Dot Earth on O10
Posted by Jeff Condon on December 30, 2010
Andy Revkin of Dot Earth has published a post on our Antarctic results among other things. Eric Steig has placed some far more detailed comments on our results on line with it.
2) The results appear to confirm all of the key results in our 2009 Nature paper, notably the significant warming in West Antarctica, with especially strong warming — very widespread and including all of West Antarctica and much of East Antarctica — in spring.
I do wonder what it means when a paper with half the calculated warming trend agrees with another but I’m just an engineer. The column marked S09 below is Steig’s result for comparison to the two primary methods demonstrated in O10. You can see the “all of West Antarctica” comment is interesting when the EW version came up with statistically insignificant trends. RLS though is probably more correct but compared to S09′s very high trend of 0.2C/decade or 2C per century it’s quite a bit lower. What it means though is that we can barely tell if any real warming happened at all vs the certainty of S09′s 0.2 +/- 0.09.
|Region||RLS C/Dec||E-W C/Dec||S09 C/Dec|
|Continent||0.06 ± 0.08||0.04 ± 0.06||0.12 ± 0.09|
|East Antarctica||0.03 ± 0.09||0.02 ± 0.07||0.10 ± 0.10|
|West Antarctica||0.10 ± 0.09||0.06 ± 0.07||0.20 ± 0.09|
|Peninsula||0.35 ± 0.11||0.32 ± 0.09||0.13 ± 0.05|
Another issue which needs to be looked at is the seasonality of trends. Ryan noted differences in the peninsula matching seasonal trends which again pointed to the blending of peninsula information across the continent in S09.
Well, I think Ryan said it better anyway:
Certainly other portions of S09 are confirmed by our paper (such as overall positive trends). However, we note that earlier studies also showed the same things, so these were not newly introduced with S09. The results that were newly introduced with S09, on the other hand, are all shown to be artifacts.
There is another comment by Dr. Steig which I would like to discuss:
4) The only other major difference in the findings, compared with our paper, is that we show significant warming over the eastern Ross sea region, while O’Donnell et al. show cooling, particularly in winter. I’m very skeptical of this result, because it is distinctly at odds with other assessments of trends in that region.
Addressing this properly would require a little time with the offset calculations to separate and plot Winter trends, I may do that later but for today it is worthwhile to simply place the offset reconstructions next to the results. The Ross region is in the southern portion of the graph attached. There is a bay like region which is covered with the Ross ice shelf year round. Note the white coloration of the temperature stations immediately to the West of this bay. This information is taken directly from temperature stations which have been offset such that averaging adjacent stations with trends don’t skew the results. Basically, it’s as simple a method as possible for knitting temp anomaly together reasonably. The actual data produces a nice red in Ross and nice White next door to the West.
Below is what the visually very similar O10 result looked like this for RLS:
Now this doesn’t address the Winter critique specifically but this shows that visually our methods spatially distribute the temperature station information with respect to the actual measured ground data. There is nothing in the method which would allow an accurate trend distribution overall and an inaccurate one in Winter. However, I believe the area weighted version may actually be more accurate than the satellite information methods due to distortions created by Chladni patterns. Small areas of the analysis can become exaggerated or minimized in trend based on their fit to the resonant modes of PCA decomposed spatially autocorrelated data. Basically, Steig may be partially correct but I’m skeptical of any large differences from the O10 and area weighted analysis.
Anyway, Nic had some particularly strong comments which are published in full at Dot Earth.