Workgroup A

---

This group will meet in the Hoyle Building's Sackler Lecture Theatre.

Portal to use:

http://ag02.ast.cam.ac.uk:8081/astrogrid-portal/

If you have problems connecting to the above URL, try the following address:

http://131.111.69.195:8081/astrogrid-portal/

Notes of Activities

Initially planned ...

Science Cases to be dealt with:

1. Get multiple stp datasets from portal (via CEA applications)

Generally "getting data from different places" - Kp, 10.7, Bx,y,z, Dst...

2. Reduce to a common stp format, this can be passed to a model, such as the CTIP model, or to a plotting application

3. CTIP model as CEA app

4. Plotting tool or javaws app that gives an overview of parameters: "Quicklook plots"

Issues

• different timescales
• missing data

Actual development leads to ....

Step 1

Get data files. Initially thought we might make up "mock data" since we might not be able to access all required datasets. Decided in the end not to use mock data - access RAL/WDC instead and use the data we down-load from there. Wrote CEA application that runs a Perl script to get data files Called WDC-dataget. It retrieves Kp (3 hour), Dst (1-hour), the F10.7 cm flux (1day) - or Bx, By, Bz (1-hour). One parameter called at a time using the same script (GetWDCdata ). Can be extended to any geophysical parameter held in the WDC

Step 2

Originally "convert data to 'common stp format' " Not done Instead a new pair of workflows written - the first inputs start time and end time and calls the application in step 1 to retrieve Dst and Kp and put them into MySpace. Workflow 2 will take these parameters and produce plot files of them in MySpace. See QuicklookGenerator.

Step 3

wrote a script to run CTIP on the RAL machine. Using fixed inputs for now. Input file can be specified. Programme will produce ten standard output files. These output files can be used by the visualisation script below. Run CTIP script in CEA - application called "CTip"

See below about producing a movie from these output files.

Step 4

This has now become two visualisation workflows - one (see Step 2 above) uses GnuPlot to plot out the WDC retrieved parameters. The other visualisation application we can now call step 5.

Step 5

CTIP visualisation - this is a script running under CEA on the MSSL site. This is a movie maker (called ctip-plot) - which takes one of the standard output files (see Step 1) from CTIP and generates 24 .png files (one per hour) and then runs these through Movie-maker to produce a movie of 24 hours of results. Plotting routine runs in idl.

Outcomes

CTIP as an application (limited in scope depending on input file) Gnu-plot workflow which plots Kp and Dst from WDC database Perl script which calls RAL WDC database to return STP parameters (limited to specific parameters now but can be genertalised) Plotting application for CTIP output files - can do any one at a time of the ten produced by the CTIP application

Key Points for Future Work

Can put the parameter fetch together with the CTIP application in one workflow so that the parameters supplied in the fixed input file to CTIP are generated by the perl script separately. As a long-term goal this can be generalised even further and other boundary condition/paremeter-set inputs can be produced to feed into CTIP. Also CTIP to be replaced by CMAT2 in medium term.

What to do with the Gnu-Plot? This might be generalisable to a more general purpose tool that would be useful to the STP community generally

group A comments

* good *workflow is a good idea *not having to reinvent the wheel *interoperability between data and applications within one environment

* bad * need to improve reliability and speed * error reporting poor

* future*

• Deploy datasets of UKSSDC via DSA
• Complete CEA installation at RAL so CTIP can be run as a CEA ap there
• Fix generic movie maker
• Join all steps together into 2 workflows: "Workflow 1" retrieves UKSSDC data, calls plotting program with sliding window, generates mpeg movie. "Workflow 2" retrieves UKSSDC data, runs CTIP model, runs CTIPmovie application to generate a movie of output data.

Resources

• Science cases

People

• AlanAylward, MartinGrill, MatthewHarris, PeterKirsch, SteveMarple, GeorgeMillward
• ElizabethAuden, KevinBenson, SilviaDalla

---

-- NicholasWalton - 10 Jul 2005

• demo_parallel.wf: Workflow to download and plot Kp & Dst in parallel