Surface shortwave radiative forcing of different aerosol types in the central Mediterranean

Transcription

1 Click Here for Full Article GEOPHYSICAL RESEARCH LETTERS, VOL. 35, L02714, doi: /2007gl032395, 2008 Surface shortwave radiative forcing of different aerosol types in the central Mediterranean A. di Sarra, 1 G. Pace, 2 D. Meloni, 1 L. De Silvestri, 1 S. Piacentino, 3 and F. Monteleone 4 Received 18 October 2007; revised 6 December 2007; accepted 28 December 2007; published 26 January [1] Ground based measurements of aerosol optical depth, t, and shortwave irradiance at the Mediterranean island of Lampedusa during 2003 and 2004 were used to estimate the surface aerosol shortwave radiative forcing. The shortwave forcing efficiency (FE) was derived at various solar zenith angles, q, as the derivative of the shortwave irradiance with respect to t. Values of FE for different classes of particles, namely desert dust, DD, biomass burning/industrial aerosols, BU, and for the whole dataset are derived. At the summer solstice the daily average FE is 86.4 W/m 2 for DD, 70.5 W/m 2 for BU, and 94.0 W/m 2 for the whole dataset. The daily aerosol forcing of DD is much larger than for the other aerosol classes due to the combination of larger forcing efficiency and largest optical depths. The estimated average daily forcing at the summer solstice and equinox for DD is 30 and 24 W/m 2, respectively. Citation: di Sarra, A., G. Pace, D. Meloni, L. De Silvestri, S. Piacentino, and F. Monteleone (2008), Surface shortwave radiative forcing of different aerosol types in the central Mediterranean, Geophys. Res. Lett., 35, L02714, doi: /2007gl Introduction [2] The climate forcing determined by the aerosol represents one of the main uncertainties in climate both at regional and global level. In the last years a great effort has been dedicated to improve the estimates of the aerosol direct forcing, through dedicated measurement campaigns and integrated analyses [e.g., Yu et al., 2006]. Large uncertainties still exist, due to the large spatial and temporal variations of aerosol properties. Different methods, based on measurements, models, and on the combined use of both, have been applied in the quantification of the aerosol forcing. In this study we use the method developed by Satheesh and Ramanathan [2000] based only on observational data to quantify the direct surface aerosol radiative forcing. We use simultaneous measurements of column aerosol optical properties and surface shortwave radiative fluxes obtained during 2003 and 2004 in Lampedusa, a 1 Dipartimento Ambiente, Cambiamenti Globali e Sviluppo Sostenibile, Ente per le Nuove Tecnologie, l Energia e l Ambiente, Rome, Italy. 2 Dipartimento Ambiente, Cambiamenti Globali e Sviluppo Sostenibile, Ente per le Nuove Tecnologie, l Energia e l Ambiente, Bologna, Italy. 3 Dipartimento Ambiente, Cambiamenti Globali e Sviluppo Sostenibile, Ente per le Nuove Tecnologie, l Energia e l Ambiente, Lampedusa, Italy. 4 Dipartimento Ambiente, Cambiamenti Globali e Sviluppo Sostenibile, Ente per le Nuove Tecnologie, l Energia e l Ambiente, Palermo, Italy. Copyright 2008 by the American Geophysical Union /08/2007GL032395$05.00 small island in the Southern sector of the central Mediterranean sea. The adopted method allows the determination of the forcing produced by different classes of particles. 2. Instruments and Data Analysis [3] This study is based on measurements obtained at Lampedusa during 2003 and 2004 with a multi filter rotating shadow band radiometer (MFRSR) and two pyranometers, a Kipp and Zonen CM-11 and an Eppley PSP (Precision Spectral Pyranometer). Lampedusa is a small island (22 km 2 surface area, maximum elevation of 130 m) in the Southern Mediterranean (35.5 N, 12.6 E). The instrumentation is installed at the ENEA Station for Climate Observations, located on a 40 m high plateau on the North- Eastern coast of Lampedusa. The instruments have the horizon free of significant obstacles. The MFRSR [Harrison et al., 1994] measures global and diffuse irradiance at six 10 nm wide channels, and at one broadband channel ( nm), with a sampling rate of 15 s. Data are averaged over 1 minute and stored for analysis. The direct irradiances at 416, 496, 615, 672 and 869 nm are calculated as differences between global and diffuse irradiances and are used to derive the aerosol optical depth, t, which is calculated applying the Beer-Lambert law and subtracting the contributions of Rayleigh scattering and ozone absorption to the atmospheric optical depth. A detailed description of the retrieval of t and of the measurement errors is given by Pace et al. [2006]. The Ångström exponent, a, defined as the negative slope of t versus l in logarithmic scale, is calculated from the values