1 Introduction

The long-range transport of pollen grains and spores has been reported and interpreted since the times of Erdtman (1937) and Gregory (1945). Such redistribution can be considered from two aspects: short-term changes in pollen concentrations over receptor regions and, important in the long-term, a large-scale redistribution of genetic material along atmospheric pathways (Lindgren et al. 1995). Recently, substantial progress has been achieved in understanding the short-term component, which is responsible for sudden and irregular modifications in the pollen seasons over large areas (Hjelmroos 1992; Rantio-Lehtimäki 1994; Corden et al. 2002; Latałowa et al. 2002; Damialis et al. 2004; Ranta et al. 2006; Sofiev et al. 2006a). Recent research has also shown that, although each individual episode is stochastic in nature, there are systematic patterns in the pollen distribution in Europe in spring. Attempts to reveal such patterns via multi-annual analysis have been made for, e.g., Finland and Russia by Sofiev et al. (2006a) and Siljamo et al. (2007a, b) and for, e.g., Denmark by Skjøth et al. (2007). Nevertheless, currently the phenomenon is still insufficiently understood, and there is no overall picture of pollen distribution across Europe due to continental-scale atmospheric transport.

Birch is one of the most important species with respect to long-range transport. First of all, the wide habitation area of birches allows an extensive dispersal of pollen. The natural distribution of silver birch (Betula pendula Roth.) and downy birch (B. pubescens L.) extends from the mountainous regions of southern Europe to northernmost Fennoscandia, and across Siberia to the east coast of Asia (Atkinson 1992; OECD 2003). In the European part of Russia, the total area of birch forests amounts to 3.66 × 107 ha. Second, birch pollen grain can be transported more than a thousand kilometers (Sofiev et al. 2006a) and, third, birch is an important allergenic plant in nearly all parts of Europe. In Northern Europe, birch pollen is the most abundant and important allergenic airborne pollen type, with approximately 15% of the population being sensitized to its allergens (WHO 2003).

Birches are broad-leaved and relatively fast-growing early-successional species preferring open and light habitats, such as areas disturbed by forest fires, storms, or clear-cut or abandoned fields (Achard et al. 2006). Apart from being an important part of the forest ecosystem in boreal and temperate zones, birches are economically important forest trees and common ornamental species.

Birches, and the other species (e.g., Alnus, Carpinus, Corylus) belonging to the order Fagales, are wind-pollinated trees. Pollen grain, which is the male gametophyte, must reach the stigma of the female flower for fertilization to take place. As the efficiency of fertilization largely depends on the amount of airborne pollen (Sarvas 1952), the production of male inflorescences is very large.

Birch pollen grains can be transported over hundreds or even thousands of kilometers if weather conditions are suitable (Sofiev et al. 2006a). Usually, most of the pollen is deposited near the source trees, but the total number of released grains is so high and the threshold levels for provoking symptoms in sensitized people are so low that comparatively small fractions of long-range-transported pollen can constitute a significant allergen presence. The health impact of an episode is considered strong if daily mean pollen counts exceed 100 grains m−3, which is the threshold limit used in Finland for all sensitized patients to show symptoms (Viander and Koivikko 1978). Long-range-transported pollen has been shown to cause allergic symptoms (Wallin et al. 1991). However, because the allergen content of birch pollen observed before the onset of local flowering varies (Matikainen and Rantio-Lehtimäki 2002), it is likely that in some cases long-range-transported pollen is lacking in allergenic material. The allergen content of long-range-transported pollen was not considered in this study.

In this study we consider three stages of the pollen season: pollen observed before local flowering (early season, long-range-transported pollen), pollen mainly from local sources (main season), and late pollen after local flowering (late season, long-range-transported or resuspended pollen). The main attention is paid to early-season pollen.

In Southern Finland birches usually flower a few days later than in the Southern and Western part of European Russia (Siljamo et al. 2008), i.e., Finland can receive pollen from Russia before and during the early part of the local flowering season. The extensive birch forests in Finland constitute a strong source of pollen that is also important on a European scale; there is a possibility that Russia could also receive pollen from northernmost Europe outside its own territory, and this could prolong or strengthen the late part of the local season. Indeed, according to the observations, near the end of the season, the central European part of Russia seems to receive some pollen from the north, possibly from North-Western Russia, Karelia, and Central and Eastern Finland. Such transport could explain the characteristic long tail in Moscow pollen counts.

