Calibration Scenarios for PICASSO-CENA

1
Calibration Scenarios for PICASSO-CENA
J. A. REAGAN, X. WANG, H. FANG
University of Arizona, ECE Dept., Bldg. 104,
Tucson, AZ 85721
MARY T. OSBORN
SAIC, NASA Langley Research Center, M.S. 475,
Hampton, VA 23681
Objective
This poster presents scenarios for calibration of
the PICASSO-CENA 532 nm and 1064 nm lidar
channels. The calibration approaches are
presented, including uncertainty assessments and
demonstrated calibrations obtained from LITE data.
2
Background and Strategy
PICASSO-CENA is being developed as a partnership
between NASA and the French space agency CNES and
is planned for launch in early 2003. Calibration
of the lidar is very essential to the
quantitative retrieval of the aerosol and cloud
properties in the PICASSO-CENA mission. Following
the procedures employed for LITE, in situ
calibration of the 532 nm lidar channel will be
accomplished via normalization to high altitude,
nearly molecular scattering regions. The
molecular backscatter will likely be too weak
and/or too aerosol contaminated to permit such
calibration for the longer wavelength 1064 nm
channel. Rather, calibration of the 1064 nm
relative to the 532 nm channel calibration will
be obtained via comparisons of the 532 nm and
1064 nm backscatter and integrated attenuated
backscatter from known/quantifiable scatters
such as cirrus clouds.
3
Molecular Normalization Calibration
The lidar calibration factor or constant, C,
appears in the lidar equation and normalized
equation as follows where E0
transmitted laser pulse energy r range
or distance from the lidar to the point of
scattering P(r) instantaneous lidar signal
from range r ?(r) atmospheric backscatter
coefficient (m-1sr-1) T(r) atmospheric
transmittance through range r
4
The calibration constant may be extracted from
the lidar signal P(rc) obtained at a reference
calibration range, rc, by where, in addition to
the terms defined above, ?m(rc) molecular
(Rayleigh) atmospheric backscatter (for the lidar
wavelength) at range rc R(rc)
?(rc) / ?m(rc) total to molecular backscattering
mixing ratio at range rc For rc selected to be
around 30 km above ground, R(rc) 1 and T2(rc) ?
0.99 are reasonable approximations for 532 nm
enabling accurate retrievals of C providing the
signal uncertainty is sufficiently small and
?m(rc) can be accurately computed(driven by how
accurately the air density can be determined).
Using ancillary meteorological data along the
satellite track, it is estimated that ?m(rc) can
be determined within ?3 uncertainty.
PICASSO-CENA simulations predict that the
shot-noise limited uncertainty in P(r) and X(r)
for nighttime, negligible
5
background conditions and a pure molecular
atmosphere at a height of 30 km above ground
should be less than 2 for vertical averaging
over 3 km and horizontal averaging of about 750
km. This should permit C532 to be determined
within ?5 uncertainty. Results from LITE
demonstrate that horizontal averaging over 1000
km and more without significant horizontal
inhomogeneity biases is quite feasible. This is
demonstrated in Fig.1 which shows retrievals of
C532 from LITE data for 300 m vertical averaging
and 100 shot (70 km) sequential horizontal
averaging over a total horizontal extent of 1000
km. The standard deviation of the mean for the
total horizontal extent is less than 1.
6
(No Transcript)
7
Fig.2. The cloud image of orbit 24 from NASA,
Langley Research Center.
8
1064/532 Ratio Calibration from Cirrus Cloud
Returns
Cirrus clouds offer good candidate targets by
which the calibration ratio C1064/C532 can be
estimated from the ratio of the normalized
returns for the two wavelengths, because, to
first order, the backscatter and extinction
should be nearly the same for both wavelengths.
Also, as cirrus occurs at high altitudes,
corrections for 1064/532 spectral transmittance
differences between the satellite and the cloud
top are relatively small and fairly predictable.
The normalized cloud return, Xc, defined as the
total normalized return minus the non-cloud
background normalized return, is given
approximately by Xc
CT2cT?cT2c
where C lidar calibration factor T2cT
round-trip transmittance to cloud top at range
rcT ?c cloud backscatter for r gt rcT T2c
cloud round-trip transmittance from rcT to r gt rcT
9
Assuming ?cT2c is the same for 532 nm and 1064
nm, the ratio of Xc for the two wavelengths at
any r within the cloud, or the ratio of the
integrals of Xc through the cloud for the two
wavelengths, will be approximately
and
For rcT at about 12 km above ground, the
round-trip transmittance ratio is given
approximately by Cirrus cloud returns from
LITE, Orbit 24, have been analyzed to assess the
feasibility of the cirrus cloud calibration
approach. Orbit 24 was chosen because the
C1064/C532 calibration ratio was estimated from
ground returns at Edwards AFB, thereby providing
something to compare against. Figure 2 shows the
LITE cloud image data used for this analysis.
Example Xc profiles for 532 nm and 1064 nm
obtained from 10 shot averages and 150 m vertical
10
averaging are shown in Figs. 3 and 4. The
background for 532 nm follows a molecular
scattering line, while the 1064 nm background is
a vertical, system noise defined line. Also, the
ratio XC1064/XC532 ( normRatio) and the
ratio of the integrals of Xc ( intRatio)
for this cloud example are shown in Fig. 5. The
plots in Figs. 5 and 6 display the expected ideal
behavior of both the normalized Xc ratio and the
ratio of the integrals at Xc being equal and
constant with height, with the integral smoothing
removing some of the fluctuations exhibited in
the Xc ratio. Not all cases were so ideal and
screening was required to identify other useable
segments and useable height ranges within a
segment. Other acceptable cases are shown in
Figs. 7 and 8. Table 1 also lists the results for
some 18 cloud cases (segments), where each
numerical listing is the average over a useable
height range identified for the particular
segment (some segments have multiple useable
height ranges). The averages for all cases yield
little difference between the results for the Xc
ratio versus the ratio of the integrals of Xc
approaches. Multiplying these averages by 0.9
yields the estimates for C1064/C532, as shown in
the table, including the ? one standard deviation
limits from the averaging.
11
Table 2 shows a comparison of the LITE C1064/C532
calibration ratios obtained by the cirrus cloud
approach (mean of the two values in Table 1),
pre-launch instrument specifications/measurements,
and surface returns from the Rogers dry lake bed
at Edwards AFB, CA.
Table 2. LITE 1064/532 Calibration Determinations
Calibration Determinations C1064/C532
From Instrument Specifications 93
From Edwards Surface Returns 79 ? 10
From Cirrus Cloud Returns 89 ? 12
12
Fig.3
13
Fig.4
14
Fig.5
15
Fig.6
16
Fig.7
17
Fig.8
18
Conclusions Calibration of the PICASSO-CENA 532
nm channel by molecular normalization, analogous
to what was done for the LITE shuttle mission, is
quite feasible and should yield calibrations with
uncertainties of ?5 or less. Calibration of the
PICASSO-CENA 1064 nm channel in terms of , or as
a ratio to, the 532 nm calibration factor by
using cirrus cloud returns appears quite
feasible. The accuracy with which this can be
achieved is still open to question, but the
results presented here suggest that it may be
possible to reduce the uncertainty to 10 or
less.