of t at 496 and 869 nm. The CM- 11 has a uniform spectral responsivity throughout the region micron. It is installed on an automatic weather station operational at Lampedusa since 1999, and its signal is acquired every 60 seconds. The CM-11 has not been calibrated since its installation. [4] From June 2003 to June 2004 a ventilated PSP has been installed for periods of 3 5 days, every 2 months, on the roof of the laboratory, about 30 m from the weather station and the CM-11. Starting from June 2004, the PSP has operated continuously. It has a uniform responsivity between 0.3 and 2.8 micron; its signal is sampled every 30 seconds. The PSP was calibrated at Eppley in August 2002, and has not been used before the installation at Lampedusa. It was re-calibrated at the World Radiation Centre at Davos, Switzerland, in The instrument sensitivity changed by less than 6% from 2002 to The PSP observations were used to derive an updated calibration for the CM-11 and to verify its performance. The small difference in the spectral responsivity interval of the two pyranometers produces negligible effects on the L of5

2 removes cloudy periods. Data were further screened eliminating data with a relatively large variability in t (as described by Pace et al. [2006]) and in irradiance (see section 3.2). A residual contamination from very thin clouds may still remain, and may affect cases of very low optical depth. These cases constitute a minor fraction of the used data. 3. Determination of the Aerosol Surface Forcing [7] The surface radiative forcing of a given atmospheric constituent is the difference between the observed, F net, and the pristine net radiative flux, F p net. The surface net flux is defined as the difference between downward and upward irradiances at the Earth s surface. The pristine flux is the one occurring in the absence of the atmospheric constituent being considered. The net flux is usually calculated assuming a modelled or a measured value of the surface albedo, A, and is defined as F net ¼ ð1 AÞI m ð1þ where I m is the measured downward irradiance. The shortwave surface aerosol radiative forcing RF is defined as follows: RF ¼ F net F p net ¼ ð1 AÞ I m I p ð2þ Figure 1. Behavior of the shortwave surface net flux versus aerosol optical depth at solar zenith angle of 30 and 60 for different classes of particles (see text). Least square fits are also shown for the different classes. measurements, and the two signals were compared without corrections for the spectral response differences. The method used in this study is weakly sensitive on the absolute irradiance scale, and the CM-11 data from 2003 and 2004 have been scaled to the PSP irradiance scale of year 2002 reduced by 3%. [5] The CM-11 corrected irradiances were compared with PSP irradiances in the available 58 days with simultaneous measurements. The two datasets agree within about ±2%. Beside differences in the instrumental cosine response and spectral responsivity, also the cleanness of the pyranometer dome affects measured irradiances. The domes are regularly cleaned by an operator approximately once a week. However, the CM-11 is not ventilated, and its signal is subject to a stronger influence from dew, rain, or intense desert dust deposition than PSP, and part of the deviation between the two pyranometers may be due to this effect. Considering the measurement uncertainty of PSP and the uncertainty associated with the transfer of the calibration, we estimate the measurement error of the CM-11 to be ±2.5%. Before further analysis, all irradiances were reported at the mean Sun-Earth distance. [6] Only cloud-free intervals may be used for this analysis. We applied the cloud screen algorithm described by Meloni et al. [2007]. It uses MFRSR broadband global and diffuse irradiances to identify cloud-free conditions in a large fraction of the sky. The algorithm was verified against visual observations and total sky images, and effectively where the fluxes are in the shortwave spectral range, and I p is the downward irradiance in the absence of aerosols. Different methods are used to estimate RF. In this study we use the direct method [Satheesh and Ramanathan, 2000; Bush and Valero, 2002] to derive the forcing efficiency FE, i.e. the radiative forcing produced by aerosols with an optical depth equal to 1. FE is calculated at fixed solar zenith angle as the derivative of F net (q) with respect to t at 496 nm, FEðÞ¼dF q net ðþ=dt q ðþ q The value of RF is derived multiplying FE by the corresponding value of t. Measurements of I m and t, and an estimate of A are needed to calculate F net ; the knowledge of F net and t for pristine conditions is not required. F net depends on the surface albedo and on the water vapor amount, and reliable