Black and brown carbon over central Amazonia: long-term aerosol measurements at the ATTO site

Abstract

Abstract. The Amazon rainforest is a sensitive ecosystem experiencing the combined pressures of progressing deforestation and climate change. Its atmospheric conditions oscillate between biogenic and biomass burning (BB) dominated states. The Amazon further represents one of the few remaining continental places where the atmosphere approaches pristine conditions during occasional wet season episodes. The Amazon Tall Tower Observatory (ATTO) has been established in central Amazonia to investigate the complex interactions between the rainforest ecosystem and the atmosphere. Physical and chemical aerosol properties have been analyzed continuously since 2012. This paper provides an in-depth analysis of the aerosol's optical properties at ATTO based on data from 2012 to 2017. The following key results have been obtained. The aerosol scattering and absorption coefficients at 637 nm, σsp,637 and σap,637, show a pronounced seasonality with lowest values in the clean wet season (mean ± SD: σsp,637=7.5±9.3 M m−1; σap,637=0.68±0.91 M m−1) and highest values in the BB-polluted dry season (σsp,637=33±25 M m−1; σap,637=4.0±2.2 M m−1). The single scattering albedo at 637 nm, ω0, is lowest during the dry season (ω0=0.87±0.03) and highest during the wet season (ω0=0.93±0.04). The retrieved BC mass absorption cross sections, αabs, are substantially higher than values widely used in the literature (i.e., 6.6 m2 g−1 at 637 nm wavelength), likely related to thick organic or inorganic coatings on the BC cores. Wet season values of αabs=11.4±1.2 m2 g−1 (637 nm) and dry season values of αabs=12.3±1.3 m2 g−1 (637 nm) were obtained. The BB aerosol during the dry season is a mixture of rather fresh smoke from local fires, somewhat aged smoke from regional fires, and strongly aged smoke from African fires. The African influence appears to be substantial, with its maximum from August to October. The interplay of African vs. South American BB emissions determines the aerosol optical properties (e.g., the fractions of black vs. brown carbon, BC vs. BrC). By analyzing the diel cycles, it was found that particles from elevated aerosol-rich layers are mixed down to the canopy level in the early morning and particle number concentrations decrease towards the end of the day. Brown carbon absorption at 370 nm, σap,BrC,370, was found to decrease earlier in the day, likely due to photo-oxidative processes. BC-to-CO enhancement ratios, ERBC, reflect the variability of burnt fuels, combustion phases, and atmospheric removal processes. A wide range of ERBC between 4 and 15 ng m−3 ppb−1 was observed with higher values during the dry season, corresponding to the lowest ω0 levels (0.86–0.93). The influence of the 2009/2010 and 2015/2016 El Niño periods and the associated increased fire activity on aerosol optical properties was analyzed by means of 9-year σsp and σap time series (combination of ATTO and ZF2 data). Significant El Niño-related enhancements were observed: in the dry season, σsp,637 increased from 24±18 to 48±33 M m−1 and σap, 637 from 3.8±2.8 to 5.3±2.5 M m−1. The absorption Ångström exponent, åabs, representing the aerosol absorption wavelength dependence, was mostly <1.0 with episodic increases upon smoke advection. A parameterization of åabs as a function of the BC-to-OA mass ratio for Amazonian aerosol ambient measurements is presented. The brown carbon (BrC) contribution to σap at 370 nm was obtained by calculating the theoretical BC åabs, resulting in BrC contributions of 17 %–29 % (25th and 75th percentiles) to σap 370 for the entire measurement period. The BrC contribution increased to 27 %–47 % during fire events under El Niño-related drought conditions from September to November 2015. The results presented here may serve as a basis to understand Amazonian atmospheric aerosols in terms of their interactions with solar radiation and the physical and chemical-aging processes that they undergo during transport. Additionally, the analyzed aerosol properties during the last two El Niño periods in 2009/2010 and 2015/2016 offer insights that could help to assess the climate change-related potential for forest-dieback feedbacks under warmer and drier conditions.