Posts Tagged ‘clouds’

Guest post at Realclimate on aerosol nucleation and climate

April 24, 2009

Guess it’s a little late notice, but I have a guest post at RealClimate on the potential effects of aerosol nucleation and cosmic rays on climate. For the whole article please see RealClimate. The bottom line is as follows.


Freshly nucleated particles have to grow by about a factor of 100,000 in mass before they can effectively scatter solar radiation or be activated into a cloud droplet (and thus affect climate). They have about 1-2 weeks to do this (the average residence time in the atmosphere), but a large fraction will be scavenged by bigger particles beforehand. What fraction of nucleated particles survives to then interact with the radiative budget depends on many factors, notably the amount of condensable vapor (leading to growth of the new particles) and the amount of pre-existing particles (acting as a sink for the vapor as well as for the small particles). Model-based estimates of the effect of boundary layer nucleation on the concentration of cloud condensation nuclei (CCN) range between 3 and 20%. However, our knowledge of nucleation rates is still severely limited, which hampers an accurate assessment of its potential climate effects. Likewise, the potential effects of galactic cosmic rays (GCR) can only be very crudely estimated. A recent study found that a change in GCR intensity, as is typically observed over an 11 year solar cycle, could, at maximum, cause a change of 0.1% in the number of CCN. This is likely to be far too small to make noticeable changes in cloud properties.

Aerosols, clouds and climate

April 16, 2009


Clouds can cool or warm the underlying surface. They reflect some of the incoming shortwave solar radiation back to space (cooling). They also absorb infrared radiation and re-radiate it to all directions, with a wavelength (and thus energy) that depends on cloud temperature (and thus height). The latter is a quasi-greenhouse effect (warming).


The colder (=higher) the cloud, the less energy will be re-radiated to space, so the more the atmosphere will warm. (The same happens with greenhouse gases: the higher re-radiation into space takes place, the less energy is lost, and the more the atmosphere will warm. In that sense, the sky is the limit in terms of a warming effect of CO2: no matter if the absorption bands at the surface would be saturated, it would just move up the re-radiation to a higher altitude, thereby still adding to the warming effect.)

At night, there’s no solar radiation to reflect, and so clouds mainly warm the surface. However, the net effect of clouds is a cooling.


The basic response of liquid water clouds to an increased CCN population is expected to be a change in the droplet size spectrum: Assuming the liquid water content remains relatively unchanged, more cloud droplets will form, having a smaller average size. Absorption depends on particle volume, whereas reflection depends on its geometric cross-section, and thus the ratio of reflection over absorption of the cloud increases as the droplet size decreases. Dependent on the altitude of the cloud droplets, and the albedo of the underlying surface, this will usually result in cooling. A smaller average droplet size may increase cloud lifetime, via the suppression of rain. Cloud cover could be susceptible to CCN concentration, especially in remote areas. In reality, different feedback mechanisms involving micro-physics and radiative transfer make the picture more complicated. For example, the liquid water content of a cloud is highly variable and not much is known about the response of liquid water content to increasing CCN concentrations.


The effects mentioned here concern only “warm” clouds; much less is known about the response of ice clouds. If through human activities the concentration of ice nuclei has also decreased (e.g. black carbon), there could be a “glaciation indirect effect”: more ice nuclei cause a cloud to rain out quicker, and may thus lead to a decrease in cloud cover (i.e. opposite to the “warm” indirect effect). This is still very hypothetical, however.


From local measurements of ship tracks, relations were found between particle emissions, cloud droplet diameter and cloud albedo. Still, more simultaneous measurements of aerosol particles, cloud droplets and radiative properties are necessary to gain a more accurate understanding of climate change. There are strong indications that besides inorganic salts, some organic compounds may also contribute to the CCN activity of an aerosol. Ultimately though, size may matter more than chemistry for the cloud nucleating ability of aerosols.

