2011-07-22 08:48:41Rignot 2011 overestimated Antarctic ice loss?
Alex C

coultera@umich...
67.149.101.148

Reported at Science, here's the reanalysis.

Quote from report (emphasis mine):

"Zwally and Giovinetto reassessed pub- lished altimeter and gravity estimates. They also recalculated an earlier input-minus-output estimate published by radar scientist Eric Rignot of NASA’s Jet Propulsion Laboratory in Pasadena, California, and his colleagues. The pair challenged some of Rignot’s assumptions, including some underlying his estimate of ice loss along the 15% of the ice sheet’s periphery where no one had made measurements. And whereas Rignot had estimated all East Antarctic snowfall from a weather model, Zwally and Giovinetto substituted field measurements for 6% of the area. Those changes reduced the loss estimated using input-minus-output by 90%."

Alley also weighed in a bit for the article, here's something he had to say, with extra context:

"The new analysis is a “perfectly reasonable reinterpretation,” glaciologist Andrew Shepherd of the University of Leeds, U.K., says. “The paper’s main contribution is a very convincing argument that one needs to account for uncertainties in a consistent way.” Not that this is the final word. “I wouldn’t say this shows Eric is wrong,” Alley says, “nor that Jay is wrong.” In fact, Shepherd says, “we’ve always felt the number is somewhere in between” the extreme estimates."

2011-07-22 13:57:45
Albatross
Julian Brimelow
stomatalaperture@gmail...
199.126.232.206

Some "good" news fopr a change.

This is clearly a rapidly evolving and developing field and the new Zwally paper is by no means the final word.

2011-07-22 21:45:11Comment
Robert Way

robert_way19@hotmail...
142.162.27.99

I'm not a big fan of this article to be honest. Zwally and Givenetto have far too much fair in their ERS-radar altimetry. The fact that they did not even once mention Thomas et al (2008) shows this. They argue that the other techniques have errors and when fixed that they support their own preferred technique the altimetry method. In their discussion of the method they discuss how there have been errors in it for 3 reasons in the past. They explain away two of the errors but neglect the single biggest contributor to ERS errors that basically nullifies their argument. Thomas et al (2008) showed that in steep/dynamic areas ERS significantly underestimated ice loss in Greenland. In fact it was by greater than 50% that it underestimated. Another problem is that radar altimetry (until cryosat 2) had a coarse spatial resolution of around 1 km. That means you can't see the dynamics of what is going on inside that 1 km. For regions such as the Antarctic peninsula this is very important because most glaciers are smaller than that in that region. Plus we have direct evidence of tidewater glacier retreat across 90% of the Antarctic peninsula... and the extreme acceleration after Larsen B collapsed.

Another argument I find weak is the weather model one. The model used by Rignot et al (2008) was created using field measurements and that was how it was calibrated. They look at 6% and say that data is unreasonable so they put in field data. That's not how things work with models. Over the whole ice sheet the model should give a realistic depiction. Localized regions will undoubtedly have some outliers.

2011-07-23 02:07:38
Albatross
Julian Brimelow
stomatalaperture@gmail...
199.126.232.206

Robert, thanks for that interesting review.  So maybe the news is not so good-- but I'm sure the 'skeptics' will like this paper for all the wrong reasons.

IO tried to speed read the paper, and that was a mistake, becasue I did not really understand how they incorporated field precip measurements or what the nature of these measurements was (I'm sure it is all in there, I just found the paper very "busy".

I had forgotten about the problem with detecting ice loss on steep margins using the altmitery data-- ignoring that is huge, especially for Greenland and the Peninsula IIRC

When I was at Halley we had an array of stakes that we frquently meaured to detemrine the accumulation (or loss) of snow.  That seems the best way b/c gauges simply do not work in polar environments. That said, it was for a point, and while I share the author's desire to use in-situ data over model data they seem to ahve gone about it in a rather odd way. 

I said earlier that this paper will not be the final word-- hopefully Rignot et al will publish a rebuttal.

Hopefully Cryosat can shed some light on this....

2011-07-24 05:43:16
MarkR
Mark Richardson
m.t.richardson2@gmail...
92.24.250.176

I encountered Andrew Shepherd last week.

He seems like a very good, skeptical scientist and he gave a great talk on his Greenland work, showing how efficient drainage has helped slow glacier flow recently. It's promising news, but he made it clear that there's lots more work needed before we'll know if we're out of the woods in terms of Greenland ice sheet collapse...

2011-07-24 05:46:23Comment
Robert Way

robert_way19@hotmail...
142.162.27.99

Oh I think he's a very good scientist. I just challenge his reliance on altimetry data to attack other methods when the radar altimeters themselves have been shown (by thomas et al. 2008) to significantly underestimate losses in dynamic regions.

2011-07-24 06:07:57
Alex C

coultera@umich...
67.149.101.148

I don't think Shepherd was involved with the reanalysis, just that he agrees it is a good interpretation (unless that's not what you were implying).

2011-07-24 08:18:16Comment
Robert Way

robert_way19@hotmail...
142.162.27.99

Yeah I initially thought Shepherd was involved in the reanalysis. Either way the reason I associated him with it is because his previous work has used radar altimetry surveys just like the reanalysis. I stand by what I said though regarding his views on the methods used.