WOVO/WOVO-dat meeting
Sunday, July 21 2013, 7.00am to 5pm
Kagoshima, Japan

MINUTES

Present: Andrew Tupper (Bureau of Meteorology, Australia); Jiandong Xu (Changbaishan, CEA); Jennifer Jay, Matt Pritchard (Cornell University); Christelle Wauthier (Carnegie Institution of Washington); Yo Fukushime (DPRI,Kyoto University); Fidel Costa, Christina Widiwijayanti, (EOS); Thomas Walter, Jacqueline Salzer, Luigi Passarelli, Mehdi Nikkhoo, Nicole Richter (GFZ, Potsdam, Germany); Shinji Takarada, (GSJ, AIST); Supriyati Andreastuti (CVGHM, Indonesia); Martin Hensch (Icelandic Meteorological Office); Selva Jacopo, Tierz Pablo, (INGV - Bologna); Alessandro Bonforte (INGV-Osservatorio ETNEO Catania, Italy); Joan Marti, Rosa Sobradelo (Institute of Earth Sciences, CSIC, Barcelona, Spain); Haiguan Wei (Institute of Geology China Earthquake Administration); Patricia Mothes (Instituto Geofisico, Escuela Palitecnica Wacional Quito, Ecuada); Eisuke Fijita, Hideki Ueda (NIED); Antonius Ratdomopurbo, Nang Thin Zar Win (NTU.Singapore); Renato Solidum, Mariton Bornas (PHIVOLCS-Dost); Steve Saunders (Rabaul Volcano Observatory); Richard Robertson, Robert Watts (Seismic Research Centre, University of West Indies); Luis E. Lara (Sernageomin); Ben Andrews (Smithsonian G.V.P); Masato Iguchi (SVO.DPRI.Kyoto University); Juliet Biggs, Sarah K. Brown, Steve Sparks,(University of Bristol); Magnus T. Gudmundsson, Vincent Drouin (University of Iceland); Falk Amelung (University of Miami); Roberto Carniel (University of Udine, Italy); Valerio Acocella (University Roma Tre, Italy); Christine Neal (USGS, Alaska Volcano Observatory Anchorage AK-USA); Tamar Elias, (USGS, Hawaiian Volcano Observatory); Angie Diefenback, John Pallister (USGS, VDAP); Emile Jansons (VAAC Darwin); Chris Newhall.

Meeting chaired by: Gill Jolly (WOVO co-chair and GNS Science, New Zealand)

1. Introduction:
Gill welcomed all participants to the meeting.

2. Volcano observatories and ICAO responsibilities
Gill explained that new ICAO guidelines that had resulted from the International Volcanic Ash Task Force (following the Eyjafjallajokull eruption) requested that states with volcanoes should monitor them and that there should be a strong driver now to have partial cost recovery from airlines to meteorological agencies and volcano observatories. Some VOs were uncertain about how to start on the pathway on how to get cost-recovery.
Chris Newhall asked whether most national met agencies have cost recovery but Andrew Tupper was not sure although WMO was a strong advocate.

ACTION: Gill to email Andy (as WMO representative) to get guidance on how VO should approach airlines for cost recovery.

Chris Newhall (on behalf of WOVO) had also put forward project to the International Air Transport Association (IATA). CN had met with IATA who are concerned about quality and timeliness of information from VO's to aviation. There are notable shining stars, but globally there are a number of weak spots. IATA were interested in a gap analysis - where were the weak links in communication chain from VO's to VAAC to cockpit, such as volcanoes without monitoring, others with monitoring? Information doesn't flow efficiently from volcanoes to VO to VAAC - language problems. IATA suggested submitting a proposal to identify the gaps in the first instance, with possible later funding to plug the gaps. A proposal was submitted but IATA wouldn't fund the requested amount and asked for a smaller proposal with a partner organization. It was possible that ICAO would be the partner; details are still to be worked out. Chris will go back to IATA with a smaller proposal. Any gap study would be face-to-face.

ACTION: Chris to investigate partnership with ICAO and re-submit proposal to IATA.

