Composition Dependence of Stratospheric Aerosol Shortwave Radiative Forcing in Northern Midlatitudes

Abstract

Both observations and recent model results reveal a significant organic component in stratospheric aerosols. However, intrinsic optical properties (i.e., complex refractive index) of this organic component and the mixing state of organic and ubiquitous sulfuric acid components are quite uncertain. We examine the effects of different complex refractive indices and particle mixing states on shortwave radiative forcing (RF) of stratospheric aerosols in northern midlatitudes. In the absence of large volcanic sulfur emissions, our calculations show that organic components may have substantial impacts on stratospheric aerosol optical depth (AOD). Compared to pure sulfuric acid/water aerosol, organic-containing aerosols could cause ±100% change in shortwave RF (for low AOD conditions) depending on the refractive index and mixing state. The range found here of shortwave RF results, for different scenarios of organic complex refractive index and mixing state, call for better understandings of chemical and transport processes determining aerosol optical properties in the stratosphere.

Key Points

• Organics may have significant impacts on the stratospheric aerosol shortwave radiative forcing in periods of minimal-moderate volcanic activity

• Substituting organics for sulfate may cause ±100% change in shortwave radiative forcing depending on refractive index and mixing state

• Radiative impacts of organic-containing stratospheric aerosols are strongest in the lower stratosphere

Plain Language Summary

While it is generally assumed that stratospheric aerosol is dominated by pure sulfuric acid plus water aerosols, recent modeling studies and in situ measurements suggest that organic matter makes up a significant fraction of lower stratospheric aerosol. The implications of this organic component are uncertain but may require significant revision of our understanding of the stratosphere's climate influence. Here, we perform a sensitivity study of shortwave RF of stratospheric aerosols in northern midlatitudes, examining the influence of different plausible values of complex refractive index and particle morphologies. A radiative transfer model indicates that both complex refractive index of organics, and particle morphologies, could have a consequential influence on the shortwave RF of stratospheric aerosols. Currently, however, there is very little data to constrain the mixing state and refractive index of these organic-containing aerosols. Reducing uncertainty in details of these aerosols requires a concerted effort to bring models, observations, and laboratory studies together in a unified framework.

1 Introduction

Aerosols play a critical role in climate by influencing the radiative balance of the Earth both directly and indirectly (Stocker et al., 2013). In the stratosphere, aerosols modify radiative fluxes by scattering and absorbing solar radiation. It has been suggested that stratospheric aerosols have contributed about 21% of the total direct aerosol radiative forcing (RF) since 1850 (Yu et al., 2016). It has long been believed that stratospheric aerosols are sulfuric acid/water (sulfate) solution droplets that can include nitric acid under some conditions (Arnold et al., 1998; Junge & Manson, 1961; Myhre et al., 2004; Rosen, 1971; Wilson et al., 2008). Most previous studies on radiation effects of stratospheric aerosols have assumed a pure sulfate composition (Andersson et al., 2015; Myhre et al., 2004; Solomon et al., 2011). However, recent aircraft studies have revised this view and shown the importance of organic aerosols (OA) in the lower stratosphere, where OA accounts for 5–50% of particle mass, especially at midlatitudes (Froyd et al., 2009; Martinsson et al., 2019; Murphy et al., 2007, 2014). Modeling results also suggest that organic constituents contribute 30–40% of the stratospheric aerosol budget, primarily in the lowermost stratosphere (Yu et al., 2016). Recent Australian and Canadian wildfire smoke reaching the middle stratosphere injected significant aerosols (Hirsch & Koren, 2021; Torres et al., 2020), which generally include a large amount of OA (Andreae, 2019). Observational data show that stratospheric aerosols below 15 km contribute up to 80% of nonvolcanic total stratospheric aerosol optical depth (AOD) at middle to high latitudes (Ridley et al., 2014). Organic material in the stratosphere, however, has been largely ignored by the stratospheric modeling community, partly because that organic material in the lower stratosphere is hard to distinguish from sulfate by satellite instruments. Furthermore, little is known about its physical and chemical properties, making meaningful inclusion in models challenging. Modeling investigations suggest that radiative influence of OA and dust must be included to obtain consistency with satellite measurements of stratospheric aerosols in terms of the seasonal and interannual variability (Schallock et al., 2017). Organic aerosols mixing with sulfate and nitrate, depending on the aerosol composition and mixing state, could significantly change aerosol radiative effects and thus the RF (Klingmüller et al., 2014).

