The influence of extreme high and low temperatures and precipitation totals on pollen seasons of ...

13 Pages·2012·1.14 MB·English

Grana,2012;51:215–227
The influence of extreme high and low temperatures and precipitation
totals on pollen seasons of Ambrosia, Poaceae and Populus in Szeged,
southern Hungary
LÁSZLÓ MAKRA1, ISTVÁN MATYASOVSZKY2, ANNA PÁLDY3 & ÁRON JÓZSEF DEÁK4
2
01 1DepartmentofClimatologyandLandscapeEcology,UniversityofSzeged,Hungary,2DepartmentofMeteorology,Eötvös
r 2 LorándUniversity,Budapest,Hungary,3FodorJózsefNationalInstituteofEnvironmentalHealth,Budapest,Hungary,
e
b 4DepartmentofPhysicalGeographyandGeoinformatics,UniversityofSzeged,Hungary
o
ct
O
9
1
1
4 Abstract
0: Extreme high and low temperatures and precipitation totals may have important effect on daily and annual pollen
0
at concentrations.Theaimofthisstudyistoanalysetheassociationsbetweenpollencharacteristicsandmeteorologicalvari-
] ables, furthermore between the rank of pollen characteristics and the rank of annual values of meteorological variables
a
kr for Szeged, southern Hungary. Pollen characteristics include pollen count parameters (TPA, total annual pollen amount;
Ma APC, annual peak pollen concentration) and pollen season parameters (start, end and duration of the pollen season).
ó Meteorological variables are temperature and precipitation. The data set used covers a 14-year period (1997–2010) and
zl containsdailyvaluesofAmbrosia(ragweed),Poaceae(grasses)andPopulus(poplar)pollenconcentrations,aswellasthose
s
á oftemperatureandprecipitation.BothPearsonandSpearmanrankcorrelationswerecalculated,becausetherankcorrela-
L
[ tionislesssensitivethanthecorrelationtooutliersthatareinthetailsofthesample.OurresultssuggestthatAmbrosiaand
d], Populusarereverselyrelatedtotemperature(negativecorrelations),whilePoaceaeexhibitaparallelrelationshipwithprecip-
e
g itation(positivecorrelations).Onthewhole,pollencountcharacteristics(TPAandAPC)indicateadecreaseforAmbrosia
e
z andPoaceae,whileforPopulusanincreaseisexpected.
S
f
y o Keywords: Ambrosia,Poaceae,Populus,pollencounts,pollenseason
sit
r
e
v
ni
U Weather related daily variability of pollen concen- over Europe in summer 2003 substantially influ-
[
y trations has a wide literature, among them studies enced pollen phenology and pollen production in
b
d of therelationship of meteorological parameters and Switzerland with its mean temperature exceeding
e
ad dailypollenconcentrations(e.g.Bartková-Šcˇevková, the 1961–1990 mean by about 5 ◦C in June, July
o
nl 2003; Rodríguez-Rajo et al., 2005; Štefanicˇ et al., and August (Gehrig, 2006). The grass pollen sea-
w
o 2005; Kasprzyk, 2008; Recio et al., 2010), while sonwasmostaffectedstarting1–2weeksearlierand
D
others, based on meteorological data, use different ending 7–33 days earlier than in general. Extreme
techniques for predicting pollen characteristics (e.g. highChenopodium,PlantagoandPoaceaedailypollen
Galánetal.,2001;Aznarteetal.,2007;García-Mozo concentrations were measured in that pollen season.
etal.,2009). Carinˇanos et al. (2000) analysed the yearly distribu-
Theroleofextremeweathereventsregardingdaily tion and severity of Artemisia and Chenopodiaceae-
pollen concentrations has received low attention so Amaranthaceae pollen indicating the highest and
far. Frei (2004, 2006) studied the occurrences of particularly high pollen levels in a rural area with
extreme events (storms, floods or droughts) with sub-desert climate and extreme dryness. Hart et al.
extremebirchandgrasspollenconcentrationsinthe (2007) studied the effect of the six warmest months
data set from Basel, Switzerland. The heat wave onthepollenconcentrationsinSydney,Australia.
Correspondence:LászlóMakra,DepartmentofClimatologyandLandscapeEcology,UniversityofSzeged,POBox653,HU-6701Szeged,Hungary. E-mail:
[email protected]
(Received4July2011;accepted18January2012)
ISSN0017-3134print/ISSN1651-2049online©2012CollegiumPalynologicumScandinavicum
http://dx.doi.org/10.1080/00173134.2012.661764
216 L.Makraetal.
As a result of the recent climate change, vari-
ability of temperature and precipitation increases
(Tank&Konnen,2003)andextremeweatherevents
(e.g. cold or hot days, as well as droughts or rainy
periods) can be persistent and can keep up to sev-
eral weeks and even they can be repeated several
times. Global warming may facilitate the extension
of certain herbaceous and arboreal plant habitats by
contributingtotheincreaseofpollenlevelsandtothe
exacerbationoftheiradverseeffects,hencetotherise
of pollen sensitivity and respiratory admissions due
toapollenallergy(d’Amato&Cecchi,2008;Ariano
etal.,2010;Ziskaetal.,2011).Thus,theanalysisof
theeffects oflong-lasting extremeweather eventson
2 the daily or annual pollen concentrations is of ever
1
0 increasingimportance.
