Science Briefs

AeroCom Project Compares Organic Aerosol Handling in 31 Global Models

Atmospheric aerosols have significant impacts on both air quality and climate. These minute airborne particles, and also gases that lead to their formation, can come from factory and vehicle emissions, volcanic eruptions and biomass burning, sea spray and desert winds, and even natural forest and soil emissions.

Photo of hazy Great Smoky Mountains

Figure 1. Aerosol "blue haze", as seen here in the Great Smoky Mountains in North Carolina, is formed by the oxidation of volatile organic compounds emitted by trees. This effect may be strengthened by emissions from human activities. Image courtesy Frank Kehren via Flickr.

The majority of fine aerosols consists of non-refractory material and has been found to contain large amounts of organic matter. They are known as organic aerosols, or OA. Organic aerosols that are directly injected in the atmosphere in the particulate phase are called primary organic aerosols (POA), while those that are created by the oxidation of volatile organic compounds are known as secondary organic aerosols (SOA).

For simplicity and due to lack of experimental data, global computer models of aerosols have historically considered just one category of OA with the properties of POA. This is the approach that was taken during Phase I of the Aerosol Comparisons between Observations and Models (AeroCom) project.

Comparisons of individual models with aerosol observations have shown a large underestimate of the ΟΑ component by models, especially in polluted areas. In order to capture the magnitude, seasonality, and the OA physical and chemical properties, several global models now treat SOA as semi-volatile, based on an continuously increasing amount of experimental data on SOA formation in the atmosphere. Still, OA simulations have many degrees of freedom due to incomplete knowledge on their behavior and fate in the troposphere. This means modelers must make several assumptions, which leads to parameters that may vary greatly between models.

'Box-and-whisker plot of model aerosol sources

Figure 2. POA, SOA and OA global budgets, with comparison to AeroCom Phase I results. Bar colors are POA (brown), SOA (green), OA (blue), AeroCom A (red), and AeroCom B (orange). The boxes represent the first and third quartile range (50% of the data), the line is the median value, the star is the mean, and the error bars cover 9% to 91% of the data. Outliers are presented with X-symbols, and the numbers of models participating in each bar's statistics are presented with a grey number at the top.
+ View larger image.

In Phase II of AeroCom, a large-scale model intercomparison was performed to document the current state of OA modeling in the global troposphere, evaluate the OA simulations by comparison with observations, identify weaknesses that still exist in models, explain the agreements and disagreements between models and observations, and attempt to identify and analyze potential systematic biases in the models. The study quantified the uncertainties in the simulated surface OA concentrations and attributed them to major contributors. The ensemble of the simulations was used to build an integrated and robust view of our understanding of organic aerosol sources and sinks in the troposphere.

The study, led by Kostas Tsigaridis, an associate research scientist at Columbia University and at the NASA Goddard Institute for Space Studies, and Maria Kanakidou, professor at the University of Crete, Greece, brought together 71 scientists from 49 institutions, 31 different global models of various degrees of complexity, and observational data from more than 1000 stations around the globe acquired over the past few decades.

Although many models still use the simplistic approach of AeroCom Phase I, about half of the 31 models participating in Phase II included explicit treatment of semi-volatile SOA formation in the atmosphere. Four models also accounted for chemistry in the atmospheric aqueous phase (both cloud and particulate) and six models accounted for natural sources of POA from land and sea.

The models in Phase II presented a much higher variability in their total OA sources than those from Phase I (see Fig. 2). This is primarily attributed to variability in SOA chemical production. The comparison between Phase I and Phase II model results indicates that the significant uncertainties in the POA emissions of global models in AeroCom Phase I have not been reduced.

