Guest Post by Bob Tisdale
This post will serve as part 2 of the 2015 update of the model-data comparisons of satellite-era sea surface temperatures. The 2014 update is here. This, the second part, contains time-series graphs. But the data and model outputs are being presented in absolute, not anomaly, form.
The climate models used by the Intergovernmental Panel on Climate Change (IPCC) are not simulating climate as it exists on Earth. That reality of climate models will likely come as a surprise to many climate laypersons.
We presented in part 1 of this series how the spatial patterns of the modeled warming rates for the surfaces of the global oceans from 1982 to 2015 (the era of satellite-enhanced sea surface temperature observations) show no similarities to the spatial patterns of the observed (data-founded) warming and cooling. And we discussed why it’s important that the models used by the IPCC are capable of simulating where and when and why the temperatures of the ocean surfaces vary. It’s relatively easy to understand. Where and when the surfaces of the oceans warm, don’t warm, or even cool naturally and by how much—along with other naturally occurring factors—dictate where and when land surface air temperatures rise and fall and where precipitation increases or decreases…on annual, decadal and multidecadal bases.
In part 2, we’re presenting the model-data comparisons in time-series graphs globally and for a number of subsets. And as noted earlier, the data and models are being presented in absolute form. The use of sea surface temperatures instead of anomalies helps to illustrate addition problems with the models.
WHY ABSOLUTE TEMPERATURES?
The actual ocean surface temperatures, along with numerous other factors, dictate how much moisture is evaporated from the ocean surfaces, and, in turn, how much moisture there is in the atmosphere…which impacts the moisture available (1) for precipitation, (2) for water vapor-related greenhouse effect, and (3) for the negative feedbacks from cloud cover. In other words, failing to simulate sea surface temperatures properly means the atmospheric components of coupled ocean-atmosphere models are also flawed.
The sea surface temperature dataset being used in this post is the weekly and monthly version of NOAA’s Optimum Interpolation (OI) Sea Surface Temperature (SST) v2 (a.k.a. Reynolds OI.v2). We’re using it because (1) it is the longest running satellite-enhanced sea surface temperature dataset available, (2) its satellite data are bias adjusted based on temperature measurements from ship inlets and from buoys (both moored and drifting), and (3), most importantly, NOAA’s Reynolds OI.v2 data have been called “a good estimate of the truth”. See Smith and Reynolds (2004) Improved Extended Reconstruction of SST (1854-1997). The authors stated about the Reynolds OI.v2 data (my boldface):
Although the NOAA OI analysis contains some noise due to its use of different data types and bias corrections for satellite data, it is dominated by satellite data and gives a good estimate of the truth.
This is not the (over-inflated, out-of-the-ballpark, extremely high warming rate) high-resolution, daily version of NOAA’s Reynolds OI.v2 data, which we illustrated and discussed in the recent post On the Monumental Differences in Warming Rates between Global Sea Surface Temperature Datasets during the NOAA-Picked Global-Warming Hiatus Period of 2000 to 2014. Based on that post, we also are not using the NOAA ERSST.v4 “pause buster” data, which makes up the ocean portion of the NOAA and GISS combined land+ocean surface temperature products. NOAA has adjusted the numerous parameters in their ERSST.v4 model so that the warming rates for the NOAA-selected periods of 1951-2012 and 2000-2014 are at the extreme high ends of the trend uncertainty ranges for those periods. And as we illustrated and discussed in that post, the original (weekly/monthly, 1-deg resolution) version of the Reynolds OI.v2 data we’re presenting herein has, basically, the same warming rate as the UKMO HADSST3 data for the NOAA-selected hiatus period of 2000-2014.
The climate models presented in this post are those stored in the CMIP5 archive, which was used by the IPCC for their 5th Assessment Report (AR5). The CMIP5 climate model outputs of sea surface temperature are available through the KNMI Climate Explorer, specifically through their Monthly CMIP5 scenario runs webpage, under the heading of Ocean, ice and upper air variables. Sea surface temperature is identified as “TOS” (temperature ocean surface). For consistency with past posts, the “CMIP5 mean” and the Historic/RCP6.0 scenario are being used. The RCP6.0 scenario is the closest to scenario A1B used for the CMIP3 models (used by the IPCC for the 4th Assessment Report). And once again we’re using the model mean because it represents the forced component of the climate models. Basically, the model mean represents the consensus of the climate modeling groups for how the surfaces of the oceans should warm if they were warmed by manmade greenhouse gases and the other factors used to drive the models. For a further discussion, see the post On the Use of the Multi-Model Mean.
