Forecast model of allergenic hazard using trends [620204]

ORIGINAL PAPER
Forecast model of allergenic hazard using trends
ofPoaceae airborne pollen over an urban area in SW
Iberian Peninsula (Europe)
Santiago Ferna ´ndez-Rodrı ´guez1•Pablo Dura ´n-Barroso1•
Inmaculada Silva-Palacios2•Rafael Tormo-Molina3•
Jose´Marı´a Maya-Manzano3•A´ngela Gonzalo-Garijo4
Received: 14 April 2016 / Accepted: 31 May 2016 / Published online: 11 June 2016
/C211Springer Science+Business Media Dordrecht 2016
Abstract Cities are becoming bigger, being necessary the knowledge of associated natural
hazards from organic and inorganic aerosols. This hazard could be included in the context
of urban air pollution and climate change as environmental risk factors for allergy. Overall,
grass pollens are the most important cause of pollinosis in Europe due to its high aller-
genicity and extensive distribution. The main objective of this work was to model daily
average Poaceae airborne pollen concentrations from an urban area placed in a city in the
SW of the Iberian Peninsula, taking into account the temporal distribution of five different
meteorological variables from 23 years of continuous recording. This was achieved using a
combination with the Shuffle Complex Evolution Metropolis Algorithm using as an
optimisation function the root mean square error. Aerobiological sampling was conducted
from 1993 to 2015 in Badajoz (SW Spain) using a 7-day Hirst-type volumetric sampler.
The Poaceae Main Pollen Season lasted, on average, 89 days, ranging from 41 to
144 days, from April 17 to July 14. The model proposed to forecast airborne pollen
concentrations is described by one equation composed of two terms. The first term rep-
resents the resilience of the pollen concentration trend in the air according to the average
concentration of the previous 10 days, and the second term is obtained from considering
the actual pollen concentration value, which is calculated based on the most representative
meteorological variables multiplied by a fitting coefficient. The fit of the model was
examined for a forecast horizon of 1, 7, 15 and 30 days. The R2values obtained were 0.70,
&Santiago Ferna ´ndez-Rodrı ´guez
[anonimizat]
1Department of Construction, School of Technology, University of Extremadura, Avda. de la
Universidad s/n, Ca ´ceres, Spain
2Department of Applied Physics, Engineering Agricultural School, University of Extremadura,
Avda. Adolfo Sua ´rez s/n, Badajoz, Spain
3Department of Plant Biology, Ecology and Earth Sciences, Faculty of Science, University of
Extremadura, Avda. Elvas s/n, Badajoz, Spain
4Department of Allergology, Infanta Cristina University Hospital, Avda. Elvas s/n, Badajoz, Spain
123Nat Hazards (2016) 84:121–137
DOI 10.1007/s11069-016-2411-0

0.69, 0.62 and 0.57, respectively, which show a trend in decreasing order. These results
confirm the suitability of the proposed model.
Keywords Airborne natural hazards /C1Poaceae pollen /C1Allergy /C1Time series analysis /C1
Forecasting /C1Temporal modelling
1 Introduction
A rise in the global average temperature happened between 1880 and 2012 and it has been
anticipated that this trend will continue in the following years according to IPCC reports
(2013 ). The impact of climate change has been studied with different natural indicators
such as runoff and soil losses in a rainfed basin (Ramos and Martı ´nez-Casasnovas 2015 ),
vegetation and animal husbandry (Miao et al. 2016 ), weed pollen (Bogawski et al. 2014 )
and allergens and allergic diseases (Beggs 2014 ). There are several environmental factors
related to plants such as temperature changes, the interaction with local meteorologicalparameters (mainly precipitation and wind) (Aboulaich et al. 2013 ; Myszkowska 2014 ),
edaphic factors (Wielgolaski 2001 ) and photoperiod (Iba ´n˜ez et al. 2010 ). The backspin of
climate on plant phenological response may be related by the impact of human activity onplant distribution with changes in land use. Garcı ´a-Mozo et al. ( 2016 ) have studied this
relation for Poaceae pollen. Garcı ´a de Leo ´n et al. ( 2015 ) have highlighted the Poaceae
pollen dynamics and how climate change might impact the future evolution of airborne
Poaceae pollen concentrations and thus the future evolution of related pollen allergies.
Poaceae family comprises more than 700 genera containing around 10,000 species
(Wang et al. 2013 ), with over 420 of these occurring in Europe (Gala ´n et al. 1989 ). Daily
average grass pollen concentrations and features of the grass pollen season such as startdate and intensity have been studied (Ugolotti et al. 2015 ; Zhang et al. 2015 ). Several
papers have examined the relationship between airborne grass pollen concentrations withsymptoms of allergic diseases (Annesi-Maesano et al. 2012 ; Erbas et al. 2012 ;S a´nchez-
Mesa et al. 2005 ) and the pollution air (Annesi-Maesano et al. 2012 ). Grass pollen is the
main cause of pollinosis in several regions of the world, but its frequency may changebetween zones (Tormo-Molina et al. 2015 ). Poaceae pollen is one of the four major
allergenic pollen families (Smith et al. 2014 ), inducing allergic rhinoconjunctivitis in
Europe (D’Amato et al. 2007 ). A statement of the World Allergy Organization have
discussed about meteorological conditions, climate change, new emerging factors andasthma and related allergic disorders (D’Amato et al. 2015 ).
Other papers have developed comparative studies examining temporal and spatial
variations in Poaceae pollen counts (Gala ´n et al. 1995 ; Smith et al. 2009 , Garcı ´a-Mozo
et al.
2009 ; Ferna ´ndez-Rodrı ´guez et al. 2015a ), and have estimated of the potential allergy
risk of a city (Rojo et al. 2016 ). Influence of Poaceae pollen in urban air has been also
studied (Moreno-Sarmiento et al. 2016 ; Vesterberg 2001 ; Peel et al. 2014 ). Urban green
infrastructure is a concept that includes not only ornamental street trees, but any area withnatural vegetation in urban environments. It has been developed in the last two decades andit is clear that improves quality of life for citizens (Breuste et al. 2015 ). Due to demo-
graphic trends of population from rural to urban areas, cities are becoming bigger (Sch-
midheiny and Suedekum 2015 ), being necessary the knowledge of associated natural
hazards from organic (fungal spores) (Srivastava et al. 2012 ) and inorganic aerosols122 Nat Hazards (2016) 84:121–137
123

