The MODIS meeting (October 10-11, 1996) was chaired and called to order by Vince Salomonson, MODIS Team Leader. The complete set of these minutes, along with attachments, is available in Portable Document Format (PDF) on the MODIS Home Page, at http://modarch.gsfc.nasa.gov/MODIS/MODIS.html.
EOS Project Science Status Report
Michael King, EOS Senior Project Scientist, announced that four of ten chapters to be included in the EOS Science Implementation Plan have been received. The plan is expected to be completed by Spring 1997 and will be available both as hard copy and as a document available via World Wide Web.
King told the Team that the EOS Data Products Handbook is in the final stages of review. The handbook describes all EOS AM-1 and TRMM data products and presents data flow diagrams. He noted that the MODIS sections of the handbook are still needed.
EOS AM Platform Status
Chris Scolese, EOS AM Project Manager, reported that the EOS AM-1 platform is still on track for a June 1998 launch. Flight hardware is currently being delivered and installed, and spacecraft integration has begun. The launch vehicle and launch facilities are also progressing well. Scolese said that deliveries of ASTER and CERES are imminent. MISR is making good progress, but still awaits system level tests. Scolese is concerned that the MOPITT schedule is currently very tight, and MODIS development is now lagging.
Scolese showed a chart listing the top ten issues facing EOS AM-1. Issue #2 is that the MODIS schedule contingency is eroding. Testing of the MEM (main electronics module) continues and the resistor networks were to be replaced in November. Scolese assured the Team that we will not launch a compromised instrument. He feels that although it will not be perfect, MODIS will be one of the best Earth sensors ever flown. Scolese pointed out that Valley Forge is closing down in June 1998. This drives the EOS spacecraft schedule somewhat to be fully integrated, tested, and ready to launch before that date.
Scolese stated that the MODIS flight model 2 (FM-2) was deleted as directed in the POP 96 budget; however, the Project Office is still pursuing an advanced technology MODIS instrument, or AMODIS. He stated that the AM Project Office is planning to implement a small yaw-and-roll maneuver for calibration purposes. He said that in order to move forward on planning for maneuvers there needs to be a consensus among all EOS instrument teams in favor of the maneuver. Scolese stated that it is cheaper to implement a joystick approach to the spacecraft maneuvers, rather than creating new software at this point. Yoram Kaufman, EOS AM Project Scientist, stated that he will deliver a consensus letter on October 18 after the Science Working Group for the AM Platform (SWAMP) meeting.
EOSDIS Plans to Support Development of At-Launch Production Software
John Dalton, ESDIS deputy project manager, reported that the replanning of EOSDIS is in progress, but no final decisions have been made yet. He stated that the replanning process includes a detailed review and feedback from the EOS instrument teams and DAACs. Dalton presented an overview of the replan strategy. There has been a significant slip--from December 1996 to May 1997--in the scheduled delivery of Release A of the science processing segment of ECS (EOSDIS Core System) in the data server portion. According to Dalton, one factor contributing to the slip was underestimation of the code size required for the data server at the Critical Design Review (CDR). Another factor was that there were COTS (computer off-the-shelf) software product incompatibilities across vendor platforms. Also, there has been difficulty recruiting and retaining expertise in key technologies, such as C++ and Distributed Computing Environment (DCE). Dalton concluded that planned recovery through parallel development is not feasible due to serial dependencies within the system. As a result of the above problems, ECS failed its Test Readiness Review (TRR) on August 5-6.
Mark Abbott asked how users will access EOS data--via the Internet? Dalton responded affirmatively. Abbott said he is concerned that Internet throughput will go down too far, greatly hampering electronic data distribution. Specifically, he is concerned that the system won't deliver Level 3 products, which he feels users will want.
New EOS AM Project Scientist
Salomonson introduced Yoram Kaufman, the new EOS AM Project Scientist, succeeding Piers Sellers. Kaufman began his presentation by showing an organization chart of the EOS AM Project Office personnel. Jon Ranson is the Deputy AM Project Scientist, Jim Collatz is the Associate Project Scientist in relation to IDS investigations, and Francesco Bordi is the AM project liaison.
