--Bob Kannenberg,
rkannenb@pop900.gsfc.nasa.gov
Science Systems & Applications, Inc.
Introduction
Vince Salomonson, MODIS Team Leader, convened the MODIS Science Team
(MST) meeting and welcomed participants. Salomonson stated that the
instrument is currently undergoing consent-to-ship review at SBRS.
Thermal vacuum testing was completed successfully, and polarization
issues have been resolved. The instrument will soon be shipped to Valley
Forge, and it is expected to launch in 13 months.
EOS Project Science Report
Michael King, EOS Senior Project Scientist, reported that the biennial
review process has been instituted to look at things like how MTPE
programs are addressing science themes, processes for incorporating new
technologies, partnerships with commercial and international entities,
etc. EOS is implementing an interagency validation strategy,
incorporating satellites, surface networks and field campaigns. King
presented a timeline showing Earth system model development; the goal
here is to have fully coupled Earth system models by the year 2010.
King presented a table comparing the numbers of EOS data products in May 1993 to the numbers in April 1997. The total number has increased from 239 in 1993 to 282 in 1997; this jump is attributable to certain algorithms, most notably data assimilation, that were mistakenly neglected in the 1993 database. King dispelled the perception that there is an astronomical overload of data products; many of these products can be accounted for as front-end calibration.
King announced that the Standard Data Products Resources Board, chaired by Skip Reber, has been established to assess and manage changes to the EOS data products list. Currently the BoardÍs main dilemma is the accuracy of the data products database. Hughes ECS is building its database to the February 1996 baseline, although the 1997 Data Products Handbook differs substantially. Discrepancies have emerged as some interim products have drifted to standard. King expressed concern that a mechanism be put into place to provide a reality check between actual MODIS sizing requirements and what Hughes understands the requirements to be.
AM-1 Project Science Status and Early Science
Yoram Kaufman, EOS AM Project Scientist, announced that David Herring,
recently hired as Outreach Coordinator, has created an EOS AM Bulletin
Board System (BBS) and will be creating an EOS AM Web site as well.
Kaufman encouraged use of the BBS to discuss issues like the calibration
maneuver and early science. He stated that he is actively seeking to hire
an outreach scientist to coordinate MTPE PAO outreach efforts like
brochures, museum exhibits, articles and press releases. Herring will
assist in promoting a contest to name the AM-1 platform. Kaufman
indicated that AM-1 is assessing how best to perform the calibration
maneuver, or deep space look, for MODIS and CERES.
PM-1 Status
Claire Parkinson, PM Project Scientist, reported that the spacecraft is
now in month 13 of its 54-month development period. The work by TRW, the
spacecraft contractor, was delayed by 6 months (September 1995 through
April 1996) because of a protest to the Common Spacecraft contract, but
launch is still scheduled for December 2000, and overall progress has
been rapid since TRW came onboard. Parkinson announced that TRW has
signed off on interface control documents (ICD) for all but one (Humidity
Sounder for Brazil [HSB]) of the instruments planned for PM-1.
Parkinson indicated that on May 6 a meeting was held to discuss the feasibility of conducting a PM-1 calibration maneuver. MODIS and CERES are in favor of the maneuver, while AMSR and AIRS are opposed. TRW will examine the feasibility and potential impacts associated with performing the maneuver one time on day 45 of the mission, with MODIS and CERES powered on, and AMSR and AIRS powered off.
EOSDIS Status Report
Rick Obenschain, EOSDIS Project Manager, reported that Release B.0, due
May 15, 1998, will have limited product generation capability but will
allow MODIS all the capabilities necessary to support early mission
calibration and algorithm testing and refinement. Release B.0', an early
incremental build, will be demonstrated in August; it will provide the
capabilities necessary for critical functions. If B.0 is delayed and not
delivered as scheduled, then B.0' can at least provide the capabilities
critical for launch. Release B.1 is scheduled for delivery January 15,
1999. Obenschain expressed his concern that the August demo of B.0',
initially intended as an internal demo, has evolved into a "go/no-go"
decision point. As such, the scope of the demo has been expanded to the
point where preparation is impacting development of B.0. This impact,
estimated to be about 4 to 6 weeks, will be taken into consideration by
Sara GravesÍ review committee. Funds have been allocated and held to
support AM-1, Landsat-7 and SAGE III emergency backup plans. MODIS is
currently revising its backup plan proposal.
