/failure

Robert F. Wimmer-Schweingruber Kiel University, Kiel, Germany	Continental power outages Widespread radio communication issues Widespread loss of internet Continental power outages Widespread radio communication issues Widespread loss of internet

Anonymous	Regional power outages Continental power outages Widespread radio communication issues Widespread loss of internet Space hardware damage/failure Regional power outages Continental power outages Widespread radio communication issues Widespread loss of internet Space hardware damage/failure

Anonymous	Regional power outages Space hardware damage/failure Regional power outages Space hardware damage/failure

Evangelos Paouris George Mason University	Regional power outages Widespread radio communication issues Widespread loss of internet Space hardware damage/failure Significant health risk to airline crew/passengers Regional power outages Widespread radio communication issues Widespread loss of internet Space hardware damage/failure Significant health risk to airline crew/passengers
With the term “Carrington type” event we assume an extremely strong solar flare, a very fast CME moving towards Earth, and very energetic particles. The results will be a radio blackout, an extreme solar storm, and an extreme geomagnetic storm.
Karim Meziane University of New Brunswick	Widespread radio communication issues Widespread radio communication issues

Anonymous	Regional power outages Continental power outages Widespread radio communication issues Widespread loss of internet Space hardware damage/failure Significant health risk to airline crew/passengers Regional power outages Continental power outages Widespread radio communication issues Widespread loss of internet Space hardware damage/failure Significant health risk to airline crew/passengers

Anonymous	Regional power outages Widespread radio communication issues Space hardware damage/failure Significant health risk to airline crew/passengers Regional power outages Widespread radio communication issues Space hardware damage/failure Significant health risk to airline crew/passengers

Vincenzo Romano Istituto Nazionale di Geofisica e Vulcanologia	Regional power outages Continental power outages Widespread radio communication issues Widespread loss of internet Space hardware damage/failure Regional power outages Continental power outages Widespread radio communication issues Widespread loss of internet Space hardware damage/failure

Delores Knipp University of Colorado Boulder	Regional power outages Widespread radio communication issues Space hardware damage/failure Regional power outages Widespread radio communication issues Space hardware damage/failure
There is no correct answer for Q3 for me. I study coupled geospace. For Q5 I think loss of internet is a possibility, but tied to regional power outages.
Anonymous	Continental power outages Widespread radio communication issues Widespread loss of internet Space hardware damage/failure Significant health risk to airline crew/passengers Continental power outages Widespread radio communication issues Widespread loss of internet Space hardware damage/failure Significant health risk to airline crew/passengers

Anonymous	Regional power outages Widespread radio communication issues Space hardware damage/failure Regional power outages Widespread radio communication issues Space hardware damage/failure

Anonymous	Continental power outages Widespread radio communication issues Widespread loss of internet Space hardware damage/failure Continental power outages Widespread radio communication issues Widespread loss of internet Space hardware damage/failure

Ankush Bhaskar Space Physics Laboratory, VSSC, India	Regional power outages Continental power outages Widespread radio communication issues Widespread loss of internet Space hardware damage/failure Significant health risk to airline crew/passengers Regional power outages Continental power outages Widespread radio communication issues Widespread loss of internet Space hardware damage/failure Significant health risk to airline crew/passengers

Anonymous	Regional power outages Continental power outages Widespread radio communication issues Widespread loss of internet Space hardware damage/failure Significant health risk to airline crew/passengers Regional power outages Continental power outages Widespread radio communication issues Widespread loss of internet Space hardware damage/failure Significant health risk to airline crew/passengers

Anonymous	Regional power outages Widespread radio communication issues Space hardware damage/failure Significant health risk to airline crew/passengers Regional power outages Widespread radio communication issues Space hardware damage/failure Significant health risk to airline crew/passengers

Anonymous	Regional power outages Continental power outages Widespread radio communication issues Widespread loss of internet Space hardware damage/failure Regional power outages Continental power outages Widespread radio communication issues Widespread loss of internet Space hardware damage/failure

Anonymous	Continental power outages Widespread radio communication issues Widespread loss of internet Space hardware damage/failure Significant health risk to airline crew/passengers Continental power outages Widespread radio communication issues Widespread loss of internet Space hardware damage/failure Significant health risk to airline crew/passengers

Anonymous	Regional power outages Widespread radio communication issues Widespread loss of internet Regional power outages Widespread radio communication issues Widespread loss of internet

Anonymous	Regional power outages Widespread loss of internet Space hardware damage/failure Significant health risk to airline crew/passengers Regional power outages Widespread loss of internet Space hardware damage/failure Significant health risk to airline crew/passengers
I think the above selection doesn't capture the interdependency of societal infrastructure. The internet isn't necessarily vulnerable (depending how you define it) but it required powering at various levels. If the grid fails, the internet fails.
Anonymous	Regional power outages Continental power outages Widespread radio communication issues Space hardware damage/failure Significant health risk to airline crew/passengers Regional power outages Continental power outages Widespread radio communication issues Space hardware damage/failure Significant health risk to airline crew/passengers

Mirko Piersanti University of L'Aquila, Italy	Regional power outages Widespread radio communication issues Widespread loss of internet Space hardware damage/failure Regional power outages Widespread radio communication issues Widespread loss of internet Space hardware damage/failure

Anonymous	Widespread radio communication issues Space hardware damage/failure Widespread radio communication issues Space hardware damage/failure
Significant risk to astronauts on the ISS / elsewhere (e.g. Artemis 2)
Anonymous	Regional power outages Widespread radio communication issues Space hardware damage/failure Regional power outages Widespread radio communication issues Space hardware damage/failure

Anonymous	Regional power outages Space hardware damage/failure Regional power outages Space hardware damage/failure

Anonymous	Space hardware damage/failure Space hardware damage/failure

John Haiducek US Naval Research Laboratory	Regional power outages Widespread radio communication issues Space hardware damage/failure Significant health risk to airline crew/passengers Regional power outages Widespread radio communication issues Space hardware damage/failure Significant health risk to airline crew/passengers
Hard to answer this with confidence given the many uncertainties involved, but I'd imagine that industry's increased awareness of potential space weather impacts would result in mitigation of some of the most severe possibilities (such as continental power outages), but the impacts would still be considerable.
Daniel Welling University of Michigan	Regional power outages Continental power outages Widespread radio communication issues Regional power outages Continental power outages Widespread radio communication issues

Anonymous	Continental power outages Widespread radio communication issues Widespread loss of internet Space hardware damage/failure Significant health risk to airline crew/passengers Continental power outages Widespread radio communication issues Widespread loss of internet Space hardware damage/failure Significant health risk to airline crew/passengers

Anonymous	Regional power outages Widespread radio communication issues Widespread loss of internet Space hardware damage/failure Significant health risk to airline crew/passengers Regional power outages Widespread radio communication issues Widespread loss of internet Space hardware damage/failure Significant health risk to airline crew/passengers

Anonymous	Regional power outages Widespread radio communication issues Space hardware damage/failure Regional power outages Widespread radio communication issues Space hardware damage/failure

Simon Machin Met Office	Regional power outages Widespread radio communication issues Space hardware damage/failure Regional power outages Widespread radio communication issues Space hardware damage/failure
The question of loss of internet would be closely tied to the scale and location/distribution of impact to power outages. There would be losses, but how widespread would vary depending on a number of variable factors (e.g. geography, geology, grid/transformer characteristics, effective mitigation procedures etc).
Anonymous	Continental power outages Widespread radio communication issues Widespread loss of internet Space hardware damage/failure Significant health risk to airline crew/passengers Continental power outages Widespread radio communication issues Widespread loss of internet Space hardware damage/failure Significant health risk to airline crew/passengers

Anonymous	Regional power outages Widespread radio communication issues Widespread loss of internet Space hardware damage/failure Regional power outages Widespread radio communication issues Widespread loss of internet Space hardware damage/failure

Phillip Chamberlin University of Colorado	Widespread radio communication issues Space hardware damage/failure Widespread radio communication issues Space hardware damage/failure

Anonymous	Continental power outages Widespread radio communication issues Space hardware damage/failure Continental power outages Widespread radio communication issues Space hardware damage/failure

Jingnan Guo USTC	Regional power outages Widespread radio communication issues Space hardware damage/failure Significant health risk to airline crew/passengers Regional power outages Widespread radio communication issues Space hardware damage/failure Significant health risk to airline crew/passengers

Sean Bruinsma CNES	Regional power outages Widespread radio communication issues Space hardware damage/failure Regional power outages Widespread radio communication issues Space hardware damage/failure
Much exageration around this topic in order to get funding
Anonymous	Regional power outages Space hardware damage/failure Regional power outages Space hardware damage/failure

Xiukuan Zhao Institute of Geology and Geophysics, Chinese Academy of Sciences	Regional power outages Widespread radio communication issues Space hardware damage/failure Significant health risk to airline crew/passengers Regional power outages Widespread radio communication issues Space hardware damage/failure Significant health risk to airline crew/passengers

Anonymous	Regional power outages Widespread radio communication issues Space hardware damage/failure Significant health risk to airline crew/passengers Regional power outages Widespread radio communication issues Space hardware damage/failure Significant health risk to airline crew/passengers

Xinan Yue Institute of Geology and Geophysics, Chinese Academy of Sciences	Regional power outages Widespread radio communication issues Regional power outages Widespread radio communication issues

Anonymous	Regional power outages Widespread radio communication issues Space hardware damage/failure Regional power outages Widespread radio communication issues Space hardware damage/failure

Anonymous	Regional power outages Continental power outages Widespread radio communication issues Widespread loss of internet Space hardware damage/failure Significant health risk to airline crew/passengers Regional power outages Continental power outages Widespread radio communication issues Widespread loss of internet Space hardware damage/failure Significant health risk to airline crew/passengers

Dean Pesnell NASA GSFC	Regional power outages Space hardware damage/failure Regional power outages Space hardware damage/failure
Is the Internet more susceptible to power outages or comm outages? WiFi connections use both and might be the dominant connection method for consumer Internet.
Anonymous	Regional power outages Widespread radio communication issues Widespread loss of internet Space hardware damage/failure Significant health risk to airline crew/passengers Regional power outages Widespread radio communication issues Widespread loss of internet Space hardware damage/failure Significant health risk to airline crew/passengers

Anonymous	Widespread radio communication issues Widespread loss of internet Space hardware damage/failure Widespread radio communication issues Widespread loss of internet Space hardware damage/failure

Anonymous	Regional power outages Widespread radio communication issues Widespread loss of internet Space hardware damage/failure Significant health risk to airline crew/passengers Regional power outages Widespread radio communication issues Widespread loss of internet Space hardware damage/failure Significant health risk to airline crew/passengers
"without any warning" is not likely. We have fairly good situational awareness overall. Hence, it's not clear what will be learned by this question.
Anonymous	Regional power outages Widespread radio communication issues Space hardware damage/failure Regional power outages Widespread radio communication issues Space hardware damage/failure

Anonymous	Regional power outages Widespread radio communication issues Space hardware damage/failure Significant health risk to airline crew/passengers Regional power outages Widespread radio communication issues Space hardware damage/failure Significant health risk to airline crew/passengers

Anonymous	Regional power outages Widespread radio communication issues Space hardware damage/failure Regional power outages Widespread radio communication issues Space hardware damage/failure
A lot of parameters can be used to describe a solar/geomagnetic storm. Maybe, the use of the term "Carrington event" is too simple...
Alexander Mishev University of Oulu	Widespread radio communication issues Widespread loss of internet Widespread radio communication issues Widespread loss of internet

Leon Golub Smithsonian Astrophysical Observatory	Regional power outages Widespread radio communication issues Space hardware damage/failure Significant health risk to airline crew/passengers Regional power outages Widespread radio communication issues Space hardware damage/failure Significant health risk to airline crew/passengers

Anonymous	Continental power outages Widespread radio communication issues Widespread loss of internet Space hardware damage/failure Significant health risk to airline crew/passengers Continental power outages Widespread radio communication issues Widespread loss of internet Space hardware damage/failure Significant health risk to airline crew/passengers

Noe Lugaz University of New Hampshire	Regional power outages Widespread radio communication issues Space hardware damage/failure Regional power outages Widespread radio communication issues Space hardware damage/failure

Anonymous	Regional power outages Continental power outages Widespread radio communication issues Widespread loss of internet Space hardware damage/failure Significant health risk to airline crew/passengers Regional power outages Continental power outages Widespread radio communication issues Widespread loss of internet Space hardware damage/failure Significant health risk to airline crew/passengers

Anonymous	Regional power outages Widespread radio communication issues Widespread loss of internet Space hardware damage/failure Significant health risk to airline crew/passengers Regional power outages Widespread radio communication issues Widespread loss of internet Space hardware damage/failure Significant health risk to airline crew/passengers

Anonymous	Regional power outages Widespread radio communication issues Space hardware damage/failure Regional power outages Widespread radio communication issues Space hardware damage/failure

Eelco Doornbos KNMI	Regional power outages Space hardware damage/failure Significant health risk to airline crew/passengers Regional power outages Space hardware damage/failure Significant health risk to airline crew/passengers
For airline crew/passengers, the answer to this question relies on the definitions of "risk" and "significant". Effects would perhaps be equivalent to radiation exposure from those from certain types of medical scans. I've checked this box because I think it would still be worthwhile to work to prevent it.
Anonymous	Regional power outages Continental power outages Widespread radio communication issues Widespread loss of internet Space hardware damage/failure Significant health risk to airline crew/passengers Regional power outages Continental power outages Widespread radio communication issues Widespread loss of internet Space hardware damage/failure Significant health risk to airline crew/passengers
Carrington evant would be a catastrophic disaster in the modern society. We need to be prepared.
Michael Liemohn University of Michigan	Regional power outages Space hardware damage/failure Regional power outages Space hardware damage/failure

Marianna Korsos University of Catania	Continental power outages Widespread radio communication issues Widespread loss of internet Space hardware damage/failure Significant health risk to airline crew/passengers Continental power outages Widespread radio communication issues Widespread loss of internet Space hardware damage/failure Significant health risk to airline crew/passengers
the U.K. government(opens in new tab) lists adverse space weather as one of the most serious natural hazards in its National Risk Register, and companies have contingency plans to deal with severe events — as long as they have sufficient warning of them. Lloyd's of London and the Atmospheric and Environmental Research agency in the U.S. have estimated that a Carrington-class event today would result in between $0.6 and $2.6 trillion in damages to the U.S. alone, according to NASA spaceflight.
Anonymous	Regional power outages Widespread radio communication issues Widespread loss of internet Space hardware damage/failure Significant health risk to airline crew/passengers Regional power outages Widespread radio communication issues Widespread loss of internet Space hardware damage/failure Significant health risk to airline crew/passengers

Jean Uwamahoro University of Rwanda	Regional power outages Regional power outages

Anonymous	Regional power outages Widespread radio communication issues Space hardware damage/failure Regional power outages Widespread radio communication issues Space hardware damage/failure

Anonymous	Widespread radio communication issues Widespread radio communication issues
The effect on the neutral density - with consequences on satellite should be important
Anonymous	Regional power outages Widespread radio communication issues Space hardware damage/failure Regional power outages Widespread radio communication issues Space hardware damage/failure

Matthew Lang British Antarctic Survey	Regional power outages Widespread radio communication issues Widespread loss of internet Space hardware damage/failure Regional power outages Widespread radio communication issues Widespread loss of internet Space hardware damage/failure

Anonymous	Regional power outages Widespread radio communication issues Space hardware damage/failure Regional power outages Widespread radio communication issues Space hardware damage/failure
I've ticked regional power outages as there is a chance for that but I think general widespread disruption and local outages is more likely.
Anonymous	Regional power outages Widespread radio communication issues Space hardware damage/failure Regional power outages Widespread radio communication issues Space hardware damage/failure

Anonymous	Regional power outages Widespread radio communication issues Widespread loss of internet Space hardware damage/failure Significant health risk to airline crew/passengers Regional power outages Widespread radio communication issues Widespread loss of internet Space hardware damage/failure Significant health risk to airline crew/passengers
Different from the ground severe weather events, the impact of severe space weather events hasn't yet been well known by the general public. More research and promotions are needed.
Anonymous	Regional power outages Widespread radio communication issues Widespread loss of internet Space hardware damage/failure Regional power outages Widespread radio communication issues Widespread loss of internet Space hardware damage/failure

Jonny Rae Northumbria University	Regional power outages Widespread radio communication issues Widespread loss of internet Space hardware damage/failure Significant health risk to airline crew/passengers Regional power outages Widespread radio communication issues Widespread loss of internet Space hardware damage/failure Significant health risk to airline crew/passengers
I am honestly not sure what the impact would be. It is very difficult to get sufficient information out of companies to understand how badly they will be affected (compared to other impacts)
James Adams University of Alabama in Huntsville	Regional power outages Widespread radio communication issues Space hardware damage/failure Regional power outages Widespread radio communication issues Space hardware damage/failure
The impact on the power grid would be only regional because of the protections that have been instituted. It is reasonable to suppose that the ionospheric disturbances would lead to widespread communication issues for a range of wavelengths. It seems likely that there would be a large SPE causing spacecraft anomalies, damage and failures.
Anonymous	Continental power outages Widespread radio communication issues Space hardware damage/failure Continental power outages Widespread radio communication issues Space hardware damage/failure

Anonymous	Regional power outages Widespread radio communication issues Widespread loss of internet Space hardware damage/failure Significant health risk to airline crew/passengers Regional power outages Widespread radio communication issues Widespread loss of internet Space hardware damage/failure Significant health risk to airline crew/passengers

Lucilla Alfonsi Istituto Nazionale di Geofisica e Vulcanologia (INGV)	Widespread radio communication issues Space hardware damage/failure Significant health risk to airline crew/passengers Widespread radio communication issues Space hardware damage/failure Significant health risk to airline crew/passengers

Anonymous	Regional power outages Widespread radio communication issues Space hardware damage/failure Regional power outages Widespread radio communication issues Space hardware damage/failure

Anonymous	Regional power outages Continental power outages Widespread radio communication issues Widespread loss of internet Space hardware damage/failure Significant health risk to airline crew/passengers Regional power outages Continental power outages Widespread radio communication issues Widespread loss of internet Space hardware damage/failure Significant health risk to airline crew/passengers
It's not anymore wether but when a major SWx strom will hit us.
Ed Thiemann LASP, University of Colorado	Regional power outages Continental power outages Widespread radio communication issues Widespread loss of internet Space hardware damage/failure Significant health risk to airline crew/passengers Regional power outages Continental power outages Widespread radio communication issues Widespread loss of internet Space hardware damage/failure Significant health risk to airline crew/passengers

Anthony Mannucci Jet Propulsion Laboratory, California Institute of Technology	Regional power outages Continental power outages Widespread radio communication issues Space hardware damage/failure Significant health risk to airline crew/passengers Regional power outages Continental power outages Widespread radio communication issues Space hardware damage/failure Significant health risk to airline crew/passengers

Anonymous	Regional power outages Continental power outages Widespread radio communication issues Widespread loss of internet Space hardware damage/failure Regional power outages Continental power outages Widespread radio communication issues Widespread loss of internet Space hardware damage/failure
Most of those effects are mitigated by warning, of course.
Mike Hapgood RAL Space	Regional power outages Widespread radio communication issues Space hardware damage/failure Significant health risk to airline crew/passengers Regional power outages Widespread radio communication issues Space hardware damage/failure Significant health risk to airline crew/passengers
Risk to airline crew/passengers includes disruption of aircraft control by single event effects in avionics - immediately dangerous whereas radiation dose is long-term health issue. Also disruption of control systems for road transport is a key risk (including to life) as noted by DfT in UK.
Anonymous	Regional power outages Widespread radio communication issues Widespread loss of internet Space hardware damage/failure Significant health risk to airline crew/passengers Regional power outages Widespread radio communication issues Widespread loss of internet Space hardware damage/failure Significant health risk to airline crew/passengers

Anonymous	Regional power outages Widespread radio communication issues Widespread loss of internet Space hardware damage/failure Regional power outages Widespread radio communication issues Widespread loss of internet Space hardware damage/failure

Anonymous	Regional power outages Widespread radio communication issues Widespread loss of internet Space hardware damage/failure Regional power outages Widespread radio communication issues Widespread loss of internet Space hardware damage/failure

Sean Elvidge University of Birmingham	Regional power outages Widespread radio communication issues Space hardware damage/failure Regional power outages Widespread radio communication issues Space hardware damage/failure

Anonymous	Regional power outages Continental power outages Widespread radio communication issues Widespread loss of internet Space hardware damage/failure Significant health risk to airline crew/passengers Regional power outages Continental power outages Widespread radio communication issues Widespread loss of internet Space hardware damage/failure Significant health risk to airline crew/passengers

Martin Mlynczak NASA Langley Research Center	Continental power outages Widespread loss of internet Space hardware damage/failure Continental power outages Widespread loss of internet Space hardware damage/failure
The real question is what would the effect be WITH warning. What could be done to mitigate an event if we knew it was coming? What confidence do we have in our decision processes?
Allison Jaynes University of Iowa	Regional power outages Widespread radio communication issues Space hardware damage/failure Regional power outages Widespread radio communication issues Space hardware damage/failure

Anonymous	Continental power outages Widespread radio communication issues Widespread loss of internet Space hardware damage/failure Significant health risk to airline crew/passengers Continental power outages Widespread radio communication issues Widespread loss of internet Space hardware damage/failure Significant health risk to airline crew/passengers

