The NFBAP exhibited clear seasonal patterns, with the overall highest NFBAP concentrations in the summer and the lowest ones in the winter, respectively (P<0.0001). The NFBAP data covered approximately two full fall seasons, two winter seasons, approximately one-and-a-half spring seasons, and one full summer season (see Table 2). Figure 1 shows NFBAP at a 5 min sampling resolution (green) and 7 d median (magenta). Seasons according to the SMHI are delimited by vertical lines.

The median monthly NFBAP concentration over the full campaign was 0.005 cm-3 but varied significantly (P<0.0001) over the full campaign. Monthly and seasonal distributions of NFBAP, NTAP, and NFBAP/NTAP are shown in Figs. 2 and 3 (Fig. S2 in the Supplement shows the full NFBAP/NTAP time series with seasons indicated). The monthly median varied by a factor of ∼ 13, from a minimum of 0.0015 cm-3 in December 2021 to a maximum of 0.019 cm-3 in July 2021. The period with the highest NFBAP abundance was initiated by a steep increase in May, where NFBAP increased by a factor of ∼ 8 over a few weeks. The period with high NFBAP ended with a steep decrease over a few weeks between November and December 2021, when the concentration decreased by a factor of ∼ 8. A similar decrease by a factor of ∼ 4 was observed in winter 2020. NFBAP varied a lot over the summer months but was more stable during fall and winter, when the concentrations were lower. The overall NFBAP agreed with other long-term measurements conducted in Finland, the USA, Germany, France, India, and China (Huffman et al., 2010; Schumacher et al., 2013; Toprak and Schnaiter, 2013; Valsan et al., 2016; Yu et al., 2016). The NFBAP concentrations in winters were, as expected, lower compared to the other seasons. Cold temperatures, lowered biological activity, low absolute humidity, and snow coverages are suggested to cause a lower generated amount of bioaerosols in winter but also reduce bioaerosols' abilities to be lofted into the air (Schumacher et al., 2013; Toprak and Schnaiter, 2013; Saari et al., 2015; Huffman et al., 2010). At the beginning of fall, spores from fungi are dispersed in the air, which increases the fall NFBAP (Schumacher et al., 2013; Toprak and Schnaiter, 2013). This also explains the pattern observed here, where an increase in bioaerosols in fall is plausibly due to a combination of pollen dispersed from the late summer in combination with fungi and spores in late summer and early fall (September–October primarily).

Not all biological material will have a sufficiently strong fluorescence signal to be detected as FBAPs. This is because the fluorescence signal is a function of the concentration of fluorophores in PBAPs, the ability to be excited by the laser of the instrument, and the presence of opaque or absorbing material that may fluoresce only very weakly. Moreover, the detection threshold of ∼ 1 µm biases the reported FBAPs low. Therefore, fluorescence measurements, such as the ones made here, have the risk of underestimating the total biological material present, and the NFBAP measurements should be viewed as lower limits of PBAPs. This was also suggested by Huffman et al. (2012, 2010) when using a UV-APS for FBAP measurements. Also, further investigations are needed to better understand the response of BioTrak to different types of biogenic aerosol particles and to quantity potential interferences with non-biogenic particles.

The BioTrak instrument remains to be further characterized through comparisons with other bioaerosol monitoring instruments using well-characterized fluorescent particles and by comparison to other offline techniques. Atmospheric NFBAP depends a lot on local sources. Moreover, NFBAP discrepancies between different studies (see Table 1) can also be explained by differences between different types of instruments. To understand how the bioaerosol data collected here compare to other long-term studies would require thorough intercomparisons between the different types of instruments that have been employed in different studies. The UV-APS measures autofluorescence between 420 and 575 nm after excitation with a 355 nm laser at a flow rate of 1 L min-1, while the WIBS measures autofluorescence between 310–400 and 400–600 nm upon excitation with 280 and 370 nm xenon lamps at a flow rate of 0.3 L min-1. The BioScout excites particles with a 405 nm laser and measures autofluorescence at >442 nm with an adjustable flow rate and a default value of 2 L min-1. The BioScout has been suggested to detect ambient bioaerosols more efficiently than the UV-APS (Saari et al., 2014). BioTrak was developed for the purpose of detecting bioaerosols in pharmaceutical production and clean environments; thus, it operates at a higher flow rate, 28 L min-1, than other LIF instruments. To further understand the potential of using the BioTrak in future studies, the BioTrak needs to be compared to other new technologies for automatic bioaerosol monitoring. Recently a potential standardized validation method for assessment of counting efficiency and fluorescent measurement of bioaerosol instruments that could be used for validation of the BioTrak instrument was presented (Lieberherr et al., 2021). Validations could also be performed with calibration techniques for optical particle counters (Iida et al., 2014). To fully understand how the BioTrak data compare to other bioaerosol measurement techniques, BioTrak measurements need to be benchmarked against offline filter analyses.

