This is not bad science. An instrument failed in a subtle way, the error was caught and corrected a month or two after it happened.
Data doesn't arrive at the lab engraved on a golden tablet. Continual error correction is a big part of science, and the only way to get data we can count on is to do the sort of double-checking and correction that's the subject of this article.
We'd rather just keep on using outdated modes of measurement and forecasting that give incorrect results every year because the results fit our hypothesis better.
It's got nothing to do with political bias. Inaccurate data taken for a long time is often better than accurate data taken for a short time.
Example: Suppose I ask you to measure the temperature of a cup of coffee, and tell me if the coffee is getting colder over time. I give you a thermometer which is inaccurate, it always reads about 5 degrees too low.
You take your data for an hour or so, and make a nice graph. Then I give you a shiny new thermometer, which is much more accurate: it always reads about 1 degree too high.
You start using the new thermometer because it's better. Tell me, did the temperature of the coffee suddenly increase by 6 degrees?
This is a simple example: combining data from different sources without introducing errors is a tricky business. To do it right, you need to compare the data sets over a time period where they overlap (in the coffee experiment, you use *both* thermometers for a while), but in the case of AMSR-E we haven't got enough data to do that reliably yet.
One standard technique is to employ both sensors simultaneously for some time - in other words, the two data series would overlap for that time.
The guys responsible for this dataset know that, and they do just that sort of thing for their archival datasets. But that kind of analysis takes time, effort, and human double-checking, so the "real-time" data is released without some of these checks, causing it to be much less reliable but more immmediately available.
This is more complicated than we are led to believe...in the popular press. and current reports oversimplify the problem....in the popular press.
We *do* have a lot of the proper data (but not all), and you *can* gain a holistic understanding of the problem so you can understand how to address it... but you won't get there by reading Slashdot and watching CNN. I suggest realclimate.org for a start.
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.
We climate scientists try to give a sense of the certainty of our predictions, but our discussions of certainty are almost always deleted when the news is presented to the public. The two predictions you mention are vastly different -- we're pretty sure about ice, but flooding New York has a gigantic pile of "ifs" in front of it.
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.
Real scientists *do* say "I could be wrong" all the time, and try to estimate the odds of being wrong. We're happy with greater certainty, but we know that 100% guarantees don't exist.
The confident, absolutely certain authoritarian scientists you see on TV are just talking heads: they may have PhDs, but they're not acting as true scientists.
"This is complicated. I don't understand it. And we shouldn't implicitly trust anyone who claims to know. So how can we ever know?"
You've got two options. Either trust the guys who say they know, or learn it yourself. Rejecting an idea because you can't take the time to learn about it is pretty ridiculous.
I'd consider myself "close enough" to a sea ice expert.
What we end up with this fall depends on both how much ice there is now, and how warm it is over the next seven months.
The more ice there is now, the *less likely* it is to all disappear in one year, but I'm not prepared to say it's impossible, if summer 2009 is a record-breaking scorcher.
"a big step toward ending the chicken-and-egg problem of a dearth of good games that run natively under Linux."
So was the Quake port in the late '90s. So was Loki Software around 2000. So was Uplink in the mid-2000s. So was EVE a couple years ago.
People have been predicting the imminent end of crappy Linux gaming for ten years now; every new game is heralded as the savior of Linux gaming, and a year later we're pretty much back where we started.
Me neither. But before I saw Rain Man I had a kid in my elementary school classes who was never really "with us" emotionally, doing terribly in classes, etc... and another couple kids who just seemed a little distant and odd. And then one day I learned what the word "autism" meant, and said "oooh, yeah, I've seen that."
Multiply that experience by 300 million Americans, and you've got yourself an "autism epidemic".
One man's chaff is another man's wheat. One of the most entertaining ways to waste a couple hours on the Internet is to search Slashdot for "whatcouldpossiblygowrong"?
You can't hide the U.S. President in his daily business. It's just impossible. Motorcades, helicopters, public speeches, platoons of serious men in dark suits... the dude wins the prize for Most Obvious Man in America. forget about it.