Bernard V Jackson University of California, San Diego	Regional power outages Continental power outages Widespread radio communication issues Widespread loss of internet Space hardware damage/failure Significant health risk to airline crew/passengers Regional power outages Continental power outages Widespread radio communication issues Widespread loss of internet Space hardware damage/failure Significant health risk to airline crew/passengers
I forecast these things, and recently forecast April 23, 2023 geomagnetic storm and that it would quite dramatic. It was fun to watch the forecast come true.
Anonymous	Regional power outages Continental power outages Widespread radio communication issues Widespread loss of internet Space hardware damage/failure Significant health risk to airline crew/passengers Regional power outages Continental power outages Widespread radio communication issues Widespread loss of internet Space hardware damage/failure Significant health risk to airline crew/passengers

Yaqi Jin University of Oslo	Regional power outages Space hardware damage/failure Regional power outages Space hardware damage/failure
Nowadays people are more prepared for the possible space weather impact than in 1859, even though still not enough, so widespread power outages are less likely to happen. Regional impact can be experienced.
Anonymous	Continental power outages Widespread radio communication issues Space hardware damage/failure Continental power outages Widespread radio communication issues Space hardware damage/failure

Anonymous	Regional power outages Widespread radio communication issues Widespread loss of internet Space hardware damage/failure Regional power outages Widespread radio communication issues Widespread loss of internet Space hardware damage/failure

Anonymous	Regional power outages Widespread radio communication issues Widespread loss of internet Space hardware damage/failure Regional power outages Widespread radio communication issues Widespread loss of internet Space hardware damage/failure
More could be done with information on major space age events for which solar and upstream (L1) information is sufficient to do serious 'postcast' modeling. The results would allow applications experts to evaluate the ranges of consequences (e.g. for the July 2012 STEREO event)
Dario Del Moro University of Rome Tor Vergata	Regional power outages Widespread radio communication issues Space hardware damage/failure Regional power outages Widespread radio communication issues Space hardware damage/failure

Huw Morgan Aberystwyth University	Regional power outages Widespread radio communication issues Widespread loss of internet Space hardware damage/failure Significant health risk to airline crew/passengers Regional power outages Widespread radio communication issues Widespread loss of internet Space hardware damage/failure Significant health risk to airline crew/passengers
Considerable uncertainty in degree and range of disruption since we have not experienced such an event. Note internet becomes useless after a short period of power outage!
Mateja Dumbovic Hvar Observatory, University of Zagreb	Regional power outages Widespread radio communication issues Space hardware damage/failure Regional power outages Widespread radio communication issues Space hardware damage/failure
We should also keep in mind that we are making our technology increasingly resistant to protect it against negative space weather effects, it’s not all about the timely warnings
Yuhao Wang Nanchang University	Regional power outages Continental power outages Widespread radio communication issues Widespread loss of internet Space hardware damage/failure Significant health risk to airline crew/passengers Regional power outages Continental power outages Widespread radio communication issues Widespread loss of internet Space hardware damage/failure Significant health risk to airline crew/passengers
If the Carrington incident occurred in modern times, it would be a global disaster, seriously affecting the networks and communication satellites that humans rely on.
Anonymous	Regional power outages Widespread radio communication issues Space hardware damage/failure Significant health risk to airline crew/passengers Regional power outages Widespread radio communication issues Space hardware damage/failure Significant health risk to airline crew/passengers

Rajkumar Hajra University of Science and Technology of China	Continental power outages Widespread radio communication issues Widespread loss of internet Space hardware damage/failure Continental power outages Widespread radio communication issues Widespread loss of internet Space hardware damage/failure

Anonymous	Regional power outages Regional power outages

Jayachandran P.T. University of New Brunswick	Regional power outages Widespread radio communication issues Regional power outages Widespread radio communication issues
We have more resilient technologies now. Therefore, the impact will be minimal.
Mike Lockwood University of Reading	Regional power outages Widespread radio communication issues Widespread loss of internet Space hardware damage/failure Significant health risk to airline crew/passengers Regional power outages Widespread radio communication issues Widespread loss of internet Space hardware damage/failure Significant health risk to airline crew/passengers
On health risks - integrated doses may be more the issue than large, Carrington-scale events.
Anonymous	Continental power outages Widespread radio communication issues Space hardware damage/failure Continental power outages Widespread radio communication issues Space hardware damage/failure

Ciaran Beggan British Geological Survey	Regional power outages Widespread radio communication issues Space hardware damage/failure Regional power outages Widespread radio communication issues Space hardware damage/failure
There are probably lots of secondary effects which we might not recognise until after the event e.g. trains stuck on lines or loss of specific internet services
Anonymous	Regional power outages Space hardware damage/failure Regional power outages Space hardware damage/failure

Anonymous	Continental power outages Widespread radio communication issues Widespread loss of internet Space hardware damage/failure Significant health risk to airline crew/passengers Continental power outages Widespread radio communication issues Widespread loss of internet Space hardware damage/failure Significant health risk to airline crew/passengers

Mark Moldwin University of Michigan	Regional power outages Widespread radio communication issues Space hardware damage/failure Regional power outages Widespread radio communication issues Space hardware damage/failure
There could be cascading failures in communication and other infrastructure.
Daniel Brandt Michigan Tech Research Institute	Regional power outages Widespread radio communication issues Widespread loss of internet Space hardware damage/failure Significant health risk to airline crew/passengers Regional power outages Widespread radio communication issues Widespread loss of internet Space hardware damage/failure Significant health risk to airline crew/passengers

David R. Themens University of Birmingham	Regional power outages Widespread radio communication issues Space hardware damage/failure Regional power outages Widespread radio communication issues Space hardware damage/failure
Heath risk would be measurable but is unlikely to cause permanent long-term health issues. Space hardware damage/failure is a pretty big range; even modest events cause some measure of solar cell degradation, so it's hard not selecting that one regardless of severity.
Anonymous	Regional power outages Continental power outages Widespread radio communication issues Widespread loss of internet Space hardware damage/failure Significant health risk to airline crew/passengers Regional power outages Continental power outages Widespread radio communication issues Widespread loss of internet Space hardware damage/failure Significant health risk to airline crew/passengers

Anonymous	Regional power outages Continental power outages Widespread radio communication issues Widespread loss of internet Space hardware damage/failure Significant health risk to airline crew/passengers Regional power outages Continental power outages Widespread radio communication issues Widespread loss of internet Space hardware damage/failure Significant health risk to airline crew/passengers

Anonymous	Regional power outages Widespread radio communication issues Space hardware damage/failure Regional power outages Widespread radio communication issues Space hardware damage/failure

Anonymous	Continental power outages Widespread radio communication issues Widespread loss of internet Space hardware damage/failure Significant health risk to airline crew/passengers Continental power outages Widespread radio communication issues Widespread loss of internet Space hardware damage/failure Significant health risk to airline crew/passengers

Anonymous	Regional power outages Widespread radio communication issues Space hardware damage/failure Regional power outages Widespread radio communication issues Space hardware damage/failure
The question is ill-posed. What constitutes a significant health risk? An increase in likelihood of cancer - or death? Widespread loss of internet - % in space vs ground? Power grid impacts on internet? etc
Anonymous	Regional power outages Continental power outages Widespread radio communication issues Widespread loss of internet Space hardware damage/failure Significant health risk to airline crew/passengers Regional power outages Continental power outages Widespread radio communication issues Widespread loss of internet Space hardware damage/failure Significant health risk to airline crew/passengers

Anonymous	Regional power outages Continental power outages Widespread radio communication issues Widespread loss of internet Space hardware damage/failure Regional power outages Continental power outages Widespread radio communication issues Widespread loss of internet Space hardware damage/failure
I am assuming that we are speaking of a solar flare > X30 associated with a significant, Earth-directed CME.
Anonymous	Regional power outages Widespread radio communication issues Space hardware damage/failure Significant health risk to airline crew/passengers Regional power outages Widespread radio communication issues Space hardware damage/failure Significant health risk to airline crew/passengers
Such strong events are indeed likely to have significant impacts on our technological society, but we would not be unprepared for the event, and I believe many steps could be taken to successfully mitigate impacts.
Larry C Gardner Utah State University Eastern	Continental power outages Widespread radio communication issues Widespread loss of internet Space hardware damage/failure Continental power outages Widespread radio communication issues Widespread loss of internet Space hardware damage/failure

Piers Jiggens ESA	Regional power outages Widespread radio communication issues Space hardware damage/failure Significant health risk to airline crew/passengers Regional power outages Widespread radio communication issues Space hardware damage/failure Significant health risk to airline crew/passengers
GNSS outages would be important. Space hardware upsets and reboots would be widespread but failures would be concentrated on new space I would guess
Bernd Heber Christian-Albrechts-Universität	Regional power outages Widespread radio communication issues Widespread loss of internet Space hardware damage/failure Regional power outages Widespread radio communication issues Widespread loss of internet Space hardware damage/failure

Jacob Bortnik UCLA	Regional power outages Widespread radio communication issues Space hardware damage/failure Significant health risk to airline crew/passengers Regional power outages Widespread radio communication issues Space hardware damage/failure Significant health risk to airline crew/passengers
It's really (really) hard to tell because we don't have anything like this event to even calibrate.
Anonymous	Continental power outages Widespread radio communication issues Widespread loss of internet Space hardware damage/failure Significant health risk to airline crew/passengers Continental power outages Widespread radio communication issues Widespread loss of internet Space hardware damage/failure Significant health risk to airline crew/passengers
The specific impacts would vary based on the exact geometry of the Earth-Sun system.
Anonymous	Regional power outages Widespread radio communication issues Widespread loss of internet Space hardware damage/failure Regional power outages Widespread radio communication issues Widespread loss of internet Space hardware damage/failure

Anonymous	Continental power outages Continental power outages
You did not have a category of "all of the above". That would be my choice

Question 2

Cosmogenic isotope data – such as 14C and 10Be – suggest that significantly larger solar energetic particle events occurred in the last few thousand years than in the last few centuries, but it is not clear if this also translates to larger geomagnetic storms in the past. Does the nearly 200-year record of geomagnetic observations cover the likely range of impacts to the power distribution grid over the next 50 years?

• Yes, the geomagnetic record covers the likely range of power grid impacts
• No, but likely range is only slightly (10-50%) higher
• No, the likely range is moderately (50 -100%) higher
• No, the likely range is significantly (>100%) higher

Results

Participant	Response	Confidence
Edmund Henley Met Office (views my own)	Likely range is slightly higher
If an option, I’d have chosen “higher, but don’t know how much”. I suspect high end likely undersampled. But I’m not sure where to put it! Going conservative based just on gut, without looking at isotope literature to get sense of scaling there. Or literature extrapolating current window.
Anonymous

Ryan McGranaghan NASA Jet Propulsion Laboratory	Likely range is moderately higher
'Likely' is a difficult term to use hear. I believe it is quite possible that what we have observed in the 200-year record is not representative of the tails of the distribution of possible activity. The challenge with tails of distributions is that they may be unlikely in the way we think of probabilities, but they are indeed possible. I'm uncertain if the question is asking about likelihood or possibility. If the latter, my answer might have been the significantly higher one.
Anonymous	Likely range is slightly higher

Dr Mario M. Bisi UKRI STFC RAL Space	Likely range is moderately higher
Here I wanted to choose a "not sure" option. This is partly because the question is poorly worded confusing geomagnetic storms with SEP events - the two are not necessarily coupled. I think it was ~1976 where the largest SEP event in recent times was recorded, and this was a tiny geomagnetic storm compared with the 1859 Carrington event - and then vice-versa for the SEPs.
Anonymous	Likely range is moderately higher

Ian Mann University of Alberta	Yes
Hard to establish the impacts in the next 50 years in the context of expected variance observed over the previous 200 years. I would suspect some likelihood of potentially larger events on a longer timescale than 50 years.
Anonymous	Likely range is moderately higher

Anonymous	Likely range is moderately higher
Energetic particle levels do not correlate well with other geo-effectiveness parameters. they don't provide reliable bounds for GICs
Anonymous

Anonymous	Likely range is moderately higher
This is an educated guess based on extreme value analysis. It gives some indication that high-latitude regions (with regular high geomagnetic activity) have already nearly experienced the worst case, whereas this is not the case for lower latitudes.
Anonymous	Yes

Anonymous	Likely range is slightly higher

Tamas Gombosi University of Michigan	Likely range is moderately higher

Gabor Toth University of Michigan	Likely range is moderately higher
To the best of my knowledge the distribution is a power law with no apparent break (saturation). But the probability for a fixed time interval becomes small for (very) extreme events. The question above can be answered quantitatively. I only provided a guess.
Harlan Spence University of New Hampshire	Likely range is significantly higher
Our ability to unravel the past's largest geomagnetic storms is seriously limited by access to only imperfect proxies in the geologic record. That leads to big uncertainties in predicting the likely occurrence, size, and hence impacts of events that impact specifically the power grid, but also most other space weather effects.
Anonymous	Likely range is moderately higher
Stellar flare observations also suggest that the drivers of space weather could be significantly stronger than what we've observed.
Robert F. Wimmer-Schweingruber Kiel University, Kiel, Germany	Likely range is moderately higher
We know that there are longer cycles than the 11 (22)-year solar activity cycle and that we are currently trending towards lower amplitudes at activity maximum. What we don't know is when we will move back to higher activity maxima again.
Anonymous	Likely range is slightly higher

Anonymous	Likely range is slightly higher

Evangelos Paouris George Mason University	Likely range is moderately higher
There is always a chance for a "perfect Space Weather storm" with an impact moderately higher than the historical records.
Karim Meziane University of New Brunswick	Yes

Anonymous	Likely range is significantly higher

Anonymous	Likely range is moderately higher

Vincenzo Romano Istituto Nazionale di Geofisica e Vulcanologia	Likely range is moderately higher

Delores Knipp University of Colorado Boulder
This question is not well posed. It ties super GLE/SEP events to geomagnetic storms and then to the changing vulnerability of power grid/distribution
Anonymous	Likely range is significantly higher

Anonymous	Yes

Anonymous	Likely range is slightly higher

Ankush Bhaskar Space Physics Laboratory, VSSC, India	Likely range is slightly higher

Anonymous	Likely range is moderately higher

Anonymous	Likely range is slightly higher

Anonymous	Likely range is moderately higher

Anonymous	Yes

Anonymous	Likely range is slightly higher

Anonymous	Likely range is moderately higher

Anonymous	Likely range is slightly higher

Mirko Piersanti University of L'Aquila, Italy	Yes

Anonymous	Likely range is slightly higher

Anonymous

Anonymous	Likely range is slightly higher

Anonymous	Likely range is moderately higher

John Haiducek US Naval Research Laboratory	Likely range is moderately higher
It's hard to imagine that the last 200 years is totally representative of the potential variations, but it's probably at least partially representative since some of the underlying factors seem to be subject to gradual change.
Daniel Welling University of Michigan	Likely range is slightly higher

Anonymous	Likely range is moderately higher

Anonymous	Yes

Anonymous	Likely range is moderately higher

Simon Machin Met Office	Likely range is moderately higher
There would be an expected physical-bound to CME extent and impact magnitude, which may be less than 2 x measured 200-year recorded range, although I feel that considerable doubt remains and that there is a need for more research on this theme.
Anonymous	Likely range is slightly higher

Anonymous
I would be guessing to select from choices.
Phillip Chamberlin University of Colorado	Yes

Anonymous	Yes
Statistically, I think you could argue for any answer here, depending on what you consider a "likely range" over 50 years.
Jingnan Guo USTC	Likely range is moderately higher

Sean Bruinsma CNES	Yes

Anonymous	Yes
When a GIC causes a power outage it doesn't matter if the geomagnetic storm that triggered the GIC was a Dst = -300 nT or Dst=-1000 nT (never observed)
Xiukuan Zhao Institute of Geology and Geophysics, Chinese Academy of Sciences	Yes

Anonymous
Don't know about #7.
Xinan Yue Institute of Geology and Geophysics, Chinese Academy of Sciences	Likely range is moderately higher

Anonymous	Likely range is moderately higher

Anonymous	Likely range is moderately higher

Dean Pesnell NASA GSFC	Likely range is moderately higher

Anonymous	Yes

Anonymous

Anonymous	Likely range is moderately higher

Anonymous	Likely range is moderately higher
The distribution of flares is a power law that does not cover the full range. Solar-like G dwarfs have been observed to produce much larger events.
Anonymous

Anonymous	Likely range is moderately higher
My answer is clearly "NO" but it is difficult to choose between the last three answers.
Alexander Mishev University of Oulu	Likely range is slightly higher

Leon Golub Smithsonian Astrophysical Observatory	Likely range is significantly higher

Anonymous	Likely range is moderately higher

Noe Lugaz University of New Hampshire	Likely range is moderately higher
I would say in the 30-60% range, so between slightly and moderately.
Anonymous	Likely range is moderately higher

Anonymous

Anonymous	Yes
Extreme storm statistics and extreme value theory. You can extrapolate statistics.
Eelco Doornbos KNMI	Likely range is moderately higher

Anonymous	Likely range is moderately higher
Observation coverage is a key to understanding and preparing for the space weather impacts.
Michael Liemohn University of Michigan	Yes
As you state in the wording of question 7, this shift is at the thousands of years timescale...not 50 year timescale.
Marianna Korsos University of Catania	Likely range is slightly higher

Anonymous	Likely range is slightly higher

Jean Uwamahoro University of Rwanda	Yes
I am not quite sure
Anonymous	Likely range is slightly higher

Anonymous	Likely range is slightly higher

Anonymous	Likely range is significantly higher

Matthew Lang British Antarctic Survey	Likely range is slightly higher

Anonymous	Likely range is moderately higher
We have so little data in the wider context of geomagnetic storms its very hard to know what the largest impacts could be. I doubt we've seen the full range in the last 200 years so it is hard to predict whether the most extreme storms are just scaled up versions of what we have measured or something completely different.
Anonymous	Likely range is moderately higher

Anonymous	Likely range is slightly higher

Anonymous	Likely range is moderately higher

Jonny Rae Northumbria University	Likely range is moderately higher
just guessing here, as I know very little about the isotope data and no information is given on how "significantly larger" the SEP events were....
James Adams University of Alabama in Huntsville	Likely range is slightly higher
My reply to this question is just a guess.
Anonymous	Likely range is moderately higher

Anonymous	Likely range is slightly higher

Lucilla Alfonsi Istituto Nazionale di Geofisica e Vulcanologia (INGV)	Yes

Anonymous	Likely range is significantly higher

Anonymous	Likely range is moderately higher

Ed Thiemann LASP, University of Colorado	Yes

Anthony Mannucci Jet Propulsion Laboratory, California Institute of Technology	Likely range is significantly higher

Anonymous	Yes

Mike Hapgood RAL Space	Likely range is moderately higher
In other environmental sciences, e.g. hydrology, rule of thumb is >=5T years of independent data are needed to estimate the 1-in-T year risk. Current geomagnetic data are too limited, hence my assessment. This is particularly for severe substorms where we have good data for <60 years.
Anonymous	Likely range is significantly higher
Tree ring data based over thousands of years show evidence of solar storms far greater than witnessed in the modern times
Anonymous	Likely range is moderately higher

Anonymous	Likely range is slightly higher

Sean Elvidge University of Birmingham	Yes
200-year record of geomag obs clearly covers the likely next 50-year geomag obs. But observations alone do not tell you about impacts on the power distribution grid (it wasn't around for most of that 200 year period). Plus grid resilience is not constant in time.
Anonymous	Likely range is significantly higher

Martin Mlynczak NASA Langley Research Center	Likely range is slightly higher
It is really impossible to tell from only 200 years of observations. If there are longer cycles in the Sun we will have to wait and see.
Allison Jaynes University of Iowa	Likely range is slightly higher

Anonymous	Likely range is slightly higher

Bernard V Jackson University of California, San Diego
This is not my area of expertise - no opinion.
Anonymous	Likely range is slightly higher

Yaqi Jin University of Oslo	Likely range is moderately higher

Anonymous	Likely range is slightly higher

Anonymous	Likely range is significantly higher
During the same period Earth's dipole moment has changed dramatically, which also changes the impact. In addition, these two hundred years does not cover high latitudes to good enough extent.
Anonymous	Likely range is moderately higher
This is really a pure guess. A big problem is that multiple events often occur during active times that interact in ways we are still trying to unravel. There is also a role of ambient conditions that can 'enable' extremes. These make estimations of worst case scenarios highly uncertain. Extreme events are a 'game of chance' in many respects, so difficult to foresee or simulate given all variables. (Maybe a good target for AI/probablistic approaches?)
Dario Del Moro University of Rome Tor Vergata	Likely range is moderately higher

Huw Morgan Aberystwyth University	Likely range is significantly higher
My answer to 7 is educated opinion rather than based on data. I believe the isotope data points to very large events, even compared to Carrington. Given the risk posed, we must be prepared even if likelihood is very small.
Mateja Dumbovic Hvar Observatory, University of Zagreb
We are possibly missing some extreme events that go beyond the likely range of impacts, but these would be very rare and thus not very likely
Yuhao Wang Nanchang University	Likely range is moderately higher
No.
Anonymous	Likely range is slightly higher

Rajkumar Hajra University of Science and Technology of China	Yes

Anonymous	Likely range is moderately higher

Jayachandran P.T. University of New Brunswick	Yes

Mike Lockwood University of Reading	Likely range is significantly higher
The above answer on the likely range can only be a guess. It is a "known unknown"
Anonymous	Likely range is slightly higher

Ciaran Beggan British Geological Survey	Likely range is moderately higher
It's so hard to know. That's why we do research on it.
Anonymous	Likely range is moderately higher

