Data Management in Earth Observation - The Potential of Open Digital Twins of the Earth
This is a modal window.
The media could not be loaded, either because the server or network failed or because the format is not supported.
Formal Metadata
| Title |
| |
| Title of Series | ||
| Number of Parts | 351 | |
| Author | ||
| License | CC Attribution 3.0 Unported: You are free to use, adapt and copy, distribute and transmit the work or content in adapted or unchanged form for any legal purpose as long as the work is attributed to the author in the manner specified by the author or licensor. | |
| Identifiers | 10.5446/69225 (DOI) | |
| Publisher | ||
| Release Date | ||
| Language | ||
| Production Year | 2022 |
Content Metadata
| Subject Area | ||
| Genre | ||
| Abstract |
| |
| Keywords |
FOSS4G Firenze 202220 / 351
1
7
13
22
25
31
33
36
39
41
43
44
46
52
53
55
58
59
60
76
80
93
98
104
108
127
128
133
135
141
142
143
150
151
168
173
176
178
190
196
200
201
202
204
211
219
225
226
236
242
251
258
263
270
284
285
292
00:00
Open setSatelliteTwin primeType theoryOrder (biology)ForceCartesian coordinate systemState of matterMultiplication signState observerPoint (geometry)Task (computing)Computer programmingSatelliteService (economics)Domain nameSpacetimeLatent heatMultiplicationMeasurementDependent and independent variablesObservational studyData managementPredictabilityDigitizingOnlinecommunityBitOperator (mathematics)Source codeOnline helpImage resolutionFrame problemPerspective (visual)WordGoodness of fitComputer animation
05:10
SatelliteProduct (business)Green's functionFigurate numberFile archiverVolume (thermodynamics)TowerPhysical systemTransformation (genetics)Population densitySatelliteMereologySet (mathematics)Service (economics)State observerPerspective (visual)Cartesian coordinate systemFunctional (mathematics)Domain nameBitSoftwareData managementKey (cryptography)Message passingChainLatent heatInformationRule of inferenceLevel (video gaming)Different (Kate Ryan album)Order (biology)OnlinecommunityTerm (mathematics)LengthSpacetimeInteractive televisionObservational studyComplex (psychology)Limit (category theory)Line (geometry)Point (geometry)FreewareFloppy diskService-oriented architectureOpen setComputer animationSource code
10:21
ConsistencyProcess modelingDistanceForm (programming)MultimediaStructured programmingPhysical systemState observerKey (cryptography)Operator (mathematics)Order (biology)Product (business)Variety (linguistics)Service (economics)Term (mathematics)Set (mathematics)1 (number)Latent heatNumberBoundary value problemFocus (optics)Electric generatorMorley's categoricity theoremOrientation (vector space)Reverse engineeringComputer programmingLine (geometry)Service-oriented architectureGoodness of fitVelocityComputer animation
12:39
Process modelingPredictionAreaDisintegrationComputer forensicsDecision theoryDigital signalElectronic visual displayPoint cloudNetwork topologyEvent horizonFocus (optics)Parameter (computer programming)Observational studyNumbering schemeMetrologieComputing platformForceInteractive televisionSoftware developerCollaborationismData integrityScheduling (computing)SupercomputerPresentation of a groupElement (mathematics)Cartesian coordinate systemCentralizer and normalizerProcess modelingDomain nameOpen setProduct (business)Twin primeService (economics)Data fusionHypermediaForm (programming)TwitterInformationCloud computingPredictabilityDigitizingFunctional (mathematics)Machine visionSimulationView (database)Key (cryptography)Set (mathematics)Point cloudAdaptive behaviorExtreme programmingAddress spaceOrder (biology)Business modelSlide ruleVisualization (computer graphics)Point (geometry)Computer architectureSoftwareChannel capacityHarmonic analysisPower (physics)Event horizonDecision theoryPhysical systemPhase transitionVector potentialComputer programmingLevel (video gaming)Basis <Mathematik>Artificial neural networkScientific modellingElectronic data processingMathematical optimizationDegree (graph theory)Lattice (order)Connectivity (graph theory)State observerOnlinecommunityFundamental theorem of algebraResultantSource codeComputer animation
19:47
CASE <Informatik>Standard deviationEndliche ModelltheorieDigital signalSurfaceMathematical analysisUser interfaceCASE <Informatik>Twin primeEndliche ModelltheorieComplex systemOrder (biology)Element (mathematics)Observational studyContext awarenessPhysical systemProcess modelingKey (cryptography)Dynamical systemWater vaporCartesian coordinate systemComputer animation
20:57
Cloud computingProcess modelingData managementDenial-of-service attackMoistureFunction (mathematics)SimulationData terminal equipmentDemo (music)Domain nameDenial-of-service attackWater vaporLevel (video gaming)Cartesian coordinate systemEndliche ModelltheorieData managementCASE <Informatik>Online helpPanel painting
21:29