There are no major differences between the dates of the pollination seasons in Central Russia and Finland, and the seasons in Moscow and southern Finland usually overlap at the end of April or the beginning of May. In principle, this would facilitate a gene redistribution in a north-south direction, which is facilitated by birch dependence on spring temperature accumulation (e.g., Häkkinen et al. 1998; Linkosalo 1999, etc.). Due to the stochastic features of weather conditions, the heat accumulation threshold for flowering may occasionally be met earlier in northern regions. Further, it has been shown that the leaf bud burst (taking place almost simultaneously with male flowering; Linkosalo 1999) may occur among the genotypes of one birch population at one site within two weeks (Rousi and Pusenius 2005).

As a result, depending on the meteorological conditions of the particular year, the patterns of daily pollen counts in the Finland and Moscow areas can be almost synchronous, such as those in 1999 and 2006 (in both cases the synchronization was due to long-range transport from Russia to Finland). By contrast, in 2004 the peak of concentrations was earlier at Turku than in Moscow (the grains at Turku were partly transported from the south and south-west, but also the pollination season started earlier at Turku than in Moscow) (Table 1). Such cases evidently facilitate gene exchange across areas that are thousands of kilometers in size.

Table 1 Birch pollen long-range transport episodes in Finland in 1999, 2002, 2003, 2004, and 2006
Full size table

Usually the first pollen grains are observed in Moscow in the first half of April, whereas the flowering there usually starts at the very end of the month. The vast birch forests of Central Russia result in a quite definite start to the local season: even if some long-range transport does take place before the season, the local impact is usually five to ten times stronger, so the concentrations will rapidly increase after the start of local pollination. The high-concentration period lasts for two to four weeks, after which low levels can be still observed until the end of summer—either due to transport from the north or due to the remaining pollen in the catkins of some trees in the region.

The overall objective of the current study was to evaluate the redistribution of natural allergenic and genetic material in Eastern and Northern Europe, using birch as a prominent example. In particular, we tried to identify the source origins of the birch pollen in the Moscow region, especially before the beginning of the local pollinating season. These results can be used to refine previous assessments of the pollen budget in Finland (i.e., the frequencies of the origin and receiver areas), and to evaluate the mutual exchange of birch pollen between Finland and Russia.

2 Model, methodology, and input datasets

We have combined observations of phenological stages, pollen concentrations in the atmosphere and dispersion modeling of pollen transport, both forward and inverse.

Concentrations of airborne allergenic pollen have been monitored in Finland since 1972 and in Russia since 1992. The operational aeroallergen network currently includes nine stations in Finland (http://aerobiologia.utu.fi/en/index.html), five stations in the European part of Russia, and two in Siberia (Fig. 1). In this study, we used all the observations in Finland and the data from one observational site that was available in the Moscow region. Other Russian stations have only been operating since 2003 and do not provide daily data.

Fig. 1
figure 1

Fraction of birch in Europe used in the SILAM model (revised on the basis of Sofiev et al. 2006a). Location of the pollen monitoring stations in Finland and Russia: Kevo (KE), Rovaniemi (RO), Oulu (OU), Vaasa (VA), Kuopio (KU), Kangasala (KA), Joutseno (JO), Turku (TU), Helsinki (HE), St Petersburg (StP), Yaroslavl (YA), Moscow (MO), Smolensk (SM) and Krasnodar (KR)

Full size image

At all the Finnish and Russian sites, pollen sampling was performed with the volumetric Burkard-spore trap (Hirst 1952) on an open rooftop. The height of the trap location varied from 10 m above ground level in Moscow to 15 m in Turku. The observational site in Moscow is assumed to be representative of the Moscow region; this is an area with an approximate diameter of 200–300 km. The technique of volumetric trapping is standard throughout most of Europe (British Aerobiology Federation 1995). At Russian sites, pollen grains are counted and identified by microscopic observation in 12 transverse strips every second hour and in Finland with the method of randomized fields (Mäkinen 1981). The results of these two different types of counting procedure have been shown to be comparable (Käpylä and Penttinen 1981). The pollen concentrations are the daily average pollen counts per m3 of air.