The contribution of stratospheric aerosols to the RF depends on wavelengths and aerosol optical properties, which are determined by the aerosol complex refractive index (𝑚 = 𝑛 + 𝑘𝑖), morphology, and size distribution. Values of the complex refractive index of OA vary depending on the precursor(s), oxidation pathway, and reaction/aging duration (Moise et al., 2015). There are three pathways contributing to the OA in the midlatitude lower stratosphere, i.e., deep convection from the troposphere, air mass descent and particle sedimentation from the middle stratosphere, and isentropic transport from the tropical lower stratosphere. Note that the Brewer-Dobson circulation transports OA entering the tropical tropopause upward into the middle stratosphere and poleward. OA from these pathways is expected to be different in the terms of precursor(s), oxidation pathway, and aging, and thus values of complex refractive index. Laboratory and in situ studies suggest a range of possible complex refractive indices of OA in the stratosphere. The real refractive index () could vary over a range that brackets the value of for sulfate (Cappa et al., 2011; Lambe et al., 2013; Liu et al., 2013, 2015; Nakayama et al., 2012, 2013). The imaginary refractive index () of OA is highly uncertain, suggesting aerosols that range from being purely refractive to significantly absorbing (e.g., brown carbon, BrC) (Flores et al., 2014; Lambe et al., 2013; Liu et al., 2014; Zeng et al., 2020). BrC comes from both primary sources, such as incomplete biomass/biofuel combustion (Chakrabarty et al., 2010; Feng et al., 2013), and secondary sources, such as photooxidation of volatile organic compounds and aqueous-phase reactions (Nakayama et al., 2010; Nguyen et al., 2012). Additionally, observational and laboratory studies show that both primary and secondary BrC bleach by photochemical aging, which reduces the BrC light absorption (Forrister et al., 2015; Wong et al., 2017, 2019). There is also laboratory evidence that some high-molecular-weight chromophores from primary biomass/biofuel emissions survive photochemical bleaching (Wong et al., 2017, 2019). Given these complicated processes influencing optical properties of BrC, there are not enough observations to specify the variable value for BrC from different sources, nor to quantify the relative contribution of these sources to stratospheric OA. Although understanding of the contribution of organics to stratospheric aerosols continues to improve, the role organic compounds play in stratospheric aerosol mixing state is still quite uncertain (Kremser et al., 2016; Nguyen et al., 2008). Depending on the complex refractive index of the organic component, the morphological variation could have a significant influence on aerosol radiative properties (Klingmüller et al., 2014).

Because we lack observational constraints on the complex refractive index and mixing state of organic-containing stratospheric aerosols, we use a sensitivity study approach to calculate the shortwave RF. For large volcanic eruptions, longwave RF can reduce shortwave RF by 40% due to longwave absorption of large sulfate particles (Andronova et al., 1999). Such large organic-containing particles are not observed in the stratosphere, and are therefore neglected as a source of longwave RF in this study. In situ measurements of aerosol size (vertical profile data up to 30 km from balloon-borne optical particle counter (OPC) measurements) and composition (aircraft measurements below 20 km) are used to derive representative size distributions, vertical profiles, and organic mass fractions for the computation of aerosol optical properties. These profiles of aerosol optical properties are used as inputs to a radiative transfer model to calculate profiles of shortwave fluxes and radiative heating rates for standard model atmospheres.

