NASA'S EARTH OBSERVER MAGAZINE ARTICLE
Introduction

The MODIS Airborne Simulator (MAS) continues to evolve to
meet scientific requirements for airborne remote sensing
of clouds, aerosols, water vapor and land. During its
development, MAS has flown periodically on both national
and international campaigns. These missions have provided
valuable experimental data to determine the capability of
narrow bandpass channels in examining the Earth's
atmosphere and in performing scientific research. These
early indications suggest that MAS will successfully
fulfill its goal, which is to aid in defining algorithms
and build an understanding of the ability of MODIS to
remotely sense atmospheric conditions for assessing
global change.
History
Prior to 1991, the MODIS Atmosphere Discipline Group
performed atmospheric investigations with data from the
Multispectral Cloud Radiometer (MCR) and the
Multispectral Atmospheric Mapping Sensor (MAMS). These
instruments have both flown on the NASA ER-2, a high-
altitude research aircraft based out of Ames Research
Center. By 1991 the limited channel capability and
problematic performance of the MCR and MAMS could not be
altered to fit the needs of the users. A study of the
latest scanners was conducted to determine which would be
appropriate for conducting MODIS simulations. Funding and
design schedules were also considered. Of all the sensors
reviewed, the Wildfire infrared imaging spectrometer was
found to be the most adaptable to the science
requirements. It possessed fifty narrow bandwidth
channels, a blackbody calibrator system and could mount
in the wing pod of the ER-2. It was under development and
due to be operational in 1991. The principle investigator
of WILDFIRE (Jim Brass) and John Arvesen (Branch Chief of
the High Altitude Missions Branch at Ames) were amenable
to our proposal to modify the instrument immediately
after development to meet the schedule for the First
ISCCP (International Satellite Cloud Climatology Project)
Regional Experiment (FIRE) cirrus campaign, to be
conducted near Coffeyville, KS during November 1991. This
was to be the first MAS field deployment. Ames'
cooperation was a major factor in the selection and
scheduling.
An Overview of WILDFIRE
In 1987, Dedalus Enterprises, Inc., began developing the
WILDFIRE infrared imaging spectrometer as a
shortwave-infrared, near-infrared and infrared
spectrometer to monitor gases carried into the upper
atmosphere, nutrient transport phenomena, and the
intensity and size of forest fires. The instrument is a
50 channel cross track scanning spectrometer consisting
of 4 spectrometer and detector subassemblies or ports
with linear array detectors (InGaAs, InSb, InSb, and
HgCdTe). The wavelength range of the WILDFIRE
spectrometer was 1.17-12.4 um. This spectrometer is
joined to a scanner subassembly that collects radiometric
imaging data with the following characteristics:
-
Spatial resolution of 2.5 mrad, corresponding to a
ground resolution of 45 m at nadir from a nominal
ER-2 altitude of 18 km,
-
Cross track scan of +-43 degree, equivalent to a
swath width of 34 km with 716 pixels per scan,
- Scan rate of 375 rpm (6.25 Hz).
A photograph of the WILDFIRE spectrometer mated to the
scanner subassembly is shown above. There are up to 16
channel outputs dispersed from each of the four detector
assemblies.
With inputs from the MODIS, CERES and FIRE II Science
Teams, the opportunity arose to modify this very capable
system to more fully meet the needs of the earth science
research community.