3. Global Volcano Model
Steve Sparks explained GVM (as a UK NERC funded project) and that for a part of GVM, a report was being undertaken for UNISDR (GAR15), which included a gap analysis relating to volcano monitoring capability and capacity around the world. Steve sees GAR15 as very influential and this might be able to generate some funding. The GAR15 report is a first step, and will be relatively short, backed up by more details. This could also feed into the IATA proposal. It was a tight time frame, and so had to be based largely on existing information. Evidence will be collected regionally through WOVO re: volcanoes and volcanic risks in each country. GVM has the resources to collect this, but they want to work personally with observatories. GVM will work with Gill over next few months to do this.
Rene Solidum thanked GVM/WOVO for working with observatories. Their past experience with GAR was that observatories were left out and the information was accordingly not what PHIVOLCS would have contributed.
Steve said that this is an opportunity for VOs to demonstrate that we are a community and to set a benchmark for future work. He has at least 2 people to work on this including Sarah Brown.

ACTION: Gill will send out weblinks to previous GAR.

4. Volcano Observatory Best Practice (VOBP) workshop, November 2013
Gill talked about the first VOBP in Sept 2011. The topic was eruption forecasting. Many VO representatives were present and there was excellent discussion. The next one would be focussing on communicating hazard and risk and was to be held on 2 - 6 November in Erice, Sicily. The same general approach was to be used and an announcement and invitations are coming shortly. It is supported by GVM, VUELCO, INGV and others, especially for developing countries.
How many VO's are there? CN said that there about 80 listed in WOVO directory, although there are different ways of counting.
Gill said we are encouraging young scientists to come too. There are a limited number of places.
A summary of the first VOBP workshop was almost finalized.

5. WOVO website
This hadn't been updated for a long time but is being tidied up currently.

ACTION: Let Gill know if contact information is OK or needs changing.

6. Volcano deformation database
Matt Pritchard talked about the new volcano deformation database project with Juliet Biggs. Three collaborations were outlined: working closely with WOVOdat, with GVP (Smithsonian), and for GAR15.
Matt's project will put raw interferograms and raw time series into WOVOdat. There has been an "explosion" in the amount of available data. In 1997, 44 deforming volcanoes known. We now know of >200, and the number will increase with new ALOS2 and Sentinel satellites.
What does deformation means re: behaviour of volcanoes? How do we process these data and get results to observatories before going public?
With GVP, the project will add derived, summary information (such as yes/no a volcano is deforming).
Matt is looking for volunteers to work on the project. Others so far included are: Falk Amelung, Paul Lundgren, Fukushima-san, Andy Cooper, and more. There's already a template for the project. It will take years to figure out how to automate the data processing and get useful products to those who need them.

A best practices document will be developed on remote sensing for volcano observatories.
The intention is to include INSAR, tilt, EDM, levelling, GPS. All of these data are going into WOVOdat (long-term goal). A thousand interferograms aren't useful; time series are. The go to place for data should still be WOVOdat.
HVO is already doing this; Matt hoping to learn from them, and from AVO.

CN stated that there is sensitivity at VO's about feeding information in realtime. WOVOdat delighted to receive in realtime but we would feed first to VO's. However, Matt said that satellite data are available to all and it is difficult to stop those who are processing and releasing in real time, e.g. MODVOLC.

Juliet Biggs said that ESA Sentinel will be launched next year; this is the first radarsat designed for monitoring, not science. It will pass over every point on the earth's surface every 6 days, hence yielding a huge volume of data, so there will need to be automation. The distribution of the data are not clear yet, possibly ESA to UK Space Agency. Ideally, everything will be free. Juliet needs input on what is needed: does WOVOdat want data from all 1500 Holocene volcanoes or just a subset? How to search these data? How about raw data vs derived parameters, e.g., deformation rate, LOS, etc? Is interpretation needed? Comments are welcomed from VOs.

John Pallister said that the raw data are also valuable (not just interferograms), e.g., for tracking growth of a lava dome.

Falk Amelung said there was also an opportunity for volcanoes to be assigned as a "supersite" to enable (?) free image capture. Also in October a new satellite TerraSAR-X will start. This would work with growing lava domes. This will also have high-res DEM coming soon (order of magnitude better than SRTM).

7. WOVOdat
Purbo presented the latest news on WOVOdat, which was now in year 4. The main points are:
1. The main WOVOdat online system is now ready and accepting data, with the data policy as discussed in COV7.

2. A WOVOdat standalone package is available. Some observatories already have their own, e.g., DIVO, but others still need a system. The WOVOdat basic principle is to use open source software, MySQL, so it's freely downloadable. Thanks to PHIVOLCS, NIED, Indonesia, NZ for early data and test runs. Jacopo Selva, Ueda-san were pioneers of WOVOdat.