Chris Justice voiced his concern that ESDIS is considering the reduction of instrument interdependencies and, from the Land point of view, these interdependencies are critical. Obenschain replied that ESDIS had been asked to look at ways to cut costs in the out years. ESDIS is evaluating the savings associated with data volume reductions of 25% and 50%. The likely scenario is that the number of Level 2, 3 and 4 products will be reduced, at least for the initial phases of each flight mission, and 100% of the Level 1 products will be produced.
Coordination with Integrated Program Office (IPO)/National
Polar-orbiting Operational Environmental Satellite System (NPOESS)
Stan Schneider reported that NPOESS intends to fly in three orbits: 0530
sun-synchronous, 1330 and EUM (between 930 and 1030). Of most interest to
MODIS is the Visible and Infrared Imaging Radiometer Suite (VIIRS), which
will be a single sensor or suite of sensors that address visible IR
imagery and radiometric measurements. VIIRS meets critical NPOESS
requirements for imagery, sea surface temperature (SST) and soil
moisture. Delivery of the first flight unit is slated for January 2004.
The RFP for VIIRS differs from the typical NASA RFP in that the vendors were asked to deliver the algorithms as well as the product. However, most of the expertise in algorithm development resides in universities and the government, so Operational Algorithm Teams (OAT), consisting of DoD, NOAA and NASA representatives, have been formed to oversee the algorithm development process. Schneider announced that the NPOESS Web site is located at: www.laafb.af.mil/SMC/PK/NPOESS/rfp.htm
Level 1B Validation Plans
Bob Murphy reported that the Level 1B plan is not as well-developed as we
would like it to be. The validation approach will incorporate aircraft,
ground-based, ship-based and mooring platform sensors. MODIS derived TOA
spectral radiances will be validated in the 16 VIS/NIR and 4 SWIR bands
using high radiance sites (i.e., White Sands) and low radiance sites
(i.e., Lake Tahoe). TIR bands 31 (11 µm) and 32 (12 µm) will be validated
directly; other TIR bands will be validated by referring to 31 and 32 via
the onboard blackbody. Long-term stability will be monitored by lunar
looks through the space view port 3 to 7 times per year. Radiometric
calibration change over days to weeks will be checked with solar diffuser
measurements. Murphy noted that in the Level 1B draft algorithm, no
follower algorithm is incorporated. Until the MODIS Science Team allows a
follower algorithm, the solar diffuser will not be integrated. He
anticipates that the Level 1B Validation Plan will be revised and
available for review sometime in early 1998.
Version 2 Delivery Schedule
Ed Masuoka, SDST Leader, reported that software will be received from
April through October, and that science software need dates are driven by
synthetic data, product dependencies, time to test at TL-SCF, and time to
integrate and test at the DAACs. Version 2 software requirements are
completed. SDST is finishing up metadata work with ECS, and Masuoka
stated that he would like to put a freeze on metadata this month (May),
and the earlier the better.
Phase-in Plan for Data Products
Masuoka indicated that the phase-in plans for data products are being
worked by the discipline groups. He cited the dependencies between Land
and Atmosphere as a key issue that needs to be worked.
Direct Broadcast/Reception
Murphy reported that the Project has a Level 1 requirement to provide
direct broadcast (DB) of data, but there is no requirement for reception
of same. EOS AM-1 Direct Access System (DAS) capabilities include
real-time DB of MODIS data and ancillary data, as well as real-time
direct downlink (DDL) of ASTER (Vis-NIR-Thermal High Resolution) data.
Murphy stated that DB would promote the use of MODIS data by our
international colleagues, and enable their participation in our
validation programs. NASA HQ is interested in DB, and a joint EOS-HQ
effort may produce a cost-effective way to receive, process and
distribute data.
Geolocation Validation
Strahler reported that the Geolocation Validation meeting held May 13 was
very successful. Overall, good progress has been made; there are no
showstoppers, although the schedule is tight. Participants reviewed
Version 3.0 of the Geolocation ATBD. Strahler anticipates that Version 2
production code should be delivered as scheduled on June 9, 1997. The
Version 2 base algorithm should be delivered between June and September
1997. The Geolocation Group will take the action to work the issue of
band-to-band registration in the event of SRCA failure.
Data Assimilation Office (DAO) Status Report
Yong Li reviewed the methodology behind GEOS-3 Gridded Data Products. The
GEOS-3 system is expected to be operational in June 1998. He encouraged
Science Team members to contact him for copies of the file specification.
