That's one of the things that Squyres has mentioned in the past iirc. Usually, at the end of a talk, people will ask 'how old are the rocks' - and he replies with something alluding to the fact that the instruments to do that science, no one's figured how to put them in a shoebox yet.

Doug

True. But you'd be amazed at the number of theories that floated about which had the feel of last-ditch attempts to either shoot down the magma ocean concept (the only way in which the lighter anorthositic rocks could have separated out and floated to the surface to form a lunar crust) or to maintain some cherished pet theory (that the Moon must be completely chondritic, that it must have never experienced differentiation, or even that it must have once sported real seas and oceans).

It took a combination of orbital sensing (plotting aluminum concentrations from orbit using the J-mission SIM bay spectrometers) and ground truth (sampling a wide variety of anorthositic rocks) to finally prove the existence of an anorthositic original lunar crust, and to banish once and for all the defensive, "no, the Moon must be exactly as I have envisioned it" theories that grasped at more straws than you'll find in a barnyard in order to explain away the feldspar-rich fragments found in mare soils.

That whole episode reminds me a lot of the current state of speculation on Martian geology and planetary genesis. There are literally hundreds of theories out there about Mars which try to explain away ubiquitous observed phenomenah as the results of rare and non-characteristic "special circumstances." As with the resistance to accept an anorthositic crust on Luna, which requires acceptance of a magma ocean, there is still resistance to a Mars that was once warm and wet enough to have generated the water flow features observed there and which leads those who resist to postulate very transitory special circumstances to account for these features. It may require the discovery of, for example, a hidden cache of carbonate rocks to finally pin down the details of Mars' early history -- and it may require far more than "a few" samples to uncover those rocks. And that history.

-the other Doug

Let's assume that all instruments can be brought to Mars, and that we can do the laboratory measurements as well there as here. (Obviously not the case for many measurements, but folk like to argue about those comparisons even though it's not all that relevant to what in my mind is the strong argument for sample returns, which follows.)

Let's say we can do ten such measurement types per rover on multiple locally collected samples. To first order, we can figure that one such set of locally collected samples brought back to Earth will be the cost of, say, five rovers. Then if we want to do 50 such measurements, it's a wash. If we want to do 100 such different measurements, then the sample return is a win. (The Earth measurements will be so much cheaper than the mission, I neglect that cost here, but if you include that cost, the answer will be about the same.) I'd like to know how many different types of analyses were done on the Lunar samples after return. Just in the first three years. I'd wager it's in the thousands.

Now let's look at schedule. This is the real kicker for me, since I'm impatient. Whenever we send an instrument somewhere, it answers some questions, and raises many more new questions. For Mars, we could respond to that with a new in situ instrument at best three opportunities, or 6.5 years later. It happens again, another 6.5 years. And so on. The pace of discovery can be excruciatingly slow, even if you're lucky enough to have something like a Mars Program that keeps sending things there. (I've been through a few of these cycles, and it can really wear you down.)

However, if you have samples here, you do a measurement. It raises questions. You do a new experiment months or weeks later in your own lab or in your colleague's different lab with different equipment. You publish a paper and within a year other labs are answering more questions you raised. While you're waiting for the next Mars in situ mission to be developed and flown, our knowledge of Mars has been completely revolutionized by the equivalent of dozens of in situ missions. The cycle time of discovery is one to two orders of magnitude faster.

So in my mind the strongest arguments for Mars sample return are simply cost and schedule. Sample return is much more cost effective and much more schedule efficient for extensive laboratory investigations than sending in situ instruments to Mars. Even if I completely set aside the issues of packaging entire Earth laboratories in a shoe box, and even if I completely set aside the continuing improvement in the capabilities of Earth laboratories and their application to already collected samples.

OD, nice history of resistance to the lunar magma ocean hypothesis, but might you have the "warm, wet" Mars argument slightly backwards? After all, was early Mars ever much bigger or closer to the Sun than it is now? Could early Mars (except for abundant ice and an atmosphere) have been more like the Moon than most here prefer to envision it, in that all of those craters we see ("ubiquitous observed phenomena") have something fundamental to do with its history (including rock layering, spherules, water flow channels, crater-filling lakes, subsurface clays, abundant salts, and so on)? Did the LHB (a.k.a. "lunar cataclysm") affect Mars, at the same time as it apparently affected the Moon, Earth, Venus, and Mercury? Would such a "hot" bombardment event tapering off at about 3.8 Ga qualify as your "rare and non-characteristic 'special circumstance'" (analogous to the special circumstance that appears to have, e.g., formed the Moon or killed off the dinosaurs)? Just another hypothesis, already discussed in other posts and publications. By all means, Mars Sample Return!