(Particulate Matter 10-PM10) (Ambade 2016 ). This hazard could be included in the
context of urban air pollution and climate change as environmental risk factors for allergy(35). In 1998, the Conference ‘‘Ragweed in Europe’’ was dedicated to pollen allergy of
ragweed as natural hazard according UE ordinances (Comtois 1998 ;J a¨ger1998 ;J a´rai-
Komlo ´di1998 ; Makovcova ´et al. 1998 ; Peeters 1998 ; Thibaudon 1998 ).
Predicting the onset and duration of pollen season benefits allergic patients, providing
information by forecasting pollen models and their levels at different spatiotemporal scales(Zhang et al. 2015 ). To forecast Poaceae pollen is necessary to consider daily values of
airborne Poaceae pollen and the Season Pollen Index (SPI) (Ugolotti et al. 2015 : Fer-
na´ndez-Rodrı ´guez et al. 2016a ; Kasprzyk and Walanus 2010 ; Piotrowska 2012 ; Puc and
Wolski 2013 ), as well as trends in airborne related to climate change (Bogawski et al.
2014 ; Jato et al. 2009 ; Tormo-Molina et al. 2010 ; Garcı ´a-Mozo et al. 2010 ).
MATLAB (Matrix Laboratory) is a language of technical computing that has allowed to
model green areas (Hao et al. 2015 ; Rivera et al. 2015 ) and pollen data (Gong et al. 2015 ;
Senyuva et al. 2009 ) between time series of airborne pollen: Cupressaceae (Silva-Palacios
et al. 2015 ),Quercus (Ferna ´ndez-Rodrı ´guez et al. 2016a ) and Olea (Ferna ´ndez-Rodrı ´guez
et al. 2016b ) and meteorological parameters. This tool has verified the close agreement
between observed and predicted mean concentrations. The statistical tools used to analysepollen season trends have included time series analyses (Damialis et al. 2007 ), correlation
with meteorological variables (Erbas et al. 2007
), regression analyses (Stach et al. 2008 ),
process-based models (Garcı ´a-Mozo et al. 2009 ) and multiple regression (Myszkowska
2014 ; Brighetti et al. 2014 ; de Weger et al. 2014 ) for Poaceae pollen. Forecasting of
Poaceae pollen has been widely developed in Europe: Spain (Garcı ´a-Mozo et al. 2009 ),
France (Cassagne et al. 2008 ), Italy (Tassan-Mazzocco et al. 2015 ), Greece (Voukantsis
et al. 2010 ), Turkey (Kizilpinar et al. 2011 ), UK (Smith and Emberlin 2005 ; Smith and
Emberlin 2006 ) and Poland (Piotrowska 2012 ). Other tools used include methods based on
artificial intelligence (AI) for Poaceae pollen applied to a coastal Atlantic climate region
(Rodrı ´guez-Rajo et al. 2010 ) and in polluted areas in Europe (Cse ´pe et al. 2014 ).
The main objective of this work is to model daily average Poaceae airborne pollen
concentrations from an urban area placed in a city in the SW of the Iberian Peninsula, in
relation to the temporal distribution of five different meteorological variables from
23 years of continuous recording. This was achieved using a combination with the ShuffleComplex Evolution Metropolis Algorithm (SCEM-UA) using as an optimisation functionthe root mean square error (RMSE).
2 Materials and methods
2.1 Sampling site
Aerobiological samples were collected from 1993 to 2015 in Badajoz (SW Spain) using a7-day volumetric sampler of the Hirst design (Hirst 1952 ) with an intake hole located
1.5 m above ground on an open terrace situated 6 m above ground at the AgriculturalEngineering School of the University of Extremadura (38 /C17653
04500N, 6/C1765800700W). Petro-
latum White (CAS number 8009-03-8) was used as an adhesive. Standardised data man-agement procedures were used as described by the Spanish Aerobiology Network (REA)(Gala´n et al. 2007 ). The main pollen season (MPS) was determined using the 5–95 % range
(Nilsson and Persson 1981 ) of the data. The data were used in the form of the SeasonalNat Hazards (2016) 84:121–137 123
123

Pollen Index (SPI), calculated as the sum of the daily average pollen counts recorded in
each MPS for Poaceae pollen. Figure 1includes a 12 910 km map with the main land
uses located around the spore trap.
2.2 Climate data
Badajoz has a Mediterranean climate with the maximum rainfall in autumn and winter(AEMET 2016 ). The highest temperatures are recorded during the summer dry period.
Annual values of rainfall (447 mm), mean temperature (17.1 /C176C) and relative humidity
(64 %) were registered in the period 1981–2010. Wind data available (1993–2008) providean average wind speed of 9.5 km/h with SW (35.4 %) and NW (26.4 %) winds beingpredominant. Climate data were obtained from Badajoz Airport (38 /C17653
0N, 6/C176490W)
(Fig. 1).
Spore trap Meteorological stationN
Fig. 1 Maps of location of spore trap and meteorological station and land use. Spore trap Meteorological
station124 Nat Hazards (2016) 84:121–137
123

2.3 Statistical analysis
The normal distribution of the data was tested using the Kolmogorov–Smirnov and Sha-
piro–Wilk tests. These tests showed that the daily data did not follow a normal distribution.The tests also showed that the log 10-transformed data also did not follow a normaldistribution. A Spearman correlation test was then used to analyse associations betweenselected variables. The statistical analysis included the entire daily data set and was per-formed with the SPSS 15.0 statistical package. The use of meteorological information topredict the Poaceae pollen concentration was evaluated with a mathematical model. The
model was calibrated using the Shuffled Complex Evolution Metropolis algorithm (SCEM-UA) with the root mean square error (RMSE) as the optimisation function. The suitability
of the SCEM-UA has been previously verified in other areas of science, such as hydro-
dynamics (Vrugt et al. 2003 ; Efstratiadis and Koutsoyiannis 2010 ). This algorithm is based
on an automatic search of the feasible parameter space and starts by generating randomsamples of unknown calibration parameters. From this search, the set of solutions thatoptimises the goodness-of-fit criterion is found. Meteorological parameters used wererainfall (R), relative humidity (RH), maximum (Tmax), mean (Tmean) and minimumtemperature (Tmin). Probability distribution function of parameters and R
2has been
analysed by MATLAB.
The model proposed to forecast the airborne Poaceae pollen concentration (CP) is
described by Eq. ( 1). This expression is composed of two terms: the first term represents
the resilience of the pollen concentration trend in the air according to the average con-
centration of the previous 10 days; the second term is obtained based on the actual pollen
concentration value, which is composed by the most representative climatic variablesmultiplied by a fitting coefficient. Finally, considering the effect of previous days on thedaily value of each variable the average temperature and the cumulative rainfall of theprevious 10 days to the equation:
CP
tțDt¼a/C1Pi¼t
i¼t/C010CPi
10
țCPtb/C1Tt
maxțc/C1Tt
meanțd/C1Tt
mințe/C1Pi¼t
i¼t/C010Ti
mean
10țf/C1Rt
țg/C1Xi¼t
i¼t/C010Rițh/C1RHt!
ð1Ț
where a,b,c,d,e,f,gandhare the coefficients to be calibrated for our time series of
climatic and airborne pollen concentration and Dtis the time step considered for the
forecasting process, which is fixed at 1 day. It is composed by the integration of differentclimatic variables for each time step regarding the actual airborne pollen concentrationvalue joined to the mean concentration value of the previous 10 days for each time step.Besides the fact that the value of each variable is evaluated for every time step, the
aggregated variables of the mean temperature for 10 previous days as well as the cumu-
lative rainfall of the 10 previous days have been included to analyse the influence of thetemporal variation of these variables over the airborne pollen concentration. Regardless ofthe fact that the degree of influence of some meteorological variables could be previouslyneglected by its apparently low relevance, it has been therefore considered appropriate tomaintain them due to the fact that the range of variations are different among them as wellNat Hazards (2016) 84:121–137 125
123

