Spatiotemporal Variation of Particulate Fallout Instances in Sfax City, Southern Tunisia: Influence of Sources and Meteorology

Particles deposition in the main industrial zone of Sfax City (southern Tunisia) was studied by weekly monitoring particulate fallout instances at twenty sites from November 11, 2012, to April 15, 2013. Very high fluctuation in those particle fluxes, ranging from 0.376 to 9.915 g/m, was clearly observed. Spatiotemporal distribution of the deposited particulate fluxes and the exposure of each site to the main industrial plumes (i.e., phosphate treatment plant “SIAPE,” soap industry “SIOS-ZITEX,” and lead secondary melting industry “FP Sfax Sud”) indicated the concomitant effects of surrounding industrial sources. In addition, the highest particulate deposition seemed to be associated with predominant strong cyclonic situations. Those deposition rates exceeded the levels recorded in the case of strong stabilities, considered as responsible for pollutant accumulation.


Introduction
Air pollution usually results from diverse factors including mining activities, metallurgy, chemical industries, road and air traffic, household waste incineration, and industrial wastes [1,2]. The composition of various gaseous and particulate pollutants that are emitted to the atmosphere depended on the origin, the emission conditions (i.e., flow of distribution near sources), and the exposure to the industrial plumes. In addition to the above-mentioned factors, weather conditions (e.g., wind speed and direction, temperature inversion, anticyclones, and cyclones) may effectively contribute to air pollution. It is well known that pollutants distribution in the atmosphere depended on dispersal phenomena that are, in turn, directly related to meteorological parameters (i.e., daylight hours, wind, temperature and humidity of the air, atmospheric stability, and precipitations [3]). During the atmospheric transport, the composition of original pollutants can be modified by mixtures of substances and aggregates from different origins, with various sizes and chemical compositions depending on the sources, the weather, and the environment [4]. In fact, chemical reactions between pollutants and atmospheric compounds may take place to produce secondary pollutants. In these conditions, the origin of atmospheric pollution becomes quite ambiguous [5]. Therefore, it may be more convenient to study not only the atmospheric composition but also the fallout settling. Within this context, the present work has been undertaken to evaluate the atmospheric particulate fallout instances near the main industrial zone of Sfax City (southern Tunisia). To achieve this goal, solar saltern was chosen as receiving environment because of (1) the homogenous support of the deposits (aquatic surface) and (2) the exposition to the main sources of pollution.

Material and Methods
Spatial and temporal variation of the deposition fluxes above Sfax solar saltern was recorded by monitoring the weekly rate deposition of particulates using collectors installed in twenty sites (i.e., PI2 , TS2, TS35, TS36, TS38, S6, TS3, TS16,  TS20, TS25, PM2, R2, R3, PI1, TS7, TS10, TS12, TS13, TS32, and TS42, Figure 1). These site abbreviations were conserved as predetermined by the technical staff of the solar saltern. Study samples were collected during the period extending from November 11, 2012, to April 15, 2013, during which Sfax area received limited rainfall storms and low Saharan dusts 2 Advances in Meteorology input [22]. Sampling procedure was carried out according to the NF standard 43-007, 2008, by using DIEM plates. "Air Quality-Ambient Air-Determination of the Mass of Dry Atmospheric Depositions-Sampling on Deposit Plates-Preparation and Treatment" clearly described the placement of plates, coated with petroleum jelly, to analyze and subsequently treat the collected samples. The particulate phase dry deposition flux was measured using two smooth deposition plates (22.2 × 7.5 cm 2 for each one) that were pointed at 1.5meter height into a metal holder. The dimensions of each greased strip placed on the deposition plate were 10 × 5 cm 2 . Two greased strips mounted on deposition plates with a total collection area of 100 cm 2 were used for each deposition sample. This type of deposition plate was used successfully as a surrogate surface by others to directly measure particulate dry deposition [23,24].
For the purpose of the present study, deposition plates and strips were washed with alcohol and dichloromethane to remove all the impurities. Finally, strips were dried at 105 ∘ C, adequately numbered, and stored in separate containers for subsequent use. Those strips are made of aluminum (5 × 10 cm 2 ) and covered with petroleum jelly. According to the NF standard 43-007, 2008, strips have to be placed. Afterward, GPS coordinates of all sampling points were registered for further mapping.
After the sampling period of 7 days, strips were collected and submitted to the following laboratory treatments to collect dust samples: (i) Connect the Erlenmeyer flask with screw to the pump.
(ii) Weigh the virgin filter in a precision balance (three weightings).
(iii) Place the filter on top of the Buckner funnel; place the top part and the spring that allows the joining of both parts.
(iv) Using a pipette, place a small amount of dichloromethane on the surface of the strip containing dust.
(v) Scrape with a spatula.
(vi) When the strip is completely cleared, proceed to the filtration connecting the pump.
(vii) Remove the filter from the Buckner funnel and weigh the sample with a precision balance (three weightings).
Nine campaigns of measure were carried out by covering the selected period on a weekly basis. The weekly fluxes of atmospheric fallout instances were expressed in g/m 2 . Meteorological data such as air temperature, humidity, atmospheric pressure, wind velocity, and dominant wind direction was obtained from the meteorological station of Sfax Airport, the closest station to the sampling sites.

