The media has been abuzz with claims that the January 2015 New England Blizzard was worsened by human-induced global warming. One of the outspoken activist members of the climate science community who has been quoted often on the storm is NCAR’s Kevin Trenberth. An example of Trenberth’s interviews can be found in the ClimateNexus post Blizzard of 2015: Normal Winter Weather, Amplified by Climate Change. The subtitle is actually quite funny, bringing back the old “consistent with climate models” nonsense: “Above average sea surface temperatures increase snowfall, consistent with model projections.”
Kevin Trenberth is reported to have written about the January 2015 New England blizzard:
The number 1 cause of this is that it is winter. In winter it is cold over the continent. But it is warm over the oceans and the contrast between the cold continent and the warm Gulf Stream and surrounding waters is increasing. At present sea surface temperatures are more than 2F above normal over huge expanses (1000 miles) off the east coast and water vapor in the atmosphere is about 10% higher as a result. About half of this can be attributed to climate change.
Interesting. Trenberth noted that only “about half” of the warming of the surface of the North Atlantic off of the east coast of the United States and only “about half” of the additional water vapor in the atmosphere there “can be attributed to climate change”. One has to assume Trenberth is referring to the human-induced type of climate change with that statement and that the other half was caused naturally—in response to the coupled ocean-atmosphere process of the Atlantic Multidecadal Oscillation.
I don’t think anyone will disagree with Trenberth’s opening sentence, which associates New England blizzards with winter. But let’s look at the number of ways the rest of Kevin Trenberth’s statement is incorrect. And we’ll also show the obvious flaws in the ClimateNexus author’s declaration “consistent with model projections”.
NOTE: For more information on the Atlantic Multidecadal Oscillation, refer to the NOAA Frequently Asked Questions About the Atlantic Multidecadal Oscillation (AMO) webpage and my blog posts:
- An Introduction To ENSO, AMO, and PDO — Part 2
- Multidecadal Variations and Sea Surface Temperature Reconstructions
That NOAA FAQ webpage confirms that the Atlantic Multidecadal Oscillation can contribute to global warming and suppress it. [End note.]
OVER THE PAST 27 YEARS, BOREAL WINTER WATER VAPOR HAS DECREASED OVER THE NORTHWEST ATLANTIC OCEAN
Dr. Roy Spencer addressed the water vapor portion of Trenberth’s statement in his post Northeast Snowstorms & Atlantic Water Vapor: No Connection in Last 27 Years. Roy Spencer found no correlation between the boreal winter (DJF) snowstorm index for Northeast United States and satellite-based water vapor data for the Northwest Atlantic (30N-50N, 80W-50W) since 1988. The data also indicate that, while water vapor in December has increased in 27 years based on the linear trend, it decreased in January and in February, so, overall, there has been a decrease in boreal winter (DJF) water vapor over the Northwest Atlantic.
Dr. Spencer was very clear that water vapor annually in that region has increased in that time, but during the boreal winter months since 1988, atmospheric water vapor has decreased. Kevin Trenberth’s speculations about water vapor must refer to annual, not winter data associated with blizzards.
BASED ON THE LINEAR TREND OF THE BOREAL WINTER DATA, NORTHWEST ATLANTIC SEA SURFACES SHOW NO WARMING IN 95 YEARS
Figure 1 illustrates the boreal winter (DJF) sea surface temperatures of the western extratropical North Atlantic, based on NOAA’s ERSST.v3b data. I’ve used the same coordinates as Dr. Spencer (30N-50N, 80W-50W). If you’re wondering about the size and location of that region, I’ve highlighted those coordinates on the map here, which was linked to the ClimateNexus post. (You’ll note on that map that the region with warmer-than-normal water actually stretches farther east than the 1000 miles noted by Trenberth, before we run into cooler-than-normal water. As a reference, at 40N, the longitudes of 80W-50W reach almost 1600 miles.) The first data point in Figure 1 is for the 3-month season of December 1880 through February 1881 and the last data point is for December 2013 through February 2014. As shown, the sea surfaces for the western extratropical North Atlantic were warmer for boreal winters (DJF) in the 1930s and 1950s than they have been in recent years. That, of course, is a response to the Atlantic Multidecadal Oscillation.
NOTE: There may be an uptick in 2014/15, but we’ll have to wait until early March to see how this winter compares. Adding one year of data, though, is not going to impact the following trend comparison to any great extent. [End note.]
So, working back in time in 5-year increments, I determined the longest time period where the boreal winter data for the western extratropical North Atlantic showed no surface warming, based on the linear trend. Turns out it was 95 years…since the boreal winter of 1919/20. See Figure 2.
I’ve also illustrated in Figure 2 the average of the sea surface temperatures for that region as simulated by the dozens of climate models included in the CMIP5 archive. Those are the models the IPCC used for their 5th Assessment Report. We present the model mean because it best represents the groupthink (the consensus) about how the sea surfaces of that region should have warmed if manmade greenhouse gases and other contributors (that drive the models) caused the warming. In other words, the average (the multi-model ensemble member mean) represents how the sea surfaces in that region should have warmed if they were warmed by manmade greenhouse gases. (Basically, the model mean represents the forced component, not the “noise”, of the models. See the post On the Use of the Multi-Model Mean for more information.) According to the climate models used by the IPCC, the boreal winter sea surfaces of the western extratropical North Atlantic should have warmed almost 0.6 deg C in those 95 years, but the data show no long-term warming in that time based on the linear trend. Note also that the modeled sea surfaces are too warm over the entire period, on average by more than 1.0 deg C.
