Slashdot Log In
Arctic Ice Extent Understated Because of "Sensor Drift"
Posted by
samzenpus
on Thu Feb 19, 2009 08:57 AM
from the give-it-a-few-taps dept.
from the give-it-a-few-taps dept.
dtjohnson writes "The National Snow and Ice Data
Center (NSIDC) has been at the forefront of predicting doom in the arctic as ice melts due to global warming. In May, 2008 they went so far as to predict that the North Pole would be ice-free during
the 2008 'melt season,' leading to a lively Slashdot discussion. Today, however, they say that they have been the victims of 'sensor drift' that led
to an underestimation of Arctic ice extent by as much as 500,000 square
kilometers. The problem was discovered after they received
emails from puzzled readers, asking why obviously sea-ice-covered
regions were showing up as ice-free, open ocean. It turns
out that the NSIDC relies on an older, less-reliable method of tracking
sea ice extent called SSM/I that does not agree with a newer method called AMSR-E. So why doesn't NSIDC use the newer AMSR-E data? 'We do not use AMSR-E data in our analysis because it
is not consistent with our historical data.' Turns out that the AMSR-E data only goes back to 2002, which is probably not long enough for the NSIDC to make sweeping conclusions about melting. The AMSR-E data is updated daily and is available to the public. Thus far, sea ice
extent in 2009 is tracking ahead of 2005, 2006, 2007, and 2008, so
the predictions of an ice-free north pole might be premature."
Related Stories
[+]
North Pole Ice On Track To Melt By September? 978 comments
phobos13013 writes "Recently released evidence is showing the North Pole ice is melting at the highest rate ever recorded. As a result, the Pole may be completely ice-free at the surface and composed of nothing but open water by September. As reported in September of last year, the Northwest Passage was ice-free for the first time known to man. The implications of this, as well as the causes, are still being debated. Are global warming experts just short-sighted alarmists? Are we heading for a global ice age? Or is the increase in global mean temperature having an effect on our planet?"
This discussion has been archived.
No new comments can be posted.
The Fine Print: The following comments are owned by whoever posted them. We are not responsible for them in any way.
Full
Abbreviated
Hidden
Loading... please wait.
Not consistent? (Score:5, Insightful)
What's the point of being consistent with a flawed methodology? I would have thought the thing to do would be to collect the new data, base newer model off that and then perform a statistically weighted correction to the older dater. Both data sets can be maintained if required.
Am not sure I see a point in consistency for consistency's sake, when you in the light of newer information you now know the original measurements are flawed.
Cheers,
Ian
Re:Not consistent? (Score:5, Insightful)
It really is frustrating how intensely climate science is doubted and denied. Economics - a far softer science with a (so far) vastly greater impact on human society - gets a staggering amount of leeway by comparison. And when it's practitioners (who outnumber climate scientists 100 to 1) get things catastrophically wrong, as in the case of the recent Wall Street collapse, there is surprisingly little criticism of the theoretical underpinnings, nevermind little details like bad data.
The science of climate change, by contrast, is on very solid theoretical footing; but sometimes every science has to deal with bad data, as in this case. The notion that this somehow discredits the theoretical basis of radiative forcing and the greenhouse effect is sheer lunacy. Simple stock-and-flow box models are enough to understand that anthropogenic climate change is inevitable. If you can understand how a bathtub overfills when you leave the faucet running, you should be able to understand that climate change is real and unavoidable.
The reactions of laymen and the ignorant masses who follow Limbaugh et al can only be explained as propaganda-induced hysteria, to which only the profoundly ignorant and/or fearful are vulnerable. The reactions are similar to those of the North Atlantic fishermen who vehemently asserted that since they'd been fishing the Georges Bank for 250 years it was 'obvious' it could never be depleted. Changes in fish populations, if there were any, were 'natural'. They ignored scientists and continued to produce record catches - right up until the entire fishery collapsed a few years ago.
Any one who is genuinely interested in learning about how and why complex systems change catastrophically should read "Limits to Growth" - the classic by the MIT team headed by Donella Meadows.