Total aerosol particle number concentration did not exhibit the same patterns as the NFBAP concentrations. The NTAP concentrations are shown on monthly and seasonal scales in Figs. 2b and 3b (Fig. S1 in the Supplement shows the full NTAP time series with the seasons indicated). Compared to the NFBAP concentrations, the TAP concentrations were more homogeneous over the full campaign. No distinct peak of NTAP during warmer periods was observed. The average highest NTAP was measured in October and November 2020 and in February and March 2021. The lowest NTAP was observed in January 2021 and in December 2022. The NTAP varied more in fall and winter compared to summer, as indicated by the 5th and 95th percentiles in Fig. 3b. No significant difference in NTAP was observed over the seasons. The NFBAP/NTAP ratio had its highest values in June, July, August, and September. The average highest relative contribution of FBAPs (0.006) was observed in July.

The seasonal-average NTAP was between 3.42 and 4.96 cm-3, which is higher than in most other similar studies where both NFBAP and NTAP have been measured. For instance, Schumacher et al. (2013) reported an average NTAP between 0.41 and 0.47 cm-3 in Finland and between 0.20 and 0.73 cm-3 in the USA measured with a UV-APS. Valsan et al. (2016) reported a mean NTAP between 0.96 and 2.66 cm-3 on a monthly scale in India measured with a UV-APS, and Toprak et al. (2012, 2013) reported a mean NTAP between 0.47 and 0.69 cm-3 in Germany measured with a WIBS-4. Comparisons suggest that our NTAP measurements were about ∼ 2–10 times higher than the NTAP previously reported in combination with NFBAP measurements. Note that these differences also influence the difference in ratios between NFBAP and NTAP. One factor that could possibly help explain these differences is that the BioTrak instrument operates at a higher sample flow (28 L min-1) compared to other LIF instruments, which commonly have sample flows ranging from 0.3 to 2 L min-1. However, a comparison made between the BioTrak and a Grimm OPC showed that the BioTrak was closely correlated with the TAP concentrations measured with the Grimm OPC (see Figs. S7 and S8 in the Supplement). Therefore, the major difference between the TAP measured in this campaign and previous long-term bioaerosol and TAP measurements is presumably due to local sources of the particle concentrations. It should also be noted that we only counted particles with diameters between 1 and 12 µm, which means that results from studies where other particle diameter ranges were studied might not be adequately comparable. To understand these differences in more detail, further analysis of local and regional sources is needed. Since the measurement site is an integrated ACTRIS and ICOS station, there are other aerosol measurements available. However, these measurements are only of fine particles (≤1 µm) and therefore are not eligible for comparison with the NTAP and NFBAP measurements. There are plans to also install micron-sized aerosol measurements at the site. In the future we will continuously compare BioTrak measurements to these.

Variations in the daily bioaerosol concentrations were studied by averaging the hourly NFBAP abundance for each season (see Figs. S3 and S4 in the Supplement). Daytime NFBAP was not significantly different from NFBAP measured during the night, and the daily variations were numerically very small. On an hourly basis, daily NFBAP and NFBAP/NTAP peaked in the afternoon or evening, when relative humidity was relatively low and the temperature high in winter, spring, and summer. In fall, NFBAP peaked later in the evening.

