Top Banner
Web-based, GPU-accelerated, Geospatial Platform for Real-time Monitoring, High-performance Analysis, and Early Warning - Tsunami Showcase - Spring 2015 [email protected]
28

Web-based, GPU-accelerated, Geospatial Platform for Real-time Monitoring, High-performance Analysis, and Early Warning

Jul 21, 2015

Download

Science

Welcome message from author
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Page 1: Web-based, GPU-accelerated, Geospatial Platform for Real-time Monitoring, High-performance Analysis, and Early Warning

Web-based, GPU-accelerated, Geospatial Platform forReal-time Monitoring, High-performance Analysis, and Early Warning

- Tsunami Showcase -

Spring 2015

[email protected]

Page 2: Web-based, GPU-accelerated, Geospatial Platform for Real-time Monitoring, High-performance Analysis, and Early Warning

TRIDEC Cloud

• Cloud-based and web-based platform to open up new prospects for Tsunami Early Warning and Mitigation Systems

• Combination of several complementary cloud-based services merged into one platform for– Web-mapping of hazard specific geospatial data,

– Handling, sharing, and communicating threat specific information

– Collaborating in a distributed environment

• Modern web-based Graphical User Interface– Operators in warning centers and stakeholders of other involved parties (e.g. CPAs,

ministries) just need a standard web browser

Page 3: Web-based, GPU-accelerated, Geospatial Platform for Real-time Monitoring, High-performance Analysis, and Early Warning

Features

Page 4: Web-based, GPU-accelerated, Geospatial Platform for Real-time Monitoring, High-performance Analysis, and Early Warning

TRIDEC Cloud – Earthquake Information

• Integrates latest earthquake information from http://geofon.gfz-potsdam.de

• Simple REST API enables integration of other national and international systems

• List displays earthquake events with information on– Magnitude, location, time,

– Dip, strike, rake,

– Classification based on simplified decision matrix,

– Prediction based on attached simulation either computed automatically or manually in the background

• Map displays location of earthquake events

• http://youtu.be/C56L4qj9aRA

Page 5: Web-based, GPU-accelerated, Geospatial Platform for Real-time Monitoring, High-performance Analysis, and Early Warning

TRIDEC Cloud – Sea Level Measurements

• Integrates historic and real- time sea level data from http://ioc-sealevelmonitoring.org

• Simple REST API enables integration of other sea level stations and systems

• Users select the sea level stations based on their needs• Map displays position of the sea level stations• Small diagrams below the map display the measurements for sea

level each station– In default mode diagrams show the real-time data for the stations– For events selected by the user diagrams display the relevant sea level data for the selected

event– If a simulation for an event has been computed either automatically or manually diagrams

display additionally the predicted sea level data– Comparison of real data with simulation data for sea level stations

• Interactive diagrams for a detailed analysis and picking– Select and open with a click on a small diagram– Sliders for picking period and amplitude values or form to configure values manually– Picked values in the interactive diagram can be enabled for integration in warning

• http://youtu.be/C56L4qj9aRA

Page 6: Web-based, GPU-accelerated, Geospatial Platform for Real-time Monitoring, High-performance Analysis, and Early Warning

TRIDEC Cloud – Simulation Computation

• On-demand tsunami predictions based on actual source parameters

• Manual and automated background computation using graphics processing units (GPU) for instant tsunami propagation calculations

• On-demand tsunami simulations are computed for an event either automatically or manually by a GPU Cloud– With tsunami travel times (TTT), estimated times of arrivals (ETA), estimated wave

heights (EWH/SSH), and mareograms

– For visualization purposes with isochrones and tsunami-jets on a map

– For visualization and analysis with time series in diagrams

– For further analysis and processing purposes with tsunami forecast points (TFP) and tsunami forecast zones (CFZ) important for generating warning messages

Page 7: Web-based, GPU-accelerated, Geospatial Platform for Real-time Monitoring, High-performance Analysis, and Early Warning

TRIDEC Cloud – Virtual Scenario Player

• Enables training and exercises with virtual scenarios

• With special accounts operators and others involved are able to train skills and to prepare for large exercises and real events

• For training purposes virtual scenarios are playable with an acceleration to run prepared scenarios faster than in reality

• Selected scenarios in NEAM region and other regions in the world are already available