2
er Our primary aim is to analyse the relationship
b
o between pollen characteristics of Ambrosia, Poaceae
ct
O and Populus with meteorological variables, and, fur-
9 thermore, between the rank of ordered pollen char-
1
1 acteristics of these taxa and the rank of ordered
4
0: annualvaluesofmeteorologicalvariables,forSzeged
0
at in southern Hungary. On the basis of our results, a
a] potential change in the pollen count characteristics
r
k
a of Ambrosia, Poaceae and Populus is concluded due
M
toglobalwarming.
ó
zl Pollen concentration database of 24 taxa is avail- Figure1. MapsofEuropeandHungaryshowingthestudylocal-
s
Lá able in Szeged. Taxa with the highest pollen levels ityinthecentreofSzegedwiththepositionsofthedatasources:
], [ include Ambrosia (32.3%), Poaceae (10.5%) and 1–meteorologicalmonitoringstation;2–aerobiologicalstation.
d
e Populus(9.6%),whichtogetheraccountfor52.4%of the winter (December, January, February), respec-
g
ze the total pollen production in the city (Makra et al., tively. Mean daily summer temperatures are around
S
f 2011); therefore, these three earlier mentioned taxa 22.4 ◦C, while mean daily winter temperatures fluc-
o
y wereselectedforthisstudy. tuate around 2.3 ◦C. The irradiance exhibits an
rsit average of 656.1 and 133.1 MJ/m2 in summer and
e
v winter days, respectively. The total annual precip-
ni
U Materialandmethods itation amounts may fluctuate quite substantially.
[
y In this period, its extreme values for Szeged were
b Studylocalityanditsclimate
d minimumprecipitation(P )=203mm(2000)and
e min
ad Szeged(46◦ 15(cid:3) N,20◦ 10(cid:3) E),thelargestsettlement maximum precipitation (Pmax)=838 mm (2010).
o
nl in southern Hungary, is located at the confluence of The most frequent winds blow along the north-
w
the rivers Tisza and Maros (Figure 1). The area is northwest (NNW)–south-southeast (SSE) axis with
o
D
characterised by an extensive flat landscape of the prevailing air currents arriving from NNW (42.3%)
Great Hungarian Plain with an elevation of 79 m and south-southwest (SSW) (24%) in the summer
above sea level. The city is the centre of the Szeged and from SSE (32.6%) and NNW (30.8%) during
regionwith203000inhabitants. the winter. Due to its unique geographical posi-
The climate of Szeged belongs to Köppen’s tion, Szeged is characterised by relatively low wind
Ca type (warm temperate climate) with relatively speeds with average daily summer and winter values
mild and short winters and hot summers (Köppen, of 2.8 and 3.5 m/s, respectively. The highest hourly
1931). According to the climate classification of wind speeds were recorded in spring with a rate of
Trewartha(1968),SzegedisassociatedwithclassD1 5 m/s (Hungarian Meteorological Service, personal
(continental climate with a long warm season). The communication,2011).
climate of Szeged is characterised by hot summers
andmoderatelycoldwinters.Forthe30-yearperiod
Pollensampling
1981–2010, the distribution of rainfall is fairly uni-
form during the year with a contribution of 29% In Szeged, the pollen content in the air has been
and 19% for the summer (June, July, August) and measured since 1989 using a seven-day Hirst-type
MeteorologicalinfluenceonpolleninHungary 217
Figure2. Dailypollenconcen-
2
1 trations of Ambrosia, Poaceae
0
2 and Populus in Szeged, Hun-
er gary, in the years 1997–2010
b
o (black curves – mean daily
ct pollen counts 1997–2010
O
9 except for 2003; grey curves –
1 dailypollencounts2003).
1
4
0:
0
at
] volumetric pollen trap Lanzoni VPS 2000 (Hirst, Conservancy Inspectorate of Lower-Tisza Region,
a
kr 1952).Theairsamplerislocatedontopofthebuild- Szeged) located in the downtown area of Szeged,
a
M ingoftheFacultyofArts,UniversityofSzeged(20m about10mfromthebusiestmainroad(Figure1).
ó above ground) in the downtown area (Figure 1; In order to determine the relationship between
zl
ás Makra et al., 2005). Pollen sampling was performed extreme high and low temperatures and precipita-
L
[ as follows: A specific tape was made adhesive by tion totals on the one hand and pollen counts of
d], washingitwithsiliconeoil.Thesamplerabsorbedair the three selected taxa on the other hand, the daily
e
eg atarateof10l/min(=14.4m3/day,whichiscorre- values of mean temperature, precipitation total and
z
S spondingtothedailyrequirementofanadultperson) daily pollen concentrations of Ambrosia (ragweed),
f
o and was supplied with a timer, to which a rotating Poaceae (grasses) and Populus (poplar) were con-
y
sit drumwasfitted.Thedrummovedtheadhesivetape sidered. The selection of these taxa is justified by
er (2mm/h)wherepollengrainsadhered.Afteraweek their high or medium allergenicity [in a four-score
v
ni ofexposure,thetapewasremovedandcuttoalength scale to be find at Hungarian pollen index (www.