Anyone in need of DAO data should complete the DAO data survey form; the
DAO will then respond as to whether their data are suitable for the
user's purposes as described in the survey. Justice questioned the need
for the survey, saying that by this time the DAO should be addressing the
requirements that it has already been given. Li encouraged more direct
communication between the MODIS Discipline Groups and himself.
Li presented a sample test data set that DAO had already created, which covers the period July 31 through September 1, 1996. This data set was created for MODIS software testing and based on the requirements received in that time frame. The test data set is available now in HDF-EOS format. ECS is currently working on the metadata for this data set.
Early Results from the Ocean Color and Temperature Scanner
(OCTS)
Hajime Fukushima, Leader of the OCTS Sensor Team, reported that OCTS is
collecting global VIS and IR data with a 700 m resolution. Overall
instrument performance has been satisfactory, although he noted a couple
of problems (tilt mechanism and sensor degradation in the IR bands).
Fukushima stated that calibration numbers for the visible bands will soon
be updated using AVIRIS data from observations made in April and May. He
noted that on April 10 the calibration system was changed to allow for
the vicarious calibration factor and, so far, the vicarious calibration
results appear to be good.
Early Results from Polarization and Directionality of Reflectances
(POLDER)
Didier Tanr reported that the POLDER instrument utilizes a camera
composed of a two-dimensional CCD detector array, wide FOV telecentric
optics and a rotating wheel carrying spectral and polarized filters.
Planned products include ocean color and marine aerosols, land surfaces
and aerosols over land, radiation budget, water vapor content, and
clouds. Calibration is being done with radiometric methods such as
absolute calibration using Rayleigh scattering, polarization calibration
over sunglint and clouds, inter-band calibration over sunglint and
clouds, multitemporal calibration over desert sites, OCTS/POLDER
intercalibration and statistical relative calibration over clouds.
Salomonson inquired about the accuracy that POLDER aims to achieve, and
Tanr responded that an accuracy of ± 2-3% is anticipated. The spec for
polarization is better than ± 1%. Tanr stated that POLDER is looking at
the cloud mask by using the reflectance threshold in IR and NIR data,
polarization and the oxygen-A absorption band. Data will be available
late 1997 (point of contact:Anne.Lifermann@cst.cnes.fr).
MODIS-like 1-km BRDF and Albedo Retrievals over New England
Strahler presented the work of Robert dÍEntremont at Boston University,
whose objective is to prototype the MODIS/MISR BRDF/albedo algorithm
using satellite observations, as well as to demonstrate the retrieval of
BRDF and spectral albedo at a 1-km spatial resolution. Strahler explained
that the research plan is to obtain AVHRR data and make comparisons. The
algorithm is a kernel-driven semi-empirical BRDF model based on view and
illumination angles. Reflectance is based on 3 terms: a constant for
isotropic scattering; a weight applied to a BRDF shape for volume
scattering; and a weight applied to a BRDF shape for surface
(geometric-optical) scattering. Strahler presented two examples of kernel
shapes, the Ross (thick) kernel and the Li (sparse) kernel. (The first is
for volume scattering, and the second for surface scattering.) The data
set comes from AVHRR and GOES visible band data. Strahler showed a number
of images taken over New England, and demonstrated how the Ross kernels
have BRDF with a moderate-to-strong bowl shape, and no hotspots. These
are associated with forests and dense crops. The Li kernels are more
dome-shaped, with hotspot peaks. These are associated with urban and
suburban regions. Strahler concluded that this work is moving beyond the
theoretical stage, and he is looking forward to using the algorithms on
real MODIS and MISR data.
Evaluation of MODIS Fire Algorithm Using SCAR-B Results
Kaufman explained the methodology that he and Justice used to construct
the MODIS fire algorithm. MAS data gathered over fires in Brazil indicate
that MODIS fire detection capability will be much better than that of
AVHRR. Kaufman indicated that MODIS data will be used to generate a fire
energy product that will distinguish the order of magnitude of fires. He
noted that the combination of both AM- and PM-data will be essential to
doing real science related to fires.
MODIS Cloud Mask Progress and Current Status
Steve Ackerman announced that there have been two significant changes
regarding cloud mask since the ATBD review. The first change involves
replacing the wetland bit with a desert scene bit in the land water flag.
