Thursday, March 6, 2008

Misleading headline of the day: sea level is falling!

Headline from National Geographic: Sea Levels to Plunge Long Term, Study of Dino Era Says.

It paints a scary view of the Cretaceous: forget T. Rex. Did you know that sea levels were 560 feet higher than they are now? That would make the tip of the Washington Monument barely above sea level! (And, for that matter, the high points of Florida, Delaware, and Louisiana would be only accessible by scuba diving!) That's right... and it's all natural - it's due to the sea floor pulling apart!

Umm. Actually, I did know that. In fact, where I live, at 6500 feet above sea level, I can go see the shoreline of the Cretaceous Interior Seaway that flooded North America. The high sea levels of the Cretaceous are hardly big news to geologists, and the role of mid-ocean ridges in controlling global sea levels was accepted by the 1980's, at least.

The story is about an article coming out in Science tomorrow. Science is supposed to published ground-breaking, cutting-edge research. They wouldn't publish something that's been accepted for decades, would they?

Of course they wouldn't.

The paper discussed by National Geographic (Muller, R.D., Sdrolias, M., Gaina, C., Steinberger, B., Heine, C. (2008). Long-Term Sea-Level Fluctuations Driven by Ocean Basin Dynamics. Science, 319(5868), 1357-1362. DOI: 10.1126/science.1151540) deals with a far more subtle question. We know sea level was high during the Cretaceous, but how high was it, exactly? Different methods of estimation give different answers. Estimations of mid-ocean ridge volume gave numbers in the neighborhood of 250 meters (820 feet - above the high point of Rhode Island). Sedimentary rocks in New Jersey, on the other hand, imply that the highest Cretaceous sea level was only 40 meters (131 feet) - less than the sea level rise that would occur if we melted all the ice on the planet. Why the discrepancy?

Muller and colleagues started by assuming the sea level rise was mostly the result of tectonics. The young rocks at mid-ocean ridges ride higher than the old, cold rocks of the abyssal plains, so the first thing they had to do was figure out just how old the ocean floor was in the Cretaceous. That's not that easy, because several oceanic plates in the present-day Pacific have gone down subduction zones and disappeared. Then the authors had to account for hot spots (like Hawaii), and for sediments, and for changes in the area of oceans versus continents (such as that caused by the widening of Nevada during formation of the Basin and Range). And finally, yes, they calculated how much sea level would rise if our ice caps all melted.

They conclude that Cretaceous sea levels were 170 meters (~560 feet) above present-day sea level.

So why do the sedimentary rocks from New Jersey give such a different answer?

They blame the mantle. Actually, they blame the Farallon plate - the oceanic plate that subducted beneath North America during the Mesozoic (and whose remnants are still sliding under the Pacific Northwest, Central America, and South America). The old Farallon slab is still down there, sinking... and apparently it is dragging New Jersey (and presumably the rest of the East Coast of the US) down with it.

I'm not sure how to evaluate the explanation for New Jersey's low elevation. If it's correct, it implies that the mantle drives major changes in relative sea level. Now, heat in the mantle is blamed for the high elevations in rift zones (such as the East African Rift), but I hadn't heard of ancient subducting plates dragging their overriding continents down. (Is there a similar effect in eastern South America? Is northern Asia affected by subduction of the Indian plate?)

Whether their explanation for New Jersey is correct or not, however, the National Geographic article completely misses the point. Yes, I know that "Discrepancy Resolved Between Cretaceous Sea Level Estimates" doesn't have the same ring. But the Science article does not say anything about future drops in sea level. It certainly doesn't try to extrapolate the trend 80 million years into the future, as the National Geographic article does:

When this trend is extrapolated out 80 million years from now, it suggests that even if all of today's ice caps were to melt, sea levels would be 230 feet (70 meters) lower than they are today.

This sounds like the advice of a bad stock analyst in the 90's: if it's going up now; it will always go up. And if it's going down now, it... will continuing decreasing forever.

My advice: if you are hoping for subduction of mid-ocean ridges to save your city from flooding due to global warming, don't hold your breath. Even if tectonics causes global sea levels to fall, the ocean floor isn't going to change very fast.

Edit: Reuters has an equally misleading headline. It sounds from the news release like the authors of the paper are claiming that the paper predicts the future. Hmmm. I guess it's possible to predict the future subduction of the East Pacific Rise and the Gorda/Juan de Fuca/Explorer Ridges. But is it possible to predict future hot spots/large igneous provinces, especially given that there's a debate about their origin? And do we know exactly why the Pacific plate appeared at the triple junction between three other oceanic plates? (Is it the result of some kind of triple-junction instability?)

Meanwhile, Ole has a good discussion of long-term versus short-term sea level changes. I can't comment on his site (it requires registration), so I'll say here: read it. It's a good discussion.

10 comments:

hypocentre said...

Mantle affecting sea level sounds sensible to me. The geoid is correlated with mantle velocities (harmonics 2-3 with lower mantle, 4-7 with upper)

Chuck said...