U
[ correspondingto24hpollensampling,coveredwith pollenindex.hu), allergenicity of both Ambrosia and
y
b agelmountingagentcontainingfuxinasastainand Poaceae is the highest indicated by score four, while
d
e put on a microscope slide. Afterwards, the samples that of Populus is medium indicated by score two]
d
oa wereexaminedunderalightmicroscopeatamagnifi- and by their more or less permanently high pollen
nl cationof400×todeterminepollentypesandcounts. concentrations(Figure2).
w
o Five horizontal sweeps were analysed on each slide.
D
Horizontal sweeps were used because the variation
in the concentration during the day can be observed Consideredtaxaandaccompanyingflora
along this axis (the direction of the tape shifts in
Regarding the taxa to the highest pollen con-
thesampler). Theaccuracy ofthemeasurement was
centrations, Ambrosia is presented by only one
proportional to the number of sweeps and the con-
species, namely Ambrosia artemisiifolia L. (Common
centration of particles. Counting was done using a
Ragweed)intheSzegedregionappearingbothinthe
standard sampling procedure. Pollen concentrations
were expressed as number of pollen grains/m3 of air urbanenvironmentandinthecountryside.Ragweed
especially frequently occurs in the western part of
(Käpylä&Penttinen,1981;Peterneletal.,2006a).
the city. The dominant north-western winds can
easily transport pollen into the city. Ambrosia can
Meteorologicaldata
spread unchecked in the sandy region (northwest of
Meteorological data were collected at the mete- Szeged), because stubble stripping is not necessary
orological monitoring station (operating by the for ground-clearance due to the mechanical proper-
Environmental and Natural Protection and Water ties of sandy soils. Owing to newly-built motorways
218 L.Makraetal.
Figure3. Dailypollenconcen-
trations of Ambrosia in years
with extreme temperature
2
1 (warmest year: 2003, coldest
0
2 year: 2001); examination
er period of the years: days
b
o 132–280.
ct
O
9
1
1 aroundSzeged,severalfarmlandareashavebeenleft Trin.(CommonReed)prevailingalllandscapetypes
4
0: untouched for a long time, which also favours the around Szeged. Furthermore cereals, i.e. wheat
0
at expansion of Ambrosia. Several species of Poaceae (Triticum aestivum L.), maize (Zea mays L.) and bar-
a] family are common in and around Szeged: namely ley (Hordeum vulgare L.) also belong to Poaceae that
r
ak Agropyron repens (L.) P.Beauv. (Common Couch), cover a huge proportion of the landscape surround-
M
Poa trivialis L. (Rough Meadow-grass), Cynodon ing Szeged as arable lands, thus being the most
ó
zl dactylon (L.) Pers. (Bermuda grass), Bromus sp. commonhabitatsofalllandscapetypesinthisregion.
s
Lá (Brome species) as well as Poa bulbosa L. (Bulbous For Populus, natural species of Populus alba
[
], Meadow-grass) and Poa angustifolia L. (Narrow- L.(WhitePoplar)andP.canescensSm.(GreyPoplar),
d
e leavedMeadow-grass).Theyaretypicalinweedydry as well as cultivated poplars such as I-273 Poplar
g
e grasslands, while Echinochloa crus-galli (L.) P.Beauv. and P. canadensis Castigl. (P. deltoides × P. nigra;
z
S
f (Cockspur) is common in wet weedy grasslands Canadian Poplar) and its variants, are the most fre-
o
y of inland-water covered depressions of fallows con- quentinthecity.Theyalsodominatethenaturaland
sit vertedintoarablelands.Severalgrasslandsremained plantedforestsoffloodplainsandarecommoninthe
r
e
v aroundSzegedbothinsandandloessandfloodplain sandland as well, because they were planted inten-
Uni landscapes – now mainly connected to the depres- sively both in active floodplains and in sandlands
y [ sions – that are dominated by the representatives of duringthelastdecades.
b
d Poaceae. Alopecurus pratensis L. (Meadow Foxtail) is Theanalysiswasperformedforthe14-yearperiod
e
d the main grass of floodplain meadows, which cov- 1997–2010.Thepollenseason isdefined byitsstart
a
o ers also the dykes in Szeged. Fescue species (mainly and end dates. For the start (end) of the season,
nl
w Festuca pseudovina Hack. ex Wiesb. and F. rupicola we used the first (last) date on which one pollen
o
D Heuff.) prevailing sand, loess and short-grass alkali grain/m3 air is recorded and at least five consecu-
steppe grassland types, whereas Molinia hungarica tive (preceding) days also show one or more pollen
Milk.(HungarianPurpleMoor-grass)formstherare grains/m3 air (Galán et al., 2001). Evidently, the
Molinia fens in those depressions of Kiskunságian pollen season for all three pollen types varies from
Sandland where upwelling-points of groundwater yeartoyear.