Parent
Re:Not consistent? (Score:5, Insightful)
"they just don't believe the hysterical rantings of scientists who can't predict the weather or a hurricane season with any certainty"
It's far easier to predict average temperatures will decline through summer, fall, and winter (in the northern hemisphere) than it would be to predict whether the temperature will go up or down on a given day or week.
In other words, not being able to reliably predict the daily weather or hurricane season with great precision says nothing about the reliability of predicting longer-term climate change.
Yes, there's nothing wrong with skepticism, but at least try to compare equivalent problems.
There is no ambiguity about which way arctic sea ice has been headed for the last few decades, and little difficulty predicting the continuation of that trend unless something truly fundamental changes about the climate system. Ask anyone who lives in the arctic or who travels there by sea -- it has changed a lot. Icebreakers in the region have a very different experience now versus 20 years ago. Changes in instrument interpretation/calibration doesn't change the existence of that trend. It's freaking obvious when you do some "ground truthing". Predicting what sea ice will do in a given year or exactly when the trend might reach its ultimate end (total melting) is much harder to predict, but the trend is obvious and the implications of its ultimate conclusion are pretty reliable. Could the multi-decadal trend reverse and the sea ice grow more extensive year after year to attain its past summer extent again? I doubt it. It's going to be ice-free in summer eventually. The debate is about when it will finally happen first, but there's no point in getting hung up on whether it's going to be 2008, 2009, or 2012, because the exact date doesn't matter.
End of the world? No. But it's going to be a pretty radical change nonetheless. Yes, we know that climate change happens (it has changed throughout Earth history). The question is whether this kind of change and this rapid is a good thing for human interests. It doesn't help the debate when climate change deniers make a big deal about an instrument calibration change and start implying we know nothing or that the trend isn't real. There's problems on both sides of this debate, but it is normal to have problems with scientific data. They are worth finding and fixing. But the basic interpretation hasn't been compromised because it affects only a small part of the data.
Parent
Re:Not consistent? (Score:5, Insightful)
When are the smart-arses who reckon they've spotted some humoungous flaw in the actual science going to actually publish? Oh, right, it's all a crock of shite by delusional Daily Mail readers, and journals of record don't print papers that arrive for review written in green crayon.
If you're so smart and the world's climatologists are so dumb, for the love of god stop yammering about it on Slashdot, publish, and collect your Nobel Prize.
Parent
Multiple Data Sources (Score:5, Insightful)
OK, I can see their point, but using the EOS sensor may have given pause to researchers doing a comparison to current conditions using the traditional sensor, i.e. cross-reference current conditions to be more confident that your data is correct. Nothing like screaming "the sky is falling" due to bad data. Any science experiment, especially one that can produce sensationalist news, should not just rely on one piece of data.
Re:Multiple Data Sources (Score:5, Insightful)
<i>"Sensor drift is a perfect but unfortunate example of the problems encountered in near-real-time analysis. We stress, however, that this error in no way changes the scientific conclusions about the long-term decline of Arctic sea ice, which is based on the the consistent, quality-controlled data archive discussed above."</i>
If the media outlets and attention seekers sensationalise the real-time output, then unreasonable conclusions might well be drawn. What's the alternative though? To not make this real-time data available? Scientific hypotheses will be tested against the corrected data, so this sensor drift doesn't affect them. These are preliminary measurements only, not full-blown experiments with scientific conclusions. The polar bears are still going to have to become better swimmers.
It's a little ironic that this data will have been used by one group with an agenda to sensationalise climate change, and now will be used by the (perhaps overly) sceptical amongst us to poo poo it. Some have an agenda to sensationalise both. All serve to cloud the real message.
Parent
Tosh. (Score:5, Insightful)
"n May, 2008 they went so far as to predict that the North Pole would be ice-free during the 2008 'melt season,' "
No they didn't.
They said it was a remote possibility.
This was taken up by the anti global climate change, altered, and then used to "prove" that global warming wasn't happening when it didn't happen.
The fact it is presented that way by the story submitter shows which way they think, and thus how reliable the overall story is.
