Inﬂuence of an Extreme Saharan Dust Event on the Air Quality of the West Region of Portugal

: This paper describes how an extreme Saharan dust event that took place in March 2022 affected the Iberian Peninsula and was noticed not only by the outdoor air quality monitoring stations measuring PM 2.5 and PM 10 but also by indoor air monitoring systems in Fatima, central Portugal. The observed particulate matter concentrations clearly show the inﬂuence that such an event has on the indoor air quality inside buildings and that the magnitude of that inﬂuence is also dependent on the speciﬁc characteristics of the buildings, mainly the ventilation conditions, as should be expected. Therefore, this study alerts us to the necessity of integrating indoor and outdoor air quality monitoring systems to achieve automated air conditioning systems capable of efﬁciently controlling both temperature and air cleanliness.


Introduction
Particulate Matter (PM) aerosols originate frequently from semiarid or arid regions as a consequence of continuous soil erosion produced by winds [1]. The strong warming of desert areas during the daytime period produces vertical thermal turbulences that can reach altitudes of about 5 km, followed by subsequent periods of nocturnal stability [2]. Therefore, the resuspension of huge amounts of particulate aerosols is thus produced, and these can be transported over long distances by several different mechanisms. About 40% of the aerosol mass emitted into the troposphere is attributed to desert dust and is currently considered to be the second-largest source of natural aerosols, after sea salt, much more than the amounts usually attributed to anthropogenic PM pollution (which is mostly related to the fine and ultrafine fractions) [3][4][5]. The main desert dust source is the Sahara Desert since it is responsible for more than half of the world's atmospheric particulate dust [6][7][8][9]. Under specific meteorological situations, large amounts of Saharan dust can be transported toward the Mediterranean basin over which the planetary boundary layer is usually relatively shallow [10][11][12][13]. From the point of view of air quality, it has been demonstrated that African dust is the main particle source, contributing to the regional background levels of PM 10 across the Mediterranean (accounting for 35 to 50% of PM 10 ) with maximum contributions up to 80% of the total PM 10 mass [14]. These sporadic but huge natural contributions of PM have been responsible for a high number of excedencies of the PM 10 daily limit value (50 µg/m 3 , according to the 2008/50/EC European Directive [15]) as registered in different rural and urban monitoring sites across the Mediterranean Basin, particularly in Portugal and in Spain [16,17]. The specific origin of the air masses could be determined by backtrajectories simulation using the NOAA/ARL hybrid single-particle Lagrangian integrated trajectory (HYSPLIT) model. That would allow the air mass origin (that could, eventually, be from the Sahara desert and/or from a continental origin in the South of Europe, in this type of event) to be ascertained with certainty. However, the Portuguese Environment Agency (APA, Amadora, Portugal) presented a forecast relating to 15 March 2022, which is one of the days monitored in this study [18]. This study comprises data from the model NMMB/BSC-DUST MODEL of the Barcelona Supercomputing Center, along with HYSPLIT model back-trajectories or the Copernicus data (shown in Figure 2), which are considered necessary for ascertaining the origin of the air mass [19,20]. Additionally, the origin of air masses could be also confirmed by performing chemical characterization of PM 2.5 and PM 10 sampled in filters, since the North of Africa air masses are, typically, characterized by high concentrations of crust elements, such as Al, Fe, and Si [21].
Nevertheless, it can be said that there is a direct relationship between the occurrence of these extreme events in the desert, and the increase in recorded PM levels regarding outdoor air quality [22]. As indoor air quality derives, mainly, from ambient (and thus, outdoor) air, it can be expected that these events will also influence indoor PM levels, but the nature of effects on indoor air quality remains to be investigated, particularly with regard to the recorded increases per size range of PM [23].

