Global options
This page describes the processing options that are used by
chilplot,
chil2a,
chil2nc and
loadchil.
Hereafter, PROGRAM is used to mean any of these
programs. The recommended values for each radar system are given at
the end.
Usage
- PROGRAM
- With no options, some basic usage information is returned.
- PROGRAM file1.dat rasterrange1 [file2.dat rasterrange2 ...] [options]
- Read the specified rasters from a set of files, where
rasterrange can be a single raster number
(e.g. 1), a range of rasters (e.g. 1-10) or a
single hyphen to indicate all rasters in the file (-).
- PROGRAM -all file1.[dat|raw] [file2.[dat|raw] ...] [options]
- Read all rasters from the files specified.
- PROGRAM -date date filepattern1.dat [filepattern2.dat ...] [options]
- Search the summary files corresponding to the files that match the
specified patterns for rasters recorded on the date specified (which
can be of the form YYYYMMDD or YYMMDD. A file
pattern can contain the wildcards * and ?.
General options
These options can be used with both scanning and cloud data. Note that
n means an integer, f means a
floating-point number and s means a string.
- -verbose
- Display information on files and rasters as they are opened.
- -quiet
- Suppress progress messages; display only warnings and errors.
- -version
- Display version and quit.
- -Zcal f
- Add f dB to all the Zh values.
- -autocal
- Take the Zh and Zdr calibration values from the
summary file. Beware that the values in the
summary files may not be correct.
- -vfold f
- Declare that the folding velocity for all velocity parameters is
f m s-1. Thus when averaging is performed,
it can be done in such a way as to account for the fact that the data
may be folded, and the mean velocity reported is accurate.
- -v1fold f
- Declare that the folding velocity for the first velocity
parameter (v1) is f m s-1.
- -v2fold f
- Declare that the folding velocity for the second velocity
parameter (v2) is f m s-1.
- -averays n
- Average groups of n rays together; the parameters
Zh and Ldr are converted into linear units before
taking the average. If the radar is measuring v in cloud mode
and averaging is performed then sigma_v, the standard
deviation of the velocity, is calculated as well.
- -avegates n
- Average every n gates. The comments for -averays
above apply here.
- -day n
- Accept data only with a day-of-the-month n.
- -sumdir s
- Look in directory s for summary files if they are not found
in the same directory as the data files. By default, s is set
to `.', the current working directory.
- -maxfields n
- Retrieve at most n radar parameters; for example,
if a file contained Zh, Zdr and Ldr then
specifying -maxfields 2 would result in only Zh and
Zdr being retrieved.
- -rangeoffset f
- Perform or re-perform range correction with a range offset of
f metres. Note that in recent cloud data the
appropriate range correction figure is given in the information record
of each raster so that only -correctrange need be used.
- -clean
- This option will remove any cloudy/precipitation pixel that does
not have a signal above or below it, and thus reduce speckle noise.
It is also particularly useful for removing anomalous `lines' of cloud
in vertically-pointing cloud data.
- -doubleclean
- This option is most useful with vertically-pointing cloud data;
often when vertical averaging is not applied (-avegates is
not selected), the anomalous `lines' of cloud mentioned above can be
two pixels thick. This option will remove any cloudy pixel or pair
of cloudy pixels that do not have cloud above or below
them. Provided you don't mind losing any real cloud that is 2 pixels
thick, and are not doing any vertical averaging, you should probably
select this option.
- -customclean n1 n2 n3
- This option removes anomalous `lines' of cloud, but only from
certain heights: isolated groups of n3 or fewer cloudy
gates are tested every n1 gates, offset by
n2.
- -reversev
- Negate all the velocities in the file, so that the convention of
positive towards radar becomes positive away from radar (or vice
versa).
- -minelev f
- Skip any rays with an elevation less than f
degrees. This is often useful to screen out occasions when the dish is
scanning but the operator has forgotten to turn the cloud data
acquisition program off. It is also useful in scanning data to remove
data recorded when the radar is looking below the horizon, as this can
make unfolding of the velocities difficult.
Processing options specific to scanning data
- -scan n
- Retrieve only scan n of the raster.
- -ZDRcal f
- Add f dB to all the Zdr values.
- -LDRcal f
- Add f dB to all the Ldr values.
- -PHIDPcal f
- Add f degrees to all the phi_dp values.