as its mean value. The fitting parameters for this model were estimated using the SCEM-
UA algorithm.
3 Results
3.1 Trends
ThePoaceae MPS lasted, on average, 89 days, ranging from 41 to 144 days, from April 17
to July 14 (Table 1). Not statistically significant correlation was found between the SPI and
the number of days in the MPS ( r=- 0.260, p=0.232) (Fig. 2). The mean daily average
pollen concentration was 116 grains/m3, with a range of 26–303 pollen grains/m3. Figure 3
shows a trend towards a shorter pollination period by delaying the start day and advancingthe last day of the MPS. The day with maximum concentration (day maximum) is pro-gressively closer to the day end, and the correlation between the day maximum and day
Table 1 Poaceae pollen season characteristics (day start, day end, day maximum, number of days, daily
average count and SPI)
Day start
(DOY)Day end
(DOY)Day maximum
(DOY)Number of
days (days)Daily average count
(pollen grains/m3)SPI (pollen
grains)
1993 18-5 2-8 30-5 77 82 6345
1994 25-4 23-7 4-5 90 119 10,7441995 27-3 23-7 17-5 115 46 52881996 15-4 18-7 31-5 95 185 17,571
1997 18-3 8-8 20-5 144 60 8572
1998 23-3 16-7 15-5 116 119 13,7651999 24-4 29-7 25-5 97 68 65532000 12-5 13-7 18-5 61 281 16,613
2001 26-4 1-7 25-5 67 303 20,272
2002 27-4 1-7 20-5 66 152 10,0512003 7-5 16-7 20-5 71 217 15,3772004 23-3 11-7 26-5 111 108 11,970
2005 26-3 20-7 5-5 117 26 3004
2006 20-4 1-7 17-5 73 139 10,1682007 25-3 3-6 19-5 80 134 10,6812008 30-4 30-7 18-5 92 69 6366
2009 26-3 6-8 8-5 130 27 3481
2010 28-4 15-7 30-5 79 108 85262011 26-3 6-7 14-5 103 50 51222012 10-5 23-7 25-5 75 38 2882
2013 25-4 2-7 22-5 69 54 3750
2014 10-5 26-7 17-5 78 107 83642015 28-4 7-6 14-5 41 189 7731
Start, end and maximum dates of the Poaceae pollen season as day of the year (DOY)126 Nat Hazards (2016) 84:121–137
123

start ( r=0.336, p=0.117) is higher than between the day maximum and the day end
(r=0.001, p=0.995). The number of days of Poaceae MPS is statistically significant
correlated with day start ( r=- 0.776, p=0.001), day end ( r=0.599, p=0.003) and
daily average concentration ( r=- 0.596, p=0.003).1992 1994 1996 1998 2000 2002 2004 2006 2008 2010 2012 20145060708090100110120130140150160170180190200210220230240250
Starting day Ending day Maximum dayDays of a year [Days]
Year
Fig. 2 Trends in the pollen season for Poaceae pollen
Fig. 3 Poaceae pollen season characteristicsNat Hazards (2016) 84:121–137 127
123

3.2 Relationship with meteorological variables
The first analysis performed was a Spearman correlation analysis of the relationship
between the Poaceae SPI and the selected meteorological variables. Table 2shows the
results obtained in each year. Temperature (maximum, minimum and mean) and relativehumidity were the meteorological variables with the highest positive correlation coefficientwith the SPI. The correlation with relative humidity, rainfall and temperatures (maximum,minimum and mean) fluctuated between positive and negative values and was only sig-nificant for 1 year (1998). Nevertheless, 11 years showed a significance level greater than95 % for the cumulative rainfall and temperature mean of 10 previous days. Separately, 15and 17 years showed significance level for rainfall and temperature mean of 10 previous
days, respectively. All years have shown significance level for cumulative Poaceae con-
centration of 10 previous days.
3.3 Model
In Table 3, the values obtained for each parameter for the 23 years analysed, and the value
obtained from the coefficient of determination ( R2) is shown. The model has been validated
Table 2 Correlation coefficient between daily SPI and meteorological parameters
Tmax Tmin Tmean Rain RH Rain
(10 days)Tmean
(10 days)CPoa
(10 days)
1993 -0.383** -0.433** -0.474** 0.134 0.343** 0.664** -0.767** 0.741**
1994 -0.259* -0.283* -0.302** -0.095 -0.007 0.047 -0.399** 0.740**
1995 0.181 0.107 0.019 -0.105 -0.252** 0.218* -0.036 0.737**
1996 0.034 0.052 0.037 -0.232* 0.056 -0.247* -0.087 0.825**
1997 0.219 * 0.228** 0.079 -0.311** -0.257** -0.015 0.094 0.681**
1998 0.552** 0.548** 0.490** -0.257** -0.323** -0.055 0.456** 0.852**
1999 -0.287 ** -0.341** -0.381** -0.189 0.027 0.343** -0.596** 0.805**
2000 -0.219 -0.222 -0.184 0.138 0.230 0.630** -0.731** 0.790**
2001 -0.178 -0.182 -0.191 -0.015 0.230 0.269** -0.435** 0.808**
2002 -0.105 -0.163 -0.280* -0.185 -0.077 0.467** -0.614** 0.681**
2003 0.019 -0.131 -0.279* -0.176 -0.465** -0.363** -0.501** 0.819**
2004 0.410** 0.466** 0.542** -0.046 -0.217* 0.220* 0.407** 0.838**
2005 0.203* 0.151 0.094 0.070 -0.286** 0.114 -0.045 0.759**
2006 -0.165 -0.426** -0.613** -0.200 -0.309* -0.450** -0.740** 0.871**
2007 0.112 -0.103 -0.408** -0.303* -0.265* -0.251* -0.503** 0.767**
2008 -0.413** -0.483 -0.521** 0.118 0.176 0.498** -0.616** 0.750**
2009 0.115 0.132 0.101 -0.314** -0.209* 0.092 -0.021 0.772**
2010 -0.231 -0.239* -0.302* -0.253* 0.038 -0.158 -0.552** 0.792**
2011 0.151 0.188 0.060 -0.332** -0.099 0.246* -0.042 0.678**
2012 -0.290* -0.172 -0.093 0.205 0.323** 0.485** -0.386** 0.807**
2013 0.045 -0.011 -0.244 -0.217 -0.180 0.053 -0.588** 0.559**
2014 -0.397** -0.484** -0.491** -0.104 0.011 0.155 -0.645** 0.799**
2015 -0.285 -0.326 -0.213 0.000 0.086 0.577** -0.650** 0.732**
* Significance at level 95 %; ** Significance at level 99 %128 Nat Hazards (2016) 84:121–137
123

for the last 5 years considering the rest of the data for the calibration process (Fig. 4),
obtaining a correlation between the observed data (1993–2010) ( R2=0.70) and simulated
data (2011–2015) ( R2=0.69). Moreover, Table 3also includes the results obtained with a
forecast time step of 1, 7, 15 and 30 days to examine the suitability of the model to predictconcentrations for longer time gaps. For this purpose, Eq. ( 1) is modified with Dtequal to 7
for a forecast time step of 7 days, 15 for a forecast time step of 15 days and 30 for a
forecast time step of 30 days. Parameter ‘‘ a’’ had the highest value, showing the influenceTable 3 Parameters and statisti-
cal analysis of the model pro-
posed to predict airborne pollenconcentration (D: previous day)Time prediction delay
1-day 7-day 15-days 30-days
Parameters
a 0.491 1.286 0.941 0.819
b 0.006 0.074 -0.003 -0.003
c 0.082 -0.092 0.032 0.027
d -0.066 0.061 -0.007 -0.007
e -0.041 -0.079 -0.035 -0.028
f -0.017 -0.015 0.001 -0.009
g 0.004 0.002 0.001 0.001
h 0.451 0.570 0.028 0.430
Comparison criteria
R
20.704 0.689 0.622 0.573
RMSE 95.47 95.34 92.92 65.73
CC obs 103.677
CC sim 100.869 98.194 82.279 68.370
2011 2012 2013 2014 20150.00.10.20.30.40.50.60.70.80.91.0Coeffcient of determination R2
Year Time prediction delay: 1 day
Time prediction delay: 7 days
Time prediction delay: 15 days
Time prediction delay: 30 daysCalibration period: 1993-2010 Validation period: 2011-2015
Fig. 4 Validation of the model for the last 5 yearNat Hazards (2016) 84:121–137 129
123