• New scenarios might be added for– Internal training, e.g. weekly, or

– Extended monthly communication tests

• http://youtu.be/CaRXYR4PE8s

Page 8: Web-based, GPU-accelerated, Geospatial Platform for Real-time Monitoring, High-performance Analysis, and Early Warning

TRIDEC Cloud – Timeline

• Tracing of information and activities related to an initial event sorted by time– Earthquake events with first information and follow-up refinements

– Simulations automatically or manually processed for earthquake event updates

– Warning messages sent

• http://youtu.be/Q5Znr61qUEw

Page 9: Web-based, GPU-accelerated, Geospatial Platform for Real-time Monitoring, High-performance Analysis, and Early Warning

TRIDEC Cloud – Message Dissemination

• Generation of warning messages with static and dynamic information based on instant computation of a tsunami simulation

• Review, modification, and addressing of generated messages in one form

• Dissemination via– E-mail

– Fax

– FTP (and thus GTS)

– SMS

– Cloud Messages

– Shared Maps

• Messages follow the format defined by ICG/NEAMTWS (initial and ongoing information-watch-advisory messages and messages for tsunami end or cancellation)

• http://youtu.be/ByVpGh40Tqg (mail), http://youtu.be/pGLyoQ5M6HA (fax), and http://youtu.be/bEIq3HY4gI8 (FTP/GTS and SMS)

Page 10: Web-based, GPU-accelerated, Geospatial Platform for Real-time Monitoring, High-performance Analysis, and Early Warning

TRIDEC Cloud – Cloud Messages

• Cloud messages are sent internally within the TRIDEC Cloud to other users

• Complementary to e-mail, fax, FTP/GTS, and SMS

• Contain text information and integrate interactive and dynamic event data and simulation data

• Recipients are enabled to interact dynamically with the map and diagrams beyond traditional text information

• http://youtu.be/l8XJpql_tLM

Page 11: Web-based, GPU-accelerated, Geospatial Platform for Real-time Monitoring, High-performance Analysis, and Early Warning

TRIDEC Cloud – Shared Maps

• Maps and related data are shared with others by using URL addresses– Either to a list of closed recipients via e-mail,

– Or to the public via twitter, Google Plus, and Facebook

• Recipients open the web address of the map in a web browser and interact immediately with the map

• Works even if the recipient is not a TRIDEC Cloud user

• If the recipient is a TRIDEC Cloud shared maps can be copied into the recipient's user account– Users are enabled to fully interact with shared maps as they know it from own events

– Modify and reprocess the shared map or use it for warning message dissemination

• http://youtu.be/3QhxC0PBTkg

Page 12: Web-based, GPU-accelerated, Geospatial Platform for Real-time Monitoring, High-performance Analysis, and Early Warning

TRIDEC Cloud – Monitoring and Training

• Monitoring mode– Provides important functionality required to act in a real event (see above)

– Integrates historic and real- time sea level data from http://ioc-sealevelmonitoring.organd latest earthquake information from http://geofon.gfz-potsdam.de

– Simple REST API enables integration of other systems

• Exercise and Training mode– Enables training and exercises with virtual scenarios

– Integrates Virtual Scenario Player

– Corresponds to Monitoring mode

– Operators and others involved are able to train skills and prepare for large exercises and real events

Page 13: Web-based, GPU-accelerated, Geospatial Platform for Real-time Monitoring, High-performance Analysis, and Early Warning

Exercise Runsin the Fringe of NEAMWave14

http://ioc-unesco.org/index.php?option=com_oe&task=viewDocumentRecord&docID=14357

Page 14: Web-based, GPU-accelerated, Geospatial Platform for Real-time Monitoring, High-performance Analysis, and Early Warning

Background

• Opportunity provided by NEAMWave14 to test the TRIDEC Cloud

• Collaborative activity based on previous partnership and commitments at the European scale– GFZ German Research Centre for Geosciences (GFZ)

– Kandilli Observatory and Earthquake Research Institute (KOERI)

– Portuguese Institute for the Sea and Atmosphere (IPMA)

• Not officially involved in Part B of the Turkish, Portuguese, and Greek NEAMWave14 scenarios

• Official scenarios have been used by GFZ, KOERI, and IPMA for testing with extended and modified versions in exercise runs in the fringe of NEAMWave14 on– Monday, October 27, 2014 (14:30-17:30 UTC) with the Turkish scenario

– Tuesday, October 28, 2014 (8:00-12:00 UTC) with the Portuguese scenario

• Official scenario has been used by GFZ only on– Wednesday, October 29, 2014 (13:00-16:00 UTC) with Greek scenario