Typical spin job (Score:5, Informative)
In May, 2008 they went so far as to predict that the North Pole would be ice-free during the 2008 'melt season,'
Er, no, they said it was possible and later quote "a 59% chance of a new record minimum this year". How the media chose to report this is another matter... Oh yes, note the date: May 2008.
Today, however, they say that they have been the victims of 'sensor drift' that led to an underestimation of Arctic ice extent by as much as 500,000 square kilometers.
And if you read TFA, the sensor drift started in January 2009, was spotted within a few weeks and only affected their daily images which are effectively "live" and hence haven't gone through QA.
So how exactly does an error which occurred in Jan/Feb 09, was almost immediately spotted and declared affect a (misreported) prediction made last May?
<irony>Meanwhile, I'm sure the little fairies are hard at work ensuring that the geological era's worth of sequestered CO2 we're in the process of releasing back into the atmosphere magically changes its physical properties. After all, it is made from special carbon that God put there in 4004BC for us to burn, unlike that nasty communist CO2 that exhibits the greenhouse effect in godless laboratories.</irony>
Re:alternative solutions (Score:5, Informative)
Both the AMSR_E and SSM/I data are satellite derived products.
http://www.ncdc.noaa.gov/oa/satellite/ssmi/ssmiproducts.html
http://www.aqua.nasa.gov/about/instrument_amsr.php
Parent
Re:Rocket science? (Score:5, Insightful)
Science isn't an exact science, people are involved and people make mistakes. Scientist need to remember that they are human too and they will make mistakes. Data can be off or altered, I remember a local weather channel use to use a point for the local temperature until they built a Dunin Donuts next to it, and heat escaping the building or cars or something (it was a long time ago) raised the temperature 5 or 6 degrees warmer then the actual weather.
Climate Scientists are trying to make very accurate predictions where they don't have the data to do so. We can probably say global warming is real, however that is a very broad statement. But to say The Polar caps will be gone in 2008 or by 2012 NYC will be flooded, is grossly misunderstanding the complexity of the earths environment.
Granted that most of the fear mongering worst case scenarios stuff isn't from the people doing the real science but from activists groups who pick and choose data to make people afraid so they do what they want. But still the Scientists don't like saying to people Hey I could be wrong, but thats OK because with the scientific process being wrong takes us the next step closer to the real answer. They want to in general portrait themselves like the Sci-Fi scientist who know what is going on and is always right.
Parent
Re:Rocket science? (Score:5, Insightful)
There is no question that data analysis and prediction is subject to errors, sometimes quite large ones. The real question is whether these errors are due to researcher bias. Unfortunately climate data and predictions are apparently more motivated by political beliefs and biases than hard facts.
Many people have strong feelings that disaster is about to occur. Perhaps this comes from childhood recollections of maternal warnings about running with scissors or touching hot stoves. These fears can be reinforced by religious beliefs that portend the end of the world. Because this psychological factor is so prevalent we need to be especially skeptical of predictions of future disasters.
Parent
Re:Rocket science? (Score:5, Insightful)
I think it goes back to the showing the DDT is a harmful chemical (other then the old commercial "DDT its good for you and good for me!") it was widely used and was considered safe and effective a marvel in scientific advancement. Then it was shown to have effect on the eagle population and perhaps humans as well. So we had to stop using the chemical, but that opened peoples minds to think other things will have such problems, and to question every advancement and study the effects of it. In general it is a positive thing but it has lead to fear mongering and a belief that we should stop advancements as every advancement has some cost to it (however in my opinion they usually forget to factor in the benefits). Such as the immunization and the possible link to autism, lets say it creates a 1% increase in autism how ever it saves 25% from death, the benefits out weigh the risks and the parents who avoid this are poor judges on risk assessment.
Parent
Re:Rocket science? (Score:5, Informative)
So, to be clear, this issue has arisen over the last 4-6 weeks. The records for the last decade, clearly showing a significant trend towards less sea-ice, are unaffected.
Parent
Re:Rocket science? (Score:5, Insightful)
But to say The Polar caps will be gone in 2008 or by 2012 NYC will be flooded, is grossly misunderstanding the complexity of the earths environment.
