Grass pollen seasons in Poland against a background of the meteorological conditions

The paper refers to the estimation of Poaceae pollen seasons in Poland in selected areas. The aim of the study was to present the long-term variability of the start, end and duration of grass pollen seasons and the seasonal pollen index (SPI) in Poland against a background of the meteorological conditions over pollen seasons. The study was performed in eight Polish cities in 1992-2014 (the common seasons were 2003-2012). Pollen season start was relatively stable in the studied period, the seasons began about the 10th of May, a bit earlier in the south part of Poland. Pollen season ends were more changeable in comparison to the season start and fluctuated from the middle of July to the middle of September. SPI clearly depended on temperature and precipitation in April-August. Daily maximum pollen concentrations were achieved between the end of May and the first decade of July and no evident relationship between this day and weather conditions was found, apart from 2004.


Introduction
Weather changes during a year influence the natural environment, including the pollen season dynamics of dif ferent taxa, especially the start of the season and the pollen season intensity.Particular attention is paid to grass pollen seasons, because Poaceae are the main components of meadow vegetation, grassland and steppe and are the source of the strongest pollen allergens in West and Central Europe [1].In Poland about 300 species of Poaceae occur, belonging to the most frequent genera: Festuca (36 species), Bromus (26 species) and Poa (20 species) [2][3].Because the great number of grass species flower in different time periods, grass pollen is observed in the air during the whole vegeta tion season.Grasses release pollen from May to October, and according to Szczepanek (1994) their pollen seasons are determined as "long seasons" [4].
Poaceae pollen allergens are the main cause of the allergic rhinitis (AR) provoked by plant pollen.In Poland, up to 90% of patients with AR are sensitive to grass pollen [5].The allergenic clinical symptoms depend on the allergens occurrence in the air, and the symptoms intensity correlate with the concentration of allergenic particles [6,7].
In the majority of regions in Poland the grass pollen concentration increases in late May and early June, achiev ing the highest concentrations in mid-July, while the last pollen grains are observed at the end of September and at the beginning of October [8].In temperate climate zone, characteristic for Poland, plants show a wind-pollination cycle associated with the seasons of the year, but climate changes have modified the start and duration of the seasons in recent years.Gehrig [9] reported, that the exceptionally warm summer in 2003 in Basel (Switzerland) influenced the grass pollen season timing strongly.Air temperature in June-August was 5-times higher than the long-term mean value and rainfall during June-October achieved only about 60% of the common sum.The indicated meteorological conditions had an evident impact on the earlier grass pol len occurrence (1-2 weeks earlier than mean value in the multi-year period), while the end of the season occurred 7-33 days earlier than mean value.
Piotrowicz and Myszkowska [10] indicated cloudiness, sunshine, precipitation and wind direction as factors sig nificantly influencing the grass pollen concentration over the pollen season.In the sunny days, the anthers start to dehisce and pollen is released.Piotrowska [11] reported the statistically significant relationship between the grass pollen season start and temperature (min., max., mean) and also the number of days with minimum temperature above 5.6°C.The impact of weather conditions on the pollen season dynamics is considered in the regional and wider scale.A multi-center studies on Poaceae pollen seasons in a country scale were performed in the Netherlands [12], Spain [13] and Great Britain [14].Sanchez-Mesa et al. [15] published the comparative results obtained in Spain and Great Britain.It must be also mentioned about the comparative study on the grass pollen season dynamics in 12 European aerobiological centres of a great geoclimatic diversity [16,17].Smith et al. [18] indicated the influence of NAO (North Atlantic Oscil lation) on Poaceae pollen seasons in Europe.
The aim of the study was to present the long-term vari ability of the start, end and duration of grass pollen seasons and the seasonal pollen index (SPI) in eight Polish cities against a background of the meteorological conditions over pollen seasons.

Aeropalynological data
The studies were performed in eight cities in Poland in 1996-2014 (Tab.1).The volumetric method was employed in pollen monitoring using a Hirst type trap (Lanzoni s.r.l., Burkard Ltd.).The traps work continuously taking samples of airborne particles with a flow rate of 10 l min-1.The particles were placed on the Melinex adhesive tape wrapped around the moving drum.Pollen grains were stained with Basic Fuchsine or Safranin, with no impact on the counting quality.Every week Melinex tape was changed and cut into seven segments corresponding to 24 h periods.Microscopic slides were prepared according to the instruction by Stach and Kasprzyk [19].For pollen grain identification the atlas of Moore et al. [20] as reference book was used.Twelve vertical lines or four horizontal lines on each slide were analyzed.Results were expressed as the number of pollen grains (PG) per cubic meter of the air over 24 hours (PG/m3).In the majority of cities the traps were installed at heights recom mended by the EAS (European Aerobiology Society, Quality Control Working Group; 15-25 m above ground level) [21].
Pollen seasons were calculated using the 95% method, where the start of the season was defined as the date when 2.5% of the seasonal pollen total was trapped and the end of the season as the date, when 97.5% of the seasonal pollen total was reached [22].The basic statistics, like arithmetic mean, minimum, maximum, standard deviation, coefficient of variation were applied in the descriptive statistics of the season characteristics (season start, season end, season duration, SPI value, maximum concentration and the day, when it was achieved).The SPI value is described as a sea sonal total concentration in the determined season.Due to the fact that the different data series were used, the mean values of the season start, end and duration were calculated for the common period 2003-2012 to compare these season characteristic in all studied stations.

