In our last World Climate Report article, we detailed a recent paper that showed that climate models which fail to account for the evolution of stratospheric aerosols (that is, reflective particles in the earth’s upper atmospheric) during the past decade or two project less warming than they would have had they included the influence of stratospheric aerosols in their calculations. This means that the discrepancy between the observed warming trend during the past 10-15 years (which is near zero) and climate model projections should be even larger than it appears (and it is already quite large).
Now comes along a new paper which hints at another reason why the climate models should actually be projecting more warming than they currently do—again, meaning that the models are faring even worse than it appears.
The new paper is by Stephen Hudson of the Norwegian Science Institute, is soon to be published in the Journal of Geophysical Research, and deals with the impact of Arctic ice loss on global temperatures.
Hudson points out that this topic is significant because it seems to be a pretty straightforward demonstration of a positive feedback resulting from global warming, going something like this: atmospheric carbon dioxide increases lead to higher temperatures which lead to the melting of the highly reflective Arctic sea ice which leads to more sunlight being absorbed which leads to higher temperatures, and so on.
In fact, this example is a favorite of Al Gore. Hudson explains:
In general introductions to the topic of climate change, the sea ice–albedo feedback (SIAF) is often singled out for use in explaining the concept of climate feedbacks (e.g., it is the only feedback mentioned by Gore ), something that can give the impression to the interested public that it is the most important feedback process, while its popularity likely stems from the relative ease with which it can be explained and grasped.
Hudson’s intent is to find out just how much of an influence Arctic sea ice really has on the earth’s temperature (or at least its radiative effect, which is directly related to temperature). And he is intent on doing so without overly involving climate models. He describes his motivation:
This study focuses directly on the changes in the amount of solar radiation absorbed by Earth due to the loss of Arctic sea ice. It focuses only on the Arctic because that is where significant changes have been observed in recent decades. The estimates here are based mostly on observations, rather than on the results of climate models. Furthermore, they are kept relatively simple, to make the uncertainties and assumptions that go into the calculation of the increased absorption of solar radiation as clear as possible.
Included in Hudson’s calculations are the observed sea ice values from 1979 through 2007, the climatological cloud cover over the sea ice regions, various characteristics of ice and cloud reflectivity, and the changing angle of the sun over the seasons. Using these factors, Hudson calculated the total global radiative forcing anomaly for each year based on the amount of sea ice that year (Figure 1). The radiative forcing anomaly is basically the change in radiation that is absorbed at the surface and goes into heating the earth—positive forcing anomalies indicate a tendency towards higher temperatures. Plotted along side the radiative forcing anomaly iin Figure 1 is the observed Arctic sea ice extent in September of each year from 1979 to 2007 (the left-hand axis flipped so less ice is upwards), which shows that September sea ice is a pretty good indicator of the radiative forcing changes—the less September sea ice cover in the Arctic, the greater the radiative forcing anomaly and the greater the pressure imparted to raise the earth’s average temperature.
Figure 1. The top series (line with dots) shows the yearly anomalies in globally and annually averaged radiative forcing caused by that year’s anomalies in Arctic sea ice area. All anomalies are from the 1979–1998 twenty year mean. The horizontal lines give 5 year averages of the radiative forcing anomalies (4 years for the last one because 2008 ice concentration data were not yet finalized). The bottom series, plotted against the right‐hand axis, is the mean Arctic sea ice extent in September of each year, with the scale inverted (from Hudson, 2011).
In order to convert radiative forcing changes into actual global temperature changes, we need to use some value for the “climate sensitivity”—that is, how much the earth’s average surface temperature changes for a given forcing change. In general, the climate sensitivity that climate models spit out is about 0.75°C per Watts per square meter. We, along with a growing number others, think that climate models overestimate the climate sensitivity, and that in fact, it is probably less than half of this value. But for the sake of this article in which our goal is to assess climate model performance, we will use the model estimate of climate sensitivity. So to convert the radiative forcing changes in Figure 1 to global temperature changes, we multiply by 0.75°C/W/m2. In Figure 2, we show this result (updated with our best guess at the values through 2011) and add a trend line through the temperature change. The magnitude of this trend, which represents the rate of global warming that climate models would likely project from the decline in Arctic sea ice is 0.034°C/decade (from 1979-2011). If we only look at the last 10-15 years of the record, a period of time in which sea ice decline has hastened, the trend increases to about 0.06°C/decade.
