Climate Feedback

Atmospheric aerosols: correlation is not causation

Posted by Olive Heffernan on behalf of Kevin E. Trenberth

Aerosols in the atmosphere have been hot topics in several recent climate studies but one wonders if the pollution has made not only the atmosphere murkier but also the scientific reasoning?

In March a widely reported study in PNAS by Zhang et al. linked changes in storm tracks over the North Pacific to Asian pollution. In this case effects on radiative forcing by atmospheric aerosols supposedly increased deep convective clouds over the Pacific Ocean in winter, a finding based on long-term satellite cloud measurements (1984–2005). The blame was assigned to the aerosol effect from Asian pollution, which supposedly leads to intensified storms. Unfortunately the authors did not seem to be aware of the major problems in the satellite-based cloud data that give spurious trends owing to changes in satellites and associated instruments. In particular there were major changes in 1994 in all 3 geostationary satellites (GOES West, GOES East and GMS (Japanese)) observing the Pacific Ocean. All are known to be associated with spurious changes in cloud present in the International Satellite Cloud Climatology Project (ISCCP) data. The ISCCP record is simply not reliable for trends. The time series of the paper are perhaps more a measure of the problems with the data than they are of climate change or evidence of effects of aerosols?

Another recent study by Lau and Kim in the February 27 issue of EOS claims that nature foiled the 2006 hurricane season by increasing aerosols over the tropical North Atlantic in the summer of 2006 (see also Wu 2007 for more general analysis). There is no doubt that sea surface temperatures were lower in 2006 than in the record breaking year of 2005 and that aerosols were more plentiful. It is also clear that there were fewer tropical storms in 2006. But is the former the cause of the latter, as inferred in the article? Or is it more likely that the absence of storms led to less rain and thus less washout of aerosols leading to the increase in atmospheric aerosols? Of course, even in the latter case, there may be a feedback, but estimates of the aerosol effects places them much smaller than effects from changes in cloud and sunshine and from evaporative cooling as winds change.

[Note the contrasting findings between the above two studies with one finding more convection and the other finding less, all due to aerosols!]

Now someone else thinks they can use aerosols to change hurricanes, (as reported by Joshua Zaffos in the Rocky Mountain Chronicle in June 2007). Bill Cotton, a professor from Colorado State University, thinks that dumping dust into a hurricane could be used to weaken the storm. The irony in this case is that Cotton is a skeptic on the issue that humans are changing the climate in spite of overwhelming evidence. Human changes in the greenhouse effect amount to about a 1% change in the energy flow through the climate system. But the energy in a hurricane is enormous; it is typically equivalent in energy to a 1 megatonne nuclear bomb going off every 5 seconds, and Cotton thinks he can mess with a hurricane! What if he is actually right and instead it strengthens the storm?


  1. Report this comment

    Amato Evan said:

    I published a little paper in GRL last Fall about the correlation between dust and TCs (being very careful to never imply that we had uncoverd a causal link). So, just a few weeks ago, someone from the Department of Homeland Security called us, basically offering funding to study TC modification by way of African dust!!! I can’t help but think that this money could be so much better spent in helping people/cities to be better preparred for a hurricane!

    On another note, you may find interesting my comment on the Lau and Kim paper, that was published in EOS this week (



  2. Report this comment

    sureshkumar.s said:

    Higher sea surface mean temperatures should naturally mean higher evaporation and greater moisture content in the atmosphere which on condensation on aerosols and other condensation nucleii should form increased cloud coverage, part of which comes down and part reflects back solar radiation ,thus countering the warming effect due to solar forcing.

    Also the presence of dust and aerosols can heat up the upper atmosphere causing condensation nucleii to migrate upward due to convection, causing cloud formations ,which dissipate when the aerosols themselves get heated up by absorbing radiation.

    Thus we see two balancing counterforces in operation that drives global warming and cooling simultaneously causing wide fluctuations in precipitation and storms.

    There is also the possibility for intense upward convection currents being set up by the aerosols and dust particles due to natural solar forcing causing more water vapour being sucked up from large terrestrial water body sources like oceans driving the storm and precipitation potential or disturbing the hydrological cycle causing huge wave and wind phenomena in regions near the tropics,perhaps even simulating giant tsunami like wave phenomena in and around the mid tropics

Comments are closed.