Anonymous	Likely range is slightly higher

Mark Moldwin University of Michigan	Yes
You don't know what you don't know, but difficult to extrapolate from beyond past experience. Several recent natural disasters showed that large events can appear but would not want to put a more likely number on it.
Daniel Brandt Michigan Tech Research Institute	Likely range is slightly higher

David R. Themens University of Birmingham	Yes
This question is a bit weirdly posed; by the very nature of the extreme value problem, it is not "likely" that a 1 in 200 year event or greater would occur over the next 50 years.
Anonymous	Likely range is slightly higher

Anonymous	Likely range is moderately higher
The 200-year record of geomagnetic observations is limited by both lack of geographic coverage and lack of data collected at high temporal sampling rates. Both factors are known to affect hazard analysis. It's difficult to know what events we're missing given these limitations.
Anonymous	Likely range is moderately higher

Anonymous	Likely range is moderately higher

Anonymous
What is the connection between SEP and power grid impacts? The GIC connection is not clear and a GIC does not dictate that a power grid issue will occur.
Anonymous	Likely range is slightly higher

Anonymous	Likely range is moderately higher

Anonymous	Yes

Larry C Gardner Utah State University Eastern	Likely range is significantly higher

Piers Jiggens ESA
I am unsure
Bernd Heber Christian-Albrechts-Universität	Yes

Jacob Bortnik UCLA	Likely range is slightly higher
while there may well be larger events, we see the general range of solar activity every solar cycle.
Anonymous	Likely range is slightly higher
This question is a bit confusing.
Anonymous	Likely range is significantly higher

Anonymous	Likely range is significantly higher
See Tsurutani and Lakhina GRL 2014

Question 3

Which factors most limit our ability to make accurate space weather forecasts with 1-day lead time? (Rank from most limiting to least limiting)

Results

Participant	Response	Confidence
Edmund Henley Met Office (views my own)	Observational limitations: 1 Incomplete physics knowledge: 4 Inherent unpredictability of the system: 3 Insufficient computational capability: 5 Lack of routine forecast verification and benchmarking: 2
~Implicit in verification: operational space weather’s ~overdependence on academia for underpinning models/tools. Divergent goals: researchers/funders aren’t incentivised for incremental improvements which yielded “quiet revolution” in terrestrial weather modelling (Bauer+ 2015). Nor to coalesce resources on fewer, better community models. Unpredictability: maybe, assuming solar & thermosphere system ends are chaotic.
Anonymous	Observational limitations: Incomplete physics knowledge: Inherent unpredictability of the system: Insufficient computational capability: Lack of routine forecast verification and benchmarking:
lead time for solar wind conditions, especially Bz model skill
Ryan McGranaghan NASA Jet Propulsion Laboratory	Observational limitations: 1 Incomplete physics knowledge: 2 Inherent unpredictability of the system: 3 Insufficient computational capability: 5 Lack of routine forecast verification and benchmarking: 4

Anonymous	Observational limitations: 1 Incomplete physics knowledge: 2 Inherent unpredictability of the system: 3 Insufficient computational capability: 5 Lack of routine forecast verification and benchmarking: 4
Needs better observations of the sun (incl. far side) and understanding of solar active regions / flaring.
Dr Mario M. Bisi UKRI STFC RAL Space	Observational limitations: 1 Incomplete physics knowledge: 2 Inherent unpredictability of the system: 4 Insufficient computational capability: 5 Lack of routine forecast verification and benchmarking: 3
By observational limitations, I also include measurement limitations (the two are quite different!)... We desperately need more observations and measurements of the inner heliosphere (between 0.1 and 1.1 AU) to better-understand the outflow of plasma from the Sun to just beyond Earth's orbit to enable improvements to both empirical and physics-based models. In addition, we need more data assimilation to be included into the models we already routinely use, and this thus stems back to needing more observations and measurements.
Anonymous	Observational limitations: 1 Incomplete physics knowledge: 2 Inherent unpredictability of the system: 5 Insufficient computational capability: 3 Lack of routine forecast verification and benchmarking: 4

Ian Mann University of Alberta	Observational limitations: 2 Incomplete physics knowledge: 1 Inherent unpredictability of the system: 4 Insufficient computational capability: 5 Lack of routine forecast verification and benchmarking: 3

Anonymous	Observational limitations: 1 Incomplete physics knowledge: 5 Inherent unpredictability of the system: 4 Insufficient computational capability: 2 Lack of routine forecast verification and benchmarking: 3

Anonymous	Observational limitations: 1 Incomplete physics knowledge: 3 Inherent unpredictability of the system: 2 Insufficient computational capability: 5 Lack of routine forecast verification and benchmarking: 4
measurements at key points in the inner heliosphere and within geospace are the most important limitations
Anonymous	Observational limitations: 1 Incomplete physics knowledge: 2 Inherent unpredictability of the system: 3 Insufficient computational capability: 4 Lack of routine forecast verification and benchmarking: 5

Anonymous	Observational limitations: 2 Incomplete physics knowledge: 3 Inherent unpredictability of the system: 1 Insufficient computational capability: 4 Lack of routine forecast verification and benchmarking: 5

Anonymous	Observational limitations: Incomplete physics knowledge: Inherent unpredictability of the system: Insufficient computational capability: Lack of routine forecast verification and benchmarking:

Anonymous	Observational limitations: 1 Incomplete physics knowledge: 2 Inherent unpredictability of the system: 3 Insufficient computational capability: 4 Lack of routine forecast verification and benchmarking: 5

Tamas Gombosi University of Michigan	Observational limitations: 1 Incomplete physics knowledge: 2 Inherent unpredictability of the system: 4 Insufficient computational capability: 5 Lack of routine forecast verification and benchmarking: 3

Gabor Toth University of Michigan	Observational limitations: 1 Incomplete physics knowledge: 3 Inherent unpredictability of the system: 2 Insufficient computational capability: 4 Lack of routine forecast verification and benchmarking: 5
Solar wind monitors placed around r=0.5au orbit would provide reliable prediction of solar wind at 1au and geospace impact forecast with a ~1-day lead time for moderately fast CME. For ultra fast CME-s 1-day forecast is not possible as they arrive in ~18 hours after the eruption. But a very useful prediction with ~8-12hr lead time would be still possible.
Harlan Spence University of New Hampshire	Observational limitations: 1 Incomplete physics knowledge: 2 Inherent unpredictability of the system: 5 Insufficient computational capability: 4 Lack of routine forecast verification and benchmarking: 3
If we could have a further upstream (~1 day) monitor of an approaching ICME, then we would do a much better job in predicting than at present. The physics is mostly understood but we lack key observables such as the direction of the IMF Bz component - that's the tall pole.
Anonymous	Observational limitations: 1 Incomplete physics knowledge: 2 Inherent unpredictability of the system: 4 Insufficient computational capability: 3 Lack of routine forecast verification and benchmarking: 5
All of these are important but perhaps the biggest challenge right now is the lack of sustained financial support to enable any/all of their development.
Robert F. Wimmer-Schweingruber Kiel University, Kiel, Germany	Observational limitations: 2 Incomplete physics knowledge: 3 Inherent unpredictability of the system: 4 Insufficient computational capability: 5 Lack of routine forecast verification and benchmarking: 1

Anonymous	Observational limitations: Incomplete physics knowledge: Inherent unpredictability of the system: Insufficient computational capability: Lack of routine forecast verification and benchmarking:

Anonymous	Observational limitations: 1 Incomplete physics knowledge: 2 Inherent unpredictability of the system: 3 Insufficient computational capability: 4 Lack of routine forecast verification and benchmarking: 5

Evangelos Paouris George Mason University	Observational limitations: Incomplete physics knowledge: Inherent unpredictability of the system: Insufficient computational capability: Lack of routine forecast verification and benchmarking:
We need continuous observations of the Sun from different viewpoints
Karim Meziane University of New Brunswick	Observational limitations: 3 Incomplete physics knowledge: 2 Inherent unpredictability of the system: 1 Insufficient computational capability: 5 Lack of routine forecast verification and benchmarking: 4

Anonymous	Observational limitations: 1 Incomplete physics knowledge: 4 Inherent unpredictability of the system: 5 Insufficient computational capability: 2 Lack of routine forecast verification and benchmarking: 3

Anonymous	Observational limitations: 2 Incomplete physics knowledge: 1 Inherent unpredictability of the system: 3 Insufficient computational capability: 4 Lack of routine forecast verification and benchmarking: 5

Vincenzo Romano Istituto Nazionale di Geofisica e Vulcanologia	Observational limitations: 2 Incomplete physics knowledge: 1 Inherent unpredictability of the system: 4 Insufficient computational capability: 5 Lack of routine forecast verification and benchmarking: 3

Delores Knipp University of Colorado Boulder	Observational limitations: Incomplete physics knowledge: Inherent unpredictability of the system: Insufficient computational capability: Lack of routine forecast verification and benchmarking:
I'm leaving the ordering as is, and adding an additional limiting element: The inability to treat the Sun-Geospace system as a system and in an ensembled-manner
Anonymous	Observational limitations: 1 Incomplete physics knowledge: 2 Inherent unpredictability of the system: 3 Insufficient computational capability: 5 Lack of routine forecast verification and benchmarking: 4

Anonymous	Observational limitations: 3 Incomplete physics knowledge: 2 Inherent unpredictability of the system: 5 Insufficient computational capability: 4 Lack of routine forecast verification and benchmarking: 1

Anonymous	Observational limitations: 5 Incomplete physics knowledge: 1 Inherent unpredictability of the system: 4 Insufficient computational capability: 3 Lack of routine forecast verification and benchmarking: 2
These answers are for my own, not what I know is available at other institutes or in online database.
Ankush Bhaskar Space Physics Laboratory, VSSC, India	Observational limitations: 1 Incomplete physics knowledge: 2 Inherent unpredictability of the system: 4 Insufficient computational capability: 3 Lack of routine forecast verification and benchmarking: 5

Anonymous	Observational limitations: 1 Incomplete physics knowledge: 5 Inherent unpredictability of the system: 3 Insufficient computational capability: 2 Lack of routine forecast verification and benchmarking: 4

Anonymous	Observational limitations: 3 Incomplete physics knowledge: 1 Inherent unpredictability of the system: 2 Insufficient computational capability: 4 Lack of routine forecast verification and benchmarking: 5

Anonymous	Observational limitations: 1 Incomplete physics knowledge: 2 Inherent unpredictability of the system: 3 Insufficient computational capability: 4 Lack of routine forecast verification and benchmarking: 5

Anonymous	Observational limitations: 1 Incomplete physics knowledge: 2 Inherent unpredictability of the system: 3 Insufficient computational capability: 5 Lack of routine forecast verification and benchmarking: 4

Anonymous	Observational limitations: 2 Incomplete physics knowledge: 1 Inherent unpredictability of the system: 3 Insufficient computational capability: 5 Lack of routine forecast verification and benchmarking: 4

Anonymous	Observational limitations: 3 Incomplete physics knowledge: 1 Inherent unpredictability of the system: 2 Insufficient computational capability: 5 Lack of routine forecast verification and benchmarking: 4
My answers are predicated on the top factor - we don't fully understand the physics. The system may well be inherently unpredictable, but I don't think we know enough yet to say that for sure.
Anonymous	Observational limitations: 2 Incomplete physics knowledge: 1 Inherent unpredictability of the system: 3 Insufficient computational capability: 5 Lack of routine forecast verification and benchmarking: 4

Mirko Piersanti University of L'Aquila, Italy	Observational limitations: 5 Incomplete physics knowledge: 1 Inherent unpredictability of the system: 3 Insufficient computational capability: 2 Lack of routine forecast verification and benchmarking: 4

Anonymous	Observational limitations: 1 Incomplete physics knowledge: 3 Inherent unpredictability of the system: 2 Insufficient computational capability: 5 Lack of routine forecast verification and benchmarking: 4

Anonymous	Observational limitations: 2 Incomplete physics knowledge: 4 Inherent unpredictability of the system: 1 Insufficient computational capability: 3 Lack of routine forecast verification and benchmarking: 5

Anonymous	Observational limitations: 3 Incomplete physics knowledge: 2 Inherent unpredictability of the system: 1 Insufficient computational capability: 4 Lack of routine forecast verification and benchmarking: 5

Anonymous	Observational limitations: 1 Incomplete physics knowledge: 3 Inherent unpredictability of the system: 2 Insufficient computational capability: 4 Lack of routine forecast verification and benchmarking: 5

John Haiducek US Naval Research Laboratory	Observational limitations: 1 Incomplete physics knowledge: 3 Inherent unpredictability of the system: 2 Insufficient computational capability: 4 Lack of routine forecast verification and benchmarking: 5
The inability to obtain solar wind observations more than a few hours in advance is the limiting factor for lead time of many space weather forecasts. Heliospheric models could fill this gap, but are poorly constrained due to limitations of solar and heliospheric observations.
Daniel Welling University of Michigan	Observational limitations: 2 Incomplete physics knowledge: 4 Inherent unpredictability of the system: 5 Insufficient computational capability: 3 Lack of routine forecast verification and benchmarking: 1

Anonymous	Observational limitations: 1 Incomplete physics knowledge: 4 Inherent unpredictability of the system: 2 Insufficient computational capability: 5 Lack of routine forecast verification and benchmarking: 3

Anonymous	Observational limitations: 1 Incomplete physics knowledge: 3 Inherent unpredictability of the system: 2 Insufficient computational capability: 5 Lack of routine forecast verification and benchmarking: 4

Anonymous	Observational limitations: 3 Incomplete physics knowledge: 2 Inherent unpredictability of the system: 1 Insufficient computational capability: 4 Lack of routine forecast verification and benchmarking: 5

Simon Machin Met Office	Observational limitations: 2 Incomplete physics knowledge: 1 Inherent unpredictability of the system: 3 Insufficient computational capability: 5 Lack of routine forecast verification and benchmarking: 4

Anonymous	Observational limitations: 3 Incomplete physics knowledge: 1 Inherent unpredictability of the system: 2 Insufficient computational capability: 4 Lack of routine forecast verification and benchmarking: 5

Anonymous	Observational limitations: 1 Incomplete physics knowledge: 2 Inherent unpredictability of the system: 5 Insufficient computational capability: 4 Lack of routine forecast verification and benchmarking: 3

Phillip Chamberlin University of Colorado	Observational limitations: 3 Incomplete physics knowledge: 2 Inherent unpredictability of the system: 1 Insufficient computational capability: 5 Lack of routine forecast verification and benchmarking: 4

Anonymous	Observational limitations: 1 Incomplete physics knowledge: 3 Inherent unpredictability of the system: 2 Insufficient computational capability: 5 Lack of routine forecast verification and benchmarking: 4

Jingnan Guo USTC	Observational limitations: 3 Incomplete physics knowledge: 1 Inherent unpredictability of the system: 2 Insufficient computational capability: 4 Lack of routine forecast verification and benchmarking: 5

Sean Bruinsma CNES	Observational limitations: 1 Incomplete physics knowledge: 2 Inherent unpredictability of the system: 4 Insufficient computational capability: 5 Lack of routine forecast verification and benchmarking: 3

Anonymous	Observational limitations: 1 Incomplete physics knowledge: 3 Inherent unpredictability of the system: 5 Insufficient computational capability: 4 Lack of routine forecast verification and benchmarking: 2
Our ability is mostly limited by the fact that the funding agencies have, for the last 2-3 decades, confounded space weather with space physics and have focused on programs that (possibly) improved our understanding of fundamental space plasma physics, mostly by means of physics-based simulations. As a matter of fact, none of these simulations are predictive. Only recently (some) project managers in funding agencies have started realizing that simulations does not mean predictions and that the statistical and machine learning tools will make the biggest impact on our ability of predicting space weather in the near future.
Xiukuan Zhao Institute of Geology and Geophysics, Chinese Academy of Sciences	Observational limitations: 1 Incomplete physics knowledge: 2 Inherent unpredictability of the system: 4 Insufficient computational capability: 5 Lack of routine forecast verification and benchmarking: 3

Anonymous	Observational limitations: 2 Incomplete physics knowledge: 1 Inherent unpredictability of the system: 3 Insufficient computational capability: 4 Lack of routine forecast verification and benchmarking: 5

Xinan Yue Institute of Geology and Geophysics, Chinese Academy of Sciences	Observational limitations: 2 Incomplete physics knowledge: 1 Inherent unpredictability of the system: 3 Insufficient computational capability: 4 Lack of routine forecast verification and benchmarking: 5
Space weather is not paid sufficient attention by the research community.
Anonymous	Observational limitations: 2 Incomplete physics knowledge: 1 Inherent unpredictability of the system: 3 Insufficient computational capability: 5 Lack of routine forecast verification and benchmarking: 4

Anonymous	Observational limitations: 1 Incomplete physics knowledge: 2 Inherent unpredictability of the system: 3 Insufficient computational capability: 4 Lack of routine forecast verification and benchmarking: 5

Dean Pesnell NASA GSFC	Observational limitations: 1 Incomplete physics knowledge: 2 Inherent unpredictability of the system: 4 Insufficient computational capability: 3 Lack of routine forecast verification and benchmarking: 5

Anonymous	Observational limitations: 3 Incomplete physics knowledge: 5 Inherent unpredictability of the system: 1 Insufficient computational capability: 4 Lack of routine forecast verification and benchmarking: 2

Anonymous	Observational limitations: 1 Incomplete physics knowledge: 2 Inherent unpredictability of the system: 3 Insufficient computational capability: 4 Lack of routine forecast verification and benchmarking: 5

Anonymous	Observational limitations: 4 Incomplete physics knowledge: 2 Inherent unpredictability of the system: 1 Insufficient computational capability: 5 Lack of routine forecast verification and benchmarking: 3
There's a lot left out of this list. "insufficient use of present data to develop improved forecast methodology", for example.
Anonymous	Observational limitations: 2 Incomplete physics knowledge: 4 Inherent unpredictability of the system: 1 Insufficient computational capability: 5 Lack of routine forecast verification and benchmarking: 3

Anonymous	Observational limitations: 1 Incomplete physics knowledge: 5 Inherent unpredictability of the system: 2 Insufficient computational capability: 3 Lack of routine forecast verification and benchmarking: 4

Anonymous	Observational limitations: 4 Incomplete physics knowledge: 1 Inherent unpredictability of the system: 2 Insufficient computational capability: 3 Lack of routine forecast verification and benchmarking: 5
Note that sometimes the factors mentioned there are closely related to each other.
Alexander Mishev University of Oulu	Observational limitations: 3 Incomplete physics knowledge: 2 Inherent unpredictability of the system: 4 Insufficient computational capability: 5 Lack of routine forecast verification and benchmarking: 1

Leon Golub Smithsonian Astrophysical Observatory	Observational limitations: 1 Incomplete physics knowledge: 4 Inherent unpredictability of the system: 3 Insufficient computational capability: 2 Lack of routine forecast verification and benchmarking: 5

Anonymous	Observational limitations: 2 Incomplete physics knowledge: 1 Inherent unpredictability of the system: 5 Insufficient computational capability: 4 Lack of routine forecast verification and benchmarking: 3

Noe Lugaz University of New Hampshire	Observational limitations: 1 Incomplete physics knowledge: 3 Inherent unpredictability of the system: 2 Insufficient computational capability: 5 Lack of routine forecast verification and benchmarking: 4
It depends a lot on what. For GICs, it is observational limitations and incomplete benchmarking. For SEPs, it is incomplete knowledge and inherent unpredictability. Getting a 1-day lead time for GICs is significantly easier than getting a 1-day lead time for SEPs.
Anonymous	Observational limitations: 1 Incomplete physics knowledge: 3 Inherent unpredictability of the system: 4 Insufficient computational capability: 5 Lack of routine forecast verification and benchmarking: 2

Anonymous	Observational limitations: 1 Incomplete physics knowledge: 5 Inherent unpredictability of the system: 3 Insufficient computational capability: 2 Lack of routine forecast verification and benchmarking: 4

Anonymous	Observational limitations: 1 Incomplete physics knowledge: 2 Inherent unpredictability of the system: 3 Insufficient computational capability: 4 Lack of routine forecast verification and benchmarking: 5

Eelco Doornbos KNMI	Observational limitations: 2 Incomplete physics knowledge: 4 Inherent unpredictability of the system: 1 Insufficient computational capability: 3 Lack of routine forecast verification and benchmarking: 5
The lack of routine forecast verification and benchmarking is largely due to observational limitations, but I've treated it here independent of that. This resulted in higher weight to observational limitations and lower for the lack of verification and benchmarking.
Anonymous	Observational limitations: 1 Incomplete physics knowledge: 2 Inherent unpredictability of the system: 3 Insufficient computational capability: 4 Lack of routine forecast verification and benchmarking: 5

Michael Liemohn University of Michigan	Observational limitations: 2 Incomplete physics knowledge: 1 Inherent unpredictability of the system: 3 Insufficient computational capability: 4 Lack of routine forecast verification and benchmarking: 5
My preference on this ordering is not strong. We need improvements in all 5 areas you list.
Marianna Korsos University of Catania	Observational limitations: 1 Incomplete physics knowledge: 2 Inherent unpredictability of the system: 3 Insufficient computational capability: 5 Lack of routine forecast verification and benchmarking: 4
no routine magnetic field (LOS and vectors) from the lower solar atmosphere, where the solar eruptions occur. they are the driving force of the most dangerous sw event
Anonymous	Observational limitations: 2 Incomplete physics knowledge: 1 Inherent unpredictability of the system: 3 Insufficient computational capability: 5 Lack of routine forecast verification and benchmarking: 4

