Innovative methods to retrieve PM concentrations using high spatiotemporal resolution data

• Κύριος συγγραφέας: Κοσμόπουλος, Γεώργιος
• Άλλοι συγγραφείς: Kosmopoulos, Georgis
• Γλώσσα: English
• Έκδοση: 2023
• Θέματα: Particulate matter; Low-cost sensors; Αισθητήρες χαμηλού κόστους; Αιωρούμενα σωματίδια

The aim of this thesis focuses on the investigation of new and innovative techniques to quantify Particulate Matter (PM) levels and properties in various environments. Since PM is an ubiquitous constituent of the atmosphere, accurate measurements in high spatial and temporal resolution is fundamental. PM mass concentrations and AOD values are examined., using low – cost systems established in various Greek cities. Moreover, particle number concentrations were also quantified in three different areas in Greece and Spain. It is fundamental to investigate the basic PM properties and processes in the atmosphere and retrieve valuable information about aerosols size distribution along with their effect on human health and climate. PM-related pollution levels in Europe, Mediterranean Basin and Greece along wint mapping of PM sources and hotspots is still insufficient and innovative methods are necessary to fill PM knowledge gaps. During the last years several innovative, methodologies and instrumentations have been developed to quantify PM mass and number concentrations. A Grimm analyzer and PAir low – cost sensors were used to measure PM concentrations, while AOD was estimated using a custom-made real-time sampler (AMODv2). A thorough evaluation of the performance and precision of low – cost PAir sensors compared to a reference instrument (Grimm) is crusial. Consequently, a 18-month measurement campaign was conducted in Patras ambient environment for varying atmospheric (PM2.5 ranged from 0.1 to 95 μg m−3) and meteorological (temperature ranged from 3 to 39 °C and RH from 9 to 97%) conditions. The low - cost sensors showed high intraprecision with intra units discrepancies lower than 1.47 μg m−3. Adequate correlation was found among the raw low-cost sensors readings and the reference instrument. Moreover, increased biases were found during periods that coarse PM were dominant. Hence, the measurements during that periods were excluded from the analysis. For the remaining dataset, a calibration factor was calculated using a linear regression model against the collocated reference instrument. Implementation of the calibration factor reduced the relative root mean square error to 24.5% and 24.6% for hourly PM1 and PM2.5 measurements. The impact of relative humidity was also examined but the effect was minimal. Low – cost sensors constitute a great tool to to deploy dense air quality monitoring networks ore supplement existing ones. Sixteen sensors were installed in the greater Patras area and the results of the 3-year measurements are presented. Beneficial insights about the spatiotemporal variations across the city were extracted. For the first time such a dense network was deployed in the city providing representative data for all the examined areas. This is crucial to obtain valuable information on citizens exposure to PM air pollution. A season variation was identified in most of the measurement locations, with higher concentrations during the winter and lower during summer, indicating the impact of biomass burning emissions related with domestic heating systems. Additionally, increased concentrations were found to the areas that are closer to the urban core compared to suburban areas. An interesting finding was that the area located near restaurants showed the most impaired air quality conditions, suggesting the effect of cooking emission. The contribution of local and regional sources was also quantified. The local sources represented only the 30% of total PM2.5 concentrations within the urban core while the rest was transported from other areas. The fraction of local sources also showed a seasonal variation with lower values from April to October (18%) and higher from November to March (52%). The capabilities of such dense monitoring networks allow us to investigate PM2.5 concentrations along urban environmnetns in neigborhood scale. Air quality conditions in three Greek cities were also studied, with the use of PAir sensors. Measurements were conducted in Kalamata, Thermi and Agrinio, during different time periods and the spatial and temporal fluctuations are discussed. Exceedances of the regulated by WHO PM2.5 daily limits were recorded at many measurement sites. The examination of data collected indicated that seasonal and spatial variations are related the characteristics of the measurement sites. Air pollution levels were more pronounced during winter due to heating emissions. Differentiations were reported between neighboring sites, indicating the necessity to implement the existing monitoring sites for a more accurate PM mapping. Particles number concentration measurements were conducted within urban suburban and semi - arid environments in Greece and Spain. The investigation low-cost optical counters efficiency and response for particles with different diameters (0.3 – 10 μm) reported ambiguous results. The sensors size selectivity was affected by particles source and size. At the urban and background areas (Patras, Greece), PAir sensors reported NC for particles with diameter 0.3 – 1 μm with a good agreement with the reference instrument (Grimm) while a systematic overestimation was observed for coarser particles (>1 μm) though. At the semi-arid area (Almeria, Spain), biased number concentration measurements were reported. The increased coarse particles concentrations in this location biased the observations. The number concentrations reported by PAir were lower than the Grimm ones. Overall, the evaluation camping showed that the sensor was insensitive to number concentration fluctuations during periods when coarse particles are dominant. Aerosol optical properties can be derived from ground-based or satellite observations. The first ones provide high quality data, however, their spatial resolution is sparse, while the latter have sufficient spatial resolution but increased uncertainties. Thus, there is still demand for instruments that could supplement current monitoring methods. The performance of a custom-made sensor of PM2.5 and AOD was examined. A field campaign was conducted in Patras and Thessaloniki, to explore the capabilities of the low-cost AMODv2 sampler. The sensor is able to provide simultaneous PM2.5 and AOD measurements at fine temporal scale; 20 minutes time interval. The unit was collocated with reference instrument and promising results were shown. AOD measurements across four wavelengths, 440, 500, 675 and 870 nm, reported good agreement with the reference instruments (AERONET and Microtops) at both sites. PM2.5 measurements are derived from an integrated PMS 5003 sensor. The findings from the validation campaign showed close agreement between the sensor and identical sensors built in different enclosure. The results in general underline the AMODv2 capabilities to provide simultaneous AOD and PM2.5 data in high temporal resolution.