are found. Dactylis glomerata L. (Cocksfoot) is
common in sand steppe grasslands and regenerat-
Statisticaltreatmentofrecordeddata
ing sandy fallows. Agrostis stolonifera L. (Creeping
Bent) is the most common grass of the sandland- The periods examined for the three taxa are indi-
type saline meadows, whereas Puccinellia limosa catedbythedaysoftheyear:Ambrosiadays132–280,
Homlb. (Common Saltmarsh Grass) forms the Poaceae days 62–282 and Populus days 35–113. The
Puccinellia swards and grasslands with surface salt- end of these periods was selected according to the
accumulation,appearingmainlyintheKiskunságian average end date of the annually varying pollen
Sandlands.Swamps,artificiallakes,oxbowlakesand seasons. Choosing the starting days (132, 62 and
channels are surrounded by Phragmites communis 35, respectively) is slightly more complicated. When
MeteorologicalinfluenceonpolleninHungary 219
Figure4. Dailypollenconcen-
trations of Ambrosia in years
2
1 with extreme precipitation
0
2 (wettestyear:2001,driestyear:
er 2000); Examination period of
ob theyears:days132–280.
ct
O
9
1
41 analysingpollencountcharacteristics,itisafrequent were ordered from the highest to the lowest cumu-
0: task to predict the starting date of the pollen sea- latedvalues,andtheseranksofyearswererelatedto
0
at son using meteorological information. A generally theranksoforderedannualpollencharacteristicsvia
a] used technique includes the method of accumu- correlations(Spearmanrankcorrelation).
r
k
a lated temperatures when daily mean temperatures As the number of data for calculating correla-
M
ó are cumulated from a starting date. Specifically, if tions is very small (14 or just six in some cases),
zl the actual cumulated temperature amount exceeds the commonly used approximation to the proba-
s
Lá a given threshold on a day the start of the pollen bility distributions of estimated correlations with
[
], season is identified with that day. The threshold t-distributions is not advisable. Therefore, the inter-
d
e and the starting day of temperature accumulation val for accepting the null-hypothesis of zero cor-
g
ze are estimated from an available data set as to min- relation was determined by a permutation test as
S
f imise the mean squared error of estimated pollen follows. The N number values of meteorological
o
y season starts (Laaidi et al., 2003). Our starting days variables (cumulated temperature, cumulated pre-
sit mentioned earlier were taken equal to starting days cipitation, ranks) corresponding to the N years were
r
e
v obtained from the procedure of cumulated daily reordered while keeping the pollen characteristics.
ni
U meantemperatures. Correlation between these two new data sets was
[
y Daily mean temperatures and daily precipitation then calculated. Reordering and computation of
b
d amounts were cumulated over the earlier mentioned the correlation were done in every possible case
e
d periodsforeveryyearseparately,regardingthethree except for the original non-reordered case. Having
a
nlo taxa. These quantities were then related to annual (N!–1) number of correlations, the interval to be
w pollen characteristics via correlations. Additionally, found for a significance level ε·100% is (q , q ),
o ε/2 1–ε/2
D the annual course of both daily mean tempera- whereq istheε-quantileofthesecorrelations.
ε
ture and daily precipitation amount was described
by fitting sine and cosine waves of one year and
half year periods to the entire 14-year data set.
Results
The half year period was used to reproduce the
asymmetry of the annual cycle. The mean squared The pollen season for Ambrosia artemisiifolia in
deviation (MSD) between the actual daily mean Hungary begins between 13 July and 4 August and
temperature/precipitation amounts and the annual lasts for almost three months. The maximum pollen
course was calculated for each year separately and release of flowers appears 1–1.5 weeks after their
fortheperiodscorrespondingtothethreetaxa.Years blossom (Szigetvári & Benko˝, 2004). Pollination
withthehighestannualMSDvalueswerethentaken of Poaceae is species-dependent. The pollination
as extreme years. An extreme year is warm/cold peak of Poaceae species represented in the Szeged
(wet/dry) if its cumulated temperature (precipita- region begins in late spring (especially in May)
tion) is above/below the 14-year mean cumulated and can last for almost the whole summer (mainly
temperature (precipitation). Years (extreme years) June–July), but a time-shift can be observed in the
220 L.Makraetal.
Figure5. Dailypollenconcen-
trations of Poaceae in years
2 with extreme temperature
1 (warmest year: 2009, coldest
0
2 year: 1997); examination
er period of the years: days
b
o 62–282.
