WP11: Improving the accuracy of aerosol light absorption determinations

WP11 Objectives

This research activity focuses on reducing the uncertainty in the determination of the aerosol light absorption by optimizing the advanced remote sensing and in-situ methods adapted to the RI. The specific objectives are:
•    To test novel in-situ instrumentation for reducing uncertainty in the retrieval of the multi-wavelength light absorption and to assess the relationship between absorption and black carbon concentration.
•    To optimize the inversion of remote sensing observations for the derivation of vertically-resolved absorption coefficient and single-scattering albedo (SSA), utilizing the synergy of daytime and night-time photometry and lidar techniques.
•    To evaluate the accuracy of absorption coefficient profiles retrieved by the inversion of remote sensing observations using collocated vertically-resolved in-situ measurements (deploying tethered balloons and UAVs).
•    To utilize all the above in order to reduce the uncertainties associated with the SSA parameter (currently of the order of 0.05), which is one of the most important uncertainties of the aerosol radiative impact.
To establish potential relationships between aerosol sources and light absorption in order to provide trustworthy absorption characteristics for predominant aerosol light absorbers of anthropogenic and natural origin, targeting the needs of the model and satellite communities.

Description of work

Absorption is a key climate-relevant aerosol property of particular importance for aerosol-radiation and aerosol-cloud interactions (IPCC, 2013). Specifically, aerosol absorption can (i) directly modify the global radiation budget, (ii) indirectly modify cloud properties and abundance (e.g. Bond et al., 2013) and (iii) modify the atmospheric stability in the boundary layer and free troposphere (e.g. Babu et al., 2011). However, the magnitude of absorption on the global scale is subject to considerable uncertainties for aerosol particles of both anthropogenic and natural origin (IPCC, 2013). While the AOD and aerosol particle size distribution are relatively well-constrained from measurements, uncertainties in the SSA (e.g. Loeb and Su, 2010), and especially the vertical profile of the black carbon (BC) concentration (e.g. Zarzycki and Bond, 2010), contribute significantly to the overall uncertainties of the aerosol radiative effect. More specifically, the light absorption of BC, which is the predominant anthropogenic absorber, is most probably poorly represented in atmospheric models due to the variability of the BC mass absorption cross section (MAC), which depends on a wide variety of additional variables such as the size, morphology and mixing state of the particles, amount of scattering material, and relative humidity. Furthermore, the absorption of  natural aerosols such as mineral dust, is subject to considerable uncertainties as well since the imaginary part of the refractive index depends crucially on the dust mineralogical phase (e.g. Petzold et al., 2009), affecting the variability of dust absorption globally and resulting in diverse assumptions in atmospheric models and satellite retrievals.
The novel instrumentation operated in the RI can serve for a comprehensive absorption characterization for anthropogenic and natural aerosols, utilizing urban and remote stations of the network. In-situ instrumentation and methods can be used to assess the relationship between absorption and BC concentration at representative sites. The vertical distribution of the absorption coefficient and the SSA, crucial for climatic applications, can be retrieved by both in-situ airborne methods and remote sensing techniques. Closure studies between remote sensing retrievals and in-situ measurements can lead to a better aerosol absorption characterization with respect to the aerosol type and contribute to reducing the related uncertainties. Finally, multi-wavelength absorption in-situ reference methods as well as retrievals can significantly add to improving aerosol characterization, speciation, and source apportionment.
This research activity will therefore utilize the RI potential to perform further developments towards (1) improving in-situ absorption coefficient measurement and reference techniques related to BC, (2) optimizing inversion schemes for the derivation of the absorption coefficient and SSA from remote sensing observations, (3) combining remote sensing with in-situ aerosol profiling to harmonize and validate the different methodologies and reduce absorption uncertainties and (4) consolidating the aforementioned outcomes to provide typical absorption values for different aerosol types/components, targeting the needs of the model and satellite communities which require more accurate absorption and SSA estimates.

