Climate Audit

Rick Ballard

It’s just a flesh wound.

Real Richard Sharpe

Do the SH and NH include all of each hemisphere, or only the extra-tropical areas?

John S.

Has anyone noticed the drastic reductions in the number of stations employed throughout 2007 in GISSTEMP analysis on the way to their conclusion that 2007 was the second-warmest year on record? And now the number of stations is no longer displayed.

Steve: Station numbers have been discussed here on many occasions.

steven mosher

anyone trust hansen to adjust june 2008?

Basil

It is not as apparent in Steve’s graphs, though it is there if you know what to look for, but there’s a divergence between the tropics and NH even in the satellite temps. It is, however, more pronounced in the land-sea series. Here’s a chart I did a few days ago in which I overlaid smoothed versions of UAH tropics and NH over the HadCRUT data.UAH NH begins to diverge during the 1990’s at about the same time as HadCRUT’s Northern Extra Tropics. The latter is probably accentuated by the station number issues, UHI effects, and station quality issues. But these wouldn’t affect the UAH’s NH trend. So what’s besides all the issues that probably bias the land-sea NH record, what’s causing the NH warming, relative to the tropics, seen even in the satellite record?

lucia

Did you get Roy Spencer to email you the data? I cleared cache and I don’t see June.

Steve: No.

(Just for fun, I’m answering your question in litigation examination for discovery mode.)

steven mosher

re 12. thats, funny. was it predict may 2008 when june 2008 came out?
or predict may 1934

SOM

Are the red dots cherries?

Basil

#13

Steve: No.

(Just for fun, I’m answering your question in litigation examination for discovery mode.)

That’s funny. I’ve been there, done that. Not just in discovery, but at trial, I make the cross-examining attorney earn their money by answering “yes” or “no” and nothing else, to the extent possible, forcing them to follow up. It really flusters them if you answer “no” to a question they were sure they were going to get a “yes” from. But often there is no way you can answer “yes” without qualifying the answer, and they will try to prevent the qualified responses (“Just answer ‘yes’ or ‘no,’ dangit!”) This then forces them to follow up, giving you the chance to qualify your answer. I once had a cross-examining attorney “object” (and that’s rare, the objection coming from the attorney asking the questions) that my answers were not “forthcoming.” The judge, who was sympathetic to the other side, agreed. That may be the only time in history an expert witness was chastised by a judge for not talking enough, as opposed to talking too much.

James Erlandson

You agitatin’ my dots?

Larry T

re 21. I like his error estimates of .1 degree from data that was basically measured to the nearest degree even leaving out any biases. We are starting with a minimum of .5 degree errors + bias + uncertainty. Bias I estimate from the Station Survey to be close to a degree and i will assume the uncertainty to be what he showed for a .95 error estimate of 1.6 degrees.

–

Larry T

oops re 20

Anthony Watts

Steve, did you get it from RSS data, which IS out here

…and then calculate the offset to arrive at UAH? They both use the same source data, but different methods of processing it.

Steve: Nope.

You beat me on RSS !

Steve McIntyre

#12. GISS May temperatures are out now. When did April’s come out?

Anthony Watts

RE28 Steve, Last month RSS came out on the 6th so yes, 4 days earlier this month

Anthony Watts

OK it must be (or had been and taken down) on his new “coming soon” website

http://gavinschmidt.com/

I guess he just could not get enough talk time on RC.

Luis Dias

Steve, there’s a kind of a graph that you or anyone else could make that would present the changes in a dramatically better visual manner. I’ve seen it in the blog of Mr. William Briggs, the statistician, where he plotted the CO2 increase in such a fashion, here

The trick is to plot the points of monthly avg temperatures and only connect the same months in a sequence of years. There will be 12 lines, each one for a single month and it will run for the decades of data available. January 1999 connects with jan 2000 and jan 2001, etc, while feb 1999 connects with feb 2000, etc. You get the idea.

In this graph, seasonal changes are diminished greatly. One can be left with a much better grasp of the changes whe are going through the last years and specially this year.

