…PLUS AN OBVIOUS ERROR IN THE NEW NOAA ERSST.V4 PAPER
Guest Post by Bob Tisdale
We’ve discussed NOAA’s new ERSST.v4 “pauses-buster” sea surface temperature reconstruction in a number of posts this year. They are linked at the end of this post. We can add yet another curiosity to the list…this time relating to the global ERSST.v4 data during the first half of the 20th Century. Additionally, there is an error in a new paper about the NOAA ERSST.v4 that I want to discuss as well.
This post was prompted by a paragraph in NOAA’s recent paper about the uncertainties of their ERSST.v4 sea surface temperature reconstruction. That paper is Huang et al. (2015b) Further Exploring and Quantifying Uncertainties for Extended Reconstructed Sea Surface Temperature (ERSST) Version 4 (v4), which is presently available (paywalled) as an Early Online Release. Two things in the following paragraph caught my eye (my boldface):
A recent study (Karl et al. 2015) indicated that the trend of globally averaged SST in ERSST.v4 in the most recent decades (0.99°C century-1; 2000-2014) is as large as in the longer period of 1951-2012 (0.88°C century-1). Figure 6b shows the histogram of the trend during the longer period. The trend ranges from 0.7°C to 1.0°C century-1, which is higher than the long term trend shown in Figure 6a, indicating stronger oceanic warming since the middle of the 20th Century. Factor analyses indicate that the major contributor to this trend uncertainty is the ship-buoy adjustment (the 9th parameter; Fig. 7b).
FIRST, THE ERROR IN THE NEW HUANG ET AL. (2015b) PAPER
For the period of 1951-2012, Karl et al. (2015) listed the trend of 0.088 deg C/decade (0.88 deg C/century) for their “old” ERSST.v3b data, not the “new” ERSST.v4 data. The 1951-2012 trend for the “new” ERSST.v4 data shown in Table S1 from the Supplementary materials for Karl et al. (2015) is listed as 0.100 deg C/decade (1.00 deg C/century). Their Table S1 is included as my Table 1. NOAA repeated that “typo” throughout Huang et al. (2015b).
Looks like the peer reviewers missed an obvious mistake.
In Figure 1, I’ve added vertical red lines to Figure 6 (SSTa trend uncertainty histograms) from Huang et al. to show the data trends listed in Karl et al. (2015) for the periods of 1951 to 2012 and 2000 to 2014. (Refer again to Karl et al. Table S1, my Table 1.) I’ve also shown the 0.88 deg C/Century trend that was erroneously listed in Huang et al. (2015b) for the period of 1951 to 2012.
Curiously, but not surprisingly, the actual sea surface temperature data trends align with, or are near to, the extreme high end trends in the uncertainty histograms.
And if you’re wondering about the ERSST.v4 data trend for the period of 1901 to 2014 (Cell a in Figure 6 from Huang et al. 2015b), the data for the latitudes of 60S-60N (global excluding the polar oceans) present a linear trend of 0.073 deg C/decade (0.73 deg C/century)….once again toward the high end of the histogram. (Data available through the KNMI Climate Explorer.)
SO THE WARMING RATE FOR 1951-2012 IS HIGHER THAN THE RATE FOR 1901-2012
The above quote from Huang et al. (2015b) included (my brackets):
The trend ranges from 0.7°C to 1.0°C century-1 [for the period of 1951-2012], which is higher than the long term trend shown in Figure 6a [for the period of 1901-2014], indicating stronger oceanic warming since the middle of the 20th Century.
There’s nothing unusual about global sea surface temperatures since 1951 having a higher warming rate than the data since 1901. All sea surface temperature and night marine air temperature datasets (end products) available from the KNMI Climate Explorer show that simple relationship. See Figure 2. The difference between the shorter-term (1951-2012) and the longer-term (1901-2012) trends depends on the dataset.
Figure 2 (Typo Corrected in ERSST.v3b Graph)
Figure 2 (and Figures 3 through 5) include global (60S-60N) ocean surface temperature anomalies referenced to the period of 1971-2000, from top to bottom:
- NOAA’s ERSST.v4 Sea Surface Temperature (NOAA’s new “Pauses-Buster” data)
- NOAA’s ERSST.v3b Sea Surface Temperature (NOAA’s “old” data)
- UKMO’s HadNMAT2 Night Marine Air Temperature (Which NOAA uses to adjust their ERSST.v4 data to account for ship-based temperature biases, from buckets and ship inlets)
- UKMO’s HADSST3 Sea Surface Temperature (An uninterpolated product used in the UKMO HadCRUT4 global land+ocean surface temperature product), and
- UKMO’s HadISST Sea Surface Temperature (the most-often-used sea surface temperature dataset used in scientific studies, even though it has the lowest warming rates.)