System programmingProcess modelingFrequencySatelliteBulletin board systemEstimationTwin primePhysical systemPredictionCategory of beingPhysical systemEndliche ModelltheorieHeat waveComplex systemDomain nameCartesian coordinate systemBitComputer animation
22:07
DisintegrationTwin primePlastikkarteWordSystem programmingComputer configurationData integrityComplete metric spaceElement (mathematics)Computing platformFigurate numberPoint (geometry)PlastikkarteCentralizer and normalizerTwin primeCartesian coordinate systemService (economics)Context awarenessMeasurementPhysical systemDigitizingVector potentialDecision theoryINTEGRALProcess modelingIntegrated development environmentLogicSlide ruleBitObservational studyLatent heatArithmetic meanPerspective (visual)Natural numberoutputSet (mathematics)Image resolutionResultantInteractive televisionLevel (video gaming)Cycle (graph theory)Key (cryptography)Open setLattice (order)Product (business)SimulationSystem callInterface (computing)Functional (mathematics)Complete metric spaceComputer configurationGoodness of fitScaling (geometry)Mathematical analysisFile archiverCore dumpAuthentication
Transcript: English(auto-generated)
00:00
Thank you very much. It's a real pleasure to be here. I will try to do my best to relaunch your metabolism after, hopefully, a good lunch. And maybe just a few words of introduction. So I'm with the European Space Agency, and I'm responsible at the European Space Agency
00:23
for the Earth observation satellites and operations, for everything relating to the satellites and operations, to the ground segments, the data management aspects. And I'm also coordinating for the European Space Agency a topic that I see as a framing topic for a lot of the things that you are discussing here,
00:43
a framing theme that is emerging under the label digital twins. And I will try to explain a little bit how from the challenges of Earth observation data management, we have reached the ability of what we call predictive,
01:03
to do what we call predictive Earth observation with the help of so-called digital twins. But before I do this, I want to take you on a little voyage into the history and the very source of the data
01:21
that most of you are dealing with. So I will talk a little bit about the infrastructure that we are operating. And we'll just paint a little bit the picture from 15 years ago towards what we are seeing today. So I'm taking a little bit a look from the European perspective and from a European Space Agency perspective.
01:43
The big programs that we are operating, of course, are the Earth Explorer program and the Copernicus program. And this has completely changed our approach to how we handle data, how we deal with the communities, and how the infrastructures and how we involve industry,
02:03
industrial infrastructures in the task that we have in front of us. Earth observation is by far the largest directorate of the European Space Agency. We have an annual budget of about 1.7 billion. And this is a very important point in time for us
02:21
because in November this year we will have a ministerial conference in which the 22 member states will decide on the contribution to ESA for the next three years, the funding for ESA for the next three years. And, of course, it's the applications that are driving the attractivity and the need for the Earth observation topics
02:44
that we are putting in front of the ministers to sign up on the budgets that we will need in order to do all the things that we want to do. The look 15 years ago, from our perspective,
03:00
we had a few big individual satellites for a few types of measurements and for very specific communities, typically scientific communities very much focused, very much grown in a specific discipline. And to give examples, big satellites with multiple instruments,
03:22
ERS, NVSAT in particular, representing an era where we were trying to optimally serve these specific communities and scientific domains. The real origin and one of the earliest fleets of Earth observing satellites were, of course, dedicated to meteorology and are today operated
03:45
also by our colleagues at Jumetsat in Germany. Now the picture has completely changed. As I mentioned, we have a large fleet of Copernicus Sentinels, seven big satellites. We have a fleet of Earth Explorer satellites,
04:01
another six satellites. We have Scout missions. We have Fisats. We have, of course, still a meteor satellite fleet that has expanded. And we have a lot of new space players in Europe, which are especially important to us because one of the key tasks that we have been fulfilling
04:21
over the decades is to integrate commercial data, in particular very high resolution data that is not generated by Copernicus and the Earth Explorers into the services and applications that now shape the overall domain of Earth observation services and applications.