of the previously explained trend in the pollen concentration. Among the parameters
related to daily temperature, parameter ‘‘ b’’ related to maximum temperature with the
prediction to 7 days has shown the highest value; however, parameter ‘‘ c’’ linked to mean
temperature for forecast to 1 day was higher than 7, 15 and 30 days. Parameter ‘‘ d’’
associated with minimum temperature showed the greatest value for 7 days as the meantemperature. Rainfall (parameter ‘‘ f’’) has shown lower values. The value of parameter
‘‘g’’, linked to the average rainfall of the 10 previous days, was higher than ‘‘ f’’ parameter
for a forecast time step. Finally, the value of parameter ‘‘ h’’, which is linked to daily
relative humidity, was the higher parameter. Figure 5shows the posterior probability
distribution function of all parameters (‘‘ a’’ to ‘‘ h’’) and R
2. Red lines show the most
accurate combination of parameters. The position of each red line is close to the value withhigher probability for all parameters. Consequently, any variation around these valueswould not generate high differences in terms of R
2.
4 Discussion
Badajoz is a historic city and is the urban centre most populated of SW Iberian Peninsula.
Urban area extends about 8 km E–W and 6 N–S with approximately 18 km2, river Gua-
diana crosses the city from NE to SW, and river banks were recently modified as greenornamental areas. Urban surrounding are covered by irrigated crops at NE and NW
Fig. 5 Posterior probability distribution function of parameters ( a,b,c,d,e,f,gandh) and R2.Red line
means the more accurate combination of parameters130 Nat Hazards (2016) 84:121–137
123

(cornfields, fruit trees crops and other vegetable crops); dried crops appear at S, N and NW
(wheat and oats); some small areas of holm oak and cork oak dehesa appear mainly at S,some time shrubbery. Urban green areas represent 5.5 % or urban surface.
Recent papers have reported that airborne pollen recorded in Badajoz comes from
median and distance sources. In Badajoz, Ferna ´ndez-Rodrı ´guez et al. ( 2015a ,2015c ) have
reported that airborne organic particles (fungal spores and pollen grains) recorded comesfrom local and regional sources with winds of West direction (Silva-Palacios et al. 2007 ).
Grass pollen concentrations recorded in a site depends of environment specific conditions
(Mun˜oz-Rodrı ´guez et al. 2010 ), being representative emission sources up to 100 km
around the sampler (Domı ´nguez et al. 1993 ; Fornaciari et al. 2000 ). Phenology of wind-
pollinated plants such as grasses is a tool to study aerobiological results (Tormo-Molinaet al. 2015 ) and the relationship of spatiotemporal evolution of phenological phases, and
pollen counts in sampling site (Tormo-Molina et al. 2011 ) allow a better knowledge of
Poaceae pollen.
Trends in Poaceae pollen concentration displayed negative slopes, indicating a decrease
in airborne grass pollen in Badajoz (SW Spain) as in Italy (Ugolotti et al. 2015 ) and in
Poland (Stach et al. 2008 ), nevertheless not being in line with the trends in Andalucı ´a( S
Spain) (Garcı ´a-Mozo et al. 2010 ) and in the UK (Smith and Emberlin 2005 ). Changes in
land use and a general reduction in grassland area have been used to explain a marked
decrease in the length and severity of the grass pollen (Emberlin et al. 1993 ). A completely
reverse effect of humidity was noted in the UK where in wetter years, grass pollen releasewas delayed (Emberlin et al. 1994 ). Smith et al. ( 2014 ) have showed that of the intensity of
Poaceae pollen seasons has changed at several European cities, but this change was not
consistent. The Poaceae MPS lasted about 2 months and, although trends are not statis-
tically significant, it seems that there is a small delay trend in the pollination. This factcould be explained because the number of MPS days is correlated with the end day. Thedelay or advance in pollination differs from other studies and it is quite dependent on theyears analysed. Barbosa et al. ( 2015 ) have shown the relationship between NDVI and
rainfall might occur not in time neither in space, with a delay of both parameters. This factcould be justified by the fact that the occurrence of El Nin ˜o-Southern Oscillations (ENSO)
events at the end of a year would cause decreased or increased rainfall in the rainy season.
A delay was found of the pollen season in Extremadura for Cupressaceae pollen (Silva-
Palacios et al. 2015 ),Quercus pollen (Ferna ´ndez-Rodrı ´guez et al. 2016a ) and Olea pollen
(Ferna ´ndez-Rodrı ´guez et al. 2016b ). However, an advance of spring flowering days per
decade was found in Europe for trees (Menzel et al.
2006 ) and of the pollen season in
Betula (Zhang et al. 2013 ),Olea (Aguilera et al. 2015 ) and for herbs as Poaceae (Ugolotti
et al. 2015 ; Garcı ´a-Mozo et al. 2010 ).
Not statistically significant correlation was found between the SPI and the number of
days in the MPS; however, Garcı ´a-Mozo et al. ( 2016 ) recorded an increase in Poaceae
pollen concentration for 15 years in Co ´rdoba (S Spain). Piotrowska ( 2012 ) found a neg-
ative correlation between SPI and season duration, peak value and end date. On the other
hand, we have found statistical significant trend in MPS length and day start, day end and
daily average concentration. In contrast, Garcı ´a-Mozo et al. ( 2010 ) have found a length-
ening mostly due to the advance of the pollen season start. A trend to appear earlierbetween onset and peak of pollination was found. This may be in line with other pheno-logical studies of trees that display a general advance in flowering of Poaceae pollen
(Myszkowska 2014 ; Stach et al. 2008 ). However, other studies have shown advance in the
pollen season start (Garcı ´a-Mozo et al. 2010 ; Stach et al. 2007 ).Nat Hazards (2016) 84:121–137 131
123

