D Case Studies
The committee examined several cases in which small satellites and streamlined procurement and management approaches were employed to perform Earth observation missions. Candidate missions included the SeaWiFS (Sea-Viewing Wide Field-of-View Sensor), the TOMS-EP (Total Ozone Mapping Spectrometer Earth Probe), the Lewis and Clark missions under the SSTI (Small Spacecraft Technology Initiative), and the QuikSCAT (Quick Scatterometer) mission. For some of these missions, the data available to the committee were limited due to proprietary or other considerations, and only brief synopses are presented for the insights they provide. More complete case studies are included where data were available.
Program Objectives and Context
In support of global change research, the National Aeronautics and Space Administration (NASA) has been monitoring changes in the Earth's ozone layer with a series of ozone mapping spectrometer instruments on various spacecraft. The mission objective for the TOMS-EP satellite was to fill a gap and ensure continuity of data between similar instruments on the Russian METEOR satellite and Japanese ADEOS (Advanced Earth Observing Satellite). Continuity of coverage allows for better correlation of the thickness of the ozone layer with events on the Earth and Sun.
The TOMS-EP program was originally planned as an in-house Goddard Space Flight Center (GSFC) project; after some deliberation, the center decided that a competitive procurement was a better approach to meet the mission goals. The procurement was for a dedicated small satellite to accommodate a previously developed ozone mapping spectrometer from Perkin Elmer (now Orbital Sciences Corporation [OSC]). The contract was let in September 1991. The payload was procured by GSFC and supplied as government furnished equipment. GSFC selected the Pegasus XL to launch the satellite from the western range at Vandenburg Air Force Base into a 955 km Sun-synchronous orbit during the summer of 1994.
TRW was selected as the TOMS-EP contractor. Its proposed design was based on its Space Test Experiment Program (STEP) bus, upgraded to meet the reliability and life goals (0.90 at 3 years) of this critical mission. The
upgrade was achieved through design improvements, parts upgrades, addition of redundancy, and a more robust quality assurance program.
The need to upgrade the bus led to an interesting programmatic trade-off. TRW's STEP spacecraft were being produced in facilities operating efficiently with streamlined processes appropriate to low-cost technology demonstration missions where higher levels of programmatic risk are acceptable. In contrast, processes at TRW's primary Space Park facilities were appropriate to the more demanding requirements of the high-reliability, performance-critical spacecraft produced at that site. Because its design was based on the STEP bus, consideration was given to developing TOMS-EP within the STEP facilities, using processes and controls modified to meet the more demanding TOMS-EP requirements. This plan was rejected, and the Space Park facility was selected for development, primarily to avoid technical and schedule risks to the program that might accrue from imbedding a high-reliability development program within a more informal culture.
The key technical issues in design approach involved structure, solar array orientation/articulation, battery size, distributed versus centralized architecture for the data system, and the design complexity of the spacecraft safing mode. The finished product resulted in an aluminum structure with fixed arrays, a centralized data system, and a safing system that relied heavily on ground operations for recovery.
Recurring design trade-offs for TOMS-EP were the degree of design flexibility and the type of design margins to incorporate. Flexibility and large margins reduce risk and increase the potential for reuse of the bus design on future missions, but at increased cost for TOMS-EP. Because cost was an important issue on TOMS, most trades were decided in favor of limiting flexibility and margins to that needed to ensure the mission. Nonetheless, the TOMS-EP bus provided the heritage for several later spacecraft buses, including the Republic of China's ROCSAT-1, the Republic of Korea's KOMPSAT, and the SSTI Lewis satellites.
Status and Evaluation
The original plan was to launch TOMS-EP during the summer of 1994. Problems with the Pegasus XL launch vehicle delayed the launch to July 1996. At this time, the launch of the multisensor ADEOS with another TOMS instrument was imminent (it was launched August 17, 1996). The flexibility inherent in dedicated small satellite missions gave NASA the opportunity to reoptimize TOMS-EP to take better advantage of its concurrence with ADEOS. Thus, the TOMS-EP orbit was lowered from 955 km to 500 km where it would provide higher resolution data and augment the ADEOS science data return.
TOMS-EP was successfully launched and deployed on July 2, 1996. By mid-August, the spacecraft had gone through its integral propulsion system firings to get into the correct orbit, instruments were turned on, and TOMS became fully operational with real-time data available to the science community. TOMS-EP continues to be operational as of this writing.