Anyway, it seems a refreshing kind of graph and I wonder if any mathmatician in here is interested.

kim

Wisconsin?
=======

John Goetz

#29 Steve

I think Lucia is referring to the new May value when June comes out. The modifications to history begin immediately. Remember the ruckus over the high GISS March anomaly? March has come down every month since.

lucia

John Goetz

I think Lucia is referring to the new May value when June comes out.

Yes. That’s what people were joking about betting on. Whether the previous month’s data rises or falls seems like a crap-shoot, so that makes for good meaningless bets.

Steve McIntyre

I self-snipped a couple of snarky sentences in the post.

Richard deSousa

Dr. Spencer’s website reports that the June 2008 MSU data are still below normal…

http://www.weatherquestions.com/Roy-Spencer-on-global-warming.htm#satellite-temps

snip

Andrew

42 (Steve): I should have guessed, as I know you and John have had e-mail correspondence before. And here I thought you were making stuff up.

Tony Brown

Harold #48

I have plotted all the CET figures myself and tabulated them by hand and Mr Masterman is not always accurate with his monthly charts and comments. In particular his winter ones seem off, and I wonder if he has plotted the calendar year-jan feb dec, rather than the consecutive months dec/jan/feb

Very many of the records for warmth still remain from before the modern era-not surprising as regards cold, but more surprising for warmth, as this is all supposed to be in the (intermittent)Little ice age. The similarity with the 1730’s in particular is very marked. The overall analysis is fair enough though, in as much the average mean temperatures are broadly similar over the centuries, but that is because of a shifting of the characteristics of months. For instance we don’t tend to have the cold winters any more and October is generally warmer than ‘normal’ but counterbalanced by other months changing their traditional form. This was observed way back in the Anglo Saxon Chronicles whereby the winters had ‘less bite’ and summers became ‘hot and dry’ (the latter condition a situation we in the UK can only dream of!)

Tony B

streamtracker

Steve,

For someone who is well versed in statistics, it makes me wonder why you are so fixated on two data points. Given the noise in the data, it is really a statistically meaningless exercise. Your dislike for Hansen seems to be getting in the way of your grasp of trend analysis.

Steve McIntyre

#51. Look through the Surface Record category (left frame). A couple of things are going on. In the US, there are two different USHCN versions, with NOAA being more up to date than CDIAC. GISS used a stale source, though they seem to be responding to our suggestions, thereby improving their timeliness (no thanks to us, of course.) The USHCN stations get incorporated with a lag – the 2006-2007 values get picked up. I think that 2006 is picked up now.

GISS uses GHCN data for the ROW. GHCN collected a lot of data under special programs in 1990–1992 and this accounts for the high station populations prior to then. MAny people have thought that the stations have been shut down duie to funding; this is untrue. Many stations that appear in GHCN (And thus GISS) to have been shut down in 1990 or so can be located online. We’ve made fun of this on many occasions (Wellington NZ, Cobija Bolivia, many Canadian stations.

There’s a lot of information that’s been time consuming to collect. If you’rte interested, spend the time to wade through the Category.

TonyA

Here in Australia Prof Ross Garnaut has today released his Climate Change Review Draft Report. In chapter 5 he uses a graph (Fig 5.1) Average global average air temperature anomalies, 1850 to 2005 – Source: Brohan et al. (2006, updated 2008).

Would someone be kind enough to direct me to any CA discussion or criticism of this graph or its methodology. (Sincere apologies if there is a more appropriate thread for this enquiry.)

Nylo

Latest news, charles: Earth’s sea ice area is now bigger than in the previous years. Earth’s global temperature is, however, going down.

charles

Nylo

Look at the figures over several years and stop kidding yourself, you know denialism can only go on for so long, after a while it’s going to become so obvious the only ones that will be able to keep it up are those that believe the natural world only exist in out imagination so why worry.

Breaking news.

charles

Dave
Nasa good enough http://earthobservatory.nasa.gov/Newsroom/NasaNews/2007/2007092525664.html.

Melting ice and temperature happen to be related. When the heat is going into melting the ice, the heat isn’t going into raising the temperature. Have a look at the triple point of water if your interested in the concept.

