Introduction

The AEROCOM-project is an open international initiative of scientists interested in the advancement of the understanding of the global aerosol and its impact on climate. A large number of observations (including MODIS, POLDER, MISR, AVHHR, SEAWIFS, TOMS, AERONET and surface concentrations) and results from more than 14 global models have been assembled to document and compare state of the art modeling of the global aerosol. A common protocol has been established and models are asked to make use of the AEROCOM emission inventories for the year 2000 and preindustrial times. Results are documented via interactive websites which give access to 2D fields and standard comparisons to observations. Regular workshops are held to discuss findings and future directions.

Background

Multi-component aerosol modules in global models promise a much needed better seasonal and regional characterization of aerosol. However, the added complexity may have introduced many (potentially offsetting) errors. Thus, a rigorous validation effort is needed. Initial comparisons model evaluation efforts to remote sensing data (e.g. Kinne et al and Penner et al) illustrated the need for more detailed comparisons. Only a much expanded model output will permit process studies, like pioneering comparisons of near surface sulfate mass (COSAM). This study also indicated that unwanted difficulties arise from differences in model initialization (e.g. source strength or meteorology). Concepts of a new model-intercomparison and model-evaluation effort were discussed during IAMAP 2001 and introduced to modelers or interested by-standers during IGAC 2002. A driving force behind the new inter-comparison is availability of more accurate aerosol products from satellite, a tighter ground network of aerosol measurements and a multitude of results from field experiments.

Objectives

The joint initiative AEROCOM shall document and understand the differences apparent in current global aerosol models. It seeks answers to key modeling questions, such as consequences of differences in

  • source strength
  • vertical transport or removal
  • water uptake
  • aerosol composition
  • aerosol size
  • broadband radiative transfer

on the overall quality of the aerosol simulation and estimates of the direct radiative effect of the aerosol.

The project shall make use of new remote sensing measurements of aerosol from ground and space. Satellite data will be applied to investigate to what extent patterns of global aerosol distributions are correctly represented in aerosol models. Ground statistics will be applied to provide quantitative references in terms of composition, size, concentration and altitude. To address not only regional but also seasonal differences, those measurements and models shall be used which cover at least an entire year. For comparability models shall be forced by analyzed meteorology. Regional models shall be included, if they can fit in the output framework set for this global model intercomparison.

The range of model results documented with AEROCOM shall help to establish and reduce the uncertainty in aerosol climate forcing estimates.

More specifically, AEROCOM shall document and intercompare among models and to available measurements for a given year on a seasonal and regional basis

  • Near surface winds (critical for the production of for natural aerosol)
  • Parameterization schemes in aerosol processing
  • Diffusive properties of the transport model
  • Mass (balances) for each aerosol type (dust, seasalt, sulfate, bc, org.matter)
  • Mass compositional mix
  • Humidification and water uptake for each aerosol species
  • Assumptions on size and refractive index
  • Optical depth (balances) for each aerosol type
  • Optical depth compositional mix
  • Assumptions in the broadband radiative transfer scheme

Aerocom also aims to provide

  • Control simulations with prescribed sources
  • Control simulations with prescribed ambient rel.humidity for swelling
  • Reduced output at times of model-scale ‘cloud-free’ satellite observations
  • Simulations at different spatial resolutions

Future

AeroCom Phase II Planning Document

To achieve the goals of the AeroCom phase II, work is required on four areas detailed below:

Goals

The primary goal of the AeroCom initiative is to evaluate global aerosol models in order to obtain reliable estimates of the present and future aerosol impact on climate and air quality. Model output documentation and comparison shall be facilitated to help upcoming assessments, relevant for policy decisions related to climate change and air quality. Primary scientific questions concern the ability to reproduce and predict aerosol constituent loadings, their optical properties, the aerosol-cloud interactions and the radiative impact on different spatial and temporal scales. While more models become able to study the climate response to anthropogenic aerosol, it is time to assess the postulated feedback processes in the framework of a model intercomparison.

 

Structure

Preface: It is recognised that considerable effort is underway on national and international levels to better understand the ambient aerosol. The AeroCom initiative thus primarily seeks to link and look for synergy with other programs.

Organising committee: To improve the coordination of the analysis it is proposed to create a broader coordination committee, comprised of representatives active in the field of analysing aerosol models and observational data. A balanced international representation is seeked. Propositions shall be send to Michael Schulz and Phil Rash (on behalf of IGAC) before the Virginia Beach meeting.

Task groups: To further analyse the AeroCom model data assembled it is proposed to form task groups, in which one member has direct access to the data base and presents the planned work and the analysis results to the AeroCom participants.

Link to IGAC and WCRP: AeroCom has been solicited to participate in a new activity being initiated jointly by WCRP-SPARC and IGBP-IGAC on the issue of "Atmospheric Chemistry and Climate (AC&C)". At a recent workshop in Boulder (August 2006) it has been proposed to link the work on model evaluation in three related fields (1) stratospheric chemistry, 2) tropospheric chemistry and 3) aerosols) in a joint "Atmospheric Chemistry and Climate" initiative. AeroCom and ccmVAL have been identified as possible major contributing components, if not subprojects, to such an initiative. The report from the Boulder workshop and further cooperation shall be discussed in the forthcoming AeroCom and ccmVAL workshops and in a AC&C meeting in January 2007.

Link to HTAP: Synergy is proposed by cooperating with the Task Force on Hemispheric Transport of Air Pollution. A set of experiments, planned to analyse long range transport and source receptor relationships, shall be analysed in connection with previous AeroCom experiments. An exchange of tools and harmonized formats are being developped. Aerocom modellers are thus encouraged to participate especially in experiments SR1 and SR5.