That change will be contained in Release 2. The second change will not
show up in the 48-bit cloud mask file spec. This change involves putting
a cloud adjacency effect into bit field 12. This change will not show up
until after launch. Ackerman encouraged MST members to obtain the
available code to read MAS HDF files, run the cloud mask on a MAS scene,
and let him know the results.
MODIS Instrument Status
Bill Barnes stated that sometime in the next 3 to 6 months MCST intends
to hold a workshop to present an in-depth analysis of instrument test
data. Overall, though, we believe we have a good data set, and that we
can get all of the characterization that we need. Barnes concluded by
saying that Dick Weber has retired and will be succeeded by Ken Anderson
as MODIS Instrument Manager. Barnes thanked Weber for all of his efforts.
Tom Pagano stated that comprehensive environmental tests have validated the instrument's principal design features and demonstrated that MODIS is an excellent spectroradiometer. He reported that all ambient and thermal vacuum tests were performed at hot, cold and nominal temperatures, ten degrees beyond what we expect to see in the space environment. Pagano showed a video entitled "MODIS Calibration and Environmental Testing," and indicated that he would make copies available if requested.
Pagano reported that most major performance areas have some non-compliances but, in most cases, the instrument complies with specifications. Waivers are out for non-compliant areas. Pagano reiterated that overall instrument performance is excellent, and he expects improvements in many noncompliant areas with further data analysis. Concerns include crosstalk, which is higher than expected. It may be possible to improve this by a factor of 4. Another concern is the Near Field Response, which is mostly noncompliant. Finally the dynamic range is noncompliant on several bands, although Band 21 represents the biggest concern.
Algorithm Developers Meeting Summary
Joe Glassy reported that overall the Algorithm Developers feel they have
made progress working as a team amongst themselves and with their SDST
contacts. Programmers are encouraged to see a reduction in SDST staff
turnover, as this aids "institutional memory." Some long-standing issues
persist, however, such as frequent specification changes within a given
development cycle. Glassy suggested that better communication between
algorithm developers and SDST personnel would be mutually beneficial, and
encouraged use of Mike HeneyÍs Programmer BBS where appropriate.
Ocean Group Splinter Summary
Wayne Esaias, Ocean Group Leader, enthused that MODIS looks great, and
added that we probably know more about MODIS than we even thought to ask
about previous sensors. Ocean algorithms and software are on schedule.
Esaias stressed that emergency backup planning (and testing) is
essential. The Ocean Group would like to see a full end-to-end data flow
demonstration included in the August ECS demo. Murphy added that, within
MODIS, we need to define our own success criteria-independent of the
criteria defined by the Graves committee-for the August demo.
Atmosphere Group Splinter Summary
Michael King, Atmosphere Group Leader, announced that the Atmosphere QA
Plan was submitted to the ESDIS Project Office in March 1997; the next
iteration, to include Level 3 QA, is expected in July 1997. King reviewed
figures showing the data volume increases due to QA; cloud mask accounts
for the most significant increase. Liam Gumley at UW demonstrated a
prototype of a MAS online visualization tool, which will be very useful
for cloud mask development. King reported that since the ECS February
1996 baseline, Atmosphere storage volume requirements have grown by
roughly 30%, and this increase has been driven by cloud mask and QA.
Atmosphere data still represents a relatively small fraction of overall
MODIS data volume.
Land Group Splinter Summary
Chris Justice, Land Group Leader, cited the need for a MODIS instrument
"performance/operation" log to establish a link to QA. Land has asked
MCST to look at this. Land has suggested that SDST assess establishing a
similar platform-wide performance log. Justice reported that the Land
Version 2 code delivery schedule currently meets SDST needs. Land needs
simulation and DAO test data sets for Version 2. Justice indicated that
the Land Group is aware that it is the ñtall poleî when it comes to
volumes and loads, and the Group is currently looking at how to refine
these areas. Justice discussed Land validation activities, including the
upcoming Grassland PROVE field campaign at the Jornada Experimental Range
in New Mexico.
MODIS Science Team Meeting Summary
Salomonson stated that in his mind the intensity of the MODIS effort has
now shifted from the instrument itself to the software, and it will stay
that way until launch. The issues surrounding software and validation
require attention from us all. If not now, soon flexibility with
requirements will be clamped, and we will have to become more creative.
We have an exercise to look at validation and products, and we need to
establish these numbers. We must provide specifics to EOSDIS in order to
justify and obtain what we need. The next MST meeting will be held in
October 1997 in the GSFC area.