Friday, November 28, 2008

Warring Views on MSL

Science magazine today published a letter from James Garvin, former chief scientist of NASA's Mars exploration program. Garvin strongly disputes Alan Stern's position on the Mars Science Laboratory (MSL) budget increases.

First, here are the latest public information as to the full cost of MSL. From a Science article on 9/25/08: "The science laboratory, currently slated for launch in the fall of 2009, is four times heavier than the current rovers trundling across the planet's surface. It features a plethora of advanced tools and instruments designed to analyze rocks, soil, and atmosphere. But that complexity has led to technical troubles and higher costs. When proposed in 2004, the lab was expected to cost $1.2 billion. By this summer, that price tag had climbed to $1.9 billion, and last week NASA space science chief Edward Weiler warned that "there is another overrun coming." Another NASA official put the latest increase at approximately $300 million."

Here is Stern's view from his letter to Science (10/31/08): "When the National Research Council's Planetary Science Decadal Survey recommended the Mars Science Laboratory (MSL) mission for priority funding, it assigned a cost level of $650 million. This value, rather than $1.4 billion, is the true metric for seeing the deep damage that MSL's profligately overrunning cost--now likely to top $2.1 billion--has inflicted on NASA's Mars and wider planetary science budget. Also, the story focused its overrun discussion on instrument costs. Although certainly part of the problem, instrument cost increases have been considerably smaller than overruns in the rest of MSL's budget, which was severely mismatched to the project's complexity from its inception. This mismatch sowed the most fundamental seeds of MSL's cost problems."

And Garvin's view from today's Science: "Stern also claims that MSL was "assigned" a cost level of $650 million. He fails to mention when and by whom. The $650 million cost was a placeholder assigned to a medium-class Mars rover mission by the National Research Council Solar System Decadal Survey committee in 2002, before NASA had developed a basis of cost estimate for MSL. This served as input to NASA studies from 2000 to 2004 to fully define the MSL mission and culminated in the competitive selection of its science payload in late 2004.

"At that time, the overall mission was baselined at a cost of $1.4 billion, not including several costs associated with the radioisotope power system. Given the experience with the cost of the Mars Exploration Rovers and the increased scientific and technical scope of the MSL mission, the so-called assigned value of $650 million is not credible. Stern's own New Horizons flyby mission to Pluto cost NASA more than $650 million; it is unrealistic to expect that a 700-kg analytical laboratory that must soft-land on Mars and drive around with 100 kg of scientific instruments could possibly cost less than a planetary flyby mission.

"Indeed, MSL's 2 years of intensive surface science operations are difficult to compare to any missions in the $650 million price class given typical science-per-dollar metrics. The established NASA cost to implement MSL as of the time of its confirmation review was $1.55 billion (August 2006), which grew due to NASA-wide issues with thermal protection system materials in 2007 to approximately $1.7 billion. The total cost growth of the MSL mission development since NASA confirmed the mission is typical of other Mars exploration missions successfully flown over the past decade. The cost to fly MSL in 2009 will be less than the cost (in today's dollars) of flying a nonmobile Viking Lander laboratory to Mars, and MSL includes a whole new generation of instruments and mobility."

My take is that the two letters talk past each other. ~$2.1B for a rover of MSL's capability is seems a fair price and money well spent for the science (and joy of exploration) that it will return. However, when the mission was fit into a roadmap of science missions by the Decadal Survey, it was as a $650M mission, which would have represented a modest increase in funding (on a per rover basis) over the MER rovers ($820M (per Wikipedia) for two rovers). The difference between that initial guesstimate and the likely final cost is $1.65B. The important question in my mind is whether or not the Decadal Survey would have included MSL in its roadmap if it has assumed that the true cost was >$2B. For only the cost difference, NASA could have flown 3 Discovery missions, 3 Scout missions, 4 MER rovers, or two New Frontiers missions. The problem appears not to be with the management of the MSL program, which as Garvin points out, has experienced cost increases in line with other Mars programs. The problem is with the way mission costs are estimated for Decadal Survey type exercises. NASA took the priorities of that Survey and retained them after it was clear that the cost estimate used in the Survey was off by over a factor of 2X. MSL went from a medium-class to a flagship-class rover. As I said, MSL will be money well spent for the science it returns. However, the Decadal Surveys will be meaningless exercises if they costs they assume in recommending the pieces of the roadmap are off by such large amounts.

No comments:

Post a Comment