Converting WILDFIRE into MAS
Converting the WILDFIRE infrared imaging spectrometer
into a MODIS Airborne Simulator (MAS) was a gradual
metamorphosis. First, there was the FIRE II
configuration, focused on cirrus clouds in the central
U.S. and the Gulf of Mexico. Less than 6 months later the
upgraded MAS was deployed to the Azores, Portugal as part
of the Atlantic Stratocumulus Transition Experiment
(ASTEX), where the focus was on marine stratocumulus
clouds over the eastern Atlantic Ocean. The MAS is
currently deployed to Townsville, Australia as part of
the Tropical Oceans- Global Atmosphere Coupled
Ocean-Atmosphere Response Experiment (TOGA-COARE),
focused on high-altitude cirrus clouds and convective
cloud systems of the western Pacific. These changes in
WILDFIRE have primarily been driven by the need to
increase the performance, reduce the saturation
temperature in the thermal infrared, and to extend the
wavelength coverage in the visible as well as the thermal
infrared for the remote sensing of clouds, aerosols, and
water vapor, and finally to increase the number of bands
that have corresponding channels on MODIS. These changes
were considered and prioritized in light of available
resources and flight schedules, and were conducted
between field campaigns so that the MAS was able to
complement other sensors (e.g., lidar, infrared sensors)
needed to support atmospheric missions.
The MAS configuration for the FIRE II campaign differed
from the original WILDFIRE configuration in that one SWIR
set of channels was replaced with a single 0.68-um
channel and several other spectrometer channels were
readjusted. The channels of the MAS for the FIRE II
campaign are given in Table 1. Eleven flights were
conducted over the FIRE ground site at Coffeyville, KS
and/or the Gulf of Mexico. The aircraft data system could
only handle a limited number of channels, so eleven of
the most significant ones were selected. Of these, seven
channels were recorded at 8 bits, while the remaining
four channels were recorded at 10 bits. These flights,
coordinated with the NCAR Sabreliner, University of North
Dakota Citation, and NOAA 11 and Landsat-4 satellites,
were all successful. Two thermal infrared channels--3.75
and 4.50 um--were noisier than desired, thereby limiting
their measurement capability. Subsequent development has
improved the sensitivity and accuracy in this wavelength
region.
Table 1
Paul Menzel (NOAA/NESDIS, University of Wisconsin)
examined the engineering flight data and set the gains of
the instrument for the first research flights. He
subsequently processed the FIRE II data and converted it
to the spatial resolution of MODIS. In this way he was
able to reduce the instrument noise as well as simulate
the scene characteristics likely to be obtained using
MODIS. The onboard thermal blackbodies and the
ground-based visible and near-infrared calibration
obtained by viewing the Goddard integrating hemisphere
were used to calibrate the data. Finally, cold-chamber
tests were conducted at Ames Research Center to assess
the effect of temperature on the visible and
near-infrared calibration coefficients, since the ER-2
flies at a much colder temperature than the surface where
the ground-based calibration was performed. All data were
processed post-flight by the MODIS Science Data Support
Team, thereby providing calibrated MAS radiances for
every pixel on every scanline and geolocation data
(latitude, longitude, solar zenith angle, zenith and
azimuth angles of the observations) for every tenth pixel
on every scanline. These output data files also contain
the spectral channels that were recorded, as well as the
calibration coefficients and navigation data (altitude,
heading, etc.), and are available in netCDF format on
Exabyte 8500 tapes through the NASA Climate Data System
(NCDS) at the Goddard DAAC (now known as DAAC/Climate).
Conversion for ASTEX
The MAS was returned to Dedalus Enterprises for further
modification in January 1992. At that time the visible
and thermal detector/spectrometer ports were to be
completed. The single visible silicon (Si) detector was
replaced by a nine-element visible and shortwave infrared
linear array detector. The thermal (HgCdTe) detector was
also replaced. However, due to problems with the detector
and the thermal lens, the modifications were incomplete
when MAS was delivered for the ASTEX campaign in May. A
gain change due to the hostile high altitude environment,
which was observed during FIRE II cirrus campaign, and
the aircraft 400-Hz interference noise on the thermal
detector, were further reduced. The fully-configured MAS
for ASTEX was a 50-channel spectrometer with 19 channels
having spectral characteristics similar to those of
MODIS. Again, the aircraft data system could only handle
eleven channels, with the ones selected for this
deployment listed in Table 2.