Yeah, but climatologists don't suffer from that misunderstanding. They're the ones who have to actually slog through that complexity.
Granted that most of the fear mongering worst case scenarios stuff isn't from the people doing the real science but from activists groups who pick and choose data to make people afraid so they do what they want.
An astute point, and really it was both the activists and the -- well, whatever the opposite of an activist is, someone who wants to do nothing is called -- who with the help of the media took the scientists claims of a chance for a minimum and an ice free north pole (just the pole, btw, not all arctic ice gone entirely) and run with it as if the scientists are saying it's a sure thing. The activists to cry doom and gloom, and the inactivists to pounce on as proof the whole thing is bunk if the results come out on the "wrong" side of the scientist's claimed 60:40 odds.
But still the Scientists don't like saying to people Hey I could be wrong, but thats OK because with the scientific process being wrong takes us the next step closer to the real answer. They want to in general portrait themselves like the Sci-Fi scientist who know what is going on and is always right.
They didn't just say "Hey I could be wrong", they attempted to quantify the chances of them being wrong. Then they voluntarily report on censor errors that briefly screwed up their data, while I'm sure knowing that this would be blown way out of proportion and used to "prove" that they never know what they're talking about at all. In other words, the opposite of trying to appear like they can't be wrong. So I'm not buying it at all.
Just because news headlines omit the qualifiers does not mean they do not exist.
Parent
Re:Rocket science? (Score:5, Insightful)
[citation needed]. Oh, wait, you just invented those doom mongers yourself, bravely defeating a horrible strawman.
Parent
Re:How can people expect... (Score:5, Informative)
Parent
Re:How can people expect... (Score:5, Insightful)
'Climate Change' is a politically hot topic, and plenty of governments give grants into this kind of research.
"The overwhelming majority* of the world's climate scientists" know what side their bread is buttered on. It's on the side of giving governments more excuses to tax and define their citizens activities ever more closely.
Follow the money. Isn't that what you'd say about the report of any climate scientist who is a global warming 'denier'?
(*I don't know how many climate scientists there are, but hundreds have signed on to papers with adverse positions.)
Parent
Re:How can people expect... (Score:5, Insightful)
While I respect your skepticism of the scientific press, I think this reasoning suffers from two flaws. First
All I'm saying is that most of the 'studies' I've seen floating around the press smell fishy to me.
Relying on the press to get your scientific information is going to be incomplete. The press reports particularly sensationalistic doom & gloom stuff, whereas most science goes out of its way to take a neutral tone. It is too much to ask a non-scientist to pay attention to the leading journals (I'm thinking of Science and Nature here), but we are also at a point in our history where science needs to inform our politics. This is obviously troublesome for democracy, and why I sympathize with your skepticism of science.
Second,
I'd rather we follow simple common sense and watch out for our planet because it's the frickin' right thing to do instead of running around like headless chicken being afraid of our children being cooked alive by the sun
This reasoning is suspect because, aside from global warming effects, green house gas emissions aren't very harmful. It is relatively easy to see the pollutant effects of particulate emissions: they make things dirty and also hard to breathe. GHG emissions, on the other hand, are fairly clean and only have a mediated effect on human health (through climate change). Your strategy would have us fix only the easy to see problems even if there are more important environmental matters that require advanced scientific techniques to understand.
Parent
Re:We only use data that support our hypothesis (Score:5, Insightful)
"We do not use AMSR-E data in our analysis because it is not consistent with our historical data."
And our historical data shows terrible calamity awaiting us at every turn, and even if reality doesn't bear this out, it makes sense that we should continue to sound the alarm because if we do decide to face reality people may not take us and our hysterical blatherings seriously.
We'd rather just keep on using outdated modes of measurement and forecasting that give incorrect results every year because the results fit our hypothesis better. And what better to support a hypothesis than data that will back it up?
Right! If the data doesn't back our conclusions, use different data!