The flowering period for birch species is determined from phenological observations. The dates of leaf bud burst, leaf unfolding, and first flowering are used to determine the start of flowering. For Finland, observations of leaf bud burst were used, because no corresponding dataset for the onset of the male flowering of silver birch is available. According to Linkosalo (1999), the difference between the observed first date of male flowering and leaf bud burst is 1.1 days, with male flowering observed first. This result is based on over 750 observations of leaf bud burst and male flowering during 55 years in South and Central Finland. Observations are made using a standardized method repeatedly on the same tree individuals (Kubin et al. 2004) two to three times per week in the phenological network of the Finnish Forest Research Institute. Only silver birch was used in this study, because it flowers earlier than, or at the same time as, downy birch (Luomajoki 1999). The birch trees used are adult trees growing in exposed habitats representing typical conditions with regard to regional climate. The phenological observation sites were located within a range of 10–35 km of the pollen-monitoring sites.

Russian phenological observations, described by Siljamo et al. (2008), were originally collected by the Ecological Centre Pasva and Moscow State University. Only in 1994, 1996, 1997, and 2005 was the first flowering date available. In other years, leaf unfolding observations were used from five stations around Moscow. In the Finnish dataset, leaves are registered as unfolded when 50% of the leaves on five trees are unfolded (Kubin et al. 2004), whereas in the Russian dataset the figure is 10%. Phenological observation data were used without any unification adjustment, because the uncertainties in the data are higher than those due to the differences in methodologies (Rousi and Heinonen 2007; Siljamo et al. 2008).

The basis of the modeling approach to evaluation of pollen dispersion from known sources, and to delineating the sources of observed concentrations, is described by Sofiev et al. (2006a). Following these considerations, pollen is treated by the SILAM dispersion model as a coarse (about 20 μm in diameter) but light (800 kg m−3) atmospheric aerosol with a typical sedimentation velocity of ~0.01 m s−1 and a moderate (~0.02 m s−1; Chamberlain 1953) wet deposition.

Pollen simulations are done with the SILAM dispersion modeling system (Sofiev et al. 2006b). It is capable of both forward and inverse simulations, for which a Lagrangian random-walk particle model and Eulerian dynamics cores are available. The simulations presented here used SILAM v. 3.7 with a Lagrangian random-walk model.

For all simulations the meteorological data have been taken from the archives of the European Centre for Medium-Range Weather Forecast (ECMWF) with a time interval of 6 h and spatial resolution varying from 75 km for the earliest-considered years to 40 km for more recent periods.

Forward model simulations required description of the emission source term comprising the timing of the start of pollen emission (assumed to coincide with the birch flowering), its duration, intensity, and vertical and horizontal distribution. The map of European birch forests used (Fig. 1) was the same as that described by Sofiev et al. (2006a) and the vertical distribution of the emission flux was limited to a layer from 2 to 20 m above the surface. Forward simulations for the year 1999 were made using a passive tracer (air parcels) assuming continuous emission.

For the source delineation problem, the adjoint simulation mode was used. With this mode, the input for the model is compiled from the observations, while the output outlines the areas from where the observed substances can originate—represented in terms of a spatial probability distribution for the sources to be located within a specific region (footprint of the observations).

From previous work it is known that long-range pollen transport is episodic, and strongly varies from year to year and from day to day. Therefore, straightforward long-term simulations aiming at a semi-climatological “mean exchange” description would not provide a useful outcome. The variability would exceed the value itself. We have therefore focused on those years when episodic long-range transport was to be expected, based on the analysis of the observations, and the aimed at the quantification of the pollen exchange and its consequences for the local pollen seasons in both countries.

3 Re-analysis of the main episodes during 1994–2005

We have selected cases in which it was to be expected that long-range transport would probably influence the pollen seasons in Finland and in the Moscow region. Some of these, for pollen seasons in Finland, have been analyzed in previous studies by Ranta et al. (2006), Sofiev et al. (2006a). and Siljamo et al. (2007a, b). These analyses are not repeated here, but their outcomes are used throughout the paper. Selected characteristics of these previously-studied cases are summarized in Table 1.