The models are in no way “consistent with” reality.
MULTIDECADAL VARIATIONS IN SEA SURFACE TEMPERATURES THERE ARE NOT A FORCED COMPONENT OF THE MODELS
True-blue believers in the hypothesis of human-induced global warming say we need to look at 30-year trends, so let’s do that. See Figure 3. It shows the 30-year trends (in deg C/decade) for the boreal winter sea surface temperatures of the western extratropical North Atlantic. The first data point at 1910 illustrates the slight negative trend (cooling rate) of the boreal winter (DJF) sea surface temperatures for the 30-year period of 1880/81 to 1909/10, and the last data point in 2014 shows the boreal winter warming rate for the 30-year period of 1984/85 to 2013/14. Between the first and last data points, the sea surfaces of the western extratropical North Atlantic show a wide range of multidecadal variations in 30-year trends. Note how the (boreal winter) warming rate of the 30-year period ending in 1938/39 was more than 1.5 times faster than the warming rate of the most recent 30-year period ending in 2013/14.
But the vast majority of the warming during the 30-year period ending in 1938/39 was not forced by manmade greenhouse gases. We can show this by adding the modeled trends in sea surface temperatures for the western extratropical North Atlantic to the graph. See Figure 4. The observed warming rate of 0.37 deg C/decade for the 30-year period ending in 1938/39 was more than 5-times faster than the modeled rate of 0.07 deg C/decade. Obviously, the naturally occurring multidecadal variations in the warming and cooling rates of the surface of the western extratropical North Atlantic are not a response to the forcings used by the climate models. And that suggests that it is very likely that most of the warming there in recent decades was also a response to natural variability.
Once again, the models show they are in no way “consistent with” reality.
THE CONTRAST BETWEEN THE COOL LAND SURFACES AND WARM OCEANS IS DECREASING, NOT INCREASING
Many of you noticed the likely flaw in Trenberth’s early claim:
In winter it is cold over the continent. But it is warm over the oceans and the contrast between the cold continent and the warm Gulf Stream and surrounding waters is increasing.
Everyone knows that annual, decadal and multidecadal variations in land surface air temperatures are, in part, simply exaggerations of the variations in local sea surface temperatures. If the sea surfaces for the western extratropical North Atlantic warmed in recent decades, and they have, then the air over land should have warmed a little more. So Trenberth’s statement seems to contradict the instrument temperature record. But maybe he was correct for the boreal winter. Let’s look.
The big-3 global surface temperature data suppliers (GISS, NCDC and UKMO) only provide their data in anomaly form, but NOAA’s sea surface temperature data (used by GISS and NCDC) are available in absolute form. So for the land air temperatures in absolute form, we have to refer to a reanalysis to get an idea of the temperature difference between land and ocean during the boreal winter. The reanalysis is called GHCN-CAMS. It has to be kept in mind that a reanalysis is the output of a computer model that uses data as inputs, so, in other words, it’s not data. The authors of the paper that supports the GHCN-CAMS reanalysis also caution against using it for evaluating climate models. That’s fine. We’re not using it for that purpose. We’re simply using the GHCN-CAMS reanalysis outputs to get an idea of the magnitude of the temperature difference between the sea surfaces of western extratropical North Atlantic and the land surface air temperatures for the eastern U.S. and Canada, using the coordinates of 30N-50N, 80W-50W for ocean and land surface temperatures. See the top cell of Figure 5.
The GHCN-CAMS reanalysis starts in 1948, and since that time, there has been about a 20 deg C difference between boreal winter (DJF) land and sea surface temperatures (ocean minus land) for those coordinates. We know, however, that the sea surfaces of the western extratropical Atlantic were warmer in the 1950s than at present times, so the start date of the GHCN-CAMS reanalysis skews the trend results. Thus, the bottom cell shows the temperature difference between the sea and land surfaces (ocean minus land) for the coordinates of 30N-50N, 80W-50W, starting in 1975, which has been determined through breakpoint analysis to be the start year for the recent global warming period. The temperature difference between land and oceans in that part of the world is decreasing, not increasing as claimed by Kevin Trenberth.
Some readers may not feel confident with the GHCN-CAMS reanalysis. That trend difference does look a little steep. In Figure 6, for the same regions, we’re using the sea surface temperature anomalies and land surface air temperature anomalies based on the GISS Land-Ocean Temperature Index (LOTI). It includes the same ERSST.v3b sea surface temperature data used in this post, but they’re in anomaly form in the GISS LOTI. The differences (ocean minus land) also show land surface temperatures rising much faster than ocean surface temperatures in those regions since 1975, but not as fast as the with the GHCN-CAMS reanalysis.
And even climate models contradict Trenberth’s claim. They too show a decrease in the temperature difference for those ocean and land regions since 1975. See Figure 7.
Looks like the climate models got that relationship right since 1975. Too bad they can’t simulate sea surface temperatures over any time frame. Now recall that the oceans cover about 70% of this planet. For more insight into how poorly climate models simulate sea surface temperatures, see the posts:
- Alarmists Bizarrely Claim “Just what AGW predicts” about the Record High Global Sea Surface Temperatures in 2014
- IPCC Still Delusional about Carbon Dioxide
Apparently, alarmist climate-change advocates are still willing to furnish misinformation to the public about the contribution of human-induced global warming to weather events. Somehow, I don’t think many readers will find that surprising. It’s been the norm for many years.
The data, reanalysis and climate model outputs are available through the KNMI Climate Explorer.