ct
O
9
1
41 pollination of the different species resulting in a Echinochloa crus-gallii begins at the end of July and
0: pollen season that can continue until early autumn lasts even until October. Phragmites communis has
0
at even until September and October. In this way, the latest pollen season beginning in August and
a] different species are responsible for the measured finishing in October. The pollination of all poplar
r
k
a pollen concentrations during the vegetation period. species is short term and concentrates on the spring
M
ó Alopcurus pratensis – being dominant in the flood- period between March and April without any plant-
zl plainmeadows,dykesandCrisicumtypeofsecondary specific exceptions (Figure 2). Correlations between
s
Lá salinemeadowsofsaved-sideformerinundatedareas the pollen characteristics and the cumulated daily
[
], – and Poa bulbosa (species of disturbed grasslands) values of meteorological variables (temperature and
d
e begin the pollen season for Poaceae in April. The precipitation)aresummarisedinTableI.
g
ze pollen season of Alopecurus pratensis lasts until July,
S
of whereas that of Poa bulbosa keeps up just until June. Table I. Correlations between daily pollen characteristics and
y Formostofthegrassspecies,pollenseasonbeginsin cumulateddailyvaluesofmeteorologicalvariables.
sit May, but it can last for different periods. The short-
r
e Temperature Precipitation
v est pollen period can be detected for the main dry
Uni grassland species of Festuca spp. and the semi-wet, Pollen
[ characteristics, Six
by somewhat disturbed condition favouring Poa angus- taxa Everyyear Sixyearsa Everyyear yearsa
d tifolia that pollinates just for one month until June.
de Albeit the disturbed grassland favouring Bromus TPA,Ambrosia −0.39 −0.88(1.8%)∗ 0.24 0.28
nloa species, namely, Poa trivialis and some main cereals ASPPSC,,AAmmbbrroossiaia −00..2501(6%)∗ −00..9079(0.2%)∗ 00..3187 00..6216
w (wheat and barley) pollinate until July, the grass of EPS,Ambrosia −0.27 −0.07 −0.07 −0.02
o
D the disturbed grassland (i.e. Agropyron repens) pol- DPS,Ambrosia −0.35 −0.20 −0.15 −0.20
linates until September, whereas Dactylis glomerata TPA,Poaceae −0.2 0.24 0.24 0.71
APC,Poaceae −0.15 0.10 0.48(8%)∗ 0.68
(grass of sand steppe grasslands and regenerating SPS,Poaceae −0.41 −0.70 −0.19 −0.55
fallows) even until October. The pollen season for EPS,Poaceae 0.20 0.51 −0.18 0.00
the grasses of the sandland-type saline grasslands DPS,Poaceae 0.30 0.68 0.01 0.52
TPA,Populus 0.22 0.11 0.16 0.42
begins in early summer, in June and for the most
APC,Populus 0.12 −0.16 0.14 0.31
halophytic Puccinellia limosa, it lasts just for one SPS,Populus −0.59(3%)∗ −0.72(10%)∗ −0.07 −0.21
month, until July, whereas for Agrostis stolonifera EPS,Populus −0.36 −0.77(8%)∗ −0.15 −0.10
it lasts until August. Maize starts its one-month DPS,Populus 0.46(10%)∗ 0.56 0.00 0.33
pollination in July. As the continental precipitation Note:TPA–totalpollenamountduringthepollenseason;APC–
peak in June is essential for the growth of the main annualpeakconcentration;SPS–startofthepollenseason;EPS
grass of Molinia fens (the wettest moor-type vege- –endofthepollenseason;DPS–durationofthepollenseason.
aThe three warmest/coldest and the three wettest/driest years,
tation in the surroundings of Szeged) in this way,
respectively.
the pollen season of Molinia hungarica starts in July
∗Significancelevelsforcorrelationsbeingnon-zeroareshownfor
and can last until September. The pollination of levelsnohigherthan10%(inparentheses).
MeteorologicalinfluenceonpolleninHungary 221
Figure6. Dailypollenconcen-
trations of Poaceae in years
2
1 with extreme precipitation
0
2 (wettestyear:2001,driestyear:
er 2000); examination period of
ob theyears:days62–282.
ct
O
9
1
1
4 For Ambrosia, only temperature related associa- of data pairs. Taking the three warmest and cold-
0:
0 tions are important. In a first approach, correlations est (wettest and driest) years from the data sets,
at were determined for every year, and then computed the correlation based on just the six most extreme
]
ra for the three warmest and coldest as well as for yearsistailoredtoextremes.Whenconsideringevery
k
Ma the three wettest and driest years (extreme years), year,thecorrelationbetweentheannualpeakpollen
ó respectively(TableI).Thisisdonebecausecorrelat- concentration (APC) and temperature is inversely
szl ingthe14datapairstheroleofwarmest/coldestand proportional. For the extreme years, both the total
á
L wettest/driest years can be lost as correlation mea- annual pollen amount (TPA) and the APC are
[
], sures the overall relationship between the entire set in substantial negative connection with temperature
d
ge (Figures3,4,TableI).
e
z For Poaceae, considering every year, the APC
S TableII. Correlationsbetweentherankofdailypollencharacter-
of istics and the rank of years based on cumulated daily values of is positively correlated with precipitation. However,
y temperatureandprecipitation. thereisnosignificantassociationbetweenthispollen
rsit variableandtemperature(Figures5,6,TableI).
