Changes Between Versions 8a and 8b

[The following information is not intended to be understandable without referring to the corresponding source files.]


Summary of Major Changes

[The changes summarized here are expected to affect the final velocity profiles. Changed default values are not included here.]


Tweakable Parameters

p.xmv_min(loadrdi.m) replaces p.xmc_min; has to do with low-battery warning.
p.saveplot_png(resolve.m) new parameter, controlling generation of png figures, in addition to PostScript.
ps.std_weight(getinv.m) new parameter, selecting the way of weighing velocity data; NEW METHOD ENABLED BY DEFAULT; see documentation/description.txt for some comments.
ps.btrk_weight_nblen(getinv.m) new parameter allowing to weigh BT data as function of distance off bottom; weighing disabled by default.
ps.up_dn_looker(getinv.m, getshear2.m, plotinv.m) new parameter allowing exclusion of uplooker or downlooker data during processing.
p.btrk_below(getbtrack.m) defaults to 0.5; center BT-velocity search this many bins below target-strength max.
p.btrk_range(getbtrack.m) changed semantics; in V8a this gives the max allowed range for BT detection; in V8b, this gives max/min; defaults to [300 50].
p.btrk_wlim(getbtrack.m) max allowed difference between reference layer w and BT w; defaults to 0.05m/s.
p.ts_att*(getbtrack.m) relationship btw. echo amplitude and relative target strength (in db/m); default is frequency dependent.
ps.dragfac(getinv.m) default changed back to 0 (as in V7b) from tanh(p.maxdepth/2000)*0.2.
ps.smoofac(getinv.m) default changed from 0.01 to 0.
ps.smallfac(getinv.m) changed semantics and format (V8a was number, V8b is 2x1 vector); default changed from 0 to heuristically set using water depth.
p.wlim(loadrdi.m) changed default from 0.08 to 0.2.
p.weighbin1(loadrdi.m) changed default from 0.5 to 1.
p.deviat(prepinv.m) removed.
p.superens_std_min(prepinv.m) added.
p.nav_error(getinv.m) changed default from 20 to 30.
p.ctd_maxlag(loadctd.m) changed default from 100 to 150.
p.ctd_maxlagnp(loadctd.m) how many lag chunks to try (defaults to 600).
att.tim_hour.long_name
att.tim_hour.unitsnew attributes for archival format (mainly CDF).


Files

Deleted Files

apptilt.mapply tilt-sensor offset; called from misc/gettilt.m (still there there V8b); not used for standard processing.
deviat.mrotate velocities; called from misc/ladeviat2.m, misc/ladeviat3.m, prepinv.m; not used in V8b any more (but misc routines are still there).

Added Files

documentation/description.txtdocumentation of i) major changes between V7 and V8b; ii) outline of processing steps; iii) description of ProcFigs.
battery.mLow-battery warning; experimental (two LDEO instruments hardcoded); called from resolv.m.
besttlag.mnew time-lagging routine to merge CTD with LADCP data; called from loadctd.m; uses old routine bestlag.m.
checkbtrk.mcompare RDI to post-processed bottom track; called from resolv.m; generate ProcFig.13.
checkbtrk2.malternative to checkbtrk.m; not used in standard processing; generate ProcFigs.13 and 14.
checkinv.mcheck inversion for consistency (errors vs. noise); called from resolv.m; generate ProcFig.12.
dragmodel.mbeginning of an experimental wire drag model; NB: THIS IS NOT THE DRAGMODEL THAT IS ENABLED BY SETTING p.dragfac! not called from anywhere.
fixcompass.mapply compass corrections; called from laproc.m; disabled by default (I think).
pcolorn.mreplacement for pcolor using imagesc; called from geterr.m, plotraw.m.
switchbeams.mcorrect for switched beams; approximate only; called from laproc.m; disabled by default.


