Even if you assume a low value for the variables of SO2 cooling and the rate of heat uptake by the deep ocean. Yes, that caught our eyes too.

“Studies of these [volcanic eruptions and El Niño] effects using climate models have improved understanding of the climate system and increased confidence in projections of global warming from anthropogenic greenhouse gases” – Alan Robock, Rutgers University in Encyclopedia of Global Environmental Change (2002)

Volcanoes (especially explosive tropical volcanoes) are interesting, for many reasons. When they erupt tropical volcanoes emit all sorts of aerosols into the atmosphere, often penetrating the troposphere-stratosphere boundary. When this happens, large-scale temperature effects can take place due to increased albedo (solar reflection) and global dimming. What is also striking is that volcanic activity seems to sometimes come in clusters – making volcanoes very useful for climate model calibration. The historical record shows that tropical volcanoes had been relatively quiet for over 5 decades before a period of higher activity started somewhere around the 1960s. The 1980s showed slightly higher volcanic activity than the 1970s, which was mostly due to the eruption of El Chichón in Mexico in 1982 – a devastating event that killed 2,000 people, destroyed 9 villages and created a 1 kilometer wide and 300 meter deep crater. Prior to the eruption the volcano had kept quiet for some 600 years. El Chichón injected some 20 megatonnes of aerosols into the stratosphere, together with 7 megatonnes of sulphur dioxide (SO2 – a gas initially, reacting with water to form sulphate, an aerosol). In the above graph please note the logarithmic scale on the y axis. The Chichón eruption was of a similar magnitude on the northern hemisphere as the eruption of super volcano Mount Pinatubo in 1991 (blue line), but Pinatubo also emitted a large amount of aerosols into the southern hemispheric stratosphere (green line), which El Chichón did not. (Of course for global land temperature effects the northern hemisphere is of larger direct influence due to the position of the continents.) There are several well-established historical volcanic datasets at hand for climate model calculations, for instance Sato (1993) and Hammann (2003).

It’s a promise. Soon we’ll leave climate model studies for what they are – and start looking at far more basic paleoclimate comparisons, in our quest to better understand the ‘Real Global Temperature Trend‘.

But there’s still a couple of model studies we think deserve attention beforehand. This one is exceptional, performed by Chris Forest, Peter Stone and Andrei Sokolov, researchers of the Department of Earth, Atmospheric & Planetary Sciences of MIT, and published in Geophysical Research Letters in 2006.

When El Niño isn’t enough to explain a lack of observed cooling + a jump in warming…

The MIT research group tried to let their 2D statistical-dynamical climate model explain why global temperature datasets do not show major cooling of the Earth in the 1980s – when tropical volcanic activity was above average, and cooling sulphur dioxide (SO2) emissions to the stratosphere (where their cooling potency is greatest) where high.

On average, see for instance this NASA GISS graph, global average temperatures were not cooler in the 1980s than in the 1970s, but rather some 0.2 degrees Celsius higher.

The authors assume two factors to explain this lack of observed atmospheric cooling/relative jump in atmospheric warming. Firstly they assume a low figure for the (uncertain, debated) cooling potency of SO2 – a value of just -0.74 to -0.14 W/m2. (For comparison, this Nature study that investigated total aerosol forcing (including global dimming by non-SO2 aerosols) states a cooling figure of -4.4 W/m2 – a BIG difference.)

Then they also use ‘below climate model average’ deep-ocean-mixing values. This means less of the extra absorbed atmospheric energy gets transported to deeper waters. See image below for comparison to various other climate models.

Climate sensitivity versus ocean heat uptake for various climate models. White zone indicates MIT study results.

Decreased deep ocean heat uptake is of course a situation that mimics El Niño dominance. Indeed in 1983 a strong El Niño event happened, that may have compensated for much of El Chichón’s expected climate cooling.

Although 1984, 1985 and 1988 were (weak) La Niña years, the 1980s had a slight El Niño dominance – with two other El Niño events in 1986 and 1987. All in all this ‘low rate for deep ocean heat uptake’ does not let the MIT climate model explain the temperature developments. Something else must have happened in the atmosphere. Perhaps something the relatively high rise in atmospheric CO2 concentrations in the 1980s – over 1.6 ppm per year – could explain.

Back to the study. When you insert low aerosol cooling and low deep ocean heat uptake in your model, and you still don’t get to recreate the observed warming – you’re essentially left with one other option: increased atmospheric heat absorption.

“With additional new forcings, a larger climate sensitivity and a reduced rate of ocean heat uptake below the mixed layer are required to match the observed climate record in the 20th century. The primary factor leading to this change is the strong cooling forcing by volcanic eruptions through the stratospheric aerosols. Similarly, there is a small change in the aerosol forcing which tends to offset the volcanic cooling. When using uniform priors on all parameters, these new results are summarized by the 90% confidence bounds of 2.1 to 8.9 K for climate sensitivity,” the authors conclude.

Yes. You’ve read it. 2.1 to 8.9 degrees climate sensitivity.

The study is shown as the black line and box in the graph above. The 90 percent confidence interval for climate sensitivity lies between 2.1 and 8.9 degrees Celsius, with best estimate (the black diamond) above 4 degrees. This is high, compared to other studies.

© Rolf Schuttenhelm | www.bitsofscience.org