Interaction of Dust Aerosols with Land/Sea Breezes over the Eastern Coast of the Red Sea 1 from LIDAR Data and High-resolution WRF-Chem Simulations

18 With advances in modeling approaches and the application of satellite and ground-based data in dust-related research, our understanding of the dust cycle has significantly improved in recent decades. However, two aspects of the dust cycle, namely the vertical profiles and diurnal cycles, 21 are not yet adequately understood, mainly due to the sparsity of direct observations. Measurements of backscattering caused by atmospheric aerosols have been ongoing since 2014 at the (KAUST) using a micro- pulse LIDAR with a high temporal resolution. KAUST is located on the east coast of the Red Sea (22.3° N, 39.1° E), and currently hosts the only operating LIDAR system in the Arabian Peninsula. We use the data from this LIDAR together with other collocated observations and 27 high-resolution WRF-Chem model simulations to study the following aspects of aerosols, with a 28 focus on dust over the Red Sea Arabian coastal plains. Firstly, we investigate the vertical profiles 29 of aerosol extinction and concentration in terms of their seasonal and diurnal variability. Secondly, we evaluate how well the WRF-Chem model performs in representing the vertical 31 distribution of aerosols over the study site. Thirdly, we explore the interactions between dust 32 aerosols and land/sea breezes, which are the most influential components of the local diurnal 33 circulation in the region. We found a substantial variation in the vertical profile of aerosols in 34 different seasons. We also discovered a marked difference in the daytime and nighttime vertical 35 distribution of aerosols at the study site, as revealed by the LIDAR data. The LIDAR data also 36 identified a prominent dust layer at ~5–7 km during the nighttime, which represented the long- 37 range transported dust brought to the site by the easterly flow from remote inland deserts. The 38 vertical profiles of aerosol extinction in different seasons were largely consistent between the 39 LIDAR, MERRA-2 reanalysis, and CALIOP data, as well as in the WRF-Chem simulations. The 40 sea breeze circulation was much deeper (~2 km) than the land breeze circulation (~1 km), but 41 both breeze systems prominently affected the distribution of dust aerosols over the study site. We observed that sea breezes push the dust aerosols upwards along the western slope of the Sarawat Mountains, which eventually collide with the dust-laden northeasterly trade winds coming from nearby inland deserts, causing elevated dust maxima at a height of ~1.5 km above sea level over 45 the mountains. Moreover, the sea and land breezes intensified dust emissions from the coastal region during the daytime and nighttime, respectively. The WRF-Chem model successfully captured the onset, demise, and height of a large-scale dust event that occurred in 2015, compared to LIDAR data. Our study, although focused on a particular region, has broader environmental implications as it highlights how aerosols and dust emissions from the coastal plains can affect the Red Sea climate and marine habitats.