Overall, the daily relative humidity curve was smooth and repeated the same pattern in all the seasons: relative humidity was low in daytime and increasing at night due to the decrease in temperature. In summer, the minimum relative humidity was aligned with the NFBAP peak. In fall and spring, however, the NFBAP peak was preceded by the minimum relative humidity. The wind speed maximum was on average aligned with the daily temperature peak, and average wind directions were on average coming from the south in the morning and more from the southwest later during the day, the evening, and the night.

Over the full campaign, the smallest particles, 1–3 µm, on average made up ∼ 70 % of the total NFBAP. As a comparison, particles with sizes 3–5 and 5–12 µm contributed ∼ 25 % and ∼ 5 %, respectively. The contributions from larger particles, 3–12 µm, were higher in spring and summer compared to fall and winter. The constantly elevated NFBAP in the 1–3 µm range was consistent with previous observations (Artaxo and Hansson, 1995; Huffman et al., 2010; Schumacher et al., 2013; Healy et al., 2014; Valsan et al., 2016). The seasonal difference also agreed with previously reported results (Schumacher et al., 2013; Huffman et al., 2010). Figure 4 displays the NFBAP particle size distribution for all seasons divided into the three differently sized bins.

While LIF measurements do not distinguish between different bioaerosol types, particle size can give some indication of the kind of particle. Single bacterial and fungal cells and fungal spores typically have sizes of 1–3 µm. Particles with sizes 3–5 and 5–12 µm correspond to the sizes of larger fungal spores and smaller-sized pollen grains, although most pollen is too large to be sampled by the instrumentation used. The higher contribution of larger particles during spring and summer can potentially be explained by the spread of pollen during these seasons. Further confirmation and identification of the origin and the sizes of the bioaerosols measured at Hyltemossa will be performed in follow-up studies with microscopic analysis but are beyond the scope of this work.

The air temperatures ranged from -2 to 20 ∘C in fall, from -10 to 10 ∘C in winter, from -2 to 18 ∘C in spring, and from 6 to 30 ∘C in summer. Figure 5 shows the observed relationship between NFBAP and temperature: median NFBAP binned based on air temperature separately for each season (Fig. 5a) and the 7 d Pearson rolling correlation coefficient r for NFBAP with air temperature for the full campaign (Fig. 5b). From the air-temperature-binned data, NFBAP was significantly positively correlated with increasing ambient air temperature in spring (r=0.88, P<0.01; ρ=0.83, P<0.01) and summer (r=0.88, P<0.0001; ρ=0.94, P<0.0001). No significant associations between NFBAP and air temperature were identified in fall and winter. The 7 d rolling correlation coefficient r, based on hourly mean values shown in Fig. 5b, indicates that for most of the campaign (60 % of the rolling correlation periods studied) NFBAP increases were significantly (P<0.05) correlated with increasing air temperatures. It should be noted that the correlation coefficient for the most part indicated only a moderate correlation (r between 0.25 and 0.5), and the correlation was rarely strong (r above 0.7). Negative correlations between NFBAP and air temperature were rarely observed (only in 13 % of the periods studied) but consistently only observed during winter months. The rolling correlation was nonsignificant during a considerable part of the campaign (27 %). The presence of positive, negative, and nonsignificant correlations between NFBAP and temperature in fall and winter can explain why no overall consistent relationship between air temperature and NFBAP was observed. The data presented in Fig. 5 suggest that the processes that determine the release of FBAPs are strongly dependent on season and on mechanisms that potentially require a minimum temperature or that are at least correlated with increasing air temperatures.

