A31F-01 INVITED
Sources, Paths and Scales of Transport and Transformation of PM in the Mediterranean and North Atlantic
The Greater Mediterranean Region (GMR) is known for its unique climatic conditions and the associated air quality characteristics. The prevailing weather conditions especially during the warm period of the year define characteristic paths and scales of transport and transformation of anthropogenically and naturally produced air pollutants of gaseous and/or particulate form. Both types of sources (natural or anthropogenic) are considerable. The work presented here has two major objectives: one is to summarize the existing knowledge on the transport paths of particulate matter in the GMR and the other one is to illustrate some new findings related to the PM transport and transformation properties in the GMR. Findings from previous studies indicate that anthropogenically produced air pollutants from European sources can be transported over long distances reaching Africa, Atlantic and North America. Particulate matter of natural origin, like Saharan dust, can be transported towards Atlantic and North America mostly during the warm period of the year. Recent model simulations and studies in the area indicate that specific long range transport patterns of aerosols have limited or negligible contribution to air quality degradation in the GMR such as the transport from Asia and Indian Ocean, Central Africa or America, as compared with the other sources. Also, new findings from this work suggest that the imposed EU limits on PMs cannot be applicable for Southern Europe without the discretization of the origin (natural or anthropogenic). The impacts of PM levels in the GMR are not limited only to air quality but there are other serious implications on water budget and regional climate. These are issues that need extensive investigation since the processes involved are rather complicated. At a framework of a recently-funded EU project (CIRCE) further model development is under way in order to cover critical issues related to feedbacks between regional climate and air quality.
http://forecast.uoa.gr
A31F-02
African aerosols and Atlantic tropical cyclone activities
Previous studies have shown that the Atlantic basin major hurricane (MH) activity is associated with western Sahelian monsoon rainfall, while rainfall in the Sahel is found to be highly anti-correlated with the African dust storms. So if the Atlantic basin MH activity may be anti-correlated with the African dust aerosols? In order to investigate the relationship between the African dust and the tropical cyclone (including both tropical storms and hurricanes) activities in the Atlantic basin, we explore how the African dust may link to Atlantic TC activity by using the long-term (1982-2005) NCEP Reynolds sea surface temperature (SST) product, and tropical cyclone (TC) data from the National Hurricane Center Best Track Files, and the TOMS aerosol index (AI) data, because the TOMS AI positive values are associated with UV-absorbing aerosols, like dust and smoke. Although no significant negative correlation between the TOMS AI and the Atlantic TC or MH frequency and duration is found, the initial locations of the Atlantic tropical cyclones did occur over the ocean where the aerosol loading was low. Our analysis shows that SST over the north tropical Atlantic ocean is anti-correlated with the TOMS aerosol index. This may be due to the radiative forcing of the aerosols. The effects of the dust aerosols carried across the West African region led to a lowering of SST and therefore inhibited tropical cyclogenesis. During 2005, the aerosol loading along the western African coast was unusually low, while the SST over the main development region (MDR) was abnormally high, and the Atlantic TC/hurricane activities became record strong. We propose future observations to test these results.
A31F-03
What Causes Aerosol Growth and Ozone Production in Smoke Plumes?
The growth of aerosol particles and production of ozone in smoke plumes is the result of a complex interaction between horizontal diffusion, gas-phase oxidation, coagulation, and mass transfer between phases. Models allow us to separate the effects of these processes and predict their impact on the global environment. We present the results of a new model of gas and aerosol chemistry applied to young biomass burning plumes. The model includes heterogeneous chemistry, kinetic mass transfer, coagulation and the formation of secondary organic and inorganic aerosol. Comparison with measurements from SAFARI 2000 (Hobbs et al., 2003, JGR, doi:10.1029/2002JD002352) suggests the baseline model underpredicts ozone formation and the growth of aerosol within the plume. We explore whether the model predictions can be improved by (1) including heterogeneous HONO production, and (2) adding in surrogates for the uncharacterized organic compounds emitted by the biomass burning. Including the heterogeneous reaction NO$_{2}$ => HONO greatly improves the match for ozone, OH, and aerosol nitrate concentration, but only when the uptake coefficient approaches 10$^{-3}$, which is over an order of magnitude higher than previously reported values (Stemmler et al., 2006, doi:10.1038/nature04603). Using the reaction NO$_{2}$ => 0.5 HONO + 0.5 HNO$_{3}$ with an uptake coefficient of 10$^{-3}$ (the top of the range recommended by Jacob, 2000, Atm. Env.,34, 2131-2159) provides an even better match for aerosol nitrate, but produces less O$_{3}$ and OH than the first reaction. Direct measurements of HONO and OH in young biomass plumes would help determine if this chemistry is taking place. We used two surrogates to model the uncharacterized compounds: long chain alkanes and monoterpenes, representing primary and secondary sources of condensable compounds respectively. Complete condensation of the long-chain alkanes can account for nearly all of the observed increase in organic carbon. However, the accommodation coefficient must be near 10$^{-3}$ or the alkanes will condense too quickly or too slowly. This value is reasonable when compared to measured accommodation coefficients of organic vapors on organic films (Donaldson et al, 2005, Faraday Discuss, doi:10.1039/b418859d). Monoterpenes gradually increase condensed organic carbon even with accommodation coefficients of 0.1, but produced only 35% as much as the alkanes. Production of sulfate and total particulate matter is below observations in all simulations described above. Better characterization of the composition of the condensed organic matter as a function of downwind distance could help determine the source of the organic aerosol growth.
