Comments and Questions

I want to join the SHINE-2005 campaign. Coronal mass ejections is the most exciting topic for me, and I want to know more about CME. I have also published some research papers. Can I join this workshop?

CCMC offers any scientist the opportunity to request simulation runs of solar coronal and heliospheric models (in addition to Earth-magnetospheric and ionospheric models). For example, the Magnetohydrodynamics Around a Sphere ("MAS") model of the solar corona (by Z. Mikic and J. Linker of SAIC) is driven by the solar surface magnetic field pattern obtained over a full 27-day rotation period of the Sun which can be selected for a simulation run among other parameters. A simulation run of this model creates a steady-state representation of the magnetic field and plasma flows under the assumption of stationary magnetic fields on the surface of the Sun.

Modeling runs of the Heliospheric Tomography model (by B. Jackson of UCSD) are based on scintillation observations from radar on Earth that determine line-of sight plasma densities and velocities in the heliosphere. The densities are then advanced in time to allow tomography by correlating with earlier and later measurements to create a three-dimensional representation of the inner heliosphere.

However, at the CCMC, no time-dependent magnetohydrodynamic simulations of solar CME events are available at this time. We have done some runs that covered time periods in the SHINE campaign with the models mentioned to create model representations of the coronal and heliospheric environment of CMEs.

In the future we plan to offer coupled coronal-heliospheric simulations (possibly of simple CME structures) with the arrival of the heliospheric magnetohydrodynamic model "ENLIL" (by D. Odstrcil) and a new version of MAS that allows time-dependent inputs at the Sun's surface and higher numerical resolution.

There is also the option to not project the data in the case that you want to perform your own time shifting.

• Select "Radius r-constant" as a "Cut Plane"and enter the maximum radius (20 or 30 R_S, depending on the run).
• Keep the "Plot Area" (ranges for Lon, Lat.) at 0 to 360 deg. and -90. to 90 deg, respectively.
• Select all three magnetic field components as plot variables "Q1", "Q2", "Q3" and keep the "Plot Mode" set at any one of the 2D plot options ("Color Contour", for example, ALL above variables will be printed regardless of wheher they are actually seen in the plot image).
• At the bottom of the CCMC plot interface (below the "Update Plot" button) check "List Data" and keep "List Data from Plot" selected.
• Click "Update Plot" to generate the plot and get the listing.
  The r,phi,theta positions ("r", "p", "t") and the respective magnetic field components "B_r", "B_"p, "B_t" should come out as rows of 6 numbers after some header information.
• Paste the listing into a file and you are in business.

March 27, 2006 - Answered
We have received developer's feedback on the possibility of lowering the default altitude. The developer does not encourage us to use values for the altitude of the base of a field line (altmin) lower than 85 km (default value). The code has not been designed/tested for such cases and many effects might not be taken into account if used with such low altitudes.

Files will contain contour line information representing the polar cap (boundary between open and closed field lines) in magnetic local time and co-latitude for each ionosphere time step.

I have a few questions about the ENLIL run.

1. What is the coordinate system used (X,Y) for the ENLIL runs?
2. Where is the Earth in this coordinate system?

March 8, 2006 - Answered
1) Coordinate system: X points to the Earth at the time listed for the output file. Z is along the ecliptic north pole and Y is perpendicular to X and Z (right-handed coordinate system).
2) Where is Earth:
For a time series plot created from a stationary run with ENLIL, one would compute the Carrington Longitude at the time of the model output (year,month,day,hour,minute shown on plots) and the longitude at the desired time (year2,month2,day2,hour2,minute2) and then take the difference: earth_longitude=carr_lon(year,month,day,hour,minute)
- carr_lon(year2,month2,day2,hour2,minute2)

Earth's position would approximately be (R=Sun-Earth Distance =~1 AU): X~R*cos(earth_longitude) Y~R*sin(earth_longitude). The Earth "moves" about -13 degrees per day (-360/27.2).

By looking at the shifts in the flow direction, we are wondering if we can get the orientation of the stream interface. We were hoping that the ENLIL run would be useful to test out our data-analysis methods, but it seems that the resolution of the run is not high enough: the stream interface is quite rippled in the simulation.