The ADEOS spacecraft failed in orbit on June 29, 1997; lost with it were the data from the TOMS and other instruments it carried. Because the TOMS-EP spacecraft carries on-board propulsion, NASA could raise its orbit closer to that of ADEOS, both to increase coverage of the instrument and to reduce drag (and extend orbit life). The boost maneuver was performed in December 1997 and TOMS-EP was raised from a 500 km to a 750 km orbit. This will extend the mission's orbit life beyond the 2-year requirement and 3-year goal to as long as 5 years.
The TOMS-EP project embraced a low-cost, small satellite approach to flying a TOMS instrument over the 1994–1997 time frame as a potential gap filler to ensure continuity of ozone measurements between instruments on the Russian METEOR and Japanese ADEOS satellites. Key programmatic decisions were made, and the program plan was developed, to meet the performance and cost objectives on the desired schedule. TOMS-EP is
a successfully operating system yielding valuable data. However, the program experienced both launch delays and cost growth. Some of the lessons learned or relearned were as follows:
Early agreement and a freeze on requirements are essential for cost and schedule control. Both customer and in-house requirements ''creep" are to be avoided.
TRW used its Eagle Test Bed, an engineering model "spacecraft on a table," to expedite the project schedule through early verification of subsystem interfaces, validation of flight software, etc. This test bed proved very effective in reducing the development schedule and cost. It would have been even more effective had it been built with full redundancy rather than with single-string subsystems. Issues of redundancy management had to await the actual protoflight hardware for resolution.
Fixed price commercial launch vehicles require discipline—in terms of early resolution and documentation of all requirements, interfaces, and other issues—in dealing with launch facilities, integration procedures, safety, and team working relationships. This ensures that the terms of the fixed price contract are clearly understood. A launch vehicle liaison—a designee on both sides of the interface responsible for working on and resolving all issues—should be identified early in the program. This liaison should document all results in a formal launch vehicle Interface Control Document; this too should be accomplished as early as possible.
Program schedule and cost depend on all program elements. For TOMS-EP, the instrument and spacecraft were available on time to support the mission, but the launch vehicle was not. Significant expenditures made to deliver the satellite on time were not effective in meeting the mission schedule. Additional costs were then incurred to store the satellite and reconfigure the team to support the launch 2 years later.
Risk must be carefully assessed for all program elements when defining the system, particularly for schedule-critical missions. For greatest cost-effectiveness, risk should be continuously assessed, progress monitored, and plans adjusted to keep the total program in balance.
The TOMS-EP program clearly demonstrates the programmatic flexibility inherent in dedicated small satellite missions. The planned orbit was lowered to maximize synergy with the ADEOS mission; later, upon ADEOS failure, the actual orbit was raised to reoptimize the mission and extend its life.
Program Objectives and Context
The first research-quality space-based ocean color remote sensing instrument was the Coastal Zone Color Scanner (CZCS), which was launched on board Nimbus-7 in 1978. CZCS imagery revealed the presence of intense mesoscale variability in the spatial patterns of phytoplankton biomass; with the emergence of research on global biogeochemistry, it was realized that satellite ocean color data were an essential component of an ocean observing system. Planning for a replacement ocean color sensor soon began in the early 1980s as the CZCS was designed to operate for only 2 years.
The specifications for an ocean color sensor were well understood, based on analyses of CZCS data as well as in situ observations. By the mid-1980s, the International Geosphere Biosphere Program had assumed sponsorship of the International Joint Global Ocean Flux Study (JGOFS) which was focused on the role of the ocean in global biogeochemistry. One of the JGOFS objectives was the global study of ocean primary productivity and its variability, and satellite ocean color measurements were deemed to be an essential element of the program. This increased scientific support for a satellite ocean color mission culminated in an agreement whereby NASA agreed to purchase from OSC high-quality global ocean color data to be acquired by the SeaWiFS instrument. SeaWiFS would fly on OSC's SeaStar satellite. OSC would be responsible for SeaWiFS development as well as SeaStar development, launch, and operations and the provision of ocean color data to NASA. The SeaWiFS sensor would be built by Hughes (now Raytheon) Santa Barbara Research Center (SBRC).
