Comparative aerobiological study between two stations located at different points in a coastal city in Southern Spain

Due to the increase in allergies, aerobiological studies carried out in cities are essential to keep the population informed about the pollen atmospheric concentrations detected. However, the high cost and complexity of aerobiological studies often mean that the information is generated from a single sampling point that may not be representative of the entire city. In this study, the data obtained by two volumetric pollen traps, located in the coastal city of Malaga (Spain) were analyzed. One of the pollen traps was situated in the city center while the other was located on the city outskirts, 5 km away from the first. This was complemented with a meteorological and land use analysis to determine their influence on the pollen concentrations. Despite being located within the same city, the data obtained from both collectors showed significant differences in the relative abundance and annual integrals of the main pollen types, as well as in the periods in which elapse their main pollen seasons. These differences were more notable in the case of Amaranthaceae, Casuarina, Parietaria and Plantago pollen types due to the asymmetric distribution of green areas, agro-forestry areas and urban surfaces within the city, as well as the influence of local wind dynamics on the airborne pollen detected. Despite that, some differences were also observed in the other pollen types. For all the above, we consider that it is important to keep operational several sampling points in cities of a certain magnitude to provide more detailed information about atmospheric pollen concentrations.


Introduction
Pollen is one of the main bioaerosols present in the atmosphere, also being one of the main causes of allergic rhinitis worldwide (Akdis et al., 2015). It is estimated that between 15 and 40% of the population in Europe (depending on the country) suffer from an allergy to some pollen type. This percentage has shown an upward trend in the last 30 years and, in general, use to be higher in children (Akdis et al., 2015;Buters et al., 2018;Clot et al., 2020). Additionally, recent studies show that the symptoms of allergic rhinitis increase with air pollution in cities, due to its interaction with the allergens present in the atmosphere (Bédard et al., 2020). This respiratory disorder affects the quality of life and daily Supplementary Information The online version contains supplementary material available at https:// doi. org/ 10. 1007/ s10453-023- 09786-7. activities of the dwellers of the cities, especially work and school activities, which causes great economic losses to society (Akdis & Agache, 2014;Zuberbier et al., 2014).
The incidence caused by airborne pollen grains in the sensitive population could be reduced by applying appropriate treatments, but also implementing methods for early detection of the pollen concentrations and warning mechanisms, as well as prevention measures (Clot et al., 2020). To implement them, it is crucial to study the effects of meteorological conditions on airborne pollen concentrations, given that they are well known to influence the airborne pollen concentrations detected (Bilińska et al., 2019;Cariñanos et al., 2022;Recio et al., 2010Recio et al., , 2018. Due to the long dedication time that aerobiological studies require, the pollen present in the atmosphere of the different cities is usually estimated from data recorded by a single sampling point. However, given the high complexity of urban environments, the concentrations detected at a given sampling point may not be representative of the entire city or extrapolable to the different points of the same. Besides this, it is necessary to take into account, not only where the population lives but also where the inhabitants carry out their daily activities (Charalampopoulos et al., 2021). For this reason, different authors recommend sampling the air at several city points to increase the precision and usefulness of the released information (Charalampopoulos et al., 2021;Fuentes Antón et al., 2020;Rojo et al., 2020;Werchan et al., 2017Werchan et al., , 2018. The creation of green urban areas, such as parks and gardens, can help to improve the lifestyle of the dwellers and widely favours physical activities, social relationships and reduces stress (Jonker et al., 2014;Lee et al., 2015;Maas et al., 2009;Mitchell & Popham, 2008). Such green spaces also reduce the urban heat island effect, noises and air pollutants related to traffic (Dadvand et al., 2012(Dadvand et al., , 2015. However, they can also constitute a problem for public health given that some of the plants which are cultivated in them produce allergenic pollen grains that can trigger allergy symptoms in the sensitized population (Cariñanos et al., 2017). On the other hand, rural exodus, land abandonment and industrial development are favouring the growth of the cities, as well as changes in land use and environmental degradation, which widely affects the vegetation around the cities, favouring the growth of grasses and weeds (Caldas et al., 2007).
Malaga (Andalusia, Southern Spain), the capital of the "Costa del Sol", is one of the largest coastal cities in the south of the Iberian Peninsula. National and international tourism has increased notably in Malaga in the last decades, which has led to a fast growth of the city as well as the metropolitan areas. Due to being a coastal city surrounded by mountains, this growth has not occurred homogeneously in all directions. The expansion has mainly taken place towards the flat area of the Guadalhorce river valley (west and northwest of the city), with surrounding vegetation dominated by crops and agroforestry areas. This has led to different land use in different points of the city, which has also resulted in a diversity of urban environments with different vegetation (Andrade et al., 2021). Therefore, these disparities in the distribution of pollen sources inside cities are enhancing the differences in the pollen concentrations detected in the different areas. This is why the city of Malaga constitutes an excellent and unusual example to study the influence of land uses and urban environments on the atmospheric pollen content registered in different areas of the cities.
In the present work, a comparative study of the pollen concentrations registered in two different sampling points has been carried out in the city of Malaga, one of them located in the old town (city center) and, the other, in a more modern area of the city outskirts. This study had a threefold aim: (I) to analyse the land use differences in the surroundings of the two pollen sampling locations, (II) to study the relationships between the meteorological variables and the airborne daily pollen concentrations in the city, and (III) to quantify the differences detected between the pollen counts in both sampling points, relating them to differences in land use and the influence of nearby vegetation.

Area of study
Malaga presents an oceanic pluviseasonal Mediterranean bioclimate, with a thermomediterranean thermotype and a dry ombrotype (Rivas-Martinez & Rivas-Saenz, 2020). The average annual temperature is 18.4 °C, one of the warmest in Europe. Rainfalls are scarce, around 540 mm per year on average, mainly concentrated in autumn and winter, with periods of heavy drought alternating with episodes of heavy rainfalls (Domínguez-Amarillo et al., 2020).
The city of Malaga presents great urban contrasts between the center and the periphery due to the rapid growth suffered in the recent years (Andrade et al., 2021). The city center is made up of small and medium-size buildings, and generally narrow streets. It is common to find some abandoned lots or buildings in ruins in which nitrophilous herbaceous plants are detected, such as species of Amaranthaceae and Urticaceae. There are also some parks and gardens with ornamental species. Among them, Malaga Park stands out, situated just 800 m from the pollen trap. In it abounds a great diversity of ornamental species from all over the world, mainly of tropical and subtropical origin (del Cañizo Peralte, 1975;Trigo, 1992;Trigo et al., 1996). Among the wind-pollinated species, shade plane trees (Platanus x hispanica Mill. ex Münchh.), palm trees and some grasses are common. 500 m east of the sampling point we find the Gibralfaro hill, with large areas of pine repopulations (Pinus halepensis Mill.).
On the contrary, on the periphery of the city, large avenues and higher buildings are frequent. There are also many building lots and agro-forestry areas in transition between the typical Mediterranean scrub and cultivated areas (supplementary material, Figure  S1). To the northwest, in the fertile plain of the Guadalhorce river, there are abundant citrus crops, olive groves and orchards (Llamas et al., 2020).

Land use
Land use was obtained from the Corine Land Cover database (version v2020_20u1, data from the year 2018; Copernicus Europe's Eyes on Earth) with a spatial resolution of 100 m. The data were reclassified into 12 categories: Agricultural areas (all crops except olive trees), agroforestry areas (pastures, agroforestry areas and land mainly occupied by agriculture with significant areas of natural vegetation), coniferous forests, green urban areas (urban green areas, sports and leisure facilities), inland water bodies, mixed forests, olive crops, sclerophyllous forests, sea, shrublands and grasslands (including the categories natural grasslands, sclerophyllous vegetation and transitional woodland-shrub), sparsely natural areas (beaches, bare rocks, sparsely vegetated areas and burned areas) and urban areas (all buildings and urban factories but excluding green urban areas).
Using the R software (R Core Team, 2021), the percentage of surface occupied by each land use was calculated in a radius of 1 km around each pollen trap, as well as in rings of 1 to 5 km, 5 to 10 km and 10 to 25 km. Detailed information about the land use distribution in the area can be consulted in the supplementary material ( Figure S1).

Meteorological data
The meteorological data used for this study were obtained from the Malaga-Airport station, belonging to the Spanish National Agency of Meteorology (AEMet), located 4.64 km and 8 km away from the pollen sampling locations in the city outskirts and center, respectively. This station was the nearest official meteorological station having complete records for the sampling points and period studied. For this study, daily records of maximum, minimum and mean temperatures were considered, as well as relative humidity, precipitation, wind speed and frequencies of winds blowing from the different quadrants.
Spearman correlation tests were performed between the daily records of the aforementioned meteorological variables and the daily pollen concentrations during the MPS at both sampling locations in order to determine possible associations between them. Benjamini and Hochberg (1995) post hoc correction was applied to the significance level to reduce type I errors in the results. Given that meteorological variables usually have opposite effects on the pollen concentrations during the pre-peak and post-peak periods of the MPS (Recio et al., 2018), the correlation analysis considered both periods separately.

Pollen data
The airborne pollen samplings were carried out uninterruptedly during the period 2017-2019, by using two volumetric Hirst-type pollen traps (Hirst, 1952) installed at different points of the city, separated by a distance of 5 km. One of them was set in the old town (city center), on the plain roof of the multi-services building of the Malaga City Council in Dos Aceras st. (36°43′32.54"N 4°25′15.08″ W; 29 m.a.s.l.; 12 m.a.g.l.); while the other (Teatinos) was located in the western part of the city, on the roof of the Faculty of Sciences of the University of Malaga, situated in the Campus of Teatinos, a peripheral area, on the edge of the city (36°42′58.09"N 4°28′21.86″ W; 62 m.a.s.l.; 20 m.a.g.l.) (Fig. 1).
The samples were mounted and counted following the standardized methodology proposed by the Spanish Aerobiology Network (REA) and the European Aeroallergen Society (EAS) (Galán et al., 2007(Galán et al., , 2014. The pollen traps were adjusted to have a continuous flow of 10 l/min, with silicone fluid being used as adhesive substance and glycerine jelly as mounting mean. Four longitudinal sweeps were analyzed for each sample at a magnification of X400 with the aid of a light microscope. The data used in the figures and tables correspond to the daily mean pollen concentrations, expressed as the number of pollen grains per cubic meter of air (pollen grains/ m 3 ). The results were analyzed using the statistical program SPSS v21 and the AeRobiology package, implemented in the R software (R Core Team, 2021; Rojo, Picornell, et al., 2019a. The study is focused on the most abundant pollen types in the atmosphere of Malaga, which were defined as those that exceeded 1% of the annual total pollen detected: Amaranthaceae, Cupressaceae, Olea europaea, Pinus, Plantago, Platanus, Poaceae, Quercus, Urtica membranacea, Parietaria and Casuarina, the last one only in the case of the peripheral station. The pollen type Parietaria corresponds to the Meteorological data for calculating wind frequencies were obtained from the nearest sampling station of the Spanish National Agency of Meteorology (AEMet) (Malaga airport). CRS WGS 84 whole Urticaceae, excluding U. membranacea which presents a different pollen type.
The relative abundance of the main pollen types for the 2017-2019 period was calculated based on the total pollen detected during the three years of the study. The daily mean concentrations from the two sampling points were compared using Spearman's correlations and Mann-Whitney-Wilcoxon tests. For these statistical tests, the days included in the main pollen season (MPS) at both sampling points were considered.
For calculating the MPS, the 95% method was used, which defines the beginning as the day in which 2.5% of the annual pollen integral is reached, and the end when 97,5% of the annual total is accumulated (Andersen, 1991). In the case of Casuarina and Cupressaceae, since they are pollen types that present their highest concentrations between the end of one year and the beginning of the next one, the period August to July (overlapping years) was used instead of natural years. To calculate the MPS of Casuarina, Parietaria and Urtica membranacea, as they are pollen types with long tails at the beginning and the end of their pollination period, the 90% method was used (Nilsson & Persson, 1981). The peak date was defined as the day when the highest annual pollen concentration is detected in each sampling location. This date splits the MPS into pre-peak and post-peak periods.
The annual pollen integrals of the main pollen types were also calculated in order to establish differences between the two sampling stations as well as to detect variations among years (i.e. interannual fluctuations).

Land use
The types of land use detected in the surroundings of the two pollen traps were similar, but the percentage of surface occupied by each of them was variable depending on the distance to the sampling point ( Fig. 2). Within a 1 km radius around the collector, we can find in the city center a clear predominance of built-up areas, together with a small percentage of green urban areas corresponding to the central park (Parque de Málaga) and other gardens distributed throughout the old city. On the contrary, in the city outskirts, parks and gardens had less prominence, with an increase in the percentage of the agroforestry areas, corresponding, in part, to the Guadalhorce Valley where agricultural areas are mixed with natural vegetation, in which herbaceous plants and some shrubs predominate.
Within the ring situated between 1 and 5 km away from the sampling sites, remarkable differences were observed in the land uses. The city center of Malaga has a greater proportion of the area occupied by coniferous forests, due to the repopulation of pines carried out in the mountains near the historic center of the city. In addition, due to its proximity to the coast, the Fig. 2 Comparative graph of the percentage of surface occupied by the different land uses in the surroundings (radius of 1 km, rings of 1-5 km, 5-10 km, and 10-25 km) of the two aerobiological sampling stations: center and periphery, in the city of Malaga percentage of surface occupied by the sea is notably greater in the center than in the periphery. In the city outskirts, the percentage occupied by urban areas is greater, as it also covers the areas of industrial estates situated on the outskirts of the city and other nearby coastal towns. The percentage of agro-forestry use is also notably higher in the periphery than in the city center, as well as the percentage of surface occupied by olive groves.
Within the ring located between 5 and 10 km away from the pollen traps, we observe how in both sampling points the percentages of land use are similar in terms of urban areas, coniferous forests, scrublands and grasslands, agro-forestry areas, and the influence of the sea. However, in the city outskirts, the use of areas dedicated to agriculture predominates. Finally, in the range between 10 and 25 km, the percentages of land use are similar in both sampling points, with a predominance of the area occupied by the sea, followed by agricultural and scrub and pasture areas.
Comparing the differences observed in land use with a similar study carried out in central Europe with two pollen traps in Munich , we found that, in our case, such differences were much more notable in the area of 1 km around the pollen trap and the ring within 1-5 km. This highlights the notable effect of the asymmetric growth that the city has experienced in the last years.

Correlations with meteorological variables
The correlations obtained between meteorological variables and daily mean pollen concentrations showed very similar results in both sampling locations (Fig. 3). In general, negative and significant correlations were detected between the airborne Significance levels: p ≤ 0.05 (*), p ≤ 0.01 (**), p ≤ 0.001 (***). RelHum, daily mean relative humidity; Rainfall, daily total rainfall; Tmean, daily mean temperature; Tmax, daily maximum temperature; Tmin, daily minimum temperature; WindSpeed, mean daily wind speed; Quad1, 2, 3 and 4, frequency of winds blowing from the 1st, 2nd, 3rd and 4th quadrants, respectively; Calm, wind calm frequency pollen concentrations and the relative humidity in most pollen types. High relative humidity promotes binding among airborne particles (pollen included), inducing the precipitation of pollen grains and the consequent reduction in their airborne concentrations. A similar effect is observed with rainfall, since it cause atmospheric washing and favours pollen precipitation, which is reflected in the negative significant correlations obtained (Piotrowska-Weryszko et al., 2021;Ravindra et al., 2022;Recio et al., 2018;Rojo et al., 2015;Schramm et al., 2021).
Regarding the correlations with temperatures, opposite results were generally observed during the pre-peak and post-peak periods. Increasing temperatures favour the flower aperture and anther dehiscence once they are developed increasing the pollen concentrations. On the contrary, when flowers are already opened, high temperatures promote premature senescence, reducing the pollen concentrations detected (Picornell et al., 2019aRecio et al., 2018;Subba Reddi & Reddi, 1985). In addition, most pollen types present their MPS in spring, when temperatures are rising. During the pre-peak period, temperatures are increasing at the same time as pollen concentrations do, while during the post-peak period temperatures continue increasing when pollen concentrations decrease. This explains the positive significant correlations observed between the temperatures (i.e., mean, maximum and minimum) and the airborne pollen concentrations for most pollen types during the pre-peak, as well as the negative significant ones detected during the post-peak. This general pattern was observed in all pollen types except Casuarina, which had opposite correlations. Casuarina usually flowers during autumn when temperatures are dropping.
The wind had a determinant role in the pollen concentrations. High wind speed was associated with days in which high pollen concentrations were registered (i.e., a positive significant correlation). On the opposite, in days with a high percentage of wind calm, the pollen concentrations were lower (i.e., negative significant correlations). The only positive significant correlations detected with the calm frequency were in the case of Amaranthaceae pollen in the city outskirts during the pre-peak period, probably due to the pollen concentrations detected during this period come from the ruderal vegetation surrounding the pollen trap whose pollen grains does not require to be transported far away by the wind.
On the contrary, high pollen concentrations during the post-peak period seemed to be related to pollen coming from further distances (positive correlation with the wind speed and with the winds blowing from the 3rd and 4th quadrants).
In days when the wind blew from the sea (2nd quadrant), pollen concentrations are, in general, lower (i.e. negative significant correlation), as reported in previous studies (Recio et al., 1996(Recio et al., , 2018Trigo et al., 1997). Cupressaceae pollen at the city center was the exception to this pattern, showing a positive significant correlation with the winds blowing from the 2nd quadrant. This may be explained because the Cupressaceae pollen usually came from ornamental trees that are abundant in the parks and gardens of the city center, most of them being located at the southeast of the pollen trap ( Figs. 1 and S1).
The correlations with the other pollen types, when significant, generally were positive with the 3rd and 4th quadrants, matching the dominant winds in the city as well as the distribution of the main natural and seminatural vegetation (Figs. 1 and S1). In the case of Parietaria and Urtica membranacea these patterns may be explained by the presence of old buildings in the case of the city center, and nitrophilous wastelands in the case of the city outskirts.

Differences in pollen detection between both sampling locations
Regarding the total annual pollen records, no clear pattern was observed for all the sampling years (Fig. 4). In 2017, a greater APIn was obtained in the periphery, while in the other years, the opposite occurred. A longer time series would be necessary to exclude the particularities of each year and discern if a sampling point generally registers higher integrals than the other. However, the directionality of the interannual changes in these values was the same at both sampling points (i.e. they increased or decreased at the same time at the two sampling points). The 11 main pollen types constituted, as a whole, 89.43% of the total pollen collected during the study period. However, despite these pollen types were the same during the 3 years at both sampling points (except for Casuarina), slight differences were detected in the order of abundance from one year to another. Furthermore, the order of abundance of the different pollen types changed depending on the sampling location. In general, almost all pollen types had a higher relative abundance in the city center, compared to the periphery (Fig. 5). These differences between sampling sites in a same city were much more evident for almost all pollen types than in a previous study carried out in Cordoba (southern Spain) (Velasco-Jiménez et al., 2013).
As previous studies highlighted, the abundance of the pollen types detected in certain city areas is highly conditioned by the abundance of the emission sources in the surroundings, ornamental trees and local meteorological conditions (Bilińska et al., 2019;Charalampopoulos et al., 2021;Fuentes Antón et al., 2020;Pecero-Casimiro et al., 2019;Velasco-Jiménez et al., 2013;Werchan et al., 2018). This may explain the differences observed in the order of abundance of the pollen types detected as previous studies suggested (Bilińska et al., 2019;Fuentes Antón et al., 2020). Besides that, the relative abundances are also influenced by the amounts of total pollen detected in each sampling location, so the analysis of these results must be done jointly with those of the annual integrals (Fig. 6).
As can be seen in Fig. 6, the values of the annual pollen integrals showed differences from year to year, depending on the sampling station and the pollen type. Parietaria stands out, always presenting much higher values in the city center, probably due to the greater abundance of ruined buildings, which favours the proliferation of nitrophilous herbaceous plants, as is the case of Parietaria, which usually grows in abandoned lots and old walls. This explains not only its higher annual integrals in the center of the city but also its greater relative abundance (Figs. 5 and 6). Similar results were observed when comparing the inner city and the outskirts of Cordoba (Velasco-Jiménez et al., 2013). On the contrary, Amaranthaceae, Casuarina and Plantago always presented higher values in the city outskirts. Amaranthaceae and Plantago are abundant plants in agro-forestry and crop transition zones that are more abundant in the outskirts of the city (supplementary material, Figure S1), which may explain the higher integrals detected at this sampling point. On the other hand, Casuarina is a common ornamental tree in the Teatinos university campus, where the pollen trap is installed, which would explain the highest integrals detected in the city outskirts. A previous Fig. 6 Values of the annual pollen integral in both sampling sites during the period studied (2017-2019). Units expressed in pollen grains*day/m3 study carried out in the center of the Iberian Peninsula also found similar patterns in the distribution of the airborne pollen concentrations of these pollen types in the city due to differences in land use and the distribution of ornamental trees (Fuentes Antón et al., 2020).
Regarding the remaining pollen types, they reached higher annual pollen integrals in one or another sampling location depending on the year and did not present so clear patterns as in the case of the mentioned above, despite the fact that some of them presented important differences in terms of relative abundance. This may be explained by a similar abundance of the emission sources of these pollen types at both sampling sites, as well as by prevailing wind dynamics during the different years, which would be determinant to the airborne pollen concentrations detected, as suggested by previous studies (Bilińska et al., 2019;Fuentes Antón et al., 2020).
In the case of Cupressaceae, the annual integrals presented scarce differences between the sampling points, being slightly higher in the city center in the years 2017 and 2019, and slightly higher in the periphery in 2018. This similarity is probably due to the fact that cypresses are common ornamental plants in the vicinity of both sampling sites, and several authors have suggested their influence on the local airborne pollen concentrations (Charalampopoulos et al., 2021;de Linares et al., 2021;Monroy-Colín et al., 2020;Velasco-Jiménez et al., 2020). A similar pattern was observed for Cupressaceae pollen type in a previous study carried out in Cordoba (Velasco-Jiménez et al., 2013).
The annual integrals of Olea europaea were similar in both sampling points in 2018 and 2019, but much higher in 2017 in the city outskirts. The differences observed during this last year could be related to different olive groves' s surfaces within the 1-5 km radius in both sampling locations, together with punctual variations in the meteorological conditions, especially wind dynamics. Something similar happened with Quercus pollen type: its annual integral was higher on the periphery in 2017, similar in 2018 at both sampling points, and higher in the center in 2019. The forests that contribute to the increase in this pollen type are found in similar abundances around both pollen collectors (Fig. 3), so the differences detected over the years are probably also conditioned by wind dynamics.
In the case of Pinus, the annual integrals were similar in both locations in 2017, higher in the periphery in 2018 and higher in the center in 2019. Despite their proximity, the repopulation pine forests located in the Montes de Malaga, towards the northwest of the city do not seem to have had a great influence on the quantities detected in the city center pollen trap. This may be due to the fact that the prevailing winds in Malaga are those blowing from SE and NW, and they would not contribute to the transport of pollen from these pine forests to the sampling sites (Recio et al., 2018). In this case, the main source of pollen would be small copses located within the parks and peri-urban areas, frequent in the vicinity of both sampling sites (supplementary material, Figure S1).
The annual integral of Platanus was higher in 2017 in the city center but in 2018 and 2019 it was higher in the periphery. The integral of Poaceae was practically similar in the two sampling points throughout the period studied, and the integral of U. membranacea was higher in the center in 2017, higher in the periphery in 2018 and similar in both sampling sites in 2019. Platanus is represented in the area by ornamental trees whose abundance is similar in both sampling points, and they are the unique emission source of this pollen type. Therefore, the APIn detected are expected to be similar, with punctual changes depending on wind dynamics (Alcázar et al., 2004;Cariñanos et al., 2020;Lara et al., 2020;Maya-Manzano et al., 2017;Pecero-Casimiro et al., 2019).
Grasses, due to their wide distribution and the high number of species, are present in a large number of different land uses, therefore, their pollen concentrations in the air tend to be similar at both locations, as previously reported by Fuentes Antón et al. (2020) in the city of Salamanca (central-nortwest Spain). Something similar happens in the case of U. membranacea, as it is a nitrophilous ruderal plant, it does not present a special predominance in one or another sampling point, apart from the influence that atmospheric conditions and wind trajectories may exert.
Although several studies have evidenced the influence that different heights above ground level could have on the pollen concentrations detected in the atmosphere, these differences tend to be insignificant when the collectors are located more than 10 m above ground level Rojo, Oteros, et al. 2019). Therefore, we consider that the influence of the scarce difference in the heights of the pollen traps used in this study has been of little relevance to the results obtained. On the other hand, in other localities, land use and pollen emission sources located in the vicinity of cities have also been identified as the main variables responsible for the differences detected between nearby sampling points (Fuentes Antón et al., 2020;Pecero-Casimiro et al., 2019;Picornell et al., 2019aPicornell et al., , 2019bRojo et al., 2020;Velasco-Jiménez et al., 2013).
At both sampling points, about 85% of the total annual pollen is concentrated from February to June (both inclusive), this being the period in which the highest daily pollen concentrations of most main pollen types took place, with the exception of Casuarina, which is typically autumnal (Fig. 7). In general, May is the month in which the highest concentrations of the total pollen are detected at both sampling points, a typical behaviour in the Mediterranean basin. Although almost all the pollen types presented a single period of maximum concentrations, some of them such as Cupressaceae or Amaranthaceae, presented two peaks of different intensities. In the case of Cupressaceae, two periods of high concentrations were detected, the first, more intense, from January to May and the second, of lesser intensity, from September to December, due to the presence of different species, some of them with autumn flowering. Something similar occurs in the case of Amaranthaceae, whose maximum concentrations occur between April and May and, after a marked decrease, they slightly increase again at the end of summer (August-September). In any case, sporadic detection of pollen types is frequent throughout the year, due fundamentally to re-buoyancy phenomena.
As previously observed for relative abundances and annual integrals, mean daily concentrations, in general, were also higher in the periphery for the pollen types Amaranthaceae, Casuarina, and Plantago, and higher in the city center for Parietaria (Fig. 7). This last pollen type showed the greatest differences between the two sampling points, especially from March to July. The rest of the pollen types presented similar daily average concentrations throughout the year at both sampling points, despite the fact that some peaks of greater intensity were sporadically detected at one or another site. In general, the Fig. 7 Daily mean pollen concentrations of the main pollen types for the period studied (2017-2019) at the two sampling sites timing and intensity of the daily average concentrations detected were similar to those of other nearby cities located in the south of the Iberian Peninsula, such as Granada, Ronda or Cordoba (de Linares et al., 2019;Picornell et al., 2019aPicornell et al., , 2019bVelasco-Jiménez et al., 2018).
Comparing the graphs of seasonal behaviour of the different pollen types in both sampling points, we observe that there is a great coincidence in terms of their profiles. In fact, the correlation studies carried out between the mean daily concentrations of the different pollen types at both sampling points (center and outskirts), resulted in positive and significant correlation coefficients for p ≤ 0.001 in all cases (Table 1). However, after applying Mann-Whitney-Wilcoxon tests, significant differences (p < 0.05) were observed in 6 of the 11 pollen types studied (i.e. Amaranthaceae, Casuarina, Olea europea, Plantago, Quercus, and Parietaria). Significant differences were also observed when comparing daily airborne pollen concentrations in other cities of Spain (Fuentes Antón et al., 2020;Velasco-Jiménez et al., 2013).
Given the proximity of both sampling points, we can expect that the meteorological conditions that influence pollen concentrations in the atmosphere have similar variations and therefore, the concentrations oscillate in the same direction, even when the absolute values are different due to differences in emission sources or air flows (Table 1). The significant differences observed according to the Mann-Whitney-Wilcoxon tests were probably due to the fact that, repeatedly, concentration peaks detected in one of the sampling points were not detected in the other, and vice versa.
Regarding the MPS of the pollen types studied (Table 2), differences between both sampling stations have been observed in the dates of start and end, in the duration of the whole MPS and the periods pre-and post-peak, as well as in the dates and intensity of the peak days. Actually, few coincidences in dates have been found, and sometimes these differences are of up to 30 days between one and another sampling site. Differences in the beginning, end and duration of the MPS may be related to the methodology used, which is based on percentages and, therefore, is dependent on the value reached by the APIn of the pollen type in question, but this would not be the case of the peak day or their intensity and they have also been variable.
Some pollen types present their maximum daily peaks earlier in the periphery, as occurred in the case of Amaranthaceae and Parietaria. The rest of the pollen types did not present a clear behaviour regarding the date on which the peak day takes place, occurring first in the center or the periphery depending on the year. In the case of Quercus, the differences are less than 10 days. However, for the rest of the pollen types, the fluctuations were greater. For example, in Platanus the fluctuations oscillated between 1 and 19 days; in U. membranacea between 6-27 days; and in Parietaria, between 14-81 days (Table 2). Usually smaller differences in the MPS parameters were observed in other sampling locations of Spain and central Europe (Fuentes Antón et al., 2020;Rojo et al., 2020).
