Multi-Input Convolutional Neural Networks for Automatic Pollen Classification

Pollen allergies are a cause of much suffering for an increasing number of individuals. Current pollen monitoring techniques are lacking due to their reliance on manual counting and classification of pollen by human technicians. In this study, we present a neural network architecture capable of distinguishing pollen species using data from an automated particle measurement device. This work presents an improvement over the current state of the art in the task of automated pollen classification, using fluorescence spectrum data of aerosol particles. We obtained a relative reduction in the error rate of over 48%, from 27% to 14%, for one of the datasets, with similar improvements for the other analyzed datasets. We also use a novel approach for doing hyperparameter tuning for multiple input networks.

Development and application of a method to classify airborne pollen taxa concentration using light scattering data

Although automated pollen monitoring networks using laser optics are well-established in Japan, it is thought that these methods cannot distinguish between pollen counts when evaluating various pollen taxa. However, a method for distinguishing the pollen counts of two pollen taxa was recently developed. In this study, we applied such a method to field evaluate the data of the two main allergens in Japan, Chamaecyparis obtusa and Cryptomeria japonica. We showed that the method can distinguish between the pollen counts of these two species even when they are simultaneously present in the atmosphere. This result indicates that a method for automated and simple two pollen taxa monitoring with high spatial density can be developed using the existing pollen network.

Earlier Flowering of Betula pendula Roth in Augsburg, Germany, Due to Higher Temperature, NO2 and Urbanity and Relationship with Betula spp. Pollen Season

Flowering and pollen seasons are sensitive to environmental variability and are considered climate change indicators. However, it has not been concluded to what extent flowering phenology is indeed reflected in airborne pollen season locally. The aim of this study was to investigate, for the commonly represented in temperate climates and with highly allergenic pollen Betula pendula Roth, the responsiveness of flowering to different environmental regimes and also to check for commensurate changes in the respective pollen seasons. The region of Augsburg, Bavaria, Germany, was initially screened for birch trees, which were geolocated at a radius of 25 km. Random trees across the city were then investigated during three full flowering years, 2015–2017. Flowering observations were made 3–7 times a week, from flower differentiation to flower desiccation, in a total of 43 plant individuals. Data were regressed against meteorological parameters and air pollutant levels in an attempt to identify the driving factors of flowering onset and offset. Flowering dates were compared with dates of the related airborne pollen seasons per taxon; airborne pollen monitoring took place daily using a Hirst-type volumetric sampler. The salient finding was that flowering occurred earlier during warmer years; it also started earlier at locations with higher urbanity, and peaked and ended earlier at sites with higher NO2 concentrations. Airborne pollen season of Betula spp. frequently did not coincide locally with the flowering period of Betula pendula: while flowering and pollen season were synchronized particularly in their onset, local flowering phenology alone could explain only 57.3% of the pollen season variability. This raises questions about the relationship between flowering times and airborne pollen seasons and on the rather underestimated role of the long-distance transport of pollen.

‘Pollen potency’: the relationship between atmospheric pollen counts and allergen exposure

Pollen allergies are responsible for a considerable global public health burden, and understanding exposure is critical to addressing the health impacts. Atmospheric pollen counts are routinely used as a predictor of risk; however, immune responses are triggered by specific proteins known as allergens, which occur both within and on the surface of the pollen grain. The ratio between atmospheric pollen counts and allergen concentrations (‘pollen potency’) has been shown to be inconsistent, with potentially important implications for pollen monitoring practice. Despite this, there has been no previous synthesis of the literature and our understanding of the factors that influence pollen potency remains poor. We conducted a scoping review with the aim of deriving a current understanding of: (a) the factors that influence pollen potency; (b) its variation through time, between taxa and by location; and (c) the implications for pollen monitoring practice. Our synthesis found that pollen potency is highly variable within and between seasons, and between locations; however, much of this variability remains unexplained and has not been deeply investigated. We found no predictable pollen potency patterns relating to taxon, geography or time, and inconclusive evidence regarding possible driving factors. With respect to human health, the studies in our synthesis generally reported larger associations between atmospheric allergen loads and allergy symptoms than whole pollen counts. This suggests that pollen potency influences public health risk; however, the evidence base remains limited. Further research is needed to better understand both pollen potency variability and its implications for health.

Analysis of changes in Betula pollen season start including the cycle of pollen concentration in atmospheric air

Birch belongs to the most important allergenic taxa in Europe, therefore information on the start dates of the pollen season is very important for allergists and their patients as well as for climatologists. The study examined changes in the start of the birch pollen season as well as determined the trend of these changes. Pollen monitoring was performed in Lublin (eastern Poland) in the period 2001–2019 using the volumetric method. The Makra-test was used to detect periods with significantly higher or lower average of the onset than the average for the whole dataset. Two significant falls in the average of the pollen season start were found in 2007 and 2014. Besides, taking into account the 2-3-year rhythm of high and low concentrations of birch pollen in the atmospheric air, linear trends were fitted for the subsets of high and low abundance seasons. Significant changes in Betula pollen season start dates were only determined for the highly abundance seasons, while the results for seasons with a low concentration did not allow rejecting the hypothesis about the lack of a linear trend in the changes in the studied parameter. Moreover, a significant polynomial relationship was found between the beginning of a pollen season and the average values of monthly temperatures preceded a season. These analyses show that the start dates of the Betula pollen season are getting significantly earlier. The dynamics of changes differ between seasons with high and low concentrations of pollen.