e Temperature Precipitation
v
ni Pollen For Populus, precipitation based associations are
U irrelevant, but for the extreme years, the end of
[ characteristics, Six
by taxa Everyyear Sixyearsa Everyyear yearsa the pollen season (EPS) is in significant negative
d correlationwithtemperature(Figures7,8,TableI).
de TPA,Ambrosia −0.39 −0.81(5%)∗ 0.15 0.14
oa APC,Ambrosia −0.49(8%)∗ −0.90(1%)∗ 0.43 0.67 Correlations between the rank of pollen char-
nl SPS,Ambrosia 0.16 0.05 0.39 0.56 acteristics and the rank of years based on cumu-
ow EPS,Ambrosia −0.25 −0.22 0.00 0.06 lated temperature and precipitation were calculated
D DPS,Ambrosia −0.28 −0.32 −0.23 −0.42
for all three taxa since the rank correlation is less
TPA,Poaceae −0.21 0.29 0.37 0.67
APC,Poaceae −0.17 0.21 0.45(10%)∗ 0.57 sensitive than the correlation to outliers that are
SPS,Poaceae −0.40 −0.62 −0.14 −0.39 in the tails of the sample. This is important when
EPS,Poaceae 0.28 0.54 −0.12 −0.06 analysingextremeyearsastheycorrespondtothetail
DPS,Poaceae 0.43 0.70 0.01 0.29
areas of the sample. Specifically, only five significant
TPA,Populus 0.20 −0.09 0.20 0.61
APC,Populus 0.13 −0.37 −0.04 0.39 correlationscanbeobserved(TableI),buteightrank
SPS,Populus −0.04 −0.71 −0.07 −0.26 correlations are significant (Table II). This might be
EPS,Populus −0.16 −0.81(5%)∗ −0.16 0.01 due to the smaller sensitivity of the rank correlation
DPS,Populus −0.08 0.56 −0.08 0.36
tooutliers,ortherelationshipispossiblynotlinear.
Note:TPA–totalpollenamountduringthepollenseason;APC– For Ambrosia, only temperature related correla-
annualpeakconcentration;SPS–startofthepollenseason;EPS tions are relevant. When considering every year, the
–endofthepollenseason;DPS–durationofthepollenseason.
rank of APC is inversely proportional to the rank
aThe three warmest/coldest and the three wettest/driest years,
of the annual temperature data. However, for the
respectively.
∗Significancelevelsforcorrelationsbeingnon-zeroareshownfor extreme years, the ranks of both TPA and APC are
levelsnohigherthan10%(inparentheses). negatively associated with the rank of the annual
222 L.Makraetal.
Figure7. Dailypollenconcen-
trations of Populus in years
2 with extreme temperature
1 (warmest year: 2007, coldest
0
2 year: 2003); examination
er period of the years: days
b
o 35–113.
ct
O
9
1
1
4 temperature data. In more detail, in the warmest which occurred in the warmest year (2003) as well.
00: year (2003), the TPA was the second smallest and So in warm years, Ambrosia concentrates on its sur-
at the APC was the third smallest. In the coldest year vivalratherthanonpropagation.Cooleryearsinvolve
]
a (2001),theTPAwasthesecondhighestandtheAPC high precipitation in early summer or even in late
r
k
a wasthethirdhighest(Figures3,4,TableII). spring (May–July). As an example, in 2010, the
M
ó For Poaceae, considering every year, the rank far above average May–July precipitation in Szeged
zl of the APC is proportional to the rank of years regionhelpedtheincreaseoftheAmbrosiapopulation
s
á
L based on precipitation. However, there is no even in natural open sand vegetation types forming
[
], significant association between the rank of any beltarounddepressions(Figures3,4).
d
e pollen characteristics and the rank of the extreme AwarmerspringcanhelpbothgrowthofPoaceae
g
ze years based on either temperature or precipitation andtheirearlierflowering.Ifsufficientwaterisavail-
S
f (Figures5,6,TableII). able, the late spring pollen peak can be high even
o
y For Populus, only temperature related substan- in semi-wet or dry conditions. This is the case
sit tial correlations have been detected. When regard- even if the early spring temperatures are higher
r
e
v ing every year, the ranks for both the start of the than the average. As low spring temperatures can
ni
U pollen season (SPS) and the duration of the pollen hinder the growth of these warm-tolerant continen-
[
y season (DPS) are associated with the rank of the tal grass species, they require proper temperatures.
