|
Post by zer0th on Jun 13, 2009 14:06:15 GMT
|
|
|
Post by glc on Jun 13, 2009 17:13:30 GMT
glc writes "Yes but this could be related to the GISS extrapolation of the arctic."
I wish people would read what others write. That is EXACTLY what I said. "e.g. extrapolating to the poles ".
Sorry.
|
|
|
Post by glc on Jun 13, 2009 17:23:36 GMT
So, yes, it does bother me that gistemp is consistently just enough of an outlier to generate global headlines, that its data quality is demonstrably poor, and that its leadership regards this as a non-problem (Schmidt calls the idea that GISS staff would be personally responsible for data quality "laughable").
For the 48 states the trends over the last 20 years are as follows:
UAH +0.26 deg per decade GISS +0.25 deg per decade
In what way is GISS an "outlier"? And what is "demonstrably poor" about it's data?
If there is an outlier at all, it's UAH. Between 1979 and 1992 UAH were recording warmer temperatures than RSS (information provided by Roy Spencer) and it is this that has meant the UAH trend is less than Hadley, GISS and RSS.
Since 1992, there is barely a few hundredths of a degree between the trends. Over the past 20 years the trends have been remarkably consistent.
|
|
|
Post by woodstove on Jun 13, 2009 19:42:42 GMT
|
|
|
Post by tacoman25 on Jun 13, 2009 20:25:01 GMT
"We are talking" ?
Ahhh, you are talking....
GISS is like a watershed for highly polluted data. So whenever GISS takes a flight from reality I find it interesting and of note. I predicted we'd get this divergence nonsense a month or so back. It was only a few months ago that GISS was falling and UAH rising. In fact when we compared the datasets over the same period (1979-1998), GISS was about 0.2 deg below UAH. Now UAH is falling relative to GISS. There is actually a good reason for this. The surface records appear to respond much quicker to ENSO fluctuations probably because they take a current 'snapshot' of SST, while the satellites are measuring the atmosphere which takes a longer time to adjust to SST changes. Trends since 1992 are virtually identical for all 4 main data sets. This would all be fine and dandy except that GISS consistently ends warmest overall, as of late. They were far warmer than the other sources in 2007, warmest in 2008, and will likely end up warmest again in 2009. And of course, this is all after adjustments for their different baseline.
|
|
|
Post by tacoman25 on Jun 13, 2009 20:29:07 GMT
JimC Somewhere around the start of the 21st century, GISS values started to tend significantly higher that the other data sets. Yes but this could be related to the GISS extrapolation of the arctic. You can argue about whether the extrapolation is a good enough approximation or not, but they do appear, to have been consistent, at least. In the early months of this year, arctic temperatures dropped from their recent highs and this was reflected in the GISS anomalies. It's worth noting that summer arctic ice extent since ~2002 provides strong support for the GISS version of events. I disagree. There is very weak correlation the last decade between global temperatures and Arctic summer ice extent.
|
|
|
Post by tacoman25 on Jun 13, 2009 20:36:23 GMT
Right I guess you are suggesting that in the past month millions of AC units got installed next to temperature sensors, or something equally ridiculous. At least think about your arguments before you make them. That way you won't make silly arguments. We are talking about a divergance in one single month between GISTEMP and UAH. The obvious answer for this is that SST have increased significantly in the past 3 months and this hasn't yet affected UAH due to lag time, but immediately affects GISTEMP. Yes we can expect UAH and RSS to jump up significantly in the next few months. However, in 2008 SST warmed significantly from March to June, and yet GISS cooled significantly over that time-frame. So it's not always as simple as you make it out to be.
|
|
|
Post by glc on Jun 14, 2009 10:48:28 GMT
I disagree. There is very weak correlation the last decade between global temperatures and Arctic summer ice extent.
You need to explain that. We have posters complaining that GISS has diverged from other datasets since 2003. We have low arctic ice extent since ~2002. There doesn't need to be a correlation between ice extent and global temperatures - just a correlation between the ice extent and the GISS divergence. Actually we don't really need that. We need a correlation between arctic temperatures and the GISS divergence, but as arctic temperatures are difficult to come by, ice extent could act as a proxy.
|
|
|
Post by socold on Jun 14, 2009 11:07:08 GMT
|
|
|
Post by socold on Jun 14, 2009 11:20:24 GMT
Right I guess you are suggesting that in the past month millions of AC units got installed next to temperature sensors, or something equally ridiculous. At least think about your arguments before you make them. That way you won't make silly arguments. We are talking about a divergance in one single month between GISTEMP and UAH. The obvious answer for this is that SST have increased significantly in the past 3 months and this hasn't yet affected UAH due to lag time, but immediately affects GISTEMP. Yes we can expect UAH and RSS to jump up significantly in the next few months. However, in 2008 SST warmed significantly from March to June, and yet GISS cooled significantly over that time-frame. So it's not always as simple as you make it out to be. We'll see
|
|
|
Post by glc on Jun 14, 2009 17:38:27 GMT
There's a graph on the GISTEMP site here which shows how much the warming trend is reduced if you remove higher latitudes: data.giss.nasa.gov/gistemp/graphs/ArcticEffect.pdfActually this shows the arctic influence on GISS perfectly. Take out the arctic and you have what looks to be a flat/falling trend similar to the other datasets.
|
|
|
Post by icefisher on Jun 14, 2009 17:51:42 GMT
We need a correlation between arctic temperatures and the GISS divergence, but as arctic temperatures are difficult to come by, ice extent could act as a proxy. Probably true if your definition of arctic temperatures includes water temperatures.
|
|
|
Post by tacoman25 on Jun 15, 2009 4:31:11 GMT
However, in 2008 SST warmed significantly from March to June, and yet GISS cooled significantly over that time-frame. So it's not always as simple as you make it out to be. We'll see There's nothing to "see"...it's a fact. Sometimes GISS/HadCRU seem to closely follow SST changes, but sometimes they don't.
|
|
|
Post by tallbloke on Jun 16, 2009 9:21:04 GMT
If you're going to tell a lie, best tell a big one.
|
|
|
Post by glc on Jun 16, 2009 11:31:43 GMT
GISS is in at +0.55.
If you're going to tell a lie, best tell a big one.
Why do you believe this is a lie? and what makes you think it is a big one?
The GISS anomaly is +0.34 above the 1979-1998 average for May. In February, the UAH anomaly was +0.35 above the 1979-1998 average for February, while GISS were less than 0.2 deg above. Were UAH lying?
Over the next few months, satellite anomalies are likely to be relatively cooler than the surface anomalies. This has nothing to do with fraud by Hadley and GISS, but is more likely due to the fact that there appears to be a longer lag time in response to fluctuations in ocean temperatures, i.e. the troposphere is still 'seeing' the colder SST from earlier months.
Since ~1990 there is only a few hundredths of a degree between the trends of the 4 main datasets.
|
|