A31F-04
Studying dust using AIRS thermal infrared radiances
The radiances measured at the top of the atmosphere (TOA) by high spectral resolution sounding instruments such as AIRS are significantly affected large particle size mineral aerosols (dust). The broad, semi-continuous wavelength coverage of AIRS in the 8-12 micron region allows both detection and retrieval of dust aerosol parameters. We focus on the ability of AIRS to monitor large-scale transport of dust and summarize the strengths and weaknesses of thermal infrared for these purposes. Preliminary results using AIRS to study the Outgoing Longwave Radiation (OLR) forcing by dust will also be presented as will efforts to remove the effects of dust on AIRS retrievals of temperature and humidity fields.
A31F-05
Simulations of Mineral Dust Content With CHIMERE-Dust Model
Simulations of the mineral dust cycle have been performed whith CHIMERE-Dust model over a domain that includes North Africa, the Mediterranean basin and the North Tropical Atlantic Ocean (10S-60N and 90W-90E) with a 1�x1� resolution using the ECMWF (European Center for Medium-Range Weather Forecasts) meteorological fields for two years, 2000 and 2001. As a validation, we compare the simulated dust concentration fields with photometric data from the AERONET network. From the comparisons between the simulated and measured aerosol optical depth for several stations of the Mediterranean basin, the model appears to reproduce correctly the intensity and occurrences of the dust events. Over Western Africa, the results are not as satisfying since some of the most intense dust events observed on the continent and downwind are not captured by the model. In addition, the simulated events are generally underestimated compared to the measured ones. It appears that these differences in the model performances are connected to the origin of the dust plumes. For example, dust plumes coming from Libya are well simulated while dust plumes originating from the Bod\'el\'e depression not as frequent as intense as the observations suggest. Soil properties in these two regions are comparable and typical of very erodible surfaces. We thus focused on the comparison between the ECMWF 10m wind speed fields and 10m wind speed measured at the meteorological stations located in both areas. We noticed that over Libya, the measured and ECMWF 10m wind speed are in very good agreement, while the meteorological model does not reproduce the extrema of the measured wind speed in the Bod\'el\'e depression. We found that a crude empirical correction of the 10m wind field in the Bod\'el\'e Depression significantly improve the simulations in terms of occurrence and of intensity.
http://euler.lmd.polytechnique.fr/menut/chimeredust/
A31F-06
Constraining IMPACT Biomass-burning Source Model Predictions with Level 2 Satellite Data Using the AMAPS Distributed Science Network
The IMPACT Model (Integrated Massively Parallel Atmospheric Chemical Transport model) simulates trace gases and aerosols comprehensively throughout the troposphere and stratosphere. IMPACT biomass burning simulations have been verified for consistency with field measurements and retrievals from AVHRR and TOMS. However, these exercises focus almost exclusively on monthly and annual means over global grids. While agreement in monthly mean is required, it is not sufficient to guarantee agreement in distribution of the model generated and observed data. In order to preserve the data distributions of our data sources, we use optical depth and particle properties from MISR and MODIS satellite observations at their native resolutions, rather than grid cell averages. In the past, the task of gathering high resolution (Level 2) satellite data from different instruments and corresponding model output has been prohibitively difficult. To streamline this process, NASA has funded development of the Aerosol Measurement and Processing System (AMAPS). AMAPS uses grid technology combined with special python data structures to access, manipulate and analyze large volumes of aerosol data stored in diverse, remote locations. We compare IMPACT predictions of biomass burning source activity over southern Africa with MISR and MODIS observations of aerosol optical depth, angstrom exponent, and particle optical depth fraction information. We explore the spatial and temporal patterns present in MISR and MODIS data, focusing on land-ocean differences. The IMPACT model is used to make predictions of Angstrom exponent at wavelengths corresponding to MISR and MODIS measurements, and these are compared. In addition, AERONET data may be used to constrain aerosol type and assess differences between column integrated and ground-based properties.