Figure 2. The likely change in global temperature that climate models would attribute to the change in radiative forcing resulting from Arctic sea ice declines as indicated in Figure 1 (assuming a climate sensitivity of 0.75°C/W/m2). The filled circles are calculated directly from the radiative forcings given in Hudson (2011), the open circles are our extension of that data through 2011 (using the least squares regression fit between September sea ice extent and radiative forcing anomalies from Hudson (2011).
So what does this have to do with climate models and their projections?
We are glad you ask!
A couple of years ago, Julienne Stroeve and colleagues from the National Snow and Ice Data Center analyzed the rate of Arctic sea ice loss projected by a host of climate models used by the IPCC and then compared the model projections against the observations. In Figure 3, we reproduce the major finding from Stroeve et al.’s study (including updating the September sea ice extents through (an estimated value for) 2011). Figure 3 demonstrates that the observed sea ice is declining at a rate about twice as fast as climate models projected. A fairly larger body of scientific work has been focused on explaining why this is the case, with most studies concluding that natural variability (in such factors as wind and ocean circulation patterns) has been responsible for the extra amount of ice loss over the past several decades.
Figure 3. The decline in Arctic sea ice as projected by climate models used in the IPCC (thin dark lines), compared with observations (thick, red line) (updated from Stroeve et al., 2007).
But regardless of the cause, it means than only about 50% of the warming from the loss of Arctic sea ice has been correctly captured by the climate models.
In other words, over the past 10-15 years, about 0.03°C/decade of warming has been missed by the models. Had they correctly captured it, the model-projected warming rate for the past decade or more would have been even greater than it is now, and the resulting discrepancy between model projections and the observed rate of global warming would have been even larger than it is currently.
In other words, the models are doing worse than we realize.
Currently, a bevy of researchers are furiously scrambling to make excuses why the models aren’t working so well by pointing out potential influences such as a slight decline in solar radiation (Lean and Rind, 2007), and a decrease in upper atmospheric water vapor (Solomon et al., 2010) which may be acting to impart a cooling pressure on surface temperatures and thus offset some warming from increasing greenhouse gas concentrations. Others are contending that the observed lack of warming is perfectly consistent with model projections (e.g. Santer et al., 2011).
But there are few researchers who are pointing out that the models are missing some influences that are acting to raise the rate of warming over the past decade or two—to a rate over and beyond whatever warming may be the result of increasing greenhouse gas concentrations.
In our two most recent World Climate Report articles, we highlighted two different mechanisms that are each contributing warming over and beyond what is included in the climate models. It is interesting to note, that neither of the two research teams made note of this fact themselves. It was only your obedient servants here a WCR that have sleuthed out the true implications of these results (in fact, some of the original researchers have even tried to spin their results as being “pro” model performance).
The reason that all of this is important is that climate models which produce too much warming quite possibility are doing so because they are missing important processes which act to counteract the warming pressure exerted by increasing greenhouse gas concentrations—in other words, the climate sensitivity produced by the climate models is quite possibly too high.
If this proves to be the case, it means that there will be less future warming (and consequently less “climate disruption”) as greenhouse gas emissions continue to increase as a result of our use of fossil fuels.
Evidence continues to mount that this is indeed the case.
Stay tuned to these pages to keep track of the latest developments in this important topic.
Hudson, S.R., 2011. Estimating the global radiative impact of the sea ice-albedo feedback in the Arctic. Journal of Geophysical Research, 116, D16102, doi:10.1029/2011JD015804.
Lean, J., and D. Rind, 2008. How natural and anthropogenic influences alter global and regional surface temperatures: 1889 to 2006. Geophysical Research Letters, 35, L18701, doi:10.1029/2008GL034864.
September sea ice extent data, ftp://sidads.colorado.edu/DATASETS/NOAA/G02135/Sep/N_09_area.txt
Santer, B. D., et al., 2011. Separating Signal and Noise in Atmospheric Temperature Changes: The Importance of Timescale. Journal of Geophysical Research, in press.
Solomon, S., et al., 2010. Contributions of Stratospheric Water Vapor to Decadal Changes in the Rate of Global Warming. Science, 327, 1219-1223.
Stroeve, J., et al., 2007. Arctic sea ice decline: Faster than forecast. Geophysical Research Letters, 34, L09501, doi:10.1029/2007GL029703.