2 Methods

The aerosol optical properties required for the shortwave RF calculation are the AOD, single scatter albedo, and asymmetry parameter. These quantities are calculated assuming spherical particles and using Mie theory. The Mie calculation requires a particle size, complex refractive index, and mixing state. The particle radii are taken from balloon-based in situ size distribution observations. The complex refractive index of OA is allowed to vary over a wide range to bracket the plausible values from laboratory measurements cited in Section 1. The real part is varied between 1.3 and 1.65, and the complex part ranges from 0.0001i to 0.02i. The mixing states studied are core-shell, external mixing (EM), and internal mixing (IM). The core-shell mixing states include organic core-sulfate shell (OCSS) and sulfate core-organic shell (SCOS), and are calculated using the approach described in Bohren and Huffman (2008) implemented in MATLAB with the code of Mätzler (2002). The core-shell mixing state optical properties are calculated using the coated sphere solution. The internal and external mixing states use a homogeneous sphere with a complex refractive index value calculated from standard mixing rules.

The computation of stratospheric shortwave RF requires a profile of AOD spanning the stratospheric layers containing significant quantities of aerosol. In situ observations of size distribution made from sounding balloons therefore provide an ideal snapshot for this purpose of a profile of stratospheric aerosol at a particular time and location. We utilize size distributions retrieved from the OPC measurements that began in 1971 (Deshler et al., 2003, 2019; Hofmann et al., 1975). These retrievals provide a best estimate starting from the tropopause up to about 20 hPa of the fit of a bimodal lognormal distribution to the cumulative distribution measurements. Each of the two modes is characterized by total particle number, median radius, and modal width. Mie calculations are integrated over the bimodal distribution. The empirical bimodal fitting is different from classical aerosol mode characterization (i.e., nucleation, Aitken, accumulation, and coarse mode) but it fits well to measurements. The aerosol profiles are averaged over 0.5 km layers between 12.5 and 29 km (180-15 hPa). For the purpose of computing optical properties, the smaller first mode was assumed to be a mix of OA and sulfate aerosol, while the larger second mode was assumed to be pure sulfate aerosol, as indicated by Particle Analysis by Laser Mass Spectrometry (PALMS) observations (Murphy et al., 2014; Murphy et al., 2021). A speculative case of 20% volume fraction organic content in the second mode results in extinction changes of −2–3% for quiescent and −1–1% for moderate volcanic conditions relative to a pure sulfate second mode, as assumed for the rest of this work. We calculate optical impacts of including organic components against the baseline case of pure sulfate aerosol. Measurements to date indicate that the vast majority of OA is below 500 K potential temperature (∼21 km) (Murphy et al., 2014). Our analysis (see Figure S1 in Supporting Information S1) shows that this region is also the location of the most optically significant aerosol. Based on this understanding, and for computational simplicity, a constant organic fraction was applied throughout the atmosphere column. Only small radiative differences were found when calculations over a full or partial column were compared. Our analysis focuses on the period of 1991–2013. Regular balloon soundings made during this span covered a wide range of volcanic conditions, which we divided into a period dominated by the eruption of Mt. Pinatubo (1991–1996), a period of relative volcanic quiescence (1997–2004), and a period characterized by intermittent, moderate-sized eruptions (2005–2013). The relative similarity of the retrieved distributions within each of these three periods provided a climatology to fill missing data and provide vertically complete profiles.

The shortwave RF of these aerosols was computed using the shortwave component of the Rapid Radiative Transfer Model for General Circulation Model Applications (RRTMG_SW) developed by Clough et al. (2005). RRTMG_SW accepts profiles of aerosol extinction optical depth, single scatter albedo, and asymmetry parameter, to perform two-stream calculations of shortwave radiative fluxes. We computed shortwave bulk aerosol optical properties, weighted within bands by the top of atmosphere (TOA) spectral solar flux, from the two modes and interpolated these to the levels of the midlatitude summer standard atmosphere. We then calculated the difference between the net shortwave radiative flux with and without the stratospheric aerosols at a solar zenith angle of 30° to estimate the stratospheric aerosol shortwave RF.

