Back to search

Aristotle University Of Thessaloniki (AUTH)

Development of a modeling system to study and forecast biological weather

Abstract

dc:description

The scope of the present Doctoral Dissertation is the enhancement and implementation of a fine-scale modeling system, with a spatial resolution of 2 km, for the prediction of daily pollen concentrations of selected taxa in the broader area of Thessaloniki, Greece. The selected taxa are abundant in the Mediterranean region. Οlive (Olea europaea L.) is a primary source of airborne allergenic pollen, while Quercus and Pinaceae contribute substantial quantities of pollen grains to the atmosphere.In recent decades, the prevalence of pollen allergies has increased markedly, particularly in urban populations and among children. The high allergenicity of some tree species can negatively affect pollen allergy sufferers during their flowering period. In the first stage of this research, the Urban Green Zone Allergenicity Index (IUGZA) was implemented to characterize the potential allergenicity of the street trees most commonly used as ornamentals in Thessaloniki. The index is based on allergenic potential, pollen emissions, principal pollination period and surface area covered by each tree. Two case studies were conducted in the city center: the block of streets surrounding Agia Sofia church (AS) and the junction of Agiou Dimitriou (AD) and Ethnikis Aminis (EA). Results revealed that in the AS case, 38% of the species displayed high allergenic potential or are recognized as major local allergens, compared to only 8% in AD-EA. The lack of female trees was evident in both cases (1% and 6% respectively), although such trees are generally classified as non-allergenic. In AS, the allergenic potential of the street trees reached IUGZA = 0.56, exceeding the critical threshold of 0.5, with Fraxinus, Ulmus and Platanus being the most important contributors. In AD-EA, the index was significantly lower (IUGZA = 0.12), with Cedrus and Platanus as the dominant contributors. Accurate prediction of the Main Pollen Season (MPS), i.e. the time period within which anemophilous pollen is released, is crucial for both public health and environmental management. In the second stage, a comparative evaluation of multiple predictive models for estimating MPS in Thessaloniki was carried out for the period 2016-2022. The predictive approaches were divided into two categories: (i) methods based on tree phenology and meteorological variables, and (ii) data-driven and statistical methods. The first category included Thermal Time (TT) models, which integrate chilling and heating requirements, as well as Partial Least Squares Regression (PLS) and Temperature-Photoperiod (TP) models. Predictive performance was generally reliable, with a combination of TT and PLS methods providing the most reliable forecasts. PLS regression analysis highlighted the strong influence of spring and preceding autumn temperatures on the MPS. The chilling period for both taxa typically began in late November to early December and ended around early January, with an average requirement of approximately 1187 chilling hours. Heating requirements varied substantially, from 435 oC to 2100 oC depending on the taxon and model used.The second category of models included cumulative temperature-based approaches (DT), Logistic Models (LM), Distribution Methods (DM) and Machine Learning Techniques (MLTs) such as Random Forest (RF), Neural Networks and Ensemble Learning. Results showed that DT and LM models effectively captured the end of the pollen season, with differences from observed values ranging from 0 to 7 days. In contrast, MLTs, particularly RF, exhibited high accuracy in predicting the onset of the season, with deviations ranging from 0 to 10 days. The DT approach, which incorporates transition ranges, further enhanced prediction reliability. Based on these evaluations, specific predictive approaches were selected for integration into the final modeling framework: the DT temperature sum method and TT model (depending on the taxon) for predicting the onset of MPS and the LM approach for determining its end. The final objective of the dissertation was the implementation of an enhanced high-resolution modeling system WRF-NEMO-CAMx, to predict airborne pollen concentrations for Quercus, Pinaceae and Olea. This was the first time such a system was applied in Greece at a spatial resolution of 2 km. Three emission scenarios (“Base”, “Max” and “Min”) were implemented, using literature-derived pollen production rates. The system was evaluated against ground-based meteorological observations and daily pollen concentration measurements for the year 2019. The evaluation of predicted pollen concentrations confirmed that the system captured daily variability satisfactorily. Among the three emission scenarios, the “Min” simulation yielded the best agreement for Quercus (IOA = 0.81) and Pinaceae (IOA = 0.76), while for Olea, the “Max” scenario achieved superior performance (IOA = 0.89). Subcategory analyses revealed that the QRPR group was the dominant contributor to Quercus concentrations, while PMSC accounted for the largest share among Pinaceae subgroups. Spatial emission maps showed higher values in areas with dense vegetation for Quercus and Pinaceae, and in coastal and agricultural zones for Olea, consistent with known species distributions. These findings collectively demonstrate the feasibility and accuracy of a fine-scale pollen modeling system capable of simulating airborne concentrations for highly allergenic and abundant taxa. The study highlights the potential for integrating such systems into health-warning frameworks, complementing thermal stress and air pollution forecasts with pollen exposure predictions. This dissertation thus contributes to the development of biological weather forecasting, with important implications for public health and environmental management.

Degree

thesis:*
Grantor dc:publisher
Aristotle University Of Thessaloniki (AUTH)
Year dc:date
2025

Author and committee

dc:creator, dc:contributor.*
Authors dc:creator
  • Papadogiannaki, Sofia
  • Παπαδογιαννάκη, Σοφία

Subjects

dc:subject × 13

Rights

Language dc:language
eng

Identifiers

dc:identifier.*
Identifier
10.12681/eadd/60160

Chain of custody

source
Harvested from
Greek National Archive of PhD Theses
Base URL
phdtheses.ekt.gr/eadd_oai/request
Last updated
2026-11-18
Source record
OAI-PMH GetRecord
citation

Papadogiannaki, Sofia; Παπαδογιαννάκη, Σοφία. Development of a modeling system to study and forecast biological weather. Aristotle University Of Thessaloniki (AUTH), 2025. http://hdl.handle.net/10442/hedi/60160