values for these two quantities are needed in the analysis. It is worth emphasizing that the precision of the instrument is more critical then its absolute accuracy in the application of this method. The uncertainty on the retrieved FE thus depends on the instrumental error and on the variability of surface albedo and water vapor columnar content, wvc. In the next sections we will examine the role played by surface albedo and water vapor in the retrieval of FE Albedo and Water Vapor Content [8] FE depends linearly on the albedo (see expressions 1 and 2). Lampedusa is a small and rocky island with a surface of 22 km 2 and sparse vegetation, especially from May to October, when the precipitation is very low. The albedo was calculated as the weighted average of land and ocean albedo. During 2003 and 2004 the land albedo ð3þ 2of5

3 Figure 2. Evolution of the surface aerosol shortwave forcing efficiency versus solar zenith angle for the considered aerosol classes (see text). The forcing efficiency is calculated with respect to the optical depth at (a) 496 nm and (b) 671 nm. measured by MODIS at Lampedusa has a minimum at 0.14 in winter, and a peak at 0.22 in late summer and autumn. In this study we use data acquired in the period 20 May 10 November, when A varies between approximately 0.18 and The ocean albedo is calculated according to Briegleb et al. [1986], and depends on q. [9] The land and ocean albedo are weighted by the fraction of surface occupied by land and ocean within a circle of 5 km radius. The size relevant for the determination of the albedo is therefore assumed of the order of 10 km, as discussed by Charlock et al. [2003]. Changes of land albedo between 0.18 and 0.22 have a very limited impact on the weighted albedo, and a fixed value of 0.2, independently of the solar zenith angle, was used. The calculated albedo ranges from to 0.17, these values corresponding to q = 20 and 75, respectively. The station at Lampedusa faces the open sea surface to the East. Thus, data collected during the morning are strongly affected by the Sun glint, and a more sophisticated treatment of the surface albedo would be required to correctly estimate the net fluxes. Consequently, only afternoon data, for which we have a reduced influence from Sun glint, are used in the analysis. [10] Changes in water vapor column, wvc, produce a change in I m, due to absorption by water vapor in the shortwave range. Water vapor column averages for years 2003, 2004, and over the period at Lampedusa, were derived from the NCEP/NCAR reanalysis dataset. The wvc annual variation is about 1.2 cm, with wvc ranging from 1.48 in February to 2.72 in August. In the period of investigation (May November), the monthly average wvc is comprised between 2.0 and 2.7 cm. Starting from 2005, the wvc has been derived at Lampedusa from MFRSR measurements at 940 nm with a 1-minute time resolution. In 2005 the water vapor column spans between 1.2 and 2.9 cm during summer and autumn. Short-term (within few hours) changes are always smaller than 0.5 cm. Values of wvc derived from MFRSR observations in September October 2005 were compared to those derived from NCEP/NCAR reanalyses. The linear correlation coefficient between the two datasets is 0.77; the bias is 0.28 cm (NCEP overestimates MFRSR), while the root mean square deviation between the two datasets is 0.13 cm. The MFRSR retrieval is based on a calibration with local radiosonde profiles, and is totally independent of the NCEP reanalysis. [11] The surface irradiance dependency on wvc has been studied using the SBDART radiative transfer model [Ricchiazzi et al., 1998]. Model calculations of surface shortwave irradiance at q = 20, 40, 60, 75 for different values of wvc were performed. A 1.0 cm increase (decrease) in wvc around 2.5 cm produces a 2% decrease (increase) in irradiance. The variation is slightly larger at q = 75. Thus, the effect of the expected short-term wvc variability on the surface shortwave irradiance is smaller than ±2%. Due to the lack of continuous wvc observations during 2003 and 2004, wvc variations are taken into account by including an additional 2% random error on the measured irradiances Determination of Surface Aerosol Forcing Efficiency [12] Data corresponding to the period 20 May 10 November, i.e. in late spring, summer, and early autumn, were used in the analysis. In this time interval the frequency of cloud-free conditions is high, and the variability in water vapor and surface albedo is relatively small. [13] The FE is calculated at q =20, 30, 40, 50, 60, 65,70,75, as the derivative of the average F net, F net, over intervals of ±1.5 solar zenith angle, with respect to the corresponding mean value of t. To remove possible residual contamination by