Meteorological data
Characteristics of meteorological conditions were performed using mean daily and monthly temperature, precipitation in the studied cities, obtained in the local meteorological stations.In case of Sosnowiec, the meteoro logical data came from meteorological station in Katowice.The other sources used to complete data were: Monthly Climate Monitoring Bulletin developed by the Institute of Meteorology and Water Management -National Research Institute (IMGW-PIB; http://www.imgw.pl).
The Spearman rank correlation test was used to find the relation between the pollen season characteristics and the meteorological elements (Statistica, Version 10).

Pollen seasons start
In Poland, grass pollen seasons started usually in May.In the southern part of the country (Wrocław, Sosnowiec, Kraków, Rzeszów) grass pollen seasons began about the middle of May, while in the stations located in central and northwestern regions the pollen seasons started between 20th to 24th of May (Fig. 1).However, the latest pollen season start was detected in Lublin, where the season start was relatively stable (according to the SD value 5.71 days) (Tab.2).In comparison, the most variable grass pollen season start was found in Kraków, probably because of the longest data series (23 years; Tab. 2).Analysis of the impact of weather conditions on the grass pollen season start showed that this season charac teristic coincidence with the course of monthly average air temperature in April and May (Tab.3).This phenomenon was the most evident in 2010, when grass pollen seasons started in the south-eastern Poland within the long-term average (mid-May), but in Wrocław, Poznań and Szczecin the grass pollen seasons were one of the latest (Tab.2, Fig. 2).In these regions (western Poland) mean monthly temperature in May 2010 was relatively low.According to the classification of IMGW-PIB May 2010 was described as "cool" (Fig. 3).Isolines of average air temperature anomalies coincided with the dates of the grass pollen season starts in all analyzed stations.

Pollen season ends
Pollen season ends in the studied time were more variable then season starts, both in the temporal and spatial scale.The seasons have finished the earliest in Rzeszów, the lat est in Szczecin (Fig. 1, Tab. 2).In the southwestern part of Poland, the grass pollen season end vary in time the most evidently (Wrocław), in opposite to the northwestern Poland (Szczecin; V% = 7.13 vs. V% = 2.87; Tab. 2).

Pollen season duration
Grass pollen seasons lasted the shortest in the eastern part of Poland (82-85 days on average), and have been longer in the south and central part of the country (86-94 days), the longest seasons were observed in the northwestern regions and in Sosnowiec (105-109 days; Fig. 1).
The duration of pollen seasons was associated with the data series (Tab.2).However, analysis of the extreme values of the seasons length leads to the conclusion that extremely short seasons (less than 70 days) occurred mainly in the south, especially in the south-western part of the country.The longest pollen season was determined in Szczecin (Tab.2).

Seasonal pollen index (SPI)
The analyses of the seasonal pollen index indicated that the least intensive pollen seasons occurred in Kraków and Wrocław (2500 PG/m3 on average; Tab. 2, Fig. 4), which could have resulted from the low SPI values in Kraków in the 90s of the 20th century.The highest pollen load was found in Lublin and Szczecin in all studied seasons (Tab.2, Fig. 4).
Analyses of monthly temperature during the season indicated the significant correlation between SPI value in June and temperature in this month (Tab.3).The most evident relationship was found in Łódź (r = 0.713; P < 0.05) and Rzeszów (r = 0.689; P < 0.05), while the weakest, not statistically significant relationship was obtained in Poznań (r = 0.13; P > 0.05).In case of Szczecin, temperature in July proved to be the best correlated with SPI value (r = 0.553; P < 0.05), however, the highest negative correlation was found between the SPI and temperature in July-August value in Łódź (r = -0.519).Considering the sum of precipitation in May-August and June-August, the positive correlation between precipitation and SPI value was found in the most stations (Kraków, Łódź, Poznań, Lublin, Sosnowiec; r > 0.40; P < 0.05).In general, the highest pollen load occurred in all analyzed stations in warm and wet months, what was clearly seen in July 2010, which was classified as "extremely warm" with the higher than the average (Fig. 5).

