Influencia de la meteorología en las concentraciones polínicas en la atmósfera de Cartagenaaplicación de modelos estadísticos predictivos

  1. María Dolores Galera Martínez
Supervised by:
  1. José María Moreno Grau Director
  2. Luis Negral Director

Defence university: Universidad Politécnica de Cartagena

Fecha de defensa: 20 March 2023

  1. Stella Moreno Grau Chair
  2. Rosa María Pérez Badia Secretary
  3. Carmen Noemí Zoroa Alonso Committee member

Type: Thesis


This doctoral dissertation has been presented in the form of thesis by publication. A study has been carried out on pollen concentrations in the atmosphere of the city of Cartagena (Spain), located on the south-eastern Mediterranean coast of the Iberian Peninsula, in the Region of Murcia. Although 62 pollen types have been studied and classified in the aerobiological database for the period 1993-2016 (24 years), this study focuses on the flowering of Olea. This pollen type has the highest prevalence of skin sensitisations and is the main cause of allergies in the Region of Murcia, where its annual pollen index exceeds 500 grains/m3. The aerobiological samples have been processed following Hirst’s methodology and developed by the Spanish Network of Aerobiology. The aerobiological database has been complemented with meteorological data provided by AEMET (Spanish Meteorological Agency). The characteristics of the Main Pollen Season (MPS), the behaviour of the meteorological variables and the threshold values for the Olea pollen type in the city of Cartagena have been defined from the databases. The evolution of the Olea_MPS and Olea_peak indices have been described over the 24 years analysed, together with the meteorological variables. The Olea pollen concentration series has been used to characterise the MPS. From the characterisation of the Olea MPS, a longer MPS and an increased amount of Olea pollen grains have been observed both in the MPS and on the peak day. Accordingly, the categories Olea pollen exposure have been adjusted to improve preventive measures. A predictive model has been developed to evaluate the intensity of flowering and the characterisation of the MPS in advance. One of the objectives has been to estimate the two pollen indices for the next pollen season at least one month in advance. Additional estimates are also needed during the MPS to improve accuracy. Estimates of Olea pollen in its MPS and on its peak day have been obtained using regression methods and climate factors. The analysis reveals that the Bagging for Regression Trees (BRT) method is a good predictive alternative since it also establishes the importance of each meteorological variable. Olea Pollen concentrations have been studied in relation to the typology of air masses, pollen grain sources and the marine nature of air masses during advections in the coastal enclave of Cartagena. Given that Spain is the world’s leading olive producer and olive cultivation extends throughout the Mediterranean basin, this location is ideal for studying long-distance transport events (LDT). Air masses have been classified using the back trajectory calculation for 48 h at 250, 500 and 750 m above ground level using the HYSPLIT model. Once classified, the frequency of LDT events from Africa and Europe was 8.7 % of the MPS. In contrast, regional air masses have been found in 38.6 % of the MPS. Accordingly, significantly higher pollen concentrations (p-value < 0.05) have been detected on days with regional air masses than on days with air masses originating in Europe. The importance of air masses coming from nearby regions with dense olive groves, such as Andalusia, has been confirmed. The joint review of air masses, source areas and pollen concentrations has allowed the peak day dates of Olea grain concentration and the LDT associated with these dates to be contextualised. The distortions caused by back trajectories have explained alterations in the pollen concentrations on consecutive days. Recirculation and air mass back trajectory loops have modified the pollen load that each type of air mass originally contained. The study was carried out by applying a Linear Model Analysis of the Variance of one Factor (ANOVA I) to the aerobiological database covering the period 1993-2014. Meteorological data were added and information on the origin of air masses and Olea pollen sources was incorporated. Finally, the study has investigated whether earthquakes occurring between 2010 and 2019 had an impact on the pollen concentrations in ambient air monitored by the Aerobiological Network of the Region of Murcia (REAREMUR). To this end, atmospheric pollen concentrations have been studied coinciding with earthquakes of magnitudes up to Mw = 5.1 and intensities up to grade VII on the European Macroseismic Scale (EMS). Olea pollen concentrations and the total pollen recorded during the Olea_MPS have been studied in the three cities of REAREMUR: Cartagena, Lorca and Murcia. Earthquakes were detected on 100 of the 4,457 days of the Olea_MPS in the three cities: 12 of the 1,535 days in Cartagena, 49 of the 1,481 days in Lorca and 39 of the 1,441 days in Murcia. The maximum-intensity earthquake (EMS grade VII) during the period under study was in Lorca, on May 11th 2011. It did not cause an immediate impact on pollen concentrations in the city, possibly because an African dust outbreak also occurred on this day. However, Olea concentrations were statistically higher on days when an earthquake occurred than on days when no earthquake was felt in Lorca (p-value < 0.05). Although higher pollen concentrations associated with increased seismic intensity have been found, the limitations of the time series and the relative scarcity of seismic events force us to take these findings with caution. Therefore, they should be ratified by collecting more data to extend the time series and by studying other sites with seismic activity. Statistical analyses of pollen concentrations and meteorological and seismic variables have highlighted the preponderance of phenology over seismic movement in air pollen concentrations. In addition, African dust outbreaks have clearly been associated with lower pollen concentrations of Olea and all the pollen types in the Olea_MPS in the three cities of REAREMUR