Last Update: 09.01.2007


 *To do tracking of level 1.8 MDI data
 *This was first required to do exactly the same tracking fro already 
tracked Doppler tiles. So this process first reads the 'fastrack.*.map' 
files, describing how the tracking was done, from given directories, of 
already fastracked data at MPS, and extracts the details required to
 create 'mapfiles'.

MPS  /data/seismo1/schunker/STANFORD/

unix> source ~CM/script/soi_exe.setup

unix> emacs push_script -> make sure the arguments DIR and OUT are as you want:
			-> DIR: where to look for files to read
			-> OUT: where to put the mapfiles
			-> check the "echo" commands at the bottom that they
			   contain the correct directories for Stanford output

          		-> OUTPUT: $OUT/mapfile*



unix> scp $OUT/mapfile* schunker@solarport.stanford.edu:/scr20/schunker/somewhere/here 
			->Copy the mapfiles to Stanford



unix> ssh gizon@sonar.stanford.edu
			-> go to Stanford to run the fastrack

*Now we have all the mapfiles to run fastrack.c at Stanford
STANFORD
unix> cd /scr20/schunker/somewhere/here/

unix> emacs /scr20/schunker/somewhere/here/track_script
			-> find track_script and make sure it searches in the 
			   correct directory for the mapfiles

unix> track_script
			-> run track_script, does the fastracking using
			-> every mapfile in the directory, takes about
			-> 48 hours to do full set of data, one every minute.

			-> OUTPUT: tracked datacubes in a specified directory

What WE get fastracking to do: 
DESCRIPTION

     fastrack extracts from a series  of  solar  images  multiple
     regions  fixed  in  co-moving  frames  at a set of initially
     specified heliographic coordinates.  
     (specified by the user by giving heliogrpahic latitude and 
     longitude of central point as well as size of the region).

     The tracked regions are
     of  a  common specified size, and the same map projection is
     used, only the central  point  varying 
     (which happens to track the longitude and latitude.)   

     Depending  on  the
     nature  of the tracking, the heliographic coordinates of the
     map centers will be continuously adjusted  as  functions  of
     time,  matching the specified coordinates only at the start.
     If a tracking matching the Carrington rate is selected,  the
     heliographic  coordinates  of  all  map  locations  will, by
     definition, remain fixed.

     If images are missing from the input data series  or  other-
     wise  rejected,  fastrack  will linearly interpolate in time
     across the gaps in the mapped output.   (Missing  images  at
     the  end  points  result in continuation backward or forward
     from the first or last valid map.)  Partial input images may
     produce missing values in the mapped images.


!!!!!!SO WHY DO MY FASTRACKED CUBES HAVE ZERO IMAGES??? !!!!!!!

     The input data series is normally assumned to be continuous,
     so  if  an  image selection range is made (...,series:XXX[I-
     J],sel:[M-N]) the first selector (M)is assumed to apply only
     to  the first data set (I) and the last selector (N)to apply
     only to the last data set (J); all  intervening  images  are
     accepted.   The  -c flag forces the image selection to apply
     independently to each data set in the series.

     The images are tracked at the Carrington rate. The coefficients 
     of the sin^2(latitude) expansion of rotation minus Carrington
     rotation (urad/sec) are specified. Default values are from
     Snodgrass 82/84, but these can be changed and some alternate
     options are given in the comments of fastrack.c
     Snodgrass values are:
     a0 Float -0.02893
     a2 Float -0.3441
     a4 Float -0.5037

     THE PARAMETERS
     in = input files, the dataset is assumed to include in its 
          ancillary data the parameters required for  heliographic  
          mapping,  namely the  observation  time  and  heliographic 
          location.
     out = output files
     bckgn = we DO NOT subtract an average image (necessary only for 
             Doppler calculations
     lat = latitude of centre of region for tracking
     lon = longitude (or list of) for the centre of region for tracking
     map = "Postels" although there are many options for mapping, again
           refer to the comments of fastrack.c
     scale = 0.125, which is the scaling of the pixels (degrees/pixel)
	     at a location appropriate for the selected  mapping  
     	     option;  a  0.0   value implies  autoscaling  to  best  
             scale of image. 
     cols = 128. (width of image in pixels)  but these are easily changed
     rows = 128. (height of image in pixels) but these are easily changed
     map_pa = 0.000000, used to change the direction of heliographic
	    North from vertical in the output image.
     radius = 0.000000
     a0 = -0.028930 
     a2 = -0.344100 
     a4 = -0.503700
     merid_v = 0.000000, is the meridional velocity
     xc = nan0xf0000000 
     yc = nan0xf0000000 








     OTHER TRACKING OPTIONS
     The -l flag sets a "lagrangian" tracking option, forcing the
     tracking  to  be at the differential rotation appropriate to
     the instantaneous latitude for each  point  in  the  region,
     rather than at a common rate for the whole region.