-- HDP Don

I'd bet that there were less than a dozen critical measurements made (age dating and isotopic abundance stuff, primarily), and a whole bunch of essentially trivial and unimportant ones. But I'm not a geologist.

I've been doing pretty much nothing but Mars stuff for twenty years, and I, perhaps more optimistically, view it as job security.

I think you make a very good argument, but I think to really do the tradeoff you have to have a better idea of the kinds of measurements you want to make at a given time and how best to make them. It's often hard to sell big expenditures without knowing what you hope to learn more specifically. It would be more clear-cut if sample return could be done for the cost of, say, 3-4 rover missions, instead of the current estimates, which are in the tens at least.

Has anyone studied the costs and feasibility of putting testing equipment in orbit around Mars vs. all the complications of atmospheric entry and landing? I would venture that you could put more equipment and larger equipment and more fragile equipment into orbit in some kind of automated Mars lab than you ever could land on the surface. The sample retrieval then becomes one of merely getting the sand and rocks up to a defined orbit and nothing more. It would have the benefit of sample analysis from multiple locations at a much smaller cost than either multiple sample returns or multiple rover/lander-labs.

Just shooting from my armchair here, though Mark makes a real good case for Earth sample return.

I know that a case has been made that a human mission to Mars might be justified even if you never landed the humans on the surface. They could hang out in their pressurized, non-dusty, climate-controlled station in orbit (or on Phobos) and, without the hassle of lengthy two-way light-times, or the risk, complexity, and cost of landing and eventually taking off again, they could joystick vehicles across the surface, retrieve samples for analysis, etc. But I seriously doubt that an early human mission to Mars will include more than a few "shoebox-sized" pieces of science equipment, so that doesn't really solve the problem we're discussing here.

--Emily

"Cost effective"? Well, I guess that all depends on whether putting a large number of eggs into one rather expensive basket can be said to be"cost effective"!

Some on this thread seem to be expecting a Rolls-Royce mission. For example, back on February 22 JRehling posted "I'd say the right architecture would be to drop landers onto several top sites with the capability of blasting their samples into orbit. The number of samples so launched may exceed the capacity of the return vehicle to return. Let's say five samples were launched into Mars orbit, with the choice perhaps made of which three should be sent back to Earth."

If each of those landers were carrying a rover the size of the MSL or ExoMars ones (as others on this thread seem to have in mind) then on the face of it we're talking about multiple launch vehicles just to get to the sample collector missions to Mars. As far as I can discover NASA itself is, at least at present, only envisaging a single lander and a single rover. (Just as nobody seems to be suggesting there will ever be an MSR 2.)

In any case, if NASA didn't have the money to send 2 MSLs in 2009 what are the odds it will have enough to send 3+ in 2018 (or whenever)? Without the money then unless Congress and the president suddenly become more generous than they seem to be at the moment (see this graph which Ryan Anderson mentioned over on Martian Chronicles site) it will surely only be able to do so by cannibalising other areas of its Mars program (as indeed some at the recent MEPAG panel foresaw; see below); and maybe even other non-Mars unmanned missions, especially flagship ones.

IMHO MSR will be lucky to get off the ground by 2018, let alone with multiple rovers. According to the Science article that Vikane summarised on this thread back on February 29 on the MEPAG panel "[e]arly estimates put the overall cost of such an ambitious mission at $5 billion to $6 billion". Given that new technology especially for the sample return part of the mission will have to be developed, then given the given past blowouts in cost for such things (eg MSL) I will not be surprised to find the cost will be more like $8-10+ billion dollars by the time the thing actually flies.

