A.3 Laboratory calibration
The laboratory calibration of TRMM LIS is discussed in detail inKoshak et al. [2000], and this was the same calibration
approach applied to the spare LIS unit (i.e., ISS LIS). The calibration
consisted of four main efforts: (1) a static response test, a (2)
transient response test, (3) a spectral test, and (4) an FOV test.
Additional elements of the calibration pertinent to LIS performance
characteristics are discussed in Boccippio et al. [2002].
The calibration tests, which are referred to here as the original
calibration (OC), were carried out on both the original TRMM LIS and the
spare unit (the present ISS LIS) in the summer of 1997. TRMM LIS was
subsequently launched to orbit, while the spare unit was stored in a
safe box in an environmentally controlled facility for many years, until
it was integrated on STP-H5. In the summer of 2014 and prior to the
integration on STP-H5, a retest calibration (RC) was performed on the
spare unit to determine if there were any significant changes in the OC
given the many years that the unit was in storage. The RC
instrumentation and procedures employed were made as similar as possible
to those employed in the OC, but unfortunately the OC and RC
methodologies were not identical.
A brief overview of the OC tests applied to both the TRMM LIS and ISS
LIS are provided below:
- Static Response Test : The OC static response test provided
the linear response of each pixel, and hence also quantified
uniformity across the charge coupled device (CCD) array. It employed
an 8-inch integrating sphere calibration standard. The sphere lamp
source emitted a static radiance that was nearly isotropic and uniform
over the 2-inch diameter exit port (source stability at 3000 K color
temperature was specified at ± 0.5% over a 1-hr duration, and ± 2.0%
over 100 hr). The radiance was continuously adjustable over a range of
five orders of magnitude without changing the color temperature. Since
the sphere output could not fill the sensor FOV, a motorized
positioning system (containing precision Newport/Klinger rotation
stages) was used to yaw and pitch the sensor head to effect full FOV
coverage.
- Transient Response Test : The purpose of the OC transient
response test was to determine the transient response of the sensor to
optical pulses of various integrated energies, against several
different levels of steady-state background radiance, and for several
different pixels across the CCD array. Pulse energy was changed by
varying the pulse duration within a 2-ms LIS frame. The primary
component of the test system was a 2-inch integrating sphere
containing a near-infrared light emitting diode (LED) and a small
quartz tungsten halogen (QTH) lamp. The LED was mounted behind a
pinhole in the far surface of the sphere. Background radiance levels
were adjusted by a variable aperture in the lamp input port, thus
maintaining a constant color temperature.
- Spectral Response Test : The OC spectral test employed a
high-resolution grating monochromator (500-mm focal length, f/5
aperture, and 0.1-nm resolution) as the primary component. The
attached source module contained a QTH lamp and a krypton rare gas
discharge lamp as a wavelength reference. The monochromator output was
fed through a fiber-optic cable whose output was approximately
collimated by a small off-axis paraboloid mirror. By scanning in
wavelength, the spectral test determined the sensor end-to-end
relative spectral response. This test covered only the wavelength
region near and within the passband of the narrowband interference
filter.
- FOV Test : The FOV test in the OC employed a 9-inch diameter,
off-axis paraboloid mirror and an infrared LED. The LED was used to
illuminate a total of 31 pixels that were evenly spaced across the
CCD, and the associated source incidence angles for each pixel was
computed. The LED incidence angles to the lens could be viewed
equivalently as lightning source angles. The geometrical mappings were
mathematically unique and were used to build a lens transfer function
(i.e., boresight angle vs. pixel distance from center of the CCD).
These results are fundamental to the process of geolocating lightning.
The overall sensor FOV (approximately 78.5° x 78.5°) was determined by
simply illuminating pixels on the CCD perimeter.
As mentioned previously, the RC methodology and equipment were not
identical to the OC. In the fall of 2013, prior to beginning the RC, it
was deemed necessary to upgrade much of the equipment employed in the
OC, since these were out of date and were no longer compatible with
current technology. For the static response test, two integrating sphere
systems were procured. The first was a 12-inch sphere with a large
aperture which was intended to allow uniform illumination of a larger
portion of each quadrant of the ISS LIS instrument. The second was a
6-inch sphere that was comparable but not the same size as the 8-inch
sphere used in the OC. This second sphere was acquired for the purposes
of transient testing, as it had a removable port in the back for
attaching an LED.
Ultimately the 12-inch sphere was deemed unusable for the static
response test because the large opening allowed for “hotspots” where
the luminosity was greater surrounding the tungsten bulbs. The 6-inch
sphere worked for testing the static response of the instrument;
however, it only covered a fraction of each of the quadrants. After
comparing the results with the legacy calibration, it was noticed that
there was a form factor difference, so the legacy 8-inch sphere was then
used to determine if the values were still the same as the legacy
calibration. Using the legacy sphere as a one-to-one comparison with the
previous calibration was a success.
Use of the 6-inch sphere for transient testing was originally the plan;
however, after inserting the newly painted LED insert there were
noticeable differentials in the reflectivity of the new insert,
especially around the edges. It was decided to use the LED as the source
for the backgrounds for each pixel and then on top of that send the
pre-programmed transient signal. This allowed for a very controllable
and fluid process for the transient calibration.
For the spectral test, the grating monochromator used in the OC could no
longer be used because the controller was no longer available, and the
connections were obsolete. Instead, a new monochromator was employed for
the RC. The FOV test for the RC was performed in much the same way as in
the OC, and the results were essentially identical.
ISS LIS alignment measurements were conducted in the RC. The alignment
measurements were obtained by illuminating different sides of a
mirror-faced alignment cube (attached to the outside of the LIS sensor
head assembly) with a theodolite; this allowed determination of the
overall rigid alignment of the LIS lens/CCD systems with the STP-H5
module and ISS platform.
While the RC procedure was acceptable, it was not optimal. The intent of
the RC was to ensure that nothing significantly changed with the
instrument during the years in storage. For expediency, and because the
RC results showed no significant changes from the OC results for TRMM
LIS (Fig. A1), the OC results for the TRMM LIS were applied in the ISS
LIS processing code, which produces the version 1 dataset available at
the GHRC.
However, a plan is being implemented to replace the TRMM LIS OC results
in the ISS LIS processing code with the ISS LIS OC results, since
identical procedures could not be followed due to the passage of time.
In particular, the RC static response test was incomplete because it did
not cover all pixels in the CCD array; it only illuminated a small
circular portion in the center of each quadrant. This leaves some
lingering calibration uncertainties in the version 1 ISS LIS dataset.
The ISS LIS science team is working to retrieve the digital OC
calibration data for the spare unit, and before updating the ISS LIS
processing code the ISS OC and RC results will be compared in detail, as
well as with the OC results for TRMM LIS. Based on initial analysis, an
improvement of 2-5% in key instrument performance parameters (e.g.,
flash detection efficiency, flash false alarm rate, geolocation
accuracy, and optical amplitudes) is expected after a future switch to
the ISS LIS OC.