Introduction
Dust aerosols, which mainly originate from natural deserts and disturbed soils such as 54 agricultural areas, have implications for air quality (Prospero, 1999 Peninsula represents a key area within the global dust belt where significant dust emissions take 58 place in all seasons. However, the spatio-temporal characteristics of dust emissions in the region 59 have not yet been fully described, partly because of the sparsity of observations. Although our 60 understanding of the dust cycle and the related physical processes has substantially improved in 61 recent decades (Shao et al., 2011), in the present context, two aspects of dust aerosol dynamics   Because the site is located exactly at the land-ocean boundary, some unique small-scale 143 processes exist that affect the local climate of this region. For instance, land and sea breezes 144 affect the distribution of dust in the atmosphere over the study site. The desert land heats up 145 during the day, which consequently heats the surface air above the land. This warm air mass rises 146 due to convection, creating a local low pressure at the surface. The cooler and more moist air 147 over the Red Sea then flows towards the low pressure, thus forming sea breezes (Simpson, 1994; over the study region are presented in Fig. 1, which we discuss in detail later.  156 We use several datasets, described below, to derive the climatology of the season profile and 157 surface winds for the years 2015-2016. 158 Datasets 159 We collected meteorological data, including wind speed, temperature, and humidity from a tower Institution) (Farrar et al., 2009;Osipov et al., 2015). 162 We use cloud-free aerosol extinction profiles retrieved from a CALIOP (Cloud-Aerosol Lidar 163 with Orthogonal Polarization) instrument onboard CALIPSO (Cloud-Aerosol Lidar and Infrared 164 Pathfinder Satellite Observations) for analyzing the vertical structure of aerosols at the study site. 165 CALIPSO is flown in a sun-synchronous polar orbit and is a part of NASA's Afternoon (A-train) 166 constellations (Stephens et al., 2018). CALIOP acquires observations during both the day and 167 night portion of the orbit with a 16-day repeat cycle. We use level-3 day/night aerosol data 168 v3.00, which are monthly aerosol products generated by aggregating level-2 monthly statistics at 169 2° (lat) × 5° (long) resolution (Winker et al., 2013). The data have 208 vertical levels up to a 170 height of 12 km above sea level. 171 We also analyze aerosol optical depth (AOD) data from AERONET station at KAUST (Holben 172 et al., 1998). We use a level 2.0 version of directly measured AOD values (direct sun algorithm), 173 which are cloud-screened and quality-assured. From AERONET, we also use an aerosol number 174 density and a particle size distribution (PSD) obtained by inversion (Dubovik et al., 2000) to 175 characterize the aerosol particles in the region. We use the AERONET V3, level 2.0 product, 176 which provides volume concentration of aerosols in the atmospheric column in 22 bins between 177 0.05 and 15 microns in radius (Dubovik et al., 2000;Parajuli et al., 2019;Ukhov et al., 2020). 178 We use Moderate Resolution Imaging Spectroradiometer (MODIS) level-2 Deep Blue AOD data 179 (Hsu et al., 2004), which are available daily, for the whole globe, at a resolution of ~ 0.1°× 0.1°. 180 We use the latest version of the MODIS dataset (collection 6) (Hsu et al., 2013) because of its 181 extended coverage and improved Deep Blue aerosol retrieval algorithm, compared to its earlier 182 version (collection 5). We process AOD data of both Terra and Aqua satellites on a daily basis, 183 and use the average of the two data products for our analysis. From MODIS, we also use the true 184 color images for a qualitative analysis of a dust event. 185 We adopt the Modern-Era Retrospective Analysis for Research and Applications version 2  We also employ 555nm column AOD from MISR onboard Terra satellite archived under 192 collection MIL3DAE_4, which is a daily product available at 0.5x0.5 degree resolution (Diner,193 2009). Because MISR has a wider view with nine viewing angles, MISR identifes thin aerosol 194 layers more accurately and is more sensitive to the shape and size of particles (Kahn et al., 2005). 195 We also use the RGB composite from SEVIRI (Spinning Enhanced Visible and Infrared Imager) 196 instrument onboard the geostationary Meteosat satellite, which is a composite prepared from 197 specific infrared channels that are sensitive to the presence of dust in the atmosphere (Ackerman,198 1997; Schepanski et al., 2007). Dust appears 'pink' in these composite images and is thus 199 distinguishable from clouds, which are usually shown in yellow, red, or green. 201 Micropulse LIDAR is a fully autonomous active remote-sensing system in which a laser 202 transmitter emits light vertically upward, and an optical sensor receives the backscattered signals.

203
The numbers and the detection time of the backscattered photons provide information about the 204 aerosols and clouds in the atmosphere. The LIDAR located on the KAUST campus, which is 205 also a part of the MPLNET network, operates at a wavelength of 532nm. The data from this 206 LIDAR (hereafter called KAUST-MPL) is the main basis of this paper.

207
The colocation of the KAUST-MPL and AERONET station provides a more comprehensive 208 microphysical picture when combined with AERONET sun-photometer measurements. We

215
We use cloud-screened AERONET radiances and LIDAR backscatter signals combined to 216 retrieve aerosol properties during the daytime. As the AOD data are unavailable during the night, 217 for nighttime retrievals, we use a so-called multi-pixel approach, first introduced by Dubovik et 218 al. (2011) and realized in GRASP. According to this approach, retrieval is implemented for a 219 group of observations coordinated in time or/and in space (e.g., several satellite pixels).

220
Correspondingly, the quality of the retrievals can be improved by using some additional a priori 221 constraints on the time-varying aspect of the retrieved parameters. For example, in this study, we 222 invert the closest AERONET measurements obtained the day before and the day after, together 223 with the nighttime LIDAR backscatter data, under some constraints on the temporal variability of 224 columnar parameters (size distribution, complex refractive index, and sphericity fraction) 225 provided by AERONET measurements. In contrast to other similar but simpler retrieval 226 approaches used currently, multi-pixel concept constraints, but do not eliminate possible 227 variability between parameters. For example, in this study, the implemented retrieval allows us  The retrieved aerosol data has 100 levels in the vertical dimension with a resolution of 75m from 232 505m to 7700m above sea level. The processed LIDAR extinction data has some data gaps 233 because of the quality constraints applied and cloud filtering. To achieve a complete diurnal 234 picture, we also analyze the raw data of the normalized relative backscatter (NRB), which gives 235 the total backscatter from both aerosols and clouds at a fine, 1-min resolution.  Table 1. 246 The model top is set at 100 hPa, and the model has 30 vertical levels between ~20 m to 16 km.

247
To better represent winds, we apply 'grid nudging' on the u and v components of wind above the    We use high-resolution operational analysis data from ECMWF (~15 km) to provide initial and

278
(1986) and Gong (2003). In this parameterization, the rate of sea salt emissions produced via 279 whitecaps and wave disruption is given as a function of particle size and 10-m wind speed. We activate both gas and aerosol chemistry in our simulations (gaschem_onoff = 1, 298 aerchem_onoff = 1) and apply the aerosol chemistry options in all three domains.

