An Efficient Ensemble Data Assimilation Approach and Tests with Doppler Radar Data
Efficient Data Assimilation with OpenDA
-
Upload
nilsvanvelzen -
Category
Documents
-
view
551 -
download
7
description
Transcript of Efficient Data Assimilation with OpenDA
Efficient data assimilation with OpenDAAn introduction
10-06-11
Nils van Velzen, Martin Verlaan, Stef Hummel, Julius Sumihar, Erwin Loots, Albrecht Weerts, Ghada El Serafy, Herman Gerritsen
Outline
● What is OpenDA● How to link models to OpenDA● Example of Calibration● Example of Data assimilation● Parallel computing● Ongoing developments
10-06-11 2Efficient data assimilation with OpenDA
OpenDA
• A generic toolbox for data-assimilation
– (real time) data assimilation
– model calibration
– impact observations
– reconstruction of sources
10-06-11 3Efficient data assimilation with OpenDA
OpenDA
• Content:
– set of interfaces that define interactions between components
– library of data-assimilation algorithms
– philosophy
10-06-11 4Efficient data assimilation with OpenDA
OpenDA
• Open source (LGPL)
• Why OpenDA?
– More efficient than development for each application
– Shared knowledge between applications
– Development of algorithms with e.g. universities
– Easier to test, which should result in fewer bugs
– Optimized building blocks
– Development template
10-06-11 5Efficient data assimilation with OpenDA
www.openda.org
10-06-11 6Efficient data assimilation with OpenDA
• Downloads
– Users
– Developers
• Documentation
• Wrappers for models
• OpenDA association
OpenDA Main Application
10-06-11 7Titel van powerpoint
Single program for:
Editing input files● Running simulations and calibrations● All models and observation sources● Gui and command line, sequential and parallel
Object Oriented Approach
10-06-11 8Titel van powerpoint
Models in OpenDA
• A model wrapper is needed for each model:
– Blackbox
– Native:
• Simple, only a few functions
• Full, the (whole) interface
• parallel
– Java
• Simple java models for testing and research
• Direct implementation of java interface
• Stochastic extension
10-06-11 9Efficient data assimilation with OpenDA
Black Box Wrapper
• Uses input and output files of the model
• No source code of the model is needed
• Well suited for calibration
• Sometimes suited for Data Assimilation
10-06-11 10Efficient data assimilation with OpenDA
Native Wrapping
• Implement a specified set of functions (C/Fortran)
• Create a dynamic library
10-06-11 11Efficient data assimilation with OpenDA
SWAN 3rd generation wave model
10-06-11 12Efficient data assimilation with OpenDA
Wave breaking and interaction over a bar
0.4
0.3
0.2
0.1
0
0.5m
X
Z0
1:20
10
1:10
20 30m
WAVES
STATIONS01 02 0403 0705 06 08
totS =wind input+ non-linear interactions (quadruplets & triads)+ whitecapping + bottom friction+ depth induced wave breaking
SWAN 3rd generation wave model
10-06-11 13Efficient data assimilation with OpenDA
SWAN 3rd generation wave model
10-06-11 14Efficient data assimilation with OpenDA
Operational storm surge forecasting
10-06-11 15Efficient data assimilation with OpenDA
Operational Storm-Surge Model Steady-state Kalman Filter
10-06-11 16Efficient data assimilation with OpenDA
OpenDA
10-06-11 17Efficient data assimilation with OpenDA
New operational model DCSMv6-Zuno
10-06-11 18Efficient data assimilation with OpenDA
Parallel Computing
• Model timestepping often most time consuming part of algorithm (90-99%)
• Often natural parallelism in the model timestepping
• Parallel time stepping for all native OpenDA models (OO)
10-06-11 19Efficient data assimilation with OpenDA
Parallel computing
• Parallelized using threads
• Parallelized using multiple processes
– master worker
– worker worker
10-06-11 20Efficient data assimilation with OpenDA
1x 2x 3x
4x 5x 6x
Ongoing Developments
• OpenDA core:
– New major release of OpenDA
– Development of course material
– Parallelization of the algorithms
• With OpenDA
– Operational wind induces waves forecasting with swan
– Coupling with FEWS
– Observation sensitivity and design of monitoring networks
– Lotos-Euros
– Coupling with OpenFOAM, …..etc
10-06-11 21Efficient data assimilation with OpenDA