University of Birmingham Analysis of radiative properties and direct radiative forcing estimates of dominant aerosol clusters over an urban-desert region in West Africa

The strategic location of the AERONET site (Ilorin) makes it possible to obtain information 26 on several aerosol types and their radiative effects. The strong reversal of wind direction 27 occasioned by the movement of the ITCZ during the West Africa Monsoon (WAM) plays a 28 major role in the variability of aerosol nature at this site. Aerosol optical depth (AOD) (675 29 nm) and Angstrom exponent (AE) (440-870 nm) with 1st and 99th percentile values of 0.08 30 and 2.16, and 0.11 and 1.47, respectively, confirms the highly varying nature of aerosol at 31 this site. Direct radiative forcing (DRF) and radiative forcing efficiency (RFE) of aerosol as 32 retrieved from the AERONET sun-photometer measurements are estimated using radiative 33 transfer calculations for the period 2005-2009 and 2011-2015. The DRF and RFE of 34 dominant aerosol classes - desert dust (DD), biomass burning (BB), urban (UB) and gas 35 flaring (GF) - have been estimated. Median (±standard deviation) values of DRF at top-of- 36 atmosphere (TOA) for the DD, BB, UB and GF aerosol classes are -27.5±13.2 Wm − 2 , - 37 27.1±8.3 Wm − 2 , -11.5±13.2 Wm − 2 and -9.6±8.0 Wm − 2 , respectively. While that of RFE for 38 DD, BB, UB and GF aerosol classes are -26.2±4.1 Wm − 2 δ − 1 , -35.2±4.6 Wm − 2 δ − 1 , -31.0±8.4 39 Wm − 2 δ − 1 and -37.0±10.3 Wm − 2 δ − 1 , respectively. The DD aerosol class showed the largest 40 DRF but the smallest RFE, arguably, due to the high SSA and asymmetry factor values for 41 this aerosol


A C C E P T E D M A N U S C R I P T
by six-hourly ERA-Interim (European Centre for Medium-Range Weather Forecasts Interim 124 Re-Analysis) wind analyses data. The trajectories of particles are calculated backward in time 125 by interpolating these wind analysis to the current particle position. The position (latitude, 126 longitude) and pressure were output every trajectory time step of 0.6 hours. The choice of 7-127 day back trajectory length is due to the atmospheric lifetime of between 5 and 9 days 128 estimated for black carbon (BC) and particulate organic matter (POM), respectively (

A C C E P T E D M A N U S C R I P T
7

AERONET data analysis 145
The absolute magnitude of aerosol radiative forcing is determined, predominantly, by the 146 values of aerosol optical depth (AOD) and single scattering albedo (SSA), while its sign is 147 dependent on the SSA and surface albedo. Both AOD and SSA vary significantly with the 148 source of the aerosol (Pani et al., 2016). In this study, Version 3 Level 1.5 of AERONET data 149 released in January 2018 to which improved cloud screening and new quality controls have 150 been applied were used to estimate the DRF and RFE of anthropogenic and natural aerosol 151 classes in the West Africa sub-region. Adequate knowledge of aerosol SSA, hemispheric 152 backscatter fraction (b) and AOD can be used to calculate the mean TOA aerosol radiative 153 forcing for optically thin, partially absorbing aerosol (Haywood and Shine, 1995). For sites 154 like Ilorin, where differences in the diurnal variation of aerosol properties (extensive and 155 intensive) could be highly pronounced, the use of monthly averages of aerosol parameters 156 will only provide highly generalised estimates of the optical and microphysical properties of 157 the aerosol at such a site. 158 One of the key properties that determine the climate forcing ability of an aerosol is the 159 angular distribution of the light scattered by the aerosol particles (Marshall et al., 1995). The 160 angular distribution of scattered light intensity at a specific wavelength is referred to as the 161 phase function (P). The asymmetry parameter, g, an important intensive parameter of 162 aerosols for estimating its climate forcing ability could be derived from P. Values of g range 163 between -1 for entirely backscattered light to +1 for entirely forward scattered light (Andrews 164 et al., 2006). The fraction of backscattered light is the ratio of the integral of the volume 165 scattering function over the backward half solid angle divided by the integral of the volume 166 scattered function over the full solid angle (Horvath et al., 2016). 167 ultraviolet wavelengths (Holben et al., 1998;Eck et al., 1999). For all sky radiance 171 wavelengths (that is, 440, 675, 870, and 1020 nm), the uncertainty in SSA is expected to be 172 ±0.03 based on Version 1almucantar retrieval computations (Dubovik et al., 2000;Holben et 173 al., 2006). to be incorporated into radiative forcing calculations and radiative transfer codes that can 206 produce representative and accurate estimates of radiative forcing with one or two 207 wavelength regions (Blanchet, 1982). In their study to examine the possibility of replacing 208 aerosol parameters by wavelength-independent parameters and the accuracy and 209 representativeness of such average parameters for the complete solar spectrum, Blanchet 210 (1982) found out that results of calculations with average parameter are in close agreement 211 with corresponding terms at a wavelength (λ) of 700 nm. Haywood (1995), using detailed 212 radiative transfer codes, tested the representativeness of average aerosol parameter and found 213 that results at around λ=700 nm were quite similar to those of using the entire solar spectrum.

Variability of Angström exponent and (AOD) for the aerosol classes 317
The median (±standard deviation) values of the optical and microphysical properties of the 318 identified aerosol classes are presented in Table 1

Urban aerosol (UB) 336
Aerosol signature in the urban aerosol class is prominent in the WAM months when the 337 south-westerly moist monsoon wind is prevalent in the region. For this class, the median 338 AOD 440 and AE values are 0.53(±0.35) and 0.52(±0.34), respectively. Even though this AE 339 value is low, it is still higher than that for the DD aerosol class. The DD aerosol class is 340 expected to contain a higher fraction of coarse aerosol. This class of aerosol (Urban), with a 341 median value of SSA of 0.93(±0.04), is partially absorbing arguably due to increased 342 carbonaceous particle content from anthropogenic sources in the urban area. 343

Gas flaring aerosol (GF) 344
This class is similar to the urban class but has a lower median value of AOD 440 and an 345 average AE value, which is higher than that of the urban aerosol by a factor of ∼2.   African sub-region using AERONET retrievals from Ilorin, Nigeria. The DD aerosol class is 447 characterised by high AOD and low AE. The BB aerosol class is characterised by high AOD 448 and high AE while the GF class is characterised by low AOD and high AE. The direct 449 radiative forcing of the various dominant aerosol types has been estimated using aerosol 450 parameters from AERONET retrievals as inputs in a simplified radiative transfer equation 451 proposed by Haywood and Shine (1995). Due to differences in methodologies and varying 452 aerosol sources/nature, it is difficult to directly compare results (average DRF values) from 453 literature. Desert dust (DD) and biomass burning aerosols were found to be the most effective 454 cooling aerosol at the TOA in the region. UB and GF aerosol classes which are suggested to 455 be rich in emissions from the combustion of fossil fuel (i.e. black carbon and sulphate) have 456 less cooling effects. The more absorbing aerosols (GF and BB) show the higher forcing 457 efficiency; and, GF aerosol class, the largest variability in RFE. These results suggest the 458 need for concerted efforts to adequately characterise and quantify emissions from real-world 459 gas flares as they make significant contributions to the radiative transfer in the Earth-