Page 15: Web-based, GPU-accelerated, Geospatial Platform for Real-time Monitoring, High-performance Analysis, and Early Warning

KOERI scenario: Oct. 27, 14:30-17:30 UTC

Page 16: Web-based, GPU-accelerated, Geospatial Platform for Real-time Monitoring, High-performance Analysis, and Early Warning

KOERI scenario: Oct. 27, 14:30-17:30 UTC

• TRIDEC Cloud operated by GFZ staff

• Messages sent to restricted group of recipients incl. KOERI, IPMA, and GFZ via E-mail, Fax, FTP (no GTS), SMS, Cloud Messages, Shared Maps

• Issues identified (have been fixed after exercise run)– Operator missed manual analysis and editing of ETA for measurements

– Cloud messages 001 and 002 have not been send to the cloud account of one recipient due to the usage of a wrong username / mail address – fixed directly by the operator for cloud messages 003-005

– Message content for ongoing, and end messages included text intended for first messages

– Sorting of TFP list not implemented properly

Page 17: Web-based, GPU-accelerated, Geospatial Platform for Real-time Monitoring, High-performance Analysis, and Early Warning

KOERI scenario – modified MSEL

Page 18: Web-based, GPU-accelerated, Geospatial Platform for Real-time Monitoring, High-performance Analysis, and Early Warning

KOERI scenario – overview and analysis

Page 19: Web-based, GPU-accelerated, Geospatial Platform for Real-time Monitoring, High-performance Analysis, and Early Warning

IPMA scenario: Oct. 28, 8:00-12:00 UTC

Page 20: Web-based, GPU-accelerated, Geospatial Platform for Real-time Monitoring, High-performance Analysis, and Early Warning

IPMA scenario: Oct. 28, 8:00-12:00 UTC

• TRIDEC Cloud operated by GFZ staff

• Messages sent to restricted group of recipients incl. IPMA, KOERI, and GFZ via E-mail, Fax, FTP (no GTS), SMS, Cloud Messages, Shared Maps

• Issues identified (have been fixed after exercise run)– Sorting of TFP list not implemented properly

Page 21: Web-based, GPU-accelerated, Geospatial Platform for Real-time Monitoring, High-performance Analysis, and Early Warning

IPMA scenario cont’d – modified MSEL

Page 22: Web-based, GPU-accelerated, Geospatial Platform for Real-time Monitoring, High-performance Analysis, and Early Warning

IPMA scenario – overview and analysis

Page 23: Web-based, GPU-accelerated, Geospatial Platform for Real-time Monitoring, High-performance Analysis, and Early Warning

NOA scenario: Oct. 29, 13:00-16:00 UTC

Page 24: Web-based, GPU-accelerated, Geospatial Platform for Real-time Monitoring, High-performance Analysis, and Early Warning

NOA scenario: Oct. 29, 13:00-16:00 UTC

• TRIDEC Cloud operated by GFZ staff

• Messages sent to restricted group of GFZ recipients via E-mail, Fax, FTP (no GTS), SMS, Cloud Messages, Shared Maps

• So far, no issues identified

Page 25: Web-based, GPU-accelerated, Geospatial Platform for Real-time Monitoring, High-performance Analysis, and Early Warning

NOA scenario cont’d – modified MSEL

Page 26: Web-based, GPU-accelerated, Geospatial Platform for Real-time Monitoring, High-performance Analysis, and Early Warning

NOA scenario – overview and analysis

• Involvement of GFZ only, without any other participating site– Messages have been delivered within GFZ network timely and thus

– Analysis of the performance is skipped

Page 27: Web-based, GPU-accelerated, Geospatial Platform for Real-time Monitoring, High-performance Analysis, and Early Warning

Outlook

Page 28: Web-based, GPU-accelerated, Geospatial Platform for Real-time Monitoring, High-performance Analysis, and Early Warning

Activities and future prospects

• Activities– Support institutional tasks and responsibilities

– Contribute to future projects

– Enable collaboration across domains

– Supply commercial services

• Areas– Monitoring, analysis, and early warning for

• Natural hazards such as tsunami, earthquakes volcanic activities,cyclones, flash floods, floods, …

• Spread of pollutants, smog and carbon oxides

• …

– Analysis of disaster risk, exposure, vulnerability, and resilience

– Scientific Computing in Research Infrastructures and Virtual Research Environments