3 Results and Discussion

Figure 1 shows the mean stratospheric aerosol surface area density (SAD) distributions of first and second modes averaged over time and altitude (180-15 hPa) in volcanically quiescent and moderate volcanic periods. The SAD distribution of first mode particles is quite similar in volcanically quiescent and moderate volcanic periods. In contrast, the SAD of second mode particles is substantially enhanced during the moderate volcanic period (and the Pinatubo period, see Figure S2 in Supporting Information S1), which is consistent with the significant mass increase of stratospheric sulfate particles from volcanic SO₂ oxidation followed by nucleation and/or condensation (Thomason & Peter, 2006). The multipeaked second mode in the volcanic period is a result of averaging over a long period of time and over multiple levels. The episodic volcanic eruptions which perturb the second mode size distribution create these transient narrow peaks that show up clearly against the smoother background distribution (see Figure S3 in Supporting Information S1). The radius dependence of scattering (upper panel) and absorption (bottom panel) efficiencies are also shown in Figure 1 for various cases including pure sulfate particles and organic-containing particles at 532-nm wavelength and averaged over RRTMG band 10 (442–625-nm wavelengths). Note that scattering and absorption efficiency plots are independent of the time period. The organic-containing particles presented here are in the OCSS mixing state, which is consistent with the predominance of OA transported from the lower atmosphere and coated by ubiquitous sulfate components, as inferred from PALMS observations (Murphy et al., 2014) below 20 km (50 hPa). Compared to smooth structures of the radiative transfer model band-averaged scattering and absorption efficiency distributions, the Mie efficiencies at a single wavelength show a complicated resonance structure due to partial-wave interferences (Chýlek, 1990). At larger radii, these resonance phenomena lead to a decrease in the overall band-averaged scattering efficiency compared to the single wavelength case. The band average case is more relevant to the RF.

Figure 1

Averaged stratospheric surface area density (SAD) distribution in volcanically quiescent (upper) and moderate volcanic (bottom) periods. The SAD distributions for first and second modes are also presented. The radius dependences of scattering (upper) and absorption (bottom) efficiencies are shown for cases including pure sulfate particles and 50% (volume fraction) organic-containing particles at 532-nm wavelength and averaged over RRTMG band 10 (442–625-nm wavelength). For organic-containing particles, m_org = 1.55 + 0.001i is used for organics. The mixing state is organic core-sulfate shell (OCSS).

Although the radius of the maximum of the SAD distribution does not match that of the maximum scattering efficiency, there is still a sizable overlap between radii of high scattering efficiency and significant SAD, and thus stratospheric aerosols are capable of significant optical scattering. Since the tails of the two distributions overlap, the optical impacts of stratospheric aerosols are highly sensitive to the details of both SAD distribution and the organic fraction and properties (see Figure S4 in Supporting Information S1), which affects the tail of the scattering efficiency. Due to the measurement uncertainty of the OPC, the higher moments of the size distribution, e.g., the SAD, are uncertain to ±40% (Deshler et al., 2003). Figure S4 in Supporting Information S1 shows an example of the change in the optical properties caused by a notional change in size distribution. Figure 1 shows that organic content shifts the maximum of the scattering efficiency curve toward that of the first-mode SAD distribution, which increases their overlap and thus increases the extinction coefficient (cumulative SAD distributions are in Figure S5 in Supporting Information S1). Based on these considerations, knowing the stratospheric aerosol burden is not sufficient to determine the aerosol optical impacts in the stratosphere. It is essential to quantify the contribution of any aerosol organic component to reduce uncertainty in the column optical properties.