Huang et al. included trends for the periods of 1901-2014 and 1951-2012, using two different end years. For ease of illustration, I ended the data for both periods in 2012. That also agrees with the periods shown in Table 2.5 from Chapter 2 of the IPCC’s 5th Assessment Report. The exception to the 2012 end year is the HadNMAT2 data, which run through 2010 because the dataset ends then.
The data suppliers account for sea ice differently…thus the use of the latitudes of 60S-60N.
In the quote above, Huang et al. (2015b) are suggesting that global warming is accelerating. But Huang et al. (2015b) failed to note something important.
THE WARMING RATE FOR 1901-1950 IS ALSO HIGHER THAN THE RATE FOR 1901-2012
Figure 3 includes trend comparisons for the five ocean surface temperature products, but this time comparing the warming rates for 1901-1950 and 1901-2012.
Figure 3 (Corrected typo on HADSST3 graph)
“ew” “new” and “old” NOAA sea surface temperature products show slightly higher warming rates for 1901-1950 than they do for 1901-2012. On the other hand, the UKMO’s HadNMAT2, HADSST3 and HadISST show noticeably higher warming trends for the early period of 1901-1950.
ONLY THE NOAA “PAUSES-BUSTER” DATA SHOW A LOWER WARMING RATE IN THE FIRST HALF OF THE 20TH CENTURY
Figure 4 compares the warming rates for 1901-1950 and 1951-2012 for the 5 ocean surface temperature products.
Figure 4 (Corrected typo on HADSST3 graph)
With the “old” NOAA ERSST.v3b data, the warming rates for the early (1901-1950) period are comparable to those for the later period (1951-2012). The early period has a slightly higher warming rate than the later period in the HadNMAT2 and HadISST datasets…though it could be argued that they’re comparable. The HADSST3 data have a noticeably higher warming rate in the early period.
The exception is NOAA’s new ERSST.v4 “pauses buster” data, which show a noticeably lower warming rate in the first half of the 20th Century.
DIVIDED INTO TWO EQUAL-LENGTH (50-YEAR) PERIODS
Someone is bound to note that we’re not comparing periods of equal length. So, for Figure 5, I ended the later period in 2000, breaking the 20th Century in two.
Figure 5 (Corrected typos in ERSST.v4 and HADSST3 graphs.)
THE DATASET USED BY NOAA FOR SHIP-BASED BIAS ADJUSTMENTS DOES NOT SUPPORT THE LOWER WARMING RATE IN THE FIRST HALF OF THE 20TH CENTURY
Figure 6 compares the warming rates for the periods of 1901-1950 and 1951-2010 using the global (60S-60N) HadNMAT2 and the new NOAA ERSST.v4 “pauses-buster” sea surface temperature data. NOAA adjusts their ERSST.v4 sea surface temperature data using HadNMAT2 night marine air temperature data to account for ship-based temperature measurement biases (buckets of various types and ship inlets).
Curiously, NOAA’s ERSST.v4 data have a noticeably lower warming rate than the reference HadNMAT2 data during the first half of the 20th Century (and a noticeably higher warming rate from 1951-2010). Those additional tweaks are the reasons why the NOAA ERSST.v4 are the outlier, showing a lower warming rate in the first half of the 20th Century than for the period from 1951-2012.
Did NOAA adjustment (modify, tweak, manipulate, etc.) their ERSST.v4 “pauses-buster” sea surface temperature data during the first half of the 20th Century so that the ocean surfaces showed a slight acceleration in global sea surface warming for 1951-2012? The other ocean surface temperature datasets don’t show the same disparity over those two time periods. If fact, the UKMO’s HADSST3 dataset (which like ERSST.v4 is also adjusted for bucket, ship inlet and buoy biases) shows a noticeably higher warming rate for 1901-1950 than for 1951-2012.
ADDITIONAL POSTS ABOUT NOAA’S NEW ERSST.v4 “PAUSES-BUSTER” SEA SURFACE TEMPERATURE DATA
In previous posts at WattsUpWithThat and at my blog ClimateObservations, I’ve discussed the new NOAA “pause-buster” sea surface temperature dataset (ERSST.v4) a number of times since the publication of Karl et al. (2015)—latest to earliest:
- Busting (or not) the mid-20th century global-warming hiatus
- Pause Buster SST Data: Has NOAA Adjusted Away a Relationship between NMAT and SST that the Consensus of CMIP5 Climate Models Indicate Should Exist?
- Open Letter to Tom Karl of NOAA/NCEI Regarding “Hiatus Busting” Paper
- More Curiosities about NOAA’s New “Pause Busting” Sea Surface Temperature Dataset
- NOAA/NCDC’s new ‘pause-buster’ paper: a laughable attempt to create warming by adjusting past data
Many thanks to those readers on the cross post at WattsUpWithThat who found all of the typos. My apologies.