Maps of model aerosol mean and standard deviation

Figure 3. At left, the composite annual mean organic aerosol surface air concentrations of the models, defined as the median of the model fields on a 5°×5° grid. At right, the model diversity, defined a the ratio of the standard deviation of all models from their corresponding mean values.
+ View larger image

The models' composite annual mean OA surface air concentrations exceed 0.5 microgram of carbon per cubic meter across most continental regions, as shown at the left of Fig. 3, with maximum concentrations primarily over biomass burning regions and secondarily over industrialized areas. The model diversity (right side of Fig. 3) is smallest over and downwind of continental regions, with ratios below 1 over most continental areas and above 1 over the remote oceans. Diversity less than 2 is calculated over the northern Pacific and Atlantic oceans and exceeding 2 over most of the oceanic regions south of 30°S and over Antarctica. The latter difference is a result of marine OA sources that are taken into account by only a few models.

Plot showing vertical distribution of aerosols in the various models

Figure 4. Annual global mean vertical profile of organic carbon (OC) at ambient conditions, from the surface to 50 hPa (about 20 km altitude). Most models assume that OA=1.4×OC.
+ View larger image

The models' vertical distribution of OA shows an even higher diversity, spanning more than an order of magnitude. The simulated global mean organic carbon (OC) concentrations increase with altitude up to a mean pressure level of about 800-900 hPa, and then decrease with altitude (see Fig. 4). Some models simulate a secondary maximum at around 100-200 hPa, with concentrations much lower than the maximum near the surface; this is primarily due to condensation of semi-volatile SOA at low temperatures but also due to OA accumulation above clouds, where dry and wet deposition is absent. The models that explicitly simulate SOA calculate a slower removal rate of SOA from these altitudes than is assumed in the other models.

The increasing complexity did not significantly improve the model performance. However, model complexity is necessary to quantify the anthropogenic impact on climate via the aerosol direct and indirect effects. The strength of secondary OA sources that are enhanced by interactions of natural and anthropogenic emissions remains an open question that cannot be answered by a simple parameterization. Furthermore, the organic aerosol impact on climate depends on aerosols' physical, chemical, hygroscopic and optical properties, and their distribution in the atmosphere. Both properties and distribution are affected by continuous evaporation and condensation processes of semi-volatile organic material which must be simulated.

New information from dedicated recent and future field campaigns is expected to shed light on organic aerosol formation processes and how they are altered in the presence of anthropogenic pollution. Model development related to OA is expected to accelerate in the near future and must be performed in parallel with extensive model evaluation.


Tsigaridis, K., N. Daskalakis, M. Kanakidou, P.J. Adams, P. Artaxo, R. Bahadur, Y. Balkanski, S.E. Bauer, N. Bellouin, A. Benedetti, T. Bergman, T.K. Berntsen, J.P. Beukes, H. Bian, K.S. Carslaw, M. Chin, G. Curci, T. Diehl, R.C. Easter, S.J. Ghan, S.L. Gong, A. Hodzic, C.R. Hoyle, T. Iversen, S. Jathar, J.L. Jimenez, J.W. Kaiser, A. Kirkevåg, D. Koch, H. Kokkola, Y.H. Lee, G. Lin, X. Liu, G. Luo, X. Ma, G.W. Mann, N. Mihalopoulos, J.-J. Morcrette, J.-F. Müller, G. Myhre, S. Myriokefalitakis, S. Ng, D. O'Donnell, J.E. Penner, L. Pozzoli, K.J. Pringle, L.M. Russell, M. Schulz, J. Sciare, Ø. Seland, D.T. Shindell, S. Sillman, R.B. Skeie, D. Spracklen, T. Stavrakou, S.D. Steenrod, T. Takemura, P. Tiitta, S. Tilmes, H. Tost, T. van Noije, P.G. van Zyl, K. von Salzen, F. Yu, Z. Wang, Z. Wang, R.A. Zaveri, H. Zhang, K. Zhang, Q. Zhang, and X. Zhang, 2014: The AeroCom evaluation and intercomparison of organic aerosol in global models. Atmos. Chem. Phys., 14, 10845-10895, doi:10.5194/acp-14-10845-2014.


Please address all questions about this research to Dr. Kostas Tsigaridis.