299
To determine the contribution of each aerosol species on total AOD, we modify the WRF Chem 300 code, mainly the Fortran files optical_driver.F and chem_driver.F located under the chem folder.

301
For this purpose, we calculate aerosol optical properties twice, first with the mixture containing 302 all aerosols and second after removing a specific aerosol. This calculation is implemented in the 303 subroutine "optical_averaging". Thus, we obtain the contribution of specific aerosol species on 304 total AOD by subtracting the AOD obtained without a specific aerosol from the total AOD 305 calculated when all aerosols are accounted for. 306 We calculate the total aerosol concentration (TAC) in       Note that CALIOP extinction profiles represent data averaged over a large grid box (2x5 degree) 451 that contains the KAUST site. As such, CALIOP represents the larger regional-scale vertical 452 structure of aerosols compared to KAUST-MPL, which represents a more local structure. Above dust events. Figure 9 shows the average extinction profiles for clear and dusty conditions from  The elevated dust layer during the nighttime at a height of 5.5-7 km observed earlier in summer 501 and fall (Fig. 7) is present in the 'dusty days' and is absent in 'clear days' (Fig. 9, right). We 502 suggest that these dust layers represent dust of non-local origin transported at higher altitudes 503 during large-scale dust events. Next, we explore why such a high dust loading at this altitude in 504 summer is present only in the nighttime and not in the daytime. Stronger convection in the inland 505 desert regions during the daytime carries aerosols to higher altitudes. In the summer in deserts, 506 convection is strongest in the afternoon, and the planetary boundary layer height (PBLH) can 507 reach well above 5 km (Fig. S4). By the evening, the dust is mixed thoroughly within the PBL by 508 this strong convection (Khan et al., 2015). At night, the PBL weakens and breaks the capping 509 inversion, which allows the dust-laden layer from the PBL to mix into the free troposphere. The 510 dust that lies above the PBL is ultimately carried to our site by the accelerated easterly process is evident if we look at the wind vectors at higher altitudes. As Fig. S5  (1-2 km), which is why KAUST-MPL data shows elevated dust loading at these heights.   events are more frequent in summer and fall, as seen in the KAUST-MPL data (Fig. 7).      which seems to occur at low altitudes ~500 m (Fig. 14). Some dust collects over the Red Sea 618 during the daytime in the winter also, which appears well mixed. During the day, the 619 northwesterly sea breezes move landward because of which the dust emitted from the coastal 620 region cannot move over the sea. Therefore, this dust observed during the daytime must be the 621 residual dust that accumulated overnight. The dust mobilization from the coastal area by the sea 622 breezes (daytime) is weaker during the winter.

623
In the spring, there is very high dust loading over the coastal region and the western flanks of the 624 mountains, which is much higher than in winter. This higher dust loading is consistent with 625 stronger sea breezes in spring than in winter (Fig. 13). The highest dust loading is observed over  In summer, the patterns of dust mobilization and transport are similar to those in spring but are 641 not quite as pronounced. In fall, the mobilization of dust from the coast and its ocean-ward 642 transport is very weak, and their patterns are similar to those in winter.

643
The model-simulated vertical distributions of aerosols do not exactly match the KAUST-MPL 644 profiles presented earlier (Fig. 8). Although it is difficult to identify the exact reason for this 645 discrepancy, there are several possible explanations. Although the effect of orography on dust 646 seems to be correctly resolved (Fig. 14), the transport of dust towards the KAUST site may not that we presented earlier (Fig. 14).

666
In summary, the timings and patterns of dust emission and transport in the study region are    August 09 shows a dust plume originating from northeast Africa around Port Sudan, which, after 689 being deflected by the northerly winds, experiences a marked curvature (Fig. 16b). KAUST-MPL observations. The dust is mainly confined within a height of ~2 km, which is 747 consistent in both datasets. We also observed a higher intrusion of dust into the atmosphere, 748 which is expected because the PBL is well developed in summer.

749
Note that the model data also show a high extinction at a height of ~6 km on August 09/10, 750 particularly at night (Fig. 18), which is consistent with the dust layers observed at 6-7 km height in the KAUST-MPL nighttime data (Fig. 8). Although the model data does not identify these 752 dust layers at 6-7 km in the seasonally averaged profiles presented earlier (Fig. 8), the model 753 nonetheless correctly identified these same dust layers in this event (Fig. 18). The demise timing  results, vertical profiles of aerosols can be affected by local or regional processes, which indicate 842 that the profiles can differ across different regions. Therefore, it is vital to examine the aerosol 843 vertical profiles of a region to understand the regional climate.   Although derived from actual observations, KAUST-MPL retrievals are also subject to 886 uncertainties, and their accuracy is dependent on assumptions made by the retrieval algorithms.

887
A study that compared the GRASP retrieval scheme employed here against in situ measurements 888 showed that the differences were less than 30 % for the different retrieval schemes (Benavent-   Supercomputing Laboratory for providing computing resources. We also thank Anatolii 910 Anisimov for providing SEVIRI images and for helpful discussions. We are grateful to Ellsworth