The observed relationship between NFBAP and relative humidity was complex and calls for more detailed studies. Figure 6 displays the average relationship between NFBAP and relative humidity, with NFBAP binned based on relative humidity (6a) and the 7 d rolling correlation coefficient for NFBAP and relative humidity (6b). On a seasonal level, NFBAP was positively correlated with relative humidity (r=0.21, P<0.05; ρ=0.26, P<0.05) in fall but negatively correlated with relative humidity in summer (r=-0.65, P<0.05; ρ=-0.71, P<0.0001) and in winter (r=-0.14, P<0.05; ρ=-0.12; P<0.05), while no correlation was found between relative humidity and NFBAP in spring. The observed positive correlation between NFBAP and relative humidity in fall may indicate that the bioaerosols detected in fall were potentially generated or ejected due to relative humidity-dependent mechanisms. On the other hand, the negative correlation between NFBAP and relative humidity in winter and summer suggests that the high relative air humidity may be a limiting factor in the release of bioaerosols during those seasons. In particular, the high NFBAP at relative humidity ∼ 20 %–40 % suggests that relatively dry conditions increase the concentrations of bioaerosols generated in the summer. The aerosol particles were not dried prior to BioTrak sampling, which in high outdoor temperature and humidity could lead to particle growth and hence particles being classified as larger than their actual size.

From the 7 d rolling correlation coefficient r between NFBAP and relative humidity for the full campaign, one can see that the correlation between NFBAP and relative humidity was mostly nonsignificant (46 %), while it was negative and significant for a large part of the campaign (39 %) and only rarely positive (14 %). The observed relationship that NFBAP and relative humidity are uncorrelated or negatively correlated is in contrast to what other studies report. In other long-term studies, the connection between NFBAP and relative humidity has been reported as overall positive (Huffman et al., 2010, 2012; Schumacher et al., 2013; Toprak and Schnaiter, 2013; Valsan et al., 2016), although in some of the cases only low correlation coefficients (but not reported how low) were identified (Huffman et al., 2010), in other cases correlation coefficients and significance levels were not reported (Huffman et al., 2012), and the NFBAP and RH relationship was inconsistent (Schumacher et al., 2013).

Relative humidity correlates inversely with air temperature since the temperature affects the saturation water vapor pressure in air. This relationship was also observed for the relative humidity and air temperatures measured (r=-0.30, P<0.0001; ρ=-0.28, P<0.0001), and the relationship can be observed in the daily trends in Fig. S3. Based on the positive relationship NFBAP had been observed to have with increasing air temperatures, it might be plausible to think that part of the negative correlation observed between relative humidity and NFBAP could be related to the relationship between relative humidity and air temperature. However, the relationships are complex, and the causality between the different effects can only be interpreted after more detailed studies. The data reported here underline the need for further studies of the relationship between relative humidity and NFBAP under different conditions and in different types of environments, especially since different types of bioaerosols have been observed to be both positively and negatively correlated with relative humidity.

The NFBAP was positively correlated with increasing wind speed. Figure 7 displays the average relationship between NFBAP and wind speed for the four seasons (a) and for most of the campaign as assessed with the 7 d rolling correlation between NFBAP and wind speed (b). The NFBAP was positively significantly correlated with wind speed in the winter (r=0.89, P<0.0001; ρ=0.95, P<0.0001), spring (r=0.70, P<0.05), and summer (r=0.64, P=0.05; ρ=0.76, P<0.05). In fall a similar but nonsignificant correlation was observed (r=0.58, P=0.08; ρ=0.46, P=0.18). These correlations were also confirmed by the rolling correlation coefficient. Figure 7b shows that for a large part of the campaign time the relationship between NFBAP and wind speed was nonsignificant (42 %) or positive (39 %), and only rarely was there a negative and significant relationship (15 %).

Figure 8 shows NFBAP as a function of wind direction. For all seasons, the winds coming from northeast to southeast were correlated with the highest NFBAP abundances. Meanwhile, winds from southwest to northeast were in general associated with the lowest NFBAP. It was also noted in the data that in fall and winter the coldest air temperatures were correlated with winds from the north, while air temperatures in the spring and summer were more independent of wind directions. While NFBAP was not significantly different depending on wind direction, the types of bioaerosols may still vary with wind directions. Studies on long-range transport of air masses were beyond the scope of this study but could have been indicative of a better understanding of these data.