Sea ice doesn’t cause water levels to rise or fall, its the melting land ice that matters. If you wish to argue a high school experiment for your enjoyment.

http://www.teachersdomain.org/resources/ipy07/sci/ess/watcyc/icesimulate/index.html

With regard to the sea ice. Record low artic see ice in 2007.
http://www.nasa.gov/vision/earth/lookingatearth/quikscat-20071001.html
What are you on about?

rex

Australians are probably the most gung-ho on GW. They are usually wrong but it takes years for admission this will take alot of time and they will spend billions on try to prevent something thats not happening

steven mosher

RE 52 steamtracker. are you bothered in the least by the way hansens first testimoney was staged? bothered in the least?

Stephen Richards

Steve Mc I hate it when you snip. It reminds me of when I was a child and mum used to say soemthing like ” I bought you a …. no I’ll tell you later, maybe” Get what I mean?. 🙂

Steve: If people complied with blog rules not to vent, not to be angry and not to talk about policy or impacts, then I’d snip less. It’s an unfortunate tradeoff. I took pride for a long time in never snipping, but it led to too many food fights.

TerryB

Methinks it’s time for a Global Warming – Mid Year Assessment Report Special Edition.
Now can I have a prize for acronym of the year???

Chris

I used to assume that that the scare stories about CO2 were correct, but the more I’ve looked into the issue, the more I’ve become convinced that (a) the existing temperature rises have been overstated and (b) the levels of certainty of the models predicting future rises have also been overstated.

However, I’ve suddenly become very worried about the satellite measurements, and I’m hoping someone can clear up a few questions I’ve been wondering about since I started looking at the AMSU temperatures today
http://discover.itsc.uah.edu/amsutemps/
As I understand it, the significance of these is that they are used to derive the UAH and RSS figures, and use the NOAA satellite.

1. (This is probably just a technical aside). If you look at the average temperature at 3,300 ft (http://discover.itsc.uah.edu/amsutemps/execute.csh?amsutemps+001) why is it round about 25-32F? If global average temperature is close to 60F, I wouldn’t expect temperature at 3,300 ft to be much under 50F.

2. More importantly, the traces are 1-2F above the 20-year average for both 2007 and 2008 so far (albeit 2008 is significantly lower than 2007), and if you add in the traces for 1998 and 1999 (as a possible comparison re: El Nino followed by El Nina?) then it appears that they were much closer to the 20-year average than to the present.

3. So why are UAH and RSS showing such low anomalies for the last few months if the AMSU anomaly close to the surface is so high? It must be because there are some low anomalies in the AMSU higher up, and this is true. At 14,000 feet, 2008 so far is normal, whereas 2007 and 1998 were equally high, and at 25,000 feet 2008 has the lowest anomaly of the last decade by some margin with 1998 the highest and 2007 a bit below 1998.

4. But isn’t it the temperature closest to the surface that we’re most interested in, as this is what actually affects the ice, crops, storms, forests or whatever else the AGW theorists like to emphasise? (Incidentally I assume the temperatures are weighted according to the pressure of the air at different heights?)

5. Are there agreed definitions of the height boundaries of lower and upper troposphere? I’ve read a lot about these but there never seems to be any precision….

6. Is there a mechanism by which lower anomalies at higher altitudes can result in lower anomalies at lower altitudes (with a lag)? I’m really hoping there is, as otherwise could it be that once the next El Nino and solar cycle kick in, the low anomalies that are currently giving us hope might never have actually reached the surface?

I’m really hoping I’ve got all this wrong, as it’s the first time in ages I’ve started to think AGW might be worse than I thought.

Geoff Sherrington

Re Australian Garnaut report just released

I searched the 60 pages for “satellite”, found three references the word, all non-relevant temperature. I am considering starting legal action because of failure to comply with the Terms of Reference of this Inquiry.

As a reward to my hard-working friends on Climate Audit,I give you the accolades given by Garnaut.

