welc lcome
play

Welc lcome Conversations with Academia Big Data for Big - PowerPoint PPT Presentation

Welc lcome Conversations with Academia Big Data for Big Challenges: The Swiss Data Cube for Environmental Monitoring 1. Understand how Earth Observations (EO) acquired by satellites can be used for environmental monitoring Session 2.


  1. Welc lcome Conversations with Academia Big Data for Big Challenges: The Swiss Data Cube for Environmental Monitoring

  2. • 1. Understand how Earth Observations (EO) acquired by satellites can be used for environmental monitoring Session • 2. Explore the potential and challenges of EO Objectives data to drive progress against key national and international development agendas like the Sustainable Development Goals

  3. • Dr. Gregory Giuliani Senior Lecturer, University of Geneva, Head of the Digital Earth Unit, GRID-Geneva • Professor Pascal Peduzzi Speakers Director, GRID-Geneva • Dr. Claudia Röösli Senior Lecturer, University of Zurich, Group Leader, Remote Sensing Laboratories

  4. Big Data for Big Challenges: The Swiss Data Cube for Environmental Monitoring Dr. Gregory Giuliani, Dr. Claudia Röösli, Prof. Pascal Peduzzi

  5. The key to sustainable development… …is achieving a balance between the exploitation of natural resources for socio- economic development , Natural Capital and conserving ecosystem services that are critical to everyone’ s wellbeing and livelihoods. Socio-Economic needs

  6. The Challenge: E vidence-based policy-making Action: Decisions Societies decide how to use resources of our planet Understanding Meaning Value Knowledge Knowledge: inform us about the limits of the planet Information Data

  7. To better understand these changes … Our plant is under continuous observation from satellites

  8. Copernicus – Europe’s Eye on Earth Largest EO data provider in the World: 250TB/day data Archive: 250PB of data stored, daily growth rate: 220TB

  9. Sentinel-2: 22’000 images to cover the Earth Every 5 days! > 1’600’000 images per year!

  10. https://apps.sentinel-hub.com/eo-browser/

  11. Spatial Resolution = Pixel Size Freely available Commercial satellites

  12. ‘p ’ Not just photos… Courtesy of Geoscience Australia

  13. VIS Visible Light RGB NIR Near Infra-Red (IR) Spectral Resolution SWIR Shortwave IR TIR Thermal IR VIS NIR SWIR 1 SWIR 2 TIR 13 bands 11 bands 8 bands 14 bands 36 bands

  14. Spectral Resolution • Vesuv ITA • Sentinel-2 • 12-July-2017 • Red, green, blue (See also Cicalaet al., 2018) Page 16

  15. Spectral Resolution • Vesuv ITA • Sentinel-2 • 12-July-2017 • SWIR2, SWIR1, NIR (See also Cicalaet al., 2018) Page 17

  16. Temporal Resolution → a game changer Landsat 8 Landsat 8 + Sentinel-2A&B Images per year Li & Roy (2017)

  17. Sentinel-2: 22’000 images to cover the Earth Every 5 days! > 1’600’000 images per year!

  18. How to transform this large amount of data in useful information and support evidence- based decisions?

  19. Everything is in place… Why doesn’t the world use satellite data? • It requires scientific knowledge to understand what data is needed … optical (which resolution?), radar (which type?) • It is hard to access or download • It is hard to prepare … atmospheric correction, grid formats, pixel alignment, speckle filtering • It requires capacity building and training A New Solution… DATA CUBES

  20. Sampling a Data Cube Longitude Latitude A single time slice, similar to a Time standard “scene” can be used to assess a single point in time Pixels in the Data Cube are processed, aligned, and compressed and ready for data analysis Adapted from CEOS

  21. Sampling a Data Cube Longitude Latitude Several time slices can be combined Time into one to form a “ Mosaic ”. This is often used to reduce clouds or create seasonal or annual images. Typical Mosaics ... Most/Least Recent Pixel, Mean/Median, Geomedian, Min/Max NDVI Adapted from CEOS

  22. Sampling a Data Cube Longitude Latitude Examples of Time Series analyses include: Land Change Time (PyCCD), Water Change (WOFS), Parameter variation along a transect (Hovmoller plot) Time Series analyses consider the variation of data over time to assess change Adapted from CEOS

  23. • Governments have national and international reporting commitments and obligations as well as national environmental programs. • They all need information that is synoptic, consistent, spatially explicit , sufficiently detailed to capture anthropogenic impacts , and national in scope. • EO Data Cube can provide the long baseline required to determine trends, define present, and inform future . This can fit these interests to inform programs and communities.