The atmospheric aerosol

April 13, 2009


This serves as an introduction to atmospheric aerosols to accompany my guest post on RealClimate. I’ll be discussing more about aerosols in the near future, including their potential role in GeoEngineering.



Atmospheric aerosol refers to the liquid or solid particles suspended in the air. They can be emitted directly into the atmosphere (primary aerosol), or they can be formed in the atmosphere by gas-to-particle conversion (secondary aerosol). The oxidation of precursors, such as SO2, NO2, and suitable organic compounds, can produce compounds with very low vapor pressures. These can condense to form a new particle, often in combination with other species, such as water vapor and NH3, a process called homogeneous nucleation. They can also condense on existing particles and contribute to them growing in size, referred to as condensation. Furthermore, aerosol particles can undergo evaporation, deposition, and coagulation, whereby two particles collide to form one larger particle. All these processes influence the evolution of the aerosol population in time.



Atmospheric aerosol particles span several orders of magnitude in diameter, from a few nanometer to hundreds of micrometer. Cloud droplets and raindrops can grow even larger, but are usually treated separately. Though not all particles are spherical, they are usually characterized according to their equivalent spherical diameter. Most of the properties, and thus effects, of particles depend on their size.



The tropospheric number concentration of aerosol particles ranges from several tens or hundreds per cubic centimeter of air in remote locations, to more than hundred thousand a million per cubic centimeter in polluted environments or after a strong nucleation event. The number concentration as a function of particle size is described by a size distribution. An important feature of atmospheric aerosol size distributions is their multimodal character: commonly found are the aitken mode, the accumulation mode, and the coarse mode. The nucleation mode consists of freshly nucleated particles and usually occupies diameters below a few tens of nanometers. The accumulation mode can vary in size between 50 nanometer and 1 micrometer, and its name comes from the fact that particles tend to accumulate in this size range, because they bear the longest lifetime (~2 weeks): smaller particles are efficiently removed by deposition and coagulation, and larger particles are efficiently removed by gravitational settling. The coarse mode usually refers to particles larger than 1 micrometer diameter. Sometimes an Aitken mode is defined as the mode that could exist between the nucleation and accumulation mode.

The number concentration of the atmospheric aerosol is usually dominated by sub-micrometer particles, while the total particle surface area and volume (and thus mass) are more influenced by larger particles. The exact shape of the size distribution depends on the environment, as it reflects the sources, sinks, and transformations of the particles.



Atmospheric particles have both natural and anthropogenic sources.

Examples of natural (primary) particle emissions are the shattering of sea-spray into tiny droplets, that evaporate before gravitating back towards the water surface, particles caused by forest fires, soil dust, pollens, etc. Examples of natural emissions of precursors to (secondary) aerosol are SO2 and H2S from volcanoes, dimethylsulfide (DMS) from phytoplankton, volatile organic compounds such as monoterpenes from vegetation, etc.

Examples of anthropogenic (primary) particle sources are vehicular emissions, industrial emissions, and some natural sources that are significantly enhanced by human activities, such as biomass burning and soil dust (due to e.g. erosion). Anthropogenic precursors to (secondary) aerosol are emitted by industry, burning of fossil fuels, usage of solvents, etc. Primary emissions contribute relatively more to the aerosol volume (because they’re often large), while gas to particle conversion (secondary aerosol) contributes relatively more to the aerosol number.


Cloud formation

Aerosol particles are necessary for clouds to form: Dependent on their size, composition and hygroscopic properties, they can act as cloud condensation nuclei (CCN) on which the available water vapor condenses, when it reaches a certain critical super-saturation of a few tenths of a percent. This process is called (cloud droplet) activation. Without these substrates to condense on, the relative humidity would have to exceed several hundred percent in order for the water vapor to overcome the Kelvin effect and homogeneously nucleate into droplets.

Similarly, some aerosol particles act as ice nuclei on which water can freeze into an ice crystal. Aerosol particles also provide a surface for specific heterogeneous reactions to take place, and thus affect the chemistry of the atmosphere.


