Skip to main content
SAGE
Search form
  • 00:00

    [INTRO MUSIC]

  • 00:15

    VICTORIA BENNETT: There's a whole range of satellitesin space right now.And they're producing terabytes of data per day.And it's really a goldmine for doing environmental research.I work for CEDA, the Center for Environmental Data Analysis,which is based at STFC, the Science and TechnologyFacilities Council.

  • 00:37

    VICTORIA BENNETT [continued]: Our function is to support environmental scientistswith access to data and information services.And that includes getting data from people that producedata in the research community.It includes bringing data in from spaceor from climate models or from meteorological observations,ground observations-- all those sortsof things that environmental scientists are

  • 00:58

    VICTORIA BENNETT [continued]: interested in, bringing that into one big dataarchive that the scientists can access to do their studies.We've got the biggest environmental dataarchive in the UK.There are other environmental data centersand they specialize on different areas.Some of them focus particularly on the polesor on ecology, hydrology-- those sorts of things.

  • 01:20

    VICTORIA BENNETT [continued]: So they have their own complexity in their data,but none of them are quite as voluminous as the datathat we have here at CEDA for Earth observationand for atmosphere.We've put in data from space agencies, from Met offices,from scientists, observation networks,international organizations as well.

  • 01:41

    VICTORIA BENNETT [continued]: We tend to be very supportive of what the UK sciencecommunity need.So where there is a new data set out therethat the UK science community needs for a project or researchprogram, we'll investigate how to get those dataso they don't have to do it themselves.That's another big aspect of our work,is that we save other people of havingto fiddle about with licenses and data transfers.

  • 02:03

    VICTORIA BENNETT [continued]: One of the biggest areas of data that we holdis Earth observation data.Earth observation data comes from satellites,that comes in from the European Sentinel series of satellites.There are six or seven of them in space at the moment.And they measure different aspects of the environment.

  • 02:24

    VICTORIA BENNETT [continued]: Sentinel-1 data, which is radar data, the Sentinel-2, whichis imagery, and then Sentinel-3, which ison the picture behind me here.Our organization here, STFC's had a longstanding interestand involvement in these sorts of sensors.They're particularly used for surface temperature,

  • 02:47

    VICTORIA BENNETT [continued]: for understanding the climate.And we've been involved in building and calibratingthese kind of instruments for decades.We take the data, store them in our archive.And we do the science as well.Sentinel-1 data is being used to understand earthquakesand ground defamation.

  • 03:08

    VICTORIA BENNETT [continued]: The Sentinel-2 data is used to look at forests and land cover.The Sentinel-3 data is used for land surface temperature, seasurface temperature, clouds, and atmospheric aerosols,and also atmospheric composition to a certain extent.So there are lots of different thingsthat you can gather from these observations.

  • 03:35

    TIM WRIGHT: My name's Tim Wright.I'm a professor at the University of Leeds.And I'm the director of Comet, whichis the NERC center for the observationand modeling of earthquakes, volcanoes, and tectonics.Satellite radar interferometry, whichis a bit of a mouthful to say, so we say InSAR,is a technique that uses satellite radar images.

  • 03:58

    TIM WRIGHT [continued]: These satellites are flying around the Earth.They send a radar signal down to the Earth's surfaceevery 90 minutes and measure the reflection.And we can look at the difference between thosetwo images very precisely.We were contacted by the Ecuadorian governmentwho managed the volcanic hazard in the Galapagos.They were particularly concerned by these earthquakes

  • 04:20

    TIM WRIGHT [continued]: that had happened around Cerro Azulin the south of the Galapagos.They wanted to know, primarily, is this volcano going to erupt?So we had a look with what we could see with Sentinel-1 data.And this was the result. This interferogram,which is the product of this difference between the tworadar images, showed us exactly what

  • 04:42

    TIM WRIGHT [continued]: had happened in the 12-day period between the 8th of Marchand the 20th of March in 2017.What we can see is this part here in the North had subsidedand this part in the South had uplifted.And that's a very characteristic signalassociated with horizontal movement of magma.So the magma under the surface was going sideways rather than

  • 05:03

    TIM WRIGHT [continued]: going towards the surface.And the Ecuadorian government wereable to use this data in combinationwith information on the ground to lower the alert level.

  • 05:16

    VICTORIA BENNETT: We deliver the CEDA data archiveand the JASMIN facility for NERC, the Natural EnvironmentResearch Council.And JASMIN is a collaboration between RAL space and STFCin the scientific computing department.

  • 05:30

    HUGH MORTIMER: In RAL space, we'redeveloping remote sensing technologies that enable usto observe the world around us.When we have an instrument in space,it's often orbiting at very high velocities around the Earth.We orbit the Earth every 90 minutes.And that means that we're getting about 15 or 16 imagesof the Earth's surface, orbits of the Earth, every single day.