Jean Uwamahoro University of Rwanda	Observational limitations: 2 Incomplete physics knowledge: 4 Inherent unpredictability of the system: 1 Insufficient computational capability: 3 Lack of routine forecast verification and benchmarking: 5

Anonymous	Observational limitations: 2 Incomplete physics knowledge: 3 Inherent unpredictability of the system: 1 Insufficient computational capability: 5 Lack of routine forecast verification and benchmarking: 4

Anonymous	Observational limitations: 1 Incomplete physics knowledge: 2 Inherent unpredictability of the system: 3 Insufficient computational capability: 4 Lack of routine forecast verification and benchmarking: 5

Anonymous	Observational limitations: Incomplete physics knowledge: Inherent unpredictability of the system: Insufficient computational capability: Lack of routine forecast verification and benchmarking:

Matthew Lang British Antarctic Survey	Observational limitations: 1 Incomplete physics knowledge: 2 Inherent unpredictability of the system: 5 Insufficient computational capability: 3 Lack of routine forecast verification and benchmarking: 4

Anonymous	Observational limitations: 1 Incomplete physics knowledge: 3 Inherent unpredictability of the system: 2 Insufficient computational capability: 5 Lack of routine forecast verification and benchmarking: 4

Anonymous	Observational limitations: Incomplete physics knowledge: Inherent unpredictability of the system: Insufficient computational capability: Lack of routine forecast verification and benchmarking:
List as presented looks about right
Anonymous	Observational limitations: 2 Incomplete physics knowledge: 1 Inherent unpredictability of the system: 3 Insufficient computational capability: 5 Lack of routine forecast verification and benchmarking: 4

Anonymous	Observational limitations: 3 Incomplete physics knowledge: 1 Inherent unpredictability of the system: 2 Insufficient computational capability: 5 Lack of routine forecast verification and benchmarking: 4

Jonny Rae Northumbria University	Observational limitations: 3 Incomplete physics knowledge: 1 Inherent unpredictability of the system: 4 Insufficient computational capability: 5 Lack of routine forecast verification and benchmarking: 2
again, I'm not sure, but this is a great question
James Adams University of Alabama in Huntsville	Observational limitations: 1 Incomplete physics knowledge: 3 Inherent unpredictability of the system: 4 Insufficient computational capability: 2 Lack of routine forecast verification and benchmarking: 5
The fundamental limitation is the lack of observations needed to drive physics-based models. The observational requirements are daunting. Also we lack the computational power to run the existing physical models in real time. It is possible, in principle to create state machines in gate arrays or custom VLSI that would greatly speed up the computations.
Anonymous	Observational limitations: Incomplete physics knowledge: Inherent unpredictability of the system: Insufficient computational capability: Lack of routine forecast verification and benchmarking:

Anonymous	Observational limitations: 3 Incomplete physics knowledge: 2 Inherent unpredictability of the system: 1 Insufficient computational capability: 4 Lack of routine forecast verification and benchmarking: 5

Lucilla Alfonsi Istituto Nazionale di Geofisica e Vulcanologia (INGV)	Observational limitations: 1 Incomplete physics knowledge: 2 Inherent unpredictability of the system: 5 Insufficient computational capability: 4 Lack of routine forecast verification and benchmarking: 3

Anonymous	Observational limitations: 2 Incomplete physics knowledge: 1 Inherent unpredictability of the system: 3 Insufficient computational capability: 5 Lack of routine forecast verification and benchmarking: 4

Anonymous	Observational limitations: 1 Incomplete physics knowledge: 3 Inherent unpredictability of the system: 5 Insufficient computational capability: 4 Lack of routine forecast verification and benchmarking: 2
We do not have yet a proper routine observing system.
Ed Thiemann LASP, University of Colorado	Observational limitations: 1 Incomplete physics knowledge: 2 Inherent unpredictability of the system: 4 Insufficient computational capability: 3 Lack of routine forecast verification and benchmarking: 5

Anthony Mannucci Jet Propulsion Laboratory, California Institute of Technology	Observational limitations: 1 Incomplete physics knowledge: 4 Inherent unpredictability of the system: 3 Insufficient computational capability: 5 Lack of routine forecast verification and benchmarking: 2
It's a combination of insufficient observations and the capability to develop forecast benchmarking from the observations that we need to have.
Anonymous	Observational limitations: 1 Incomplete physics knowledge: 4 Inherent unpredictability of the system: 2 Insufficient computational capability: 3 Lack of routine forecast verification and benchmarking: 5
Major observational limitations exist in the realm of CME arrival and geoeffectiveness forecasting. Wide-field heliospheric imagers, especially with 3D imaging capability, will change that. Specific flare forecasting is a different story and is comparable to forecasting sandpile avalanches, another analogous unsolved (and probably unsolvable) problem in physics.
Mike Hapgood RAL Space	Observational limitations: 1 Incomplete physics knowledge: 3 Inherent unpredictability of the system: 2 Insufficient computational capability: 5 Lack of routine forecast verification and benchmarking: 4
I've ranked observations top as they drive everything else. I've put inherent unpredictability slightly ahead of incomplete physics, to encourage wider engagement with the idea of unpredictability. I've put verification next as it is key to confirming if we have a good understanding of the physics and its predictability.
Anonymous	Observational limitations: 3 Incomplete physics knowledge: 2 Inherent unpredictability of the system: 1 Insufficient computational capability: 4 Lack of routine forecast verification and benchmarking: 5
The inherent nature of the Sun indicates, at least presently, that forecasting SPE events is and will remain impossible. Effort in this area should be directed to nowcast mitigation work rather than chasing forecasting
Anonymous	Observational limitations: 2 Incomplete physics knowledge: 3 Inherent unpredictability of the system: 1 Insufficient computational capability: 4 Lack of routine forecast verification and benchmarking: 5

Anonymous	Observational limitations: 2 Incomplete physics knowledge: 4 Inherent unpredictability of the system: 1 Insufficient computational capability: 3 Lack of routine forecast verification and benchmarking: 5

Sean Elvidge University of Birmingham	Observational limitations: 1 Incomplete physics knowledge: 2 Inherent unpredictability of the system: 5 Insufficient computational capability: 4 Lack of routine forecast verification and benchmarking: 3
"Space weather forecasts" covers a range of topic areas, each could end up with a different listing to above.
Anonymous	Observational limitations: 2 Incomplete physics knowledge: 1 Inherent unpredictability of the system: 3 Insufficient computational capability: 4 Lack of routine forecast verification and benchmarking: 5

Martin Mlynczak NASA Langley Research Center	Observational limitations: 2 Incomplete physics knowledge: 1 Inherent unpredictability of the system: 5 Insufficient computational capability: 4 Lack of routine forecast verification and benchmarking: 3

Allison Jaynes University of Iowa	Observational limitations: 1 Incomplete physics knowledge: 3 Inherent unpredictability of the system: 2 Insufficient computational capability: 5 Lack of routine forecast verification and benchmarking: 4
We don't have teams of "space weather forecasters" working in space weather in the same way they have them in terrestrial weather. We have the science researchers, but not the forecasters as a separate, focused group. This is partially what hinders our progress relative to terrestrial weather forecasting. Also, we do not have distributed arrays of sensors as is done for terrestrial weather, ocean dynamics, etc. I believe that is the #1 biggest roadblock to effective forecasts.
Anonymous	Observational limitations: 3 Incomplete physics knowledge: 1 Inherent unpredictability of the system: 2 Insufficient computational capability: 5 Lack of routine forecast verification and benchmarking: 4

Bernard V Jackson University of California, San Diego	Observational limitations: 1 Incomplete physics knowledge: 2 Inherent unpredictability of the system: 3 Insufficient computational capability: 4 Lack of routine forecast verification and benchmarking: 5
The magnetic field cannot be observed in the heliosphere. For SEPs there is little remote sensing info, and little way to predict their dispersion.
Anonymous	Observational limitations: 1 Incomplete physics knowledge: 2 Inherent unpredictability of the system: 3 Insufficient computational capability: 5 Lack of routine forecast verification and benchmarking: 4

Yaqi Jin University of Oslo	Observational limitations: 1 Incomplete physics knowledge: 3 Inherent unpredictability of the system: 2 Insufficient computational capability: 4 Lack of routine forecast verification and benchmarking: 5
We have accumulated a lot of data in the space era. However, key observations for certain event chain is still difficult to obtain.
Anonymous	Observational limitations: 1 Incomplete physics knowledge: 3 Inherent unpredictability of the system: 5 Insufficient computational capability: 4 Lack of routine forecast verification and benchmarking: 2

Anonymous	Observational limitations: 1 Incomplete physics knowledge: 2 Inherent unpredictability of the system: 4 Insufficient computational capability: 5 Lack of routine forecast verification and benchmarking: 3
We need data from more upstream sensors or measurements in the solar wind. But we also lack the complete understanding of the magnetospheric response to great events.
Anonymous	Observational limitations: 3 Incomplete physics knowledge: 1 Inherent unpredictability of the system: 2 Insufficient computational capability: 4 Lack of routine forecast verification and benchmarking: 5
One has the impression that what we have to work with already could be used far more effectively if we had a more efficient and effective system of supporting research and transitioning to forecasting. e.g. the sponsoring management and bureaucracy could be much more enabling
Dario Del Moro University of Rome Tor Vergata	Observational limitations: 1 Incomplete physics knowledge: 4 Inherent unpredictability of the system: 3 Insufficient computational capability: 5 Lack of routine forecast verification and benchmarking: 2

Huw Morgan Aberystwyth University	Observational limitations: 1 Incomplete physics knowledge: 2 Inherent unpredictability of the system: 3 Insufficient computational capability: 5 Lack of routine forecast verification and benchmarking: 4
Obs. limit and incomplete physics by far most limiting, the others rank far below.
Mateja Dumbovic Hvar Observatory, University of Zagreb	Observational limitations: 1 Incomplete physics knowledge: 3 Inherent unpredictability of the system: 2 Insufficient computational capability: 5 Lack of routine forecast verification and benchmarking: 4
Observational limits and inherent unpredictability of the system will always remain an issue, even if we managed to solve the other 3 factors
Yuhao Wang Nanchang University	Observational limitations: 1 Incomplete physics knowledge: 2 Inherent unpredictability of the system: 3 Insufficient computational capability: 4 Lack of routine forecast verification and benchmarking: 5
No.
Anonymous	Observational limitations: 2 Incomplete physics knowledge: 3 Inherent unpredictability of the system: 1 Insufficient computational capability: 4 Lack of routine forecast verification and benchmarking: 5

Rajkumar Hajra University of Science and Technology of China	Observational limitations: 2 Incomplete physics knowledge: 3 Inherent unpredictability of the system: 1 Insufficient computational capability: 4 Lack of routine forecast verification and benchmarking: 5

Anonymous	Observational limitations: 1 Incomplete physics knowledge: 4 Inherent unpredictability of the system: 3 Insufficient computational capability: 2 Lack of routine forecast verification and benchmarking: 5

Jayachandran P.T. University of New Brunswick	Observational limitations: 2 Incomplete physics knowledge: 1 Inherent unpredictability of the system: 3 Insufficient computational capability: 4 Lack of routine forecast verification and benchmarking: 5

Mike Lockwood University of Reading	Observational limitations: 1 Incomplete physics knowledge: 4 Inherent unpredictability of the system: 3 Insufficient computational capability: 2 Lack of routine forecast verification and benchmarking: 5

Anonymous	Observational limitations: 2 Incomplete physics knowledge: 3 Inherent unpredictability of the system: 1 Insufficient computational capability: 5 Lack of routine forecast verification and benchmarking: 4

Ciaran Beggan British Geological Survey	Observational limitations: 1 Incomplete physics knowledge: 3 Inherent unpredictability of the system: 2 Insufficient computational capability: 4 Lack of routine forecast verification and benchmarking: 5

Anonymous	Observational limitations: 1 Incomplete physics knowledge: 4 Inherent unpredictability of the system: 2 Insufficient computational capability: 5 Lack of routine forecast verification and benchmarking: 3

Anonymous	Observational limitations: 3 Incomplete physics knowledge: 1 Inherent unpredictability of the system: 2 Insufficient computational capability: 5 Lack of routine forecast verification and benchmarking: 4

Mark Moldwin University of Michigan	Observational limitations: 1 Incomplete physics knowledge: 3 Inherent unpredictability of the system: 2 Insufficient computational capability: 4 Lack of routine forecast verification and benchmarking: 5
Depending on what you want to forecast these can bounce around, but we have very limited observations of a coupled complex system.
Daniel Brandt Michigan Tech Research Institute	Observational limitations: 1 Incomplete physics knowledge: 2 Inherent unpredictability of the system: 4 Insufficient computational capability: 5 Lack of routine forecast verification and benchmarking: 3

David R. Themens University of Birmingham	Observational limitations: 2 Incomplete physics knowledge: 1 Inherent unpredictability of the system: 3 Insufficient computational capability: 4 Lack of routine forecast verification and benchmarking: 5

Anonymous	Observational limitations: 3 Incomplete physics knowledge: 1 Inherent unpredictability of the system: 2 Insufficient computational capability: 4 Lack of routine forecast verification and benchmarking: 5

Anonymous	Observational limitations: 1 Incomplete physics knowledge: 3 Inherent unpredictability of the system: 2 Insufficient computational capability: 5 Lack of routine forecast verification and benchmarking: 4
These rankings vary depending on the type of forecast: fluence of relativistic electrons at GEO, geomagnetic disturbance in the auroral zone, geomagnetic disturbance at low latitude, etc
Anonymous	Observational limitations: 3 Incomplete physics knowledge: 2 Inherent unpredictability of the system: 1 Insufficient computational capability: 4 Lack of routine forecast verification and benchmarking: 5

Anonymous	Observational limitations: 1 Incomplete physics knowledge: 3 Inherent unpredictability of the system: 2 Insufficient computational capability: 4 Lack of routine forecast verification and benchmarking: 5

Anonymous	Observational limitations: 2 Incomplete physics knowledge: 3 Inherent unpredictability of the system: 1 Insufficient computational capability: 4 Lack of routine forecast verification and benchmarking: 5
Forecasts of what? Solar wind speed? Radiation belt intensity? SEP? ionospheric disturbances? Orbital drag?
Anonymous	Observational limitations: 1 Incomplete physics knowledge: 3 Inherent unpredictability of the system: 2 Insufficient computational capability: 5 Lack of routine forecast verification and benchmarking: 4

Anonymous	Observational limitations: 2 Incomplete physics knowledge: 1 Inherent unpredictability of the system: 3 Insufficient computational capability: 5 Lack of routine forecast verification and benchmarking: 4
The continued failure of physics-based models is most troubling.
Anonymous	Observational limitations: 2 Incomplete physics knowledge: 1 Inherent unpredictability of the system: 5 Insufficient computational capability: 3 Lack of routine forecast verification and benchmarking: 4

Larry C Gardner Utah State University Eastern	Observational limitations: 4 Incomplete physics knowledge: 3 Inherent unpredictability of the system: 2 Insufficient computational capability: 1 Lack of routine forecast verification and benchmarking: 5
The number one problem is the non-linearity of the system. We need significant development on models to incorporate advanced physics (moment equations).
Piers Jiggens ESA	Observational limitations: 1 Incomplete physics knowledge: 2 Inherent unpredictability of the system: 3 Insufficient computational capability: 4 Lack of routine forecast verification and benchmarking: 5
Depends on the phenomena, assuming CME-driven the issue of resolving the geometry and internal magnetic field is driving. For HSS limitations are probably bounded by assuming what we saw 28 days ago will be the same today so observer in L5 or similar would improve things.
Bernd Heber Christian-Albrechts-Universität	Observational limitations: 1 Incomplete physics knowledge: 5 Inherent unpredictability of the system: 3 Insufficient computational capability: 4 Lack of routine forecast verification and benchmarking: 2

Jacob Bortnik UCLA	Observational limitations: 2 Incomplete physics knowledge: 1 Inherent unpredictability of the system: 4 Insufficient computational capability: 3 Lack of routine forecast verification and benchmarking: 5
I think we're still missing some physics, and then we need large-scale computations.
Anonymous	Observational limitations: 1 Incomplete physics knowledge: 3 Inherent unpredictability of the system: 2 Insufficient computational capability: 4 Lack of routine forecast verification and benchmarking: 5

Anonymous	Observational limitations: 3 Incomplete physics knowledge: 2 Inherent unpredictability of the system: 1 Insufficient computational capability: 5 Lack of routine forecast verification and benchmarking: 4

Anonymous	Observational limitations: 2 Incomplete physics knowledge: 1 Inherent unpredictability of the system: 3 Insufficient computational capability: 4 Lack of routine forecast verification and benchmarking: 5
This question is pretty naiive.

Question 4

What is the biggest challenge to delivering actionable forecasts of an extreme (Carrington-scale) space weather event?

Results

Participant	Response	Confidence
Edmund Henley Met Office (views my own)	Distinguishing damaging events from those with negligible effect
Assuming Carrington events give significant near-Sun observational signatures, so (iff no datagaps degrading coronagraph-based speed estimates) large impacts forecastable on speed-alone basis. But impact modulation by currently operationally-unforecastable Bz is large: distinguishability still dominates? User end of forecast chain hindered by attribution challenges around historic, proprietary impact data.
Anonymous	Detection/observation

Ryan McGranaghan NASA Jet Propulsion Laboratory	Effective dialogue with end users, such as power companies and satellite operators

Anonymous	Distinguishing damaging events from those with negligible effect

Dr Mario M. Bisi UKRI STFC RAL Space	Detection/observation
An extreme, Carrington-like event, is likely to be a fast arriver from Sun to Earth, and therefore we will only have 12-18 hours warning at best, if all the current observatories and instruments are working as they should with the real-time delivery of the data as is essential for space-weather observations and measurements that feed into forecast capabilities.
Anonymous	Distinguishing damaging events from those with negligible effect

Ian Mann University of Alberta	Distinguishing damaging events from those with negligible effect

Anonymous	Detection/observation

Anonymous	Distinguishing damaging events from those with negligible effect
Extreme transients don't translate to extreme geospace responses. we don't have enough information to know either the impact properties at Earth nor the reaction to them
Anonymous	Distinguishing damaging events from those with negligible effect

Anonymous	Effective dialogue with end users, such as power companies and satellite operators
Dialogue with end users includes understanding the severity of effects too.
Anonymous	Distinguishing damaging events from those with negligible effect

Anonymous	Detection/observation

Tamas Gombosi University of Michigan	Effective dialogue with end users, such as power companies and satellite operators

Gabor Toth University of Michigan	Effective dialogue with end users, such as power companies and satellite operators
I believe that end-users are ready to take space weather forecast seriously. Some companies do have an action plan. Others may not.
Harlan Spence University of New Hampshire	Effective dialogue with end users, such as power companies and satellite operators
Though detection/observation is indeed a major problem, even with current capability we make reasonable actionable forecasts. I believe that a big gap exists between what the science community considers useful from our models and what end users need in order to take action. I believe they are willing to act, but only if a forecast meshes with their needs.
Anonymous	Distinguishing damaging events from those with negligible effect
It's not clear to me whether a false positive is worse than a false negative, but clearly delineating between these is vital to have any hope of getting people to take action.
Robert F. Wimmer-Schweingruber Kiel University, Kiel, Germany	Distinguishing damaging events from those with negligible effect

Anonymous	Distinguishing damaging events from those with negligible effect

Anonymous	End users willing and able to act on forecast advice

Evangelos Paouris George Mason University	Detection/observation

Karim Meziane University of New Brunswick	Effective dialogue with end users, such as power companies and satellite operators

Anonymous	Detection/observation

Anonymous	Effective dialogue with end users, such as power companies and satellite operators

Vincenzo Romano Istituto Nazionale di Geofisica e Vulcanologia	End users willing and able to act on forecast advice

Delores Knipp University of Colorado Boulder	Detection/observation
Need another radio button: lack of information from end users about system vulnerabilities due to classification and proprietary nature of data
Anonymous	Distinguishing damaging events from those with negligible effect

Anonymous	End users willing and able to act on forecast advice

Anonymous	Distinguishing damaging events from those with negligible effect
I believe that 2 (distinguish) and 4 are equally big challenges (forecast advice)
Ankush Bhaskar Space Physics Laboratory, VSSC, India	Distinguishing damaging events from those with negligible effect
Carrington-scale events are large and our solar monitors will pick them up in advance. However, predicting interplanetary magnetic field orientation at 1AU is the most challenging which will limit the impact forecast time horizon.
Anonymous	End users willing and able to act on forecast advice

Anonymous	Distinguishing damaging events from those with negligible effect

Anonymous	Effective dialogue with end users, such as power companies and satellite operators

Anonymous	Detection/observation

Anonymous	Detection/observation

Anonymous	Effective dialogue with end users, such as power companies and satellite operators
I think there are other important factors that need to be considered, such as frameworks of regulation and the marketplace that discourage end users from engaging.
Anonymous	Effective dialogue with end users, such as power companies and satellite operators

Mirko Piersanti University of L'Aquila, Italy	Effective dialogue with end users, such as power companies and satellite operators

Anonymous	Detection/observation
Because we haven't had an Earth-impacting Carrington-like event in the modern era we really have little knowledge as to how modern infrastructure would really withstand such an event.
Anonymous	Effective dialogue with end users, such as power companies and satellite operators

Anonymous	Effective dialogue with end users, such as power companies and satellite operators

Anonymous	End users willing and able to act on forecast advice

John Haiducek US Naval Research Laboratory	Distinguishing damaging events from those with negligible effect
Lack of observations to constrain CME characteristics prevents us from forecasting their effects in detail prior to their arrival at L1, severely limiting forecast lead times.
Daniel Welling University of Michigan	End users willing and able to act on forecast advice