Now, if the President *wants* to hide, he goes to a secure bunker somewhere where the radio waves don't shine, somewhere that even the sneakiest guy with an antenna can't get within ten miles of.
Arrow's Impossibility Theorem is a great example of a very good mathematical proof that is pretty much irrelevant to the real world.
In particular, its criteria for a "good" voting system are too strong. Even its scariest-sounding criteria, like "non-dictatorship", are often violated by real voting systems in very very benign ways.
In short, a voting system which is mathematically imperfect may still be a huge real-world improvement. So let's not let the perfect be the enemy of the good.
The American people used voting to remove him from office.
Just to be picky, the people used voting to *keep* him in office (2004). They used term limits (which was never publicly voted on) to remove him from office in 2008.
having a bare screen on an expensive device is kinda like...jock strap while playing football.
If Apple made an iQuarterback, it would have stainless steel testicles. In my experience as a serious electronics abuser, nothing short of a belt sander will damage an iPhone screen.
Slashdot Mirror
When I was in elementary school ... Clearly somebody was (or is) lying.
Most of what you learned in elementary school is not so much a lie, but definitely oversimplified. This is a case in point.
Learn more now, you're old enough to handle the complicated truth.
Your second article is 7 years old. 2002 and 2003 were unusually low years compared to the years that came *before* them.
But the first link shows that, compared to the years *after* them they're average or rather high.
All of this is consistent with a gradual but irregular loss of ice over time.
This is not bad science. An instrument failed in a subtle way, the error was caught and corrected a month or two after it happened.
Data doesn't arrive at the lab engraved on a golden tablet. Continual error correction is a big part of science, and the only way to get data we can count on is to do the sort of double-checking and correction that's the subject of this article.
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.
It's got nothing to do with political bias. Inaccurate data taken for a long time is often better than accurate data taken for a short time.
Example: Suppose I ask you to measure the temperature of a cup of coffee, and tell me if the coffee is getting colder over time. I give you a thermometer which is inaccurate, it always reads about 5 degrees too low.
You take your data for an hour or so, and make a nice graph. Then I give you a shiny new thermometer, which is much more accurate: it always reads about 1 degree too high.
You start using the new thermometer because it's better. Tell me, did the temperature of the coffee suddenly increase by 6 degrees?
This is a simple example: combining data from different sources without introducing errors is a tricky business. To do it right, you need to compare the data sets over a time period where they overlap (in the coffee experiment, you use *both* thermometers for a while), but in the case of AMSR-E we haven't got enough data to do that reliably yet.
One standard technique is to employ both sensors simultaneously for some time - in other words, the two data series would overlap for that time.
The guys responsible for this dataset know that, and they do just that sort of thing for their archival datasets. But that kind of analysis takes time, effort, and human double-checking, so the "real-time" data is released without some of these checks, causing it to be much less reliable but more immmediately available.
There are good scientists and poor scientists, but they are still people practicing science.
Scientists often do things that aren't science. Occasionally they will go hiking, play World of Warcraft, post on Slashdot, or be interviewed on CNN.
Science is not "anything done by a scientist", and the guys on TV talking about melting ice caps are *teaching*, not doing science.
This is more complicated than we are led to believe ...in the popular press. ...in the popular press.
and current reports oversimplify the problem.
We *do* have a lot of the proper data (but not all), and you *can* gain a holistic understanding of the problem so you can understand how to address it... but you won't get there by reading Slashdot and watching CNN. I suggest realclimate.org for a start.
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.
We climate scientists try to give a sense of the certainty of our predictions, but our discussions of certainty are almost always deleted when the news is presented to the public. The two predictions you mention are vastly different -- we're pretty sure about ice, but flooding New York has a gigantic pile of "ifs" in front of it.
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.
Real scientists *do* say "I could be wrong" all the time, and try to estimate the odds of being wrong. We're happy with greater certainty, but we know that 100% guarantees don't exist.
The confident, absolutely certain authoritarian scientists you see on TV are just talking heads: they may have PhDs, but they're not acting as true scientists.
So if I can summarize your argument:
"This is complicated. I don't understand it. And we shouldn't implicitly trust anyone who claims to know. So how can we ever know?"