  • 05:52

    HUGH MORTIMER [continued]: Now that means we get global perspectives.And we get timescales and repeat timescales, whichyou can't get in any other way.

  • 06:01

    VICTORIA BENNETT: The Sentinel satellitesare producing huge amounts of data, bigger than any datavolumes we've ever had to deal with before.We've already got over four petabytes of Sentinel datain our archive.We're getting around eight terabytes per day from them.So this is one area where we're reallymaking a difference because we are saving everybody

  • 06:22

    VICTORIA BENNETT [continued]: else from having to bring those dataand having them on their computers.They can just come to us and use them here.

  • 06:30

    HUGH MORTIMER: The eight terabytes of data per daythat the Sentinel missions are creatingare very difficult to manipulate and handle.So therefore, we have to collect them and store themin a very rigorous way.The supercomputers and hardware required to store and processthis amount of data is needed because of the volume of data,

  • 06:54

    HUGH MORTIMER [continued]: but also, the complexity of the processingthat's required on all of these different data sets.Effectively, what we need to be able to dois take that data set and run it through numerical algorithmsthat will allow us to extract the useful information.For example, taking atmospheric data,we will need to be able to observethat data on a daily timescale, but then we

  • 07:16

    HUGH MORTIMER [continued]: need to run it through mathematical processes thatsimulate what's happening in the atmosphere, that enable usto forecast into the future what the potential weather iswithin the UK and within Europe.

  • 07:31

    VICTORIA BENNETT: So overall at CEDA, we'reoffering two things.It's the CEDA archive, which holds the environmental datafrom satellites, from climate models,from measurements, and from other sources.And next to it, we have the JASMIN computer facility.And having those two together is what reallymakes our organization work.The JASMIN facility has a huge amount of data storage.

  • 07:53

    VICTORIA BENNETT [continued]: And that storage includes our data archivethat we look after that holds datasets that people want to use.But it also has some storage that the scientists themselvescan dump stuff.They can bring in data they've made themselvesor some products that they've generatedand collaborate with people using that data storage.And then next to that, we've got all the computing capability.

  • 08:15

    VICTORIA BENNETT [continued]: So we've got about 12,000 computing cores at the moment.And people can access that as much as they need.It's shared between over 1,000 users.And we've got a really efficient algorithmfor allowing all those machines to be shared.

  • 08:29

    HUGH MORTIMER: These different datasets are produced and then stored in data centers,such as CEDA here at RAL Space.Looking at these different data setsfrom each one of these satellitesmeans we get a greater perspectivethan you can get than just looking at one of these datasets alone.

  • 08:47

    VICTORIA BENNETT: It's not just the Earth observation.That on it's own isn't enough.They also need weather observations, for example,and possibly some ground observationsto help them ground truth the observations from space.So it's having the combination of data sets in one place--is really useful for them, and having the computing rightnext to it, and having space thento put the data that they produce, as well, then share it

  • 09:09

    VICTORIA BENNETT [continued]: with their own collaborators.The interesting thing about the workthat people use CEDA and JASMIN foris that it covers all areas of sciences.It's really extremely diverse.It's everything from earthquake detectionto modeling storms, floods, volcanic eruptions,abstract pollution and air quality, large-scale climate

  • 09:32

    VICTORIA BENNETT [continued]: fluctuation.The possibilities are endless.And it's really interesting, actually,to see how JASMIN is being used.

  • 09:39

    TIM WRIGHT: In New Zealand, we'reparticularly interested in the earthquake hazard.We most recently got involved in a big earthquakethat happened in November 2016.This was a really impressive earthquake.It caused huge amount of deformation.When we look at earthquakes, the traditional wayof looking at them is with seismology.So that's listening to the seismic waves that are

  • 10:01

    TIM WRIGHT [continued]: emanated from the fault plate.But in this case, that the seismologistsgot it completely wrong.This rectangle is where they thought the earthquake wasand the red colors are where the strong shakingwas estimated to have been.Their prediction for shaking is shown by these red colors.The estimate here was that the earthquake had taken place

  • 10:22

    TIM WRIGHT [continued]: on a fault that went underneath the South Island,deep beneath the surface.What we saw when we actually looked at the radar data,was that the surface had been broken up very dramaticallyand very shallowly in a very complex network of faults.It changes our view of who was impacted by this earthquake.

  • 10:44

    TIM WRIGHT [continued]: So this is from the seismology.When you include the data from the InSAR,it completely changes our understandingof this earthquake.The pattern, the faulting representedby those black rectangles are in completely different places.But also, the pattern of intense shaking,shown by those red lines, is also different.