  24. Updated every week! SWISS DATA CUBE in Numbers A unique Analysis Ready Data Archive 7 sensors 36 years 10-30-90 m LANDSAT 5/7/8; FROM 1984 to 2020 PIXEL RESOLUTION SENTINEL-1/2 A-B > 450 million > 1000 billion PIXELS OBSERVATIONS ~ 12500 images ~ 5 TB INGESTED ANALYSIS READY DATA ~ 10 millions CHF Giuliani G., Chatenoux B., De Bono A., Rodila D., Richard J.-P., Allenbach K., Dao H., Peduzzi P. (2017) Building an Earth COST OF DATA WITHOUT OPEN DATA Observations Data Cube: lessons learned from the Swiss Data ACCESS POLICY Cube (SDC) on generating Analysis Ready Data (ARD). Big Earth Session X: Session Title Slide 28 Data 1(1):1-18

  25. Bondo Landslide – 23 august 2017

  26. Earth Observations is useful for monitoring SDG’s http://earthobservations.org/geo_sdgs.php

  27. Water detection – Drought impact (2018)

  28. Urbanization – Bulle 1985/2018

  29. Rhône glacier – 1985/2018/2020

  30. 6. August 2015 13 .September 2020 ~410m

  31. SDG 15.3.1 Land Degradation … … is undermining the well-being of 3.2 billion people (IPBES) ESA UNCLASSIFIED - For Official Use Gregory Giuliani | ESRIN | 12/09/2019 | Slide 32 Giuliani et al., submitted http://www.geoessential.eu Collaboration: University of Geneva; GEO; ESA, CNR, JRC; UN Environment

  32. Aggregated indicators … … are not enough for public policy! ESA UNCLASSIFIED - For Official Use Gregory Giuliani | ESRIN | 12/09/2019 | Slide 34

  33. Disaggregation of indicators … … to capture spatial (maps) and temporal dynamics (graphs) How much? Where? When? Who? ESA UNCLASSIFIED - For Official Use Gregory Giuliani | ESRIN | 12/09/2019 | Slide 35

  34. SDG15.3.1 – Results from Switzerland Official value: 4.7% SDC value: 9.7% Official definition in Switzerland is based only on soil sealing and do not consider land productivity! Do not comply with the official UN definition! Giuliani G., Chatenoux B., Benvenuti A., Lacroix P., Santoro M., Mazzetti P., Monitoring Land Degradation at national level using satellite EO time-series data to support SDG15 – Exploring the potentiation of Data Cube, Big Earth Data, https://doi.org/10.1080/20964471.2020.1711633

  35. Jan. 2020 Sentinel 5P - Air Pollution Monitoring (NO 2 ) 0 (mol/m 2 ) 0.0002

  36. Mar. 2020 Sentinel 5P - Air Pollution Monitoring (NO 2 ) 0 (mol/m 2 ) 0.0002

  37. Apr. 2020 Sentinel 5P - Air Pollution Monitoring (NO 2 ) 0 (mol/m 2 ) 0.0002

  38. Drone orthomosaic: Hossein Torabzadeh

  39. Essential Biodiversity Variables (EBVs)

  40. PhenoSwiss – Joint UNIGE/UZH project

  41. PhenoSwiss – Joint UNIGE/UZH project

  42. The SDC supports the « Digital Switzerland » strategy ▪ Support innovation and growth in the digital economy ▪ Improve efficiency and effectiveness of government investments ▪ Improve management of natural resources ▪ Stimulate research ▪ Effective monitoring mechanism ▪ Generate information products ▪ Improve data access and use & enable new products/services that can transform everyday life

  43. • Open Data: Landsat 5,7, 8 ARD; Sentinel 1-2 ARD + All scientific/decision- ready products are freely, openly available & FAIR compliant • Open Notebooks: All algorithms are documented and openly available • Open Access: All publications • Open Source: All applications • Open Educational Resources: Bringing ODC into practice Session X: Session Title Slide 49

  44. Com ommon n Sens nsing - https://www.com ommon onsensing.org.uk

  45. Di Digit ital l Earth Australia lia - http://www ww.ga.gov.au/dea

  46. Digital Ear arth Af Africa a - https ps:// //www.di digitaleartha hafrica.or org

  47. Open pen Data Cub ube e https://www ww.opendatacube.org

  48. Data Cube on Demand (DCoD) Aims at facilitating the generation and use of an ODC instance virtually anywhere in the World . Users are only required to specify: • an area of interest on a web-based mapping application; • types of sensors between Landsat 5-7-8 and Sentinel-2; • desired temporal frame; Then automatically an empty ODC instance is instantiated and desired data are ingested. Giuliani G., Chatenoux B., Piller T., Moser F., Lacroix P. (2020) Data Cube on Demand (DCoD): Generating Earth Observation Data Cube anywhere, International Journal of Applied Earth Observation and Geoinformation 87:102035 Session X: Session Title Slide 55 https://doi.org/10.1016/j.jag.2019.102035

  49. Generating a Data Cube anywhere in the World…

  50. World Environment Situation Room – https://wesr.unep.org

Recommend


More recommend