Regarding MODIS instrument testing at Valley Forge, Kaufman asked for help from the MODIS Team in defining a simple, overall check of the system. Kaufman said he supports on-orbit maneuvers as part of the optimization of the whole MODIS calibration concept. He feels that steps should be taken to minimize risk to the platform; and alternatives to maneuvers should be prepared in case they are needed during the life of the mission.
SDST Status Report
Ed Masuoka, Science Data Support Team (SDST) leader, reported that the Version 1 software deliveries have slipped 3 to 4 months. He told the Team that all software except the Level 3 atmosphere products are on schedule for the Version 1 delivery deadline. He showed a milestone chart of the software delivery timelines. He noted, however, that there is no Release A system at the DAACs that can support MODIS' integration and testing needs.
Regarding the Release A slip, Masuoka pointed out that this creates more time for the Science Team software deliveries, and allows for adequate time for SDST to conduct its tests in the Team Leader Computing Facility (TLCF). However, the slip causes a compressed testing schedule at the DAACs. To compensate, he suggested that the MODIS Team define its most important tests and identify dedicated test strings. The Team should work with SDST to develop more-streamlined testing procedures, and plan on working with several testing shifts at the DAAC. Masuoka said that Version 1 testing will parallel the Version 2 science software delivery. The burden will be on SDST and the science software developers to support changes in the two versions of the software.
Masuoka added that he is unsure whether ECS will be able to produce MODIS products at launch, which is a serious concern. Also, network bandwidth is a concern; however, Masuoka said that the network links ESDIS plans to install next year should provide the needed bandwidth. Masuoka would like all open issues that affect science software developers to be resolved by December 1996--including production rules syntax, HDF-EOS support for MODIS' nested integerized sinusoidal grids, and the standard data product toolkit (SDP-TK) compatibility across all products.
Geolocation Workshop
Alan Strahler listed the following as potential sources of geolocation error: position of center of mass of the spacecraft, spacecraft attitude, and pointing accuracy of the MODIS instrument. He feels that the team must understand the sources of dynamic geolocation error so that these errors may be corrected with software. He strongly advocated establishing ground control points to measure and remove geolocation bias. These ground control points should be stratified with respect to cloud cover, latitude and longitude, and day/night.
Strahler recommended developing a phased approach. He stated that we must characterize MODIS' geolocation bias early so that by the time of launch a system will be in place to give the static error to within 10 meters. He believes that the EROS Data Center (EDC) has the resources needed to do this at launch. Strahler said that MODIS could use the AVHRR control points library. He asked Jeff Eidenshink to report back on EDC's control point holdings.
At-Launch Production Scenarios Workshop
Paul Menzel reported that the following question was raised during this workshop: What should be the process for changing algorithms, reprocessing, and allocation of resources? Menzel suggested that there will not be a stable production environment at launch and that the algorithms will likely change dynamically. He feels that tracking these changes will be very important. Menzel presented the National Environmental Satellite Data and Information Service (NESDIS) procedure for reviewing products in the post-launch period.
Tools for Averaging and Subsetting Workshop
At this workshop ECS gave an overview of the software tools being developed and made available for averaging, subsetting, map projections, and identifying key deficiencies of data products. John Townshend noted, however, that resampling and remapping procedures are not currently being considered. He asked who will develop the software tools for these procedures?
Townshend identified two issues that need to be addressed: 1) EOS data users may get far more data than they request from EOSDIS; and 2) geographic requests for data that cross granules will require multiple orders and only rectangular areas can be requested. Before launch, Townshend would like to test the ECS toolkit on simulated MODIS data.
Townshend added that requests involving multi-temporal images may be difficult; and he is concerned about the lack of tools facilitating multisensor data use. He stated that there is general agreement that there needs to be better communication between ECS and the Science Team. He feels that the MODIS data sets as generated through EOSDIS will likely fail to satisfy many users outside the MODIS Science Team because there is a lack of tools for pre-processing. Skip Reber pointed out that there is a forum for communication between the science teams, ESDIS, and the DAACs, but perhaps the right people are not attending that forum (Data System Working Group).
Quality Assurance Workshop
Wayne Esaias stated that the goal of this workshop was to assess the status of the MODIS Quality Assurance (QA) Plan, which is required by ECS in about one month, so that the team can implement the functionality which will provide MODIS product developers with the information and system for assessing the quality of the MODIS data products. Esaias reported that MODLAND has a draft of their QA plan out for comment; they expect to release their revised draft by November 15. The Ocean Group has met and discussed QA, but has yet to produce a draft plan. The Atmosphere Group has not yet begun to address quality assurance (QA). Esaias stated that it is possible to get convergence within the MODIS Science Team on metadata quality flags. He said he expects that all Version 1 code will contain flag information and software needed to generate QA flags.