Climatic Characteristics of Sfax City
Sfax, a southern city of Tunisia (latitude 34 ∘ 43 N; longitude 10 ∘ 46 E), borders the Mediterranean Sea ( Figure 1). Its semiarid climate is therefore influenced by marine exposure generating a well ventilated area by low to moderate wind, rarely exceeding 5 m/s [22]. However, the influence of both continental and maritime air masses may lead to hot and dry summer with relatively cold and wet winter. Average temperature was estimated to be 12.3 ∘ C for winter season and 24.9 ∘ C for summer. It is to be mentioned that hot season is much longer than cold one, lasting more than five months per year (May-September) with an average temperature >24.2 ∘ C. As for rainfall, the region received about 217 mm/year with high seasonal and annual fluctuations.

Industrial Activities
Sfax is considered as the most industrial city of Tunisia. The main industrial activities included phosphate treatment ("SIAPE" plant), soap manufacturing ("SIOS-ZITEX"), and lead secondary melting industry ("FP Sfax Sud"). Most of those activities are located to the southern edge of the city; they threatened the environment because of their high emission of sulfur oxides (SOx) and particulate matter that largely exceeded the Tunisian standards [21]. SIAPE factory released about 4.5 tons/day of particulate matter with high amounts of sulphate, phosphorous compounds, and SOx, largely exceeding the permissible emission standards (8 and 20 times for particulate matter and SOx, resp.). For instance, soap industry discharged about 5.2 tons per day of particulate matter (>10 times the permissible emission standards value). Concerning the main source of particulate matter emissions, one can adopt the following sequence: SIOS-ZITEX (46%) > SIAPE (39%) > lead melting industry (15%) (Figure 2). In this regard, Azri et al. [25] stated that SIAPE and the lead melting industry were the main sources of heavy metals, as confirmed by the high enrichment factor of particulate matter with regard to heavy metals. They demonstrated that SIAPE was the main source of heavy metals (Zn, Ni, Cd, and Cu) with more than 9 g of Zn per ton of treated raw materials (equivalent to 8 kg/day of Zn emitted in the atmosphere); the emission factors of Ni, Cd, and Cu ranged between 2 and 3.2 g/ton. As for lead foundry, the emission of Pb was evaluated to be 68 kg/ton (equivalent to 200 kg/day of Pb released in the atmosphere). In contrast, the emission factors