Materials and Methods
An extreme event resulting in dust transport from the Saharan region to Portugal took place from the early morning of the 15th of March to the end of the 17th of March, 2022 and was described as causing a considerable decrease in visibility and the appearance of an "orange sky", as can be shown in Figure 1 depicting Lisbon. Figure 2 is a forecast of the PM movement over the Iberian Peninsula for the 15-16th of March, clearly affecting Portugal and Spain.  Outdoor air quality is monitored in Portugal through the National Air Quality Network, operated by the Portuguese Environment Agency (APA, Amadora, Portugal), covering the national territory, as shown in Figure 3. It should be noted that, according to the current European legislation, Directive 2008/50/EC of 21 May 2008 [15], the limit value for PM 2.5 is 20 µg/m 3 (annual average), and for PM 10 it is 50 µg/m 3 (daily average), which must not be exceeded more than 35 times a year. The WHO daily guideline values for PM 2.5 and PM 10 are 15 µg/m 3 and 45 µg/m 3 , respectively, not to be exceeded more than 3-4 days per year [24]. Therefore, the dust event could be said to be noticed from 15 to 17 March per measured PM 10 values, and from 15 to 16 March per measured PM 2.5 . This observation could be related to the size range nature of the transported dust, lying mainly with the PM 10 fraction, instead of the PM 2.5 fraction.
Indoor air quality is routinely monitored in three buildings (designated as 1, 2, and 3) located in Fátima, Leiria (latitude: 39.621820; longitude: −8.688008), corresponding to point D in Figure 3, which are more closely located near station A (but still there is a distance of about 50 km from A to D), and equipped with air conditioning systems. This monitoring is made by specifically calibrated sensors which comprise the automatic patented system Innovair24 that includes Alphasense OPC-N3 (Braintree, United Kingdom) LASER-based particle counters for particulate matter according to the following size ranges: PM 1 , PM 2.5 , and PM 10 . These particle counters were calibrated against a reference dust monitor GRIMM PM 10 /PM 2.5 /PM 1 , model 1.107 (Hamburg, Germany). The estimated accuracy of the Alphasense sensors is, respectively, 11-14%; 17-24%; and 4-5% for PM 1 , PM 2.5 , and PM 10 . This indoor air quality monitoring does follow the most recent ones recommended within the scope of Cost Action CA17136 [25,26]. The Portuguese regulation [27] prescribes the following indoor limit values: (i) 50 µg/m 3 for PM 10 , and (ii) 25 µg/m 3 for PM 2.5 .
Regarding the specific characteristics of the three buildings, building 1 has natural ventilation and the air conditioning system (AC) was not operating for most of the time. Building 3 had the AC functioning, and building 2 had only a mechanical extraction system (no AC) and had people entering and exiting through a door. Table 1 shows the measured concentrations of PM 2.5 and PM 10     The above-named sensors monitored the indoor PM 1 , PM 2.5, and PM 10 concentrations every 5 min during the whole day, in the three buildings, where the dust event is particularly noticeable from the 15th of March to the 17th of March. Figures 5-7 show the evolution of the hourly indoor PM concentrations depicted together with the respective fraction for outdoor PM concentration in station A (note that PM 1 is not measured in station A). Figure 5 clearly shows the measured indoor PM 10 values exceeding the Portuguese limit value in building 2, but not for buildings 1 and 3. Regarding PM 2.5 (Figure 6), the Portuguese limit value is again exceeded for building 2, but not for the other two buildings. Figure 7 shows lower increases over the usual trendlines but still some increase on the measured indoor PM1, while the outdoor PM concentration increases.  The measured results allowed for the calculation of PM 2.5 /PM 10 ratios, which are important to assess the proportion of PM 10 composed of fine particles. As expected, this showed a predominance of the coarse fraction during the dust event, ranging from 0.20 to 0.50, both for indoor and outdoor values, and this is the tendency usually observed for outdoor situations in several cities [28,29]. These ratios were calculated for the three studied buildings and also for outdoor measured values at station A, showing somewhat different profiles as shown in Figure 8.  The indoor/outdoor ratios were not calculated, as the only available outdoor measurements were taken at station A-Ervedeira, which is about 50 km distant from D-Fátima, where the analyzed buildings were located.