At the present highest resolution (1 degree) the grid cell size at Earth will still be comparable to the size of the entire magnetosphere and the data extracted from the ENLIL data will only vary on a time scale of hours (Earth moves 13 degrees in Carrington Longitude per day; a 1-degree grid cell is crossed in about 2 hours).

1. For the same events, with similar initial conditions two codes provide very different outputs: much less details (1-3 Re scale structures in the current sheet of the kind responsible for flapping) came from my BATSRUS run, as contrasted to latest OPEN model run which displays a lot of structure.

   What is the reason of so different results concerning the mesoscale structures? Much lower grid resolution in BATSRUS? Where all standard settings can be learned?

2. Two questions specific for OPEN(UCLA) code run, which I made the last (and possibly was not prepared to face with big differences between the request orders for two models)

   Can the simulation time series at specified position (specified spacecraft) be generated afterwards, after the completion of a run ? (it seems I did not find this option when making a run request, although it existed for BATSRUS??). What would you recommend me to do in this situation? I think the comparison of time series will be best way of starting comparisons with observations..

   Can the results be displayed in GSM coordinates instead of GSE with the standard tools? (It seems I could miss to mark the button, if it existed. Having GSM output I would easily visualize the plasma transport near the neutral sheet, which is difficult to do in GSE where I can not specify the tilted plane to make 2D plots with existing tool)

3. What could be approximate volume of my run outputs (4hours) if I'll decide to ask them for doing processing at home?

1. OpenGGCM and BATSRUS may provide different results, especially in the tail. The comparison that you are doing is very valuable for the purpose of model testing and validation. I noticed you requested high resolution run for OPen GGCM while we ran low resolution run for BATSRUS (it is a default for runs longer than 2 hours). I generated a grid for BATSRUS with similar resolution as OpenGGCM in the tail region (0.25 RE)and am rerunning your request. BATSRUS is arguably more diffusive code than OpenGGCM and can contribute to the difference. Another issue is resistivity. Reconnection in BATSRUS is supported by numerical resistivity only, while OpenGGCM adds current dependent anomalous resistivity that can increase reconnection rates in the tail.
2. At this time OpenGGCM is not writing output along satellite trajectory during simulations. We plan to work with the developer to add this feature. At the present time it is possible to generate time series along the cluster trajectory with output frequency equal to frequency of global output (2 minutes in your case). This tool is not yet available on-line. We will generate output along the Cluster trajectory for you and mail you the results.
3. At the present time you can visualize the results on-line using the model native grid (GSE for the OpenGGCM case). We are working to allow plots on oblique planes, however this feature is not yet available.

4. We can put model output in CDF format at our anonymous FTP site and send you interpolation libraries that will allow you to read model output at any specified position. OpenGGCM output is about 75 MB for each time step. Increased resolution BATSRUS that we are running right now will be about 150 MB each time step. So a 4 hour run with 2 minute output frequency requires 10 - 20 GB of disk space. Please let us know if you select to download model output to do processing at home. You may request the entire run or selected time steps.

Could you please comment why this is? It is strange because in the solar wind region, they should be only straight lines until bow shock with no interruption (I would expect).

The "detached" field lines you are seeing are actually separate lines started at regular intervals along the x-axis (Z=0) and along the top and bottom margins of the plot (Z=20, Z=-20).

Now, we have a case where we need to input our own Solar Wind data. I have some questions as to how to calculate the delay time and what I should know and be careful about while inputting ACE solar wind and field data?

We have data at ACE times. In this case, how will the delay time (i.e. solar wind convection time to Earth) be taken care of in the CCMC runs and do we need to prepare ACE data at the delay calculated times?

The simplest way of doing this (for short runs with a duration of about 2 hours) is to take the average velocity and the average of the satellite's distance from 33Re and calculate the time shift.

If it's not an option, then I need the Open GGCM model results along the trajectory for my run.

How long must I wait before the automatic input parameters for the CTIP model become available please?

We are currently working on improving the submission interface to allow automatic download of Key parameter (i.e. preliminary) data for recent events (if Level 2 data are not available) and to allow users to upload SW data themselves.

Hemispheric power data are available with very little delay. ACE Level 2 data are usually available with significant delay and we are looking for ACE Key parameter data which are available very soon (almost in real time). When ACE is having serious problems, we may try to look for ACE real-time data (which they usually store for the last 2 months).