clouds, the ratio R between the standard deviation of I m and the average I m was calculated, and cases with R > 0.01 are discarded. [14] Few outliers, i.e. those data points whose distance from the fitting curve exceeds two times the standard deviation of I m at each value of q, were removed. The fitting line was recalculated after removal of the outliers. [15] The uncertainty on FE was estimated as the error associated with the least square linear fit of the data. The propagation formula was used to determine the uncertainty on F net (q). The uncertainty on q was neglected, while the total uncertainty on the average I m (q) was calculated by taking into account the uncertainties on the CM-11 measurements (2.5%), the expected influence of wvc variability (2%), and the standard deviation of the average I m (q) (<1%). 3of5

4 Table 1. Daily Average Values of FE at the Summer Solstice and the Equinox for Different Aerosol Classes a Aerosol Type Daily Forcing Efficiency, Summer Solstice, W/m 2 Daily Forcing Efficiency, Equinox, W/m 2 Average Optical Depth at 496 nm Whole dataset 94.0 ± ± ± 0.14 DD 86.4 ± ± ± 0.14 BU 70.5 ± ± ± 0.11 a The averages and standard deviations of t for the three classes are also reported. The standard deviation on t is an indication of its variability. [16] The retrieved values of FE at different solar zenith angles were fitted with a third degree polynomial as a function of q (FE = 0 for q =90 is assumed in the fit). The uncertainties associated with FE(q) were linearly fitted as a function of q. The polynomial was integrated over 24 hours, taking into account the evolution of q with time, to calculate the daily forcing efficiency. The daily average forcing was obtained multiplying the daily average FE by the daily average t. The uncertainty on the daily FE was estimated taking into account the estimated error on FE. 4. Results and Discussion [17] Pace et al. [2006] developed a method to discriminate among different aerosol types on the basis of their optical properties and of trajectory analysis. Applying this method, the aerosol observations at Lampedusa were grouped into three classes, corresponding to cases dominated by desert dust, DD, biomass burning and urban/industrial aerosols, BU, and to all the observed data. The last class includes all BU and DD cases, as well as marine, continental, and mixed aerosol types. The irradiance measurements were grouped according to the corresponding aerosol properties, and FE(q) was determined separately for each class. Figure 1 shows the behavior of the measured irradiances at q =30 and 60, for the three classes. The derived linear fits are also displayed. [18] The possible presence of very thin clouds not detected by the cloud screening algorithm may affect the estimates of FE. As far as the cloud effect can be schematized as a constant offset in t and irradiance, it does not affect the retrieved value of FE. If thin clouds are present occasionally, the main effect on the irradiance-t relationship is an increase in the spread of the data, which, because of the small values of t, is estimated to produce a negligible effect on the retrieved FE. [19] Figure 2a shows the behavior of FE at different solar zenith angles. The FE of all aerosol types generally decreases for increasing q. The FE of the BU aerosol is smaller than DD at all solar zenith angles, except for q > 65. The forcing efficiencies of DD and of the whole dataset are very similar, except at q >60. This effect is primarily due to the influence of DD cases within the dataset: DD cases constitute the large majority of data with large optical depth, and play a large role in determining the slope of the fitting line, i.e. FE. Analyzing aerosol data for the period July 2001 September 2003 at Lampedusa, Pace et al. [2006] showed that the average Ångström exponent is 0.15 for DD, 1.77 for BU, and 0.86 for the whole dataset; the values of single scattering albedo, SSA, of DD and BU particles also differ largely [Meloni et al., 2006], indicating that large differences in size distribution and composition exist among the different classes. Large particles are dominant in DD, and produce an almost spectrally neutral behavior of t. Small particles are dominant in BU, thus producing an aerosol optical depth which rapidly decreases with wavelength. For overhead Sun, the downward radiative power between 300 and 680 nm is about 50% of the total power in the shortwave range. Thus, the optical depth at about 700 nm is crucial in determining the aerosol radiative effect, and the forcing efficiency was also calculated with respect to the aerosol optical depth at 671 nm (i.e. close to the median wavelength for the shortwave spectrum). The results are shown in Figure 2b. The FE for DD and for