The Simularium, Baskin School of Engineering
University of California, Santa Cruz, 12-16 July, 2010
The ROMS 4D-Var Workshop is rapidly approaching, and is in the advanced stages of planning. There is a lot of information in this announcement and it is important that you attend to everything mentioned here. There is a lot of information on the
workshop webpage and new information is posted there often so please check back regularly.We need a final head count for the caterers and other logistics. Please confirm your attendance as soon as possible. If you have not confirmed by Friday, July 2, we will assume you are not coming. To confirm, click
here (you will be asked to log in with your ROMS credentials).The workshop
agenda is now available for your perusal. In addition, all of the workshop exercises are now available for you to download from the ROMS repository. To download use the following command:svn checkout --username joe_roms https://www.myroms.org/svn/src/test MyDir
where MyDir is the destination directory on your local computer. It will be created if not found. If your username on your local computer is not the same as your ROMS username you will need to pass the --username option to svn. See the following
post for more informantion about this repository.
Please download all of the workshop exercises well ahead of arriving in Santa Cruz.To ensure that everything will work correctly for you, we encourage you before arriving in Santa Cruz to try running the first hands-on exercise. You can find this in the directory WC13/I4DVAR. Follow the instructions in the WC13/I4DVAR/Readme file. You should download the latest version of the ROMS code (revision 478 or higher) before doing this. If you have not already done so, you will need to build the ARPACK and PARPACK libraries before running any of the hands-on exercises. You will find instructions for building these libraries in ROMS/Lib/ARPACK/README. You can test the Matlab plotting script for first exercise by running WC13/plotting/plot_i4dvar_cost.m
We have posted the lecture notes:
- Lecture 1: PDF Primal 4D-Var
- Lecture 2: PDF The Mechanics of 4D-Var
- Lecture 3: PDF Dual 4D-Var
- Lecture 4: PDF 4D-Var Diagnostics
- Lecture 5: PDF Observation Impact and Observation Sensitivity
tutorials and exercise instructions will be posted in WikiROMS prior to the workshop so please check back often if you wish to download these before you arrive.Important Reminders:
To ensure that everything runs smoothly for you during the workshop, please attend to the following tasks before you arrive in Santa Cruz. The workshop schedule will be quite full so there will be limited time for trouble shooting if you have any configuration issues.
- Please configure a low resolution ROMS application for your own ROMS 4D-Var experiments that you will run at the workshop. 4D-Var is computationally very intensive, so starting with a low resolution version of your application is crucial in order to get the most out of the tutorials.
- Run your model application for a period that is long enough to compute meaningful circulation statistics (i.e. means and standard deviations). Save all model prognostic variables (T, S, u, v and zeta) frequently, and save all of the model history files - you will need these during the workshop for configuring 4D-Var.
- Identify a one week (7 day) period that will be the focus of your own 4D-Var configuration, and ensure that you have the surface forcing, and appropriate open boundary conditions for this period.
- Gather the available observational data for the 7 day period identified in above.
- Observational data needs to be averaged and binned in space and time as appropriate for your model grid: there should be no more than one observation per grid cell during any single time step. For example, if you have glider observations of temperature sampled every 5 minutes, these observations need to be appropriately averaged in space and time so that no more than one observation will be assimilated into a given grid cell during a single time step. The observations processed this way will be identified as super observations during 4D-Var.
- All of your processed observations must be in a format that can be read by Matlab.
- Please plan on using combinations of only satellite SST, SSH and hydrographic (T, S) during the workshop. (Other data types can be assimilated into ROMS, but present additional challenges which we will discuss during the workshop).
- Some recommended data sources include:
- SST: GHRSST OpenDAP server.
- SSH: AVISO gridded maps of absolute dynamic topography.
- Quality controlled global hydrographic observations from the UK Met Office.
- SST:
- We will post soon several scripts that will help with the processing of the satellite observations.
- We will shortly send you instructions to ensure that you can use the Matlab plotting scripts that are included in the workshop exercises.
- Please download the latest versions of the ROMS source code (revision 478 or higher) and the Matlab tools from the ROMS svn repository. These should be used for all of the hands-on activities at the workshop. Also please build the ARPACK and PARPACK libraries before running any of the hands-on exercises. You will find instructions for building these libraries in ROMS/Lib/ARPACK/README.
- Please bring a laptop computer with wireless capability and an Ethernet cable. Also, please ensure that you have System Administrator privileges for your laptop in case any last minute issues arise when configuring your computer for the local network.
- Don't forget - No local computing resources will be available, so you must be able to run all of the exercises and your own ROMS applications remotely on your own computer systems. For some of you these computer systems will be overseas, and/or may be protected by firewalls, so you should discuss remote access and security issues with your system administrators ahead of time.
- There will be a $120 registration fee: cash or cheque (payable to UC Regents) in U.S. dollars only. There are ATM machines on campus.
- The workshop will adjourn at 12:30 P.M. on Friday, 16 July.
Andy Moore (ammoore@ucsc.edu)
Hernan Arango (arango@marine.rutgers.edu)