04:43
Services, science, technology, commercial applications indicating that the user community has dramatically diversified. And there's no such thing as a specific user community anymore. Even the typology efforts that we have done over the last decade always were quickly
05:03
becoming obsolete because we realized that users that were completely out of the scope up to that point immediately took an important role in this ecosystem. From a European perspective, of course, the national missions
05:20
and, as mentioned, the commercial missions are part of that infrastructure and that ecosystem. And we should say, and I say that with a little bit of pride, that we in Europe really have one of the biggest and most sophisticated infrastructures for Earth observation in the world.
05:41
So we have grown from careful steps in the Earth observation domain 15, 20 years ago towards a very operational and a very broad application and service-oriented ecosystem of infrastructure,
06:01
but most important also of the ground systems and the data management aspects that are so important here for our discussions. So again, 15 years ago, small data in the sense that some of the satellites were generating perhaps one petabyte over 10 years.
06:21
Even if you were processing these data sets to different levels, few users in these communities, maybe hundreds of users, thousands of users, specialist users, no fusion with the data industry. We had ESA, a proprietary ESA asset infrastructure in place
06:43
that was fixed in order to support these missions. All of this has changed. Now we have a very complex data management system that is entirely industrialized. So the interaction and the partnership with European industry in order to be able to acquire,
07:01
to process, to archive, and to transform data into services and applications is one of the key features that has emerged over the last 10 to 15 years. The user communities have dramatically increased. We are now talking about hundreds of thousands of users, even millions of users.
07:23
In the spirit of this conference, I think the key message, of course, is that ESA data and Copernicus data is free and open data. So fully fulfilling the spirit of what you are discussing here and everything that we are trying to do and to develop in order to increase the uptake of this data
07:44
follows these principles very much. Specialists and non-specialists users have merged. Of course, one of the key challenges and probably the key term of reference for my own work is the ability to activate user communities
08:02
that are even not yet aware that Earth observation data is relevant for them, is usable in a commercial, in the service, in the institutional domain, but also, of course, in the scientific domain. Infrastructure and software. We have largely taken as a service function
08:23
in this big system that we have been building up. And again, we are integrating national and commercial data. From our perspective, one of the key challenges that is still remaining and is still part of the portfolio that we are looking at
08:44
is the transformation and the valorization of the data with European assets and rules into reliable and relevant information so that we can truly cover the full value chain that this huge investment that the space Earth observation infrastructure
09:03
constitutes can also be valorized at the full length of that chain. Just a picture to give you a bit of a sense when I talk about data volumes, what I actually mean. Our data dissemination systems,
09:23
meaning the first line of data dissemination because with free and open data there is no limit to the layers of data dissemination that can actually be operating and that we are actually very much welcoming, but the one that we are controlling directly is disseminating so much data each day
09:41
that if you would put it on a high-density disk would actually reach the size of the Eiffel Tower. So that's the data volume that goes from our ESA data hubs directly to the users. It's a figure that we are very proud of because in particular this is Copernicus data and Copernicus in its beginning was conceived
10:01
as a system that would only serve six specific Copernicus services of the European Union or entrusted entities, and we have now grown a community around these data access systems that is getting close to one million registered users.
10:22
Just to remind everyone what the key data challenges in Earth observation were over the last 10 years. It was the big three V's and there's also two other ones, so traditional ones of volume, velocity, so the availability, the quick availability of data
10:42
where it's needed, the variety of the data that we are seeing. One of the big revelations of the last few years was the fact that we are realizing that our scientific missions are actually becoming a very valuable operational contribution to the more operational
11:01
and service-oriented Copernicus system, so the boundary between purely scientific missions and operational-slash-service oriented missions is being blurred, and we see that the potential to generate programmatic lines for operational missions
11:22
out of an R&D-focused scientific program is very much a reality these days, and just to give you a reverse example which is showing the transparency of that categorization that was still valid a few years ago, the largest number of scientific publications
11:42
in Earth observation on space-based Earth observation is actually done today with a specific set of Copernicus sentinels. Good. Then we have the veracity of the data, something that is taking a very key role these days in terms of the robustness,
12:01
the credibility of the products and services that are being generated, and then the question of value, scientific insights, social benefits, commercial value. We are constantly in this challenge in order to illustrate these benefits, in particular to the politicians,
12:22
but also to the people that are electing the politicians in order to mobilize the budgets in order to address the big societal challenges that we are facing today, and I don't have to mention them to you.