Maximum pollen concentrations
Maximum pollen concentrations were achieved between 150 and 190 day of the year (between the end of May and the Tab. 2 Grass pollen season characteristics in the studied stations calculated using the 95% method.Tab. 3 The Spearman correlation coefficients between selected parameters and meteorological elements.Statistically significant relationships are given in bold (P = 0.05).Fig. 2 The term of the grass pollen seasons in the selected stations in the studied years.first decade of July; Fig. 6).The days in which the maximum concentration was detected differed between stations in all studied seasons and had no evident relationship with meteorological conditions, apart from 2004.In this year, in most of the stations, the maximum pollen count was observed on the 7th and 8th of July (only in Wrocław on the 28th of June).In the indicated days, the weather was shaped by the high pressure system in Poland.It was sunny (>8 sun hours/day) throughout the country, with little cloud cover.
Maximum temperature of the air on the 7th July 2004 fluctu ated from 18°C in Szczecin (one day later 24°C) to 22°C in Wrocław (one day later 27°C).The humidity at noon was low (40-50%), and the wind was weak and blow from different directions.In other years the maximum concentration was detected in days of different weather types, what made the problem to find the clear relationship between this season characteristic and meteorological elements.

Discussion
Studies on the occurrence and dynamics of pollen seasons often focus on the influence of different meteorological factors on the pollen season characteristics and the pollen concentration [23].It is crucial information for people sensitive to pollen from view of diagnostics, therapy and prophylaxis of allergic rhinitis [24][25][26][27].

Pollen season timing
In comparison to the other taxa, Poaceae pollen seasons show the stability of the appearance dates [28,29].It refers especially to the start and end of the pollen season [30,31].In the current paper, the values of the coefficient of variation (V%) did not achieve 10% in case of season start and 20% in case of season end.Grass pollen seasons in Poland are definitely more stable then tree pollen seasons, the most evident seen in case of early-pollinating tress (like alder and hazel), but the tree pollen seasons are more variable at the season start, while in case of grasses, the end of the season vary clearly from year to year [8].
It is known that the timing and magnitude of grass pol len seasons differ between regions due to factors that affect abundance and dispersal [4,16].The differences in season time in Poland are as follows: mean season start differs up to 9 days among the stations, season end differs up to one month, while season duration differs up to 20 days.It was noted on the basis of the study performed in 12 European cities, in 1985-1997, that grass pollen seasons started in the south (Spain) the earliest due to the warmer weather condi tions in the spring and early summer, followed by the West regions and finally Northern Europe (Finland), where the latest pollen season were observed [16].It was concluded that the variability of start dates tends increase with latitude.
On the other hand, Smith et al. [18] reported that latitude is the main factor describing the spatial variation in the tim ing of the grass pollen seasons at the 13 sites in Europe.This observation is related to the accumulation of temperature from the start of the growing season.This study shown also the influence of the large-scale patterns on the grass pollen seasons, including NAO, through its effect on temperature and precipitation.However, the impact of NAO is well seen in Western Europe, in the locations along the coast only.In our study the impact of longitude on the pollen season timing was rather found and the evident relationship with temperature in the months preceding and during the pollen season was revealed.
Low variability of the grass pollen season start makes this season characteristic of a great predictive value.Myszkowska [31] proposed the factor predictive models arranged using as independent variables: cumulative mean daily temperature, effective cumulative temperature ranging from 5 to 15°C and the sum of daily temperature amplitudes [S(Tmax -Tmin)].The application of factor regression analysis allowed to indicate some phenomena of the plants property: when cumulative mean daily temperature increased by 10°C, the season start was accelerated by 1 day.On the other hand, when the sum of daily temperature amplitudes increased by 10°C, the season started 3 days earlier.

Pollen season intensity
It was noted that the highest grass pollen concentrations were observed in summer, in the conditions of high tem perature and precipitation.This observation is important not only from botanical, but also from medical point of view.Most of the patients manifest the symptoms of allergic rhinitis provoked by the pollen allergens in summer, the time of the high grass pollen exposure [1, [5][6][7].The high pollen load is also responsible for the increase of allergic asthma symptoms in patients sensitive to grass pollen.Erbas et al. [32] indicated the pollen concentration of 30 PG/m3 as the threshold value at which the clinical symptoms are observed.The information of the year-to-year fluctuation of seasonal pollen intensity makes the estimation of efficacy of pollen allergy treatment more objective [33].
In the current study the pollen annual sum (presented as SPI value) differed clearly from year to year, what was not stressed in the long-scale research performed in Europe by Spieksma et al. [17].In Poland the weather conditions in summer are relatively stable.In spite of that the grass pollen severity symptoms varies strongly within seasons [34].It is worth to underline, that apart from temperature, precipita tion in summer influenced the pollen count.The relationship between pollen concentration and precipitation before the pollen season was observed in the Mediterranean regions, confirming that the development of grasses is positively

Fig. 1
Fig. 1 Mean grass pollen season start and end dates and season duration in the selected station in 2003-2012.

♦Fig. 6
Fig. 6 Dates when the maximum grass pollen concentration was achieved in the studied stations in 2003-2013.