3.1 Episode during 17–20 April 1994 in the Moscow region

In 1994, some pollen grains were observed in Moscow well before the local flowering (Fig. 2a). The first early concentration peak was on April 18 (68 grains m−3), whereas flowering started at the end of April or beginning of May. During this episode, a low-pressure area developed in Central Europe and gradually moved north-east, growing deeper and bringing southerly winds to the Moscow region. Adjoint model simulations showed that the footprint of this early episode extends to the south of Moscow (Fig. 2b), where air was warm and birch flowering was already in progress.

Fig. 2
figure 2

(a) Daily averages of Betula pollen counts in Moscow (dashed line) and Turku (solid line) in 1994. The rectangle shows the pollen counts the source origin of which was analyzed, and the grey horizontal line is a level of 100 grains m−3. (b) The cumulative sensitivity area (footprint for the observation marked with the rectangle) predicted by the SILAM model on 17–20 April, 1994. The grey shading shows the area from which air parcels were transported to Moscow within the analysis period. Moscow is marked with a plus symbol and Turku with a circle

Full size image

The second potential source of pollen in the course of this episode could have been Finland and its surrounding areas. However, the weather in Finland was cold and rainy during this episode, and local flowering had, therefore, not yet started. A few days after the episode, a high-pressure area developed east of Moscow; this resulted in light winds in the Moscow area, and pollen counts decreased to zero.

3.2 Episode during 16–21 April 1996 in the Moscow region

In 1996, relatively high early pollen counts were observed (Fig. 3a) in Moscow shortly before the main season. The first early concentration peak was observed on April 17 (444 grains m−3) and the second one two days later (179 grains m−3). Flowering started in Moscow a few days after the second peak.

Fig. 3
figure 3

(a) Daily averages of Betula pollen counts in Moscow (dashed line) and Turku (solid line) in 1996. The rectangle shows the pollen counts the source origin of which was analyzed, and the grey horizontal line is a level of 100 grains m−3. (b) The cumulative sensitivity area (footprint for the observation marked with the rectangle) predicted by the SILAM model on 16–21 April, 1996. The grey shading shows the area from which air parcels were transported to Moscow within the analysis period. Moscow is marked with a plus symbol and Turku with a circle

Full size image

Considering the weather situation, a tongue of relatively cold air extended from the Ural Mountains to the Balkan Peninsula during April 17–19. In contrast, the temperatures in Central Europe and over most of the Baltic Sea region were high and winds were light. On April 20, the cold air masses retreated to the north, allowing a sudden start of flowering in Central Russia.

On the other hand, the pollen counts measured in Moscow on April 17–19 could also have been caused by the forests in the vicinity of Moscow, where flowering was delayed for some days due to cold air masses. The short time difference between the concentration peak considered and the local pollination indicates that this may be the case.

According to the adjoint simulations, the air masses were transported to Moscow mainly from the north during the cold period, with a minor contribution from the south-west (Fig. 3b). However, it was too early for the birches in northern Europe to pollinate, while in western Russia the flowering had started just a few days earlier. Due to the warm air in the Baltic countries, flowering had also peaked there, so the pollen grains in Moscow could also have originated from the north-west.

3.3 Episode during 16–18 April 1997 in the Moscow region

In 1997, pollen grains were observed in Moscow almost three weeks before the main birch pollen season (Fig. 4a). This concentration peak occurred on April 16 (61 grains m−3), but local birch flowering in Moscow did not start until May 2–7. When the first birch pollen grains were observed, Europe lay between two strong high-pressure areas—one in Russia, east of Moscow, and the other in the Northern Atlantic. This weather situation caused southerly winds to occur in the Moscow region, and pollen was therefore transported from the south (Fig. 4b). A low-pressure area that started to develop on April 19 in Northern Europe caused the episode in Moscow to come to an end.

Fig. 4
figure 4

(a) Daily averages of Betula pollen counts in Moscow (dashed line) and Turku (solid line) in 1997. The rectangle shows the pollen counts the source origin of which was analyzed, and the grey horizontal line is a level of 100 grains m−3. (b) The cumulative sensitivity area (footprint for the observation marked with the rectangle) predicted by the SILAM model on 15–18 April, 1997. The grey shading shows the area from which air parcels were transported to Moscow within the analysis period. Moscow is marked with a plus symbol and Turku with a circle

Full size image

3.4 Episodes during 15–18 April, 20–30 April, and 3–5 May 1999 in the Moscow region and in Finland

In terms of birch pollen concentrations, the year 1999 was remarkable in both countries (Fig. 5). In Moscow, a long-duration high-concentration episode of birch pollen was observed; this was caused by a combination of long-range-transported and local pollen. For a detailed analysis, the episode is split into three parts, denoted in Fig. 5 as C1, C2, and C3.