There is no doubt about the position of most reputed specialists in climate science, in Australia and abroad, on the risks of climate change (see Chapter 3).
There is strong support for the mainstream science from the leaders of the
relevant science academies in all of the major countries.1 The outsider to climate science has no rational choice but to accept that, on a balance of probabilities, the mainstream science is right.
There are nevertheless large uncertainties in the science. While there
is a clear majority view that there are high risks, there is debate and honest recognition of limits to knowledge about the times and ways in which the risk will manifest itself. Every climate scientist has his or her views on some issues that differ from the mainstream in detail.
There are prominent dissenters on this matter, gathered under the rubric of
‘sceptic’. For the most part ‘sceptic’ is a misnomer for their position, because these dissenters hold strongly to the belief that the mainstream science is wrong. I exclude from this generalisation a small number of climate scientists of professional repute, who accept the theory of the warming effects of higher concentrations of greenhouse gases, but hold the view that these warming effects are relatively or even trivially small in comparison with many other causes of climate variations that are beyond the control of humans.
The dissent took a curious turn in Australia in 2008, with much prominence
being given to assertions that a warming trend had ended over the last decade.
This is a question that is amenable to statistical analysis, and we asked
econometricians with expertise in analysis of time series to examine it. Their response, that the temperatures recorded in most of the last decade lie above the confidence level produced by any model that does not allow for a warming trend, is reported in Chapter 5 (Box 5.1).

(Ths Box 5.1 is quoted in my # 59 above).

It’s almost as if satellite temperature data do not exist.

David Stockwell

Two issues I see are these:

1. They used a one-sided test to show a significant trend term in the unit root model. A two-sided test with at conventional 1.96 critical value would have made two of the series non-significant. A Two sided tests should be used unless there is a very good reason for doing otherwise, as it could be argued that a one-sided tests was used simply as a device to make a conventionally non-significant difference significant.

2. They used a slide-and-eyeball test (see lucia’s blog about those) for showing that “the temperatures recorded in most of the last decade lie above the confidence band of forecasts produced by a model that does not allow for a warming trend.” But in a footnote they admit “The confidence bands in these graphs are calculated using the “dynamic forecast”option in Eviews and they do not incorporate estimation uncertainty.”. A slight increase in uncertainty concerning the parameters and the start point would put all temperatures inside the confidence band.

3. Not using MSU RSS data is a big omission.

DeWitt Payne

Chris,

The numbers on the web site are the raw brightness temperatures from each channel on the satellite. There is considerable overlap of the bandwidth of each channel so the pressure reading given is for the center of band. To get an actual value at a given altitude you have to correct for this overlap. RSS and UAH use somewhat different algorithms for this, and for cross-calibration between satellites when a new satellite comes on line to replace an old one. The basic problem of calculating an atmospheric temperature profile from a limited number of channels with noise and overlap is actually ill-posed. There are an infinite number of solutions. So you have to make some assumptions or use additional data to constrain the range of solutions. This falls into the general class of inverse problems, which are frequently ill-posed, at least at some point.

Dave Dardinger

Re #71

You’re all over the place here. Now, while I did take 23 semester hrs. of math in college, I do admit I didn’t take a statistics class among them, though I did take a graduate school course in biochemistry which was largely concerned with data analysis including calculating the statistical functions thereof.

why are you telling me that temperature has nothing to do with ice

I said no such thing. I asked why you were posting on ice melting on a thread concerned with global satellite temperatures?

Now, you’ve been trying to make some issue about temperatures pausing because ice is melting, but you’ve not presented anything concerning the math of the situation.

The molar heat of fusion for water is 6.02 kJ/mol. 1 calorie = 4.184 Joule So to convert 18g of H2O (l) to 18g H2O (s) requires 6020 / 4.184 = 1439 calories. Since we typically use watts per sq m for insolation, (and also outgoing IR) it’s useful to convert this to square meters. A cubic meter of water is about 1000 Kg or 1000/.018 moles H2O. so we have 1.439*1.000/18 x 10e9 = 8.0 x 10e8 calories to freeze one cubic meter of ice. Now a watt hour is about 860 calories so we need a bit less than a million watt hours to freeze this water. Now each year has 365.25*24 = 8766 hours so it takes about 115 watts per square meter to melt a meter of ice per year. Whether there is that much inside the arctic circle I don’t know.