You can submit a request for the days in which you are interested. The submission interface will prepare all necessary input data files (including hemispheric power data input files) except the solar wind data file. We can add SW input data manually. We may use ACE real time data or you can send us SW data file via e-mail.

The time frame is 2003 April 8 18:00 to April 11 00:00 - this is ~ one Cluster orbit. Time resolution can be lower than normal, but spatial resolution should be no smaller than 1 RE.

Should I submit this as a special request?

If you are not satisfied with our standard grid for long runs -- please specify where (range of X,Y,Z) you would like to have resolution no smaller than 1 Re.

Vy and Vz are zero in the realtime run and are not plotted.

Most request runs have Vy=0 and Vz=0.

For completeness sake we have now replotted all request runs with varying solar wind conditions to also include Bx, Vy, and Vz in the images.

1. In "Choose quantity to be displayed", you don't explain ALL quantities which may be chosen, for example, what does "S" mean?
2. You may include additional options of parameters along a spacecraft trajectory, like |B| or |V|. I don't quite understand what people usually do when they want to see parameters along a trajectory for the GGCM.
3. Sometimes I want to come back to a run description after making several figures (for example, I make figures for "Magnetosphere", and then I want to see "Ionosphere" or "Input solar wind"). How can I easily get back to the run description (without pressing the browser's "BACK" button several times?)

1. We expanded the selection of variables recently and will update the descriptions soon. "S" is the plasma entropy ( P / N^2/3 ).
   UPDATE: November 9, 2006
   I have added new variables to the description page (MHD_variables.php).
2. This should be possible to implement in a similar fashion as the analysis of the 3D data. I'll look into that.
   OpenGGCM (version 2.1-1) didn't have a built-in interpolation method to generate data on satellite trajectories, but we can generate track data upon special request after a run has finished. We will work on activating the built-in method of data output along satellite tracks in the new model version (3.0) soon.
3. You could right-click the visualization link to open it in a separate window of your browser from the run's page.
   It's a good idea to keep the URL of the referring page in the visualization interface's data to provide a link back. We will add this feature as we constantly upgrade the interface.
   UPDATE: November 9, 2006.
The visualization interfaces (1D satellite, 2D ionosphere, 3D magnetosphere data) now feature a link "Go back to web page of run" that points back to the web page that linked to the interface. This link is provided after plots have been created.

Ionosphere data are in plotted in geomagnetic coordinates (field-aligned currents used to compute the ionospheric electrodynamics have been mapped from the coordinate system used in the magnetosphere part of the model, i.e. from 'GSM' for BATSRUS and 'GSE' for OpenGGCM). The magnetic pole is in the center and noon at the top. Data written to file show magnetic latitude and local time.

Data along flow lines are now available for flow lines with user-defined start points in 3D flow line plots.

After a plot has been made, you will see a link to a gzipped tar archive with text files that contain a listing of model output data along each flow line.

If you need computed variables, such as J_par and E, select them as Q2 and Q3 in the web interface so they get computed before the flow lines are traced.

We have added an option to download a tar archive containing model output data along flowlines with user-defined start points (see List Data from Plot section in the CCMC vis).

We have added the requested features for SAMI2.

The variables are now explained in more model-specific web pages linked from the visualization interface. We have added a IFM_quantities.html page in the case of IFM and USU-GAIM. More such pages will follow for other models.

With CCMC interpolater-v4-1.1, the BATSRUS-CDF interpolation library can now read one set of variables into memory from the CDF file, do interpolations, close and reopen the CDF file and read a different set of variables into memory (for field line tracing). This enables an IDL user to trace flow lines of the velocity, then of magnetic field lines within the same IDL session.

We have added the option to “Use combination of positions in a 3D grid” thus returning values at the expected 200 positions.

Here are the chosen options: I have checked 3D-flow lines in the Plot Mode option. Next, I have checked 3D-Flowline ... Plot VRML option (although not checking this option did not make any difference on the resulting field lines drawn). I have chosen Bx, By, Bz as quantity. (originally I have checked B only in Box-1 as Q1 quantity but this gave me an error saying that it is not a vector quantity). Lastly I have selected "predefined in cut plane" option to draw the field lines projected on y=0 plane (y=constant plane).