Anonymous	Detection/observation

Anonymous	Distinguishing damaging events from those with negligible effect
Especially the prediction of the Bz component is not possible with sufficient accuracy before it arrives at L1.
Anonymous	End users willing and able to act on forecast advice
End users need to be educated about understanding and how to recognise space weather impacts
Simon Machin Met Office	Distinguishing damaging events from those with negligible effect
The greatest challenge is detecting/diagnosing the magnetic characteristics of CMEs at an early stage following eruption. This would be the "magic bullet" to enabling truly actionable high confidence forecasts.
Anonymous	Distinguishing damaging events from those with negligible effect

Anonymous	Detection/observation
Information on IMF Bz in CMEs near the sun and propagation to Earth
Phillip Chamberlin University of Colorado	End users willing and able to act on forecast advice

Anonymous	End users willing and able to act on forecast advice
While I think much material is available on the impact of geomagnetic storms, end users (e.g., airline/powergrid operators) are generally unfamiliar with what protocol to follow when a significant storm is forecast.
Jingnan Guo USTC	Effective dialogue with end users, such as power companies and satellite operators

Sean Bruinsma CNES	Distinguishing damaging events from those with negligible effect
you need to know direction of Bz - and we know that only at L1
Anonymous	Distinguishing damaging events from those with negligible effect

Xiukuan Zhao Institute of Geology and Geophysics, Chinese Academy of Sciences	Detection/observation

Anonymous	Distinguishing damaging events from those with negligible effect

Xinan Yue Institute of Geology and Geophysics, Chinese Academy of Sciences	Detection/observation

Anonymous	Detection/observation

Anonymous	Distinguishing damaging events from those with negligible effect

Dean Pesnell NASA GSFC	Distinguishing damaging events from those with negligible effect
Reducing false alarms is as important as improving the reliability of true alerts.
Anonymous	End users willing and able to act on forecast advice

Anonymous	Effective dialogue with end users, such as power companies and satellite operators

Anonymous
No idea. space weather event is such a broad category, the answers are likely different according to phenomena & event.
Anonymous	Detection/observation

Anonymous	Distinguishing damaging events from those with negligible effect

Anonymous	Distinguishing damaging events from those with negligible effect
It's hard to know what the biggest challenge is...
Alexander Mishev University of Oulu	Detection/observation

Leon Golub Smithsonian Astrophysical Observatory	Distinguishing damaging events from those with negligible effect

Anonymous	End users willing and able to act on forecast advice

Noe Lugaz University of New Hampshire	Distinguishing damaging events from those with negligible effect
I am too removed from end users to have a meaningful answer.
Anonymous	Detection/observation

Anonymous	Detection/observation
We have so few cases of really intense storms to study.
Anonymous	Detection/observation

Eelco Doornbos KNMI	Distinguishing damaging events from those with negligible effect
I believe that end users would be willing and able to react, if we would be able to provide them with the right actionable information. The current lack of end user engagement is to a large extent due to the current limitations in the actionable information on offer.
Anonymous	Detection/observation
Ground-based and space-in-situ observations for space weather study is needed in a timely manner.
Michael Liemohn University of Michigan	Distinguishing damaging events from those with negligible effect

Marianna Korsos University of Catania	End users willing and able to act on forecast advice
we need more accurate sw models and show they are works properly. So the end users will act on forecast advise.
Anonymous	Distinguishing damaging events from those with negligible effect

Jean Uwamahoro University of Rwanda	Detection/observation

Anonymous	Distinguishing damaging events from those with negligible effect
I think the biggest challenge is the lack of a comparable event in recent decades for the current level of technology. We can't even provide accurate, actionable forecasts for the consequences of events from recent years.
Anonymous	End users willing and able to act on forecast advice

Anonymous	End users willing and able to act on forecast advice

Matthew Lang British Antarctic Survey	Detection/observation

Anonymous	Distinguishing damaging events from those with negligible effect
All of these points are important but until we can get a better handle on distinguishing damaging events from lesser events it is hard to get buy-in from the end users. Its costly to take action which isn't ultimately needed and it also undermines trust in the expertise if we can't provide accurate/robust forecasts with long enough lead time.
Anonymous	Detection/observation
Fast events will be covered by fewer observations, compounding the low lead time with less information to inform the forecast models.
Anonymous	End users willing and able to act on forecast advice

Anonymous	Distinguishing damaging events from those with negligible effect

Jonny Rae Northumbria University	Effective dialogue with end users, such as power companies and satellite operators
how do we know what we need to concentrate on if the end users don't share what they can?
James Adams University of Alabama in Huntsville	Distinguishing damaging events from those with negligible effect
I took this to refer to forecasts made at the time of the outburst on the Sun or earlier. Before the outburst, we can only know that there is a lot of non-potential energy available. Just after the outburst, we cannot be sure how the ICME will develop, i.e. whether the worst effects will mss Earth.
Anonymous	Detection/observation

Anonymous	Detection/observation

Lucilla Alfonsi Istituto Nazionale di Geofisica e Vulcanologia (INGV)	Distinguishing damaging events from those with negligible effect

Anonymous	Distinguishing damaging events from those with negligible effect

Anonymous	Detection/observation

Ed Thiemann LASP, University of Colorado	End users willing and able to act on forecast advice

Anthony Mannucci Jet Propulsion Laboratory, California Institute of Technology	Detection/observation
We would need to know speed of the CME and strength and orientation of Bz.
Anonymous
I am not qualified to answer this question. I know that we would detect one with the existing infrastructure. I do not know which of the last three dominates.
Mike Hapgood RAL Space	Effective dialogue with end users, such as power companies and satellite operators
I would stress the vital role of dialogue - for space weather experts to understand user needs, especially their wider operational context - and to help users understand and gain confidence in what our experts can provide.
Anonymous	End users willing and able to act on forecast advice
All of the above. Forecasts are very very unreliable, so end users are unwilling to enguage.
Anonymous	End users willing and able to act on forecast advice

Anonymous	Detection/observation

Sean Elvidge University of Birmingham	Detection/observation
I typically think of an "actionable forecast" as one which is of sufficient skill that end users could make decisions from.
Anonymous	Effective dialogue with end users, such as power companies and satellite operators

Martin Mlynczak NASA Langley Research Center	Detection/observation
There has to be high skill forecasts combined with in-place societal plans for an event, such as for hurricane forecasting. At-risk communities have evacuation and shelter plans in place in the event of a hurricane. Are there similar plans for SpWx events?
Allison Jaynes University of Iowa	Detection/observation

Anonymous	Detection/observation

Bernard V Jackson University of California, San Diego	Distinguishing damaging events from those with negligible effect
A Carrington event will probably be able to be detected and observed, but it will probably not be possible to tell how damaging it will be.
Anonymous	Distinguishing damaging events from those with negligible effect

Yaqi Jin University of Oslo	Distinguishing damaging events from those with negligible effect

Anonymous	Distinguishing damaging events from those with negligible effect
I marked "Distinguishing damaging events from those with negligible effect" even though "End users willing and able to act on forecast advice" is also important. I think because we don't have a lot of events to go by, it is currently difficult to be confident in whether an event is a CR event. Before we can convince end users to act fast when we predict a CR event to happen, we first have to be able to distinguish whether such an event is coming our way.
Anonymous	Distinguishing damaging events from those with negligible effect
The two last points are partly owing to the uncertainty of point 2
Anonymous	Detection/observation
Not seeing the farside and limited knowledge of the solar poles-especially magnetograph data. Also rapid-response time-dependent event modeling.
Dario Del Moro University of Rome Tor Vergata	Detection/observation

Huw Morgan Aberystwyth University	Distinguishing damaging events from those with negligible effect
A large event should be picked up by forecasters. The problem lies in estimating the geoeffectiveness of the event.
Mateja Dumbovic Hvar Observatory, University of Zagreb	Distinguishing damaging events from those with negligible effect
This will reduce the false alarms ratio and thus make our forecast advices more reliable and trustworthy
Yuhao Wang Nanchang University	Detection/observation
No
Anonymous	Distinguishing damaging events from those with negligible effect
On par with detection/observation, and "Effective dialogue with end users" a close third. I don't think "End users willing and able to act on forecast advice" is an issue.
Rajkumar Hajra University of Science and Technology of China	Detection/observation

Anonymous	End users willing and able to act on forecast advice

Jayachandran P.T. University of New Brunswick	Distinguishing damaging events from those with negligible effect
We, as a community, scare the end users by overselling the impact. We also put too much focus on major events and event studies.
Mike Lockwood University of Reading	Distinguishing damaging events from those with negligible effect
The answer to this does depend on the hazard and the industry involved. The cost-loss balance is different in every case and some industries/operations have the finances to withstand major losses, others do not.
Anonymous	End users willing and able to act on forecast advice

Ciaran Beggan British Geological Survey	End users willing and able to act on forecast advice
As noone living has seen or experienced such an event, we fall somewhere between chicken licken (sky falling in) and full on black swan event where we don't know what we don't know.
Anonymous	Detection/observation

Anonymous	Detection/observation

Mark Moldwin University of Michigan	Distinguishing damaging events from those with negligible effect
Until thinks break and costs an industry, folks usually don't pay attention.
Daniel Brandt Michigan Tech Research Institute	Detection/observation

David R. Themens University of Birmingham	Detection/observation

Anonymous	Effective dialogue with end users, such as power companies and satellite operators
Kind of none of the above for me. I think the answer would really be that the rarer the event, the inherently more challenging it is to predict. For terrestrial weather and for earthquakes, we've got much better detection/observation, ability to distinguish impactful from not, have effective dialogue with end users, and end users are generally willing/able to act (with some percentage always being stubborn) -- and even with all of that, those communities can't provide actionable forecasts way far out in advance of the most extreme events. I think the best we can do in all cases is build up the resilience in the systems to minimize the impacts.
Anonymous	Detection/observation

Anonymous	Distinguishing damaging events from those with negligible effect

Anonymous	Distinguishing damaging events from those with negligible effect

Anonymous	End users willing and able to act on forecast advice
Validation is the biggest issue: " delivering actionable forecasts" will require some kind of validation of the prediction of the effects. What effects are you looking at?
Anonymous	Distinguishing damaging events from those with negligible effect

Anonymous	Distinguishing damaging events from those with negligible effect
Better coverage of in-situ particle and magnetic instrumentation is critical.
Anonymous	Detection/observation
The most significant challenge to forecasting extreme events is the lack of observations of their onset/propagation, and the lack of deep physical understanding and high-quality predictive models that flow from this.
Larry C Gardner Utah State University Eastern	Detection/observation
First we need to understand what causes a Carrington-scale event, and then worry about how to predict it. Currently I am not sure we could even give any advanced warning.
Piers Jiggens ESA	End users willing and able to act on forecast advice
I don't think anyone would react, maybe civilian aircraft operators because the cost is well-bounded and they have experience. Others would likely just put more people on operations if they do anything.
Bernd Heber Christian-Albrechts-Universität	Distinguishing damaging events from those with negligible effect

Jacob Bortnik UCLA	Distinguishing damaging events from those with negligible effect
We can see larger solar flares well, but it's hard to tell exactly how they will impact Earth.
Anonymous	Distinguishing damaging events from those with negligible effect

Anonymous	Effective dialogue with end users, such as power companies and satellite operators

Anonymous	Effective dialogue with end users, such as power companies and satellite operators

Question 5

Assume that end users can take effective action for an extreme geomagnetic storm if given a sufficiently accurate forecast with 5 hours notice. Given current observation and modeling capability, what is the likelihood that forecasts will be sufficiently accurate by that time?

Results

Participant	Response	Confidence
Edmund Henley Met Office (views my own)	Somewhat likely
Assuming observable CME initial speed is a ~proxy for extreme geomagnetic storms (Richardson+ 2011). This gives >~12H notice, but no updates until ~L1: operational space weather forecasts haven’t yet drawn on HI capabilities to mimic terrestrial trade-off of lead-time for skill (e.g. narrower “cone of terror” as hurricane approaches).
Anonymous
Not sure what you mean by "sufficiently accurate by that time" so didn't answer question
Ryan McGranaghan NASA Jet Propulsion Laboratory	Roughly even

Anonymous

Dr Mario M. Bisi UKRI STFC RAL Space	Somewhat unlikely
We will still not know the magnetic-field orientation at that point, which is crucial to know how "extreme" the incoming CME/CMEs might be, but also we may have lost some of the infrastructure at that point already if there are extreme SEP events associated with the CME(s) launched from the Sun towards the Earth.
Anonymous	Somewhat unlikely

Ian Mann University of Alberta	Somewhat unlikely
Its not clear to me if this question is based on assuming current capabilities if we had a 5 hour advance warning upstream? Also "sufficiently accurate" depends on the application so is rather vague.
Anonymous	Somewhat likely

Anonymous	Somewhat unlikely

Anonymous	Somewhat likely

Anonymous	Highly unlikely
Observations do not much support accurate 5 hours notices due to the lack of in-situ measurements between the Sun and Earth. This restricts modelling capabilities too.
Anonymous	Highly likely

Anonymous	Somewhat likely

Tamas Gombosi University of Michigan	Somewhat likely

Gabor Toth University of Michigan	Roughly even
There have been seemingly large impacts with little impact and seemingly moderate events with surprisingly large impacts. The main unknown is the sign and magnitude of Bz.
Harlan Spence University of New Hampshire	Roughly even
With such a short lead time relative to transit times from L1, we could probably provide a range of outcomes of an extreme event based on the controlling factors we do not yet know. That range of predicted outcomes might be too large to be actionable by a user.
Anonymous	Highly unlikely
There's no demonstrated track record of success at this point.
Robert F. Wimmer-Schweingruber Kiel University, Kiel, Germany	Highly unlikely

Anonymous	Somewhat likely

Anonymous	Somewhat likely

Evangelos Paouris George Mason University	Roughly even
In most cases, I think this is true. But, we have to keep in mind that many eruptive events in the past seem that could produce a geomagnetic storm but nothing happened. For example, we can predict the arrival time of a CME on Earth but the geoeffectiveness is a completely different issue.
Karim Meziane University of New Brunswick	Somewhat unlikely

Anonymous	Somewhat unlikely

Anonymous	Somewhat likely

Vincenzo Romano Istituto Nazionale di Geofisica e Vulcanologia	Roughly even

Delores Knipp University of Colorado Boulder	Roughly even

Anonymous	Somewhat unlikely

Anonymous	Somewhat likely

Anonymous	Somewhat likely

Ankush Bhaskar Space Physics Laboratory, VSSC, India	Somewhat likely
Arrival times of ICMEs are highly uncertain and even the strength of ICMEs. So forecasting of impact having a 5-hour time horizon may be tricky.
Anonymous	Somewhat likely

Anonymous	Roughly even

Anonymous	Somewhat unlikely

Anonymous	Somewhat likely

Anonymous	Roughly even

Anonymous	Somewhat unlikely

Anonymous	Roughly even

Mirko Piersanti University of L'Aquila, Italy	Somewhat unlikely

Anonymous	Somewhat likely

Anonymous	Somewhat unlikely

Anonymous	Highly likely
If we speak of geomagnetic storms only, the passage time of a CME is a few tens of hours which is sufficient to make a 5-hr accurate forecast by tracing SMEs.
Anonymous	Roughly even

John Haiducek US Naval Research Laboratory	Somewhat unlikely
It depends on what is considered "sufficiently accurate," but any forecast that relies on L1 observations is unlikely to be ready 5 hours in advance, and a forecast that does not will have considerably less accuracy.
Daniel Welling University of Michigan	Roughly even

Anonymous	Roughly even

Anonymous	Somewhat unlikely
A CME with a speed of 1000 km/s needs less than half an hour from L1 to Earth. Additionally, a carrington type event is likely much faster than that. However, a fast and massive CME is able to produce a strong geomagnetic storm even if Bz is positive. So, predicting its magnetic field is highly unlikely but predicting a strong geomagnetic storm due to speed and density is possible.
Anonymous	Somewhat unlikely

Simon Machin Met Office	Somewhat unlikely
Until we have a means to characterise the magnetic characteristics of CMEs at >5 hours prior to arrival at Earth, we can only offer low-moderate confidence forecasts prior to arrival at L1.
Anonymous	Roughly even

Anonymous	Somewhat unlikely

Phillip Chamberlin University of Colorado	Somewhat likely

Anonymous	Somewhat likely

Jingnan Guo USTC	Somewhat unlikely

Sean Bruinsma CNES	Highly unlikely
the CME scoreboard at CCMC demonstrates the mediocre quality of predictions
Anonymous	Somewhat unlikely

Xiukuan Zhao Institute of Geology and Geophysics, Chinese Academy of Sciences	Somewhat likely

Anonymous
Don't know about #13.
Xinan Yue Institute of Geology and Geophysics, Chinese Academy of Sciences	Somewhat unlikely

Anonymous	Roughly even

Anonymous	Somewhat likely

Dean Pesnell NASA GSFC	Somewhat likely
The second question refers to an unspecified time to develop the forecasts but then assumed the observations and modeling would remain constant?
Anonymous	Somewhat likely

Anonymous	Highly likely

Anonymous
again, which part of an event? I can't answer this, "it depends". Radio / ionospheric disturbances vs. GICs....very different situation.
Anonymous	Somewhat unlikely

Anonymous	Roughly even

Anonymous	Roughly even

Alexander Mishev University of Oulu	Somewhat likely

Leon Golub Smithsonian Astrophysical Observatory	Roughly even

Anonymous	Roughly even

Noe Lugaz University of New Hampshire	Highly unlikely

Anonymous	Somewhat likely

Anonymous	Roughly even

Anonymous	Somewhat unlikely

Eelco Doornbos KNMI	Highly unlikely
5 hours is right in the grey zone of heliospheric simulations 1-2 days out (not actionable) and L1 in-situ observations 30 mins out (actionable).
Anonymous	Somewhat unlikely
Given the dynamic and spatial of solar wind propagation, my assessment on accuracy is not optimistic.
Michael Liemohn University of Michigan	Roughly even
1 or 2 hours notice, okay, but 5 hours...not yet.
Marianna Korsos University of Catania	Roughly even

Anonymous	Somewhat unlikely

Jean Uwamahoro University of Rwanda	Somewhat likely

Anonymous	Highly unlikely
From a power grid perspective, the current solar-wind-to-magnetosphere models are nowhere near accurate enough to give actionable forecasts - and even if they could, 5 hours is likely not enough to implement (currently undefined) counter-measures.
Anonymous	Somewhat likely

Anonymous	Somewhat likely

Matthew Lang British Antarctic Survey	Somewhat unlikely

Anonymous	Roughly even

Anonymous	Somewhat likely

Anonymous	Somewhat likely

Anonymous	Roughly even

Jonny Rae Northumbria University	Roughly even

James Adams University of Alabama in Huntsville
I don't feel that I know enough to answer.
Anonymous	Somewhat likely

Anonymous	Somewhat likely

Lucilla Alfonsi Istituto Nazionale di Geofisica e Vulcanologia (INGV)	Somewhat likely

Anonymous	Somewhat unlikely

Anonymous	Somewhat likely

Ed Thiemann LASP, University of Colorado	Somewhat unlikely

Anthony Mannucci Jet Propulsion Laboratory, California Institute of Technology	Highly unlikely
This seems like an oddly-posed question, but OK. The assumption of what end users can do with a good forecast is independent of whether a good forecast can be produced.
Anonymous
I know enough to know that I don't have the information to answer this question.
Mike Hapgood RAL Space	Somewhat likely
The elephant in the room is Bz. Are we seeking to forecast Bz? Or just that a CME is coming and it will be bad when Bz is south? My experience is that users can use a warning like that, if primed to expect surges of bad space weather.
Anonymous	Highly unlikely
SPE events are more important than Geomagnetic storms for inherent risk to earth systems. These arrive is < 20 mins from eruption.
Anonymous	Roughly even

Anonymous	Somewhat likely

Sean Elvidge University of Birmingham	Roughly even
In what systems? In the ionosphere/thermosphere maybe? In the GIC community maybe something else? In SEP community, something else again?
Anonymous	Somewhat unlikely

Martin Mlynczak NASA Langley Research Center	Somewhat unlikely
We've never had an extreme event with today's models so it is not possible to know if we will have an accurate forecast for one. Forecast skill development requires many forecasts and associated model improvements to gauge ability for future storms.
Allison Jaynes University of Iowa	Somewhat unlikely

Anonymous	Somewhat unlikely

Bernard V Jackson University of California, San Diego	Somewhat unlikely

Anonymous	Somewhat likely

Yaqi Jin University of Oslo	Roughly even
The forecast of extreme events are still not accurate. It is not easy to forecast the propagation of CMEs toward the Earth, not to say to forecast of direction of magnetic field.
Anonymous	Somewhat unlikely
I personally think the timing isn't as important for a CR like event. Being able to distinguish between a non-damaging and a damaging event is much more important. Whether we have a 5 hour accuracy or 7 hour accuracy, wouldn't matter that much, compared to the possible damage of the event. If end users would have the guarantee that we are correct in our prediction of a CR event, then the accuracy of the forecast is just a detail. (But then again, I think this is a question more for users than for scientists).
Anonymous	Highly unlikely
I am uncertain if I understood the question right. Howerver, I mean that we don't have the capability to issue high enough quality forecasts with 5 hours lead time...
Anonymous	Somewhat unlikely
depends on what is being forecasted. SEPs? GICs? Ionospheric disturbances? PCAs?
Dario Del Moro University of Rome Tor Vergata	Somewhat likely

Huw Morgan Aberystwyth University	Roughly even
Storm time of onset difficult. We cannot currently give +/-5 hours confidence
Mateja Dumbovic Hvar Observatory, University of Zagreb	Somewhat likely
This very much depends on what is considered 'sufficiently accurate'
Yuhao Wang Nanchang University	Somewhat unlikely
No
Anonymous	Somewhat likely

Rajkumar Hajra University of Science and Technology of China	Somewhat unlikely

Anonymous	Somewhat unlikely

Jayachandran P.T. University of New Brunswick	Highly unlikely
Transmission line infrastructure upgrades and modernization should prevent this from happening.
Mike Lockwood University of Reading	Somewhat likely
We will make progress with targeted observations and faster and more comprehensive modelling capabilities .... but how much progress towards ideal capabilities is far from certain
Anonymous	Somewhat likely

Ciaran Beggan British Geological Survey	Somewhat unlikely

Anonymous	Somewhat likely

Anonymous	Somewhat unlikely

Mark Moldwin University of Michigan	Somewhat likely
Depending on willingness to accept false positives...
Daniel Brandt Michigan Tech Research Institute	Somewhat likely

David R. Themens University of Birmingham	Highly unlikely
This question depends entirely on the application/user in question. Forecasts of the ionospheric/thermospheric state beyond three hours struggle to beat climatology and no model that I know of can beat climatology after more than six hours (even then, anything more than 2 hours is mainly improvement via de-biasing). If the application is HF remote sensing coordinate registration, no existing forecast is remotely close to actionable. Only nowcasts in highly instrumented regions are sufficient in an HF remote sensing context.
Anonymous	Somewhat likely

Anonymous	Somewhat likely

Anonymous	Somewhat unlikely

Anonymous	Somewhat likely

Anonymous
What end users? I assume power grid./What time are you referring to? The real question is for a particular phenomena, how long (if ever) will it take for us to be able to make an actionable forecast good for 5hrs, 12hrs, 24hrs, 72hrs.
Anonymous	Somewhat unlikely

Anonymous	Roughly even

Anonymous	Highly unlikely
Without additional observations and data-drive models, it seems unlikely we can achieve an accurate forecast five hours in advance of an event.
Larry C Gardner Utah State University Eastern	Highly unlikely
The current state of the art can not even forecast steady state situations, so a Carrington-Scale event prediction would be luck at best.
Piers Jiggens ESA	Roughly even

Bernd Heber Christian-Albrechts-Universität	Roughly even

Jacob Bortnik UCLA	Highly likely
We can already do almost 5 hr ahead forecast.
Anonymous	Roughly even
The B-z forecast is still really rough for large events.
Anonymous	Highly unlikely

Anonymous	Somewhat likely

Question 6

Given everything you are able to estimate about occurrence of extreme space weather and susceptibility of our technical infrastructure, what is the probability in the next 10 years that a space weather event will cause unplanned regional power outages?

