Talk:Satellite temperature measurements

From Wikipedia, the free encyclopedia
Jump to: navigation, search
WikiProject Meteorology (Rated C-class, Mid-importance)
WikiProject icon This article related to weather data or instruments is part of the Weather Data and Instrumentation sub-project of WikiProject Meteorology and Weather Events, an attempt to standardize and improve all articles related to weather or meteorology. You can help! Visit the project page or discuss an article at its talk page.
C-Class article C  This article has been rated as C-Class on the project's quality scale.
 Mid  This article has been rated as Mid-importance on the project's importance scale.
WikiProject Environment / Climate change  (Rated C-class, Mid-importance)
WikiProject icon This environment-related article is part of the WikiProject Environment to improve Wikipedia's coverage of the environment. The aim is to write neutral and well-referenced articles on environment-related topics, as well as to ensure that environment articles are properly categorized.
Read Wikipedia:Contributing FAQ and leave any messages at the project talk page.
C-Class article C  This article has been rated as C-Class on the project's quality scale.
 Mid  This article has been rated as Mid-importance on the project's importance scale.
Taskforce icon
This article is supported by the Climate change task force.

Step Change?[edit]

See here. Using the UAH and RSS data sets from 1980 to present, apply a linear trend analysis (similar to current chart image in the article). However run the linear trend from 1980 to 1997, then from 1998 to present. Reason -- because of the huge temperature spike attributed to the El Nino in 1998. If you do this, what jumps out of the data set is not a gradual 0.2C degree temperature increase by decade, but instead a huge step change jump. Do you all think this is a valid linear trend analysis compared to the current image linear trend? SunSw0rd (talk) 14:44, 6 February 2011 (UTC)

(Moved from own section to here):

It could be made a lot clearer exactly what the table in this section is supposed to show. It seems to purport to be the global decadal trend since Dec 1978 according to Christy et al. Is that with retrospective corrections or without?

Can anyone explain the jump in the values on the table in this section of the article

1997 0.040

1998 0.112

I know 1998 was hot but to to shift the trend over 20 years that much would take some doing!

IanOfNorwich (talk) 10:06, 26 February 2011 (UTC)

Having seen SunSw0rd's earlier comment on this and following his link, it looks like the table is accurate and the jump is due to the anomaly in 1998 divided by 20 year being large compared to the trend. This is in good agreement with the table although the site uses up-to a given year so all the years differ from the chart by 1.

So the table does seem to correctly illustrate its point. IanOfNorwich (talk) 10:55, 26 February 2011 (UTC)

While on the subject of this section, it would be nice to have a table of the most up to date RSS and UAH decadal trends for each of the channels, how about this:

Channel Start End Date RSS Global Trend (K/decade)[1] UAH Global Trend (K/decade)
TLT 1979 2011-01 0.163 0.140[2]
TMT 1979 2011-01 0.099 0.052[3]
TTS 1987 2011-01 0.008
TLS 1979 2011-01 -0.306 -0.391[4]

I was thinking to add something like this table to the MSU temperature measurements article. As a side note, to avoid confusion with the data in the article, RSS trends updated to version 3.3 are available over the graphs at the end of the page,those at the middle of the page are still the v3.2 trends:
Changes description:
tlt:+0.148 tmt:+0.091 tts:+0.001 --Giorgiogp2 (talk) 21:33, 27 February 2011 (UTC)

Thanks, Giorgiogp2, I was just beginning to scratch my head as I'd seen tlt:+0.148 on the graph there. I'll adjust and add it to the MSU article. I'm for adding it to this page also. If there's room for a list of historic trends of UAH T2LT then there should be room for this.

While on that subject I'd like to see the RSS data in that table as well. If you look at this the change is from about 0.8 to 0.15, which is a similarly sized jump but gives a slightly different impression. (Perhaps that is now different in v3.3?) I don't know if one data set is better than the other, they both seem worthwhile. I can't, however, find that data in that form. Anyone know where the UAH historic global trend averages come from? I can't find it in any nearby ref. IanOfNorwich (talk) 23:59, 27 February 2011 (UTC)

I looked to add it to the MSU article but the data is currently in the lead and nowhere obvious to put it (I won't put a table in the lede). It's too late for me to be restructuring articles right now so it's not there. I have, however, added it to the 'Trends from the record' section. There is room for a table of the latest trends in that section! IanOfNorwich (talk) 00:10, 28 February 2011 (UTC)

Now that it is 2013, go to the original link (here) and you will see the step change even more clearly. I belive this gives a much more accurate perspective of the actual data than does the current image. Also of interest is that the 2 satellite data sets diverge after the break in opposite directions regarding trend. SunSw0rd (talk) 16:19, 30 April 2013 (UTC)

Note also that the step change break is seen clearly in the HadCRUT#HadCRUT3 (sea surface, not satellite) results here. I believe this confirms the logic of showing the trends pre and post break for 1998. SunSw0rd (talk) 16:29, 30 April 2013 (UTC)
(edit conflict) Tamino had a good educational piece last year on how step changes can be both a better fit to the the data, but less significant than a linear trend. Essentially, the lower residuals are offset by the fact that there are more free parameters in the step model. In this particular case, the point of the graph is to show the match in the various surface and satellite records in terms of their overall trends, hence why a linear trend is computed for each. Sailsbystars (talk) 16:31, 30 April 2013 (UTC)
Yes, I see the Tamino article. Of interest perhaps is that he concludes with: "The step-changes model has smaller residuals, but the two-straight-lines model has lower AIC (better model)." Note that his last chart has 2 linear trend lines (not one). He is comparing it to the multi-step line models and concludes that it is superior -- and that it may be for showing clarification. I concur the last chart is best. Note however he did not show a Malacca chart with a single linear trend line -- if he had, I would say the one shown with the 2 trend lines is better (more accurate). SunSw0rd (talk) 16:45, 30 April 2013 (UTC)