The resulting figure is field line figure but it is in 3-D, i.e. y is changing even though I have chosen y=0. So the result is not in 2-D as I have expected (field lines in the xz-plane in this case keeping y =constant) but still in 3-D with all the coordinates changing.

Could you please comment on how I can draw field lines on projected plane as described above or what am I doing wrong above to get the 2D- field lines in XZ-plane?

However, we can suggest a (slightly more involved) way to plot field lines projected on a 2D plane. Here is what you can do:

1. Choose Color+Vector+Flowlines as your Plot Mode.
2. In "quantity" drop down box, choose N as Q1, N as Q2 and B_y (or any B_ component).
3. Click on "update plot" button and, after the plot is drawn, check the actual ranges of N.
4. Click on "Lock color range" box and enter Min and Max ranges that are higher than the actual maximum range for N.
5. Update the plot again. This will give you the desired projection of field lines onto a 2D plane which, by default is Y=0. Put your cut plane selection in the area titled "Choose cut plane".

Some footnotes:
• if you check on the "Plot VRML" option check box, this should produce a link to a VRML file(the link will follow the words "Plot to VRML").
• Also, the "Predefined in cut plane" option works for 3d Flowlines mode; and refers to selection of STARTING positions for your Flowlines.

The speed of the visualization has been improved by remotely running IDL on a faster Linux machine. We will transfer to a (multi-processor) Linux web server that can run visualizations locally with fast turnaround
UPDATE: (this has been done in late 2005).

I had worked to implement the vector plotting (in polar plots) and the formula for the O/N2 ratio when this error occurred. Check the page with explanations of variables for information on the variable "O/N2".

At the present time you can find the results of field line tracing to the inner simulation boundary by selecting "3D-Flowlines", "only user-defined start points", and "List Data from plot" options. We will add the results of mapping and coordinate transformation from the inner simulation boundary at 3RE (GSM) to SM ionosphere (1RE) to the Web interface.

UPDATED: November 17, 2005: The plot image is now also rendered in Encapsulated Postscript (EPS), a vector graphics format, that can be used in typesetting software such as LaTex.

Is Sigma_H the height-integrated conductivity due only to the magnetospheric FAC contributions? That's what it seems from the BATSRUS description. However, in my simulation, it says that the following parameters were set:

• Ionospheric Conductance: auroral
• Radio Flux 10.7 cm: 150.

Am I looking at a different version? If it is the case, could you kindly let me know where in your web site will I be able to find the global ionospheric potential (from the pole to the equator) from the Dr. Mei-Ching Fok model on March 31 2001 and April 17 2002?

Field-alignd currents from outside that radius are mapped along magnetic dipole field lines to within 35 degrees around the poles. Electric potentials are then calculated by a potential solver using constant or modeled ionosphere conductances.

These potentials are used as input to the Fok model. The installed version of the Fok ring current model does not compute electric potentials. You may be thinking of the Comprehensive Ring Current Model (CRCM, a newer model of Mei-Ching Fok) that CCMC has not yet installed for request runs.

BATSRUS MHD output values are saved at cell center positions and are being interpolated within the adaptive mesh blocks (arrays of 6 by 6 by 6 cell positions in X,Y,Z). However as blocks extend 1/2 of a cell size beyond those center positions, values for positions near the block boundaries have to be obtained by extrapolation or some boundary condition within the block or by using neighboring blocks.

Currently, we do not consider values in the neighboring blocks which leads to the disagreements between adjacent blocks.

Moreover, extrapolations had been disabled near block boundaries (zero-gradient boundary conditions were used) because negative densities or pressures could result. We have addressed the latter problem by using extrapolations constrained by the global minimum of the selected variable found at all cell centers. This, however, does not avoid discrepancies at block boundaries but at least the constraint prevents the appearance of negative densities, pressures or vector amplitudes.

We are planning to redesign the interpolation to use data in neighboring blocks instead of using the global constraint in the near future.

The BATSRUS adaptive mesh interpolation has been upgraded to take neighboring blocks into account when interpolating within a half grid cell of block boundaries. Values should now be continuous across block boundaries.