Because the model-data comparisons are in absolute terms, annual data are being presented. The Reynolds OI.v2 data start in November 1981 so the comparisons run from 1982 to 2015…34 years.
The linear trends in the graphs are as calculated by EXCEL.
Figure 1 presents two model-data comparisons for global satellite-era sea surface temperatures, not anomalies. The top graph for the latitudes of 90S-90N includes the polar oceans. Excluding the polar oceans, the latitudes of 60S-60N are shown in the bottom graph. It makes little difference whether we include or exclude the polar oceans in the model-data comparison. The disparity between the models and data is slightly greater if the polar oceans are excluded, but, bottom line, the models show too much warming.
The differences between models and data on a global basis highlight a very basic problem with the models. The modelers have to rely solely on manmade greenhouse gases and other anthropogenic factors to drive the climate models, because climate models do not properly simulate the timing, magnitude and duration of naturally occurring processes that can cause ocean surfaces to warm. Based on those models (that bear no similarity to the real world), the surfaces of the global oceans were expected to warm at a rate of roughly 0.17 deg C/decade for the past 34 years, but in the real world, the observed global warming rates due to the combination of natural variability and greenhouse gases was only about 0.10 deg C/decade. That’s a monumental difference between hypothesis and reality.
Another fact to consider: the data existed while the modelers were preparing their climate models for the CMIP5 archive. In other words, the modelers knew their goals. And the best the modelers could do was almost double the warming rate of the surface of the global oceans. That’s an atrocious modeling effort…even with the recent upticks in global sea surface temperatures attributable primarily to The Blob in 2014 and to the 2015/16 El Niño in 2015.
Also note that the climate model-simulated ocean surfaces are slightly warmer than observed. For the latitudes of 60S-60N, the average model-simulated ocean surface temperatures for the period of 1982 to 2015 are about 0.2 deg C higher than observed. Maybe that’s why climate models simulate too much precipitation globally. See Figure 2, which is Figure Intro-10 from my ebook On Global Warming and the Illusion of Control.
TROPICAL SEA SURFACE TEMPERATURES
The models overestimated the warming of the sea surface temperatures of the tropical oceans (24S-24N) over the past 34 years…no surprise there. The surprise is how similar the modeled and observed sea surface temperatures are in the tropics. See Figure 3. For the tropical oceans, the average model-simulated ocean surface temperatures for the period of 1982 to 2015 are only about 0.02 deg C higher than observed.
While the absolute temperatures are very similar, the models overestimated the warming by a wide margin. The data indicate the sea surface temperatures of the tropical oceans warmed at a not-very-alarming rate of about +0.08 deg C/decade, while the models indicate that, if the surfaces of the tropical oceans were warmed by manmade greenhouse gases, they should have warmed at about +0.19 deg C/decade. Now consider that the tropical oceans (24S-24N) cover 76% of the tropics and about 46% of the global oceans.
That disparity between models and data supports the June 2013 model-data comparisons of the warming of the tropical mid-troposphere prepared by John Christy. See Roy Spencer’s posts EPIC FAIL: 73 Climate Models vs. Observations for Tropical Tropospheric Temperature and STILL Epic Fail: 73 Climate Models vs. Measurements, Running 5-Year Means. The models grossly overestimated the warming rates of the mid-troposphere in the tropics.
Let’s break down the tropical oceans. As shown in Figure 4, the difference between the simulated and observed warming rates for the tropical Atlantic and Indian Oceans (24S-24N, 80W-120E) is about 0.08 deg C/decade, with the models showing too much warming of course.
But for the tropical Pacific (24S-24N, 120E-80W), Figure 5, the models are showing warming at a rate that’s almost 0.14 deg C/decade higher than observed.