See here for an update showing HadCRUT compared to RSS from 1987 to present. Again the break can be seen clearly and you can see why there have been media reports about the lack of warming for the past 15 years. SunSw0rd (talk) 12:06, 1 November 2013 (UTC)

Or see RSS MUS Lower Trop. Global Mean for the past 17 years. SunSw0rd (talk) 15:43, 4 November 2013 (UTC)

Update trends[edit]

Where did the data for the recent trend updates come from? The same sources as the previously? Then the access-dates should be updated. Otherwise it looks like they came out of thin air. Which I realize is more or less what the satellites do, but it should be documented. ~ J. Johnson (JJ) (talk) 21:18, 13 February 2012 (UTC)

Comparison Between Satellite Data Sets and Computer Models[edit]

See here. How valid is this? SunSw0rd (talk) 19:39, 8 November 2013 (UTC)

I can't reproduce it myself. I get that there was a dT of 0.5 C whereas Dr. Spencer's graph only shows an increase of 0.1 C. I note that he seems to be comparing only the 20S-20N latitudes, which I don't have an easy way of getting access to. Not sure the reason for why there's such a big change with an extra 20 deg of latitude, (I used 30S-30N) though, although the poles have been warming faster than the rest of the planet.... Sailsbystars (talk) 15:37, 9 November 2013 (UTC)
I suspect that this is closer since it is both the UAH and RSS datasets (not HadCRUT). But he applied "squares" and I don't see how to do that with the Wood For Trees site. Also WFT only provides a data start of 1980 for the satellites even if you set it to 1978. SunSw0rd (talk) 03:08, 10 November 2013 (UTC)

Problem with UAH Data Graph and Current UAH / RSS Data Graph[edit]

See UAH Satellite-Based Temperature of the Global Lower Atmosphere (Version 5.6). Now compare to the graphic labelled "Surface and Satellite Temperatures" on the top right of the page (I don't include the link here because it makes a huge image, just click on current image on page to view). There is a discrepancy between the UAH record in the two graphics. Specifically the current graphic shows UAH (red) having many recorded results between +0.2 to +0.4 between 1980 and 1985. However the UAH graphic posted in the link I provided above shows no correlation to this and only exceeds +0.1 for a single data point in 1983. Looking back and forth between the graphics it appears the current Wikipedia graphic has shifted the UAH data upwards by approximately +0.3 for the years 1980 to 1985. What is the explanation? SunSw0rd (talk) 19:13, 20 December 2013 (UTC)

You mean File:Satellite Temperatures.png? That and your fig don't have the same anomaly base William M. Connolley (talk) 20:21, 20 December 2013 (UTC)
Yes I mean that. Both figures are referencing UAH so -- what is the specific basis for the difference? Since they are referencing UAH data sets -- are the data sets different? Also please note the link below the File:Satellite Temperatures.png figure (UAH 2003; data set tltglhmam version 5.2 with 2009 updates) is broken. What would be the correct link to the UAH anomaly base for the current image? SunSw0rd (talk) 01:04, 21 December 2013 (UTC)

I suggest that this RSS versus UAH graphic is more accurate than File:Satellite Temperatures.png because it correlates directly with File:Radiosonde_Satellite_Surface_Temperature.svg. The existing graphic lower down in the article File:Radiosonde_Satellite_Surface_Temperature.svg more accurately portrays the data specific to the article which is "Satellite temperature measurements". I recommend replacing File:Satellite Temperatures.png with File:Radiosonde_Satellite_Surface_Temperature.svg and then deleting File:Radiosonde_Satellite_Surface_Temperature.svg from its current location in the article. Comments? SunSw0rd (talk) 14:53, 4 January 2014 (UTC)

You haven't understood what I said about different anomaly bases. Until you've worked that out, you've got no business discussing the accuracy of any of these pix William M. Connolley (talk) 17:55, 4 January 2014 (UTC)
O Rly? Because the reference link to the data for the current pic is broken. And the article has a superior pic already lower down the page. Did you ignore? SunSw0rd (talk) 21:47, 4 January 2014 (UTC)


The article writes: "Climate model results summarized by the IPCC in their third assessment show overall good agreement with the satellite temperature record.". Is this a case of climate models being used to actually predict the data successfully (i.e. before it was measured), or a case of climate models successfully fitting over past data? J1812 (talk) 10:40, 17 January 2015 (UTC)

See the funny-looking "[39]" coloured blue after that statement? Its what's called a "reference". You could, like, read it William M. Connolley (talk) 15:54, 17 January 2015 (UTC)

I would like to read their explanation as the RSS data to 2015 is not fit by any climate model as far as I am aware. But the link is broken so I guess the art. will have to be edited? How about this as being a bit more honest: "Although climate models initially agreed with the data they were tuned with, there was a gradual divergence between model projections and satellite measurements. This problem is not unexpected as no climate model has ever passed hindcasting tests". — Preceding unsigned comment added by (talk) 23:00, 11 February 2015 (UTC)

  1. ^ "RSS / MSU and AMSU Data / Description". Retrieved 26 February 2011. 
  2. ^ "MONTHLY MEANS OF LOWER TROPOSPHERE LT5.4". UAH. Retrieved 26 February 2011. 
  3. ^ "MONTHLY MEANS OF MID-TROPOSPHERE MT5.4". UAH. Retrieved 26 February 2011. 
  4. ^ "MONTHLY MEANS OF LOWER STRATOSPHERE LS5.4". UAH. Retrieved 26 February 2011.