Esaias stated that there is concern regarding the readiness of the DAACs to run the QA procedures. He said there is a wide range of opinions as to where QA should take place and how new QA information gets put back into the database. Another concern is the complexity of the post-production process.
Interactions for EOS PM-1
Regarding testing and characterization of the MODIS instruments, it was determined that there are no significant differences between the MODIS Protoflight Model (PFM), which will fly on EOS AM-1, and the MODIS Flight Model-1, which will fly on the PM-1 spacecraft, that make any difference to the Science Team. According to Robert Murphy, some deficiencies in the PFM were identified that could, in principle, be fixed. Among the correctable deficiencies are the D1 dichroic interaction with Bands 8 and 9, and the crosstalk in bands 31 - 36. It was also felt that the MODIS Calibration Chamber could be equipped with a zinc selenide (ZeSe) window, and that solar calibration could be re-introduced.
Regarding cross calibration, Murphy stated that the use of lunar calibration now seems unlikely. However, orbit phasing with other platforms could provide simultaneous views at the poles. Ed Zalewski, University of Arizona, will later provide better descriptions of the options for cross calibration. Murphy noted that EOS PM-1 has much lower specifications for pointing accuracy than does AM-1, which was "driven" by ASTER and MISR. He feels that the MODIS pointing specification for AM-1 and PM-1 is incorrect--it should be ten times better.
Algorithm Developers Workshop
Liam Gumley reported that the algorithm developers need a stable period for developing Version 2 software between December 1996 and the Version 2 delivery deadline. This will allow developers to integrate tools, data, and specifications and work on science algorithm development too. They will also need to better understand and plan for HDF-EOS. Gumley said there are two options for ancillary data: 1) National Centers for Environmental Prediction (NCEP) data could be obtained for Science Team members, who will be responsible for unpacking and resampling; or 2) use data from the DAO (Data Assimilation Office). Gumley stated that the DAO is responsive to users' needs. He said that good communication will be important during the algorithm test process. Developers will want to know test results in a timely manner. He suggested making certain selected results available on the Web.
Calibration Workshop
Phil Slater summarized discussions at the MODIS Calibration Workshop. Slater plans to validate the calibration coefficients using two independent vicarious calibration methods--one method is reflectance-based and the other is radiance-based. Slater proposed an intensive validation campaign during the A&E period. He said he is likely to conduct smaller campaigns at two-month intervals, and then one intensive campaign per year.
The first joint vicarious calibration field campaign--held at Lunar Lake and Railroad Playa in early June--compared TOA (top of the atmosphere) radiances. Team members from ASTER (from both Japan and JPL), MISR, MODIS, South Dakota State University, and University of Arizona participated. Preliminary results from that campaign are still being studied.
Slater introduced Ed Zalewski, University of Arizona, who proposed assisting Santa Barbara Remote Sensing (SBRS) in its absolute radiance calibration. Zalewski conducted a cross comparison of the calibration of SBRS' Spherical Integrating Source (SIS) during the month prior to their bulbs' burning out. He plans to bring back transfer radiometers to help SBRS verify its SIS and characterize MODIS' signal-to-noise ratio (SNR). He also wants to help verify the SIS' and MODIS' radiometric stability during ambient and thermal vacuum tests, as well as before and after vibration tests.
MODLAND Group Discussions
Chris Justice summarized the Land Group's deliberations. Regarding MODIS instrument tests, he said there is a need for an explicit timeline for tests from present to delivery, with scenarios to address any slips in the test schedule. He encouraged MCST and the MODIS Project to continue their current level of attention to SBRS' testing.
Following John Dalton's presentation on the delayed delivery of ECS, Justice stated that MODLAND is concerned that both team and community expectations will not be met in the first six-to-twelve months after launch. As a result, he feels there needs to be a coordinated MODIS/MODLAND strategy with timelines for securing the post-launch data stream needed for QA and validation. It is critical that the broader science community have access to MODIS data as soon as possible after launch. Justice said he is also concerned that ECS toolkits will fall short of the needs for MODIS team members and data users. Justice encouraged a Science Working Group for the AM Platform (SWAMP) initiative to address product-generation interdependencies under the phased ECS delivery.