Modified Files

bestlag.mfind temporal offset btw. 2 time series; numerous changes, presumably to make it more robust; called from loadrdi.m, besttlag.m and various CTD loading routines.
clearallbut.mauxilary routine; cosmetic change; called from mlocal/batchprocess/ladcpbatch.m, i.e. only used with Visbeck's batch processing.
cruise.mused for netCDF/html generation; cosmetic change.
default.mset some parameter defaults, comments regarding some more; version number change; many comments regarding parameters (these may or may not be correct reflections of changes elsewhere!); some parameter changes (see above).
getbtrack.mcalculate post-processed bottom-track velocities; essentially re-written; tweakable BT parameters changed.
getdpth.mcalculate depth by integrating vertical velocities; instead of setting velocities below seafloor to NaN, as in V8a, now the velocities are left unchanged but the corresponding weights are set to NaN. NB: DISABLED BY DEFAULT (getdpthi.m is used).
getdpthi.mreplacement for getdpth.m using inversion; extensive changes, all related to bottom-track handling.
geterr.mcreate ProcFig. 3; extensive changes, some cosmetic, but many having to do with vertical velocity (w_oce), which does not appear to be used anywhere else?
getfile.mheuristic to find files; fairly extensive changes, presumably to make it more robust.
getinv.mmain inversion solver; extensive changes, the most important (I think) being: 1) warn on missing magnetic deviation; 2) allow processing of single-head data; 3) allow weighing of BT data as a function of distance above bottom; 4) implemented 2nd method for weighing data selected using ps.std_weight; 5) extenisive changes to drag model; 6) smoothing changed, unclear why, unclear how; NB: V8a solution cannot be recovered by resetting ps.smoofac. 7) extensive changes to implementation of low-mode constraint (ps.smallfac); NB: V8a solution cannot be recovered by resetting ps.smallfac.
getkzprof.mexperimental diffusivity calculation; cosmetic changes.
getserial.mget ADCP head serial number; change function name in definition from getserial to getadcpserial; HOWEVER, since m-file is still called getserial.m, the function must be called using the old name!!! (in laproc.m); add support for lookup table adcpserial.tab.
getshear2.mcalculate shear solution; add support for single-head processing (ps.up_dn_looker); save w from shear method; re-scale errors (1/sqrt(2)).
htmltable.mbuild html table; additional figures; only used for Visbeck batch processing and in cruise.m.
lanarrow.mremove least consistent 1% of data for 2nd pass of processing; called from presolve.m; cosmetic changes.
laproc.mmain processing routine; restart run with previously saved parameters; add software version to log file; add support for switched beams, compass correction, serial numbers; change handling of RDI BT firmware bug (do it earlier); some cosmetic changes.
loadrdi.mload raw ADCP data files; added UNUSED parameter p.ts_signal_min; added code to allow both p.poss and p.drot (log file shows if the two disagree); changed some param defaults; added support for serial numbers; saved additional data (e.g. changed calculation of single-ping errors (old was wrong!); change warning criterion for large horizontal velocities; save start time; add low-voltage battery warning; delete ensembles with near-zero shear (suspected non-pinging); removed code to calculate post-processed bottom track; changed time-lagging between instruments; cosmetic changes.
localmax2.mfit parabola to data and find max; used in getbtrack.m; fairly extensive re-write.
num3str.mnumber to string conversion with formatting; change looks like a minor bug fix; COULD BE REPLACED BY sprintf()!!!
openwindows.mopen all windows, filling screen; added window.
outlier.mremove outliers; used in loadrdi.m, prepinv.m; added warnings when all or most data would be removed.
plotinv.mcreate ProcFig. 1; added contents to plot.
plotraw.mcreate ProcFig. 2; added contents to plot.
prepinv.mcalculate super-ensembles; changed code discarding bad BT data (2 places); add warnings on no magnetic deviation; change in code to calculate heading offset from velocities; code cosmetics; remove support for p.deviat; remove superensembles with stddeviations < p.superens_std_min.
presolve.mmain processing between laproc.m and resolve.m; added comments & debug statements.
rditype.mreturn ADCP header data; changed units of Cell_length and Pulse_length.
resolve.mfinal processing; added support for checkinv.m, checkbtrk.m; added comments & degub statements; changed ProcFig. 11 (warnings & battery voltage); changed figure writing.
savearch.msave data after processing; moved p.nav_error default elsewhere; save addl info related to battery and estimated bottom-track bias.
stdnan.mcalculate stddev, allowing for NaNs; changed semantics: new version returns NaN (instead of 0) if all input data are ~finite.
loadctd.mload CTD time series; implement more robust data lagging (with new params); changed bottom-time warning criterion; new start-time calculation; changed ProcFig. 4.
loadctdprof.mload CTD profile; cosmetic change.
loadsadcp.mload SADCP data; cosmetic change.
demo.m,demo4.mprocess example cast; remove magnetic-deviation calculation (is done somewhere else now).


© 2005 A.M. Thurnherr (e-mail)created: Tue Aug 16 11:55:59 2005 modified: Wed Aug 17 15:22:54 2005