The chilnoise program can be used to
determine the correct value to use.
- -maxLDR f
- Remove all data (probably ground clutter) with an LDR greater than
f dB (-10 is a good value).
- -noise fZH
- Subtract noise from Zh using a noise-equivalent
reflectivity at 1 km of fZH. This is important if
Zh is to be used at low signal-to-noise ratios. Note that in
cloud data you should be able to use the -process option to
automatically calculate the noise level from the uppermost gates. In
scanning data the chilnoise program
can be used to determine the value to use.
- -polar fZH fZV fZHV
- Subtract noise from Zh, Zdr and Ldr
using noise-equivalent reflectivities at 1 km in the H, V and
cross-polar channels of fZH, fZV and
fZHV (dBZ) respectively. This is important if these
parameters are to be used at low signal-to-noise ratios. Between
October 1998 and July 2000, the correct values for the CAMRa radar
were -41.7, -43.7 and -43.7 respectively. These figures are likely to
change whenever the hardware setup is modified. If you are missing
Zdr or Ldr then fZV or
fZHV are ignored (respectively). You can use the chilnoise program to determine the
correct values to use.
- -threshold f
- Set the noise-elimination threshold to f, when the
-polar option is in use. The default value is 1.19, which
essentially means that any signal less than 19% (in a linear sense)
above the mean noise floor is rejected. If anomalous echos appear in
obviously cloud-free regions then use a higher value.
- -doppler f
- Reject noisy data by looking for big jumps in v2 or
phi_dp. The threshold parameter f should be
around 400: use higher values to accept larger jumps in these two
parameters. This option is best used in combination with
-doubleclean, but I would now avoid it in favour of using the
chilunfold program, which both
unfolds the velocities and removes noise from the velocities and
phi_dp.
Processing options specific to cloud data
- -process
- Subtract noise using the Zh values in gates at the top of
the ray (which are assumed to be free from cloud) to characterise the
noise.
- -noisegates n
- Use n gates at the top of the ray to characterise
the noise. The default is 25.
- -oscillator
- Subtract noise assuming that the there is a weak receiver
interference of an oscillatory nature, and use the gates at the top of
the ray to characterise this oscillation. Hence it is important that
the correct number of gates (specified with the -noisegates
option) are used, as they must match the frequency of the
oscillation. If this option is not used with 94 GHz data then
anomalous `lines' of cloud, around 1 pixel thick, will appear in the
data around 25 gates apart.
- -noisefactor f
- When performing noise subtraction, eliminate all signals which are
less than f standard deviations above the mean noise
floor. The default value is 3, but for cloud radars such as the
Rabelais, a value closer to 5 is necessary to account for the
strangely range-dependent level of the noise floor.
- -correctrange
- Perform range correction using the range offset value encoded in
the information record (which has been estimated at Chilbolton from an
oscilloscope). If range correction has already been performed then it
will be recalculated using the new range.
- -noclutter n f
- Attempt to remove clutter using the v and sigma_v
variables; basically any data up to gate n with both a
v and sigma_v of less than f m/s is
removed. For the 94 GHz radar, the options -noclutter 15 0.1 -clean
will get rid of most clutter, although insects will not be
removed. Since sigma_v can be calculated only if some
averaging is performed, this option can take effect only if some
averaging is performed.
- -v_floor f
- Often the velocity information at low signal-to-noise levels is
garbage; this option sets the equivalent noise floor for the
v parameter to f dB above that used for
Zh.
- -sv_floor f
- Similarly, this option sets the equivalent noise floor for the
sigma_v parameter to f dB above that used for
Zh.
- -reject n f
- Sometimes the 94-GHz receiver and transmitter can get out of sync,
in which case the range to each gate is completely wrong. Since the
outgoing transmit pulse is detected strongly by the receiver at zero
range, this problem can be detected by looking for occasions when the
very large and abrupt echo from the transmit pulse occurs several km
above the surface. Selecting this option will skip rays whenever there
is a jump in reflectivity of more than f dB over two
range gates at the nth gate (zero based) or above. In
normal 94 GHz reflectivity data there is a jump of around 50 dB over
two range-gates when the transit pulse is sent out, and this should
normally occur centred on gate 4 (zero based). Thus -reject 10
30 should be OK for this radar (changes in reflectivity of 30 dB
over a vertical distance of 120 m never occur naturally). Note that
this test will be performed after any averaging of gates, so if
you use -avegates 2 then you would have to change this option
to -reject 5 30.