b
d annual temperature data. Taking into account only If the spring is cool and the early summer precipita-
e
d the extreme years, the ranks of both the SPS and tion is high (accompanied with low temperatures),
a
o
nl the EPS are inversely proportional to the rank of the pollen grains can be washed out from the air
w the annual temperature data. In more detail, in the and pollen production can be reduced to its mini-
o
D warmestyear(2007),theDPSwasthethirdlongest, mum. These cool and wet conditions, however, can
while in the coldest year (2003), it was the fifth enhance the late summer flowering and favour grass
shortest.Inthecoldestyear(2003),theSPSwasthe species such as Molinia and Agrostis spp., as well as
second latest, while in the warmest year (2007), it maize,which,consequently,producesahighernum-
wastheearliest(Figures7,8,TableII). ber of pollen grains. The warm and cool years do
Interestingly,thepollenseasonofAmbrosiafallsto not influence the number of pollen grains substan-
thedriestperiodofsummerinHungary.Anextreme tially for Poaceae during the warmest July–August
warmconditioncanlimitthepollinationofAmbrosia months, because their pollen production decreases
especially the lack of water, since even this desert- belowthelatespringpeakinordertopreservewater
origin plant group tries to preserve the water in its for survival. At the same time, there are exemptions
body. The lack of precipitation, even the occurrence forsomesalt-tolerantspeciesandforthosefavouring
of arid months before the usual late July–August earlysummerprecipitation(seeearlier)thatismainly
pollen season may lead to a decrease in its pollen responsibleforthepollenproductioninthelatesum-
productioninitsmainpollinationperiod.Hightem- mer period. Nevertheless, they produce less pollen
peratures are accompanied with low precipitation, than those species flowering in late spring. In this
MeteorologicalinfluenceonpolleninHungary 223
Figure8. Dailypollenconcen-
trations of Populus in years
2
1 with extreme precipitation
0
2 (wettestyear:2004,driestyear:
er 1998); examination period of
ob theyears:days35–113.
ct
O
9
1
41 way, Poaceae living in salt meadows and Molinia reach very high pollen production. If precipitation
0: fens can be less responsible for allergenic diseases is small and it falls rarely, the pollination period is
0
at (Figures5,6). much shorter and the pollen production is much
a] High early spring temperatures can enhance the smaller,especiallyifthereisnoflood.Laterprecipita-
r
k
a growthofPopulusanditspollenproductioncanbegin tioncanhelpthesespecieslessastheyaregenetically
M
ó earlierduetotheearlierstartofitsvegetativeperiod. adapted to spring pollen production. It means that
zl These warm conditions do not limit water resources no time-shift or secondary peak can be expected if
s
Lá for these species, especially for those stocks living theconditionsareunfavourable(i.e.norain,noflood
[
], in active floodplains receiving mass inundation dur- duringtheearlyspringperiod;Figures7,8).
d
e ing early spring floods. In this way, sufficient water
g
ze withhigherspringtemperaturescanhelpearlierblos-
S
of soming of poplars resulting in higher pollen counts. Discussion
y In the case of a cool spring, even if there is no
sit inundation in the floodplain areas, their pollen pro- A moderate warming is favourable for Ambrosia
r
e
v duction is reduced because the metabolism of these (Ziska et al., 2011). The increase of the mean tem-
ni
U speciescanslowdown.Ifthebeginningofthespring perature for the warm-tolerant Ambrosia, especially
[
y iscool,poplarsmustwaituntilthedailyaveragetem- in summer time (August), can restrict its ability to
b
d peratures increases above 0 ◦C, since a freeze-free pollinate, since the plant concentrates on preserv-
e
d period is required for their metabolism and pollen ing water and maintaining its vegetative life func-
a
nlo production. If temperatures are below 0 ◦C for a tions rather than its generative functions. This is
w long time at the beginning of the year, the pollen in accordance with the negative association between
o
D season will begin later and shift to a later period. its pollen count characteristics (TPA and APC) and
Consequently, it will last several days longer and temperature(TablesI,II).Thisgenuscanadaptwell
dailypeakpollenconcentrationwillbesmaller,com- to dry and hot conditions, but is highly influenced
pared to those in warm years. As poplars produce by future land use. If more fallows and abandoned
pollen in early spring, only a short warm period human habitats appear in the landscape, its further
is available to start their annual life-cycle. In this increase may be awaited especially on sand soils
way, a fast and early warming following the win- (Deák, 2010) in spite of the expected warming and
ter can be essential to produce more pollen. The drying summers in the Carpathian basin (Bartholy
fasterthiswarmingisandthehighertheearlyspring etal.,2008).
temperatures can be, the most effective and higher Poaceae can produce high biomass in years with
pollen-production happens. Precipitation and inun- higher than usual rainfall, which is in accordance
dation can also substantially influence the pollen withhigherpollenproduction(TablesI,II).Ontheir
productionofpoplarsandcanbeevenhigherlimita- present diverse habitats – from sand steppes to
tionfactors for Populus.Even moderate but frequent floodplain meadows – rainfall increases biomass.
rainfalls combined with floods can help poplars to Significanthabitatchangesandgrassland-zoneshifts
224 L.Makraetal.
aredetected(Deák,2010)betweentheneighbouring release. Hence, a changing climate (warming and
zonesinwetyears.Accordingly,wettercommunities drying) may partly contribute to an extension of the
with higher biomass can appear on formerly drier pollenseason(TablesI,II;Caramielloetal.,1994).