Grass pollen concentrations are greatly influenced by weather conditions (Garcı ´a-Mozo
et al. 2009 ). The primary factors, prior to flowering, such as temperature, have influenced in
plant growth. After, secondary factor such as rainfall and relative humidity influence pollenrelease (Laaidi 2001 ). Garcı ´a de Leo ´n et al. ( 2015 ) have reported that Poaceae development is
influenced by exogenous perturbations, mainly temperature and rainfall. The positive relationbetween Poaceae pollen and temperature (maxium, minimum and mean) and rainfall is in
consonance with previous papers (Garcı ´a-Mozo et al. 2010 ). Our study has shown a stronger
relationship between Poaceae pollen and rainfall of 10 previous days. Garcı ´a-Mozo et al.
(2010 ) have reported the influence with rainfall over the previous months. This fact has been
explained because the growth of grasses is related to the water availability, particulary inMediterranean regions (Clary et al. 2004 ). In relation to the wind, Garcı ´a-Mozo et al. ( 2010 )
claimed that the punctual influence of wind was not a significant factor to control temporalchanges during a long term period, even though if is accurate for short analysis of pollen. Forthis reason, in this study we have not considered this meteorological factor.
It is widely known the relevance of models for predicting Poaceae pollen concentration as
the forecast of start of the grass pollen season (Garcı ´a-Mozo et al. 2009 ; Chuine and Belmonte
2004 ) and daily grass pollen (Stach et al. 2008 ; Smith and Emberlin 2005 ; Smith and
Emberlin 2006 ). Temperature and rainfall have been used in other studies (Zhang et al. 2015 )
to forecast pollen concentration. This study also takes into account pollen records from the 10
previous days, so the predictive value may be more accurate than using meteorological
parameters alone. We have considered the more adjusted results to the real pattern of polli-nation. Relative humidity was the main meteorological factor considered in terms of its effecton the daily pollen forecast. The daily mean temperature and mean temperature of theprevious 10 days were secondary factors that could modify this forecast. This fact could beexplained due to that herbs and grasses can be less resilient to increases in temperature andrelative humidity, unlike trees (Recio et al. 2010 ). Herbaceous plants such as grasses present
more immediate response to meteorological conditions than tree species (Dahl et al. 2013 ).
To test the accuracy 1 day ahead forecast of these results, R
2value (0.70) was estimated
for the multiple regression model proposed, being a value higher than reported for the sametime series for the pollen types; Cupressaceae pollen, 0.30 (Silva-Palacios et al. 2015 ) and
0.46 (Sabariego et al. 2012 ),Quercus pollen 0.58 (Ferna ´ndez-Rodrı ´guez et al. 2016a ). A
value closed, 0.77, was reported for Olea pollen (Ferna ´ndez-Rodrı ´guez et al. 2016b )i n
Badajoz. Other studies have reported forecast models for 7-day ahead forecast valuesbetween 0.62 and 0.47 (Smith and Emberlin 2005 ), and for 30 days ahead forecast, 0.62
(Smith and Emberlin 2006 ). Stach et al. ( 2008 ) have predicted the start of the grass pollen
season to within 2 days and achieved between 60 and 70 % accuracy. Furthermore, for theanalysis of the entire time period, the fit of the model was examined for a forecast horizonof 1, 7, 15 and 30 days. The R
2values obtained were 0.70, 0.69, 0.62 and 0.57, respec-
tively, which show a trend in decreasing order with accurate results in all cases. Theseresults and the distribution of posterior probability parameters and R
2have confirmed the
suitability and the robustness of the proposed model.
5 Conclusions
This research highlights the importance to study the climate change as environmental riskfactors for allergy and to quantify the natural risks hazards from organic aerosols, Poaceae
pollen, as natural hazard in the context of urban air pollution as Badajoz (SW Spain).132 Nat Hazards (2016) 84:121–137
123

Furthermore, these results obtained may be used for benefit allergy sufferers, medical
professionals and those who produce and stock health care products in Badajoz. The modelobtained provides a good level of confidence to forecast Poaceae airborne pollen con-
centration with an R
2of 0.70, 0.69, 0.62 and 0.57 for a forecast horizon of 1, 7, 15 and
30 days, respectively. These results and the values of probability distribution function byparameters and R
2have confirmed the suitability of the proposed model.
Acknowledgments This work was made possible by research projects PRI06A190 and PRI BS10008, and
research groups aid GR15060 financed by the Regional Government, Junta de Extremadura (Spain), and the
European Regional Development Fund.
References
Aboulaich N, Achmakh L, Bouziane H, Trigo MM, Recio M, Kadiri M, Cabezudo B, Riadi H, Kazzaz M
(2013) Effect of meteorological parameters on Poaceae pollen in the atmosphere of Tetouan (NW
Morocco). Int J Biometeorol 57(2):197–205
AEMET (2016) Climate statistics. Badajoz Airport (1981–2010). Agencia Estatal de Meteorologı ´a,
Gobierno de Espan ˜a
Aguilera F, Orlandi F, Ruiz-Valenzuela L, Msallem M, Fornaciari M (2015) Analysis and interpretation of
long temporal trends in cumulative temperatures and olive reproductive features using a seasonal trenddecomposition procedure. Agric For Meteorol 203:208–216
Ambade B (2016) Characterization of PM10 over urban and rural sites of Rajnandgaon, central India. Nat
Hazards 80(1):589–604
Annesi-Maesano I, Rouve S, Desqueyroux H, Jankovski R, Klossek JM, Thibaudon M, Demoly P, Didier A
(2012) Grass pollen counts, air pollution levels and allergic rhinitis severity. Int Arch Allergy Immunol
158(4):397–404
Barbosa HA, Lakshmi Kumar TV, Silva LRM (2015) Recent trends in vegetation dynamics in the South
America and their relationship to rainfall. Nat Hazards 77(2):883–899
Beggs PJ (2014) Impacts of climate change on allergens and allergic diseases: knowledge and highlights
from two decades of research. In: Butler CD (ed) Climate change and global health. pp 105–113.
doi:10.1079/9781780642659.0105
Bogawski P, Grewling Ł, Nowak M, Smith M, Jackowiak B (2014) Trends in atmospheric concentrations of
weed pollen in the context of recent climate warming in Poznan ´(Western Poland). Int J Biometeorol
58(8):1759–1768
Breuste J, Artmann M, Li J, Xie M (2015) Special issue on green infrastructure for urban sustainability.
J Urban Plan Develop 141(3)
Brighetti MA, Costa C, Menesatti P, Antonucci F, Tripodi S, Travaglini A (2014) Multivariate statistical
forecasting modeling to predict Poaceae pollen critical concentrations by meteoclimatic data. Aero-
biologia 30(1):25–33
Cassagne E, Caillaud D, Besancenot JP, Thibaudon M (2008) Forecasting the onset of an allergic risk to
Poaceae in Nancy and Strasbourg (France) with different methods. Eur Ann Allergy Clin Immunol
40(1):14–21
Chuine I, Belmonte J (2004) Improving prophylaxis for pollen allergies: predicting the time course of the
pollen load of the atmosphere of major allergenic plants in France and Spain. Grana 43(2):65–80
Clary J, Save ´R, Biel C, De Herralde F (2004) Water relations in competitive interactions of Mediterranean
grasses and shrubs. Ann Appl Biol 144(2):149–155
Comtois P (1998) Ragweed (Ambrosia sp.): the Phoenix of allergophytes. In: Spieksma FThM (ed) Rag-
weed in Europe. Satellite symposium proceedings of 6th international congress on aerobiology,
Perugia, Italy. Alk-Abello ´A/S, Horsholm, pp 3–5
Cse´pe Z, Makra L, Voukantsis D, Matyasovszky I, Tusna ´dy G, Karatzas K, Thibaudon M (2014) Predicting
daily ragweed pollen concentrations using computational intelligence techniques over two heavilypolluted areas in Europe. Sci Total Environ 476–477:542–552
Dahl A, Gala ´n C, Hajkova L, Pauling A, Sikoparija B, Smith M, Vokou D (2013) The onset, course and
intensity of the pollen season. In: Sofiev M, Bergmann K-C (eds) Allergenic pollen. A review of the
production, release, distribution and health impacts. Springer, Dordrecht, pp 29–70Nat Hazards (2016) 84:121–137 133
123