Fig. 5
figure 5

Daily averages of Betula pollen counts in Moscow (dashed line) and Turku (solid line) in 1999. Boxes C1, C2, C3 show the pollen counts whose source was analyzed, and the grey line is a level of 100 grains m−3

Full size image

The first occasion when more than 10 grains m−3 was observed in Moscow, was on April 2; on April 18 (C1), the concentrations had increased to 2,446 grains m−3. However, local flowering started around the Moscow area on April 21–22.

The weather situation resembled the previously-considered case in 1997. In the middle of April, Europe was again between two high-pressure areas, and there were very warm air masses east of Moscow. This weather situation induced south-easterly winds, and Moscow was therefore affected by air masses originating in the Moscow region itself (not yet flowering) and in the south-eastern birch forests in Russia (already flowering) (Fig. 6a).

Fig. 6
figure 6

SILAM cumulative sensitivity area (footprint for the observation marked with the rectangles in Fig. 5) (a) April 16–18, 1999. The grey shading shows the area from which air parcels came to Moscow within analysis period C1 in Fig. 5. (b) April 20–30, 1999. The grey shading shows the area from which air parcels came to Moscow within analysis period C2 in Fig. 5. (c) May 1–5, 1999. The grey shading shows the area from which air parcels came to Moscow within analysis period C3 in Fig. 5. Moscow is marked with a plus symbol and Turku with a circle

Full size image

During April 20–30, Europe was affected by several high-pressure systems. At the beginning, there was one system in the southern North Atlantic and another in Russia that extended from the western part of European Russia over the Ural mountains. Later, at the end of April, a single high-pressure area developed over western European Russia and the Balkan peninsula. This weather structure maintained a permanent flow near the Moscow area, and the city was therefore influenced by dry and very warm air from the south-eastern and southern sources (Fig. 6b). Conditions for flowering in the source areas were fine, the transport was direct and undisturbed, and the pollen counts therefore increased to 5,000 grains m−3.

At the beginning of May, a high-pressure area developed over the Norwegian Sea. In Moscow, winds became northerly, a cold air mass reached the Black Sea and the pollen counts decreased (Fig. 6c).

In Finland, exceptionally high birch pollen concentrations were observed at all six Finnish monitoring sites during the same days (April 18–21) as in Moscow (Fig. 5). In that year spring was late in Finland, with birches starting to flower in Southern Finland on May 5–6. Local flowering was also exceptionally weak.

Our previous inverse study showed that the pollen concentrations in Finland were transported mainly from western and southern Russia, and partly from the southern Baltic or Poland, and Germany (Ranta et al. 2006). In the following text, we provide a more detailed examination of this complex case.

On April 18–19, during the birch pollen long-range transport episode in Finland, one low-pressure center was over Sweden and another in Eastern Europe. A high-pressure area was located in Russia. Between the two low-pressure areas there was a saddle surface, making the flow pattern dynamic and uncertain over Finland. On April 18, the pressure saddle surface was situated over Poland, and on April 19 it moved towards the Baltic countries. As a result, the main sources of the pollen peak C1 in Turku were almost the same as those for C1 in Moscow—the birch forests located to the south of the Moscow region. A shift in weather patterns on April 19 broke this connection: the winds in Southern Finland changed from south-easterly to south-westerly, and the observational footprint of all Finnish sites moved to the Baltic countries, where flowering had not yet started. As might be expected, pollen counts decreased.

When the low pressure over Sweden dissipated, the other low-pressure center situated in Western Russia and Poland cased the atmospheric flow in Finland to become south-easterly again, and the pollen counts again increased. The episode ended on April 22, when the low-pressure area in the Atlantic Ocean started to influence the weather in Finland.