12:40
Just to give you a few pointers, a few entry points into the question of digital twins. So what we have seen here is a set of new elements of consideration that have sprung up from this data challenge that I've been outlining. We see new and more flexible data policies,
13:01
free and open. Here is the top topic in this slide, and that's something that has been guiding and is still guiding and increasingly guiding our considerations in order to develop and to foster the ecosystem together with our partners. Licensing was mentioned in many presentations earlier,
13:23
and also the concepts, the commercial concepts, how to develop attractive business models where things like anchor customer rich and innovative schemes are being used. We are seeing a stronger focus
13:40
on the information per se than the data. That's, I think, a general trend that everybody is observing. And another element is, of course, the question of data integration. We have numerous new ways, particularly in the form of the application platforms, the public media, the service, industrial service products,
14:01
and the operational public services that are integrating Earth observation data where you wouldn't have found Earth observation previously. So that's a very strong new component to be considered. Cloud, platform-based data access,
14:21
something that is a central topic here at this meeting as well. And the realization that across the missions, but also across the agency and across agency and the industrial domain, Earth observation data can be fused to a degree that it becomes completely irrelevant for the user
14:42
where the actual data is coming from if specific things are being considered or being observed. And the openness of the process, the credibility of the process that happens to the data is one of the key aspects to make that data fusion as reliable and as optimal as possible.
15:02
And we are working very closely together with our partners at NASA, with European partners to make this data fusion a strong new component in this ecosystem. And then finally, what we are realizing is that we can learn from the lessons
15:24
and from the potential that we have seen evolving in the domain of meteorology and climate research where modeling and predictive scenarios of what will happen in the future if certain parameters are being changed, if observational data is constantly fed into this model,
15:43
that this predictive potential of system can be expanded into other domains of Earth observation. And that's basically our fundamental idea about digital twins of the future to take this concept, the power of high-performance computing
16:01
of artificial intelligence, of generating scenarios that can serve as the basis for decision-making to the next level. And we are working hand-in-hand here with the European Union. We have joined with our partners, Jumitsat and ECMWF, into a big program that is called DESTINY,
16:22
a program that is funded by the European Union and which these three entities, us together, are setting up digital twins in the first phase to particularly address questions of extreme events and of climate change adaptation.
16:41
And the underlying idea, again, is what if we create a comprehensive digital replica of the Earth that stimulates and observes human-Earth interactions with high accuracy, uses the technologies that I've just mentioned, is continuously updated so it's always fed with the most recent and relevant data sets,
17:03
enables the simulation and prediction capabilities, and provides detailed use of the past because the past feeds into the current analysis and into the prediction and the future models and simulations that are key for realizing that ambition
17:22
and that vision of digital twins. Just a few comments on the infrastructure and on the way that we structure the functionality. So cloud infrastructures, of course, play an important role. All Copernicus data is available on the cloud.
17:42
HPC infrastructures, what we are learning and what we are understanding, how to harness the power of high-performance computing centers for data processing, but most of all for the modeling aspects of what we are intending to do. The harmonization of data lakes,
18:00
the harnessing the capacity and the capabilities of simulation observation centers, and again AI software science competence centers, and the networks in order to be able to access the results or disseminate the results as needed. So these are the typical architectural elements
18:21
that we are putting together in order to make these digital twins a reality. Simulation and monitoring, past, present, future. One of the key aspects of a digital twin is digital twins should be usable by people who are not used to deal with geospatial data,
18:40
that are not used to deal with earth observation data. So the aspect of visualizing and visualization of the potential of this data and handling of the data is one of the central requirements for these systems, and we are seeing applications emerging now
19:02
all over the place, which facilitate a complete rearranging of what I would call the scientific community behind these things. Destination earth, digital twin and earth principles, again, these are a few architectural elements.
