Return to the Index.
The View archived quicklook images command displays a
page with one-month of of thumbnails images. Lidar images are
generated automatically at the end of each 12 hour period. Calibration
coefficients are computed using data from the radiosonde launched at
the begining of the time period and a default system calibration scan
file. For data acquired from Madison, WI the Green Bay, Wisconsin
sounding is used. Data from Barrow uses the BRW sounding and data
from Eureka uses the YEU sounding. MMCR, PAERI, and MWR quicklook
images are also generated for each 12-hour period. However, these data
are typically recieved one-day after they are acquired.
The customize image generation code computes images and line plots
directly from the raw netcdf lidar data archive. By default it
reads data from the last two hours; however, it can be used to
generate custom images from any time or altitude interval within the
archive. Calibration coefficients are computed from a default
calibration scan and the radiosondes nearest in time to the
data. The web interface activates
a Python script which in turn launches a Matlab image processing
program. This program: 1)locates and reads the raw data, 2) locates
and reads the nearest time radiosonde files, 3) locates and reads the
default system calibration files for the selected time period, 4)
computes the inversion coefficients used to separate particulate and
molecular scattering from the radiosonde and calibration files, 5)
performs time and altitude averaging to match data to the
display resolution, 6) generates the inverted data, 6) generates the
plots. The delay between the web request and the image display
increases as the time interval and/or altitude range is increased. Images
of MMCR, PAERI and MWR data are also generated if requested. Data for these
images are interpolated if the pixel spacing is smaller than the instrument resolution
and averaged if the pixel spacing is larger than the instrument resolution. Once
again it takes approximately one day for these data sets to appear in the archive.
If selected, data from the lidar and radar are used together to compute cloud microphysical quantities. These assume a default crystal type of 'bullet rosette'.
The generate housekeeping data command produces plots of internal lidar
parameters for the selected time period. This is mostly of interest to the lidar
operators.
TheView System Online LogBook provides a record
of lidar system hardware maintenance.
The Process and Export Data as NetCDF creates a
downloadable NetCDF data file containing data from all instruments
placed on a common time and altitude grid. Lidar data are processed
from raw data files with time and altitude resolutions selected by the
user. Data from the MMCR, PAERI, and MWR are interpolated if the
selected resolution is higher that the instrument resolution and
averaged if the selected resolution is less than the instrument
resolution. After a NetCDF is created the user can create data mask
fields using Matlab compatible logical statements to select data on
basis of variables contained in the NetCDF. Quick look images are
generated to guide the selection process. These show selected data
points in color with the rest of the image depicted in shades of
gray. The mask fields allow the user to isolate particular cloud types
or signal regimes. The mask fields do not modify the data, but are
provided in the NetCDF as arrays of ones and zeros having the same
dimension as the data arrays. Cloud microphysical properties can also
be computed from the lidar and radar data. A gamma size distribution
is assumed along with power law representation of the particle volume
and area relative to a sphere as presented by Mitchell JAS Vol 53,no
12, 1996. The parameters of the the size distribution and the power
law coefficents can be selected from preset values or entered by the
user.
The Download Archived
Instrument Data as Received option allows users to download
data received from instruments other than the lidar exactly as it was
received from NOAA or the University of Idaho.
An attenuated backscatter image observed with the combined channel of
the AHSRL on 14-Jan-04. A well calibrated conventional lidar
would produce an identical image. Notice how the
cirrus cloud at 7km by is shadowed by 3.4km water cloud which appears at
6:10 UTC. Also notice the strong lidar return seen below the clouds as
a result of the combined effect of aerosol and molecular
scattering. Data gaps at 8 and 12 UTC occur during system calibrations.
The calibration sequences will be shortened in the future. (click on image to enlarge)
The backscatter cross section image computed for the 14-Jan-04 data
shown above. This image is absolutely calibrated and molecular
scattering has been removed. Areas with insufficient signal for the
HSRL inversion are indicated by black shadows; everywhere else
attenuation has been removed. Notice the effect of attenuation
correction on the appearance of 7km cirrus cloud.
The circular depolarization image for the 14-Jan-04 case shown
above. Water clouds produce little depolarizations and appear blue in
this logrithmic display. Ice crystals produce large depolarizations
allowing easy identification of both cirrus clouds and ice crystal
virga which appear red in this image. An insturment bias is evident in
regions of small backscatter cross section where the measured
depolarizations are slightly too large(eg. in the boundary aerosol layer
before 10 UTC). Laser light scattered inside of the instrument seems
to be responsible. Corrective measures are underway.
The optical depth profile measured between 6:40 and 6:50 UTC for the
data shown above. Blue and red symbols show 150 m and 600 m vertical
averages respectively. The water cloud at 3.8k m has an optical depth
of 0.4 while an optical depth of 2.8 is measured between the base of
the cirrus cloud at 4.8km and the point where the signals fall into
the noise at ~8.5km. Column optical depth measurements are currently
limited to values of less than ~4 (even with long averages) by the
system noise floor and systematic errors. The current noise floor is
slightly larger than that which is expected based on photon counting
statistics. This suggests that optical depth measurement limit may
increase slightly as we gain more experience with the instrument.
Major improvements in the stability of system calibrations have been
achieved after additional optical isolation was installed between the
seed-laser and the main laser cavity(Dec 03). However, system
characterization and software is not complete. Known and unknown
errors exist. !!!
Unrealistically large values of depolarization appear at low altitudes
in regions of low backscatter cross section. We now believe that this is
caused by laser light scattered within the instrument. This effect
can be seen in most current images. It is very apparent below 6 km
after 5:00 UTC in the following image. Notice that more
reasonable values (but still biased to larger values) of
depolarization are seen before 5:00 UTC when the scattering cross
section is larger.
Return to the Index.
eloranta@lidar.ssec.wisc.edu