Results

Participant	Response	Confidence
Edmund Henley Met Office (views my own)	29%
Likely highballing. Naively considering 1/30 year event: examining economic value, Eastwood+ 2018 ascribed this to 1989 storm: Quebec’s regional power outage appears in Wikipedia’s major power outages; Halloween storms of 2003 (1/10 year event) does not qualify.
Anonymous	10%

Ryan McGranaghan NASA Jet Propulsion Laboratory	5%
This is a low probability, but not at all negligible. A big factor, however, is compounding effects. There could be situations where space weather is even moderate, but with the right contextual information (e.g., terrestrial weather) it causes blackouts or extreme disruption. The probability is essentially that of another Carrington-like event occurring. I believe it extremely likely that if that level of event occurred, we would see widespread blackout.
Anonymous	5%

Dr Mario M. Bisi UKRI STFC RAL Space	25%
None at this time.
Anonymous	50%

Ian Mann University of Alberta	10%

Anonymous	40%

Anonymous	0.005%
to cause an outage, the impact must happen when the grid is loaded (winter/heat wave). Such occurrences will become more likely as global warming impacts surface temperatures but not before 20-30 years in the future
Anonymous	50%

Anonymous	10%
10% is a conservative guess based on published extreme value analysis of geomagnetic variations and on the reported previous power outages, whose severity varies from very minor to one serious case (Québec blackout).
Anonymous

Anonymous	30%

Tamas Gombosi University of Michigan	30%

Gabor Toth University of Michigan	90%
Unless power grids improve and get well prepared, it is very likely to happen. It will not be the end of the world event, just a local outage, most likely.
Harlan Spence University of New Hampshire	3%
Recent studies have reached values up to 10% for the occurrence of extreme storms, but that depends very much on what you call "extreme" and how big an event you need to impact regional power outages.
Anonymous	15%

Robert F. Wimmer-Schweingruber Kiel University, Kiel, Germany	20%
This will depend a lot on where, when and how strong. Moreover, how well will we be prepared in 10 years?
Anonymous	50%

Anonymous	10%

Evangelos Paouris George Mason University	30%
This question is a difficult one. Let me explain: if you are talking about a mid/low latitude region the probability is less than 1% but for high latitude region this is very high. For the regional power outages, we have to take into account many factors like the age of the grid and what are the boundaries of the "regional", e.g. just a few kilometers, or entire countries?
Karim Meziane University of New Brunswick	10%

Anonymous	50%

Anonymous	10%

Vincenzo Romano Istituto Nazionale di Geofisica e Vulcanologia	30%

Delores Knipp University of Colorado Boulder	30%

Anonymous	100%

Anonymous	50%

Anonymous	75%

Ankush Bhaskar Space Physics Laboratory, VSSC, India

Anonymous

Anonymous	10%

Anonymous	20%

Anonymous

Anonymous	80%

Anonymous	10%

Anonymous	10%

Mirko Piersanti University of L'Aquila, Italy	50%

Anonymous	5%

Anonymous

Anonymous	25%
Since they occur roughly a few per century, 25% for the next decade.
Anonymous

John Haiducek US Naval Research Laboratory	50%
Based on the assumption that regional power outages attributable to space weather occur about once every 17 years (i.e. a 6% probability in a given year), I would estimate a 45% probability in the next 10 years. I round my estimate up to 50% since the next 10 years include solar max.
Daniel Welling University of Michigan	85%

Anonymous	50%

Anonymous	20%

Anonymous	70%
highly depends on technical details of the powergrid and ground electric conductivity
Simon Machin Met Office	90%
It would be highly unlikely if we were not to experience a G5/G5+ storm over the next decade. The last 15 years has been an anomalously low period of activity which is not expected to persist. There is little evidance to indicate that power grids globally have become more resilient to space weather and may in fact be becoming more vulnerable (e.g. due to distributed generation/renewables).
Anonymous	80%

Anonymous	10%

Phillip Chamberlin University of Colorado	5%
Predicted to be a moderate to low solar cycle
Anonymous	33%
Depends on the activity of the next Solar cycle, looks to be more active than the last 2 weak cycles at the moment, but average compared across the modern maximum
Jingnan Guo USTC	50%

Sean Bruinsma CNES	10%

Anonymous	25%
First of all, the issue is not extreme events. Those are the 'easy' to predict and to take counter actions against. The problem is medium events that are still damaging but much harder to predict. Second, machine learning will completely revolutionize the field in the next 10 years so whatever we are able to predict now (very little) is irrelevant!
Xiukuan Zhao Institute of Geology and Geophysics, Chinese Academy of Sciences

Anonymous

Xinan Yue Institute of Geology and Geophysics, Chinese Academy of Sciences	40%

Anonymous	20%

Anonymous	60%

Dean Pesnell NASA GSFC	70%
Only regional outages.
Anonymous

Anonymous

Anonymous	30%

Anonymous	0.1%

Anonymous

Anonymous	10%
No
Alexander Mishev University of Oulu	10%

Leon Golub Smithsonian Astrophysical Observatory	67%

Anonymous	70%

Noe Lugaz University of New Hampshire	25%

Anonymous	50%

Anonymous	5%
We don't have enough statistics to even estimate a probability
Anonymous	10%

Eelco Doornbos KNMI	5%
I'm taking into account the significantly lower occurrence rate of G4 storms and lack of G5 geomagnetic storms over the last 15-20 years, compared to the rates during the decades before. I think it is too early to say whether this will stay the same during the next 10 years.
Anonymous	100%
If we don’t learn from what happened in the past, history could repeat itself again and again.
Michael Liemohn University of Michigan	95%
Regional...almost certainly, as this is already the case.
Marianna Korsos University of Catania	50%
occurrence of extreme space will be occurring but the questions are when? will it be directed to earth or not? e like in the case of the 2012 event.
Anonymous	20%

Jean Uwamahoro University of Rwanda	10%

Anonymous	30%
Damage from GICs will probably become more likely with higher-voltage lines becoming more common, but at the same time awareness is also spreading, and some regions are already implementing transformer-based safety measures.
Anonymous

Anonymous

Matthew Lang British Antarctic Survey	25%

Anonymous	5%

Anonymous	50%
I estimated a high likelihood above as the definition of 'regional' is not well defined and could include outages that would be small and highly localised.
Anonymous	5%

Anonymous	30%

Jonny Rae Northumbria University	50%
It's a coin toss
James Adams University of Alabama in Huntsville	10%
This is just a guess.
Anonymous	22%

Anonymous	30%

Lucilla Alfonsi Istituto Nazionale di Geofisica e Vulcanologia (INGV)	30%

Anonymous	10%

Anonymous	50%

Ed Thiemann LASP, University of Colorado	50%

Anthony Mannucci Jet Propulsion Laboratory, California Institute of Technology	20%

Anonymous	5%
That risk is significantly mitigated by the existence of the SWPC and its space weather warning processes. Without SWPC activity I would cast it as more like 50%.
Mike Hapgood RAL Space	20%
Took likelihood of a extreme event as 10% per decade, in line with some of the higher probability values in the literature, also extreme events in 1770, 1859 and 1921. I've doubled it to reflect concerns about future grid evolution, e.g. lower inertia of solar and wind generation.
Anonymous	15%

Anonymous	30%

Anonymous	10%

Sean Elvidge University of Birmingham	5%
Carrington = 1 in 100 year. Decade 10 year. Prob ~~~10% per decade. Could cause regional power outages for up to hours. Could obviously means probability/2. Therefore 5%.
Anonymous	50%

Martin Mlynczak NASA Langley Research Center	50%
Given the relative lull in activity over the past 20 years and the active onset of SC 25, we may see much stronger storms. These will be the test of our model improvements.
Allison Jaynes University of Iowa	10%

Anonymous	20%

Bernard V Jackson University of California, San Diego	50%

Anonymous	50%

Yaqi Jin University of Oslo	50%
No idea. Just yes or no, so it is 50%. The current solar maximum seems to be stronger than expected.
Anonymous	10%

Anonymous	75%
A G4-G5, at least, if it has enough transient geomagnetic activity associated with it, will almost definitively cause power outs, especially during high load on the grid.
Anonymous
see above
Dario Del Moro University of Rome Tor Vergata	25%

Huw Morgan Aberystwyth University	10%
10% for the UK? Very hard to judge due to poor statistics on very large events.
Mateja Dumbovic Hvar Observatory, University of Zagreb	70%
We are nearing next solar maximum and I am confident a storm will occur which is able to cause regional power outages, however, it may not actually happen due to increasing resistance of our technological systems to space weather effects
Yuhao Wang Nanchang University	70%
No
Anonymous	75%

Rajkumar Hajra University of Science and Technology of China	90%

Anonymous	1%

Jayachandran P.T. University of New Brunswick	10%
No.
Mike Lockwood University of Reading	10%
Hard to answer as the level of risk varies around the globe and the robustness of infrastructure variea also. For the UK I would put the risk at 5-15%, the range being associated with the unknown effect of more grid "limbs" because of renewable power systems being increasingly brought into the network. Worldwide the number would be considerably higher.
Anonymous	75%

Ciaran Beggan British Geological Survey	5%
The Sun is in a fairly quiescent state and has been declining since the 1960s. It seems unlikely that there's enough sunspot activity to create an unprecedented extreme event but who knows.
Anonymous	10%

Anonymous	20%

Mark Moldwin University of Michigan	90%
As power grids become more stressed and aged and time between last major event, grids are more likely to suffer impacts to GIC.
Daniel Brandt Michigan Tech Research Institute	50%

David R. Themens University of Birmingham	10%
In the UK?
Anonymous	5%
The "next 10 years" part is critical. As with predictions of any extreme events, they are improbable on short time scales but inevitable on long time scales.
Anonymous	33%
We've seen a few regional power outages and power grid effects (e.g., transformer failure) in the last ~35 years at a wide range of locations from high to low/equatorial latitudes. It's at least somewhat likely another will occur soon since worldwide power grids have not been uniformly upgraded.
Anonymous	10%

Anonymous	10%

Anonymous	1%
Space weather has many practical impacts beyond GICs and power grid failures. Regional failures occur frequently due to a wide range of causes. The assumption is that a Carrington event will create a sustained impact - this is not clear at all.
Anonymous	30%

Anonymous	35%

Anonymous	20%

Larry C Gardner Utah State University Eastern	60%
Again, all of this is just guess work. Until we can expand the physics of these systems, there is no chance of ever coming close to predictions.
Piers Jiggens ESA	20%
This cycle is still not very active which is driving the number I give.
Bernd Heber Christian-Albrechts-Universität	30%

Jacob Bortnik UCLA	75%
Every solar cycle has large events, it's just a question of being in an unlucky position to experience large impact.
Anonymous	20%

Anonymous	50%

Anonymous	50%

Question 7

Improved observations of which of these physical domains would have the highest return on investment for space weather forecasting?

Results

Participant	Response	Confidence
Edmund Henley Met Office (views my own)	Near-Sun heliosphere (0.1 - 0.7 AU)
Domain specific! Some domains may need little investment for good returns. However, considering absolute cost, and extreme CMEs: having an operational ability to track CMEs beyond the coronagraph domain, as Vigil’s HI should, will let us trade-off forecast lead time for skill as extreme CMEs travel Earthwards.
Anonymous	Near-Earth heliosphere (0.7- 1 AU)

Ryan McGranaghan NASA Jet Propulsion Laboratory	Ionosphere/Thermosphere

Anonymous

Dr Mario M. Bisi UKRI STFC RAL Space	Near-Sun heliosphere (0.1 - 0.7 AU)
It is essential that we improve the observational and measurement coverage of the near-Sun and mid-travel areas of space between the Sun and Earth (and in a good cone angle around the Sun-Earth line) to better understand how structures evolve, rotate, and interact during transit between the Sun and the Earth, and also to fill in key parameters that help us to predict geomagnetic effects such as magnetic field, velocity, density, mass, size, direction, etc...
Anonymous	Near-Earth heliosphere (0.7- 1 AU)

Ian Mann University of Alberta	Magnetosphere
It is not sufficient to forecast strong Bz negative at the upstream magnetosphere. Of course that is important for developing long lead time forecasts. However, we need to understand how incidence solar wind is processed in the MI system to produce extreme effects, including hysteresis. I still think we are not very advanced in this regard.
Anonymous	Photosphere

Anonymous	Near-Sun heliosphere (0.1 - 0.7 AU)
taking direct but distributed measurements upstream of L1 would be the most straightforward way to extend forecasting horizon (for Bz, at least)
Anonymous	Near-Earth heliosphere (0.7- 1 AU)

Anonymous	Near-Sun heliosphere (0.1 - 0.7 AU)
Sparsity of in-situ observations between the Sun and Earth is presently the most restricting factor in following the progress of coronal mass ejections.
Anonymous	Corona

Anonymous	Near-Sun heliosphere (0.1 - 0.7 AU)

Tamas Gombosi University of Michigan	Corona

Gabor Toth University of Michigan	Near-Sun heliosphere (0.1 - 0.7 AU)
Solar wind is essentially a 1D MHD flow beyond the Alfven surface. Observation along the radial trajectory would provide reasonably accurate and reliable prediction.
Harlan Spence University of New Hampshire	Near-Sun heliosphere (0.1 - 0.7 AU)
We critically need the IMF Bz orientation well upstream of Earth. Studies of how large structures propagate from the inner heliosphere to 1 AU suggests that we could indeed have robust predictions of IMF Bz if we could monitor it from the inner heliosphere.
Anonymous	Near-Sun heliosphere (0.1 - 0.7 AU)
Understanding the character of earthward-propagating solar ejecta is vitally important, so getting earlier observations would be highly valuable.
Robert F. Wimmer-Schweingruber Kiel University, Kiel, Germany	Solar interior
The combination with corona and near-sun is important so that we can verify models of flux emergence and track that to ICMEs and then predict speed. Strange that I can only tick one box!
Anonymous	Photosphere

Anonymous	Magnetosphere

Evangelos Paouris George Mason University	Corona
From the CME perspective (post-event), we need observations from multiple viewpoints, especially from L5. From the Flare+CME forecasting perspective (pre-event) we need also photospheric observations and magnetograms also from L5.
Karim Meziane University of New Brunswick	Near-Sun heliosphere (0.1 - 0.7 AU)

Anonymous	Photosphere

Anonymous	Near-Sun heliosphere (0.1 - 0.7 AU)

Vincenzo Romano Istituto Nazionale di Geofisica e Vulcanologia	Photosphere

Delores Knipp University of Colorado Boulder
We should be looking at active region complexity and interaction with other structures in the corona.....and making deep investments in measuring thermosphere/ionosphere variability and mega constellation vulnerability to these
Anonymous	Near-Sun heliosphere (0.1 - 0.7 AU)

Anonymous	Corona

Anonymous	Near-Sun heliosphere (0.1 - 0.7 AU)

Ankush Bhaskar Space Physics Laboratory, VSSC, India	Near-Earth heliosphere (0.7- 1 AU)

Anonymous	Ionosphere/Thermosphere

Anonymous	Near-Sun heliosphere (0.1 - 0.7 AU)
Primarily aimed at early CME detection and propagation characterisation - greatest benefit would be achieved with multi-viewpoint observations that include views looking across the Sun-Earth line (e.g., from L5).
Anonymous	Near-Earth heliosphere (0.7- 1 AU)

Anonymous	Ionosphere/Thermosphere

Anonymous	Corona

Anonymous	Ground-based
The answer to this will depend on what timescale of forecast you are aiming for. Ultimately all of the solar/heliospheric observations in the world won't tell you how geoeffective an event will be as we don't fully understand GIC causation, which needs better ground-based measurements. But longer term forecasting needs upstream measurements. A wicked problem.
Anonymous	Near-Earth heliosphere (0.7- 1 AU)

Mirko Piersanti University of L'Aquila, Italy	Near-Sun heliosphere (0.1 - 0.7 AU)

Anonymous	Near-Sun heliosphere (0.1 - 0.7 AU)
A network of observatories stationed at Venusian lagrange points / ~Venus heliocentric orbit, would be useful
Anonymous	Near-Earth heliosphere (0.7- 1 AU)

Anonymous	Near-Sun heliosphere (0.1 - 0.7 AU)

Anonymous	Near-Earth heliosphere (0.7- 1 AU)
Solar wind
John Haiducek US Naval Research Laboratory	Corona
Better constraints on CME speed, direction, and magnetic fields in the corona would enable estimates of geoeffectiveness days in advance.
Daniel Welling University of Michigan	Near-Sun heliosphere (0.1 - 0.7 AU)

Anonymous	Near-Earth heliosphere (0.7- 1 AU)

Anonymous	Magnetosphere

Anonymous	Corona

Simon Machin Met Office	Near-Sun heliosphere (0.1 - 0.7 AU)

Anonymous	Ground-based

Anonymous	Near-Earth heliosphere (0.7- 1 AU)
L5 mission will be significant augmentation to L1 observations of the sun and in situ solar wind, L4 for SEPs.
Phillip Chamberlin University of Colorado	Corona

Anonymous	Ground-based
From a global point of view, I believe if we could accurately predict solar flares related to sunspots and there likelyhood to produce a flare, everything else would fall into place. That said, I think this is hard to implement. Regionally, the effects of space weather are very strongly tied to the geology (i.e. resistive regions produce the largest GIC), so I would say ground-based from this point of view.
Jingnan Guo USTC	Photosphere

Sean Bruinsma CNES	Near-Sun heliosphere (0.1 - 0.7 AU)
no idea which of the solar observations actually, just a guess
Anonymous	Near-Sun heliosphere (0.1 - 0.7 AU)

Xiukuan Zhao Institute of Geology and Geophysics, Chinese Academy of Sciences	Ionosphere/Thermosphere

Anonymous	Ground-based

Xinan Yue Institute of Geology and Geophysics, Chinese Academy of Sciences	Corona

Anonymous	Near-Earth heliosphere (0.7- 1 AU)

Anonymous	Near-Sun heliosphere (0.1 - 0.7 AU)

Dean Pesnell NASA GSFC	Photosphere
measurements of magnetic fields throughout the solar atmosphere
Anonymous	Near-Earth heliosphere (0.7- 1 AU)

Anonymous	Photosphere

Anonymous
See above comments.
Anonymous	Chromosphere

Anonymous	Photosphere
This answer depends on what aspect of space weather you're talking about. Near-Sun heliosphere on or near the Sun-Earth line would be extremely valuable for very-near-time forecasting, but improving forecasts in the hours- to days-out time range requires full observational coverage of the photosphere.
Anonymous	Near-Earth heliosphere (0.7- 1 AU)
No
Alexander Mishev University of Oulu	Corona
On situ observations are crucial, however the global NM network is essential to provide SW service.
Leon Golub Smithsonian Astrophysical Observatory	Corona
Early warning through evolution in main acceleration phase should help.
Anonymous	Corona