19:21
We are on a very tight schedule for these activities, and we want to go out with the system, go live with the system by the end of next year, and we are involving larger user communities and larger and larger communities over the next three years.
19:41
Openness, again, one of the key principles, and collaboration, visibility aspect, interaction aspects are very key. Just to illustrate that I'm not talking only about things that are emerging. Some of the applications, some of the potential scenarios
20:02
that we can work with have already emerged in precursor activities to the digital twins that we have been feeding with the earth observation data that we are generating. Just to go quickly through a few examples, so this is the case of Antarctica where you have a very complex system of climate, ocean, ice sheet, water,
20:25
and a lot of dynamic, Antarctica is a very dynamic place in which you can test and challenge a modeling system to its extreme, and it's one of the key applications, of course, in the context of climate change.
20:41
Again, a few schematic drawings here, how these elements are interacting. It goes from snow grain size to albedo. Everything is fed into this system in order to generate these scenarios and to optimize the models and the scenarios that are coming out of these models. Hydrology is another application domain
21:02
that is taking center stage for obvious reasons these days. In this case, here we have looked at water management, flood risk, and landslide risks associated with this topic. Forestry is another domain that we have systematically analyzed
21:21
and where we have run demo activities with the help of industry. Again, showing a little bit the various applications and outcomes that can be generated. Food systems, food supply, very much in our minds these days is another domain where this approach
21:43
seems to be highly promising and has already generated very remarkable outcomes. The ocean system, another complex system that can be looked at, and heat waves in the ocean system, something that I didn't know before, is of course closely related
22:01
and closely linked also with the models that have been developed in the overall climate domain. Let me finish with this example because from my personal experience and exposure to where digital twins are actually really entering our daily lives
22:21
on a very tangible level and where it's easy to illustrate the functionalities and the potential of digital twins, it's the example of digital twin cities. These digital twin cities are a bit special in the sense that constitute the interface between what I would call engineering industry,
22:41
sometimes call it product and production digital twins, and digital twins of natural systems. The previous slides were more focused on the natural systems, but interesting stuff and interesting applications happens of course at the interface, and in the end, the digital twins of the natural systems
23:01
that I've been showing are very much dependent on the inputs and the data sets that will be provided as a result of human activity and human interaction. Many of these digital twins are operational on a regional or local level. They are very diverse, and typically the best examples that I've seen,
23:22
they are very open, so the citizens that can profit from these digital twins have very much access to the way that the scenarios on which they, as the affected people, can decide on are generated.
23:40
Modeling, simulation AI are directly led into a circular process for decision-making scenarios, and again, fully transparent integration of citizens in scenarios and decision cycles. So various cities in Europe are actually applying all the principles that I've tried to outline theoretically here,
24:04
more based on the engineering data than the natural system data, but Earth observation data is flowing increasingly into these systems. Heat data is a big topic, for example, so that people can actually make their own educated decision
24:24
on how to improve the local environment and the regional environment they're living in, and that, I think, on the small scale, very much exemplifies and illustrates the idea and the logic of digital twins as a tool for society,
24:42
for individuals to make good decisions for the future of this planet. Just to conclude, this is my final slide to illustrate a bit where the specific role of ESA lies in the context that I've been outlining here.
25:01
So the free and openness of the data is at the core, and we are heavily defending that principle because we see the benefits and the potential of increasing and reaching out into communities by the means of keeping that data policy
25:21
as one of the leading principles. The other things, we are very much from an engineering and R&D perspective looking at novel acquisition, processing, archiving, analysis, integration approaches, application approaches, so that's another key element that's very much on our mind,
25:41
and again, we are very much focusing on the data access options because if the user is not able to access the data in the most convenient way and most convenient typically in a platform way where it's not necessary actually to download the data, so in the sense that I hope the figures of data disseminated
26:02
will actually go down at some point is clearly in the cards here. And then very high resolution data I've been mentioning. This is something that is a global competition, I would say. It's a very commercial thing these days
26:22
to deal with very high resolution data, and very high resolution data will be key for some of the most central applications and services to be grown out of the digital twin context as well. And finally, ESA supports measures to assure Earth observation data integrity,
26:43
the completeness of the data sets, the timeliness of the data, the authenticity, and finally, and I think again very much in the spirit of this meeting here, the openness of the data which I think needs to stay at the center of everything that we are doing. With that, I close and thank you very much.