the whole dataset change slightly when the optical depth at 671 nm is used, while very large changes occur for BU. This effect is primarily due to the large Ångström exponent of BU. The size distribution thus appears as the main parameter determining the different forcing of DD and BU aerosols. [20] The observed dependency of FE on the solar zenith angles is produced by the combination of different processes. The larger q, the larger is the fraction of diffuse radiation with respect to the total. At the same time, as q increases, radiation at the shorter wavelengths is reduced more strongly than at long wavelengths, thus modifying the shape of the solar spectrum. DD and BU display a different wavelength dependence of the single scattering albedo: SSA is lowest at short wavelengths for DD, while is lowest in the red portion of the spectrum for BU [e.g., Meloni et al., 2006]. The combined change in solar spectrum with q, and the dependence of the SSA on wavelength, may, at least in part, explain the overlap of the FE for DD and BU at q 70. [21] Differences in aerosol phase function, related to the particles size, and vertical distribution probably also play a role. The reduced absorption by DD at large solar zenith angles (due to the reduction of radiation at short wavelengths reaching the ground) may partly explain why the FE for DD becomes smaller than FE for the whole dataset at q 60. [22] Table 1 shows the retrieved values of the daily average (over 24 hours) FE for the different aerosol classes at the summer solstice and at the equinox. The average aerosol optical depth for the different aerosol classes in the period of investigation is also reported. The surface radiative effect of DD is the largest, because of both stronger efficiency and largest average optical depth. Using the average optical depth for each class, the daily average forcing at the equinox is about 24 W/m 2 for DD, 13 W/m 2 for BU, and 17 W/m 2 for the whole dataset. It must be emphasized that these results are relative to a low albedo surface (essentially determined by the ocean). The relative importance of the different aerosol classes may change at the top of the atmosphere, and over different surfaces. [23] Estimates of the surface aerosol FE were reported in previous studies, and span a relatively large range of values. Most of the studies carried out so far are relative to field campaigns of shorter duration compared to our analysis. Different methods to retrieve the FE are applied. Recently, Yu et al. [2006] have summarized the actual knowledge on the aerosol direct radiative effects derived from ground 4of5

5 based measurements, satellite observations, and model calculations. [24] In the Mediterranean a limited number of studies on the aerosol direct shortwave forcing at the surface have been carried out. During the Mediterranean Intensive Oxidant Study (MINOS), using measurements carried out at Crete, Greece, Markowicz et al. [2002] estimated a mean diurnal forcing efficiency of 87.9 W/m 2 for anthropogenic aerosols under the influence of fires, and 70.7 W/m 2 without fire particles. Formenti et al. [2002] derived a daily forcing efficiency of 64 W/m 2 for aged biomass burning particles in the Aegean sea. [25] Roger et al. [2006] derived a daily forcing efficiency of 107 W/m 2 in Southern France during the Expérience sur Site pour Contraindre les Modèles de Pollution et de Transport d Emission experiment for anthropogenic particles. Derimian et al. [2006] derived estimates of daily FE for desert dust and polluted aerosols in Israel. They found values of 86 W/m 2 for dust and 81 W/m 2 for polluted particles in the Negev desert. [26] Values of the daily FE of dust were derived in different regions by several authors. The retrieved values range between 60 and 94 W/m 2, depending on the possible presence of other aerosol types, and on the geographical location [e.g., Kim et al., 2005; Markowicz et al., 2003]. 5. Conclusions [27] Surface aerosol radiative forcing efficiencies were determined using measurements of shortwave fluxes and aerosol optical properties at Lampedusa in the period May November of 2003 and The forcing efficiency was calculated separately for desert dust, for industrial/urban/ biomass burning particles, and for the whole dataset, comprising different aerosol types. The aerosol FE was calculated as a function of the solar zenith angle, and estimates of the daily average forcing efficiency at the summer solstice and at the equinox were derived. The largest FE is found for the whole dataset, followed by desert dust and industrial/urban/biomass burning particles. Due to the characteristics of the size distribution, the dust optical depth is generally large throughout the shortwave spectral range, leading