Justice reported that geolocation is a major concern on the EOS PM platform. He said the bottom line is that the MODIS PM land products must have at least the same locational accuracy as the AM products.
Justice announced that by the next Science Team Meeting, MODLAND will have developed its QA plan and its QA implementation plan. Justice reported that MODLAND's validation plans are evolving rapidly towards a community validation initiative based around aircraft flights, continuous test site monitoring, and mini-campaigns. The Land Group would like to work more closely with other instrument and IDS scientists. For instance, MODLAND is planning to be an active participant in the upcoming LBA (Large-scale Biosphere-Atmosphere experiment in Amazonia ) campaigns.
In response to questions concerning validation timing and planning, Justice said that, for contingency planning in terms of validation campaigns, the EOS Project must provide a first-cut assessment of realistic launch slip scenarios. Justice also stated that the Land Group at this time has no critical external logistic dependencies for validation in the first quarter after launch. However, it is critical that ESDIS provide sufficient data to support an effective QA activity immediately post launch and enough data to permit extensive validation during the first year.
Ocean Group Discussions
Wayne Esaias reported that the Ocean Group is very encouraged with the approaches being taken by SBRS, MCST, and the MODIS Project Office to ensure that MODIS meets our challenging science requirements. He stated that the ocean group is fully committed to the Level 1 mission requirements. He said that success in achieving the sea surface temperature and polarization requirements, and the on-orbit spacecraft maneuver, will enable MODIS and MTPE to address key ocean climate questions that are beyond the capability of in situ and satellite observing systems presently available. Esaias commended SBRS and MCST for sharing preliminary test results and openly discussing the implications.
Esaias announced that the MOBY II instrumented buoy was deployed off the coast of Lanai, HI, on September 14, 1996. He told the team that the daily network transmission of data to the SeaWiFS Project is underway and so far the data quality is excellent. The MOBY instrumentation is unique in its ability to synthesize the spectral band responses of any ocean color sensor. It will play a role in the initialization of OCTS, SeaWiFS, and MODIS.
Atmosphere Group Discussions
Yoram Kaufman told the Team that all Atmosphere ATBDs have been submitted to the EOS Project Science Office, and all Atmosphere validation viewgraphs are completed. The Atmosphere QA plan is still to be determined--Allen Chu was designated to take the lead on developing that plan, as well as representing the group at the November 6 QA workshop at GSFC. Kaufman announced that the Atmosphere Group prefers MODLAND's definition of a data day--standard GMT. The group also prefers an 8-day temporal grid, in addition to the daily and monthly Level-3 averaging periods.
Kaufman stated that the Atmosphere Group would like to make some software modules available for processing direct broadcast data. He recommended that calibration coefficients and software be made available on the World Wide Web for processing direct broadcast data. He said there is also a need for a software package for visualizing images at Level 1B with latitude and longitude markers.
Regarding interactions with the Data Assimilation Office (DAO), Kaufman said the group feels that the DAO produces a very impressive list of products, and they see the DAO as very responsive to the discipline groups. Thus, the Atmosphere Group plans to work more closely with the DAO in the future. For instance, the group is planning a test of operational NCEP models to run parallel with DAO tests.
Regarding upcoming MODIS thermal vacuum tests, the group recommends characterizing the longwave infrared bands (31 with 32 - 36) crosstalk--both spectral and spatial--using existing data. Kaufman also recommends better characterization of the crosstalk between bands 27 and 5, 6, and 7 with additional thermal vacuum testing.
He encouraged MCST and SBRS to develop a model for the effect of the coating on the scan mirror's infrared properties for both EOS AM and PM MODIS instruments. He also advocates laboratory tests of the mirror for at least the EOS PM instrument.
Team Leader's Closing Remarks
Vince Salomonson noted that the MODIS Team has learned of some critical issues facing EOSDIS, and he challenged the team to help EOSDIS evolve so that they can better accommodate all of the EOS science teams.
Salomonson thanked everyone who attended this meeting and announced that the tentative dates for the next meeting are May 13 - 16, 1997.