Even $6 billion dollars is an awful lot of eggs to be committing to a single basket. If $1-2 billion dollars is reckoned as the cost of a typical NASA flagship mission then that $5-6 billion represents 3+ flagships all rolled up into one, albeit one flying at least two missions (a sample collector and sample retriever/return). NASA will doubtless be under enormous pressure to minimise the costs, not just from the usual budget cutters in Congress but maybe also from some of its own engineers and scientists who will fear that the MSR will become the unmanned equivalent of the space shuttle and the ISS: a project that will suck funding from (in this case: other) unmanned missions, especially those of the flagship class. Multiple rovers will doubtless be among the first things jettisoned to lighten the load and reduce the cost.

I mention that because the news gets worse. In what I presume to be the same article in Science that vikane summarised there is a rather sobering diagram summarising NASA's likely Mars missions for 2010-2020. Of the 5 mission opportunities over that period, two are for MSR (2018 and 2020), one has no mission at all (2011), one is reserved for the next Scout mission (2013), and the fifth is for either the Astrobiological Field Lab or 2 midsize rovers.

In other words, the shifting of the next Scout mission from 2011 to 2013 coupled with the bringing forward of MSR to 2018 has meant if the MSR lander does fly in the 2018 slot there will now be just one other US Mars mission between MSL in 2009 and MSR in 2018. Where before three missions (MSO, AFL, and the mid-sized rovers) were competing for at least two slots those same three are competing for one; and already one of them appears to have lost out judging from that diagram: MSO.

Back last year Alan Stern seems to have been non-committal back when MEPAG asked him "if there were one strategic mission opportunity between the
2011 Scout and MSR, would the Mars community choose MSO...Dr. Stern said he would request from MEPAG an analysis of the issues".

As the Science article goes on to point out (and vikane reported previously) that may not be the end of the bad news. "The group determined that NASA would have to cancel everything after MSL--including the 2013 Scout and the 2016 missions--to fly a sample return by the second half of the next decade."

To have no US Mars mission of ANY kind launched after 2009 until 2018 (when the first half of the MSR goes up) would surely be a very heavy price to pay to have a few pounds of rock conveyed back from one particular part of Mars. If MSR does end up being the lone survivor then if for any reason the MSR mission itself fails it will be the 1980s all over again: a decade lost to Mars exploration followed by an expensive all-eggs-in-one-basket mission which tanked.

Is the MSR truly that vital a mission?

I notice that losing the MSO in particular is potentially self-defeating for the MSR project itself. Without MSO then unless the MSR also contains an element which stays in orbit to act as a telecommunications relay (or alternately NASA revives the Mars Telecommunications Orbiter as a separate mission) the MSR rover will be reliant on direct Earth communications and on a fleet of aging orbiters (MRO, the youngest, will by then be well over a decade old).

Finally a query. Do the rovers which get sent to MSR collectors really need to be MSL/ExoMars ones? Could upgraded MERs do the same job, perhaps using some of the MSL technology? I mention that for two reasons:

* For those who want multiple rovers, multiple MER-class ones may be more achievable cost-wise than multiple ones of the MSL-class sort.

* I understand that currently there is a world-wide plutonium shortage that will take some time to redress. "After MSL launches, we're pretty much out of plutonium," Space.com quoted Mike Griffin back last month. Even Russia is apparently down to its last 10 kilograms. Unless and until that shortage is addressed there may not be enough for even one MSR rover. Which means solar power would need to be used which in turn will weigh against large rovers like MSL ones (as well, of course, as reducing the parts of Mars and the times of year a rover of any sort can rove in, which in turn will doubtless impact on the chances of finding the kinds of results many people seem to expect from such a mission).

======
Stephen

The more I contemplate the cost versus benefits, the more inclined I become to drop MSR indefinitely in favor of multiple rover missions. For that matter, fond as I am of Mars exploration, post-MSL I would be inclined to trade any 2-3 Mars rover missions for 1 non-Mars rover. Places like Ceres, Vesta, Europa, Callisto, Ganymede, Titan, and Triton could each use one or more rover missions, though personally I would put Europa and Titan at the head of the list, closely followed by Ceres. Ceres is likely the cheapest option; planning a Ceres rover to arrive in time to be supported by the Dawn orbiter might just be a heck of a good idea.

Is having MSR a necessary prerequisite for manned exploration of Mars? If Mars gets put back into the Exploration program, there will surely need to be unmanned precursor missions to test equipment...ISRU, etc. A manned mission would surely bring back tens if not hundreds (?) of pounds of samples...could we just not skip MSR and push stronger towards manned exploration? If MSR is looking at only bringing back 1-10lbs for $5-10 billion, would it not be just as cost effective to bring back 100+ lbs for $100 billion?