AeroCom workshops: Annual work shops are proposed to be held in autumn. For 2007 a proposition has been made to combine a french organised CNES workshop on the work-up of the A-train data with an AeroCom workshop.

Joint experiments

The following list is a tentative list of model experiments, which eventually could be assembled in the framework of AeroCom. Further discussion, reduction and detailing will be done on the 5th and forthcoming AeroCom workshop. For each of these experiments a task group must form to ensure successful analysis.

1) Standard model runs: To monitor progress and evolution in current multi-component aerosol models it would be desirable to allow for "any-time submission" of model results via an automated transferral&cataloging web based tool (see technical section below). A simplified protocol is needed, eventually concentrating on parameters needed for standard benchmark tests and annual aerosol budget control.

2) HTAP experiments: Source receptor relationships on a continental level need to be quantified to characterize long-range transport of air pollution. This will be better understood based on a multi-model evaluation. AeroCom modellers are thus encouraged to participate especially in experiments SR1 and SR5 (see HTAP experiment description) and if possible others.

3) Climate response due to aerosols: Several models have done or are about to prepare coupled (chemistry-)aerosol-climate simulations. At stake are problems of representative aerosol emission scenarios, the importance of different feedback mechanismns, the regional impact of an inhomogeneous aerosol forcing, the potential role of aerosol to explain the evolution of surface air temperature and global dimming. A set of proposed diagnostics and storage of output in a new AC&C data center for such coupled chemistry climate model runs would eventually help resolving the problems.

4) Aerosol loadings in the recent past (1980-2005): Reconstructions of the aerosol emissions and reanalysis of the synoptic meteorolgy become now available. It would be desirable to analyse the recent past in a concerted manner, especially because aerosol observations provide important constraints after 1980.

5) Ensemble prediction of global PM air quality: Prediction of PM levels is currently done with regional air quality models. While global models use higher resolutions and eventually can provide border conditions for regional models, it becomes of interest to review work in this area and eventually prepare an ensemble prediction of PM levels on a global level.

Process analysis and evalution with observational data

Model development will benefit from improved analysis of currenct concepts and a broader testing of the model output. Standard benchmark tests and in depth analysis are both needed to make progress. Use and understanding of observational data is crucial. It would be desirable to construct a table of identified standard diagnostics to document model quality. See the ccmVAL evaluation table as an example. Major fields expected to be worked upon by an AeroCom task group:

1) Indirect effect: The importance of the indirect effect is well recognised. The diagnostics and data driven evaluations are nervtheless in its infancy. A set of diagnostic parameters should be drawn from recent work such as that of Penner et al 2006 and others. Such diagnostics should develop into standard tests and documentation.

2) Aerosol dynamics: Aerosol size distribution is expected to play a big role in aerosol-cloud interaction, determining aerosol life time and radiative impact. Diagnostics in the original AeroCom protocol were not sufficient to understand eg the impact of different size assumptions on aerosol life time. New satellite retrievals and ground based observations are currently underexploited with respect to the validation of modelled aerosol size distribution.

3) Dust: Mineral dust is a major natural and possibly anthropogenic aerosol component with potentially large impact on radiation and regional climate. Considerable diversity has been diagnosed among AeroCom models with respect to simulating the dust cycle (<Textor et al. 2006). More targeted evaluation is expected to help constrain the dust cycle and radiative impact.

4) Carbonaceous aerosol: The role of the carbonaceous aerosol for radiative forcing has been shown to be a large source of diversity among models (eg Schulz et al. 2006). The analysis must be reinforced. Observations of aerosol absorption, single scattering albedo and black carbon need to be linked more closely to modelled distributions.

5) Vertical profile of the aerosol: <Initial comparison to lidar observations from EARLINET and the ARM site have shown considerable differences to the modelled vertical profile of the aerosol. The differences in the vertical profile of the aerosol are suspected to be responsible for a major part of the diversity in lifetime in AeroCom models (<Textor et al. 2006). With the upcoming Calipso data set and an improved compilation of aircraft and mountain stations it should be possible to further constrain the modelled vertical profile of the aerosol.

Technical implementation

Joint AC&C data center: The storage and analysis of an extended AeroCom data set, as outlined above, would require a large scale data centre with the commitment to maintain long-term maintenance and support. A joint storage of coupled chemistry-aerosol-climate simulations together with standard AeroCom model simulations and those from other atmospheric tracer comparisons is needed.

Automated model submission and diagnostics: AeroCom seeks to implement the AutoMod system developed by Charles Doutriaux (with OCMIP participants) at the LSCE and PCMDI. Such a system would allow via a web based interface any user to launch an upload of a new model run. Diagnostic tools can be selected to be launched after such submissions. Analysis groups, model experiments, access and model output charcteristics are stored in a meta-database. (see also recommendations in report of intercomparison IT-workshop in Ispra 2006 ).

Aerosol emission scenarios: The next IPCC runs are expected to be build on multi-component aerosol emission scenarios. It is proposed to prepare in cooperation with other initiatives and individuals three consistent 1860-2100 aerosol emission scenarios, reflecting low, mean and high emission estimates. As for the AeroCom B and PRE experiments this is expected to be a public data set.

New standards and rules for model submission: AeroCom proposes that model output complies to the CF convention. An extension of standard names is currently prepared via the standard name discussion website. Utimately we propose that modellers use CMOR fortran tools to reformat their output with the help of specific tables designed to fit each AeroCom experiment. (see IPCC example for illustration). A full set of tools will be prepared in autumn to facilitate this transition. This would allow rapid processing and stability needed for future diagnostic tools such as AeroCom tools and AutoMod.