3. Data population will be a continuous, ongoing project. WOVOdat generally deals only with 2-yr old data. WOVOdat will not publish the realtime data unless requested by the VO to do so, so as to avoid "armchair predictions".

Data formats vary from one VO to another. WOVOdat will work with you to get data in the right format and the uploading system is ready. The name of the owner of the data/ observatory is always tagged on the data. WOVOdat isn't the owner. WOVOdat deals only with processed data (e.g., catalogue).

Purbo demonstrated the database schema. Volcano information is included, but mainly the data are on unrest. Every type of data is different, e.g., deformation, gas, hydrology. There is also a bibliography.

Chris said that there has been a long gestation period for WOVOdat and it is by no means fully populated. A still relatively small % of volcanoes and data are in WOVOdat, but it is important to open it up now so that you can see what's in there, what are the possibilities. The hope is that this will be a snowball effect so people will want to add their own data, but it won't happen overnight. It is risky opening it up now, even though it is not 100% functional yet. There will be a caveat on the front page that this is still in an early stage, so if you run a search now, your results will be very small and dangerous to interpret.

Questions that followed from the audience:

Is the link to GVP dynamic and constantly updated? Updates will be periodic. It is still tricky to define eruptions and unrest?
Ben Andrews: Volcanoes of the World (VOTW) is constantly updated internally but not refreshed online continuously, perhaps once or twice a year.

How searchable is WOVOdat?
The general goal of WOVOdat is to allow ready retrieval of common information that we compare across volcanoes. We don't seek to be the source of data about any one volcano. We seek subset of data that makes for useful comparisons. There will be Boolean search capability for X,Y,Z, tied to =VOTW - will say whether it erupted, etc. Our emphasis is on things that can be searched in times of crisis or to get a research project started. To dig deeper, researchers will need to go back to the VO.

Pallister: Can you pose question of intrusion volume right now? Or, for example, Mayon in last 20 years with eq's in M range of x-y, etc?
Purbo: For DVT's we have already from IRIS but we want to add the historical data from observatories for the smaller events.
Newhall: The Boolean search capability is built into MySQL and we will shortly have some standard questions like yours to simplify the query.

Can users contribute interpretations, e.g., VLP's?
WOVOdat will simply present the data, without interpretation of the meaning of the VLP's, but the combination of other data really helps on this.
Ueda-san said that there is a server mirror in Tsukuba and synchronising will be done by the end of this year. We also want to collaborate with those who will contribute other user tools.

Is there a plan for standard tools for analysing data and interpretations?
We need to build and populate the database first, but let's discuss the future within the WOVOdat steering committee. There is one table within WOVOdat of published interpretations, e.g., dike intrusion, tied to a reference. WOVOdat won't make new interpretations, but we could make more tools available to the community, e.g., for standardization among observatories.

What sort of seismic data - is it only one station per volcano or multiple stations?
Where There's a network, we show just that, not single station counts within it. Where just one station, then counts only.

Are tremor data included?
Yes and all data has time-date-georeference. For tremor, would have start stop time, amplitude and spectra.

Is there any data about eruptions?
No, only information from GVP.

Newhall extended great thanks to Purbo, Christina and Nang, because over the last 4 years there has been very significant progress. Also to Jacopo, Ueda, Florian who helped in early pilot testing, advice, and to Dina Venezky, USGS.

8. WOVOdat data policy (discussion moderated by Newhall)

The data policy is important for building the trust of contributors and guiding those who will use it (properly).

1. Data owned by those who collected the data, and contributed it. - OK.

2. Most data are processed, reduced. Those seeking raw data, e.g., arrival times, raw GPS data, should go to the originator. If a VO wanted to provide raw data, they can't do that yet.

3. Latency period. This was first discussed 2000 in IAVCEI Bali, then Menlo Park and many times since! VO directors, in general, were not comfortable with putting data out in realtime for two reasons: (a) there was a worry about being second guessed by external people and (b) the scientists who collect the data are swamped and yet they want to be the ones to look first and publish on the data they collected.
Where's the right balance? For most uses of WOVOdat, it's ok just to have 2 year old data. There might be a few cases to speed it up (see item 4). Now in 2013, there is more movement toward more open data with shorter latencies, but we still need to provide an option for those reluctant.

4. Proposing - WOVOdat accept 3 categories of data (chosen by contributor):

a. Freely available when contributed
b. Freely available 2 years after collected. We don't fund data collection so we cannot require the contribution, but that's the target.
c. (New within the past year). Available after 2 years for searches and visualization, for crises, but not available for download.