D’Amato G, Cecchi L, Bonini S, Nunes C, Annesi-Maesano I, Behrendt H, Liccardi G, Popov T, Van
Cauwenberge P (2007) Allergenic pollen and pollen allergy in Europe. Allergy Eur J Allergy Clin
Immunol 62(9):976–990
D’Amato G, Cecchi L, D’Amato M, Liccardi G (2010) Urban air pollution and climate change as envi-
ronmental risk factors of respiratory allergy: an update. J Investig Allergol Clin Immunol 20(2):95–102
D’Amato G, Holgate ST, Pawankar R, Ledford DK, Cecchi L, Al-Ahmad M, Al-Enezi F, Al-Muhsen S,
Ansotegui I, Baena-Cagnani CE, Baker DJ, Bayram H, Bergmann KC, Boulet LP, Buters JTM,
D’Amato M, Dorsano S, Douwes J, Finlay SE, Garrasi D, Go ´mez M, Haahtela T, Halwani R, Hassani
Y, Mahboub B, Marks G, Michelozzi P, Montagni M, Nunes C, Oh JJW, Popov TA, Portnoy J, RidoloE, Rosa ´rio N, Rottem M, Sa ´nchez-Borges M, Sibanda E, Sienra-Monge JJ, Vitale C, Annesi-Maesano I
(2015) Meteorological conditions, climate change, new emerging factors, and asthma and related
allergic disorders. A statement of the World Allergy Organization. World Allergy Organ J (1):8–25.
doi:10.1186/s40413-015-0073-0
Damialis A, Halley JM, Gioulekas D, Vokou D (2007) Long-term trends in atmospheric pollen levels in the
city of Thessaloniki, Greece. Atmos Environ 41(33):7011–7021
de Weger LA, Beerthuizen T, Hiemstra PS, Sont JK (2014) Development and validation of a 5-day-ahead
hay fever forecast for patients with grass-pollen-induced allergic rhinitis. Int J Biometeorol
58(6):1047–1055
Domı ´nguez E, Infante F, Gala ´n C, Guerra F, de la Torre FV (1993) Variations in the concentrations of
airborne Olea pollen and associated pollinosis in Co ´rdoba (Spain): a study of the 10-years period
1982–1991. J Invest Allergol Clin Immunol 3(3):121–129
Efstratiadis A, Koutsoyiannis D (2010) One decade of multi-objective calibration approaches in hydro-
logical modelling: a review. Hydrol Sci J 55(1):58–78
Emberlin J, Savage M, Jones S (1993) Annual variations in grass pollen seasons in London 1961–1990:
trends and forecast models. Clin Exp Allergy 23(11):911–918
Emberlin J, Jones S, Bailey J, Caulton E, Corden J, Dubbels S, Evans J, McDonagh N, Millington W,
Mullins J, Russel R, Spencer T (1994) Variation in the start of the grass pollen season at selected sites
in the United Kingdom 1987–1992. Grana 33(2):94–99
Erbas B, Chang JH, Newbigin E, Dhamarge S (2007) Modelling atmospheric concentrations of grass pollen
using meteorological variables in Melbourne, Australia. Int J Environ Health Res 17(5):361–368
Erbas B, Akram M, Dharmage SC, Tham R, Dennekamp M, Newbigin E, Taylor P, Tang MLK, Abramson
MJ (2012) The role of seasonal grass pollen on childhood asthma emergency department presentations.
Clin Exp Allergy 42(5):799–805
Ferna ´ndez-Rodrı ´guez S, Adams-Groom B, Silva-Palacios I, Caeiro E, Brandao R, Ferro R, Gonzalo-Garijo
A´, Smith M, Tormo Molina R (2015a) Comparison of Poaceae pollen counts recorded at sites in
Portugal, Spain and the UK. Aerobiologia 31(1):1–10
Ferna ´ndez-Rodrı ´guez S, Skjøth CA, Tormo-Molina R, Brandao R, Caeiro E, Silva-Palacios I, Gonzalo-
Garijo A ´, Smith M (2015b) Identification of potential sources of airborne Olea pollen in the Southwest
Iberian Peninsula. Int J Biometeorol 58(3):337–348
Ferna ´ndez-Rodrı ´guez S, Sadys ´M, Smith M, Tormo-Molina R, Skjøth CA, Maya-Manzano JM, Silva-
Palacios I, Gonzalo-Garijo T (2015c) Potential sources of airborne Alternaria spp. spores in South-west
Spain. Sci Total Environ 533:165–176
Ferna ´ndez-Rodrı ´guez S, Dura ´n-Barroso P, Silva-Palacios I, Tormo-Molina R, Maya-Manzano JM, Gonzalo-
Garijo A ´(2016a) Quercus long-term pollen season trends in the southwest of the Iberian Peninsula.
Process Saf Environ Prot. 101:152–159. doi: 10.1016/j.psep.2015.11.008
Ferna ´ndez-Rodrı ´guez S, Dura ´n-Barroso P, Silva-Palacios I, Tormo-Molina R, Maya-Manzano JM, Gonzalo-
Garijo A ´(2016b) Regional forecast model for the Olea pollen season in Extremadura (SW Spain). Int J
Biometeorol. 1–9. doi: 10.1007/s00484-016-1141-z
Fornaciari M, Galan C, Mediavilla A, Dominquez E, Romano B (2000) Aeropalynological and phenological
study in two different Mediterranean olive areas: Cordoba (Spain) and Perugia (Italy). Plant Biosyst
134(2):199–204
Gala´n C, Cuevas J, Infante F, Domı ´nguez E (1989) Seasonal and diurnal variation of pollen from Gramineae
in the atmosphere of Co ´rdoba Spain. Allergol Immunopathol 17(5):245–249
Gala´n C, Emberlin J, Domı ´nguez E, Bryant RH, Villamandos F (1995) A comparative analysis of daily
variations in the gramineae pollen counts at Co ´rdoba, Spain and London, UK. Grana 34(3):189–198
Gala´nC ,C a r i n ˜anos P, Alca ´zar P, Dominguez-Vilches E (2007) Spanish Aerobiology Network (REA) man-
agement and quality manual. Servic io de Publicaciones Universidad de Co ´rdoba. ISBN 978-84-690-6353-8
Garcı ´a de Leo ´n D, Garcı ´a-Mozo H, Gala ´n C, Alca ´zar P, Lima M, Gonza ´lez-Andu ´jar JL (2015) Disentan-
gling the effects of feedback structure and climate on Poaceae annual airborne pollen fluctuations and
the possible consequences of climate change. Sci Total Environ 530–531:103–109134 Nat Hazards (2016) 84:121–137
123