OTOH, the amount of IR escaping from the arctic ocean is even more difficult to calculate. When there’s ice present it can be melted from below, but when there’s no ice the heat can escape to space. Thinner ice can transmit heat from the ocean below but thicker ice is more of a blanket and in the winter the surface temperature can get very, very low so that not much heat escapes to space. Thus thinner ice can perhaps let the earth’s air conditioning system work more efficiently. The question is what the balance is between additional late summer absorption of sunlight into areas where ice has melted vs the increased time of reasonably warm surfaces to radiate to space at temperatures near 0 deg C in early winter.

But of course there are many, many other things to be taken into consideration such as albedo, water temperature, thermohaline circulation, cloudiness, humidity, wind speed, air temperature, etc. Your concentration on “heat of fusion” is just a blue herring.

BTW, could someone check my math above? I hate making mistakes, but I hate worse letting them slip by and later looking silly if I push an idea and later find out I’m off an order of magnitude.

Geoff Sherrington

David Stockwell # 75, 76, 79

I suspect this is the wrong thread for this discusion but it is partly relevant and the best place I could find. Thank you for your comments, which I have read. Action time before finalisaion of this report is short, a few weeks, so I sent thefollowing email just now. Steve, I hope you did not mind me borrowing your links. Email follows to authors of paper quoted in Box 5.1:

Hi Trevor,

I refer to paper dated July 2008 with Farshid Vahid “Global Temperature Trends”. (As I write it is 5 July 1988, so I hope you are not forecasting. Ross’s report seems to reference this paper, but it was presumably published before 4th July 2008. How did such rapid events happen, with peer review etc? This is not important, I am just fascinated to know.)

Might I please ask:

1. Are you aware that satellite instruments have been measuring global temperatures since the late 1970s?

2. Were you and Farshid tempted to include these measurements in your paper?

3. If not, were you asked not to? If so, by whom?

4. The following URL leads you to a satellite dataset ending 30 June 2008, from University of Alabama at Huntsville, one of the two prime official sources of satellite temperatute data over this time span.
http://vortex.nsstc.uah.edu/data/msu/t2lt/tltglhmam_5.2

5. Would you be able to do an analysis similar to that relied upon by Ross for surface thermometry per Box 5.1 and make the results available? The satellite data has also had some adjustments, but some climate scientists regard it as more pure than the data that you used.

6.Were you/are you aware that there is extreme uncertainty about the validity of the data you used, especially the HADCRU data, where one external author (Douglas Keenan) has a peer-reviewed published paper demonstrating a high probability of fabrication of data?

7. Were/are you aware that the HADCRU, GISS and NOAA data are all subject to prior numerical and statistical adjustments, some plausibly subjective; and that there should be caution used when performing statistical tests upon data that have already undergone several layers of statistical adjustment?

8. Might you please forward the computational code you used in your 2008 joint paper so that colleagues of mine can use it and comment upon it?

9. I have appended graphs of the two main satellite data sets. (The word “satellite” is found in only 3 or so places in Ross’s report, none of them relevant to this topic). You might find a preliminary look to be challenging.

MSU data is Uni Alabama

RSS data

10. The shortness of the timetable set by politicians has caused convenient rather than formal attribution and data sources. It is my preference to properly attribute to the source.

Many thanks

Geoff Sherrington
Scientist.

Chris

Hi Anna, thanks for your reply. I think there may be a first part missing, but anyway, just to respond to your points:

1) I’m talking about figures consistently 1-2F above the 20-year average which is too short term to be explained by a retreat from the little ice age; besides, a lot of us are disputing whether temperatures are going up longer term, that’s why I’m on this thread in the first place;

2) Are you familiar with the actual figures for adiabatic lapse rates? (http://en.wikipedia.org/wiki/Lapse_rate)
The reason I said I didn’t expect temperatures at 3,300ft to be much below 50F compared to global surface average of 60F is because the lapse rate is typically about 3.5F per 1000 ft (i.e. the rate at which it cools as you go up from the surface)

3) I certainly do ask how the surface temperatures are arrived at, and that’s why I’ve started to follow satellite temperatures to try and get a truer picture. But now I’ve discovered that the satellite temperatures closer to the surface show positive anomalies that are much greater than even the most controversial surface figures, I’m starting to wonder.