Figure 7a–d shows the results of forward pollen dispersion simulations with the sources located only in Central Russia (Fig. 7, the area within the rectangle). It can be seen that Finnish pollen stations are influenced by the sources in that region twice: first during April 18, and again during April 20–21. According to the observations, these were indeed the main peaks recorded over most of Finland. Some pollen was also observed during April 19, but the concentrations were small and in fact originated in the Baltic countries (Fig. 8), where the main pollination had not yet started. Figures 7 and 8 show that the Russian forests were strong contributors to the highest concentration episodes, while the influence of the southern Baltic and Central Europe was comparatively small.

Fig. 7
figure 7figure 7

SILAM forward run with a passive tracer from Western/Central Russia. The release starts on April 16, 1999. The source area is the interior of the box masked with the birch fraction map (Fig. 1). Area of risk in 1999 (a) April 18, (b) April 19, (c) April 20, (d) April 21

Full size image
Fig. 8
figure 8figure 8

SILAM forward run with a passive tracer from the Baltic countries. Release starts on April 16, 1999. The source area is the interior of the box masked with the birch fraction map (Fig. 1). Area of risk in 1999 (a) April 18, (b) April 19, (c) April 20, (d) April 21

Full size image

3.5 The episode during 16–19 April 2002 in the Moscow region and in Finland

In 2002, some pollen grains were observed in both Moscow and Turku before the main pollen season started (Fig. 9a) but, in contrast with 1999, the seasons were unsynchronized. In Moscow, the pre-season concentrations on April 16–17 were moderate (100 grains m−3) but, according to phenological observations, flowering started around the Moscow area on April 23–26, so that the second peak on April 22 could also be caused by long-range-transported pollen. The main season was one of the strongest in the re-analysis: the highest count reached almost 6,000 grains m−3. A high-pressure area east of Moscow dominated the weather situation during the episode and winds were light, which also commonly implies that the wind direction was variable. Around Moscow and west and south-west of Moscow the air was warm, so the air mass that brought the first pollen grains (box in Fig. 9a) came mainly from the westerly component, from Belarus and west of Moscow (Fig. 9b). The northern part of the footprint was probably not able to produce pollen.

Fig. 9
figure 9

(a) Daily averages of Betula pollen counts in Moscow (dashed line) and Turku (solid line) in 2002. The rectangle shows the pollen counts the source origin of which was analyzed, and the grey horizontal line is a level of 100 grains m−3. (b) The cumulative sensitivity area (footprint for the observation marked with the rectangle) predicted by the SILAM model on 15–19 April, 2002. The grey shading shows the area from which air parcels were transported to Moscow within the analysis period. Moscow is marked with a plus symbol and Turku with a circle

Full size image

The first occasions on which Finnish monitoring stations observed long-range-transported pollen was on April 12–14 and again on April 27–30. In the second case, a low-pressure center located west or south-west of Norway brought about southerly and south-easterly flows in Finland. At the same time or slightly earlier, the main peak of concentrations was observed in Moscow (Fig. 9a), again suggesting a connection. Sofiev et al. (2006a) and Siljamo et al. (2007a) have analyzed that episode, concluding that pollen indeed came to Finland from Russia, but that Belorussian and Ukrainian forests also contributed to the peak. In 2002 there was early bud break in Finland, and in south-east Finland local flowering overlapped with long-range transport.

3.6 Episode during 20–24 April 2004 in the Moscow region and in Finland

In 2004, the pollen season started earlier at Turku than in Moscow (Fig. 10a). During the episode of April 15–19, there were warm air masses in western Europe and cold air to the east of Moscow. At Turku the first episode on April 14–15 was mainly caused by long-range-transported grains from the Ukraine, but also partly from Central Europe (Sofiev et al. 2006a; Siljamo et al. 2007a). But in Moscow, also, the early peaks were observed well before the main pollen season. The west-east temperature gradient turned south–north during April 20–22, at the same time as the first pollen peak was observed on April 22 in Moscow (the flowering started on May 4–7). Elevated pollen counts (Fig. 10a, delineated by a grey box) during April 22–24 in Moscow were due to transport from the west and the south. Noticeably, winds turned northerly by April 23, creating the northern part of the footprint (Fig. 10b), but flowering over that territory had not yet started.