Results and Discussions
Atmospheric particulate fallout instances as well as their spatiotemporal distribution were studied in detail to find out their most significant impact on the southern urban zone of Sfax area, southern Tunisia. Depending on current pollutants sources and weather conditions, the obtained results are described below.
Cluster analysis was carried out using the ITCF statistical software package [26] to classify variables (i.e., sites of collection) based on the average linkage method. Pearson correlation coefficients were also used instead of the Euclidean distance that may lead to erroneous conclusions [27]. The establishment of dendrogram related to the study sites allowed deciphering the same main groups previously identified (Figure 4). Figure 5 illustrated the average flow of each site, showing a significant spatial variability of deposits. The linear regression between particulate deposit fluxes and their respective average deviations ( Figure 6) shows that the higher the flux, the higher its standard deviation. The exposure to the industrial plumes (i.e., SIOS-ZITEX, FP Sfax Sud, and SIAPE) of each deposit site indicated that the highest fluxes values were recorded in PI2, TS35, TS36, TS38, and S6 study sites; they are strongly associated with the particulate fallout instances of SIAPE (Figure 7). It is worth noting that the exposure frequency was calculated through the occurrence of industrial plumes crossing each site downstream dominant winds. The methodology was based, first, on the emplacement of the center of the study period dominant wind rose at each point representing the industrial sources presented in Figure 1, second, on the computation of the frequency (%) of dominant Quan-Xin (Taiwan) Industrial zone 2.192 [20] wind directions joining each industrial source and study sites receiving maxima of deposit fallout instances, and, finally, on the establishment of the industrial plume trajectories (the plume axe was chosen as superimposed to the dominant wind direction) simulated by screen 3 software [28] for identifying the other exposed sites. Therefore, they are the most threatened by the aspect of the particulate deposits (quantitative and qualitative aspects). Further examination of the studied sites indicated that "SIAPE" was the main source of metals released in the atmosphere of the Sfax urban zone [25]. between sites. The increase of particulate deposit fluxes can be attributed to the combined effect of both neighboring sources and the airflow properties which has been for a long time governed by meteorological conditions. Figure 8 showed that temporal particulate fallout instances are highly variable. Based on particulate deposits, two main periods (i.e., C2 and C8) can be easily recognized for their high fluxes usually associated with the predominant strong cyclonic episodes. Those fluxes values were much higher than those recorded under stable conditions, supposed to be favourable for the accumulation of various pollutants [29]. During C2 and C8 periods, low surface pressures over Sfax (associated with relatively higher wind speeds) were predominant (Figures 9 and 10); unstable weather conditions can be adapted as possible explanation ( Figure 11). The meteorological factor in favor of high particles deposition fluxes in these periods can be found, on the one hand, in the highest frequencies of industrial plumes,  TS16  TS20  TS25  PM2  R2  R3  PI1  PI2  TS35  TS38  S6  TS2  TS7  TS10  TS12  TS13  TS32  TS36  TS42 Sites of collection Weekly average fluxes of deposits (g/m 2 ) Figure 5: Variation of weekly average fluxes (and their standard deviations) of deposits particles.  TS3  TS16  TS20  TS25  PM2  R2  R3  PI1  PI2  TS35  TS38  S6  TS2  TS7  TS10  TS12  TS13  TS32  TS36  TS42 Sites of collection their dispersion, creating hence risk zones of locally increasing pollution levels [30]. The presence of various obstacles enhanced particles deposition speed [31][32][33]. Previous works at urban Sfax carried out by JICA [34] proved that the actual heights of SIAPE chimneys (between 30 and 70 m) were lower than those necessary for the good diffusion of emissions (equivalent height ≥ 100 m). Furthermore, the existence of phosphogypsum deposit (resulting from processing plants of phosphates) having 20 m height and an area of many hectares, located on the side of Sfax solar saltern at 200 m of the SIAPE, would constitute an obstacle against the emitted effluents (especially by southwest wind directions). Studies carried out by Azri et al. [21] proved that, under extremely unstable atmospheric conditions associated with relatively high wind velocities reaching 8 m/s, very high amounts of various gaseous and particulate pollutants were recorded at both upstream and downstream phosphogypsum deposit, especially at exposed sites to SIAPE plume. Concentrations of such pollutants can be multiplied by a factor of 10. This result was also proved elsewhere [35]. Furthermore, these relatively high wind velocities may enhance earth dust transportation by exceeding the threshold velocity (7 m/s [35,36]). Succinct analysis of the synoptic maps revealed that, since November 26, 2012 (i.e., C2 period), Tunisia has been under the influence of many successive surface depressions that took place over the Mediterranean Sea (Figure 9). These depressions were associated with the Mediterranean front that separated the European cold air masses from the Saharan hot ones. A low-pressure zone over Europe (at 500 hPa) was characterized by strong winds. As for the temperature, it clearly decreased during the subsequent days ( Figure 12). On November 26, 2012, the rapidly evolving surface depression moved towards the east of Italy. From November 27 to December 1, 2012, many successive surface depressions moved towards the east of the Mediterranean Sea.

