|
HMI Ring Diagrams
Pipeline Notes
|
HMI Pipeline Notes: Ring Diagrams
Test Pipeline I
This pipeline was a first prototype for an end-to-end test using prototype
code on MDI data at MDI resolution, but with the output ingested in the DRMS.
It was specially tailored to track at multiple rates in order to address
questiosn about the sensitivity of ring fitting procedures to the data
tracking rates.
Data
The MDI full-disc Doppler data (lev 1.5 with corrections) tracked over the
standard 189-region "dense-pack" targets every "day" (15° Carrington
rotation) for 1664 minutes during the full Carrington rotation 1988. The
entire series was tracked at five different rigid rotation rates, in a range
exceeding the range of differential rotation rates in various tracers over
the latitude range analyzed.
Script:
~rick/mdi/rings/densepack
Modules:
mtrack_v09, pspec3_v08, rdfitc_v09, rdvinv_v03
Test Pipeline II
This pipeline is to test the MDI image distortion correction in mtrack.
Although this is not essential to the HMI pipelines, it provides a good
opportunity to run several of the pipeline modules together over a particular
set of targets. It also provides a first model for future pipeline scripts.
Data
The MDI full-disc Doppler data (lev 1.8) "tracked" for 1664 minutes centered on
1988:165.0 (2002.04.13_19:37:00_TAI). Only the 40 regions lying around the
perimeter of the dense pack, plus the one at the equator and central meridian
ae selected. Tracking is turned off, in order to keep each region as nearly
as possible in the same part of the image field, with a fixed distortion.
Script:
~rick/hmi/rings/scripts/smplcfmdidist
Modules:
mtrack_v08, pspec3_v07, rdfitc_v07
Problems / Issues
- The prime keys in the power spectra are in different order from those
in the tracked cubes. Probably a problem in the JSD's.
- Some important keys in the power spectra are not being set, in particular
MAPSCALE and MAP_PROJ.
- rdfitc dies with a segmentation fault when run with the -n flag;
this is because fp0 is not initialized the first time rk0[fp0 + row] and
rku[fp0 + row] are printed.
- The ring-fit module does not use proper values of ux for initial guesses
for untracked data. The ux and uy parameters for initial guesses were found
to be improperly handled in the current version of rdfitc; this resulted
in failures to fit the untracked data at latitudes below about 52.5 deg.
Timings
rdfitc: Without attempting to provide correct initial guesses for ux,
individual runs of rdfitc took the following clock times (sec):
n02 (3) maelstrom (7)
distortion corrected 3027-3155 1528-1819
uncorrected 2640-2906 1629-1790
Test Pipeline III
Script:
~rick/mdi/rings/dohr45
Modules:
mtrack_v09, pspec3_v08, rdfitc_v09