Garcı ´a-Mozo H, Gala ´n C, Belmonte J, Bermejo D, Candau P, Dı ´az de la Guardia C, Elvira B, Gutie ´rrez M,
Jato V, Silva I, Trigo MM, Valencia R, Chuine I (2009) Predicting the start and peak dates of the
Poaceae pollen season in Spain using process-based models. Agric For Meteorol 149(2):256–262
Garcı ´a-Mozo H, Gala ´n C, Alca ´zar P, De La Guardia CD, Nieto-Lugilde D, Recio M, Hidalgo P, Go ´nzalez-
Minero F, Ruiz L, Domı ´nguez-Vilches E (2010) Trends in grass pollen season in southern Spain.
Aerobiologia 26(2):157–169
Garcı ´a-Mozo H, Oteros JA, Gala ´n C (2016) Impact of land cover changes and climate on the main airborne
pollen types in Southern Spain. Sci Total Environ 548–549:221–228
Gong WH, Zhang YL, Feng F, Liu FC, He YF (2015) Complex motions of grains in dusty plasma with
nonuniform magnetic field. Wuli Xuebao/Acta Phys Sin 64(19):0195202. doi: 10.7498/aps.64.195202
Hao Y, Kang J, Krijnders JD (2015) Integrated effects of urban morphology on birdsong loudness and
visibility of green areas. Landsc Urban Plan 137:149–162
Hirst JM (1952) An automatic volumetric spore trap. Ann Appl Biol 39(2):257–265
Iba´n˜ez I, Primack RB, Miller-Rushing AJ, Ellwood E, Higuchi H, Lee SD, Kobori H, Silander JA (2010)
Forecasting phenology under global warming. Philos Trans R Soc B Biol Sci 365(1555):3247–3260
IPPC (2013) Summary for policymarkers. In: Stovker TF, D Qin, Plattner GK, Tigor M, Allen SK,
Boschung J, Nauels A, Xia Y, Bex V (eds) Climate change 2013: the physical science basis. Con-
tribution of working group I to the fifth assessment report of the intergovernmental panel of climate
change. Cambridge University Press, United Kingdom and New York, NY, USA
Ja¨ger S (1998) Global aspects of ragweed in Europe. In: Spieksma FThM (ed) Ragweed in Europe. Satellite
symposium proceedings of 6th international congress of aerobiology, Perugia, Italy 1998. Alk-Abello ´
A/S, Horsholm, pp 6–10
Ja´rai-Komlo ´di M (1998) Ragweed in Hungary. In: Spieksma FThM (ed.) Ragweed in Europe. Satellite
symposium proceedings of 6th international congress of aerobiology, Perugia, Italy 1998. Alk-Abello ´
A/S, Horsholm, pp 33–38
Jato V, Rodrı ´guez-Rajo FJ, Seijo MC, Aira MJ (2009) Poaceae pollen in Galicia (N.W. Spain): charac-
terisation and recent trends in atmospheric pollen season. Int J Biometeorol 53(4):333–344
Kasprzyk I, Walanus A (2010) Description of the main Poaceae pollen season using bi-Gaussian curves, and
forecasting methods for the start and peak dates for this type of season in Rzeszo ´w and Ostrowiec S ´w.
(SE Poland). J Environ Monit 12(4):906–916
Kizilpinar I, Civelek E, Tuncer A, Dogan C, Karabulut E, Sahiner UM, Yavuz ST, Sackesen C (2011) Pollen
counts and their relationship to meteorological factors in Ankara, Turkey during 2005–2008. Int J
Biometeorol 55(4):623–631
Laaidi M (2001) Forecasting the start of the pollen season of Poaceæ: evaluation of some methods based on
meteorological factors. Int J Biometeorol 45(1):1–7
Makovcova ´S, Zlinka ´J, Mikola ´s V, Sala ´t D, Krio V (1998) Ragweed in Slovak Republic. In: Spieksma
FThM (ed) Ragweed in Europe. Satellite symposium proceedings of 6th international congress of
aerobiology, Perugia, Italy, 1998. AlkAbello ´A/S, Horsholm, pp 27–28
Menzel A, Sparks TH, Estrella N, Koch E, Aasa A, Ahas R, Alm-Ku ¨bler K, Bissolli P, Braslavska ´OG,
Briede A, Chmielewski FM, Crepinsek Z, Curnel Y, Dahl A ˚, Defila C, Donnelly A, Filella Y, Jatczak
K, Ma ˚ge F, Mestre A, Nordli Ø, PeN ˜Uelas J, Pirinen P, Remis ˇova´V, Scheifinger H, Striz M, Susnik A,
Van Vliet AJH, Wielgolaski F-E, Zach S, Zust ANA (2006) European phenological response to climate
change matches the warming pattern. Glob Change Biol 12(10):1969–1976
Miao L, Fraser R, Sun Z, Sneath D, He B, Cui X (2016) Climate impact on vegetation and animal husbandry
on the Mongolian plateau: a comparative analysis. Nat Hazards 80(2):727–739
Moreno-Sarmiento M, Pen ˜alba MC, Belmonte J, Rosas I, Ortega-Nieblas MM, Villa-Ibarra M, Lares-Villa
F, Pizano-Nazara LJ (2016) Urban airborne pollen in a semiarid environment. Aerobiologia
32(2):363–370. doi: 10.1007/s10453-015-9401-7
Mun˜oz-Rodrı ´guez A, Silva-Palacios I, Tormo-Molina R (2010) Influence of meteorological parameters in
hourly patterns of grass ( Poaceae ) pollen concentrations. Ann Agric Environ Med 17(1):87–100
Myszkowska D (2014) Poaceae pollen in the air depending on the thermal conditions. Int J Biometeorol
58(5):975–986
Nilsson S, Persson S (1981) Tree pollen spectra in the stockholm region (sweden), 1973–1980. Grana
20(3):179–182
Peel RG, Kennedy R, Smith M, Hertel O (2014) Do urban canyons influence street level grass pollen
concentrations? Int J Biometeorol 58(6):1317–1325
Peeters AG (1998) Ragweed in Switzerland. In: Spieksma FThM (ed) Ragweed in Europe. Satellite sym-
posium proceedings of 6th international congress of aerobiology, Perugia, Italy 1998. Alk-Abello ´A/S,
Horsholm, pp 16–19
Piotrowska K (2012) Forecasting the Poaceae pollen season in eastern Poland. Grana 51(4):263–269Nat Hazards (2016) 84:121–137 135
123