The upper panel of Figure 2 shows the time series of simulated stratospheric AOD of first (nominal value of m_org = 1.55 + 0.001i is used) and second mode particles based on the balloon-borne observations between 1991 and 2013 (see Figure S6 in Supporting Information S1 for total AOD). Compared to the volcanically quiescent period (1997–2004), stratospheric AOD is greatly enhanced, especially for the second mode, in the Pinatubo (1991–1996) and moderate volcanic periods (2005–2011). The second mode dominates the stratospheric AOD during the Pinatubo period, while the organic-containing first mode dominates the stratospheric AOD during the volcanically quiescent period, which highlights the importance of better understanding the optical properties of organic components in the first mode of the background stratosphere. Depending on the volcanic influence, the relative contribution of the first and second modes varies during the moderate volcanic period. The bottom panel of Figure 2 presents relative changes in stratospheric AOD compared to the baseline case of pure sulfate aerosol for organic components in the first mode with different refractive indices. Depending on the complex refractive index of organics, the relative change in stratospheric AOD ranges from −20% to +50%. For each refractive index, the mean change is most significant during the volcanically quiescent period, which again emphasizes the importance of the composition of the background stratospheric aerosol.

Figure 2

Time series of (upper) simulated stratospheric aerosol optical depth (AOD) based on balloon-borne observations of first and second modes, and (bottom) relative changes in total stratospheric AOD when substituting pure sulfate first mode particles with 50% organic-containing particles with different refractive indices () for different periods between 1991 and 2013. The AOD data are the average of all RRTMG-SW wavelengths (200–12,195 nm) weighted by the spectrum of the extraterrestrial solar irradiance. Individual balloon-borne observations are indicated by the plus sign marker. First mode particles are 50% organic-containing with organic core-sulfate shell (OCSS) mixing state. Second mode particles are pure sulfate. To estimate total radiative forcing (RF) using shortwave AOD during large volcanic eruptions requires a reduction of up to 40% in AOD to account for longwave effects (see Section 1).

The left panel of Figure 3 presents vertical profiles of the total and second mode-only aerosol extinction coefficients for different periods. The distinct difference between the total and the second mode extinction coefficients emphasizes the relative importance of the first mode for the volcanically quiescent period. Larger values of extinction coefficient in an absolute sense are mostly found at low altitudes, especially below 450 K potential temperature for the quiescent and moderate volcanic periods, and below 600 K for the Pinatubo period, which means that aerosols in the lower stratosphere contribute the major part of aerosol light extinction. The right panel of Figure 3 shows vertical profiles of the relative change in extinction coefficient with regards to the substitution of pure sulfate first mode particles with 50% organic-containing particles at three different periods. Consistent with Figure 2, the volcanically quiescent period is most affected by organic substitution. The decrease in the wavelength location of peak scattering efficiency induced by an assumption of 50% organic particles in Figure 1 causes about 21% increase in average in extinction coefficient compared to the pure sulfate case. The Pinatubo period is dominated by the second mode.

Figure 3

Vertical profiles of (left) the total and second mode only extinction coefficients, and (right) the relative change in total extinction coefficient when substituting pure sulfate first mode particles with 50% organic-containing particles () at three different periods. First mode particles are 50% organic-containing (m_org = 1.55 + 0.001i) with organic core-sulfate shell (OCSS) mixing state. Second mode particles are pure sulfate. The shaded area represents the 95% confidence interval.