Health effects

Atmospheric aerosols have potentially far reaching effects on health, ecosystems, and climate. Aerosol particles are a major component of smog, and act to reduce visibility. Their health effects are mainly due to their adverse effects on the respiratory function, and some also act as irritants for the eyes. Small particles can enter deep into the lungs, reaching the alveoli, where the transfer of O2 and CO2 takes place. H2SO4 is often a major component of these small particles, and dependent on their history other toxic compounds, such as heavy metals and PAH’s, could also have partitioned into the particle phase. Due to their low volatility, these species would otherwise not have reached that deep into the lungs. Atmospheric particles have been found to play a major role in excess mortality in epidemiological studies in several cities. Through their acid composition, aerosol particles also contribute to acidification of ecosystems.


Climate effects

Atmospheric particles contribute to climate change directly by scattering or, in the case of black carbon, absorbing sunlight, and indirectly by changing the radiative properties and lifetime of clouds. Both direct and indirect radiative forcing are influenced by particle size, composition and relative humidity. These processes are responsible for the largest amount of uncertainty in assessing climate change. The upper end of the estimated (direct and indirect) negative forcing of atmospheric aerosols could -at least on a global and annual average- balance the positive forcing of CO2. This does not necessarily mean that on a regional scale the opposing forcing mechanisms will balance each other. Aerosols are not homogeneously mixed throughout the globe, whereas the atmospheric lifetime of most greenhouse gases is much larger, so they will get distributed equally over the globe, despite large spatial differences in emissions.



The short lifetime (days to weeks) of aerosols is an important reason for the uncertainty in their role in climate change. It causes their concentration to be highly variable in time and space, and it’s hard to even know what the global concentration is, let alone what it was in the pre-industrial era. Add to that their variability in size and chemical composition, and the poorly understood role of clouds, and it’s clear that the uncertainty in aerosol radiative forcing will remain a steady feature of climate science for some time to come.


Other introductory explanations about aerosols are here, here, and here (German) and good presentation slides here.


AGU highlights: Effects of particle nucleation and cosmic rays on clouds

December 22, 2008


This past week the annual AGU (American Geophysical Union) fall meeting was held in San Francisco. There were a number of interesting climate related talks that I attended. Here’s a short briefing of some of these, related to the climate effects of aerosol, and possibly of cosmic rays. This post is more technical than other ones on this blog. Meeting abstracts can be searched here.



Aerosol nucleation refers to the formation of a stable aerosol particle (miniscule liquid droplet of a few nanometers in diameter) in the atmosphere. These particles can grow larger in size to then affect cloud formation, and thus climate. A controversial hypothesis sais that the decreasing flux of cosmic rays from outer space has decreased the amount of particles produced by nucleation, and thus decreased the cloud cover and thereby warmed the climate. The flux of cosmic rays has remained constant over the last 50 years (apart from the 11 year cycle mirroring solar min/max conditions), so they can’t have influenced the warming during this time period. The micro-physics of the processes involved are poorly understood, but important/interesting for a number of other reasons as well, eg the climate effects of aerosols in general.


AGU highlights

The Finnish group (headed by Kulmala) gave an overview of long term measurements of particle nucleation at their Boreal forest site, reporting that there was no relation whatsoever with cosmic rays. Sulfuric acid (a prime agent in the nucleation process) had a slightly decreasing trend over the past decade, whereas both particle nucleation and growth rates slightly increased, suggesting that the role of organics in both these processes may have increased. According to their analyses, the role of ion induced nucleation (relevant to the hypothetical cosmic rays – cloud link) can explain 10 to 20% on average of the rate of production of 2 nm sized particles.


Bondo, from the Danish group headed by Svensmark, reported on laboratory (chamber) studies of ion induced nucleation under exposure of ionizing radiation from a radioactive source. The presentation focused mainly on comparing theoretical calculations with the measurements. To the speaker’s credit, no far reaching climate conclusions were drawn.