You've got two options. Either trust the guys who say they know, or learn it yourself. Rejecting an idea because you can't take the time to learn about it is pretty ridiculous.
Any sea ice experts on slashdot?
I'd consider myself "close enough" to a sea ice expert.
What we end up with this fall depends on both how much ice there is now, and how warm it is over the next seven months.
The more ice there is now, the *less likely* it is to all disappear in one year, but I'm not prepared to say it's impossible, if summer 2009 is a record-breaking scorcher.
This is a *perfect* excuse to build a Tesla coil.
"a big step toward ending the chicken-and-egg problem of a dearth of good games that run natively under Linux."
So was the Quake port in the late '90s. So was Loki Software around 2000. So was Uplink in the mid-2000s. So was EVE a couple years ago.
People have been predicting the imminent end of crappy Linux gaming for ten years now; every new game is heralded as the savior of Linux gaming, and a year later we're pretty much back where we started.
This changes nothing.
I have personally seen someone pull a mouse off a glue trap and leave the feet behind.
Mice are pretty flimsy, when it comes down to it. Rats less so.
As you might guess, I am not a programmer. Anyone want to rise to the challenge?
Okay you inspired me. I just spent an hour writing one, and a couple of hours watching it. It's unbelievably cool.
Unfortunately I wrote it in a very nonportable language (MATLAB), so you probably can't use it.
Maybe I'll write it in Java someday when I have time.
Before I saw Rain Man I had never heard of Autism
Me neither. But before I saw Rain Man I had a kid in my elementary school classes who was never really "with us" emotionally, doing terribly in classes, etc... and another couple kids who just seemed a little distant and odd. And then one day I learned what the word "autism" meant, and said "oooh, yeah, I've seen that."
Multiply that experience by 300 million Americans, and you've got yourself an "autism epidemic".
Getting rid of the chaff
One man's chaff is another man's wheat. One of the most entertaining ways to waste a couple hours on the Internet is to search Slashdot for "whatcouldpossiblygowrong"?
This is a silly discussion.
You can't hide the U.S. President in his daily business. It's just impossible. Motorcades, helicopters, public speeches, platoons of serious men in dark suits... the dude wins the prize for Most Obvious Man in America. forget about it.
Now, if the President *wants* to hide, he goes to a secure bunker somewhere where the radio waves don't shine, somewhere that even the sneakiest guy with an antenna can't get within ten miles of.
Arrow's Impossibility Theorem is a great example of a very good mathematical proof that is pretty much irrelevant to the real world.
In particular, its criteria for a "good" voting system are too strong. Even its scariest-sounding criteria, like "non-dictatorship", are often violated by real voting systems in very very benign ways.
In short, a voting system which is mathematically imperfect may still be a huge real-world improvement. So let's not let the perfect be the enemy of the good.
Arrgh!
The American people used voting to remove him from office.
Just to be picky, the people used voting to *keep* him in office (2004). They used term limits (which was never publicly voted on) to remove him from office in 2008.
Ladies and gentlemen! A full-contact legal battle for the ages!
In this corner, we have the Author's Guild, with the full weight of American copyright law behind them.
And in this corner, we've got the National Federation for the Blind, swinging a big stick: the Americans with Disabilities Act!
Gentlemen ... FIGHT!
Apple's core consumers (no pun intended... also, a half decent name for a band) ARE fanbois.
Not any more. These days, Apple's core customer is a 14-year-old girl, begging her parents for an iPod Nano.
having a bare screen on an expensive device is kinda like...jock strap while playing football.
If Apple made an iQuarterback, it would have stainless steel testicles. In my experience as a serious electronics abuser, nothing short of a belt sander will damage an iPhone screen.
Sorry, couldn't resist. ;-)
I believe you have just committed a terrorist act. Deliberate forum sabotage is a dangerous game, my friend.
If things like this weren't possible, large old alliances like BoB would be fixtures in the game, immovable and unconquerable.
Of course empires should be conquerable. But not like this.
The fact that BoB was nearly invincible reeks of bad game design. The fact that BoB was done in by a single mouseclick is also bad game design.