John M

Chris,

I see your point. Now that you mention it, I’m not sure what those graphs are plotting year-to-year. The Watts Up site has a link to the tabular data here.

I apologize for the formatting, since I don’t know how to make tables in Comments, but here’s “lower troposphere” data for 1998, 2007, and 2008. I can’t match these up with any of the low altitude graphs.

Mo 1998 2007 2008
Jan 0.582 0.594 -0.046
Feb 0.753 0.450 0.020
Mar 0.528 0.403 0.089
Apr 0.770 0.244 0.015
May 0.645 0.199 -0.183
Jun 0.562 0.203 -0.114
Jul 0.510 0.255
Aug 0.513 0.286
Sep 0.432 0.201
Oct 0.394 0.231
Nov 0.190 0.209
Dec 0.289 0.114

It would be interesting to see a detailed explanation on how the the listed temperatures are derived, at least relative to the data graphs you’re referencing.

Chris

John,
Great that someone’s finally seen my point. The discrepancy seems to be to do with how “lower troposphere” is defined. I believe this is defined as up to 10,000 metres (about 33,000 feet) but I don’t know if this is universal.
With AMSU if you look at the figures for 14,000 feet you will see that 2008 is already down to average (from a positive anomaly in 2007), and at 25,000 feet it appears it is well below (unfortunately the 20-year average is not available but you can still compare 2008 to the previous 10 years individually)
So it’s easy to see how the “lower troposphere” as a whole can have a negative anomaly (assuming the same weighting for all levels of the lower troposphere – though seeing as there is less than half as much air at 25,000 feet I wonder if these levels do or should have a lower weighting?)
What I’m struggling to see is why there appears to be so little information and discussion out there on the variations within the lower troposphere itself.
If the data was split into lower troposphere 1 (up to 5,000m) and lower troposphere 2 (5,000-10,000m) and UAH/RSS used lower troposphere 1 rather than “lower troposphere”, perhaps all the discussions taking place at the moment about the satellite data (especially vs the surface data) might be quite different?

Chris

John,

I know what you mean, it’s difficult to understand, but we do the best we can and it’s easy to assume others know more than they actually do 🙂
I just did some more googling myself and came across the following:

http://www.remss.com/msu/msu_data_description.html

It’s really useful, as it explains what DeWitt was saying, and more.
For one thing, it gives the weighting functions for the main channels, and a quick look at TLT (lower troposphere) channel which we’re concerned with here shows the kind of weighting I’d thought would be reasonable i.e. tapering substantially the higher you go. It appears the channel goes higher than 10km (up to about 12km or so, but the weighting at this level is minimal)
It’s also interesting to see the high anomalies in the Northern Hemisphere as you approach the Arctic.

Taking the channels as a whole, it’s clear that the positive anomalies become greater the closer you go to the surface (check out “Decadal Trends” section)

But I don’t want to distract from the original point re: the near surface graph, which you admit confuses you, and I admit makes me worried that AGW might be worse than I thought:

http://discover.itsc.uah.edu/amsutemps/execute.csh?amsutemps+001 (and click “1998” and “redraw graph” to see disturbing evidence that 2007 was much warmer than 1998 in the lower part of the lower troposphere, i.e. the first few thousand feet above the surface)

How can this be explained without AGW?

Len van Burgel

Chris,

Something doesn’t seem right with the graphs. The 2007 line is much higher than the 20 year record high line.

Chris

Sorry, minor error I meant to say “the UAH figures which are based on them” rather than “the UAH figures on which they are based”.
So long….for now…

anna v

Chris,

I had been block quoting part of your post when the glitch happened.

I have no problem seeing the 20 year average below the fairly constant last 8 years. The average of the rising numbers until 1998, if it is .2C per year will be .5C bellow the average of the constant since 1999, assuming that 1999 stays at the level of 1998. Averaging the two averages will still keep the twenty year average lower than the current constant state.

I have no opinion on the absolute values. It would be good if the site had provided us with a FAQ.