Noe Lugaz University of New Hampshire	Near-Earth heliosphere (0.7- 1 AU)
I am putting Near-Earth heliosphere for the GICs/geomagnetic storm. Radiation would benefit from better measurements of the interior + corona?
Anonymous	Ground-based
Cheapest build-out but poor understanding of interpretation, including due to biases about meaning of data and lack of diversity of data (for example little operational use of direct E field).
Anonymous	Ground-based

Anonymous	Corona

Eelco Doornbos KNMI	Near-Earth heliosphere (0.7- 1 AU)
I think observations from the Near-Earth heliosphere, magnetosphere and ionosphere/thermosphere will be most likely to able to result in actionable information for end-users.
Anonymous	Ground-based
Ground-based observations is the closest to end users and it will help to understand the final results and the effects caused by these results on society.
Michael Liemohn University of Michigan	Ionosphere/Thermosphere
This is a tough question because it depends on which aspect of space weather effect you focus on (SEPs, GICs, TEC/GNSS, satellite charging, satellite drag, ...). I pick GICs of this list, so you need to know the ionospheric current system patterns in high spatial and temporal resolution.
Marianna Korsos University of Catania	Chromosphere
We don't have routine magneticfield observation from the Chromosphere. Why? I don't know but we should have. Therefore we developed the first type of telescope which takes routine magneticfield measurements from the chromosphere. So the chromospheric models can be more accurate.
Anonymous	Near-Earth heliosphere (0.7- 1 AU)

Jean Uwamahoro University of Rwanda	Solar interior

Anonymous	Near-Earth heliosphere (0.7- 1 AU)
Knowing what is about to arrive at the Earth with a lead time greater than 30-90 minutes (L1) would greatly improve the usefulness of geomagnetic storm forecasts.
Anonymous	Ionosphere/Thermosphere

Anonymous	Ground-based

Matthew Lang British Antarctic Survey	Near-Sun heliosphere (0.1 - 0.7 AU)

Anonymous	Near-Sun heliosphere (0.1 - 0.7 AU)

Anonymous	Near-Sun heliosphere (0.1 - 0.7 AU)
Need to have a better understanding of heliospheric magnetic field orientation but, given evolution through the heliosphere, need something before 1AU yet not as close as PSP.
Anonymous	Near-Earth heliosphere (0.7- 1 AU)

Anonymous	Corona

Jonny Rae Northumbria University	Corona
toss up between the corona and near-Sun heliosphere
James Adams University of Alabama in Huntsville	Near-Sun heliosphere (0.1 - 0.7 AU)
My answer assumes the forecast is made when the outburst occurs on the Sun.
Anonymous	Photosphere

Anonymous	Near-Sun heliosphere (0.1 - 0.7 AU)

Lucilla Alfonsi Istituto Nazionale di Geofisica e Vulcanologia (INGV)	Ionosphere/Thermosphere

Anonymous	Corona

Anonymous	Chromosphere
For an efficient forecast we must focus on the chromosphere, the hotbed and drive of SWx.
Ed Thiemann LASP, University of Colorado	Ionosphere/Thermosphere
LEO is becoming an increasingly crowded and challenging region of near space for satellite operators and cost of thermospheric monitoring is relatively low and achievable in the near term.
Anthony Mannucci Jet Propulsion Laboratory, California Institute of Technology	Near-Earth heliosphere (0.7- 1 AU)

Anonymous	Near-Sun heliosphere (0.1 - 0.7 AU)
3D imaging and tracking of propagating CMEs is the obvious next step in forecasting and understanding geomagnetic storms.
Mike Hapgood RAL Space	Near-Sun heliosphere (0.1 - 0.7 AU)
A bit exasperated by a single choice. There are often ways to observe multiple domains with a single satellite payload, most obviously solar surface, corona, inner heliosphere.
Anonymous	Ground-based
You have not included options for Atmospheric throughout this survey. This area is neglected the most.
Anonymous	Magnetosphere

Anonymous	Solar interior

Sean Elvidge University of Birmingham	Near-Sun heliosphere (0.1 - 0.7 AU)
Close between this and thermosphere. Very difficult to map costings of mission to costings saved in forecasting. But likely near-Sun helps more domains (hedging my bet)
Anonymous	Chromosphere

Martin Mlynczak NASA Langley Research Center	Ionosphere/Thermosphere
I checked Ionosphere/Thermosphere because, like regular weather forecasts, the storm must be continually monitored and near-real time data assimilated to forecast its progression 3 to 5 days out.
Allison Jaynes University of Iowa	Near-Sun heliosphere (0.1 - 0.7 AU)

Anonymous	Ground-based

Bernard V Jackson University of California, San Diego	Near-Sun heliosphere (0.1 - 0.7 AU)

Anonymous	Near-Earth heliosphere (0.7- 1 AU)

Yaqi Jin University of Oslo	Ionosphere/Thermosphere

Anonymous	Near-Earth heliosphere (0.7- 1 AU)
CME modelling and arrival prediction needs more than 1 viewpoint. As STA is approaching Earth, we will effectively return to one viewpoint observations. So I mark near-Earth heliosphere, as I feel we need a viewpoint at some angle (60 or 90 degrees). Maybe also below 0.7au would be fine.
Anonymous	Ground-based
It would be crucial to investigate and intensify development of ground-based, remote sensing techniques of the sun and inner heliosphere for operational space weather activities.
Anonymous	Corona
really need more than one of these-at least corona, inner heliosphere, and near-Earth heliosphere
Dario Del Moro University of Rome Tor Vergata	Near-Sun heliosphere (0.1 - 0.7 AU)

Huw Morgan Aberystwyth University	Corona
I'm biased, but improved coronagraph observations crucial. Also need improved HI capabilities.
Mateja Dumbovic Hvar Observatory, University of Zagreb	Near-Sun heliosphere (0.1 - 0.7 AU)
I am possibly biased here. My next choice would be Magnetosphere.
Yuhao Wang Nanchang University	Ionosphere/Thermosphere
No
Anonymous	Near-Sun heliosphere (0.1 - 0.7 AU)

Rajkumar Hajra University of Science and Technology of China	Near-Sun heliosphere (0.1 - 0.7 AU)

Anonymous	Ground-based

Jayachandran P.T. University of New Brunswick	Solar interior
No.
Mike Lockwood University of Reading	Near-Sun heliosphere (0.1 - 0.7 AU)
We need them all. I have chose near-Sun heliosphere because it is poorly observed but the source of solar energetic particles and would help inputs into forecasting models. But I really wanted to tick all the boxes!
Anonymous	Corona

Ciaran Beggan British Geological Survey	Near-Earth heliosphere (0.7- 1 AU)
For forecasting more data at or around 0.7AU would help provide an actionable warning time. For nowcasting you would need to improve LEO and ground-based measurements
Anonymous	Near-Earth heliosphere (0.7- 1 AU)

Anonymous	Ionosphere/Thermosphere

Mark Moldwin University of Michigan	Near-Earth heliosphere (0.7- 1 AU)
Solar drivers drive geomagnetic storms. Increasing warning times and accuracy of observations along Sun-Earth line will help significantly.
Daniel Brandt Michigan Tech Research Institute	Photosphere

David R. Themens University of Birmingham	Near-Sun heliosphere (0.1 - 0.7 AU)
Different if the question is just "what would have the greatest return on investment for space weather mitigation".
Anonymous	Corona

Anonymous	Ground-based
Ground-based observations are inexpensive compared to satellites, yet there are substantial gaps in coverage needed to nowcast geomagnetic disturbance, Total Electron Content, etc and constrain space weather models. Filling these gaps would have an immediate impact on nowcast/forecast capabilities.
Anonymous	Magnetosphere

Anonymous	Chromosphere

Anonymous	Ionosphere/Thermosphere
The current 'new space' activities indicate that understanding the I/T has significant economic impacts. Power grid has to be hardened and prepped well in advance of any disturbance; exact prediction is not required.
Anonymous	Near-Earth heliosphere (0.7- 1 AU)

Anonymous	Near-Earth heliosphere (0.7- 1 AU)

Anonymous	Near-Sun heliosphere (0.1 - 0.7 AU)
Needs include both imaging AND in situ (or other innovative remote sensing techniques that can probe underlying plasma properties in the heliosphere — particularly magnetic field).
Larry C Gardner Utah State University Eastern	Ionosphere/Thermosphere
Everyone is going to say there particular area will have the highest return. The reality is that all of them are important, and significant advancements in all areas is needed.
Piers Jiggens ESA	Near-Earth heliosphere (0.7- 1 AU)
An observer at 0.7 AU would give something concrete upon which to base a forecast and it be actionable.
Bernd Heber Christian-Albrechts-Universität	Corona

Jacob Bortnik UCLA	Photosphere
We need to go back as far to the sun as we can for the longest-range predictions.
Anonymous	Photosphere
Specifically complete 4-pi photospheric magnetic field measurements.
Anonymous	Near-Earth heliosphere (0.7- 1 AU)
L1 monitoring is still managed on an ad-hoc basis rather than operational missions with established and consistent hardware on a managed replacement schedule.
Anonymous	Chromosphere

Question 8

How would you invest around $1 billion to best improve space weather forecasting, either through improved accuracy or improved lead time?

Results

Participant	Response	Confidence
Edmund Henley Met Office (views my own)
Assuming usage of Vigil HI imagery demonstrates value in improved skill for CME arrival forecasts (with lead time trade-off)… then planning for a Vigil follow-on mission, to match terrestrial weather’s meaning of operational – not just a one-off mission. Any spare change to go on polarimetry/similar for guesstimating geoeffectiveness.
Anonymous
Solar wind observation upstream of L1
Ryan McGranaghan NASA Jet Propulsion Laboratory
Start a group (free of existing institutional inertia) that could rapidly manufacture and launch microsatellites to quickly assess perceived gaps in our observational system, develop technology to effectively analyze those observations and pair with digital twins of the Heliosphere and Earth to determine new gaps, and test those. Increase the rate of hypothesis of needed observation to exploring that hypothesis. The range of technology of this group would be large, able to do microsatellites, but also creating and deploying ground-based sensors, too. The group would be extremely agile.
Anonymous

Dr Mario M. Bisi UKRI STFC RAL Space
With only $1B total, you have to carefully target what can be done. Firstly, you need to look at the people resources needed for any future infrastructures as well as for the current datasets. All too often, the nice, big, shiny "thing" gets built with little to now people funding to take advantage of the data set(s) and carry out the improvements to space-weather research or the verifications/validations of data, as well as the feeding of these data into forecast models and suchlike... Build a globally-distributed, multi-/wide-frequency, real-time radio-telescope system capable of observations of interplanetary scintillation (IPS) and Faraday rotation (FR) between around 0.075AU and 0.950AU for continuous inner-heliosphere observational coverage of the outflowing plasma with possibilities of also determining the magnetic-field rotations (along with other key space-weather and plasma parameters). In addition, this will allow for a continuous view of the Earth's ionosphere above and around each of the observing sites, plus, depending on the design of these arrays, capabilities of solar monitoring, solar and ionospheric imaging, and other key radio space-weather data sets (see: LOFAR4SW Use Cases as an example of this and the implications/insights for policy makers - http://lofar4sw.eu/wp/?page_id=1246). This will also enable more blue-skies research on what ground-based radio can do to improve space-weather forecasts (e.g. working through to a desired use of Faraday rotation in the heliosphere and the ionosphere, more IPS data to feed into tomographic and MHD models such as the UCSD tomography - IPS-ENLIL - EUHFORIA, and more-complete solar observations to look for early signatures of potential flaring/active regions and suchlike. Alongside this, more investment in the inner-heliospheric and ionospheric modelling - which would include how to better use data-assimilation techniques from the ground-based radio data (described above). And finally, a smallsat (e.g. NASA SMEX equivalent) to the Sun-Earth L1 (or if budget/telemetry costs dictate, then a suitable Earth orbit that can be Sun-pointing and minimises eclipsing, c.f. NASA SDO), with a coronagraph and heliospheric imaging set of instruments to enable viewing the plasma outflow from ~3.5 solar radio all the way to 180-degrees to view plasma all the way from the Sun to where it's passed over Earth (is in Earth orbit, this full view may need to be restricted depending on the orbit options to try to avoid the Earth being in the field of view of the heliospheric imager given the Earth is so close to the spacecraft instruments at this point. If at L1, then the critical particle and magnetic-field instruments would be needed if still within budget (priority for magnetometer and bulk velocity/density measurements, and then various energetic particles and other heavier ions after that). I could write much more on the rationale and detail of all this, but you quoted 10-15 minutes for this survey, and I've already been here over 30 minutes thinking and writing.
Anonymous

Ian Mann University of Alberta
Additional further upstream monitoring, eg L5. Constellation of research satellites in the magnetosphere plus new modelling and AI efforts for exploitation of data to deliver new knowledge.
Anonymous
Split between (i) funding blue-skies research grants, including HPC infrastructure for the scientific community and PhD studentships (with hope to develop better predictive techniques using existing data); and (ii) contribution to international satellite missions to observe the far-side of the Sun, especially with magnetograms.
Anonymous
(1) 'Space Weather diamond' upstream of L1; (2) 4pi coverage of solar phot. magn. field; (3) distributed measurements between L1-magnetopause
Anonymous

Anonymous
By adding in-situ observations between the Sun and Earth.
Anonymous

Anonymous

Tamas Gombosi University of Michigan
4π Sun observation
Gabor Toth University of Michigan
I don't know what $1B is enough for, but putting ~24 satellites around the Sun at ~0.5au would practically solve the problem.
Harlan Spence University of New Hampshire
Deploy an array of 12 spacecraft in the inner heliosphere, spaced equally in the solar equatorial plane. These spacecraft would each measure the solar wind, IMF, and SEPs. When one spacecraft moved into the sector connecting the Sun with Earth, it would provide a robust indicator of the risks associated with large transients.
Anonymous
$50M in forecast support infrastructure. $100M in research support. $850M to support a mission to address the above-mentioned observation gap.
Robert F. Wimmer-Schweingruber Kiel University, Kiel, Germany
Accuracy. "Cry wolf" will be an important factor for how trustworthy the system is deemed by potential users.
Anonymous
40% for improved accuracy, and 60% for improved lead time
Anonymous

Evangelos Paouris George Mason University
Medium to small satellites with dedicated instruments to cover the Sun as a sphere (4π) and not as a disk from different multiview points. For example, the STEREO mission changed the way of looking at the Sun and enabled the real 3D analysis of eruptive events.
Karim Meziane University of New Brunswick
Support Fundamental Research
Anonymous
Both
Anonymous
Populate the inner heliosphere with numerous cheap small spacecraft to measure the in situ plasma and magnetic field properties.
Vincenzo Romano Istituto Nazionale di Geofisica e Vulcanologia
New observing facilities and coordinating effort among the scientific community and stakeholders
Delores Knipp University of Colorado Boulder
0.5 billion to solar and 0.5 billion to near earth
Anonymous

Anonymous
Improved lead time
Anonymous
improved accuracy
Ankush Bhaskar Space Physics Laboratory, VSSC, India
Will invest in observation capabilities to have lead time o, accuracy could be tuned later.
Anonymous

Anonymous
Space mission with twin L5 and L4 spacecraft, both hosting a vectormagnetogram, inner and outer coronographs, a Sun-Earth line heliospheric imager, and in-situ particle and magnetometer instruments. Would provide 300 degree solar surface magnetic field coverage (from eastern-most limb to western-most limb) where ~260 degrees of that will be to high vectormagnetic certainty, improving inputs to global magnetic field models [accuracy] and monitoring active region complexity and structure/energy build-up for ~1 week sooner than is visible from the Earth perspective [lead time].
Anonymous

Anonymous

Anonymous

Anonymous
I'd ring fence a significant fraction to really instrument the planet with a dense network of magnetic field measurements to improve accuracy. We need to treat ground networks with the same timescale and mission planning that we do spacecraft. At the moment, it's too ad hoc, too hand-to-mouth.
Anonymous
Add more solar wind sensors at L5 and on the back side of Sun.
Mirko Piersanti University of L'Aquila, Italy
Increase the observations in the Near-Sun Heliosphere to improve the capability to understand the physics of the IMF propagation from the Sun down to the Earth. This will increase our capability of IMF dynamics forcasting, and hence of the Geomagnetic storm occurrence.
Anonymous
I would invest equally in both improved accuracy and improved lead time - this is not an either / or problem.
Anonymous

Anonymous

Anonymous
solar wind radar, ground or satellite based thermosphere/ionosphere observation and models
John Haiducek US Naval Research Laboratory
Place a solar-observing spacecraft at L4 or L5, a second spacecraft in a polar orbit around the sun to constrain CME directions, and additional spacecraft in the magnetosphere to provide better data to validate and improve magnetospheric models.
Daniel Welling University of Michigan

Anonymous
Cubesite monitoring network.
Anonymous

Anonymous
moon based network of radiation detectors
Simon Machin Met Office
Invest in a distributed constellation of operational near-sun solar monitoring missions.
Anonymous

Anonymous
Coronal imaging and in situ solar wind measurements from L5, add SEPs measurements from L4 to current planning for SWIFO L1 and GOES coronographs.
Phillip Chamberlin University of Colorado
Helio Mission to L5
Anonymous
$700M - Equipment: Telescopes, satellites with instuments, ground-based equipment, computational capacity $300M - Staff and operations, including analysis of equipment and prediction of forecasts
Jingnan Guo USTC
More observations and more jobs for young and smart people to study data and develop models
Sean Bruinsma CNES
improved accuracy. It is not sufficient presently, so having the same but earlier is not useful
Anonymous
More data
Xiukuan Zhao Institute of Geology and Geophysics, Chinese Academy of Sciences
The most crucial things are to improve the observations and make the data available to the community.
Anonymous
Don't know.
Xinan Yue Institute of Geology and Geophysics, Chinese Academy of Sciences

Anonymous
Strengthening numerical forecasting capability
Anonymous
put more space-based observatory and satellite, orbiting the Sun inside of 1 AU.
Dean Pesnell NASA GSFC
Space-based magnetographs of of the Sun-Earth line, especially viewing the far side of the Sun.
Anonymous

Anonymous

Anonymous
infrastructure for low-TRL (easy, early) method & algorithm evaluation; infrastructure for international standards of facilities needed for improved forecasting (event lists, 4pi forecasting infrastructure & standards); dont just go for "sexy & big-budget", use money to patch some serious holes in the R&D (& transitions & operations) sectors...stuff that just can't get funded, but that is impeding progress.
Anonymous
Fund basic research.
Anonymous

Anonymous
I do not know.
Alexander Mishev University of Oulu
Fleet at L1 point and on situ with new generation of space probes. Sustain the global NM network.
Leon Golub Smithsonian Astrophysical Observatory
L4, L5 and polar observations.
Anonymous
Research into the use of DA in space weather forecasting and what space missions could benefit its use.
Noe Lugaz University of New Hampshire
200M$: 2-4 smallsats with in-situ measurements somewhere ~0.05 - 0.3 AU sunward of Earth (GICs+geomagnetic storms) 800M$: remote-sensing observation platform to monitor active regions (L4+L5 or polar)
Anonymous
Cubesat constellations and more ground small detectors.
Anonymous
Satellite at L5
Anonymous
3D observations of the solar corona and magnetic field imaging.
Eelco Doornbos KNMI
First of all, invest in adoption/forcing of use of open data standards and low latency availability of currently available data (incl. e.g. real-time open SuperMAG, ionosondes, etc). Use remaining budget for a space mission to measure heliospheric conditions on the Earth-Sun line, either in-situ (solar sail?) or remote sensing (The Multiview Observatory for Solar Terrestrial Science (MOST)).
Anonymous
I don’t think these two are isolated. We will need infrastructures that provide both leading time and accuracy based on their driven mechanisms.
Michael Liemohn University of Michigan
I'll pick GIC. To predict GICs, you must understand ionospheric current flows, which means you need to understand FACs and precipitation, so you need to know the dynamics of the outer magnetosphere. You could focus on the magnetotail, with a MagCon mission, but if I had to choose then I have a slight preference for a LEO constellation to directly observe Jperp in the ionosphere.
Marianna Korsos University of Catania
Here is our ground-based telescope which is cost £20 000 (http://hspf.eu/telescope.html). So more can be installed around the world and do a 24 hrs monitoring.
Anonymous
Put a fleet fo magnetometers on CubeSats at distant retrograde orbits around the Earth to measure the CME magnetic field in real time.
Jean Uwamahoro University of Rwanda
Human resources & computation
Anonymous
Chains of cubesats at distances of 0.5-0.8 AU orbiting the Sun - for greater resolution of the solar wind structures and improved lead times. There should be enough satellites to always have one satellite within 10° of the Earth-Sun line. (But that probably costs more than $1B...)
Anonymous

Anonymous

Matthew Lang British Antarctic Survey
I would invest in more observational missions to get observations of the young solar wind. But before doing that, I would perform twin experiments need to be performed to determine optimum locations for future spacecraft missions in order for machine learning and data assimilation techniques to extract the optimum from each observation in a manner similar to how meteorological observing missions are planned. Then invest in improving and incorporating data assimilation and machine learning techniques to improve optimise usage of the new observations.
Anonymous
More space-based measurements to get a better handle on what is heading towards the Earth.
Anonymous
Need to have a better understanding of heliospheric magnetic field orientation but, given evolution through the heliosphere, need something before 1AU yet not as close as PSP.
Anonymous
improved accuracy
Anonymous