Table 2
The ASTEX campaign was conducted from Terceira, Azores
during June 1- 28, and provided many multispectral images
of marine stratocumulus in the eastern Atlantic. The
instrument with the new thermal and VIS-SWIR detector
systems operated nominally. Eleven ER-2 flights were
conducted during this campaign. The MAS operated
successfully but the data system only recorded nine of
these flights. Noise was observed in the newly- installed
CO2 slicing (13.19 and 13.95 um) channels. These channels
were noisy for several reasons, a thermal lens coating
had begun to degrade, and the detectors were picking up
radio interference noise from the unsheilded ER-2 pod
heater-blower motor.
In-Flight Performance
On June 17 the ER-2 flew a coordinated mission in which
the NASA ER-2 flew above and the University of Washington
C-131A aircraft flew within an extensive marine
stratocumulus cloud layer located approximately 1380 km
west of Lisbon. This mission also included coordination
with the SPOT satellite. Figure 1 illustrates MAS images
of reflected solar radiation at two spectral channels
(0.665 and 2.142 um), where the images have been
converted from calibrated radiances to reflection
function, formed from a ratio of the reflected radiance
and the incident solar zenith angle and solar flux. In
both of these images, the aircraft was flying from top to
bottom down the center of these images with the MAS
scanning clockwise. These images represent reflectances
over an 85 km x 35 km of the Atlantic, with a clearly
defined demarcation between marine stratocumulus (upper
portion of the image) and continental stratocumulus
(lower portion). In spite of the great distance from the
European mainland, there was a large aerosol outbreak
that led to clouds with larger numbers of small droplets
in the continental clouds. These images have been used to
derive the cloud optical thickness and effective droplet
radius (not shown), and will provide many opportunities
to test and validate remote sensing algorithms for MODIS,
CERES and ASTER.
Figure
1. Images of the reflection function of clouds at
0.665 and 2.142 um, derived from 2600 scan lines of MAS
data on June 17, 1992. These data were acquired over
marine stratocumulus clouds some 1380 km west of Lisbon,
near the Azores.
Table 3
System Upgrades
Following ASTEX, the CO2 slicing (13.19-14.30 um)
channels will be modified to increase the radiometric
sensitivity. The grating and detector will be modified to
improve the transmission and consequently the sensitivity
of those channels. Because the original HgCdTe thermal
detector has degraded, a more reliable detector was
selected as its replacement.
The instrument configuration for the upcoming TOGA-COARE
and Central Equatorial Pacific Experiment (CEPEX)
campaigns is identical to that of ASTEX, except that a
wide-band thermal lens has been selected to replace the
lens used previously. This serves to extend the useful
spectral range to 14.3 um. The channels to be recorded
for TOGA-COARE, shown in Table 3, are somewhat different
than those recorded for ASTEX. Replacing the wide
bandpass thermal lens and shielding the MAS from the pod
heater and blower system has improved the signal-to-noise
ratio for the 8.9 - 13.8 um channels. The new detector
and its discrete cold filter will be installed after the
instrument returns from the TOGA-COARE field campaign.
The cycle of development and field deployments has
continuously improved the MAS performance and the
resulting data quality.
Future Modifications
A 50 channel, 16-bit data system is currently being
developed for the MAS. This system, when completed in the
winter of 1993, will enable the MAS to record all 50
channels at 16-bit digitization, thereby allowing 12 bits
of noise free data with a dynamic gain adjustment range
of 16. With these modifications, it is expected that the
MAS will meet the needs for atmospheric, land and ocean
remote sensing applications. In addition to the CERES and
MODIS science teams, MAS will serve a broader science
community for many years to come.
Acknowledgments
The development of MAS has been the joint effort of many
people at Goddard Space Flight Center (Ken Brown, Liam
Gumley, Tom Arnold), Ames Research Center (Pat Grant, Jim
Brass, Ted Hildum, John Arvesen), the University of
Wisconsin (Paul Menzel, Chris Moeller), and Dedalus
Enterprises (Fred Osterwisch, Steve Cech).
Back to Top