DeWitt Payne

Dave,

You dropped a decimal point when you divided 1.439x10e9 by 18. It’s 8x10e7 not 8x10e8. So it only takes an excess of about 12 W/m2 to melt a meter of ice. Cloud cover is incredibly important to cooling. In the cold dry air of the polar regions, it’s possible to radiate heat to space at over 100 W/m2. Cloud cover reduces that radiation loss by about an order of magnitude.

TheDude

Charles,

You seam to want to substitute ice measurements for temperatures because the former is a proxy for the latter. This is nonsensical because we can get more accurate measurements of temperatures by measuring temperature itself. It is the same as measuring global warming by its proxies instead of actually measuring the warming. It is an unnecessary step that increases the error rate and adds no new information about the actual phenomenon in question.

Jon

TheDude (#101) writes:

You seam to want to substitute ice measurements for temperatures because the former is a proxy for the latter. This is nonsensical because we can get more accurate measurements of temperatures by measuring temperature itself. It is the same as measuring global warming by its proxies instead of actually measuring the warming. It is an unnecessary step that increases the error rate and adds no new information about the actual phenomenon in question.

This is wrong. First, all temperature measurements are made by means of proxy. Whether that proxy is the expansion of mercury in a glass tube or something else. Some proxies are better than others. Records have their own problems. So your point must be about combining proxies. If I have two sources of measurement, one noisier than the other, how should they be combined to get the most “information”? Pick the least noisy only? no, you combine them through an inverse covariance filter.

David Smith

An anomalously cool region of subsurface water (black circle) persists in the tropical North Pacific. It has strengthened in recent days.

The image is a south-north slice of the Pacific subsurface at 140W. The temperature anomaly is indicated by color.

An 8 degree C anomaly 60 meters beneath the surface is impressive.

The ocean/air pumping mechanism for making the water rise at this latitude is unknown to me and I’ll be surprised if it affects the 28C surface waters. Still, it’s interesting to watch and wonder what-if, as coolness here likely affects trade wind strength and thus may play some role in ENSO.

DeWitt Payne

More on brightness temperature.

The brightness temperature is calculated by assuming the radiation is coming from a blackbody at that temperature. Blackbody emission at any wavelength (or frequency) is calculated using the Planck equation. It’s actually simpler than that, though, because the Rayleigh-Jeans approximation holds for microwave frequencies and the brightness temperature is directly proportional to the actual temperature. The AMSU channels measure the emission from the oxygen band at 60GHz. Oxygen emits and absorbs in the microwave region because it has a non-zero magnetic dipole moment. The brightness temperature and the actual temperature will only be the same if the emission being measured is saturated (the proportionality constant is 1). Almost all the channels (all but one, IIRC) in the AMSU measure on the wing of the emission band rather than the peak so emission is not saturated and the observed brightness temperature will be lower for those channels than the actual temperature.

Michael Smith

streamtracker wrote in 108:

#55

It’s hardly unreasonable to do a comparison if the two points in a topical post.

Actually it is very misleading and statistically irrelevant.

Are you hoping that if you pretend that Steve gave us nothing but two data points, and proceeded to draw some conclusion from them, that we will forget the actual content of his post?

RomanM

#103 Jon

If I have two sources of measurement, one noisier than the other, how should they be combined to get the most “information”? Pick the least noisy only? no, you combine them through an inverse covariance filter.

Not exactly. This assumes that the sources are in fact unbiased for the parameters being estimated. Including a source which is biased (regardless of its variability) can produce an estimator with a higher mean-square error than omitting it. In this case, I suspect that, because the ice has a considerably higher variability than the mercury, it would require a demonstration that that the bias is very low before using it as a “proxy”. Considering the effects of other variables (cloud, sea currents, etc.) on the melting of ice, this sounds like quite a task.

Dinosaur

What’s GLB is that global?

I didn’t find it here
http://www.climateaudit.org/?page_id=1288

Nylo

GISS data for June’s global surface temperatures is already available:

Not bad for a warming scenario.

EW

#115 What? Even GISS is going down? Are they sure to have applied all adjustments available??

MattN

Steve, any sneak-peak at July data yet?