As Spencer notes:
As far as the race for warmest year goes, 1998 (+0.424 deg. C) barely edged out 2010 (+0.411 deg. C), but the difference (0.01 deg. C) is nowhere near statistically significant. So feel free to use or misuse those statistics to your heart’s content.
And according to Hanson:
NASA’s announcement this year – that 2010 ties 2005 as the warmest year in the 131-year instrumental record – made headlines. But, how much does the ranking of a single year matter?
Not all that much, emphasizes James Hansen, the director of NASA’s Goddard Institute for Space Studies (GISS) in New York City. In the GISS analysis, for example, 2010 differed from 2005 by less than 0.01°C (0.018 °F), a difference so small that the temperatures of these two years are indistinguishable, given the uncertainty of the calculation.
It is interesting that both point to the same difference between the preivously warmest year (0.010 C), though Spencer's is cooler that the previously warm year and Hansen's is warmer. Those who are not familiar with the different temperature data sets may be surprised that Hansen and Spencer name two different years as the "previous warmest". For a discussion of the different temperature sets see the Hansen link above. For my take on the different temperature records see this post.
Below are charts showing data and trends for the five main temperature records. I charted from 1979 because UAH and RSS start from that date. As always click on the graphics for larger and clearer versions. (For a description of the method that I used to produce these charts see the appendix at the end of this post.)
All of the data sets show a strong warming trend since 1979. Some people argue that global warming has stopped sometime in the last 12 years, either 1998, 2001 or 2002. I have discussed this issue at this post, and concluded:
There are three major reasons that disprove the claims that global temperatures are not increasing:
1. their reliance on short time preiods to produce trends - which is statistically absurd
2. their use of the complex concept of statistical significance to give a false impression to those who don't understand any statistics
3. their reliance on just one of the many indicators of warming temperatures.
For the reasoning that led to those conclusions check the post.
Even if we ignore basic statistics and fall for the skeptics' trick of using short term time spans the argument still does not hold water. The chart below shows the years that have a negative trend to the end of 2010:
| GISS | 2005 |
| HadCRU | 2001 2002 2003 2004 2005 |
| NOAA | 2002 2003 2004 |
| RSS | 2002 |
| UAH | ----- |
So bending over backwards to make the skeptic argument (and breaking some basis statistical rules in the process) we have only 1/5 of the trends from 2001 that are negative. Note also that half of the negative trends come from one data set - HadCRU. There is an enoumous irony here as climate skeptics try to discredit climate science by claiming (on the basis of stolen emails) that HadCRU distort the climate record to show excessive warming!
 |
 |
 |
 |
 |
Appendix
To produce these charts I followed the steps below:
1. I downloaded the data for each of the data sets. To do this I used programming procedures (RClimate.txt) developed by Kelly O'Day. The program that I wrote used O'Day's procedures to download each of the five data sets to a separate file. This is an easy way to ensure that data is accurate and up to date, as the producers of the data sometimes revise the values of earlier months.
The section of code below references O'Day's functions to download the data. I did not include the functions as they are quite long - they can be found in the RClimate.txt file. The major change that I made to Kelly O'Day's code was an output_link that was appropriate to my directory structure.
## The functions require the packages ggplot2 and reshape to operate so make sure that the packages are loaded
library(ggplot2)
library(reshape)
##Get and store GISS data
m <- func_GISS() head(m) output_link <- "C://Learn_R/data_GISS.csv" write.csv(m, output_link, quote=FALSE, row.names = F) ##Get and store Hadley Data m <- func_HAD() head(m) output_link <- "C://Learn_R/data_HAD.csv" write.csv(m, output_link, quote=FALSE, row.names = F) ##Get and store NOAA data m <- func_NOAA() head(m) output_link <- "C://Learn_R/data_NOAA.csv" write.csv(m, output_link, quote=FALSE, row.names = F) ##Get and store RSS data m <- func_RSS() head(m) output_link <- "C://Learn_R/data_RSS.csv" write.csv(m, output_link, quote=FALSE, row.names = F) ##Get and store UAH data m <- func_UAH() head(m) output_link <- "C://Learn_R/data_UAH.csv" write.csv(m, output_link, quote=FALSE, row.names = F)
2. I used programs that I wrote to chart the data. The code for producing the NOAA chart can be seen below. The only difference in the files to chart the other data sets is the name of the .csv file in the link line. These programs will chart data from any time in the data set up to the latest data. The date to start should be entered in the yearfrom = line.
########### NOAA data and trend analysis ##############
## STEP 1: SETUP - Source File Link
par(las=1)
link <- "C:\\Learn_R\\data_NOAA.csv"
## STEP 2: READ DATA
my_data <- read.table(link, sep = ",", dec=".", skip = 0, row.names = NULL, header = T, colClasses = rep("numeric", 2), na.strings = c("", "*", "-", -99.99,99.9, 999.9), col.names = c("yr","yr_frac", "temperature"))
## STEP 3: MANIPULATE DATA ## Change the year in the next line to the year that # you want to plot data and trends from, not before 1880!! yearfrom = 1979 Title <- c("NOAA data and trend from", yearfrom) newdata <-(subset(my_data, yr>=yearfrom))
## STEP 4:CREATE PLOT AND TREND LINE
plot(temperature ~ yr, data = newdata, type="l", col = "red", xlab = "year", ylab = "temp", main = Title)
abline(lm(temperature ~ yr, data = newdata), col = "blue")
No comments:
Post a Comment