Earlier today in part1, I posted about the new record low claimed by NSIDC: Sea Ice News – Volume 3 Number 11, part 1 – new Arctic satellite extent record. The number given is 4.1 million square kilometers:
That of course is being trumpeted far and wide, new life has been given to Mark Serreeze’s “Arctic death spiral” in the media. But, here’s a curiosity, another NSIDC product, the new and improved “multi-sensor” MASIE product, shows no record low at ~ 4.7 million square kilometers:
Note the label at the bottom of the image in red. NSIDC doesn’t often mention this product in their press releases. They most certainly didn’t mention it today.
Another product, NOAA’s National Ice Center Interactive Multisensor Snow and Ice Mapping System (IMS) plot, also shows no reason for claiming a record at all:
Their number is (for 8/22) ~ 5.1 million square kilometers. (NOTE: NSIDC’s Dr. Walt Meir points out in comments that IMS and MASIE use the same base data, but that this one product from IMS only updates weekly, unlike all other sea ice plots which are daily. They should be in sync on the next update cycle, but right now MASIE and IMS should both be at 4.7 million sqkm. -A)
Another curiosity is here. On the NATICE interactive maps on demand page (click on Arctic Daily in the pulldown menu):
The numbers they give for 80% and marginal ice add up to an extent of 6,149, 305 square kilometers.
So who to believe? It depends on the method, and who thinks their method is most representative of reality. Measuring sea ice via satellite, especially when you use a single passive sensor system that has been show in the past to have degradation problems and outright failure (which I was told weren’t worth mentioning until they discovered I was right and pulled the plug) might be a case of putting all your eggs in one basket. I suspect that at some point, we’ll see a new basket that maybe isn’t so worn, but for now, the old basket provides a comfort for those who relish new records, even though those records may be virtual.
Note that we don’t see media pronouncements from NOAA’s NATICE center like “death spiral” and “the Arctic is screaming” like we get from its activist director, Mark Serreze. So I’d tend to take NSIDC’s number with a grain of salt, particularly since they have not actively embraced the new IMS system when it comes to reporting totals. Clearly NSDIC knows the value of the media attention when they announce new lows, and director Serreze clearly knows how to make hay from it.
But this begs the question, why not move to the new system like NOAA’s National Ice Center has done? Well, it is a lot like our July temperature records. We have a shiny new state of the art Climate Reference Network system that gives a national average that is lower for July than the old USHCN network and all of its problems, yet NCDC doesn’t tell you about the July numbers that come from it. Those tasks were left to Dr. Roy Spencer and myself.
In fairness though, I asked Dr. Walt Meier of NSIDC what he thought about MASIE, and this is what he wrote to me today:
It can provide better detail, particularly in some regions, e.g., the Northwest Passage.
However, it’s not as useful for looking at trends or year-to-year
variations because it is produced from imagery of varying quantity and quality. So the analyses done in 2007 have different imagery sources than this year. And imagery varies even day to day. If skies are clear, MODIS can be used; if it’s cloudy then MODIS is not useful. Another thing is that the imagery is then manually analyzed by ice analysts, so
there is some subjectivity in the analysis – it may depend on the amount of time an analyst has in a given day.
Our data is from passive microwave imagery. It is not affected by clouds, it obtains complete data every data (except when there may be a sensor issue), it has only consistent, automated processes. So we have much more confidence in comparing different days, years, etc. in our passive microwave data than is possible using MASIE.
Finally, MASIE’s mandate is to try to produce the best estimate they can of where there is any sea ice. So they may include even very low concentrations of ice <15%. In looking at visible imagery from MODIS, in the few cloud-free regions, there does appear to be some small concentration of ice where MASIE is mapping ice and our satellite data is not detecting ice. This is ice that is very sparse, likely quite thin. So it will probably melt out completely in the next week or two.
MASIE has tended to lag behind our data and then it catches up as the sparse ice that they map disappears. This year the difference between the two is a bit larger than we’ve seen in other years, because there is a larger area of sparse ice.
You can thank the big Arctic storm of August 4th-8th for that dispersal.