places, so steppe grasslands can shift into mead- Concerning statistical approaches to forecast the
ows both in sand, loess and saline grassland domi- pollen season of Ambrosia, Laaidi et al. (2003)
nated areas. These habitat-shifts are led by grasses; applied two models, namely (1) summing the tem-
they appear first in the neighbouring habitats sev- peratures and (2) a multiple regression on 10-day
eral times. Taller grasses like Molinia and Alopecurus or monthly meteorological parameters (minimum,
have higher pollen-producing capacity compared maximum and mean air temperature, rainfall, rel-
withFestuca.However,actualgrassesofhabitatscan ative humidity, sunshine duration and soil temper-
also produce more pollen without habitat shift, as a ature), for predicting the start and the duration of
result of higher rainfall. Similarly, cultivated crops thepollenseasonofAmbrosiaforLyon,France.The
(like wheat or maize) can have higher biomass due SPSwaspredictedwithbothmethodsandtheresults
toincreasedrainfall,whichmeanshigherpollenpro- were more accurate when applying the regression
2 duction in their cases, as well (Tables I, II; Recio method (the errors between the predicted and the
1
0 et al., 2010; Sicard et al., 2010). Poaceae show observed SPS ranged from zero to three days). The
2
er high sensitivity to the amount of available water duration of the pollen season was predicted by a
b
o (Tables I, II). High temperatures can be limits for regressionmodelproducingerrorsrangingfromzero
ct
O themandcancauseregionaldecreases.Nevertheless, tosevendays.Ourmethodispartlysimilartothatof
9 asthespecies-poolofthisfamilyisthewidestamong Laaidietal.(2003)asweconsideredcumulateddaily
1
1 the studied plant groups, there will be species to values of meteorological variables (temperature and
4
0: substitute the actual grasses. Even species from the precipitation)whencorrelatingthemwiththepollen
0
at Mediterranean and more continental areas will be characteristics(TPAandAPC).
a] able to reach the Carpathian-basin in the future. Deen (1998) showed that the rate of develop-
r
k
a Although, in this case, the present species will be ment of common ragweed increased with tempera-
M
at risk; however, intra-taxonic re-assemblage could ture. Furthermore, strong associations of Ambrosia,
ó
zl solve the problem. Shortage of water (Tables I, II) Poaceae and Populus pollen counts with tempera-
s
Lá and too high temperatures can cause lower pollen ture and rainfall have been determined by several
[
], production of natural grasslands and also in arable authors. Pollen counts were found to increase with
d
e lands. This is why certain species in certain time temperature and decrease with rainfall for Ambrosia
g
e periods and places can suffer from climate change. (Bartková-Šcˇevková, 2003; Makra et al., 2004;
z
S
f Nevertheless,thechangeinspeciescompositionwill Peternel et al., 2006a; Piotrowska & Weryszko-
o
y give a good chance for the survival of this family. Chmielewska, 2006; Kasprzyk, 2008; Šcˇevková
sit Based on our data, Poaceae is sensitive to precip- et al., 2010) and for Poaceae (Bartková-Šcˇevková,
r
e
v itation changes, while is indifferent to temperature 2003; Peternel et al., 2006b; Šcˇevková et al., 2010).
ni
U variations. In a more detailed study, Makra et al. (2011) found
y [ Plantation of Populus species has not yet stopped that an association measure is negative between the
b
d duringthelasttenyears.Besideslocust-tree(Robinia annual cycles of the daily slopes of Ambrosia and
e
d pseudoacacia L.), they are the most favoured trees Poaceaepollenconcentrationtrendsontheonehand
a
o of afforestation in the Szeged region. The stocks and the annual cycles of the daily slopes of mean
nl
w planted during the last decades have grown up and temperature trends on the other hand. This mea-
o
D are in mature state, so they can pollinate on high sureispositivebetweentheearliermentionedslopes
level. Warmer, moderately humid weather in the of Poaceae pollen counts and rainfall. Laaidi et al.
spring also favours their pollination. Since Populus (2003)foundthatanincreaseintemperatureimplied
has both wet and dry tolerant species from flood- an earlier start of the Ambrosia pollen season. High
plains to bare sand, they have high environmen- daily mean Poaceae pollen levels are facilitated by
tal tolerance. Furthermore, they have low climate anticyclone ridge weather situations (influenced by
sensitivity (i.e. wide range of tolerance for climate high mean values of temperature and air pressure
conditions; Deák, 2010). However, the discrepancy as well as low relative humidity and wind speed) in
betweentheirlowclimatesensitivityontheonehand accordance with expectations (Matyasovszky et al.,
and a remarkably earlier start (Table II), later end 2011). Of the three taxa investigated, only Poaceae
(Tables I, II) and longer duration (Table II) of their show a significant increase during the pollen season.
pollen season on the other hand should be justified. Although Populus does not have any change con-
A warming and drying climate is more favourable cerning its pollen season, both the TPA and APC
for them in general, facilitating their higher pollen are definitely rising. Regarding the pollen season of