A broad range of programmatic alternatives were explored for maintaining continuity of ocean color data. At one end of the spectrum, there were many attempts to secure sponsorship for a traditional government procurement wherein NASA, the National Oceanic and Atmospheric Administration (NOAA), the Department of Defense, or some combination of agencies would fully fund the development of an ocean color sensing satellite. At the other extreme, market studies and business plans were pursued to explore the feasibility of a fully commercial enterprise that would raise its own capital investment and subsequently sell ocean color data to all users. Between these extremes, many hybrid approaches were discussed and evaluated, including various government-industry partnerships such as shared investments, advance data purchases, guaranteed data purchases in the future, and "anchor-tenant" arrangements. The purpose of all these exploratory exercises was to find the most economical and expeditious avenue for providing the ocean color data stream. Because the budget authority for a traditional development program was not available, there was an avid search for alternative approaches that would achieve the same result for a smaller taxpayer investment. A stand-alone ocean color mission was expected to cost on the order of $70 million to $100 million, so the search was on to provide the same data stream to the science community for half the cost or less.
Timeliness was another critical dimension. Major international oceanographic studies were planned for 1993–1994 (such as JGOFS), and the availability of worldwide ocean color data from space would significantly enhance theses studies.
In this context, a number of players from government, industry, an academia were engaged to varying degrees in the search for a workable solution to the ocean color mission (see Table D.1).
Fully commercial approaches proved to be unworkable because an attractive business case could not be developed. The viability of a commercial offering depends upon data sales producing a revenue stream that permits recovery of the capital investment and provides a reasonable return for the investors, and such a case could not be constructed because the future revenue stream from the sale of ocean color data was unproven and speculative.
As various business and technical approaches were explored during the 1986–89 period, the most promising and economical scheme appeared to be a "piggyback" ride for an ocean color sensor on some polar-orbiting spacecraft under development for another mission. The essence of this idea was to fly the ocean color instrument for the incremental cost of accommodation on a spacecraft rather than bear the stand-alone cost of a dedicated spacecraft and launch vehicle. Candidate spacecraft included:
NOAA/Earth Observation Satellite Company (EOSAT) Land Remote Sensing Satellite (Landsat) 6,
Table D.1 Organizations Interested in Ocean Color Data
U.S.: NASA, NOAA, Navy, National Science Foundation
Canada: Canadian Space Agency
Former Soviet Union
Earth Observation Satellite Company
Joint Oceanographic Institute
American Geophysical Union
Scripps Institution of Oceanography, University of California-San Diego
Woods Hole Oceanographic Institution
University of Miami,
University of Southern Florida,
University of Oregon,
University of Washington,
University of Rhode Island, etc.
the Canadian Space Agency's Radarsat, and
NOAA's Advanced Television Infrared Observation Satellite.
The first of these alternatives appeared to be the most attractive, and a deal was nearly signed between NASA and NOAA/EOSAT to fly a SeaWiFS on Landsat 6, but the arrangement ultimately fell through because of disagreements about liability for program delays amid concerns that the development time for the payload and modifications required to accommodate SeaWiFS would delay the Landsat 6 launch. For reference, the cost of this piggyback approach was estimated to be between $25 million and $35 million.
Selected Approach: Procurement of a Future Data Stream
Following the collapse of the Landsat 6 piggyback approach, NASA, Hughes, and OSC began exploring a stand-alone "lightsat" solution for the ocean color mission—one that would be as commercially oriented as possible. This process ultimately led to a competitive procurement for an ocean color data stream, wherein NASA would pay in advance for ocean color data to be provided in the future to the scientific community.
This was a significant departure from prior practice. Instead of purchasing hardware and software, NASA was buying only the promise of a future data stream (of specified quality). The value of the procurement was $43.5 million; most industry observers believed that this was indeed a bargain, because the actual cost of developing and fielding the system would probably exceed that figure. The balance of these costs would be recouped by the contractor through commercial sale of the data to the fishing, shipping, and offshore exploration industries, among others. This dual-use approach, where one sensor and spacecraft would serve both scientific and commercial customers, was made possible by tailoring the data distribution policy. Ocean color data are highly perishable, and their commercial utility declines rapidly with time because of the highly changeable nature of the ocean. On the other hand, most of the scientific research entailed is conducted by analyzing the data retrospectively, so it was easy to develop a scheme whereby commercial users would receive encrypted data in real time (for a fee), while scientific users would have unlimited access to the data after 48 hours.
After nearly 5 years of debate and evaluation of alternatives, a firm fixed price contract was let to OSC in March 1991. OSC subsequently contracted with Hughes SBRC to build the payload instrument, SeaWiFS. The objective was to have the spacecraft on orbit and delivering ocean color data by August 1993 in time to support JGOFS.