[...]

There's obviously a trade-off. The rover (mobile) or in-situ (e.g., orbital or on-the-ground) lab approach yields lots of imprecise, necessarily ambiguous data from a handful of instruments for a large number of sites or samples. People here, including me, love that the rovers provide superb imaging (i.e., geological context) for each site. The MSR approach yields wonderful, incredibly precise, incredibly detailed geochemical, isotopic, microscopic, mineralogic, crystallographic, etc. data (independently confirmed at multiple laboratories) on a relatively small number of samples, whose geological context may be less well constrained. Given severely limited budgets, we have to choose. If we choose the MSR approach, we have to be super-careful about geological context.

My conceptual problem - I'm not sure that 20 rovers, imaging and crudely analyzing various occurrences of cross-bedded salty fines and spherules, would unambiguously tell us much more than we know now after imaging and analyzing such rocks at 2 sites (and, in the case of the respective MER teams, coming up with completely different interpretations). If each rover studied different rocks, that would obviously be ideal. That ideal is, however, difficult to achieve when engineering constraints restrict landing sites largely to equatorial flat areas, whose surfaces tend to be dominated by deposition/erosion via late impacts and the wind. Geologists long to study more challenging features such as "young gullies" formed only on steep slopes and the layering exposed in steep canyon or crater walls. Can the present generation of rovers, however long-lived, ever satisfy that longing?

-- HDP Don

UNNECESSARY QUOTING REMOVED

Don, excellent arguments, and I believe those that the Mars community (you may be one of them!) are making. Spacecraft instruments simply are crude, and there's a limit to the finess of the of hypotheses you can test with them. The the question is, are we willing to spend $3-5B for those answers? And do we know enough yet to know where to send that one sample mission?

Also, remember that there are four separate craft that must work: Lander, rover, ascent vehicle, Mars orbiter/Earth return craft. The level of risk for this mission will always be high.

The scientist in me wants samples. The mission manager worries about the impact on the rest of the program and the consequences of failure.

Great points. Vicarious exploration via imaging rovers (and to a lesser extent, via orbiters) is certainly personally addictive, scientifically rewarding, and less risky and less expensive than MSR. Unfortunately, the results may always be ambiguous (as in any scientific investigation), owing to instrument limitations. The undeniable emotional appeal of MSR (to me at least) is to remove some of that ambiguity via unlimited instrumentation, a hope that may be somewhat unrealistic, given the history of prior investigations. The cost and risk may also be unacceptably high, as you point out.

Where and what to sample might be the most difficult choice of all, given the large number of questions that need answering (e.g., just off the top of my head, modern life? if so, where? past life? if so, where? past oceans and lakes? nature of past sedimentary deposits? past warm water carbonates? past sedimentary evaporites? past eolian rocks and desert oases? past acid seas or groundwaters? past hot springs? past hydrothermal or igneous ore deposits? past gossans (via weathering of sulfides)? past surface weathering to clays? clay oxidation? modern brines or meltwaters? modern soil and dust mineralogy? salt mineralogy? states of salt hydration? salt efflorescence? salt leaching by frost? types of volcanism? types of igneous rocks and their origin? types of glacial deposits? past climate history revealed in polar layered deposits? types of polar ices, including clathrates? impact-related spherules? impact melts and metamorphism? impact surge deposits near rampart craters or elsewhere? surface peroxides or super oxides? surface silica coatings? desert varnish or caliche analogs? dangers to potential colonists? resources for potential colonists? etc. etc.).

-- HDP Don

Or, to make a virtue of what is increasingly looking like necessity: the chances of an MSR from the most interesting (say) 0.1% of the accessible surface are greatly improved by a prolonged remote sensing campaign over several more decades - beyond 2020.

The other way to cut the Gordian knot is simple: more money! What struck me about the budget chart on MartianChronicles is the arbitrary whims with which the powers that be seem to give and then take away. The 2010 figure's gone from $1200m to $600m to $300m in three successive years. Who knows, if this particular bunch of congresscritters can decide to slash the budget by that extent, in a year or two (or five) another selection (or the same ones!) might equally quadruple the budget in the same way?