Figure 4 shows the relationship between stratospheric AOD and TOA shortwave RF for different complex refractive indices of organics and particle mixing states in the volcanically quiescent and moderate volcanic periods (see Figure S7 in Supporting Information S1 for the Pinatubo period). The substantial statistical variability of shortwave RF as a function of extinction suggests that stratospheric AOD, which is an important satellite product (Khaykin et al., 2017; Solomon et al., 2011), is not sufficient to constrain stratospheric aerosol shortwave RF. For fixed assumptions about mixing state and complex refractive index, there is however a broadly consistent relationship between column AOD and shortwave RF. A few outliers do deviate from this consistent relationship, and produce more shortwave RF (in an absolute sense) per unit column AOD. These outliers arise because certain characteristics of the size distribution produce a higher fraction of scattered radiation directed back to space. Such characteristics include the ratio of mode 1 to mode 2 optical depth and the median radius of the particles for a mode. Depending on the complex refractive index of organics and the particle mixing state, the uncertainty of TOA shortwave RF at certain stratospheric AODs could be as large as 0.6 W m⁻² at the extreme end of the difference, with a typical difference of about 0.35 W m⁻². The TOA shortwave RF could even be positive if there is strong absorption, which is consistent with satellite-based estimates after extreme wildfires by Christian et al. (2019). Although they attributed positive forcing to black carbon, we would expect similar effects for light-absorbing OA. Compared to the pure sulfate case, the 50% organic-containing case for the first mode would cause up to ±100% change in TOA shortwave RF for the extreme cases of the complex refractive index of organics in the volcanically quiescent period, depending on the particle mixing state (Figures S8 and S9 in Supporting Information S1). Both complex refractive index and mixing state considerably influence the relationship between stratospheric aerosol shortwave RF and AOD. The effects of complex refractive index are more significant, while there is very little data to constrain this quantity for the organic component of stratospheric aerosols.

Figure 4

Scatter plots of top of atmosphere (TOA) shortwave radiative forcing (RF) and total stratospheric aerosol optical depth (AOD). The AOD data are averaged over all RRTMG-SW wavelengths weighted by the solar spectrum. First mode particles are 50% organic carbon and second mode particles are pure sulfate. (left) Organic-containing particles with four different complex refractive indices in organic core-sulfate shell (OCSS) mixing state. (right) Organic-containing particles with two different complex refractive indices in four different mixing states. The baseline case of pure sulfate aerosol is presented in gray-filled circles.

4 Conclusions

Effects of the complex refractive index of organics and particle mixing state on the shortwave RF of stratospheric aerosols in northern midlatitudes are investigated via a sensitivity study. In situ measurements of aerosol size are used to provide vertical profiles of the SAD distribution (Figure S10 in Supporting Information S1). The organic volume fraction of the first aerosol mode is allowed to vary for the computation of aerosol optical properties, such as scattering and absorption coefficients, and asymmetry parameter. Results show that shortwave radiative impacts of stratospheric aerosols are highly sensitive to the aerosol SAD distribution. The vertical profiles of the aerosol extinction coefficient indicate that aerosols in the lower stratosphere contribute the major part of aerosol light extinction. Compared to the second mode particles with pure sulfate, the possibly organic-containing first mode particles dominate the simulated stratospheric AOD during the volcanically quiescent period. These profiles of aerosol optical properties are used as inputs to a radiative transfer model to calculate profiles of shortwave fluxes for a standard model atmosphere to isolate aerosol radiative impacts from those associated with thermodynamic quantities and trace species. In the volcanically quiescent period, the mean TOA shortwave RF of the pure sulfate case is about −0.2 W m⁻². Depending on the complex refractive index of organics and the particle mixing state, the relative change in TOA shortwave RF could be as high as ±100% when substituting pure sulfate first mode particles with 50% organic-containing particles. The relationship between stratospheric aerosol shortwave RF and AOD is significantly influenced by complex refractive index and mixing state, with the complex refractive index having greater impacts. Both of these quantities are highly uncertain for stratospheric aerosols. Our results highlight the uncertainty in stratospheric aerosol shortwave RF due to the presence of organic compounds and call for better understanding of the complex refractive index and mixing state of stratospheric aerosols.

Acknowledgments

We thank Atmospheric and Environmental Research, Inc. (AER) for making the source code for RRTMG available. This work was supported in part by the NASA grant 80NSSC19K0326. The in situ measurements were supported by the NSF awards 8920706, 9732547, 0216558, 0437406, 1011827.

Open Research

Data Availability Statement

The source code for RRTMG is available at https://github.com/AER-RC/RRTMG_SW. The balloon data are available at http://www.atmos.uwyo.edu/∼deshler/Data/Aer_Meas_Wy_read_me.htm.