Status and Evaluation
Following award of the contract, OSC and SBRC proceeded with the design and development of the spacecraft, payload, and associated ground system. The schedule was extremely aggressive, but project participants expected that application of commercial practices and streamlined oversight would make it possible to meet the delivery date. The critical payload had a 2-year delivery goal—nearly unprecedented for the development of a new space-qualified electro-optical sensor.
The payload contract was let in May 1991; 24 months later, the completed SeaWiFS instrument was ready for delivery. The sensor met the performance specifications that had been established at the outset of the contract. However, the test data revealed that the sensor had significantly higher off-axis response which would degrade the scientific utility of the data. This posed a challenging technical and business dilemma. Although a specification-compliant sensor had been produced, the spec had failed to capture an important aspect of performance desired by the scientific community. If this had been a typical cost-type development program, the cost of subsequent work to further improve the instrument's performance would have been borne by the government if such an improvement was desired. Since this was a firm fixed-price commercial contract, and the letter of the payload specification had been satisfied, the payload contractor had no further obligation. However, after several discussions with NASA and the scientific community, OSC and Hughes decided to proceed at their own expense with sensor modifications to improve the instrument's performance. SBRC proceeded to incorporate sensor modifications that substantially improved the performance of the instrument to the satisfaction of the science community. The modified and retested instrument was delivered to OSC for spacecraft integration in December 1993.
Subsequent problems with spacecraft development at OSC led to protracted schedule delays, and the sensor remained in storage awaiting completion of the spacecraft. More recently, problems with the Pegasus XL launch vehicle have caused further delays. The SeaStar spacecraft remained on the ground nearly 6 years after contract award and more than 3 years after delivery of the payload. On August 1, 1997, the SeaStar (now known as OrbImage-2) spacecraft was successfully launched on board the Pegasus XL. Shortly thereafter, SeaWiFS began to deliver data and has now produced a nearly continuous time series of high-quality ocean color measurements since October 1997.
A number of salient points emerge in reviewing the SeaWiFS development initiative:
The program reflected one of the classic flaws of many programs: A tremendous amount of time (nearly 5 years) was consumed in discussion, debate, and evaluation of program alternatives, leading to a highly compressed 2.5 year schedule for actual program execution—a schedule that ultimately proved unrealistic.
There is substantial risk is proceeding with unproven, untested designs for space-borne systems and particularly for launch vehicles.
The demonstrated ability of a contractor to build and fly space systems is the best assurance of success.
Cooperative international research programs (such as JGOFS) cannot depend (or be predicated) upon the timely availability of data from developmental systems.
It is difficult, if not impossible, to craft a specification that adequately constrains and establishes suitability of a product for its intended use. As a result, there is no substitute for close cooperation between data users and system builders. This ultimately worked well during the sensor development process on SeaWiFS, but there were some difficulties along the way in reconciling commercial and scientific imperatives.
It is difficult to develop business arrangements, data policies, and operational protocols that concurrently satisfy government, industry, and academia. Hybrid structures such as government-industry partnerships and advance data purchases become problematic when one or more of the parties cannot deliver as promised.
It is possible to develop space instruments on an accelerated schedule by adopting efficient design and oversight practices. Two years from contract start to flight hardware (31 months after performance-enhancing modifications) is much shorter than the traditional 4- to 5-year time line for payload instrument development programs.
SSTI (LEWIS AND CLARK)
SSTI was developed by NASA's Office of Space Access and Technology to advance the state of technology and reduce the costs associated in the design, integration, launch, and operation of small satellites. In July 1994, NASA awarded contracts to both TRW and CTA Space Systems to design and launch small Earth observing satellites named Lewis and Clark, respectively. Both contracts called for substantial new technology infusion into both payload and spacecraft bus, and for delivery of the satellites to launch within 24 months of contract start.
Both missions were unsuccessful. In the case of Lewis, the satellite development was completed within the allotted 24-month period; and, after a 1-year delay before its Athena 1 launch vehicle was deemed flight ready (see Appendix C), was successfully placed into its initial orbit in August 1997; the satellite, however, was subsequently lost. The Clark mission suffered excessive schedule delays and projected cost growth, ultimately leading to termination of the contract. Hopefully, much will be learned from their respective failures.
A retrospective on the Lewis mission was provided by NASA in a synopsis of the report by the NASA-commissioned Lewis Spacecraft Mission Failure Investigation Board (1998). The report indicates that NASA's
Earth orbiting Lewis spacecraft failed due to a combination of a technically flawed attitude control system design and inadequate monitoring of the spacecraft during its crucial early operations phase.