Some observatories are unwilling to contribute unless in category c. It is hoped there won't be many like this. If we accept in category c, VOs will be asked to convert to category b after a specified time period.For WOVOdat, we want to automate as much as possible.

For WOVOdat, we want to automate as much as possible. WOVOdat staff will go to VO's, work with them on scripts, or invite them to Singapore.

Will there be a need to register for people to download data?
To gain access, people will need to register. Access, or at least downloads, will be tracked and reported back automatically to the data contributor.

1. Good idea, but what if the user doesn't honor this? WOVOdat can cut off access to people who do not honor the terms of use.
There should be reference to the IAVCEI protocols for crisis. The data owner might want to use it for forecasting. This need to be clarified in the wording.

2. What if someone downloads a timeseries and then makes a different analysis than those made previously and publishes in a scientific journal? If it were a time series on a volcano, it would surely be the courteous thing to do to contact the observatory, share new ideas with them, e.g., would you like to work with me on it, or at least clarify your dataset.

3. The main purpose of WOVOdat is for comparisons between unrest periods and searches, not volcano specific research.

4. WOVOdat won't be looking through the literature for misuse of WOVOdat. One would hope that since all agree to the data policy before logging in, but we all know how easy it is just to click "yes" to EULA. We need to trust that most will abide.
Will there be exceptions?
Sure, but the value of making data available outweighs the issue of misuse, although this is no consolation for the aggrieved.

5. In NZ all data (including waveforms) are freely available and the VO has established itself as the source of information and the authority on volcanic activity (also legislated), so this helps to head off concerns about misuse of data.

6. Restricted may need to have a flag.

7. Tracking of users.
Could abuse be shut down?
In reality, it's hard to shut someone down - they may just login with a different user name.

8. No comments.

9. No comments.

10. This is legal language to protect WOVOdat from lawsuits from others.
Could data be on-sold to, e.g., insurance companies, or in value-added products, e.g., for sale to airlines. How will you deal with this?
This hasn't been considered.

For WOVOdat to sell a package of data of monitoring data on, e.g., Iceland would be clear violation of intent.

Insurance and reinsurance is a big market and data might be fed into a proprietary model (RMS like) for the client.
Should WOVOdat charge for this?
We can't sell access to something that's already free. There needs to be more discussion on the commercial question.

9. Future of WOVOdat.

WOVOdat has made good progress over the past few years through funding from EOS, but there is now a transition this coming year. We need to have a more distributed database and distributed funding. (Airplane taxied on runway for 20 years, just took off, and last thing we want to do is to crash!)

Realities: Purbo was on secondment from Geological Agency of Indonesia and must return back by end of year. He may or may not be allowed to continue work on WOVOdat. Chris will be returning to Philippines by end of year although he'll have more time for WOVOdat. Christina and Nang will stay in Singapore (forever!) and continue to populate WOVOdat.

But there will be more distributed work. Already there have been big contributions from PHIVOLCS, CVGHM. INGV and Japan has promised continued support. The funding from EOS was for 5 years . The Director of EOS has said he'll fund part if WOVOdat can raise the other half. The external part can be in cash or in kind (e.g., staff time in helping to build WOVOdat from the observatories). We need to make a good show to director of EOS of showing interest from VO's and it would be lovely if we could raise cash too, e.g., for airline tickets.

Several potential sources of future funding:

· Subscriptions by WOVO members - will be easier to sell after the database is more complete.

· A major partner? - Tim Ahern is willing to work with WOVOdat on programming (NSF grant to be a data hub); GVP was tried a few years ago but unsuccessful, but the door is still open; airlines (cf. IATA project)

· In kind contributions. (as above)

· Development aid funding - John Pallister is talking with USAID; there are others too DFID, JICA, AusAID, etc

· GVM - talking with Steve Sparks and Rob Muir-Wood of RMS for ideas on a business model.

· WorldBank - Andrew Tupper to send material to Chris Newhall (which he has now done).

· Private foundations? - There might be some. CN - we haven't tried yet, but we must!

· If EOS is no longer willing to support - not the case right now - WOVOdat would need 100% funding.

This will be a year of uncertainty, but we just took off and we need fuel to keep the engines running. We're short on time to get ideas. It'll be easier to sell now, with something to show, than 5 years ago when we had nothing.

Gill thanked Chris for the vision of WOVOdata and all the dedication and energy to get the plane off the ground.