to large values of the DD forcing efficiency. Desert dust produces the largest forcing, due to the high value of both FE and t in the central Mediterranean. The average forcing at the equinox is about 24 W/m 2 for DD, 13 W/m 2 for BU, and 17 W/m 2 for the whole dataset. [28] Acknowledgment. This work has been partially supported by the Aeroclouds project, funded by the Italian Ministry for University and Research. References Briegleb, B. P., P. Minnis, V. Ramanathan, and E. Harrison (1986), Comparison of regional clear-sky albedos inferred from satellite observations and model computations, J. Clim. Appl. Meteorol., 25, Bush, B. C., and F. P. J. Valero (2002), Spectral aerosol radiative forcing at the surface during the Indian Ocean Experiment (INDOEX), J. Geophys. Res., 107(D19), 8003, doi: /2000jd Charlock, T. P., F. G. Rose, and D. A. Rutan (2003), Validation of the archived CERES surface and atmosphere radiation budget at SGP, paper presented at the Thirteenth Atmospheric Radiation Measurement Science Team Meeting, Broomsfield, Colo., 31 Mar. 4 Apr. Derimian, Y., A. Karnieli, Y. J. Kaufman, M. O. Andreae, T. W. Andreae, O. Dubovik, W. Maenhaut, I. Koren, and B. N. Holben (2006), Dust and pollution aerosols over the Negev desert, Israel: Properties, transport, and radiative effect, J. Geophys. Res., 111, D05205, doi: / 2005JD Formenti, P., et al. (2002), STAAARTE-MED 1998 summer airborne measurements over the Aegean Sea: 2. Aerosol scattering and absorption, and radiative calculations, J. Geophys. Res., 107(D21), 4551, doi: / 2001JD Harrison, L., J. Michalsky, and J. Berndt (1994), Automated multifilter rotating shadowband radiometer: An instrument for optical depth and radiation measurements, Appl. Opt., 33, Kim, D.-H., B. J. Sohn, T. Nakajima, and T. Takamura (2005), Aerosol radiative forcing over east Asia determined from ground-based solar radiation measurements, J. Geophys. Res., 110, D10S22, doi: / 2004JD Markowicz, K. M., P. J. Flatau, M. V. Ramana, P. J. Crutzen, and V. Ramanathan (2002), Absorbing mediterranean aerosols lead to a large reduction in the solar radiation at the surface, Geophys. Res. Lett., 29(20), 1968, doi: /2002gl Markowicz, K. M., P. J. Flatau, P. K. Quinn, C. M. Carrico, M. K. Flatau, A. M. Vogelmann, D. Bates, M. Liu, and M. J. Rood (2003), Influence of relative humidity on aerosol radiative forcing: An ACE-Asia experiment perspective, J. Geophys. Res., 108(D23), 8662, doi: / 2002JD Meloni, D., A. di Sarra, G. Pace, and F. Monteleone (2006), Optical properties of aerosols over the central Mediterranean: 2. Determination of single scattering albedo at two wavelengths for different aerosol types, Atmos. Chem. Phys., 6, Meloni, D., A. di Sarra, G. Biavati, J. J. DeLuisi, F. Monteleone, G. Pace, S. Piacentino, and D. Sferlazzo (2007), Seasonal behavior of Saharan dust events at the Mediterranean island of Lampedusa in the period , Atmos. Environ., 41, Pace, G., A. di Sarra, D. Meloni, S. Piacentino, and P. Chamard (2006), Optical properties of aerosols over the central Mediterranean: 1. Influence of transport and identification of different aerosol types, Atmos. Chem. Phys., 6, Ricchiazzi, P., S. Yang, C. Gautier, and D. Sowle (1998), SBDART: A research and teaching software tool for plane-parallel radiative transfer in the Earth s atmosphere, Bull. Am. Meteorol. Soc., 79, Roger, J. C., M. Mallet, P. Dubuisson, H. Cachier, E. Vermote, O. Dubovik, and S. Despiau (2006), A synergetic approach for estimating the local direct aerosol forcing: Application to an urban zone during the Expérience sur Site pour Contraindre les Modèles de Pollution et de Transport d Emission (ESCOMPTE) experiment, J. Geophys. Res., 111, D13208, doi: /2005jd Satheesh, S. K., and V. Ramanathan (2000), Large differences in tropical aerosol forcing at the top of the atmosphere and Earth s surface, Nature, 405, Yu, H., et al. (2006), A review of measurement-based assessments of the aerosol direct radiative effect and forcing, Atmos. Chem. Phys., 6, L. De Silvestri, A. di Sarra, and D. Meloni, Dipartimento Ambiente, Cambiamenti Globali e Sviluppo Sostenibile, Ente per le Nuove Tecnologie, l Energia e l Ambiente, Via Anguillarese 301, I Rome, Italy. (disarra@casaccia.enea.it) F. Monteleone, Dipartimento Ambiente, Cambiamenti Globali e Sviluppo Sostenibile, Ente per le Nuove Tecnologie, l Energia e l Ambiente, Via Catania 2, I Palermo, Italy. G. Pace, Dipartimento Ambiente, Cambiamenti Globali e Sviluppo Sostenibile, Ente per le Nuove Tecnologie, l Energia e l Ambiente, Via Martiri Monte Sole 4, I Bologna, Italy. S. Piacentino, Dipartimento Ambiente, Cambiamenti Globali e Sviluppo Sostenibile, Ente per le Nuove Tecnologie, l Energia e l Ambiente, Contrada Capo Grecale, I Lampedusa, Italy. 5of5