Fig. 10
figure 10

(a) Daily averages of Betula pollen counts in Moscow (dashed line) and Turku (solid line) in 2004. The rectangle shows the pollen counts the source origin of which was analyzed, and the grey horizontal line is a level of 100 grains m−3. (b) The cumulative sensitivity area (footprint for the observation marked with the rectangle) predicted by the SILAM model on 20–24 April, 2004. The grey shading shows the area from which air parcels were transported to Moscow within the analysis period. Moscow is marked with a plus symbol and Turku with a circle

Full size image

3.7 Episode during 30 April–3 May 2005 in the Moscow region

Early pollen was again observed in Moscow in 2005. The first time when pollen counts were higher than 10 grains m−3 was on April 28, but the first peak occurred on May 1 (445 grains m−3, Fig. 11a). Usually birch flowering first starts to the south and west of Moscow, but in 2005 warm air masses were located for a long time east of Moscow, and thus birch flowering probably first started there. Therefore, also, the early observed pollen grains in Moscow could come from the east (Fig. 11b). After that, warm air penetrated into the Moscow area, local flowering started on May 3–4, and pollen counts increased.

Fig. 11
figure 11

(a) Daily averages of Betula pollen counts in Moscow (dashed line) and Turku (solid line) in 2005. The rectangle shows the pollen counts the source origin of which was analyzed, and the grey horizontal line is a level of 100 grains m−3. (b) The cumulative sensitivity area (footprint for the observation marked with the rectangle) predicted by the SILAM model on 26 April–3 May, 2005. The grey shading shows the area from which air parcels were transported to Moscow within the analysis period. Moscow is marked with a plus symbol and Turku with a circle

Full size image

4 Discussion

Some general characteristics were discovered for the predominantly early-season episodes studied (i.e., episodes that took place before the local flowering had started). Regarding the five episodes considered for which Finland was a receiver of birch pollen, in three cases the main source areas were found to be in Russia. Regarding the early-season cases in the Moscow area, none of the considered episodes was caused by the transport of pollen from Finland.

If the first pollen grains were observed in Moscow weeks before local pollination (three cases out of seven), the source areas were always located to the south of Moscow. However, if the first grains were observed shortly before local flowering (four cases out of seven), the source areas could be situated both to the west and the north-west (2/7), or to the east and the south-east (2/7).

In the Moscow region, northerly winds at the beginning of the season transport pollen-free air, as flowering in Finland and North-Western Russia starts predominantly later than in Moscow. Later in spring, such winds can transport pollen that originates in northern Europe to the Moscow region in such quantities that sensitive people are affected. However, the local birch population surrounding the Moscow region is so extensive that such episodes, which are observed almost every year, are often difficult or impossible to identify, as the local forests commonly produce pollen concentrations during the main season that are higher by one or two orders of magnitude.

There was only one episodic case (in 2005), in which source apportionment showed that the early pollen could be transported to Moscow only from the east or the north-west. The probability of early pollen transport from the north-west is small, as even in western Finland (where the flowering starts first) the concentrations were still zero, and the phenological observations did not show any flowering at that time. These pollen grains in Moscow were therefore more likely to have come from an easterly direction, where flowering could have started already, caused by warmer temperatures.

In Finland, the sources of early pollen are more diverse. Depending on the specific meteorological conditions, the source areas can be in the west, the south or the south-west. The strengths of the episodes largely depend on their origin. In central Europe, birch forests are less abundant than in southern Finland; however, these are, in turn, much less extensive than the vast birch forests in central regions of European Russia. If the early pollen arrives in Finland from the south-west or the west, the concentrations are usually fairly low or moderate, while atmospheric transport from southern and south-eastern directions can carry such amounts of pollen that the resulting concentrations are very comparable with or even exceed, typical concentrations from local origins. The most recent case took place in 2006, when vast pollen-rich plumes were transported as far as Iceland, and caused allergic symptoms over most of Central and Northern Europe (Sofiev et al. 2008b, c).

Analyses of the main flowering period and the concentrations after the season are less conclusive. Probably these stages are also less important with regard to the short-term forecasting or re-analysis of pollen seasons, but are still important for the assessment of the spatial transfer of genes. Indeed, the pollen observations cannot differentiate between local and long-range-transported pollen, while the model estimates alone are somewhat ambiguous, due to the uncertainties in the modeling of the end of season.