Jonny Rae Northumbria University
$600M on a dedicated space weather mission simply to measure flare and CME release and propagation $400M in basic research and research to operations to understand, quantify and predict all of the relevant and necessary requirements for end users You can't just spend that money on a mission, you need to exploit that mission with people, computing, models and ML.
James Adams University of Alabama in Huntsville
To improve observations of the CME and the IPM downstream of the CME.
Anonymous
Increasing observational capabilities and developing physics models of the solar interior
Anonymous

Lucilla Alfonsi Istituto Nazionale di Geofisica e Vulcanologia (INGV)
Increasing the observation coverage from ground-based and satellite assets
Anonymous
Research into underlying physics; constellation of spacecraft monitoring the sun at multiple locations
Anonymous
We need a reliable ground-based forecast network observing routinely the chromosphere that is supplementing satellite data (L1, L5).
Ed Thiemann LASP, University of Colorado
$1B should be invested in widespread monitoring of geospace. Looking to terrestrial weather forecasting as a guide, high quality forecasts were not achievable until continuous, widespread data collection of weather parameters were in place. For example, the GOES satellite program, primarily for terrestrial weather monitoring, is a $1B program, and weather forecasting in the US would be severely hampered. It's a fallacy to think the same is not the case for space weather, where currently now casting is even a challenge due to lack of data.
Anthony Mannucci Jet Propulsion Laboratory, California Institute of Technology
I'd put the funds into improved observations at L1 and L5 to give as much warning time as possible. I would also allocate some of the funds to research to optimize how the observations can be used for forecasting.
Anonymous
I'd spend a few hundred $M on building and operating operational wide-field polarizing imagers in pace near Earth, and another hundred $M on improving data assimilation from those imagers. Recent advances in 3D imaging via polarization mean that near-Earth imaging is likely to be far more cost-effective than going to L4/L5 for stereoscopic measurements. I'd spend the rest on a demonstration solar-sail mission inside L1 to provide improved ground-truth warning time of geomagnetic effects, via in-situ sampling.
Mike Hapgood RAL Space
Expand the amount of the Sun's surface and inner corona that is monitored 24/7, e.g. using Vigil as a pathfinder for more extensive network of satellites, eventually covering all 4 pi. Encourage recognition that this network is also needed to forecast space weather for human missions to Mars.
Anonymous
I wouldn't. I would focus on now-casting and public communication and awareness
Anonymous
35% first principles modeling, 35% data-ingestion/assimilation, 30% observational data quality.
Anonymous

Sean Elvidge University of Birmingham
1 near-Sun mission (measurement of B). And funded, secondary/tertiary payloads for upper atmosphere observations on all megaconstellation satellites (e.g. Starlink, OneWeb).
Anonymous
Dedicated HPC infrastructure for running physics-based model chains (like the GCM for meteorology at Earth).
Martin Mlynczak NASA Langley Research Center
I would invest for improved lead time with improved accuracy. It is not an a one or the other question. Lead time is irrelevant if there is no accuracy.
Allison Jaynes University of Iowa
More distributed sensor arrays. Sensors on geosynch, medium-Earth, and low-Earth orbits as hosted payloads on commercial s/c that measure magnetic field and crude relativistic particle fluxes (at a minimum). This could be done relatively cheaply, similar to REACH or AMPERE. More sensors and solar observing out at L1, L5, and at the near-Sun region to better understand the generation and propagation of these eruptive events.
Anonymous

Bernard V Jackson University of California, San Diego
A plan to observe heliospheric magnetic fields would be a good idea. Good instruments to provide CME tracking through the heliosphere would be great to have.
Anonymous
launch a spacecraft located ahead of the Earth
Yaqi Jin University of Oslo
I would invest both, one for long lead time and less accurate, one for high accuracy with adequate lead time for specific users.
Anonymous
1. investigate what exactly users would need from the science/operational side. This is crucial. 2. Get a spacecraft, f.e. L5 (provide info on both solar wind that is coming as well extra viewpoint for CMEs). 3. Improve modeling. 4. Verification and validation of our current prediction capabilities. (not enough time is spend on this effort).
Anonymous
More interplanetary scintillations measurements, more automatic analysis tools of already existing measurements, L5-mission with plasma and magnetic field in-situ monitoring and consolidate operations and expansion of ground based sensors of ionosphereic phenomena.
Anonymous
start and maintain a 'string of pearls' space weather network around the Sun that has spacecraft with both imaging and in-situ measurements (including a magnetograph) -plus do a one-off solar polar mission with same. Could leverage missions like SWFO operational system.
Dario Del Moro University of Rome Tor Vergata
Magnetometers on-board many deep-space cubesats at different positions in the inner heliosphere
Huw Morgan Aberystwyth University
0.5b: Two dedicated coronagraph, EUV imaging, & HI smallsat missions, one at L1 the other at L5. 0.1b: Dedicated 10-year funding ringfenced for space weather operational tools. 0.1b: Dedicated 10-year funding ringfenced for space weather research. 0.1b: Scheme to improve preparedness of industry/government. 0.2b: Aid for developing countries to help their preparedness
Mateja Dumbovic Hvar Observatory, University of Zagreb
I guess both are equally important, 50:50
Yuhao Wang Nanchang University
improve space weather forecasting accuracy.
Anonymous
Permanent satellite network between Sun and Earth.
Rajkumar Hajra University of Science and Technology of China

Anonymous
Magnetosphere-ionosphere coupling and ground conductivity
Jayachandran P.T. University of New Brunswick
Build a global network of observational capabilities. Invest in moving away from "classical" observations.
Mike Lockwood University of Reading
Developing, testing and implementing targeted new observation techniques and developing and testing new numerical models
Anonymous
50% in new observatories, 25% in R2O programs, 25% in supporting infrastructure.
Ciaran Beggan British Geological Survey
A network of STEREO-like satellites orbiting the Sun at <1AU to sample the solar wind and take imagery. Not sure how much that would cost
Anonymous
L5 in-situ and remote sensing satellite
Anonymous

Mark Moldwin University of Michigan
constellation of at and inside L1 solar wind monitors.
Daniel Brandt Michigan Tech Research Institute
I contend that the funds ought to be spent on a combination of (1) spaceborne observatories that can observe the 'rear' of the Sun and also the IMF between the Sun and Earth to study CME/CIR evolution, and (2) a constellation of satellites at multiple distances in the magnetosphere out to the magnetopause that can measure B and particle fluxes/distributions/velocities.
David R. Themens University of Birmingham

Anonymous
Honestly I'd spend it trying to build resilience in the systems (e.g., backstock of transformers ready to be swapped in). But to the actual question... I'd divvy up the $1B according to the same investment piechart that terrestrial weather forecasting has. The book "The Weather Machine: A Journey Inside the Forecast" by Andrew Blum does an excellent job describing all of it in detail.
Anonymous
Expanded, global, multi-instrument ground-based networks that provide real-time measurements L5 monitor, in addition to L1 monitor *Fundamental and applied space weather research with stronger emphasis on benchmarks, uncertainty quantification, and model-model/data-model comparisons
Anonymous

Anonymous
improved lead time
Anonymous
There are significant problems that may not be amenable to prediction due to the chaotic nature of the phenomena. For $1B in the short term we need to understand near-Earth geospace to insure safe and efficient operations.
Anonymous
Would need to think some more about this!
Anonymous
Sub-L1 small-sat particle/mag mission utilizing solar sails for station-keeping.
Anonymous
Multi-viewpoint constellation to provide 3D imaging from Sun through heliosphere, especially techniques to probe and reconstruct solar and heliospheric magnetic fields.
Larry C Gardner Utah State University Eastern
Split is 50/50 between observation and modeling. We have a long way to go to get to even a basic forecast. We need observation to understand physics, and models to with this physics to provide a forecast.
Piers Jiggens ESA
Accuracy I think many end users would choose lead time but I would choose accuracy simply because we are not reliable enough to be listened to at present.
Bernd Heber Christian-Albrechts-Universität

Jacob Bortnik UCLA
combined observations of photosphere and L1 monitoring. Also investment in numerical modeling!
Anonymous
Create a fleet of solar monitoring spacecraft to get full 4-pi coverage of the photospheric magnetic field.
Anonymous
Operational Van Allen Probes style mission with substantial data assimilation efforts.
Anonymous
Give monetary awards for improved understanding, especially observational advances. We are still learning.

Question 9

Machine learning and neural network approaches to space-weather forecasting are showing a great deal of promise. In the next decade, to what degree can such approaches replace or complement physics-based models for forecasting extreme space weather with a 1-day lead time?

Results

Participant	Response	Confidence
Edmund Henley Met Office (views my own)	A little
By their nature, extreme space weather events won’t have appeared much in past observational data. I’d assume this will hinder any machine learning approaches based on observational data – hard to extrapolate. May help complement physics-based models: e.g. as emulators, allowing larger ensembles, helping better estimate forecast uncertainty
Anonymous	A little

Ryan McGranaghan NASA Jet Propulsion Laboratory	A lot

Anonymous	A little

Dr Mario M. Bisi UKRI STFC RAL Space	A little
Maybe a little for complement - probably not to replace - my feeling is these advances would have more use to complement every-day and moderately-severe space-weather events rather than the extreme events.
Anonymous	A little

Ian Mann University of Alberta	A lot

Anonymous	A lot

Anonymous	A lot
physics-informed ML/AI models could speed up computations and/or allow multi-scale simulations thereby increasing accuracy of the driver impact on the magnetospehre
Anonymous	A lot
I am not sure machine learning and neural network approaches will (or should) replace physics-based models in the next decade, but I think they will successfully complement them
Anonymous	A little
Inherent unpredictability of the physical system is obviously a big challenge to any approach.
Anonymous	A lot

Anonymous	A lot

Tamas Gombosi University of Michigan	A little

Gabor Toth University of Michigan	A little
This is a tough question. Typical space weather events will be very well predictable with machine learning. Starting from L1 one can predict everything with machine learning as well or better than with a physics based model. Starting from the Sun is much more challenging both for ML and physics-based models. Extreme events should favor physics-based models in general, but it does not mean they will work well.
Harlan Spence University of New Hampshire	A little
Actually not too sure about this - every time I think that machine learning will provide limited new insights, I am pleasantly surprised. Perhaps there are signals present in our limiting observations that machine learning could tease out to improve predictions a lot - I wouldn't bank on it, but it is certainly worth investing effort in this area.
Anonymous	A little
Insufficient data exists at this time to really enable high-performance ML forecasting of extremes. ML is currently over-used and under-delivering in the space weather domain.
Robert F. Wimmer-Schweingruber Kiel University, Kiel, Germany	A lot
Sure this can help, but it would be better if we UNDERSTOOD what is going on.
Anonymous	A lot
It's already evident that machine learning and neural network approaches are a lot more used these days than physics-based models. The rapidly growing database and explosion in number of equipment taking measurements of the space weather will definitely call for more machine learning approaches
Anonymous	A lot

Evangelos Paouris George Mason University	Completely
The ML should be a supplement to Physics. Not the opposite. The "black boxes" which appeared recently are not very helpful to understand the physics and mechanisms taking place. For forecasting purposes with a 1-day lead time, the ML has great potential to achieve this goal. Even now there are many models (without any ML) with a 1-day lead time with good performance.
Karim Meziane University of New Brunswick	A little

Anonymous	A lot

Anonymous	A little

Vincenzo Romano Istituto Nazionale di Geofisica e Vulcanologia	A lot

Delores Knipp University of Colorado Boulder	A little
We are hampered in ML by having so few events that are extreme
Anonymous	Not at all

Anonymous	A lot

Anonymous	A lot

Ankush Bhaskar Space Physics Laboratory, VSSC, India	A lot

Anonymous	A little

Anonymous	A lot
Continued improvements can be made by using ML and/or NN approaches to complement physics-based models. Care should be taken to not develop "black-box" systems, but rather design ML implementations that allow for human scientists/operators to learn about what the system is making its decisions based upon.
Anonymous	A lot

Anonymous	Not at all

Anonymous	Not at all

Anonymous	A little
I find it hard to say - there is potential but I'm not convinced.
Anonymous	A lot

Mirko Piersanti University of L'Aquila, Italy	A lot
Machine learning will never let us understand the physics behind solar wind-magnetosphere-ionosphere coupling.
Anonymous	A little
Data assimilation models and observations at ~0.7 AU will likely be of greater benefit.
Anonymous	A lot

Anonymous	A little
AI needs to be trained on large statistics of events, which is impossible since extreme events are rare by definition.
Anonymous	A lot

John Haiducek US Naval Research Laboratory	A little
In my understanding, machine learning models tend to deliver an interpolation of their training data. Physics-based models will likely be required for extreme events and may outperform ML in other respects as well. The comparatively low computational cost of ML will make it attractive for many applications.
Daniel Welling University of Michigan	A lot

Anonymous	A little

Anonymous	A lot

Anonymous	A lot

Simon Machin Met Office	A little
I would favour something between "a little" and "a lot". More work needs to be done in these areas, but I think this could improve efficiency and consistency, although will not likely provide a "perfect" solution, given the chaotic nature of activity and phenomena.
Anonymous	A little

Anonymous	A little
Need to be pursued in parallel with preceding recommended measurements and physics based models.
Phillip Chamberlin University of Colorado	A lot

Anonymous	A little
Machine learning algorithms are usefuls at predictions, provided enough data is present, and no extreme event never seen before occurs. Machine learning would be more useful, however we simply don't have enough data, at a high cadence for machine learning to make a huge impact. Physics based models will always better consider these extremes
Jingnan Guo USTC	A little

Sean Bruinsma CNES	A lot
ML is just an intermediate solution, we must not stop working on the physics-based models. With time, say in a decade, physics-based plus DA is the way forward in my opinion
Anonymous	Completely
In the next decade we will smile (if not laugh...) at the thought that once upon a time somebody was trying to make predictions with physics-based models...
Xiukuan Zhao Institute of Geology and Geophysics, Chinese Academy of Sciences	A lot

Anonymous
Don't know about #20.
Xinan Yue Institute of Geology and Geophysics, Chinese Academy of Sciences	A little

Anonymous	A lot

Anonymous	A lot

Dean Pesnell NASA GSFC	A little

Anonymous	A little

Anonymous

Anonymous	A little
This survey was a bit tough to complete, it was too general and simplistic although I appreciate what it's trying to do.
Anonymous	A lot

Anonymous	A little

Anonymous	A lot
No
Alexander Mishev University of Oulu	A little

Leon Golub Smithsonian Astrophysical Observatory	A lot

Anonymous	A lot

Noe Lugaz University of New Hampshire	A little
Issue with ML is that for in-situ measurements we are still very data poor. Probably more opportunity for flaring/SEPs to replace empirical models.
Anonymous	A little

Anonymous	A lot

Anonymous	A little

Eelco Doornbos KNMI	A little
Machine learning models are similarly hampered by scarce observations as physics-based models. I'm also sceptical of machine learning models being able to provide reliable results for extreme events, which by their definition, are not part of their training observation data.
Anonymous	Not at all
If we can’t feed the AI with sufficient data, it would not be helpful.
Michael Liemohn University of Michigan	A lot
ML tools can help us find the relationships that we then explore, refine, and explain with physics models. For prediction, ML tools can help a lot.
Marianna Korsos University of Catania	A little
We still do not have full knowledge about the physics of how solar eruption occurs. We need farside data from the sun. We have more negative cases the positive so AI cannot really learn to music........ We need physics-based models........
Anonymous	A little

Jean Uwamahoro University of Rwanda	A lot

Anonymous	A lot
I think machine learning approaches can often replace existing models, but it's unlikely they'll lead to important breakthroughs that will really push our understanding and forecasting abilities forward.
Anonymous	A little

Anonymous	A little

Matthew Lang British Antarctic Survey	A little
The physics still needs to be considered to ensure realistic scenarios are generated from ML-models. Machine-learning will be a useful tool but until it can be properly interrogated, caution must be applied.
Anonymous	A lot
I don't think AI could (or should) ever replace pyhsics-based models but it can definitely help improve forecasts if used well
Anonymous	A little

Anonymous	Completely

Anonymous	A lot

Jonny Rae Northumbria University	A lot
It is possible that ML may replace it completely, but only when we've understood what questions we need to ask
James Adams University of Alabama in Huntsville	A little
Machine Learning (ML) is effective when used with a large and well curated database. The solar weather database is too small to support the development of accurate ML forecasting tools.
Anonymous	A lot

Anonymous	A little

Lucilla Alfonsi Istituto Nazionale di Geofisica e Vulcanologia (INGV)	A lot

Anonymous	A little

Anonymous	A little

Ed Thiemann LASP, University of Colorado	A little

Anthony Mannucci Jet Propulsion Laboratory, California Institute of Technology	A lot

Anonymous	A lot
AI networks are "fraught with peril" from a research standpoint, but they offer much more computationally effective forecasts -- enough to provide large ensemble models. With rigorous validation they are very useful tool that can supplement physics-based modeling.
Mike Hapgood RAL Space	A little
I would stress the importance of physics and ML as complementary approaches. The big issue is whether the underlying physics changes as we move to higher levels of space weather.
Anonymous	Not at all
Not enough data sets available to make this approach viable
Anonymous	A lot

Anonymous	A lot

Sean Elvidge University of Birmingham	Not at all
I think they can complement physics-based models for non-extreme spwx. But likely we have no data to train these models on extreme spwx events.
Anonymous	A lot
ML and NN will certainly improve forecasts but will not increase are understanding or insights at all. They provide black boxes. It works, but nobody understands why.
Martin Mlynczak NASA Langley Research Center	A little
This will be completely dependent on the ability to have accurate physics-based models with which to train the networks.
Allison Jaynes University of Iowa	A little

Anonymous	A little

Bernard V Jackson University of California, San Diego	A little
There really need to be better observations.
Anonymous	A lot

Yaqi Jin University of Oslo	A lot
Machine learning will improve a lot on the forecasting capability. But as long as we have curiosity, physics is needed.
Anonymous	A lot

Anonymous	A little

Anonymous	A little
would need to design a 'smart' AI experiment with well-considered inputs and questions- a good challenge for the community
Dario Del Moro University of Rome Tor Vergata	A lot

Huw Morgan Aberystwyth University	A lot

Mateja Dumbovic Hvar Observatory, University of Zagreb	A little
As noted earlier, I find that observational limits and inherent unpredictability of the system are the largest forecast issues and will always remain an issue. They equally affect all forecast methods and machine learning is not exempt
Yuhao Wang Nanchang University	Completely
No.
Anonymous	A little
I'm not in favour of black-box statistics. If anything, it shows the long way the physics-based models have to go.
Rajkumar Hajra University of Science and Technology of China	A lot

Anonymous	Not at all

Jayachandran P.T. University of New Brunswick	Not at all
None.
Mike Lockwood University of Reading	A lot
Automated systems will always need refining and testing.
Anonymous	A lot

Ciaran Beggan British Geological Survey	A little
Unless there is a major breakthrough on the data collection side, ML will not solve the issue of lack of data. Like terrestrial weather forecasts you need a lot of data to do a decent job
Anonymous	A lot

Anonymous	Not at all

Mark Moldwin University of Michigan	A lot
ML in combination with physics-based models AND higher "density" observations (big data) will make big difference in understanding missing understanding.
Daniel Brandt Michigan Tech Research Institute	A little

David R. Themens University of Birmingham	A little
It may help, but it's not going to get you 1-day lead time for SEPs or flares, at least not with current observations.
Anonymous	A lot

Anonymous	A little
~8 years ago, I recall hearing predictions that machine learning/neural networks would replace physics-based models in <10 years. This doesn't seem to be happening, and several challenges with ML-techniques remain. That said, ML has led to powerful new tools in some areas and this should continue.
Anonymous	A little

Anonymous	A lot
ML is a magic word for funding. Since money is needed to do research ML will be needed. However, I am not sure that the scientific paradigm shift will allow an increase in our prediction capabilities given the complex nature of the physical processes related to SW.
Anonymous	A lot
The assumption that 'space weather' = catastrophic events is incorrect and human exploration of space was not considered in these questions.
Anonymous	A little

Anonymous	A lot
Machine learning is not a panacea, but shows promise to improve forecasts over physics-based models.
Anonymous	A little

Larry C Gardner Utah State University Eastern	A little
Until we expand our understanding of all of the systems in question, better methods to look at the data can still only look at the data currently available. Methods such as machine learning and neural networks could help us understand observations, but improving forecast will only occur when we have significant improvements in understanding.
Piers Jiggens ESA	A little
We lack the amount of data (in general) needed for ML to replace physics-based models. The main challenge is to model the extremes for which we lack data sets and ML is not good at dealing with that as far as I know.
Bernd Heber Christian-Albrechts-Universität	A little

Jacob Bortnik UCLA	Completely
It's important to have physics-based models to develop understanding but nothing beats AI/ML for prediction
Anonymous	A little
We are still in a data sparse environment when it comes to SWx predictions... therefore the ML increase in knowledge will be fundamentally limited.
Anonymous	A little
Applicability of DL techniques to forecasting is severely limited by the data-poorness of the domain ("large" data sets are miniscule compared to those used by commercial DL users), difficulties in feature definition, and the problems of spurious correlations/"hallucinations".
Anonymous	Not at all

Space Weather July 25, 2023

Space Weather

Summary

Question 1

Results

See individual participant responses

Question 2

Results

See individual participant responses

Question 3

Results

See individual participant responses

Question 4

Results

See individual participant responses

Question 5

Results

See individual participant responses

Question 6

Results

See individual participant responses

Question 7

Results

See individual participant responses

Question 8

Results

See individual participant responses

Question 9

Results

See individual participant responses