A31F-07
New capabilities for characterizing smoke and dust aerosol over land using MODIS
Smoke and dust aerosol have different chemical, optical and physical properties and both types affect many processes within the climate system. As earth�'¢Ã¢Ã¢â‚¬Å¡Ã‚¬Ãƒ¢Ã¢â‚¬Å¾Ã‚¢s surface and atmosphere are continuously altered by natural and anthropogenic processes, the emission and presumably the effects of these aerosols are also changing. Thus it is necessary to observe and characterize aerosols on a global and climatic scale. While MODIS has been reporting characteristics of smoke and dust aerosol over land and ocean since shortly after Terra launch, the uncertainties in the over-land retrieval have been larger than expected. To better characterize different aerosol types closer to their source regions with greater accuracy, we have developed a new operational algorithm for retrieving aerosol properties over dark land surfaces from MODIS-observed visible (VIS) and infrared (IR) reflectance. Like earlier versions, this algorithm estimates the total loading (aerosol optical depth-$\tau$) and relative weighting of fine (non-dust) and coarse (dust) -dominated aerosol to the total $\tau$ (fine weighting-$\eta$) over dark land surfaces. However, the fundamental mathematics and major assumptions have been overhauled. The new algorithm performs simultaneous multi-channel inversion that includes information about coarse aerosol in the IR channels, while assuming a fine-tuned relationship between VIS and IR surface reflectances, that is itself a function of scattering angle and vegetation condition. Finally, the suite of expected aerosol optical models described by the lookup table have been revised to closer resemble the AERONET climatology, including for smoke and dust aerosol. Beginning in April 2006, this algorithm has been used for forward processing and backward re-processing of the entire MODIS dataset observed from both Terra and Aqua. ''Collection 5'' products were completed for Aqua reprocessing by July 2006 and should be complete for Terra by December 2006. In this study, we used the complete Aqua dataset (July 2002-Aug 2006) and two years of Terra (2005-Aug 2006) data to evaluate the products in regions known to be dominated by smoke and/or dust. We compared with sunphotometer data at selected AERONET sites and found improved $\tau$ retrievals,within prescribed accuracy.
A31F-08
Satellite Lidar and Aerosol Transport Model Synergism: Improved GLAS Aerosol Products
The NASA Geoscience Laser Altimeter System (GLAS) was launched in 2003 aboard the ICESat spacecraft. GLAS provides measurements of ice sheet mass balance and land topography, and global profiles of aerosol and cloud vertical structure. The largest climate forcing uncertainties are associated with aerosol-cloud interactions. Dust and smoke plume impacts are particularly difficult to quantify due to their remote locations and heterogeneous spatial-temporal distributions. GLAS lidar profiles provide global layer height information that is essential for determining where and when aerosol-cloud interactions occur. Quantification of aerosol indirect effects requires the retrieval of aerosol extinction profiles from the GLAS lidar signals. The retrieval is inherently difficult from space, particularly when ancillary data such as optical depth is not available as a constraint. In most situations, the lidar ratio for a particular layer must be assigned a priori to retrieve extinction. Large errors occur when the lidar ratio is improperly assigned. Typical aerosol lidar ratios vary from 20 to 80 sr depending upon composition and particle shape. The current GLAS aerosol product was generated with a rather simplistic lidar ratio assignment algorithm. Here we present a new method of assigning the lidar ratio using aerosol transport models to provide a means of estimating aerosol type. The lidar ratio is then assigned according to the type of aerosol present (based on numerous ground observations). This process also provides a means to assess the model performance, particularly whether a given model layer is observed and at what altitude. Results from several test cases during October and November 2003 will be shown using GLAS data and results from two aerosol transport models: the NASA Goddard Earth Observing System general circulation model and data assimilation system (GEOS-4), and the Navy Aerosol Analysis and Prediction System (NAAPS). Our method provides a means to improve the GLAS aerosol product and also validate model performance (a first step toward eventual lidar-model assimilation schemes applicable to CALIPSO data).
http://glo.gsfc.nasa.gov