     The -n option "turns off" tracking in  the  sense  that  the
     images  are  tracked at minus the Carrington rate.  There is
     no correction for observer motion in the latitudinal  direc-
     tion,  so  this  is  not  quite the same as mapping to fixed
     heliocentric coordinates.

     The -z flag directs the output maps to be organized in polar
     coordinates,  with  azimuthal angle the most rapidly varying
     coordinate, (columns)  and  radial  distance  from  the  map
     center  the  next (rows).  The length of the angle dimension
     in that case will not be a power of 2 unless set by the cols
     value.


     If the verbose option (-v)  is  set,  one  line  of  history
     information will be produced per input image.

     The mapped data sets are placed in one or more 3-dimensional
     FITS  files, with time (image serial number) the most slowly
     varying axis.




CALCULATE ANGLE FROM DISK CENTRE

* I have no idea why fastrack.c does not leave the OBS_L0 and OBS_B0 
parameters of raw MDI data headers in the fastracked cube headers. 
It would render soho_ephem redundant.

MPS

!!!!!!!!this is all becuase I had a lot of trouble extracting only 
the numbers - I would appreciate someone coming up with a way to grab 
the string of only the numbers!!!!!!!!!! which would then result in the 
following few steps being redundant	
 
unix> soho_ephem_script
	reads in all the files in all the directories in 
	directory /net/seismo1/export/data/seismo/gizon/tape1 
	or /net/seismo1/export/data/seismo/gizon/tape2 and
	outputs to /data/seismo1/schunker/STANFORD/soho_ephem_temp1
	filename and T_OBS (the middle time of the data cube)

unix> soho_ephem_script2
	reads in /data/seismo1/schunker/STANFORD/soho_ephem_temp1
	outputs T_OBS to /data/seismo1/schunker/STANFORD/soho_ephem_tape1


unix> soho_ephem_script3
	reads in /data/seismo1/schunker/STANFORD/soho_ephem_tape1
        and outputs only the time (not replicating others already done)
	to file /data/seismo1/schunker/STANFORD/soho_ephem_tape1_fin
	Now (finally) have a list of all the times to calculate soho_ephem for.


unix> scp /data/seismo1/schunker/STANFORD/soho_ephem_tape1_fin schunker@solarport.stanford.edu:~/.

(tape1 and tape2 are necessary)

STANFORD

unix> soho_ephem_main_script
	Takes as input /home/schunker/soho_ephem_tape2_fin
	and for each time in the file calculates "soho_ephem" and then
	outputs 
	OBS_DIST (au L0 (deg  B0 (deg VR (m/s) VN (m/s) VW (m/s) Carr Rot
	to /home/schunker/ephemeris_tape2
        


unix> scp /home/schunker/ephemeris_tape2 schunker@ssh.mps.mpg.de:/data/seismo1/schunker/STANFORD/.

MPS

unix> extract_ephemeris_time_script
	Extracts the time from /data/seismo1/schunker/STANFORD/ephemeris_tape2
 	and puts them in a list.


IDL> string_matching.pro
	Reads in /data/seismo1/schunker/STANFORD/ephemeris_tape2
	to get L0 and B0.
	Reads in /data/seismo1/schunker/STANFORD/final_dates_tape2.txt
	
	Outputs  the L0 and B0 and date based on matching the i's
	in the list of each file to
	/data/seismo1/schunker/STANFORD/tape2_date_L0_B0.dat




Now if we know the central L0, B0 and we have a map of the 
longitude and latitude we can calculate the heliocentric 
angle using 
Cosine(heliocentric angle)=Cosine(latitude-B0)*Cosine(longitude-L0)






The following is saved for posterity, but pointless:
IDL> disk_position.pro

	For all the files, in all the directories in data/seismo1/gizon/tape2/
        extracts the latitude (hg_lat) and longitude (hg_long) and T_OBS
	
	
	Reads in /data/seismo1/schunker/STANFORD/tape2_date_L0_B0.dat
	and then matches T_OBS to this list.

        !!! Note that TAI and UT time are not the same! They differ by 
	seconds, so when extracting strings, only match up to the
	hour (seconds may also change the minutes). This seems to be
	fine.

	Calculate hg_lat-b0 and  hg_lon-l0 for latitude and longitude
	respectively, and then the resulting angle to disk centre as seen
	by SOHO-MDI.


	Write to
	/data/seismo1/schunker/STANFORD/angle_tape2/angle_to_disk_centre_tape1_"+filename+".dat"
	date, hg_lat, hg_lon, l0, b0, angle to disk centre (degrees)
        

	-->