In some cases of rain, but not consistently for all rain events, a substantial increase in FBAP number concentrations was observed before, during, and right after rain. To test the robustness of the analysis, a rain event was defined by the threshold of 0.5 mm h-1. Although we also applied other thresholds (including 1 and 2 mm h-1), none of the thresholds that were used indicated a significant correlation between precipitation and NFBAP. For certain rain events, the FBAP concentration was observed to increase by a factor of 4–10 when compared to NFBAP outside of the rain event. The pattern was seen for all the seasons. A total number of 90, 44, 35, and 64 individual rain events were identified in fall, winter, spring, and summer, respectively. In about 50 % of all cases of a rain event, an immediate but not lasting increase in NFBAP concentrations was observed, but the effect was not statistically significant. The NFBAP concentrations varied a lot both before and after rain events, and sharp increases in NFBAP were also observed unrelated to rain. Overall, the FBAP concentration variations were large, and therefore sudden and instantaneous variations, which could plausibly be explained by rainfall, were not significant over full seasons. Figures S5 and S6 in the Supplement show examples of rain events and the distribution of NFBAP before, during, and after rain and when there is no rain. In these figures high-precipitation peaks were followed by, or occurred simultaneously with, peaks in NFBAP.

Only local rain was accounted for, and, therefore, the effect of rain upwind and possible transport of FBAPs to the measurement site was not detangled. On average, a rain event lasted a few hours, but during certain periods, the frequency of such rain events was high. The overall number of rain events and the increased relative humidity associated with such events could very well have a larger and longer-ranging effect than distinguished here. The relationships between biological aerosol particles and rain have been reported on for a long time (Gregory and Hirst, 1957; Hirst and Stedman, 1963) and call for standardized methods for the association between PBAPs and precipitation. Rainfall can be important for both scavenging of bioaerosols and the bioaerosol release.

Connecting the observed results in this study with mechanisms for aerosolization and release of biological particles allows a greater understanding of seasonal variations in NFBAP. In summer, there was a positive correlation with air temperature and wind speed, while the relationship was mostly nonsignificant or negative with relative humidity. This suggests that wind-induced bioaerosol generation was favored under dry and warm conditions. This was likely the case during spring as well, when NFBAP correlated positively with air temperature and wind speed.

During fall, the highest number of rain events was observed as well as a small positive correlation between NFBAP and relative humidity. Bioaerosols can be generated and dispersed by mechanical ejections and bubble-bursting processes when raindrops impact surfaces (Alsved et al., 2019; Kim et al., 2019), which could explain the, in some cases strong, association between NFBAP and rain events. However, rainfalls are also known to clean the air from aerosol particles in the lower troposphere (Moore et al., 2020), which can explain the inconsistent association between NFBAP and rain events. Although many studies have found a generally negative correlation between airborne spore concentrations and relative humidity, some fungal species absorb water from the air, causing swelling of the mucilage and subsequent explosive release of spores (Grinn-Gofron and Bosiacka, 2015). High relative humidity is also known to cause pollen to rupture, resulting in the release of smaller pollen fragments (Taylor et al., 2004). It is also noteworthy that different types of bioaerosols can have different relationships with humidity and air temperature, so that the common effects are masked. As noted by Oliviera et al. (2009), while some spore types have been observed to be negatively correlated with temperature but positively correlated with relative humidity, other spores showed the opposite correlations (Oliveira et al., 2009).

The lowest levels and variations of NFBAP were found during winter (Fig. 3a), likely due to the lower temperatures, sunlight, and biological activity during this season. Again, low relative humidity and wind speed were correlated with higher NFBAP, indicating wind-induced aerosolization. In southern Sweden, cold temperatures are often correlated with northern winds, which was observed for fall and winter in the meteorological data studied here. This agrees well with the lowest NFBAP levels found when winds came from the north. It should also be noted that the difference in wind direction could be indicative of different sources of bioaerosols.

For further understanding of the data, FBAP release mechanisms, and FBAP sources, detailed biological analyses, including fluorescence microscopy of PBAP filter sampling, are needed to identify the different types of bioaerosols that are present during different seasons.