  • 11:06

    TIM WRIGHT [continued]: So using satellites to tell us what moved on the surface canreally inform responders in the field to say which places haveexperienced strong shaking and which ones haven't.

  • 11:20

    VICTORIA BENNETT: JASMIN's undergoing an upgrade rightnow, as we speak.We had just under 20 petabytes of storage until very recently.And we're now expanding that to 44 petabytes,which really is huge.1,000 gigabytes gives you a terabyte.And a million gigabytes gives you petabytes.And we've got 44 petabytes in the pipeline.

  • 11:43

    VICTORIA BENNETT [continued]: And I'm fairly sure that will be filled up pretty quickly.The rate that data is coming in is just phenomenal.So we'll be asking for more before you know it.We've also got a private cloud.And the cloud allows people to build outwardfacing services-- websites, for example,

  • 12:03

    VICTORIA BENNETT [continued]: that their own users can then accessthe data or their services without us even knowing aboutit really.So there are lots of different waysthat JASMIN and the data archives are being used.Data Intensive science is really key at the moment.

  • 12:25

    VICTORIA BENNETT [continued]: Data volumes are increasing all the time.The complexity of data is increasing all the timein the same way that all our phones and camerasand digital devices are capturing and holdingmore data.It's the same for every model that's simulating things.That data is just getting bigger and bigger.But with that, the complexity of dealing with those datais becoming more challenging.So we are finding that researchers,

  • 12:49

    VICTORIA BENNETT [continued]: who need to use these vast swathes of data,either need to be pretty computer savvy themselvesto understand how to handle these huge and complex datasets or there need to be people around that build toolsto make it easier for people to deal with these massive datasets.Occasionally, there's a bit of contention of somebody

  • 13:09

    VICTORIA BENNETT [continued]: that's hogging all the computer or someone'shogging your storage.People do need to be organized.They need to write software that usesthe system in an efficient way.And they need to be fairly tidy and organized with their datamanagement.So there's a job for us to do to make sure people work togethernicely in the community.[MUSIC PLAYING]

  • 13:33

    VICTORIA BENNETT [continued]: The Natural Environment Research Council has a data policy.And they require that all the science that they fund--that they deposit the data they'veproduced within a long-term data archive.Because it's important that we can go backto the scientific record and track and alsosee what's changed.The data and all the supporting data and computing that happens

  • 13:54

    VICTORIA BENNETT [continued]: here is one of the steps along the way in understandingthe wealth of information that wehave about the planet and the environment.And it's then a question of using tools and generatingvisualizations and maps and pictures-- all these thingsthat make it easier for people in the real worldto understand what's going on and make decisions that impactpolicy and government decisions.

  • 14:16

    HUGH MORTIMER: In being able to communicate science,we're able to educate people about the importanceof science.We're able then to inspire the next generation,but also then to inform about some of the complexitiesand some of the issues associatedwith our Earth, our earth system, that in other waysthey wouldn't have access to.

  • 14:38

    VICTORIA BENNETT: In the next five years,the data is going to keep coming in.And the users are going to keep wantingmore space and more computers.So one thing that we're always doingis looking at new technologies and working out howbest to address those needs.Because we can't just keep buying more storage.It just isn't sustainable.So we're always looking at clever ways

  • 14:60

    VICTORIA BENNETT [continued]: to solve that problem.I don't know what the next clever way is,but we're evolving all the time.

Video Info

Publisher: SAGE Publications Ltd

Publication Year: 2019

Video Type:In Practice

Methods: Data collection, Data archives

Keywords: data analysis; data storage; Earth science; earthquakes; environmental analysis; environmental science; imagery; Satellite networks; Satellite technology; tectonics; volcanoes ... Show More

Segment Info

Segment Num.: 1

Persons Discussed:

Events Discussed:

Keywords:

Abstract

Victoria Bennett, division head STFC CEDA, and Hugh Mortimer, research scientist at STFC RAL Space, explain how environmental scientists have access to unprecedented volumes of Earth observations and satellite data through CEDA's data archives and the JASMIN facility. Professor and director of COMET at the University of Leeds, Tim Wright, provides examples showing how some of this data is being used to analyze earthquake hazards.

Looks like you do not have access to this content.

Archiving, Managing & Analyzing High Volumes of Environmental Data: CEDA

Victoria Bennett, division head STFC CEDA, and Hugh Mortimer, research scientist at STFC RAL Space, explain how environmental scientists have access to unprecedented volumes of Earth observations and satellite data through CEDA's data archives and the JASMIN facility. Professor and director of COMET at the University of Leeds, Tim Wright, provides examples showing how some of this data is being used to analyze earthquake hazards.

Copy and paste the following HTML into your website