So would that require sea floor spreading to stop altogether? I'm too tired to reverse-envelope the volume of the mid ocean ridge as sea level equivalent.

Kim said...

Thanks, hypocentre. Out of curiosity - I've noticed that the passive margin of Africa seems to stand higher above sea level than the east coast of the US does. I look at it when I'm about to teach intro class lectures about active and passive margins, but I've never known the reason. I've wondered if the difference related to rifting (upper vs lower plate margins), but maybe the mantle makes the difference.

Chuck - I didn't try to calculate the possible causes of National Geographic's projected sea-level fall, but that's a good idea. (It's kind of funny to see that kind of blind extrapolation applied to the conclusions of a paper that was so careful to model every controlling factor the authors could think of.)

Silver Fox said...

I'm sorry to hear that the Farallon Plate is getting such negative press. I've always liked how it (probably?) caused the Rocky Mountains, (possibly?) caused extreme extention (core complexes, etc), and (maybe?) caused the Basin & Range. Too bad (?) it's making New Jersey sink

Eric said...

Great post - ol Nat'l Geo is sometimes fairly goofy.

I'll say up front that I havn't read the Muller et al. (2008) paper yet, so maybe they address the following issues.

Miller et al. (2005), the folks who did the New Jersey sea level estimate, used a 1-D backstripping technique; that is to say, they only looked at thickness variations in the core data they had, and relied on assumptions of coastal onlap and paleodepth to constrain the overlying water column thickness. They also admit that their backstripping technique can underestimate subsidence effects due to poorly constrained thermofelxural models of miogeoclinal wedges.

Additionally, Miller et al. (2005) claimed to have recognized Late Mesozoic high-frequency and fairly high-magnitude sea level variation in their data, which they interpreted as evidence for glaciers in the Cretaceous. Did this newest paper offer any comment on that conclusion, or where they more interested in long term sea level?

Elli said...

The Boston Globe also highlighted a Science article about revising the age of the erosion in the Grand Canyon today (http://www.boston.com/news/nation/articles/2008/03/07/new_study_revises_age_of_grand_canyon/). The Science article (which admittedly, I haven't read yet) is: Atkinson et al., Science 7 March 2008: Vol. 319. no. 5868, pp. 1343 - 1344 DOI: 10.1126/science.1155286
(somehow my small liberal arts college cancelled their subscription to Science a few years ago...so I have to be creative in getting access now)

Kim said...

Silver Fox: it really does seem like the Farallon plate is the tectonic scapegoat of North America, doesn't it? Somebody should really interview it and get its side of the story.

Eric - No, the Muller et al. (2008) paper didn't discuss the details of the New Jersey sea level history (and assumed the Cretaceous was ice-free - though there is also stable isotope evidence for ice in the Cretaceous now, isn't there?). The various models of the effect of the mantle on topography gave a wide range of subsidence amounts, too - from 100 to 300 meters. So I don't think that the models have a fine enough resolution to compare them to the details of the cores.

Elli - I've seen several headlines about the Grand Canyon. I couldn't remember the ages that the other groups (Joel Pederson, in particular) working on the problem found. If I have time, I'll read that paper too.

Maria said...

What part of the African margin are you looking at? The African superplume is responsible for some uplift but I'm not sure offhand how far north you can really spread the blame.

Kim said...

I'm thinking of the coast of the big bulge of West Africa. I've never been there, and my impression is based entirely from ocean-floor maps of the Atlantic. (Especially the famous Bruce Heezen/Marie Tharp map.) I always point out the east coast of North America as a typical passive margin, and then I look at how narrow the shelf on the matching African margin is, and I wonder. (But in intro classes, I don't wonder out loud, because the whole passive/active margin thing is confusing enough.)

Old Bogus said...

One problem I have with "blaming" the Farallon Plate is that there is no subduction zone on the east coast of North America even though the Mid-Atlantic Ridge is still spreading. So North America is "riding" on the back of this plate as it tries to get to Hawaii.

The result of this is that the Farallon Plate didn't have to travel much as we banged its head passing over before it could it out of the way (making it the victim here!). In fact, the NA Plate could be visualized as "eating" Farallon as it traveled SWward, impelled by the Mid-Atlantic spreading.

I think (probably a bad idea) of NA as a large, thick sheet brownie cookie which keeps getting squeezed and has developed long N-S cracks as it folds when push comes against stop: the Mississippi Valley is a low crack; the Rockies and the Appalachians high cracks, and the tilt of the Great Plains a manifestation of the compression's effects.

The Rio Grande "Rift" is the top of the high "crack" between the Plains and the Rockies pieces of the "brownie"; being very thick, the top of the crack appears to be opening, creating a graben. But the base of the "brownie" is firmly connected so any "spreading" is only on the surface.

Or I could be full of it and a flood of molten lava could burst forth anytime and come flowing down the Arkansas River valley past my house!