Forest fire smoke layers observed in the free troposphere over Portugal with a multiwavelength Raman lidar: optical and microphysical properties

Abstract

Vertically resolved optical and microphysical properties of biomass burning aerosols, measured in 2011 with a multiwavelength Raman lidar, are presented. The transportation time, within 1-2 days (or less), pointed towards the presence of relatively fresh smoke particles over the site. Some strong layers aloft were observed with particle backscatter and extinction coefficients (at 355 nm) greater than 5 Mm(-1)sr(-1) and close to 300 Mm(-1), respectively. The particle intensive optical properties showed features different from the ones reported for aged smoke, but rather consistent with fresh smoke. The Ångström exponents were generally high, mainly above 1.4, indicating a dominating accumulation mode. Weak depolarization values, as shown by the small depolarization ratio of 5% or lower, were measured. Furthermore, the lidar ratio presented no clear wavelength dependency. The inversion of the lidar signals provided a set of microphysical properties including particle effective radius below 0.2 μm, which is less than values previously observed for aged smoke particles. Real and imaginary parts of refractive index of about 1.5-1.6 and 0.02i, respectively, were derived. The single scattering albedo was in the range between 0.85 and 0.93; these last two quantities indicate the nonnegligible absorbing characteristics of the observed particles.

Figure 1

Fire hot spots detected by MODIS on board the Terra and Aqua satellites in the period from 15 to 18 October 2011 (http://firms.modaps.eosdis.nasa.gov/firemap/). 72-hour backward trajectories arriving at Évora (black square) at the lidar measurement times, between 17 and 19 October 2011, at 2, 2.5, 3, 3.5, and 4 km agl. For 18 and 19 October 2011 the colors are the same as for 17 October 2011.

Figure 2

(a) Terra-Modis aerosol optical depth at 550 nm and (b) Terra-Modis Ångström exponent (470/660 nm) on 18 October 2011.

Figure 3

Profiles of (a) backscatter coefficients (355-blue, 532-green, and 1064 nm-red), (b) extinction coefficients (355-blue and 532 nm-green), (c) backscatter and extinction-related Ångström exponent (wavelength pair 355/532 nm), (d) lidar ratios (355 and 532 nm), and (e) particle linear depolarization ratio (532 nm) measured on (a) 17 October 2010 20:20–22:10 UTC.

Figure 4

Profiles of (a) backscatter coefficients (355-blue, 532-green, and 1064 nm-red), (b) extinction coefficients (355-blue and 532 nm-green), (c) backscatter and extinction-related Ångström exponent (wavelength pair 355/532 nm), (d) lidar ratios (355 and 532 nm), and (e) particle linear depolarization ratio (532 nm) measured on 18 October 2011 at 19:00–20:45 UTC.

Figure 5

Profiles of (a) backscatter coefficients (355-blue, 532-green, and 1064 nm-red), (b) extinction coefficients (355-blue and 532 nm-green), (c) backscatter and extinction-related Ångström exponent (wavelength pair 355/532 nm), (d) lidar ratios (355 and 532 nm), and (e) particle linear depolarization ratio (532 nm) measured on 18 October 2010 21:45–22:15 UTC.

Figure 6

Profiles of (a) backscatter coefficients (355-blue, 532-green, and 1064 nm-red), (b) extinction coefficients (355-blue and 532 nm-green), (c) backscatter and extinction-related Ångström exponent (wavelength pair 355/532 nm), (d) lidar ratios (355 and 532 nm), and (e) particle linear depolarization ratio (532 nm) measured on 19 October 2010 00:00–02:00 UTC.

Figure 7

Relative frequency distributions of (a) backscatter and extinction related Ångström exponents (355/532 nm) and (b) lidar ratios at 355 and 532 nm, considering the four different periods of measurements on 17, 18, and 19 October 2011.

Figure 8

Extinction-related Ångström exponent and effective radius of forest fires smoke, based on Raman lidar measurements versus transport time. The results from the different cases measured at Évora are shown (crosses) as well as the average values (open squares). The average values from Alados-Arboledas et al. [27] are also shown (circles). The thick lines denote the first-order exponential fitting curves and the shaded areas represent the possible ranges for the parameterization according to the fitting parameters and respective fitting errors from Müller et al. [17].