Yu and Turco analyzed data from the Finnish group, and whereas the Fins calculate a less than 10% contribution of ion induced nucleation (to the total amount of particles produced), Yu and Turco arrived at the opposite conclusion. If anything, this indicates that the nucleation processes are very poorly understood, partly because of the strong non-linearity (and thus strong dependencies on uncertain parameters) involved. Yu cautioned on the use of the nucleation theorem (which sais that the log-log slope of the nucleation rate to the sulfuric acid concentration is equal to the number of sulfuric acid molecules in the critical cluster), because other factors (if they’re not constant) may influence the slope found.


There were two talks on global modeling of aerosol nucleation. Spracklen showed that only including primary emissions of aerosol underestimates the number of aerosol particles typically measured, so nucleation contributes significantly to the aerosol number budget. Moreover, only including binary homogeneous nucleation (according to the classical nucleation theory) still leads to an underestimation of particle numbers; some sort of empirical nucleation scheme is needed to reach reasonable agreement. That is something I recognize from most other (including my own) research of both field and laboratory measurements.


Pierce, using a different model (based on the GISS II GCM), investigated the potential effect of cosmic rays on cloud condensation nuclei and cloudiness. Two different parameterizations for ion induced nucleation (Modgil et al and an ‘ion-limit’ assumption that all ions go on to form a new particle), somewhat surprisingly, both lead to a change in nucleation of about 20%, in response to a prescribed change in cosmic ray flux. This difference in cosmic ray flux was of the order of the difference between solar minimum and solar maximum, which happens to be comparable to the change in cosmic ray flux over the first half of the 20th century. If this 20% change in nucleation would translate directly in a 20% change in the number of cloud droplets, one might expect a change in cloud cover of about 2% (which is similar to the magnitude suggested by Svensmark et al as having been caused by cosmic rays). However, as one might expect, the number of cloud condensation nuclei (CCN) in the model changed by much less than 20% in response to the 20% change in nucleation. Let alone the number of cloud droplets and the cloud cover.


Penner gave a very interesting talk in which she showed that satellite derived estimates of the aerosol indirect effect (i.e. their effect on clouds and thus on climate) are substantially smaller than model calculated estimates (the former are in the range of -0.2 W/m2, whereas the latter range between roughly -0.5 and -1.5 W/m2. The negative sign means it causes cooling.) However, satellite estimates are based on the aerosol optical depth, and this may cause an underestimation of the aerosol indirect effect (AIE). The model calculated AIE is very sensitive to nucleation and to primary aerosol emissions.


Concerning the second indirect effect, based on satellite measurements a 5% difference in the probability of precipitation was found between clean and polluted clouds (where the latter has a lower probability than the former, because of the enhanced droplet number and thus reduced droplet size). Rosenfeld claims that this so called cloud lifetime effect is stronger than the cloud albedo effect, whereas others claim the opposite.


At the last EGU meeting in Vienna Philipona presented empirical evidence from several mid-European sites that the direct effect of aerosol (scattering of sunlight) may be stronger than their indirect effect via clouds, whereas modeling results usually point to the opposite. They suggested that a decrease in aerosol loading may have contributed significantly to the warming of the past 30 years, and that when the aerosol loading stabilizes again, the rate of warming will decrease as a consequence. This fits in with the global dimming – global brightening picture (the amount of solar radiation received at the surface depends on the aerosol loading because of their light scattering properties).


The large uncertainties in the aerosol effects on climate may affect estimates of the climate sensitivity (equilibrium temperature change for a given change in climate forcing) (Penner) and/or of the ocean response time (lag in temperature response caused by thermal inertia of the oceans) (Hansen). The three are connected, and their magnitudes can be traded off in trying to match the 20th century temperature record by models. For example, a larger negative aerosol forcing (and thus a weaker net positive forcing) would need to be combined with a higher climate sensitivity and/or a shorter ocean response time in order to still provide a good match. Of course, there are other constrains on these processes as well that have to be taken into account.


%d bloggers like this: