Variability and Geographical Origin of Five Years Airborne Fungal Spore Concentrations Measured at Saclay, France from 2014 to 2018

Airborne fungal spores (AFS) represent the major fraction of primary biological aerosol particles (PBAPs), and they are studied worldwide largely due to their important role within the Earth system. They have an impact on climate and human health, and they contribute to the propagation of diseases. As their presence in the air depends largely on studied ecosystems, a spore trap was used to monitor their atmospheric concentrations from 2014 to December 2018 in Saclay, a suburban area in the megacity of Paris. The main objective of this work was: (1) to understand the atmospheric variability of AFS in relation to different variables such as meteorological factors, agricultural practice, and (2) to identify their geographical origin by using a source receptor model. During our period of observation, 30 taxa have been identified under a light microscope. In order of importance, Ascospores, Cladosporium, Basidiospores, Tilletiopsis, Alternaria were found to be the most abundant types respectively (50.8%, 33.6%, 7.6%, 1.8%, and 1.3%) accounting for 95% of the atmospheric concentrations. We observed a general decrease associated with a strong interannual variability. A bimodal seasonal cycle was identified with a first maximum in July and a second in October. The main parameters driving the atmospheric concentration are temperature and precipitation. The daily variability is strongly activated by successive periods of hot weather and rainfall, multiplying the concentration by a factor of 1000 in less than 12 hours. Results from the source receptor model ZeFir point out unambiguous different origins of AFS due to specific sources impacting the observation site. Our study also indicated that a hydrological stress has a direct effect on the daily concentrations. This last point should be taken into account for every stressed ecosystem studied in a global warming context. This is particularly important for Mediterranean areas where water is a key control of the growth and dispersion of fungal spores.

(EAN), it strictly follows the recommendations regarding the minimum requirements for the counting procedure. The standard analytical method used accounted for 10% of the surface area. Standard sampling, processing and analysis techniques have been well documented [45]. The quality assurance and the quality control of theses method were regularly made during the international intercomparison campaigns [46]. The daily concentrations obtained are expressed as concentrations of fungal spores per cubic meter of air. Analogous to pollen studies, the Main Spores Season (MSS) is defined as the duration time when fungal spores are present in the atmosphere in significant concentrations in a specific ecosystem. Therefore, it defines the main season start and end points. The MSS used in this study is the period during which the sum of daily mean fungal spore concentrations stands between 5% and 95% of the total sum. The terminology and the selection criteria used in this study follow the recommendation of [47].

Investigation of the Geographical Origins of Airborne Fungal Spores
The investigation of the geographical origins of AFS was performed by coupling ambient concentrations with on-site measured wind data. At the Saclay observatory, meteorological parameters are provided by a weather station WXT520 (Vaisala, France). The measurements of wind speed (WS, m/s), wind direction (WD, Degrees), temperature (T, • C), relative humidity (RH, %), and cumulative rain (R, mm) are acquired every minute. A variant of two-dimension non-parametric wind regression (NWR) originally developed by [48], called the sustained wind incidence method (SWIM) developed by [49] has been used to identify the geographical origin of AFS. This variant takes into account the standard deviation of the wind speed and the wind direction on a daily basis.
Equation (1) below describes the calculation of SWIM, where C i , Υ i , and δ represents respectively wind speed, wind direction, and wind direction standard deviation. This actually allows downwind daily concentration values associated with high atmospheric variability to be obtained during that day. Wind direction standard deviation was estimated by the 1-pass Yamartino equations [50]. This entire study was performed with ZeFir, a user-friendly tool for wind analysis [38]. More information can be found here: https://sites.google.com/site/ZeFirproject. This method was successfully applied for the first time to determine the geographical origin of pollen by [39].

Meteorological Aspects of the Saclay Ecosystem
France belongs to the temperate climate zone and four major climate types influence the territory regarding the climate over more than 30 years period [51]. The main one is in the west, mainly influenced by the ocean (Figure 1a, Oceanic, and Oceanic degraded). This climate is characterized by significant rainfall during autumn, winter and spring periods; less precipitation is observed during the summer period. Mean temperatures vary from 5 • C in winter to 21 • C in summer, as illustrated in Figure 1b. The Saclay observation site is dominated by an oceanic degraded climate (heavy rainfall during the dormancy period), relatively mild winter conditions with less than 47 days below 0 • C per year occurring between November to March and a relatively hot summer. The Eastern region is characterized by a continental climate and the south of France exhibits a typical Mediterranean climate. The French mountains have their own climate characteristics, mainly driven by the altitude where temperatures are lower and precipitation higher on the slopes exposed to winds charged with humidity. A correct understanding of the ecosystem studied is necessary to grasp fungal spores purpose as their atmospheric concentration can be extremely variable [7].

2.5.Wind Prevalence at Saclay
Wind is an important factor for the dispersal and transport of (bio)aerosols far from their sources [35]. In order to better characterize the Saclay prevailing winds pattern between July 2014 and December 2018 the ZeFir source-receptor tool was used. The wind rose presented in Figure 2 indicate that air masses generally originate from west to south-west (oceanic), with speeds ranging between 5 to 12 km/h. A second wind regime is characterized by north (5°) to south-east (125°) winds, at speeds ranging from 2 to 7 km/h, bringing rather sunny skies, dry air and higher temperatures as reported by [39].

Results
In our study the complete identification of the fungal propagules at a genus level was not possible due to the used identification method (e.g 2.2). At this stage of our study no DNA or RNA analysis were done to complete the observations and particularly for Ascospores, Basidiospores and some optically unidentified spores. As it is the first time that these results are presented, they need to be compared with the literature to interpret the main parameters driving the AFS concentration in the Saclay ecosystem. This step is as a necessary prerequisite to use The Eastern region is characterized by a continental climate and the south of France exhibits a typical Mediterranean climate. The French mountains have their own climate characteristics, mainly driven by the altitude where temperatures are lower and precipitation higher on the slopes exposed to winds charged with humidity. A correct understanding of the ecosystem studied is necessary to grasp fungal spores purpose as their atmospheric concentration can be extremely variable [7].

Wind Prevalence at Saclay
Wind is an important factor for the dispersal and transport of (bio)aerosols far from their sources [35]. In order to better characterize the Saclay prevailing winds pattern between July 2014 and December 2018 the ZeFir source-receptor tool was used. The wind rose presented in Figure 2 indicate that air masses generally originate from west to south-west (oceanic), with speeds ranging between 5 to 12 km/h. A second wind regime is characterized by north (5 • ) to south-east (125 • ) winds, at speeds ranging from 2 to 7 km/h, bringing rather sunny skies, dry air and higher temperatures as reported by [39]. The Eastern region is characterized by a continental climate and the south of France exhibits a typical Mediterranean climate. The French mountains have their own climate characteristics, mainly driven by the altitude where temperatures are lower and precipitation higher on the slopes exposed to winds charged with humidity. A correct understanding of the ecosystem studied is necessary to grasp fungal spores purpose as their atmospheric concentration can be extremely variable [7].

2.5.Wind Prevalence at Saclay
Wind is an important factor for the dispersal and transport of (bio)aerosols far from their sources [35]. In order to better characterize the Saclay prevailing winds pattern between July 2014 and December 2018 the ZeFir source-receptor tool was used. The wind rose presented in Figure 2 indicate that air masses generally originate from west to south-west (oceanic), with speeds ranging between 5 to 12 km/h. A second wind regime is characterized by north (5°) to south-east (125°) winds, at speeds ranging from 2 to 7 km/h, bringing rather sunny skies, dry air and higher temperatures as reported by [39].

Results
In our study the complete identification of the fungal propagules at a genus level was not possible due to the used identification method (e.g 2.2). At this stage of our study no DNA or RNA analysis were done to complete the observations and particularly for Ascospores, Basidiospores and some optically unidentified spores. As it is the first time that these results are presented, they need to be compared with the literature to interpret the main parameters driving the AFS concentration in the Saclay ecosystem. This step is as a necessary prerequisite to use

Results
In our study the complete identification of the fungal propagules at a genus level was not possible due to the used identification method (e.g 2.2). At this stage of our study no DNA or RNA analysis were done to complete the observations and particularly for Ascospores, Basidiospores and some optically unidentified spores. As it is the first time that these results are presented, they need to be compared with the literature to interpret the main parameters driving the AFS concentration in the Remote Sens. 2019, 11, 1671 6 of 29 Saclay ecosystem. This step is as a necessary prerequisite to use Zefir source receptor modeling tool to identify the geographical origin of airborne fungal spores impacting the observation site.

Interannuality of the Airborne Annual Fungal Spore Integral at Saclay
The interannuality of AFS Integral (AFSIn) was studied in depth. The annual sum of fungal spores ranges from 2.5 10 6 Nb#/m 3 to 5.2 10 6 Nb#/m 3 . We observed a strong interannuality of the AFSIn and no biannual cycle has been found. Instead, a general decrease of the atmospheric concentration was found, as illustrated in Figure 3. The AFSIn data from January to June 2014 have been estimated by a linear modeling with the most basic model which is the simple linear regression where a variable X is explained and modeled by an affine function of a comparable variable Y. The interpolation of the Saclay data set from January to June 2014 has been done by comparing the daily concentrations of Paris and Saclay data set for year 2014 provided by the RNSA observations in Paris. The results obtained without the data set of the January to June 2014 were also showing a decrease of the atmospheric concentration ( Figure A1, Appendix B) and pointed out the importance to follow the AFS all over the year as it is done at the Saclay observatory.
Remote Sens. 2017, 9, x FOR PEER REVIEW 6 of 32 Zefir source receptor modeling tool to identify the geographical origin of airborne fungal spores impacting the observation site.

Interannuality of the Airborne Annual Fungal Spore Integral at Saclay
The interannuality of AFS Integral (AFSIn) was studied in depth. The annual sum of fungal spores ranges from 2.5 10 6 Nb#/m 3 to 5.2 10 6 Nb#/m 3 . We observed a strong interannuality of the AFSIn and no biannual cycle has been found. Instead, a general decrease of the atmospheric concentration was found, as illustrated in Figure 3. The AFSIn data from January to June 2014 have been estimated by a linear modeling with the most basic model which is the simple linear regression where a variable X is explained and modeled by an affine function of a comparable variable Y. The interpolation of the Saclay data set from January to June 2014 has been done by comparing the daily concentrations of Paris and Saclay data set for year 2014 provided by the RNSA observations in Paris. The results obtained without the data set of the January to June 2014 were also showing a decrease of the atmospheric concentration ( Figure A1, Appendix B) and pointed out the importance to follow the AFS all over the year as it is done at the Saclay observatory. Such results are in accordance with those observed for total species records in Europe like reported by [52] with an AFSIn in 1989 of 2.10 6 Nb#/m 3 . Similarly, in Greece, a general decrease in fungal spore levels in Thessaloniki was observed from 1987 to 2005 with a mean decrease of -52% compared to Saclay (-52%) [53]. The annual sum of fungal spores observed at Saclay is in the range of what has been observed in other oceanic environments in Europe, and is generally 3 to 4 times higher than in Mediterranean areas [54]. All the results obtained in these different ecosystems show a general decrease indicating that our observations are not linked to fluctuating changes.

Seasonality of All Taxa Combined Concentrations at Saclay
The monthly mean distribution of all AFS concentrations averaged over the five years of measurements displays a clear bimodal seasonal cycle as illustrated in Figure 4. P90, P75, P25, and P10 represent the 90 th , 75 th , 25 th , and 10 th percentiles respectively. The monthly mean found Such results are in accordance with those observed for total species records in Europe like reported by [52] with an AFSIn in 1989 of 2.10 6 Nb#/m 3 . Similarly, in Greece, a general decrease in fungal spore levels in Thessaloniki was observed from 1987 to 2005 with a mean decrease of −52% compared to Saclay (−52%) [53]. The annual sum of fungal spores observed at Saclay is in the range of what has been observed in other oceanic environments in Europe, and is generally 3 to 4 times higher than in Mediterranean areas [54]. All the results obtained in these different ecosystems show a general decrease indicating that our observations are not linked to fluctuating changes.

Seasonality of All Taxa Combined Concentrations at Saclay
The monthly mean distribution of all AFS concentrations averaged over the five years of measurements displays a clear bimodal seasonal cycle as illustrated in Figure 4. P90, P75, P25, and P10 represent the 90th, 75th, 25th, and 10th percentiles respectively. The monthly mean found for the MSS in the current study is found in May (11571 Nb#/m 3 ) and ends in November (8472 Nb#/m 3 ). The first maximum was observed in July (28,564 Nb#/m 3 ) followed by a rapid decrease in September (12,456 Nb#/m 3 ). A secondary maximum was noted in October (14,538.10 3 Nb#/m 3 ), followed by a sharp decrease in December (3439 Nb#/m 3 ). Data are available in Table A1, Appendix A. In total, the fungal spore season is significant over 7 months in a year (from May to November), representing 58% of the year and 91% of the total AFS concentrations.  Table A1, Appendix A. In total, the fungal spore season is significant over 7 months in a year (from May to November), representing 58% of the year and 91% of the total AFS concentrations.

Major Parameters Driving the Seasonality of AFS Concentrations at Saclay
The observations made on the monthly means of temperature (°C), the monthly means of precipitation (mm) and the monthly mean concentrations of fungal spores, as illustrated in Figure 5, showed that the MSS starts when the mean temperature reaches an average value above 11 °C and not below (values are reported in Table A1 and A2). This result is important to understand how the production of fungal spores occurs: (1) the need for a certain amount of water; and (2) a mean temperature above 11 °C to start the growing process and the subsequent dispersal. Those observations are in full accordance with literature [6].

Major Parameters Driving the Seasonality of AFS Concentrations at Saclay
The observations made on the monthly means of temperature ( • C), the monthly means of precipitation (mm) and the monthly mean concentrations of fungal spores, as illustrated in Figure 5, showed that the MSS starts when the mean temperature reaches an average value above 11 • C and not below (values are reported in Tables A1 and A2). This result is important to understand how the production of fungal spores occurs: (1) the need for a certain amount of water; and (2) a mean temperature above 11 • C to start the growing process and the subsequent dispersal. Those observations are in full accordance with literature [6].

Main Spore Season and Daily Variability of Airborne Fungal Spore Concentrations at Saclay
The fungal propagule pattern at Saclay was found to be different from one year to the next likely due to the meteorological factors of variability (temperature, insolation, precipitation, atmospheric pollutants) as previously reported [55,56]. For each of the five years reported in the current study, the starts and the ends of the MSS are shown in Table A3. The calculation applied was the same as what has been previously employed [57].
As shown in Figure 6, the fungal spore concentrations can be extremely variable both in time and intensity. As an example, a high concentration's episode occurred in June 2015 as the weather conditions were remarkably hot and dry for the season. Hence an increase in the AFS concentrations from 10 x 10 3 Nb#/m 3 on 22 June rising to 256 x 10 3 Nb#/m 3 on 23 June and decreasing back to 10 x 10 3 Nb#/m 3 on 24 June was noted. This exceptional episode has been characterized using several on-line techniques to understand the processes involved in creating the bloom of bioaerosols that can occur in water-stressed environments [58]. It turned out that during 2015 the temperature in the region of Saclay was greater by 1.5% from annual seasonal normal temperatures with the month of June being particularly hotter and dryer (+5%) than "normal" records. On 23 June a precipitation event occurred, and, at the end of the rain episode, the concentration of total fungal spores got multiplied by a factor 25. This observation is consistent with other studies showing that air temperature is the main physical parameter driving AFS concentrations but is not effective without the presence of water and most typically, a rain shower which increase the bioaerosol concentrations after several hours [59,60].

Main Spore Season and Daily Variability of Airborne Fungal Spore Concentrations at Saclay
The fungal propagule pattern at Saclay was found to be different from one year to the next likely due to the meteorological factors of variability (temperature, insolation, precipitation, atmospheric pollutants) as previously reported [55,56]. For each of the five years reported in the current study, the starts and the ends of the MSS are shown in Table A3. The calculation applied was the same as what has been previously employed [57].
As shown in Figure 6, the fungal spore concentrations can be extremely variable both in time and intensity. As an example, a high concentration's episode occurred in June 2015 as the weather conditions were remarkably hot and dry for the season. Hence an increase in the AFS concentrations from 10 × 10 3 Nb#/m 3 on 22 June rising to 256 × 10 3 Nb#/m 3 on 23 June and decreasing back to 10 × 10 3 Nb#/m 3 on 24 June was noted. This exceptional episode has been characterized using several on-line techniques to understand the processes involved in creating the bloom of bioaerosols that can occur in water-stressed environments [58]. It turned out that during 2015 the temperature in the region of Saclay was greater by 1.5% from annual seasonal normal temperatures with the month of June being particularly hotter and dryer (+5%) than "normal" records. On 23 June a precipitation event occurred, and, at the end of the rain episode, the concentration of total fungal spores got multiplied by a factor 25. This observation is consistent with other studies showing that air temperature is the main physical parameter driving AFS concentrations but is not effective without the presence of water and most typically, a rain shower which increase the bioaerosol concentrations after several hours [59,60]. Remote Sens. 2017, 9, x FOR PEER REVIEW 9 of 32

Airborne Fungal Spore Characteristics at Saclay
To date, the classification regarding the phylum is still under progress due to novel technics of identifications [61]. In our study, 30 types of propagules were identified and classified under the Universal Biological Indexer and Organizer (uBio) [62]. The results are reported in Table A4. During the period of observation, i.e., from July 2014 to December 2018, the relative proportion of the Ascomycota (AMC) and Basidiomycota (BMC) at a phylum level were calculated without DNA or RNA analysis in our atmospheric samples. The results were compared with the methodology proposed by [15] regarding the fungal diversity over land and oceans to determine a biogeography in the air. The ratio BMC/AMC in our ecosystem is dominated by AMC ( Figure  7). This is consistent with the predominance of BMC and AMC in the biosphere, where they account for 98% of the known species and also with the Oceanic Degraded climate characteristics of the Saclay site, in reference to marine sites as reported by [63].

Airborne Fungal Spore Diversity at Saclay
The mean concentrations and percentages of the 30 atmospheric fungal propagules found at Saclay are reported in Table A5. To better understand the variability of AFS at Saclay and the relative proportion of allergenic fungal spores, we analyzed the major classes of fungal spore

Airborne Fungal Spore Characteristics at Saclay
To date, the classification regarding the phylum is still under progress due to novel technics of identifications [61]. In our study, 30 types of propagules were identified and classified under the Universal Biological Indexer and Organizer (uBio) [62]. The results are reported in Table A4. During the period of observation, i.e., from July 2014 to December 2018, the relative proportion of the Ascomycota (AMC) and Basidiomycota (BMC) at a phylum level were calculated without DNA or RNA analysis in our atmospheric samples. The results were compared with the methodology proposed by [15] regarding the fungal diversity over land and oceans to determine a biogeography in the air. The ratio BMC/AMC in our ecosystem is dominated by AMC ( Figure 7). This is consistent with the predominance of BMC and AMC in the biosphere, where they account for 98% of the known species and also with the Oceanic Degraded climate characteristics of the Saclay site, in reference to marine sites as reported by [63].

Airborne Fungal Spore Characteristics at Saclay
To date, the classification regarding the phylum is still under progress due to novel technics of identifications [61]. In our study, 30 types of propagules were identified and classified under the Universal Biological Indexer and Organizer (uBio) [62]. The results are reported in Table A4. During the period of observation, i.e., from July 2014 to December 2018, the relative proportion of the Ascomycota (AMC) and Basidiomycota (BMC) at a phylum level were calculated without DNA or RNA analysis in our atmospheric samples. The results were compared with the methodology proposed by [15] regarding the fungal diversity over land and oceans to determine a biogeography in the air. The ratio BMC/AMC in our ecosystem is dominated by AMC ( Figure  7). This is consistent with the predominance of BMC and AMC in the biosphere, where they account for 98% of the known species and also with the Oceanic Degraded climate characteristics of the Saclay site, in reference to marine sites as reported by [63].

Airborne Fungal Spore Diversity at Saclay
The mean concentrations and percentages of the 30 atmospheric fungal propagules found at Saclay are reported in Table A5. To better understand the variability of AFS at Saclay and the relative proportion of allergenic fungal spores, we analyzed the major classes of fungal spore

Airborne Fungal Spore Diversity at Saclay
The mean concentrations and percentages of the 30 atmospheric fungal propagules found at Saclay are reported in Table A5. To better understand the variability of AFS at Saclay and the relative proportion of allergenic fungal spores, we analyzed the major classes of fungal spore propagules present in the air. Figure 8 displays the obtained pattern which takes into account 95% of the total AFS. AMC contributes to 85% of the total AFS, driven by Ascopores (51%), Cladosporium (33%), and Alternaria (1%), while BMC only represents 10% of AFSIn through the contribution of the Basidiospores (8%) and Tilletiopsis (2%) taxa. Their seasonality was investigated ( Figure A3, Appendix B) but the time variable does not appear to determine the predominance of AMC in the Saclay environment, which therefore calls for screening the air mass origins.
Remote Sens. 2017, 9, x FOR PEER REVIEW 10 of 32 propagules present in the air. Figure 8 displays the obtained pattern which takes into account 95% of the total AFS. AMC contributes to 85% of the total AFS, driven by Ascopores (51%), Cladosporium (33%), and Alternaria (1%), while BMC only represents 10% of AFSIn through the contribution of the Basidiospores (8%) and Tilletiopsis (2%) taxa. Their seasonality was investigated ( Figure A3, Appendix B) but the time variable does not appear to determine the predominance of AMC in the Saclay environment, which therefore calls for screening the air mass origins. As illustrated in Figure 8, the known allergenic taxa monitored at Saclay, which are Cladosporium and Alternaria, represent 34% of the AFSIn and in the AMC phylum. Clearly, there is a major reservoir for allergenic species but, at this stage of development, only genotyping analysis could reveal what are the species to be found in the Ascospores and Basidiospores reservoir.

Discussion
The interannual, seasonal and daily variability of total and specific AFSIn concentrations at Saclay are discussed and compared to the literature. The geographical origins of the main AFS impacting the observation site were thus assessed and discussed through the use of the Zefir source-receptor model. For the purpose of investigating the origins and point sources of AFS impacting Saclay will be also discussed

Factors Controling the Airborne Fungal Spore Concentrations
In the ecosystem of Saclay, water is not a limiting factor, which makes air temperature the only driver of the AFS seasonality, as illustrated by Figures B4 and B5. This result is in accordance with recent findings from [55]. In an attempt to model the seasonal cycle of AFS at Saclay, we explored different calculation procedures based on our understanding of the impact of rainfall on AFS concentrations. We found that by normalizing the monthly mean temperature by the monthly mean precipitation, we could reasonably reproduce the seasonal cycle of AFS at Saclay ( Figure 9) and values reported in Table A2. As illustrated in Figure 8, the known allergenic taxa monitored at Saclay, which are Cladosporium and Alternaria, represent 34% of the AFSIn and in the AMC phylum. Clearly, there is a major reservoir for allergenic species but, at this stage of development, only genotyping analysis could reveal what are the species to be found in the Ascospores and Basidiospores reservoir.

Discussion
The interannual, seasonal and daily variability of total and specific AFSIn concentrations at Saclay are discussed and compared to the literature. The geographical origins of the main AFS impacting the observation site were thus assessed and discussed through the use of the Zefir source-receptor model. For the purpose of investigating the origins and point sources of AFS impacting Saclay will be also discussed

Factors Controling the Airborne Fungal Spore Concentrations
In the ecosystem of Saclay, water is not a limiting factor, which makes air temperature the only driver of the AFS seasonality, as illustrated by Figures A4 and A5. This result is in accordance with recent findings from [55]. In an attempt to model the seasonal cycle of AFS at Saclay, we explored different calculation procedures based on our understanding of the impact of rainfall on AFS concentrations. We found that by normalizing the monthly mean temperature by the monthly mean precipitation, we could reasonably reproduce the seasonal cycle of AFS at Saclay ( Figure 9) and values reported in Table A2.
In the literature, rain has often been cited as the most efficient agent for the wet removal of bioaerosols [64]. However, precipitation alone would possibly be a poor driving parameter on regression models [35] for two main reasons 1) the annual variation of daily precipitation values does not closely correlate with the concentrations measured during dry or wet conditions, (2) precipitation on a daily basis presents high variability in time and space (rain patches). Therefore, instead of considering precipitation data, some works suggested to follow rain showers as a favorable trigger of PBAP generation [59]. The second BIOaerosol DETECTion campaign in 2015 (BIODETECT 2015) in Saclay provides an interesting illustration of the triggering property of rain showers on PBAPs. As described in Section 3.6, June in 2015 was particularly hot and dry. The AFS concentrations were low for the season compared to 2016 and 2017. A moderate rainfall event occurred on 23 June 2015 providing 4.8 mm of water in four hours (e.g., 1 mm measured is equivalent to 1 L/m 2 ). Despite this relatively low amount of water, right after the shower, the concentrations in AFS have been multiplied by 25 from 12,000 Nb#/m 3 to 226,056 Nb#/m 3 , then went back to 14,000 Nb#/m 3 the day after as reported in (Table A6) and illustrated by Figure 10. This event can be typically defined as a "splash dispersal" [65]. These spores are known to be hydrofugal and can be easily dispersed by the impact of raindrops [66]. But more interesting is the proportion of the different species. As expected, Ascospores were dominant; Cladosporium and Alternaria were not affected but Ustilago, Dydimella, and Helicomyces concentrations increased from 0 to 1131 Nb#/m 3 , 103 to 232 Nb#/m 3 , and 0 to 283 Nb#/m 3 , respectively. Ustilago is known to be a cereal and herb pathogen (Table A2) and the strong increase observed coincided with the end of the Poaceae pollen season, as reported by [39]. In this specific period, Poaceae was in early decline due to the very hot climate conditions. The rain shower suddenly favored the release of these fungi which was living and growing in this type of habitat. We also observed that the concentrations of Cladosporium increased slowly to reach a maximum 2 days after. This is consistent with the study of [67] on the growing processes. In the literature, rain has often been cited as the most efficient agent for the wet removal of bioaerosols [64]. However, precipitation alone would possibly be a poor driving parameter on regression models [35] for two main reasons 1) the annual variation of daily precipitation values does not closely correlate with the concentrations measured during dry or wet conditions, (2) precipitation on a daily basis presents high variability in time and space (rain patches). Therefore, instead of considering precipitation data, some works suggested to follow rain showers as a favorable trigger of PBAP generation [59]. The second BIOaerosol DETECTion campaign in 2015 (BIODETECT 2015) in Saclay provides an interesting illustration of the triggering property of rain showers on PBAPs. As described in Section 3.6, June in 2015 was particularly hot and dry. The AFS concentrations were low for the season compared to 2016 and 2017. A moderate rainfall event occurred on 23 June 2015 providing 4.8 mm of water in four hours (e.g., 1 mm measured is equivalent to 1 L/m 2 ). Despite this relatively low amount of water, right after the shower, the concentrations in AFS have been multiplied by 25 from 12,000 Nb#/m 3 to 226,056 Nb#/m 3 , then went back to 14,000 Nb#/m 3 the day after as reported in (Table A6) and illustrated by Figure 10. This event can be typically defined as a "splash dispersal" [65]. These spores are known to be hydrofugal and can be easily dispersed by the impact of raindrops [66]. But more interesting is the proportion of the different species. As expected, Ascospores were dominant; Cladosporium and Alternaria were not affected but Ustilago, Dydimella, and Helicomyces concentrations increased from 0 to 1131 Nb#/m 3 , 103 to 232 Nb#/m 3 , and 0 to 283 Nb#/m 3 , respectively. Ustilago is known to be a cereal and herb pathogen (Table A2) and the strong increase observed coincided with the end of the Poaceae pollen season, as reported by [39]. In this specific period, Poaceae was in early decline due to the very hot climate conditions. The rain shower suddenly favored the release of these fungi which was living and growing in this type of habitat. We also observed that the concentrations of Cladosporium increased slowly to reach a maximum 2 days after. This is consistent with the study of [67] on the growing processes. Pathogenic AFS are a key concern which requires a better understanding of their control factors as they need a host (plants or animals) for their own development. For lower case, Entomophthora species, a lethal fungal parasite affecting a type of fly, the Musca domestica [41] was observed for the first time in May 2016 at Saclay, in relatively high concentrations. The presence of this fungal parasite has been explained by the entomologist Sandrine Provots (personal communication) by the late produce egg hatching and larval formation. Indeed, the Pathogenic AFS are a key concern which requires a better understanding of their control factors as they need a host (plants or animals) for their own development. For lower case, Entomophthora species, a lethal fungal parasite affecting a type of fly, the Musca domestica [41] was observed for the first time in May 2016 at Saclay, in relatively high concentrations. The presence of this fungal parasite has been explained by the entomologist Sandrine Provots (personal communication) by the late produce egg hatching and larval formation. Indeed, the late start of the spring due to cold weather [65] created a delay in the fly reproductions at a regional scale. Consequently, an explosion of flies occurs and, a remarkable proliferation of this pathogen in the early days of May as reported by the French National Survey of Crops contamination [66]. This episode has generated a drastic increase of the atmospheric concentrations of Entomophthora's fungal spores and they were transported and detected at Saclay far from the point source. It pointed out one more time the indirect role of the local and regional meteorology in the exhalation or inhibition of AFS and particularly the dispersion of specific AFS pathogens with unknown pathogenicity consequences on human and animals. The variability of fungal spore concentrations is therefore strongly dependent on the related ecosystem and on its local and regional meteorological characteristics. In particular, some species undergo strong differences in growth following wet, dry, or "splash" discharges. These processes of discharge can subsequently affect the concentrations of the chemical tracers in the PM10 fraction. Moreover, to better understand the effects of rainfall events on fungal spore diversity, abundance and dispersal processes, models need to integrate rain and temperature as well as dew point data at the finest resolution possible.

Interannuality and Global Increase of Specific AFSIn Concentrations
The total concentrations of AFS are in general analyzed in the literature in a context of global warming and for allergenic purposes [12,53,67]. It is admitted today that fungal spores are very sensitive to temperature and water (e.g., relative humidity and rain for their proper development [15]. It is the case on Earth but also in confined and much stressed environments like the International Space Station [68]. On a year-to-year basis, the results of our 5-year study display strong variations in the interannuality of the five main fungal taxa spores identified ( Figure 11). We also analyze Ustilago as it is a strong plant pathogen on culture [36] and present on the strong allergenic Poaceae pollen We investigated the influence of the meteorological conditions that then occurred in France. During our period of observation indeed, 2015 and 2018 where hot and dry. In more details, 2015 was characterized by a deviation of +1.5 °C compared to the seasonal normal and a loss close to 20% in the total precipitations all over the country. The year 2018 was the hottest since 1959 in France with a measured anomaly of +1.4 °C (+2 °C in the continental part alone) and was relatively dry all over France (global loss in precipitation of 10%). In the north half of the country, including Saclay, a loss of more than 20% in the total rain water was observed [69]. In contrast, 2014 was the second hottest year after 2018 but benefited in the northern half of France from an additional 20% of precipitations. The years 2016 and 2017 stood in the normal seasonal range (temperature and rain), but 2017 was hot and humid in the northern part of France and quite We investigated the influence of the meteorological conditions that then occurred in France. During our period of observation indeed, 2015 and 2018 where hot and dry. In more details, 2015 was characterized by a deviation of +1.5 • C compared to the seasonal normal and a loss close to 20% in the total precipitations all over the country. The year 2018 was the hottest since 1959 in France with a measured anomaly of +1.4 • C (+2 • C in the continental part alone) and was relatively dry all over France (global loss in precipitation of 10%). In the north half of the country, including Saclay, a loss of more than 20% in the total rain water was observed [69]. In contrast, 2014 was the second hottest year after 2018 but benefited in the northern half of France from an additional 20% of precipitations. The years 2016 and 2017 stood in the normal seasonal range (temperature and rain), but 2017 was hot and humid in the northern part of France and quite close to what it has been observed for 2014.
The general decrease of the concentration observed in the AFSIn was mainly due to the predominance of Ascospores in the Saclay ecosystem. Cladosporium shows moderate biannual cycle compared to the other taxa studied. Ustilago and Basidiospores presented similar patterns: (1) a strong biannual cycle and (2) strong sensitivity to hot temperature while Tilletiopsis concentrations are strongly affected by hot weather. Interestingly, given they are all BMC and that continental climate is expected to be hotter, our observations suggest one hypothesis of the enrichment in the ratio BMC/AMC in continental ecosystem. The effect of the D a of the fungal propagules as proposed by [15] is perhaps not the only factor controlling the ratio BMC/AMC and could change in the future. Moreover, the growing kinetics is strongly dependent on temperature as reported by [70]. To understand the differences between the annual variability of the different taxa, we investigated the impact in the cycle of the land use in "Big Cultures" which include wheat, maize, ragweed, onions, and tomatoes [71]. We did not find any correlation in the annual cycles, but evidenced a probable link with the general increasing trend in the total agricultural production for humans. This last point supports idea that the presence of agricultural fields is not a limiting factor for the emission of AFS in the air, unlike rain and temperature. These differences from-to-year can explain the differences observed in the individual ratio of BMC and AMC as illustrated by the Figure A3.
Preliminary conclusions point out, as expected, that all taxa do not have the same response to an elevation of temperature and rainfalls-in particular, Ascospores and Tilletiopsis seem to be very sensitive to temperature and hydrological stress. It is noticeable that, since 1990, temperature in France has always stood between +0.5 and +1.4 • C, up +2 • C in some specific regions, over the normal annual mean computed from 1981 to 2010, as reported in [69]. Consequently, our 5-year study displays a strong interannuality of the monitored taxa. Cladosporium, Ustilago, and Basidiospores are characterized by a bi-annual cycle in accordance with that previously reported [53,67]. The general decrease observed in the total AFS is mainly due to the predominance of Ascospores followed by Basidiospores.

Seasonality of BMC and AMC
Continuous measurements of AFS are important to increase the knowledge as the seasonality can be different in time and in variability as illustrated by Figure A6. In the study, as Saclay is not referenced as a marine or coastal site, it was expected a higher abundance for BMC. One explanation for this observation can be found in the size properties of AFS. BMC are expected to have shorter atmospheric residence times, therefore are unsuitable to undergo long-range transport. The BMC/AMC ratio is thus expected to decrease with increasing distance from the sea. As Saclay is 250 km from the ocean and impacted by marine air masses it would mean that the distance effect would need to be greater than 250 km in order to see a difference in the BMC/AMC ratio as reported by [15]. During our five-year observation period, we noticed that the ratio of BMC/AMC was changing during the season ( Figure 12). This result does not affect the global ratio of BMC/AMC (e.g., Figure 7) but points out (1) the importance to consider BMC/AMC ratios by season for one given ecosystem and (2) to have a better understanding of the BMC/AMC ratio as a function of environmental considerations as proposed by [15]. In 2017, [72] showed similar and finer results regarding the seasonality of BMC/AMC using DNA analysis in several northern Europe locations and the observations suggested that the biodiversity of fungal spores can be observed after 900 km from an ecosystem to another one with consequences on the variation in the BMC/AMC ratios. BMC/AMC ratios by season for one given ecosystem and (2) to have a better understanding of the BMC/AMC ratio as a function of environmental considerations as proposed by [15]. In 2017, [72] showed similar and finer results regarding the seasonality of BMC/AMC using DNA analysis in several northern Europe locations and the observations suggested that the biodiversity of fungal spores can be observed after 900 km from an ecosystem to another one with consequences on the variation in the BMC/AMC ratios. In the literature it has been found that emissions of fungal spores from the oceans (10 Mg per year) are several orders of magnitude smaller than from inland surfaces  Tg per year) [18,19]. Emissions are characterized by specific BMC/AMC ratios: high ratios of BMC/AMC in marine or coastal sites and low ratios of BMC/AMC in continental sites [15]. The phyla richness of BMC has been explained regarding the size of the AFS. BMC are considered to be 5 to 10 μm (Da) in size and AMC to be between 2 and 5 μm (Da). BMC is enhanced in the coarse fraction (>2.5 μm), whereas the taxa richness of AMC is enhanced in the fine fraction (<2.5 μm) of continental air particulate matter due to their shorter atmospheric residence times. Our observation suggests that additional processes are involved in relation with temperature, rainfall and the ecosystem characteristics, in particular land use. As continental areas present higher temperature levels during spring and summer than marine areas, Ascospores or Basidiospores proliferation can be inhibited. In the same time, cereal and herbs are largely present in continental regions which possibly increase the presence of BMC spores like Ustilago and other Basidiospores. As described by [54] in the region of Madrid (hot and dry summer), during the MSS, AMC were not predominating due to hot and dry weather. Another preliminary In the literature it has been found that emissions of fungal spores from the oceans (10 Mg per year) are several orders of magnitude smaller than from inland surfaces  Tg per year) [18,19]. Emissions are characterized by specific BMC/AMC ratios: high ratios of BMC/AMC in marine or coastal sites and low ratios of BMC/AMC in continental sites [15]. The phyla richness of BMC has been explained regarding the size of the AFS. BMC are considered to be 5 to 10 µm (D a ) in size and AMC to be between 2 and 5 µm (D a ). BMC is enhanced in the coarse fraction (>2.5 µm), whereas the taxa richness of AMC is enhanced in the fine fraction (<2.5 µm) of continental air particulate matter due to their shorter atmospheric residence times. Our observation suggests that additional processes are involved in relation with temperature, rainfall and the ecosystem characteristics, in particular land use. As continental areas present higher temperature levels during spring and summer than marine areas, Ascospores or Basidiospores proliferation can be inhibited. In the same time, cereal and herbs are largely present in continental regions which possibly increase the presence of BMC spores like Ustilago and other Basidiospores. As described by [54] in the region of Madrid (hot and dry summer), during the MSS, AMC were not predominating due to hot and dry weather. Another preliminary conclusion suggests that the size of the spores is not the only factor that can explain the variation in the ratio of BMC/AMC. Other drivers could be (1) the climate of the ecosystem studied by determining the proportion of the main phyla and also the proportion in species in the same phyla and (2) the land use related to the culture of cereals or the presence of herbs-France among the European countries has the largest crops mainly located in the North and in its continental part (e.g., Figure 1) followed by Germany [73]. This aspect should be taken into account in the future in a context of global warming regarding the potential formation of GCCN, CCN, and IN [9,74].

Geographical Origins of AFS Impacting the Observation Site
The calculations using ZeFir from this 5-year dataset showed an interesting result regarding the general origin of AFS concentrations. The model designates a main origin from the northwest sector, that is, independent from the SW prevailing winds (Figure 2), however still in the general "wet" W sector, which mostly carries marine air masses. Two point-sources were identified, the major one is associated with wind speeds ranging from 10 km/h to 13 km/h and a minor one with wind speeds of 17 km/h ( Figure 13). We also evaluate the geographical origin of total AFS by running the model year to year as illustrated by Figure A7 to investigate the effect of the meteorology on the Saclay ecosystem.
sector, that is, independent from the SW prevailing winds (Figure 2), however still in the general "wet" W sector, which mostly carries marine air masses. Two point-sources were identified, the major one is associated with wind speeds ranging from 10 km/h to 13 km/h and a minor one with wind speeds of 17 km/h ( Figure 13). We also evaluate the geographical origin of total AFS by running the model year to year as illustrated by Figure A7 to investigate the effect of the meteorology on the Saclay ecosystem. Figure 13. Origin of total atmospheric AFS using sustained wind incidence method (SWIM) model Origin. The dotted white circles represent the wind speed scale in kilometer per hour (km/h). The color grid represents the estimated concentration (Nb#/m 3 ) for any wind speed and wind direction. Figure 14 reveals that Saclay is impacted by both local and regional sources. Among the main fungal spores (95% of the species present in the air Section 2.2) Ascospores and Basiodospores which are both wet spores are clearly originating from different areas: North West for Ascospores (marine/costal area) and North East for Basiodospores (continental area). Cladosporium and Tilletiopsis were clearly originating from the NW sector. For Alternaria, the results pointed out a regional origin, possibly related to large agricultural parcels of rapeseed and sun flower ( Figure A8) as Alternaria is known to be a plant pathogen of those crops [73,75].  Figure 14 reveals that Saclay is impacted by both local and regional sources. Among the main fungal spores (95% of the species present in the air Section 2.2) Ascospores and Basiodospores which are both wet spores are clearly originating from different areas: North West for Ascospores (marine/costal area) and North East for Basiodospores (continental area). Cladosporium and Tilletiopsis were clearly originating from the NW sector. For Alternaria, the results pointed out a regional origin, possibly related to large agricultural parcels of rapeseed and sun flower ( Figure A8) as Alternaria is known to be a plant pathogen of those crops [73,75].

Case of Studies: the Geographical Origins of Plant or Human Pathogens
To go further in our understanding on the geographical origin of specific AFS we ran the model for eight species representing 65 % on the remaining 5% of the AFS (Figure A9), thus less abundant but determining as most of them being referenced as human or plant pathogens. The result obtained for Aspergillaceae, Ganoderma, Myxomycetes and Ustilago depicts a clear North to North East origin as illustrated by Figure 15. Aspergillaceae and Myxomycetes are two taxa known to grow in open forest on decaying material (Table A4), while the habitat of Ustilago is known to be Poaceae and cereals, as already mentioned. This happens to be in close agreement with our results on the geographical origin of Betulaceae and Poaceae pollen grains impacting Saclay [39].

Case of Studies: The Geographical Origins of Plant or Human Pathogens
To go further in our understanding on the geographical origin of specific AFS we ran the model for eight species representing 65 % on the remaining 5% of the AFS (Figure A9), thus less abundant but determining as most of them being referenced as human or plant pathogens. The result obtained for Aspergillaceae, Ganoderma, Myxomycetes and Ustilago depicts a clear North to North East origin as illustrated by Figure 15. Aspergillaceae and Myxomycetes are two taxa known to grow in open forest on decaying material (Table A4), while the habitat of Ustilago is known to be Poaceae and cereals, as already mentioned. This happens to be in close agreement with our results on the geographical origin of Betulaceae and Poaceae pollen grains impacting Saclay [39].
To go further in our understanding on the geographical origin of specific AFS we ran the model for eight species representing 65 % on the remaining 5% of the AFS (Figure A9), thus less abundant but determining as most of them being referenced as human or plant pathogens. The result obtained for Aspergillaceae, Ganoderma, Myxomycetes and Ustilago depicts a clear North to North East origin as illustrated by Figure 15. Aspergillaceae and Myxomycetes are two taxa known to grow in open forest on decaying material (Table A4), while the habitat of Ustilago is known to be Poaceae and cereals, as already mentioned. This happens to be in close agreement with our results on the geographical origin of Betulaceae and Poaceae pollen grains impacting Saclay [39]. We also run the model for Didymella, Helicomyces, Botryris, and Entomophthora, which ended to originate from the NW and the SSW sectors ( Figure 16). As reported in Table A3, Didymella and Helicomyces can grow on barley and corn leaves. By comparing the geographical origin computed by the model and the map of barley and corn fields in France ( Figure A9) we observed We also run the model for Didymella, Helicomyces, Botryris, and Entomophthora, which ended to originate from the NW and the SSW sectors ( Figure 16). As reported in Table A3, Didymella and Helicomyces can grow on barley and corn leaves. By comparing the geographical origin computed by the model and the map of barley and corn fields in France ( Figure A9) we observed that the model accounted for the sources of these two fungi. In particular, Helicomyces originated from SSW brought by strong winds ranging from 15 to 20 kmh. This result is of importance for the optimization of the Zefir model to localize more precisely specific point sources by comparing the distance of the source and the spots founded by the model as a function of wind speed in km/h. Botrytris is polyphagous and the main point source was identified on the NW sector. We attempted to identify the geographical origin of Entomophthora and we found a regional origin in the NW sector. We also observed the limits of the model when the data set used to find the origin and point sources is too limited (only two days in 2016).
The use of pesticides on large agricultural parcels can inhibit the production of AFS by the crops but have no significant disturbance on production from forests or prairies, allowing the natural growth of specific pathogens. The hypothesis regarding the use of chemicals as a control factor of the proliferation of fungi on crops needs to be investigated in more details in future works, since our work suggests a same point source for Ustilago and Poaceae pollen. This last point must be taken in account for cross contamination in land in rotating land use [76].
the optimization of the Zefir model to localize more precisely specific point sources by comparing the distance of the source and the spots founded by the model as a function of wind speed in km/h. Botrytris is polyphagous and the main point source was identified on the NW sector. We attempted to identify the geographical origin of Entomophthora and we found a regional origin in the NW sector. We also observed the limits of the model when the data set used to find the origin and point sources is too limited (only two days in 2016). The use of pesticides on large agricultural parcels can inhibit the production of AFS by the crops but have no significant disturbance on production from forests or prairies, allowing the natural growth of specific pathogens. The hypothesis regarding the use of chemicals as a control factor of the proliferation of fungi on crops needs to be investigated in more details in future works, since our work suggests a same point source for Ustilago and Poaceae pollen. This last point must be taken in account for cross contamination in land in rotating land use [76].

Conclusions
The objectives of this work were, firstly, to understand the atmospheric variability of the main airborne fungal spores (AFS) present at Saclay, a suburban area of Paris and, secondly, to apply a source receptor model to attempt identifying their geographical origin. Several case studies served to investigate the link between land uses and production of AFS. In the Saclay oceanic degraded ecosystem, 30 fungal airborne spore taxa were identified following palynological methods. In relation with the classification used in this work, Ascomycota (AMC) were predominant (89%) followed by Basidiomycota (BMC, 10%). Ascospores were the main taxa (51%), then Cladosporium (33%), Basidiospores (8%), Tilletiopsis (2%), and Alternaria (1%). We observed a strong interannuality and a general decrease in the total AFSi concentrations over our 5-year study, mainly driven by the "wet spores" Ascospores and Basidiospores, while the "dry spores" Cladosporium and Alternaria had increasing trends. The year-to-year variability resulted from hot/cold temperature combined to dry/wet conditions. The AFS concentrations

Conclusions
The objectives of this work were, firstly, to understand the atmospheric variability of the main airborne fungal spores (AFS) present at Saclay, a suburban area of Paris and, secondly, to apply a source receptor model to attempt identifying their geographical origin. Several case studies served to investigate the link between land uses and production of AFS. In the Saclay oceanic degraded ecosystem, 30 fungal airborne spore taxa were identified following palynological methods. In relation with the classification used in this work, Ascomycota (AMC) were predominant (89%) followed by Basidiomycota (BMC, 10%). Ascospores were the main taxa (51%), then Cladosporium (33%), Basidiospores (8%), Tilletiopsis (2%), and Alternaria (1%). We observed a strong interannuality and a general decrease in the total AFSi concentrations over our 5-year study, mainly driven by the "wet spores" Ascospores and Basidiospores, while the "dry spores" Cladosporium and Alternaria had increasing trends. The year-to-year variability resulted from hot/cold temperature combined to dry/wet conditions. The AFS concentrations showed a clear bimodal seasonal cycle and the main spores season (MSS) is strongly affected by meteorological factors like temperature and rainfall. A mean of 11 • C has been found to allow the start of the season and the first maximum occurs in July (32,000 Nb#/m 3 ) while the second occurs in October (17,000 Nb#/m 3 ) as illustrated in Figure 4. The BMC/AMC also showed a seasonality suggesting that the size of the spores is not the only factor that can explain the variation in the BMC/AMC ratio. Preliminary results pointed out that fungal propagules do not have the same response to an elevation of temperature or can be very sensitive to temperature and hydrological stress. We also pointed out that they do not have the same geographical origin and this last point could explain the biogeography of fungal spores in the air. The source receptor model Zefir, has been successfully applied to identify the geographical origin of the total AFS impacting the observation site. Most of the fungi originated from the NW sector, hence not transported by the prevailing SW winds. The model was also applied to detect the origin of the major classes and several species of AFS. The results obtained showed that this user-friendly tool kit is accurate enough to locate individual source points and relate them to land cover and land use. The agricultural practice should be taken into account as some species undergo strong differences in growth following wet, dry, or "splash" discharges. These processes of discharge can subsequently affect the concentrations of chemical tracers in the PM10 fraction. As a recommendation, our results hence suggest that all sites dedicated to air quality monitoring providing measurement of chemical tracers in the PM10 fraction should be calibrated by measurements of AFS with the spore trap technic. Consequently, to better understand the effects of rainfall events on fungal spore diversity, abundance and dispersal processes, models need to integrate rain and temperature as well as dew point data at the finest resolution possible. To-date, the consequences of global warming on fungal growth and spore production, like on allergenic pollen production, is not documented enough. Recent studies showed similar implications for respiratory tract diseases in humans. Our results pointed out that release of fungal spores so-called wet or dry showed strong and significant interannual differences. Our hypothesis is that in following years we will observe increased levels in allergenic fungal spore production as well as changes in species diversity. This study suggests that further research is needed to revise the grouping system of fungal spores as either "dry" or "wet" and simple BMC/AMC ratios and their response to climate change. Moreover, as the Paris region is impacted by severe chemical pollution events from different origins, it is of interest to understand, in future studies, (1) how atmospheric pollutants can exacerbate the human allergenic response when fungal spores are present in the air at high concentrations during air advection from "splash dispersal", (2) the interaction between pollen and fungal spores should be studied together as we observed that Ustilago has the same source point of the Poaceae pollen and they are together showing general decreasing concentrations, (3) to identify at a genus level all the fungal spore diversity present at the Saclay observatory.                   Figure A7. Origin of total atmospheric pollen grains year by year using SWIM model Origin. The white circles represent the wind speed scale in kilometer per hour (km/h). The color grid represents the estimated concentration (Nb#/m 3 ) for any wind speed and wind direction. Figure A8. Surface of big crops in France adapted from [71]. Figure A9. AFS abundance of the 5% referenced as others from total abundance AFS. Figure A8. Surface of big crops in France adapted from [71]. Figure A7. Origin of total atmospheric pollen grains year by year using SWIM model Origin. The white circles represent the wind speed scale in kilometer per hour (km/h). The color grid represents the estimated concentration (Nb#/m 3 ) for any wind speed and wind direction. Figure A8. Surface of big crops in France adapted from [71]. Figure A9. AFS abundance of the 5% referenced as others from total abundance AFS. Figure A9. AFS abundance of the 5% referenced as others from total abundance AFS.