Generally, after pollen is released, its distribution follows the main atmospheric circulation patterns; westerly and south-westerly transport therefore dominate under European conditions. Consequently, pollen from Finland commonly reaches north-western Russia, but is rarely transported to the Moscow region. Episodes with such transport certainly exist, but these are most frequently hardly recognizable in the presence of other factors, such as the fluctuations of local flowering intensity, convective precipitation, etc.

The impact of the northern European birch forests in Finnish Lapland, Karelia, and North-Western Russia becomes detectable again in Moscow by the end of May, when the flowering is over in all other regions. An illustration of such an impact has been presented (Sofiev et al. 2008a); it was shown that during a one-day episode on 7 June, 2005, pollen from Finnish and Swedish Lapland reached Denmark and Germany (these model simulations were in good agreement with the pollen observations of the European Aeroallergen Network). In the current analysis, we have not identified other similar cases with certainty, but the temporally long periods of non-zero pollen concentrations after the local flowering during summer in Moscow may originate in such events. However, their detailed analysis would require more sophisticated modeling of the flowering than that used in this study.

Redistribution of pollen during the main flowering season may have an impact on the transport of genetic material over large distances (Lindgren et al. 1995). Such transport can be of great importance due to the changing climate and the gradual shifts of climatic zones that will force trees to adapt to the warming climate (Walther 2003). For example, Finland is expected to experience an increase in mean annual temperature of 3–5°C during the next 100 years (Carter et al. 2005). Since these changes are very rapid in comparison with evolutionary time-scales, a quick atmospheric pathway may be crucial for the adaptation processes, on the presumption that pollen remains viable during exposure in the atmosphere. The results from Lindgren et al. (1995) suggest that viable pollen of a wind-pollinated conifer, Scots pine (Pinus sylvestris), may originate hundreds of kilometers away. On the other hand, Aylor (2004), who studied the effects of atmospheric conditions on the survival of maize pollen (Poaceae: Zea mays) showed that only 60–240 min of direct exposure to sunlight and air were needed to reduce the germinability by 50%.

In summary, studies on the long-range transport of pollen are closely connected with several fields of biology and medicine, for example:

  1. 1

    reconstruction of past vegetation and evaluation of vegetation patterns around the sampling site (Ranta et al., in press);

  2. 2

    impact on agriculture and forestry due to transport of genetic material and effects of transported pathogenic organisms (Barnes et al. 2008); and

  3. 3

    transport of aeroallergens.

This study discovered systematic patterns of pollen transport; these results can therefore be used in the above-mentioned fields of science.

5 Conclusions

The current study addresses the large-scale redistribution of natural bioaerosols in Europe, using birch pollen exchange between Finland and Central Russia as a prominent example. A combination of Finnish and Russian aerobiological and phenological observations was used, together with dispersion modeling with the SILAM system. We have highlighted the main sources of non-local pollen in both southern Finland and central Russia, and presented the main patterns of the mutual exchange of pollen for Eastern and North-Eastern Europe.

The results indicate that transport of birch pollen over long distances is a common event between Russia and Finland. More generally, it can be assumed that in Northern and North-Eastern Europe, alien birch pollen may contribute substantially to the concentrations observed at nearly every site. This has implications both to human health and to evolution of the genetic structure of birch populations.

Early-spring pollen in Moscow originates extremely rarely in Finland, where flowering usually starts later. The most important source areas are located to the west and south of Moscow, sometimes also in the east. For Finland, the situation is the opposite: early-spring pollen frequently originates in Central and Western Russia, but it may also be transported from Central and Western Europe.

In Finland, long-range transport from the south-east can therefore cause sudden increases in pollen concentrations even weeks before the beginning of local flowering. For Central Russia, the reverse is true: pollen from Finnish (or northern Russian) birch forests can extend the local season or cause birch pollen outbreaks after the end of local flowering.

The long-term average dispersion of pollen during the main season largely follows the prevailing atmospheric global circulation: from west to east, and from south-west to north-east. Clearly, specific episodes can exhibit much more complicated transport patterns. Their analysis relies to a large extent on the modeling of pollen release and atmospheric transport, as the measurements cannot provide information about the origin of the pollen.

Apart from the short-term impacts on the pollen seasons, long-range transport can have substantial impacts on the dispersion of genetic material within Europe. A quick atmospheric pathway for gene transport can be important for plant adaptation in a changing climate.