Lewis was launched on August 23, 1997, with the goal of demonstrating advanced science instruments and spacecraft technologies for measuring changes in Earth's land surfaces. The spacecraft entered a flat spin in orbit that resulted in a loss of solar power and a fatal battery discharge. Contact with the spacecraft was lost on August 26; it then reentered the atmosphere and was destroyed on September 28.
The design of the Lewis attitude control system was adapted by TRW from its design for the system on the TOMS-EP spacecraft. The failure board found that this adaptation was done without sufficient consideration for applying the system's design to a different primary spacecraft spin-axis orientation on Lewis. As a result, minor rotational perturbations, possibly due to small imbalances in the forces produced by the spacecraft's attitude control thrusters, caused the Lewis spacecraft to enter a spin. This situation eventually overloaded the spacecraft's control system while it was in a safehold mode. Prelaunch simulation and testing of the spacecraft's safehold modes also were flawed because they failed to analyze this possibility.
The combination of these errors with the subsequent assumption that a small crew could monitor and operate Lewis with the aid of an autonomous safehold mode, even during the initial operations period, was the primary cause of mission failure, according to the failure board's report.
The failure board also assessed the role of the "faster, better, cheaper" project management approach in the Lewis program. "The Lewis mission was a bold attempt by NASA to jumpstart the application of the faster, better, and cheaper philosophy of doing its business," said Christine Anderson, chair of the failure board and director of space vehicles for the U.S. Air Force Research Laboratory at Kirtland Air Force Base, New Mexico (Isbell, 1998). "I do not think that this concept is flawed. What was flawed in the Lewis program, beyond some engineering assumptions, was the lack of clear understanding between NASA and TRW about how to apply this philosophy effectively. This includes developing an appropriate balance between the three elements of this philosophy, the need for well-defined, well-understood, and consistent roles for government and industry partners, and regular communication between all parts of the team.''
Dr. Ghassem Asrar, NASA associate administrator for Earth Science, said, "The Lewis failure offers us some valuable lessons in program management and in our approach to technical insight. Lewis was an extreme example of allowing the contractor to have engineering autonomy. In the end, however, NASA has the responsibility to assure that the project objectives are met, and our assurance process was ineffective in this case. NASA's Office of the Chief Engineer is developing general lessons learned from this project and other faster, better, cheaper efforts, and we intend to apply them vigorously to all of our future missions, including the second generation of spacecraft in the Earth Observing System."
NASA issued a press release on February 25, 1998, announcing the termination of the Clark Earth science mission. The mission was terminated after an investment of some $55 million "due to mission costs, launch schedule delays, and concerns over the on-orbit capabilities the mission might provide" (Steitz, 1998). NASA is retaining Clark's launch vehicle services (an Athena 1 expendable launch vehicle). At the time of termination, NASA's contract for Clark was with OSC, which had earlier acquired CTA Space Systems to whom the contract had originally been awarded.
NASA's QuikSCAT mission was developed in response to the loss of the NASA Scatterometer (NSCAT) upon the failure of the ADEOS spacecraft. NSCAT data had proven its value in weather forecasting, and a replacement source was desired as quickly as possible. NASA was able to configure a replacement mission with a planned launch late in 1998—only 15 months from the start of the effort.
This very aggressive schedule was possible as a result of special circumstances. The Jet Propulsion Laboratory (JPL) would provide a scatterometer based on the future SeaWinds mission design by using available
scatterometer hardware (with subsequent replacement for SeaWinds). Similarly, Ball Aerospace committed to provide a spacecraft bus in only 11 months using existing hardware which had been planned for an Earthwatch satellite (Quick bird) that had been put on hold following the loss of Ball's Earlybird. JPL used GSFC's Rapid Spacecraft Acquisition initiative (Chapter 4) to procure the RS2000 spacecraft bus quickly from Ball.
QuikSCAT was successfully launched in June 1999.
Isbell, D. 1998. Lewis Spacecraft Failure Board report released. Release 98-109. Washington, D.C.: National Aeronautics and Space Administration.
Lewis Spacecraft Mission Failure Investigation Board. 1998. Available online at <http://arioch.gsfc.nasa.gov/300/html/lewis_document.pdf>.
Steitz, D. 1998. NASA terminates Clark Earth Science Mission. Release 98-35. Washington, D.C.: National Aeronautics and Space Administration.