Mars Polar Lander: The South Polar Mission and the Investigation into Its Loss

Mars Polar Lander was a mission in the National Aeronautics and Space Administration's Mars Surveyor programme, paired with Mars Climate Orbiter in the 1998 project. It was intended to place a small lander near the edge of the south polar layered terrain to study volatile substances in the soil and local weather, helping reconstruct the history of water and climate on Mars. The planned landing area lay at approximately 76 degrees south latitude and 195 degrees west longitude.

The Jet Propulsion Laboratory managed the project for NASA, while Lockheed Martin Astronautics in Denver developed and operated the spacecraft. The mission carried two Deep Space 2 microprobes intended to penetrate the nearby surface at high speed and demonstrate the ability of miniature equipment to work on another planet. They formed an independent technology-demonstration mission and should not be conflated with the main lander's scientific payload.

References: [1] [2]

The complete flight system had a mass of approximately 576 kilograms, including a lander of about 290 kilograms. David Paige of the University of California, Los Angeles, led the Mars Volatiles and Climate Surveyor integrated payload, which included a stereo imager, a two-metre robotic arm with a camera, meteorological equipment and a Thermal and Evolved Gas Analyzer. The arm was intended to dig shallow trenches and deliver small soil samples to miniature ovens for heating, after which released gases would be examined to identify substances such as water and carbon dioxide.

A descent imager was to photograph the landing area after heat-shield jettison, while a Russian-provided lidar would detect dust haze and ice clouds in the lower atmosphere. The mission also carried a small microphone supplied by the Planetary Society. These instruments were prepared for approximately three months of work after landing. Continuous engineering telemetry was not planned during entry, descent and landing, so controllers would have to wait for post-landing communication periods to learn the outcome.

References: [1] [2] [3]

Mars Polar Lander launched on a Delta II rocket from Cape Canaveral's Launch Complex 17B at 20:21 Universal Time on 3 January 1999. Its cruise stage supplied power, communications and propulsion during the interplanetary journey, while the lander was enclosed in an aeroshell. After approximately eleven months of flight, it approached Mars on 3 December.

Mars Climate Orbiter had already been lost in September, eliminating the planned relay of data through that spacecraft. The polar lander could still communicate directly with Earth or relay data through Mars Global Surveyor. On arrival day, controllers performed a final trajectory correction approximately six and a half hours before atmospheric entry; telemetry indicated that the spacecraft was apparently still operating normally.

References: [2] [3]

At 12:02 p.m. Pacific Standard Time on 3 December, the spacecraft turned into its atmospheric-entry attitude, pointing its antenna away from Earth and ending the signal as planned. Controllers expected atmospheric braking, parachute deceleration and rocket-powered descent to follow, after which data transmission would resume. No signal arrived during the scheduled communications window. Neither the main lander nor either microprobe was contacted thereafter, and none returned the planned surface science data.

The search did not stop immediately. Mars Global Surveyor began photographing candidate landing areas on 16 December, while controllers repeatedly transmitted commands. The team announced the end of recovery efforts on 17 January 2000. A later Stanford University report of a possible faint signal prompted another round of attempts in late January and early February, but subsequent analysis showed that the signal originated inside the receiver rather than from the Mars lander.

References: [2]

In its report of 22 March 2000, the JPL Special Review Board judged premature shutdown of the descent engines to be the strongest explanation. Ground tests showed that deployment of the landing legs could produce a brief signal in the touchdown sensors. Software that had not ignored or cleared it as intended might have retained that transient as an indication that touchdown had already occurred.

Under this reconstruction, when touchdown detection was enabled to control the engines at an altitude of approximately 40 metres, the erroneous state would cause shutdown, leaving the lander without decelerating thrust. The board stressed, however, that no descent-flight data directly confirmed this sequence, so other failure modes could not be excluded. For Deep Space 2, investigators did not even identify a single most likely cause; the software explanation for the main lander could not simply be applied to the probes.

References: [2]

The report found that the touchdown software had not undergone the relevant testing in the complete flight configuration, and that a key restriction had been lost when requirements were carried into the software specification. Funding and schedule pressures had left some critical technical areas in the hands of a single individual. Some design-review findings had been formally closed without independent technical re-examination. These were the board's assessments of the development process.

It recommended adequate systems-engineering resources early in projects, continued peer cross-checks, and verification and fault-injection tests organized around actual flight conditions. The mission's lessons therefore went beyond a single programming error: the absence of telemetry during a critical stage both weakened fault diagnosis and prevented problems identified in ground testing from being matched individually to the actual final outcome.

References: [2]