Task 11.1: In-situ determination of absorption (PSI, CNRS, NOA, CyI, UGR, TROPOS, CNR, KNMI, JRC)

Evaluation of new instruments for in-situ absorption measurements including reference methods: The CAPS-PMssa monitor is a new promising online instrument simultaneously measuring the extinction and the scattering coefficient by combining cavity attenuated phase shift spectroscopy with an integrating sphere. From this the absorption coefficient can be calculated, which measurement does not suffer from the artefacts that filter-based methods have. It’s potential for suitability as a new absorption reference method for calibration centres as well as for long-term measurements will therefore be evaluated by TROPOS and PSI. Moreover, a prototype of a new light-weight aethalometer for airborne measurements of the light absorption coefficient at several wavelengths has recently been developed. CyI will compare and validate the instrument against existing methods that provide the absorption coefficient. If tests are successful this instrument will then be used in Task 3 for in-situ measurements of vertical profiles deploying tethered balloons or UAVs. Otherwise, the existing miniature BC aethalometer will be used.
Closure between BC mass and light absorption, and determination of mass absorption cross section (MAC): While measurements often focus on the absorption coefficient, models build on the BC mass concentration (as this is the variable that is provided by the emission inventories). However, the MAC value required to link BC mass and light absorption depends on a wide variety of additional variables, such as the size, morphology and mixing state of the particles, amount of scattering material, and relative humidity (Petzold et al., 2013). In addition, it has been suggested in the literature that filter-based light absorption measurements may suffer from condensation of semi-volatile organic compounds. While there are a number of studies reporting MAC values, a comprehensive closure between absorption coefficient and BC concentration and an assessment of the reasons for the variability of the MAC value at different sites are still missing. To close this gap, simultaneous measurements of the BC concentration, size, mixing state and chemical composition will be performed, deploying at least the following instruments: SP2 (single particle soot photometer), CAPS (cavity attenuated phase shift), aethalometer, nephelometer, OC/EC instrument (off-line determination of organic and elemental carbon from filters), size distribution by mobility particle size spectrometers and chemical composition by an AMS (aerosol mass spectrometer) or ACSM (aerosol chemical speciation monitor). The SP2 will be provided either by PSI or CNRS and the CAPS either by PSI or TROPOS and all other measurements are operational at all sites. In addition, where available, these measurements may be complemented by any of the following instruments: PAS (photoacoustic spectrometer), MAAP (multi-angle absorption photometer), PSAP (particle soot absorption photometer), or SP-AMS (soot particle aerosol mass spectrometer). Comparability between sites will be achieved by the fact that absorption coefficient instruments from all sites will be intercompared to each other at TROPOS, and all OC/EC measurements will be performed by the same group (JRC).
These measurements will be performed at the following representative sites: Melpitz (TROPOS), Bologna (CNR), Cabauw (KNMI/TNO), Athens (NOA), Finokalia (NOA) and Granada (UGR). These sites have been selected for their different aerosol typologies (three sites close to the sea and three more continental sites, with two sites in Northern and four sites in Southern Europe). The availability of the mandatory instrumentation as listed above was used as an additional criterion. This data will provide a better understanding of the links between BC and absorption coefficient, along with an improved characterization of the MAC and its atmospheric variability. The variability of the MAC values will also be discussed on the basis of available source apportionment data for BC at these sites (e.g. through the wavelength dependence of the absorption coefficient (Sandradewi et al., 2008)).

Task 11.2: 24-hour absorption coefficient profiling through inversion of remote sensing observations (UGR, CNR, CNRS, UNIVLEEDS)