The peak day values of the Urtica membranacea and Parietaria pollen types were higher during the entire study period in the city center, while Amaranthaceae, Casuarina, Pinus and Plantago obtained higher peak concentrations in the city outskirts, which seems to be related to the higher abundance of these species in the vicinity of the pollen sampler. The remaining pollen types did not show a homogeneous pattern, so the values of their peaks seem to be more conditioned by wind dynamics than by any specific land use pattern.  Previous studies also reported differences in the start, peak and end dates of the MPS when sampling at different locations inside the same city due to wind dynamics and the asymmetric distribution of the pollen emission sources (Charalampopoulos et al., 2021;Fuentes Antón et al., 2020;Velasco-Jiménez et al., 2013).
Summarising, the study carried out showed differences in the results obtained at both sampling points, not only in timing but also in the values reached by the different pollen types and their relative percentages. Due to the scarce number of sampling years, it is not possible to perform statistics with annual values such as the APIn or the MPS parameters until more data be available. However, these differences were very noticeable in the case of some pollen types, such as Casuarina, Parietaria, and Plantago. Previous studies have reported differences between sampling points located in the same city (Charalampopoulos et al., 2021;Fuentes Antón et al., 2020;Rojo et al., 2020;Velasco-Jiménez et al., 2013), but the differences observed in land use surfaces near the pollen sampling locations are unprecedented. The different land use surfaces of the surrounding places in which the two samplers are located, including parks and gardens and proximity to the sea, seem to have influenced the results obtained. In fact, the influence that the vegetation close to the pollen traps has on pollen concentrations has been already highlighted by other authors such as Kasprzyk et al. (2019) and Rojo et al. (2020). However, the influence of the prevailing winds or the structure of the city itself could explain some of the differences found since the position of the buildings constitutes barriers that can modify the local wind patterns and affect pollen dynamics (Bilińska et al., 2019;Ciani et al., 2020;Damialis et al., 2005;Gonzalo-Garijo et al., 2006;Pecero-Casimiro et al., 2019;Peel et al., 2014). This has led to unusual differences in the detection of pollen types such as Parietaria within the same city that had not been previously reported. This fact brings to the fore the necessity of performing comparative studies like this in different cities to understand the local pollen dynamics and their particularities.

Conclusions
Meteorological variables such as temperature, relative humidity, precipitation and wind patterns played an important role in determining daily airborne pollen concentrations in the city of Malaga.
Despite the short distance existing between both sampling points (5 km), the results obtained showed many differences in terms of annual pollen integrals, relative abundance, main pollen season features and dates on which the maximum peaks occur. The statistical study showed that, although there was a good correlation between the daily mean pollen concentrations of both sampling points for the 11 pollen types studied, there were significant differences in the case of 6 of them.
Some of the differences found could be explained by the different percentages of land use present in the surroundings of the sampling points, including the influence of the sea. However, the wind direction, the orography and the structure of the city itself could have contributed to the differences detected.
Based on the results obtained, in cities with a certain size and heterogeneity in terms of land use and urban characteristics, it would be necessary to install several aerobiological sampling points in order to provide more precise and detailed information to the population, which would contribute to prevent and treat respiratory allergies more effectively.
Acknowledgements This work was partially financed by the Environmental Sustainability Area of the Málaga City Council, by the Ministry of Science and Innovation of Spain and FEDER fundings inside the Operational Plurirregional Program of Spain 2014-2020 and the Operational Program of Smart Growing (Project Environmental and Biodiversity Climate Change Lab, EnBiC2-Lab), and by the University of Málaga under its program for projects leaded by young researchers (I Plan Propio de Investigación y Transferencia; B1-2021_24). A. Picornell was supported by a postdoctoral grant financed by the Ministry of Economic Transformation, Industry, Knowledge and Universities of the Junta de Andalucía (POSTDOC_21_00056).
Author contributions RR-M wrote the main manuscript, prepared the figures, and collected data. MMT reviewed the manuscript, administrated the project and was responsible of funding acquisition and resources. MR and EG-M reviewed the manuscript and collected data. AP collaborated in the elaboration of figures, reviewed the manuscript, and coordinated the study.

Conflict of interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.