- -nobreak
- Don't split cloud data up into separate days.
- -pciclean
- Sometimes the PCI data acquisition system glitches rays every few
minutes. The cause of this is unknown, but the affected rays are
nearly all removed with this option. Unfortunately it means that
the time between rays is not always constant.
Recommended options for scanning 3 GHz CAMRa data
If you have the summary file around and the calibration values are
correct then you can use the -autocal option for Zh
and Zdr calibration, and the chilnoise program for obtaining the
other parameters:
PARAMS=`chilnoise file.dat`
Then the best options are
-clean -rangeoffset -700 -clean -maxLDR $PARAMS
Alternatively, options can be selected explicitly:
Data recorded since October 1998
-Zcal +5 -PHIDPcal +42 -ZDRcal +1.5 -polar -41.7 -43.1 -43.1
-maxLDR -10 -clean -rangeoffset -700
Data recorded since around October 2000
-Zcal +6.8 -PHIDPcal +42 -ZDRcal +2 -polar -41.6 -43.1 -43.6
-maxLDR -10 -clean -rangeoffset -700
Recommended options for 3 GHz Singapore data
This is rather preliminary...
-Zcal 18 -LDRcal -11 -polar -18.35 0 -7.43 -clean -rangeoffset -500
Recommended options for 35 GHz Rabelais cloud data
There is an abundance of artifacts in 35 GHz data that my program
can't remove - in particular look out for that big band between 2 and
3 km. The system also has a poor recovery time so that anomalous echos
are often observed directly above strong signals, where from the 94 GHz
radar it is clear that there is no cloud. I've selected an increased
noisefactor of 4.5 to get rid of some of this, but you might want to
go higher.
Data recorded from 1 to 28 Oct 1998
-process -noisegates 25 -noisefactor 4.5 -clean -Zcal -15.5 -rangeoffset -480
Data recorded from 29 Oct 1998 to Jan 1999
On the evening of 28 October an attenuator setting was changed by
9.5 dB - calibration changed accordingly.
-process -noisegates 25 -noisefactor 4.5 -clean -Zcal -25 -rangeoffset -480
Data recorded since Feb 1999
Range offset encoded into the information record so doesn't need to
be specified manually. Note that early on in these data I and Q were
not centred every ray before calculating velocity so some of the
Doppler data may be very dubious.
-process -noisegates 25 -noisefactor 4.5 -clean -Zcal -25 -correctrange
Recommended options for 94 GHz cloud data
Data recorded between Oct 1998 and Jan 1999
The -oscillator option should get rid of most of those
annoying lines, but data was only recorded up to 12 km and you
need 25 gates at the top of the ray to characterise the
oscillator interference because of it's frequency - hence when cloud
extends well above 10 km this affects noise subtraction further down
the ray.
-oscillator -noisegates 25 -noisefactor 3.0 -doubleclean -Zcal +16 -rangeoffset -190
Data recorded between Feb 1999 and Oct 2000
The main uncertainty in this period is the calibration - work to
sort this out is still going on (as of 11 November 2001). The netCDF
data that has been supplied to the British Atmospheric Data Centre was
processed with -Zcal +23; it now seems that this was too low
a calibration figure (i.e. the radar is less sensitive than previously
thought).
Compared with the earlier data, there are more gates to get the
noise from so if you're not averaging any gates you can use
-noisegates 50. The range offset is encoded into the
information record as -480 m so doesn't need to be specified (you can
use the -correctrange option). However, looking at the data
suggests that this is out by about one range gate, so you might find
that -rangeoffset -420 makes rain look better very near the
ground where the values of reflectivity are most sensitive to errors
in range correction. You may not want to use the -noclutter
option as it can make plots look a little odd when insects are present
(which it cannot remove). The -maxfields 1 option suppresses
the v and sigma_v parameters which we don't
trust.
-oscillator -noisegates 50 -noisefactor 3.0 -doubleclean -Zcal +23
-correctrange -minelev 85 -reject 10 30 -noclutter 15 0.1 -maxfields 1
This is suitable for no averaging of gates; if you specify
-avegates 2 then you ought to change the rejection option
to -reject 5 30 and use -clean rather than
-doubleclean.
Data recorded from 2001
...