“The Great Arctic Cyclone of 2012″ effect on Arctic sea ice is seen in this before and after image:
Figure 4. These maps of sea ice concentration from the Special Sensor Microwave Imager/Sounder (SSMIS) passive microwave sensor highlight the very rapid loss of ice in the western Arctic (northwest of Alaska) during the strong Arctic storm. Magenta and purple colors indicate ice concentration near 100%; yellow, green, and pale blue indicate 60% to 20% ice concentration.
Credit: National Snow and Ice Data Center courtesy IUP Bremen
Trends -vs- records, just like July temperatures. One system might be better at trends, another might be better at absolutes used to determine records. In this case we have three other respected methods that show absolute values higher than that of NSIDC’s older method which they have a high confidence in. I suppose these systems are like children. In a competition, you always root for your children over the children of the other parents, so it is no surprise that NSIDC would root for their own well known media star “child” over that of NATICE’s IMS and NSIDC’s own lesser known child, MASIE.
Oh, and then there’s Antarctica, that other neglected ice child nobody talks about, with its above normal ice amounts right now:
No matter what though, its all just quibbling over just a little more than 30 years of satellite data, and it is important to remember that. It is also important to remember that MASIE wasn’t around during the last record low in 2007, and IMS was just barely out of beta test from 2006. As measurement systems improve, we should include them in the discussion.
UPDATE: Andrew Revkin reports on the issue in his Dot Earth article here
He’s a bit skeptical of the sound byte hype coming from NSIDC writing:
That’s one reason that, even with today’s announcement that the sea ice reached a new low extent for the satellite era, I wouldn’t bet that “the Arctic is all but certain to be virtually ice free within two decades,” as some have proposed. I’d say fifty/fifty odds, at best.
But is this a situation that is appropriately described as a “death spiral”? Not by my standards.
Revkin also takes Al Gore to task on Twitter:
help him out, retweet this
UPDATE2: Commenter Ron C. provides this useful information in comments that helps explain some of the differences and issues:
The main point is that NIC works with images, while the others are microwave products.
“Polar orbiting satellites are the only source of a complete look at the polar areas of the earth, since their orbits cross near the poles approximately every two hours with 12 to 13 orbits a day of useful visible data. This visible imagery can then be analyzed to detect the snow and ice fields and the difference in reflectivity of the snow and ice. By analyzing these areas each day, areas of cloud cover over a particular area of snow and ice can be kept to a minimum to allow a cloud free look at these regions. This chart can then be useful as a measure of the extent of snow and ice for any day during the year and it can also be compared to previous years for climatic studies.”
“NIC charts are produced through the analyses of available in situ, remote sensing, and model data sources. They are generated primarily for mission planning and safety of navigation. NIC charts generally show more ice than do passive microwave derived sea ice concentrations, particularly in the summer when passive microwave algorithms tend to underestimate ice concentration. The record of sea ice concentration from the NIC series is believed to be more accurate than that from passive microwave sensors, especially from the mid-1990s on (see references at the end of this documentation), but it lacks the consistency of some passive microwave time series. ”
Some have analyzed the underestimation by microwave products.
“We compare the ice chart data to ice concentrations from the NASA Team algorithm which, along with the Bootstrap algorithm [Comiso, 1995], has proved to be perhaps the most popular used for generating ice concentrations [Cavalieri et al.,1997]. We find a baseline difference in integrated ice concentration coverage north of 45N of 3.85% ± 0.73% during November to May (ice chart concentrations are larger). In summer, the difference between the two sources of data rises to a maximum of 23% peaking in early August, equivalent to ice coverage the size of Greenland.”
From Late twentieth century Northern Hemisphere sea-ice record from U.S. National Ice Center ice charts, Partington, Flynn, Lamb, Bertoia, and Dedrick
The differences are even greater for Canadian regions.
“More than 1380 regional Canadian weekly sea-ice charts for four Canadian regions and 839 hemispheric U.S. weekly sea-ice charts from 1979 to 1996 are compared with passive microwave sea-ice concentration estimates using the National Aeronautics and Space Administration (NASA) Team algorithm. Compared with the Canadian regional ice charts, the NASA Team algorithm underestimates the total ice-covered area by 20.4% to 33.5% during ice melt in the summer and by 7.6% to 43.5% during ice growth in the late fall.”
From: The Use of Operational Ice Charts for Evaluating Passive Microwave Ice Concentration Data, Agnew and Howell