Influence of the Regional Meteorology and the Effect of the Particular Conditions.
On March 13, 2013 (i.e., C8 period), surface depression was centred over Spain. It was persistent over western Mediterranean Sea until March 16, 2013. Then, it moved on March 17 toward the central and eastern Mediterranean Sea. On March 14 and 15, 2013, the meteorological situation can be described as a vertical superposition over western Spain of surface depressions and low geopotentials. The evolution of temperature at 850 hPa (as seen in C2 period) clearly witnessed a decline in the successive days ( Figure 13). This particular period, associated with relatively high wind speeds (8 m/s), was characterized, over Sfax, by a cloudy sky and low temperatures (a decrease of 3 to 6 ∘ C for maximum registered temperatures at 14:00 LT). These environmental conditions favored the dilution phenomenon that will lease to the decrease of the air constituent concentrations, except for ozone concentrations enriched by the landing stratospheric air masses [37,38]. However, significant increase of particulate deposition was observed over PI2, TS35, TS36, TS38, and S6 study sites, as seen in C2 situation. This can be explained by high frequencies of industrial plumes especially of SIAPE (23% of total observations) and the increase of surface wind velocities (about 8 m/s).
On March 16, 2013, the surface depression and the low geopotential area, under the dominance of western and northwestern winds, moved to the east. Since March 18 2013, it moved towards Eastern Europe while Tunisia returned to stable weather.
Results presented above were refined by a factorial analysis of correspondences applied both to atmospheric particulate fallout instances collected at the study sites (d (PI2) , TS32) , and d (TS42) ) through the nine campaigns and to binary numbers related to registered classes for atmospheric surface pressure, wind velocity, wind sectors, air temperature, and humidity. Three classes 1-3 related to the atmospheric surface pressure were selected (  The projection over the 1 × 2 factorial plane (presenting the maximum of inertia) of all selected particulate deposit fluxes and meteorological variables shows distinct data groups ( Figure 14): (ii) Group G2 was representative of the periods characterized by the effect of the steady to very steady atmosphere. It was pronounced in most campaigns (except for C2 and C8). During these campaigns, atmospheric surface pressure was higher than 1,014 hPa ( 2 and 3), implying a stable to very stable atmosphere. Wind speeds are relatively lower and reach values oscillating between 1 and 5 m/s ( 1 and 2).
, and d (TS42) ) in the remaining study sites (compared to those of the first group) were shown to be influenced by the dominance of the eastern, southern, and western wind's sector ( 1, 2, and 4). The temperature and the humidity were shown without significant effect.
Significant adverse effects of the local industrial activities threatened the environment in Sfax area, especially the southern edge of the city where several activities are being performed. Therefore, it is necessary to undertake a sustainable management policy that would meet the national environmental standards and regulations. More in-depth studies 11    it represented 32% of sites (TS7, TS10, TS2, TS12, TS13,  TS32, and TS42).
Such distribution patterns seemed to be governed by the combined effects between surrounding industrial sources, exposure to industrial plumes, and local airflow characteristics.
Under relatively predominant strong cyclonic situations, the increase in fluxes exceeded the levels recorded under the conditions of relatively strong stabilities that usually enhance pollutants accumulation. The meteorological factor in favour of high fluxes of particles deposition can be found in the highest frequencies of industrial plumes, associated with the relative increase of wind speeds and the effect of the phosphogypsum deposit constituting an obstacle against the emitted effluents.

Conflict of Interests
The authors declare that there is no conflict of interests regarding the publication of this paper.