El Niño and La Niña events have their greatest direct impacts on the Eastern Tropical Pacific. During an El Niño, warm surface and subsurface waters from the western tropical Pacific flood into the eastern tropical Pacific along the equator, with the warmer-than-normal subsurface waters rising to the surface. During ENSO neutral (not El Niño or La Niña) periods and during La Niña events, cool waters from below the surface are upwelled to the surface. Apparently, climate modelers haven’t grasped even the most subtle aspects of the tropical Pacific. While the data for the Eastern Tropical Pacific (24S-24N, 180-80W) show no surface warming over the past 34 years, the consensus (groupthink) of the models indicate the surfaces there should have warmed at a rate of 0.18 deg C/decade. See Figure 6.
EXTRATROPICAL SEA SURFACE TEMPERATURES
For the extratropical oceans of the Southern Hemisphere (90S-24S), Figure 7, the observed warming rate is also extremely low at about 0.06 deg C/decade. On the other hand, the climate models indicate that if manmade greenhouse gases were responsible for the warming of the sea surfaces in this region, the oceans should have warmed at a rate of 0.12 deg C/decade, effectively, doubling that observed trend. Now consider that the extratropical oceans of the Southern Hemisphere cover about 33% of the surface of the global oceans (about 23% of the surface of the planet).
On the other hand, the climate models seem to get the warming rate of sea surfaces just about right for the smallest portion of the global oceans, the extratropical Northern Hemisphere (24N-90N). See Figure 8. The extratropical oceans of the Northern Hemisphere cover only about 21% of the surface of the global oceans (about 15% of the surface of the Earth).
Unfortunately for the modelers, they underestimated the warming of the surface of the extratropical North Atlantic (Figure 9), and overestimated the warming in the extratropical North Pacific (Figure 10), so they accomplished the overall warming of the extratropical Northern Hemisphere incorrectly. No surprise there.
# # # # # # #
Now note the similarities between the modeled and observed surface temperatures (not trends) of the tropical oceans (Figure 9). The average observed tropical sea surface temperatures for the period of 1982 to 2015 are only 0.02 deg C cooler than the models. Then there are the differences between the models and observations in the extratropics of the Southern (Figure 7) and Northern (Figure 8) Hemispheres. For the period of 1982 to 2015, the modeled sea surface temperatures in the Southern Hemisphere are too warm by almost 0.8 deg C, and in the Northern Hemisphere, the modeled sea surface temperatures are too cool by 0.4 deg C. Those are further indications that ocean circulations in climate models are flawed.
Those disparities in absolute temperatures may help to explain why climate modelers cannot simulate the sea ice loss in the Northern Hemisphere and the sea ice gains in the Southern Hemisphere. See Figures 11 and 12, which compare modeled and observed sea ice area for the Northern and Southern Hemispheres. They are Figures Intro-8 and Intro-9 from my ebook On Global Warming and the Illusion of Control.
# # # # # # #
INDIVIDUAL OCEAN BASINS
The following are model-data comparisons for the individual ocean basins and their hemispheric subsets where appropriate…without commentary.
# # # # # # #
# # # # # # #
# # # # # # #
# # # # # # #
# # # # # # #
# # # # # # #
# # # # # # #
# # # # # # #
ATLANTIC MULTIDECADAL OSCILLATION
The Atlantic Multidecadal Oscillation is often described as the variations in the surface temperatures of the North Atlantic that take place over periods of 50 to 80 years. In this light, the Atlantic Multidecadal Oscillation is typically displayed by detrending the long-term sea surface temperature anomalies of the North Atlantic. The argument against the detrending is that it doesn’t consider “global warming”.
If you’re not familiar with the Atlantic Multidecadal Oscillation see the NOAA Frequently Asked Questions About the Atlantic Multidecadal Oscillation (AMO) webpage and the posts:
- An Introduction To ENSO, AMO, and PDO — Part 2
- Multidecadal Variations and Sea Surface Temperature Reconstructions
Over some multidecadal periods, the sea surfaces of the North Atlantic warm faster than the rest of the global oceans, and during other multidecadal periods, the North Atlantic sea surfaces warm slower than the rest of the global oceans and can also cool dramatically. For much of the period of November 1981 to December 2014, the sea surface temperatures of the North Atlantic were on one of their naturally occurring upswings.