Medium-Term Increases in Ambient Grass Pollen Between 1994-1999 and 2016-2020 in a Subtropical Climate Zone

Grass pollen is the major outdoor trigger of allergic respiratory diseases. Climate change is influencing pollen seasonality in Northern Hemisphere temperate regions, but many aspects of the effects on grass pollen remain unclear. Carbon dioxide and temperature rises could increase the distribution of subtropical grasses, however, medium term shifts in grass pollen in subtropical climates have not yet been analysed. This study investigates changes in grass pollen aerobiology in a subtropical city of Brisbane, Australia, between the two available monitoring periods, 1994-1999 and 2016-2020. Potential drivers of pollen change were examined including weather and satellite-derived vegetation indicators. The magnitude of the seasonal pollen index for grass showed almost a three-fold increase for 2016-2020 over 1994-1999. The number and proportion of high and extreme grass pollen days in the recent period increased compared to earlier monitoring. Statistically significant changes were also identified for distributions of CO2, satellite-derived seasonal vegetation health indices, and daily maximum temperatures, but not for minimum temperatures, daily rainfall, or seasonal fraction of green groundcover. Quarterly grass pollen levels were correlated with corresponding vegetation health indices, and with green groundcover fraction, suggesting that seasonal-scale plant health was higher in the latter period. The magnitude of grass pollen exposure in the subtropical region of Brisbane has increased markedly in the recent past, posing an increased environmental health threat. This study suggests the need for continuous pollen monitoring to track and respond to the possible effects of climate change on grass pollen loads.

Patterns in recent and Holocene pollen accumulation rates across Europe – the Pollen Monitoring Programme Database as a tool for vegetation reconstruction

The collection of modern, spatially extensive pollen data is important for the interpretation of fossil pollen assemblages and the reconstruction of past vegetation communities in space and time. Modern datasets are readily available for percentage data but lacking for pollen accumulation rates (PARs). Filling this gap has been the motivation of the pollen monitoring network, whose contributors monitored pollen deposition in modified Tauber traps for several years or decades across Europe. Here we present this monitoring dataset consisting of 351 trap locations with a total of 2742 annual samples covering the period from 1981 to 2017. This dataset shows that total PAR is influenced by forest cover and climate parameters, which determine pollen productivity and correlate with latitude. Treeless vegetation produced PAR values of at least 140 grains cm−2 yr−1. Tree PAR increased by at least 400 grains cm−2 yr−1 with each 10 % increase in forest cover. Pollen traps situated beyond 200 km of the distribution of a given tree species still collect occasional pollen grains of that species. The threshold of this long-distance transport differs for individual species and is generally below 60 grains cm−2 yr−1. Comparisons between modern and fossil PAR from the same regions show similar values. For temperate taxa, modern analogues for fossil PARs are generally found downslope or southward of the fossil sites. While we do not find modern situations comparable to fossil PAR values of some taxa (e.g. Corylus), CO2 fertilization and land use may cause high modern PARs that are not documented in the fossil record. The modern data are now publicly available in the Neotoma Paleoecology Database and aid interpretations of fossil PAR data.

Long-Term Pollen Monitoring in the Benelux: Evaluation of Allergenic Pollen Levels and Temporal Variations of Pollen Seasons

Airborne pollen is a major cause of allergic rhinitis, affecting between 10 and 30% of the population in Belgium, the Netherlands, and Luxembourg (Benelux). Allergenic pollen is produced by wind pollinating plants and released in relatively low to massive amounts. Current climate changes, in combination with increasing urbanization, are likely to affect the presence of airborne allergenic pollen with respect to exposure intensity, timing as well as duration. Detailed analysis of long-term temporal trends at supranational scale may provide more comprehensive insight into these phenomena. To this end, the Spearman correlation was used to statistically compare the temporal trends in airborne pollen concentration monitored at the aerobiological stations which gathered the longest time-series (30–44 years) in the Benelux with a focus on the allergenic pollen taxa: Alnus, Corylus, Betula, Fraxinus, Quercus, Platanus, Poaceae, and Artemisia. Most arboreal species showed an overall trend toward an increase in the annual pollen integral and peak values and an overall trend toward an earlier start and end of the pollen season, which for Betula resulted in a significant decrease in season length. For the herbaceous species (Poaceae and Artemisia), the annual pollen integral and peak values showed a decreasing trend. The season timing of Poaceae showed a trend toward earlier starts and longer seasons in all locations. In all, these results show that temporal variations in pollen levels almost always follow a common trend in the Benelux, suggesting a similar force of climate change-driven factors, especially for Betula where a clear positive correlation was found between changes in temperature and pollen release over time. However, some trends were more local-specific indicating the influence of other environmental factors, e.g., the increasing urbanization in the surroundings of these monitoring locations. The dynamics in the observed trends can impact allergic patients by increasing the severity of symptoms, upsetting the habit of timing of the season, complicating diagnosis due to overlapping pollen seasons and the emergence of new symptoms due allergens that were weak at first.