Claim: "How can you trust climate models, they don't even include clouds!"

Why this Claim is Wrong: This was a stunner when I first saw it in Fall 2009 (I have been threatened if I divulge the source, but note the direct quotation marks). Climate models have included clouds for decades. It's a gross misrepresentation. As just one example from several years ago, see Hansen et al. 2005, which includes both low and high level clouds in its climate modelling, see also earlier references therein. That said, cloud modelling remains very difficult and until recently, progress in improving the models is slow. Cloud behavior in most models in the past has been done by parameterizations based on observations. This is a good approach when the complex physics doesn't have sufficient numerical resolution to model directly from first principles. It does not mean that the models are nonsense in what they produce. The evidence shows otherwise. Clouds are important to get right since about half of all insolation reflected back out to space is reflected by clouds. The IPCC AR4 identifies cloud modelling as one of the main contributers to the remaining spread in confidence about future temperature trends. But that spread is minor compared to the absolute predictions, which are - severe warming. The radiation physics of clouds is very well known. The absorption and emission of radiation from their bottoms and their tops is well determined, given the actual cloud. What is not yet possible is to generate realistic clouds in global climate models. The problem is that clouds are small, about ~1 km in typical size. Grid sizes in global climate models are of order 100 times larger. Modelling clouds on small scales is much easier, but still has a significant uncertainty in the amount of entrainment of dryer surrounding air as convection raises an airmass through an ambient airmass to create the cloud. The solutions so far have been to embed 1- and 2-dimensional cloud model results as parameterized solutions within larger GCM's, and include observed changes in clouds from data.

Since 2006 the situation has improved substantially, with the advent of the "A-Train" satellites observing coherently the same wide regions from space across a wide range of wavelengths in order to measure simultaneously aerosol content, spectroscopic identification of components, albedo, and droplet sizes from reflection properties. Published results show the pollution aerosols make for smaller water droplets and ice crystals and less rainfall from those clouds (Jiang et.al. GRL 35, L14804 (2008)). We've confirmed also what we've known since 1989; that large concentrations of sulfate aerosols (e.g. from fossil fuel burning) lead to smaller water droplets and brighter clouds (given the same water content), and that the smaller droplets inhibit precipitation and lead to longer lived clouds. So pollution adds a net cooling, on average. Since the industrialization of Asia has significantly net added to aerosols world wide, we expect a net cooling from this effect, yet clearly global temperature continue to rise despite this. We also see that cloud feedbacks are a positive feedback, meaning that CO2-induced warming alters cloud properties so that they accentuate the warming further. Cloud models have not all been equally good in getting clouds right. According to a study by Clement et.al. (2009) studying 55 years of COADS cloud data in the southeast Pacific, rising sea surface temperatures cause a decrease in low clouds and a further increase in low level temperatures. This positive feedback, if correct (and the work of Sherwood et al. 2013 certainly indicate that it is), accentuates global warming beyond that forecast by the IPCC's 2007 report. The UK Met office's HADGem1 climate model closely matched the behavior of low clouds very well, and also is showing positive feedback. And finally, this paper (Dessler 2011), shows fatal flaws in the claims of Spencer and Lindzen that cloud cover changes are causing global warming. In fact, models, observations and theory all agree well when done self-consistently. The cloud modelling uncertainties are nowhere near large enough to change the important conclusion - that man-made GHG's are causing the observed global warming.

The two most important ideas to understand - First: if clouds are changing because of the strong global warming we are seeing, then it is a "feedback". Averaged over the globe, that feedback will be either positive (amplify global warming), negative (reduce the rate of global warming) or be zero (have no effect on the global warming). Even if, optimistically and against the data and theory, one wanted to assume cloud changes were providing a negative feedback, then the magnitude of the effect would diminish as the warming itself was reduced. Thus, there is no way for supposed negative cloud feedbacks to REVERSE the warming and bring us back cooler, as we need to be. At best, they would only reduce the rate at which we warm. Second, it is clear that past large climate changes over thousands of year time scales show that clouds are clearly NOT providing a stabilizer to global warming. The Earth's climate has oscillated up and down by about 5C going into and out of Ice Ages, due to the Milankovich cycles. CO2 feedback from those small Milankovitch cycles amplified temperature changes by bringing CO2 into and out of the ocean. The forcing from current CO2 injection by mankind vs those in the Ice Ages can be seen for comparison here. The conclusion is clear; cloud feedbacks show no evidence of moderating or halting global warming; quite the opposite. In fact, the Ice Age forcing vs. today show we can expect climate to get much hotter from here.

Return to Climate Denial Claims List

Return to Climate Science Main Page