As noted earlier, detrending the North Atlantic data does not account for the warming of the global oceans. Therefore, a better description of the Atlantic Multidecadal Oscillation is the additional long-term variations in the sea surface temperatures of the North Atlantic beyond those of the global oceans. To this end, to portray the Atlantic Multidecadal Oscillation, Trenberth and Shea (2006) subtracted global sea surface temperature anomalies (60S-60N), from sea surface temperature anomalies of the North Atlantic (0-60N, 80W-0). That method accounts for the additional variability of the North Atlantic data above the variations in the global data. But, for this post, we’re adding a minor twist to that method. We’re presenting the difference between the global and North Atlantic data in absolute terms, not anomalies.
Trenberth and Shea (2006) used HadISST data, so I used that dataset for this discussion in past updates. See Figure 22, which is Figure 25 from the post Alarmists Bizarrely Claim “Just what AGW predicts” about the Record High Global Sea Surface Temperatures in 2014. It displays the modeled and observed differences between the North Atlantic and global sea surface temperatures for the period of 1870 to 2013, a.k.a. the Trenberh and Shea (2006) portrayal of the Atlantic Multidecadal Oscillation.
Obviously, the data show that there are noticeable multidecadal variations in the sea surface temperatures of the North Atlantic that are beyond those seen in global data. That of course indicates that the natural variations in the surface temperatures of the North Atlantic can enhance or suppress global warming. Or as NOAA ended their FAQ webpage about the Atlantic Multidecadal Oscillation:
In the 20th century, the climate swings of the AMO have alternately camouflaged and exaggerated the effects of global warming, and made attribution of global warming more difficult to ascertain.
Unfortunately, the model mean of the climate models does not show that additional variability in the sea surface temperatures of the North Atlantic, indicating the Atlantic Multidecadal Oscillation is forced by the factors (primarily manmade greenhouse gases) that create global warming in climate models.
So let’s update that graph using the Reynolds OI.v2 data, which limits our view to the past 34 years. First the components:
The top graph in Figure 23 shows the modeled sea surface temperatures for the global oceans (60S-60N) and for the North Atlantic (0-60N, 80W-0) for the period of 1982 to 2015. The modeled trend of the North Atlantic is very similar to the warming rate for the global oceans, only about a 0.02 deg C/decade difference. The bottom graph shows the observed North Atlantic and global sea surface temperatures. The data indicate the warming rate of the North Atlantic is almost 0.1 deg C/decade higher than that of the global oceans.
Figure 24 includes the model-data comparison of the satellite-era Atlantic Multidecadal Oscillation, using the method presented by Trenberth and Shea (2006), except that we’re dealing in absolute temperatures. According to the model mean (which represents the consensus or groupthink of the modeling groups), the sea surfaces of the North Atlantic should only have warmed at a rate that was slightly higher than the global oceans, but the data indicate the surface of the North Atlantic is capable of warming at a rate that’s almost 0.1 deg C/decade higher than the global oceans without being forced by the factors (primarily greenhouse gases) that drive the models.
The other problem shown in Figure 24: the difference between the surface temperatures of the North Atlantic and the global oceans is too small…once again indicating that the ocean circulation in the models is flawed.
We live on an ocean-covered planet. One might have thought that one of the climate modelers’ first priorities would have been to simulate the processes that cause sea surface temperatures to vary on annual, decadal and multidecadal bases. Sadly, the modelers elected another route…they chose to create models of a planet that bear no relationship to the one where we live, no relationship at all.
Climate model simulations of sea surface temperatures are far from reality. That is, they’re modeling a virtual planet—a science-fiction planet—with no similarities to Earth. More specifically, as shown in this series of posts, the climate models used by the IPCC do not simulate (1) the actual warming and cooling rates of the ocean surfaces, or (2) the spatial patterns of those trends, or (3) the absolute temperatures. It would be nice if climate modeling agencies might try to simulate the surface temperatures of this planet, not some fairytale one. That way, their models might have some value. Right now, they don’t serve any purpose…other than to illustrate how poorly they simulate Earth’s climate.
Next in this series: ocean heat content.