Puc M, Wolski T (2013) Forecasting of the selected features of Poaceae (R. Br.) Barnh., Artemisia L. and
Ambrosia L. pollen season in Szczecin, North-Western Poland, using Gumbel’s distribution. Ann
Agric Environ Med 20(1):36–47
Ramos MC, Martı ´nez-Casasnovas JA (2015) Climate change influence on runoff and soil losses in a rainfed
basin with Mediterranean climate. Nat Hazards 78(2):1065–1089
Recio M, Docampo S, Garcı ´a-Sa´nchez J, Trigo MM, Melgar M, Cabezudo B (2010) Influence of temper-
ature, rainfall and wind trends on grass pollination in Malaga (western Mediterranean coast). Agric For
Meteorol 150(7–8):931–940
Rivera JP, Verrelst J, Go ´mez-Dans J, Mun ˜oz-Marı ´J, Moreno J, Camps-Valls G (2015) An emulator toolbox
to approximate radiative transfer models with statistical learning. Remote Sens 7(7):9347–9370
Rodrı ´guez-Rajo FJ, Astray G, Ferreiro-Lage JA, Aira MJ, Jato-Rodriguez MV, Mejuto JC (2010) Evaluation
of atmospheric Poaceae pollen concentration using a neural network applied to a coastal Atlantic
climate region. Neural Netw 23(3):419–425
Rojo J, Rapp A, Lara B, Sabariego S, Ferna ´ndez-Gonza ´lez F, Pe ´rez-Badia R (2016) Characterisation of the
airborne pollen spectrum in Guadalajara (central Spain) and estimation of the potential allergy risk.
Environ Monit Assess 188(3):1–13
Sabariego S, Cuesta P, Ferna ´ndez-Gonza ´lez F, Pe ´rez-Badia R (2012) Models for forecasting airborne
Cupressaceae pollen levels in central Spain. Int J Biometeorol 56(2):253–258
Sa´nchez-Mesa JA, Serrano P, Carin ˜anos P, Prieto-Baena JC, Moreno C, Guerra F, Galan C (2005) Pollen
allerqy in Cordoba city: frequency of sensitization and relation with antihistamine sales. J InvestigAllergol Clin Immunol 15(1):50–56
Schmidheiny K, Suedekum J (2015) The pan-European population distribution across consistently defined
functional urban areas. Econ Lett 133:10–13
Senyuva HZ, Gilbert J, Silici S, Charlton A, Dal C, Gu ¨rel N, Cimen D (2009) Profiling turkish honeys to
determine authenticity using physical and chemical characteristics. J Agric Food Chem
57(9):3911–3919
Silva-Palacios I, Tormo-Molina R, Mun ˜oz-Rodrı ´guez A (2007) The importance of interactions between
meteorological conditions when interpreting their effect on the dispersal of pollen from homoge-
neously distributed sources. Aerobiologia 23(1):17–26
Silva-Palacios I, Ferna ´ndez-Rodrı ´guez S, Dura ´n-Barroso P, Tormo-Molina R, Maya-Manzano JM, Gonzalo-
Garijo A ´(2015) Temporal modelling and forecasting of the airborne pollen of Cupressaceae on the
southwestern Iberian Peninsula. Int J Biometeorol 60(2):297–306
Smith M, Emberlin J (2005) Constructing a 7-day ahead forecast model for grass pollen at north London,
United Kingdom. Clin Exp Allergy 35(10):1400–1406
Smith M, Emberlin J (2006) A 30-day-ahead forecast model for grass pollen in north London, United
Kingdom. Int J Biometeorol 50(4):233–242
Smith M, Emberlin J, Stach A, Rantio-Lehtima ¨ki A, Caulton E, Thibaudon M, Sindt C, Ja ¨ger S, Gehrig R,
Frenguelli G, Jato V, Rajo FJR, Alca ´zar P, Gala ´n C (2009) Influence of the North Atlantic Oscillation
on grass pollen counts in Europe. Aerobiologia 25(4):321–332
Smith M, Ja ¨ger S, Berger U, S ˇikoparija B, Hallsdottir M, Sauliene I, Bergmann KC, Pashley CH, De Weger
L, Majkowska-Wojciechowska B, Rybnı ´cˇ
ek O, Thibaudon M, Gehrig R, Bonini M, Yankova R,
Damialis A, Vokou D, Gutie ´rrez Bustillo AM, Hoffmann-Sommergruber K, Van Ree R (2014)
Geographic and temporal variations in pollen exposure across Europe. Allergy Eur J Allergy Clin
Immunol 69(7):913–923
Srivastava A, Singh M, Jain VK (2012) Identification and characterization of size-segregated bioaerosols at
Jawaharlal Nehru University, New Delhi. Nat Hazards 60(2):485–499
Stach A, Garcia-Mozo H, Prieto-Baena JC, Czarnecka-Operacz M, Jenerowicz D, Silny W, Galan C (2007)
Prevalence of Artemisia species pollinosis in western Poland: impact of climate change on aerobio-
logical trends, 1995–2004. J Investig Allergol Clin Immunol 17(1):39–47
Stach A, Smith M, Prieto Baena JC, Emberlin J (2008) Long-term and short-term forecast models for
Poaceae (grass) pollen in Poznan ´, Poland, constructed using regression analysis. Environ Exp Bot
62(3):323–332
Tassan-Mazzocco F, Felluga A, Verardo P (2015) Prediction of wind-carried Gramineae and Urticaceae
pollen occurrence in the Friuli Venezia Giulia region (Italy). Aerobiologia 31(4):559–574
Thibaudon M (1998) (1998): Ragweed in France. In Lyon, France, 1987–1997. In: Spieksma FThM (ed)
Ragweed in Europe. Satellite symposium proceedings of 6th international congress of aerobiology,
Perugia, Italy 1998. Alk-Abello ´A/S, Horsholm, p 15
Tormo-Molina R, Gonzalo-Garijo MA, Silva-Palacios I, Mun ˜oz-Rodrı ´guez AF (2010) General trends in
airborne pollen production and pollination periods at a mediterranean site (Badajoz, southwest spain).
J Investig Allergol Clin Immunol 20(7):567–574136 Nat Hazards (2016) 84:121–137
123

Tormo-Molina R, Silva-Palacios I, Gonzalo-Garijo A, Moreno-Corchero A, Pe ´rez-Caldero ´n R, Ferna ´ndez-
Rodrı ´guez S (2011) Phenological records as a complement to aerobiological data. Int J Biometeorol
55(1):51–65
Tormo-Molina R, Maya-Manzano JM, Silva-Palacios I, Ferna ´ndez-Rodrı ´guez S, Gonzalo-Garijo A ´(2015)
Flower production and phenology in Dactylis glomerata. Aerobiologia 31(4):469–479. doi: 10.1007/
s10453-015-9381-7
Ugolotti M, Pasquarella C, Vitali P, Smith M, Albertini R (2015) Characteristics and trends of selected
pollen seasons recorded in Parma (Northern Italy) from 1994 to 2011. Aerobiologia 31(3):341–352.
doi:10.1007/s10453-015-9368-4
Vesterberg O (2001) Airborne birch and grass pollen allergens in street-level shops. Indoor Air
11(4):241–245
Voukantsis D, Niska H, Karatzas K, Riga M, Damialis A, Vokou D (2010) Forecasting daily pollen
concentrations using data-driven modeling methods in Thessaloniki, Greece. Atmos Environ
44(39):5101–5111
Vrugt JA, Gupta HV, Bouten W, Sorooshian S (2003) A shuffled complex evolution metropolis algorithm
for optimization and uncertainty assessment of hydrologic model parameters. Water Resour Res
39(8):SWC11–SWC116
Wang Q, Nakamura S, Lu S, Nakajima D, Suzuki M, Sekiguchi K, Miwa M (2013) Diurnal and nocturnal
behaviour of airborne Cryptomeria japonica pollen grains and the allergenic species in urban atmo-
sphere of Saitama, Japan. Asian J Atmos Environ 7(2):65–71
Wielgolaski FE (2001) Phenological modifications in plants by various edaphic factors. Int J Biometeorol
45(4):196–202
Zhang Y, Isukapalli S, Bielory L, Georgopoulos P (2013) Bayesian analysis of climate change effects on
observed and projected airborne levels of birch pollen. Atmos Environ (Oxford, England: 1994)
68:64–73
Zhang Y, Bielory L, Cai T, Mi Z, Georgopoulos P (2015) Predicting onset and duration of airborne
allergenic pollen season in the United States. Atmos Environ 103:297–306Nat Hazards (2016) 84:121–137 137
123