In the frame of ACTRIS, LOA has developed the well-known GARRLiC algorithm (Generalized Aerosol Retrieval from Radiometer and Lidar Combined, Lopatin et al., 2013) for daytime microphysical retrievals combining sunphotometer and lidar remote sensing observations. This task will focus on implementing the lidar stand-alone night-time retrievals (Müller et al., 1999) using as a constraint night-time measurements from lunar/star photometers (Barreto et al., 2013) and adapting the GARRLiC algorithm for night-time retrievals, aiming at a 24-hour aerosol absorption characterization.  
GARRLiC combines the sunphotometer measurements of the sky radiance and optical thickness with the multi-wavelength elastic lidar measurements, for the retrieval of multi-spectral absorption profiles, along with an extended product suite of aerosol microphysical properties. Based on the GARRLiC approach, LOA will develop a new algorithm for the night-time retrievals of the aerosol absorption coefficient and SSA profiles, combining the AODs from the lunar/star photometers with the elastic lidar measurements. The night-time retrievals can be augmented further from the nearest daytime retrievals from the original GARRLiC, which can be used as a-priori constraints limiting the aerosol variability in time.
Hertfordshire will combine the stand-alone lidar retrievals with lunar/star photometer AODs: the night-time measurements of the extinction and backscatter coefficients at multiple wavelengths from Raman lidar measurements will be utilized together with multi-wavelength lunar/star AOD constraints to provide an estimation of the main particle microphysical parameters with few a-priori assumptions. Furthermore, although this technique is more successful with the so-called 3+2 Raman lidar systems (named after the 3+2 available wavelengths for backscatter and extinction), the inclusion of lunar/star photometric AODs can potentially result in acceptable retrievals for Raman lidar systems employing less operating channels (i.e., 2+1 or 1+1, 3+0) at less advanced lidar sites.
The advanced night-time inversions will be initially applied utilizing the lunar/star photometer synergy with multi-wavelength Raman lidars at the ACTRIS sites in Potenza (CNR) and Granada (UGR). Special focus will be given to lunar/star photometry and related instrumentation/methods. Different methodologies applied for lunar photometry will be tested for accurately determining the night-time AOD through system intercomparisons which will be performed in Potenza (CNR). Star-photometry solutions will be tested and characterized in Granada. The different technical solutions will be compared and integrated with EARLINET multi-wavelength Raman lidars sited in Potenza and Granada.
The task also aims at extending the 24-hour aerosol absorption profile methodology to a large number of ACTRIS stations. For this purpose, the lunar photometer of Potenza will be transported to Melpitz, Athens and Finokalia to participate in the validation campaigns of Task 3.

Task 11.3: Closure studies between remote sensing and in-situ absorption retrievals for establishing an absorption model for characteristic aerosol types (NOA, CyI, CNRS, UGR, TROPOS, PSI)

This task aims at integrating the in-situ techniques employed in Task 1 (Lead PSI) with the remote sensing retrievals of the aerosol absorption profiles of Task 2 (Lead UGR), in order to produce a representative aerosol absorption model for climate studies.
First, the aerosol absorption coefficient and SSA profile retrievals from combined lidar and sun/lunar/star photometer measurements will be compared to a number of different in-situ techniques for the ACTRIS sites in Melpitz (TROPOS), Athens (NOA), Finokalia (NOA), and Granada (UGR): Tethered balloons will be employed at Melpitz and Athens, measuring the absorption and scattering coefficients, BC concentration and particle size with multi-wavelength aethalometers, nephelometers, and particle sizers. The tethered balloons will be launched to 1-1.5 km above ground level (agl), high enough to cover lidar height ranges. The unmanned aerial vehicles (UAVs) of CyI will be deployed above Athens, Finokalia and Granada. These UAVs are capable of reaching up to 3 km in altitude depending on the payload. They will carry two miniature BC aethalometers, an eight channel optical particle counter (OPC) with a particle size range of 0.3 μm to 10 μm (or larger after customization) and a condensation particle counter (CPC) for particles larger than 10 nm. Moreover, the new light-weight aethalometer for airborne measurements will be installed on the UAVs for multi-wavelength absorption retrievals if successfully validated in Task 1.  TROPOS will deploy UAVs (developed by the Technical University of Braunschweig) above Melpitz. These UAVs are equipped with an OPC (six channels > 300 nm), a CPC (>10 nm), and a mini-aethalometer (1 wavelength). They are capable of carrying a payload of 2.8 kg and perform measurements from ground to 1000 m agl above the measurement site. Finally, the station of the Sierra Nevada mountains providing in-situ measurements at 2.5 km asl will be used for evaluating inversions over the Granada lidar/photometric station.
The sites used for the comparison of the aerosol absorption profiles from remote sensing and in-situ measurements have been selected based on the following criteria: (i) different aerosol typologies, (ii) availability of state of the art remote sensing instruments (i.e., Raman lidars and lunar/star photometers) and (iii) capacity and possibility to fly UAVs or deploy tethered balloons.
The dataset that will be collected will be analysed to assess the uncertainty of the inversion retrievals of the absorption coefficient and the SSA. These optimization schemes will be studied by LOA and NOA for their consistency with the in-situ retrievals. Differences between the in-situ and remote sensing methods regarding the characteristics of the measured aerosol samples will be considered and minimized in the frame of the closure experiments (e.g. dry in-situ measurements against ambient remote sensing retrievals).
The data on the detailed vertical profiles will be offered to modellers worldwide, for validation of model results.