6
|
1 |
======================================================================
|
|
2 |
R E A D M E . P R O C E S S I N G T I P S
|
|
3 |
doc: Thu Oct 20 14:11:49 2011
|
|
4 |
dlm: Fri Oct 21 17:33:21 2011
|
|
5 |
(c) 2011 A.M. Thurnherr
|
|
6 |
uE-Info: 18 0 NIL 0 0 72 3 2 4 NIL ofnI
|
|
7 |
======================================================================
|
|
8 |
|
|
9 |
=Time Lagging Problems=
|
|
10 |
|
|
11 |
If the first pass (1Hz) fails, there are no diagnostic plots that can
|
|
12 |
be inspected. Often, however, this problem is simply due to CTD
|
|
13 |
pressure spikes, which must be removed before processing. If the CTD
|
|
14 |
pressures are good, it becomes very useful to write the
|
|
15 |
LADCP_timeseries ($out_LADCPtis =
|
|
16 |
"$data_subdir/$out_basename.LADCPtis";) and compare it to the CTD time
|
|
17 |
series (averaged to 1Hz for size), e.g. to spot obvious problems or to
|
|
18 |
determine an initial offset guess.
|
|
19 |
|
|
20 |
One possible reason for the first time-lagging pass to fail is that the
|
|
21 |
beginning of the cast is not correctly recognized. In case of
|
|
22 |
2011_IWISE (003UL, but probably also others) this is because there are
|
|
23 |
valid velocities in bin#2 even before the LADCP enters the water. The
|
|
24 |
problem was solved by increasing $refLr_firstBin.
|
|
25 |
|
|
26 |
If the 2nd pass (24Hz resolution) of time lagging fails, the diagnostic
|
|
27 |
plots (*_TL.eps, *_TLhist.eps) should be inspected. Often, the problem
|
|
28 |
can simply be solved by decreasing $TL_required_top_three_fraction to
|
|
29 |
something like 0.4.
|
|
30 |
|
|
31 |
When profiles with very little ship motion (e.g. in/near the ice) are
|
|
32 |
processed $TL_required_top_three_fraction may have to be decreased to
|
|
33 |
very small values (0.05 in case of some stations in NBP0901). With such
|
|
34 |
a low value it is important that the diagnostic plots are inspected
|
|
35 |
carefully to make sure that there are no timelagging-related issues. It
|
|
36 |
is also advisable to plot/inspect the time-series output to make sure
|
|
37 |
that the final lag is reasonable.
|
|
38 |
|