Discussion
It should be noticed, as shown in Figure 5, that outdoor PM 10 concentration is considerably higher than measured indoor PM 10 concentrations, in any of the three buildings. However, PM 10 inside buildings 1 and 3 seem to be not particularly affected by the increase in outdoor PM concentrations, which can be due to the fact that building 2 has a mechanical extraction system and air enters by the door which is frequently open while building 3 had an AC system functioning, and in building 1, the AC system was turned off most of the time. AC system filtering system thus provided some PM retention, which is more noticeable in building 1 with the AC system operating all the time.
On the other hand, building 2 indoor PM concentrations are greatly affected by the outdoor PM concentrations, which follow a somewhat similar evolution pattern, and highlight the existence of no filtering system and PM intrusion into the building itself.
Concerning fraction PM 2.5 , as shown in Figure 6, part of the previously mentioned tendency is still observed, but PM 2.5 concentration inside buildings 1 and 3 follows, a bit more closely, the same pattern observed outdoors, which is more evident for building 1 and not so much for building 3. PM 2.5 concentration inside building 2 is much more affected by outdoor concentrations, which are even higher than the measured PM 2.5 outdoor concentrations, at the beginning of the event. However, for the remaining part of the event, all PM 2.5 concentrations are lower than outdoor levels, as expected.
As the PM 1 fraction is not measured in Portuguese outdoor monitoring stations, the outdoor pattern for this fraction is unknown. However, evolution pattern differences were observed between building 3 and the other two buildings: PM 1 concentration, in this building, seems much more unstable, and higher than for the other two buildings.
It can be expected that, for this smaller fraction, filtering efficiency regarding outdoor air is much lower, which can result in observing evolution patterns different from the other size fractions. Nevertheless, the highest concentration is generally observed for building 2, possibly more influenced by the outdoor PM 1 concentrations, as already noticed for the other size fractions.
The calculated PM 2.5 /PM 10 ratios clearly show that the major proportion of PM 10 is composed of coarse particles, especially during the event, as expected and previously reported in the literature [28,29]. Nevertheless, the plots of PM 10 versus PM 2.5 , shown in Figure 8, resulted in different patterns for all studied buildings, which can be easily justified by the existence of differences in their ventilation conditions. Additionally, the obtained patterns are quite different from the one observed for outdoor air in station A, which could be explained by the existence of outdoor wind in the latter situation, and also by the fact that station A is about 50 km distant from the location of the buildings. Additionally, according to the literature [1,2,5], the Sahara dust events (and natural sources, as well) affect mainly the coarse fraction of PM, while the fine and ultrafine fractions are contributed by anthropogenic sources instead. For this reason, it can be expected that the PM 1 profile does not follow entirely the profiles of the other size fractions.
As previously referred to, I/O ratios were not presented due to the distance of the available outdoor concentrations from the indoor ones. However, it is known [26] that, as a general rule, a clear positive relationship between indoor and outdoor PM can be assumed under high ventilation conditions, but not when the ventilation rate is low. Typically, I/O ratios for PM vary from 0.7 to 1.5 [26]. If assuming the limitation of the distance, the I/O ratios were to be calculated, we would obtain, for the majority of cases, values ranging from 0.10 to 0.90, which do not clearly fit within the values to be expected.

Conclusions
This paper presents a quantitative description of how an extreme Saharan dust event that took place in March 2022 [29] affected the West Region of Portugal and was noticed not only by the outdoor air quality monitoring stations measuring PM 2.5 and PM 10 but also by indoor air monitoring systems in Fatima, central Portugal. The observed particulate matter concentrations clearly show the influence that such an event has on the indoor air quality inside buildings and that the magnitude of that influence is also dependent on the specific characteristics of the buildings, mainly the ventilation conditions, as should be expected.
This study also highlights the importance of performing continuous indoor air monitoring. Monitoring is one primary task from a set of actions such as data acquisition by an interactive system that can compare measured values to applicable limit values, and thus generate instructions to regulate and control the indoor air quality by simple actions such as the admission of more or less fresh air coming from outside. Such measures can guarantee that building occupants are not subjected to excessive concentrations of pollutants. It should be noted that this proposed procedure is already adopted in the control of air conditioning systems, but it is not generally used for indoor air quality systems. Such systems require the use of continuous monitoring sensors interactively connected as an upgrade to actual air conditioning control systems used for controlling characteristics such as temperature and moisture. A proposal for a possible architecture of an integrated system for controlling both indoor air quality and thermal comfort is shown in Figure 9.