Parent
Re:We only use data that support our hypothesis (Score:5, Insightful)
It's always a very good idea to take single quotes from a summary out of context, and make sweeping statements about that. This is especially true for science. Science really isn't one of those topics that require some in-depth knowledge to understand what's going on.
Taking off the sarcasm tag for a moment, this is one of the worst "science" pieces I've seen on slashdot in perhaps the last year. Cobbling together some serious accusations of scientific incompetence from a series of links doesn't really show anything. How the hell do I know how to interpret these statements in context? The links are all taken OUT of context and put into an entirely new argument without any further analysis or explanation. I'm left with what amounts to some hand waving and ranting about "scientific bias". Without a real analysis by someone qualified to make it this "story" is best left ignored.
Parent
There are ways... (Score:5, Insightful)
When a new sensor is thought to have greater accuracy or reliabilty than an old one, but produces data which are not entirely consistent with the older one, it does NOT prevent use of the new sensor or meaningful use of data from both sensors. One standard technique is to employ both sensors simultaneously for some time - in other words, the two data series would overlap for that time. If both series show a downward trend in ice cover, then the trend probably real, even if they always disagree about the level of ice cover or the rate of decline. Over a sufficiently long time, it should be possible to build a model to quantitatively explain the difference in readings.
Come on, guys. There must be a few PhD theses waiting to be written on how to reconcile these instruments...
Parent
Re:Oh gosh. (Score:5, Insightful)
However you feel about this issue, I think it is a bit weak to try and claim a change over 4 years constitutes a 'trend' when it comes to global climate data.
Parent
Re:Oh gosh. (Score:5, Insightful)
Looking at the new graph it's still pretty obvious that the trend is "downwards",
Err... no. What I see looking at the data is two very low years: 2007 and to a lesser extent 2008. Calling that a "pretty obvious trend" nicely reveals your bias, but not much else. I could as easily say it is a "pretty obvious oscillation", as 2008 is "clearly" recovering from the 2007 perturbation.
I can see why the guys doing this aren't using the new data, as there is no way that there is adequate statistical power here to make a judgement about trends. Unfortunately, now that the old data have been shown to be badly flawed, the dire predictions of an imminently ice-free Arctic no longer have any very robust empirical support.
THIS is the way science works: you look at the evidence, squeeze it hard and see if it breaks. There is no doubt that the evidence for a soon-to-be-ice-free Arctic is broken. Ergo, the plausibility of dramatic climate change effects in our near future has gone down, no matter what anyone's politics drives them to prefer.
The only robust signal for global climate change I'm aware of is global ocean heat content, which seems to be increasing. However, given the number of reversals of supposedly robust results in the field of climate science I want to take a much closer look at those data before being convinced by them.
I used to be very concerned about global climate change, and in open-minded arguing with "deniers" I took a hard, critical look at the data and the models, because I wanted to find a compelling, unproblematic argument to convince my opponents, whom I credit with being able to change their minds when faced with the evidence. What I found was that neither the data nor especially the models stood up to professional scrutiny. There is good science being done, but it is not the kind of stuff you'd want to base public policy on.
There are good arguments for environmental policy that do not depend on the risk of global climate change, and the environmental movement is doing itself no good by linking policy and science together they way they have, so that people think "if there is no risk of global climate change then driving my SUV must be ok."
Parent
Re:Oh gosh. (Score:5, Informative)
Parent
Re:Historical error (Score:5, Insightful)
Compare it to a game benchmark or whatever. You keep a standard, even if it's flawed, because that's the only way you get comparable results. You can't take a raw number for MIPS or millions of polygons/second or transactions per second or whatever the metric of choice is in your field (here, one is obviously ice sheath coverage) and use numbers from wildly different methods to even try to devise a historical trend. The value observed might not correlate exactly (or even very well) with old ones, but unless the flaws in the method cause great variability within that framework, the historical trends will still be accurate, or at least more accurate compared to what would happen if you changed your methodology each year and still tried to extract longer trends.
It might not be a good choice, and suggestions to run double series over the (short) timespan where overlapping data is indeed available would of course be better, but you can't just switch to the latest and greatest if you want some kind of consistency in your data.
Parent