Guest posting from Beate Liepert (LDEO)
On April 18th PBS will air the NOVA documentary “Dimming the Sun” which stirred up lively discussions among scientists and non-scientists when originally shown by BBC in the UK (under the name ‘Global Dimming’ – see our previous posts). [The NOVA version has been thoroughly re-edited and some of the more controversial claims have apparently been excised or better put into context [and we look forward to seeing it! – Ed.].
Global dimming is the phenomena of an observed reduction (about 1-2% per decade since ~1960) of sunlight reaching the surface of the Earth caused by air pollution (aerosols – small particles) and cloud changes. Some of this solar energy is reflected back out to space and this cooling effect is believed to have counteracted part of the greenhouse gas warming. The original version of the film focused mainly on the observational recognition of global dimming, but one aspect did not receive much attention in the film – namely the oft-claimed lack of global dimming in climate models. This led some to assume that climate modelers were ignoring air pollution other than greenhouse gases emissions from fossil fuel burning. Another implication was that climate models are not capable of adequately simulating the transfer of sunlight through the atmosphere and the role of clouds, sunlight extinction of aerosols and aerosol effects on clouds etc, and therefore model projections should not be trusted. The NOVA version will address this issue more prominently by adding an interview with Jim Hansen from NASA Goddard Institute for Space Studies. Along this line, I’d like to elaborate on aerosols in climate models in more detail.
It is indeed true that the first climate change simulations were predominantly concerned with greenhouse gas forced climate change. Albeit very early papers argue that man-made aerosol changes might cause a cooling (Rasool and Schneider, Science 1971 and Bryson, Science 1974). And already in 1990, J. Hansen and A. Lacis (Science 1990) published a paper where they explicitly discuss the importance of anthropogenic aerosol forcing: “Sun and Dust versus Greenhouse Gas Forcing”. The authors list direct effects of increasing concentrations of light scattering sulfate aerosols, light absorbing carbonaceous aerosols like soot and even aerosol effects on cloud properties (indirect aerosol effects). Finally they conclude that “… solar variability will not counteract greenhouse warming and that future observations will need to be made to quantify the role of tropospheric aerosols … ”. The surface dimming effect was not yet considered an important climate factor. Back then, state-of-the-art climate models changed the reflectivity at the top of the atmosphere to account for the climate effect of increases in man-made aerosol emissions. Using cloud properties from independent climate simulations and weather forecast models to provide monthly mean water vapor and temperature fields, Kiehl and Briegleb (Science 1993) estimated a top of the atmosphere global mean human-related sulfate aerosol forcing of -0.3 W/m2 in contrast to the a +2.1 W/m2 greenhouse gas forcing. A year later, Jones, Roberts and Slingo (Nature 1994) added the indirect aerosol effect – the impact of increasing sulfate aerosol concentrations on cloud droplet sizes – which make look cloud darker. They used empirical relations to link the number of aerosol particles and number of cloud droplets to cloud droplet radii for their estimates. Other groups as well, started testing new prognostic cloud schemes for general circulation models that were able to capture the microphysical processes of cloud formation (e.g. MPI in Hamburg – Lohmann and Roeckner, Climate Dynamics 1996). These climate-modeling developments were compared (Wild et al., JC 1995) with the then available observational data of the surface solar radiation (incidentally the same data sets were used by Russak (Tellus 1990), Stanhill and Moreshet (Climatic Change 1992) and myself (Liepert et al., Contr. Atm. Physics 1994) to reconstruct the history of global dimming). One conclusion was that then-current models did not include enough aerosol absorbtion in the atmosphere (and Wild and I (GRL, 1998) wrote a paper on the “Excessive Transmission of Solar Radiation Through the Cloud-free Atmosphere in GCMs”). Note that at the end of 1990s these more complex climate models with a more physically based prognostic cloud scheme were run as equilibrium experiments hence transient 20th Century changes could not be used directly for comparison. But is has always been clear that anthropogenic aerosols are so temporally and spatially variable that long-term means are not adequate in assessing the actual aerosol forcing.
Several publications on model validations and improvements based on surface solar radiation records followed and I was involved in two of these studies. We analyzed the simulated multi-decadal changes in the direct tropospheric aerosol forcing in the NASA GISS GCM and utilized global dimming time series of the United States and Germany to assess the temporal change prognosed in the model (Liepert and Tegen, JGR 2002). Ina Tegen’s aerosol model was one of the first that added time variations in carbonaceous aerosol components (including black carbon – an absorbing aerosol). We concluded that increasing absorption might actually play a stronger role than expected.
Climate simulations are the primary tools for explaining and understanding observations that might otherwise seem counterintuitive. For instance, how global dimming can go hand in hand with global warming. In 2004, at the MPI in Hamburg my coauthors and I (Liepert et al. GRL 2004) analyzed output data from a brand new version of the ECHAM general circulation model (GCM) that incorporated a fully interactive aerosol module and aerosol-cloud-scheme. This model interactively calculates aerosol chemical transformation, aerosol transport, rainout and fallout processes and even aerosol formation for some species (e.g. sulfate). We showed that, in the model, global warming caused changing rainfall patterns that fed back on aerosol distribution and composition supressing the water cycle (i.e. evaporation) as had been observed. The key to explaining the apparent contradiction was that the surface forcing changes can be very large without affecting the top-of-the-atmosphere radiation as much.
All major climate models now have some representation of aerosol physics though they range in their complexity – e.g. from top of the atmosphere aerosol forcing to highly interactive aerosol-cloud modules. The role that aerosols play in issues like the Sahel drought (Rotstayn and Lohmann, JC, 2002) or the Asian Brown Clouds (Ramanathan et al., PNAS 2005) is starting to be understood (and both these examples are featured in the documentary), but we do not as yet have a clear picture of exactly how aerosols and the other human-related forcings have affected climate.
More recently, many modeling groups ran 20th Century climate simulations in support of the Intergovernmental Panel on Climatic Change 4th Assessment Report (IPCC AR4) that include representations of the aerosol direct, indirect and semi-direct aerosol effects. The models show a global dimming effect of between 1 to 4 W/m2 over the 100 years with simultaneous global warming between 0.4 and 0.7°C (Romanou et al, under revision) which match the observational dimming quite well.
Overall, in the fifteen years from the 1990 Hansen and Lacis paper to the IPCC AR4, major steps forward have been made in implementing aerosols in climate models and hence matching observations of global dimming. However, it would be misleading to claim that the new appreciation for the surface energy balance changes implied that modelers a few years back were ignorant about the role of aerosols in other aspects of climate change. It is indeed a very complex problem.