Aristotle University Of Thessaloniki (AUTH)
Validation of an air quality numerical model and development of a post-processor filter
Abstract
dc:descriptionThe subject of this dissertation is the performance evaluation of an air quality numerical model and the development of a post-processing filter that improves model forecasts. The region of study is Southeastern Europe and emphasis will be given to Eastern Mediterranean. The purpose of the first part of this study is to examine the ability of two modelling systems to reproduce the influence of Etesian winds on near-surface ozone levels over Eastern Mediterranean. For this, the global MACC (Monitoring Atmospheric Composition and Climate) reanalysis model IFS - MOZART (Integrated Forecast System - Model for Ozone and Related chemical Tracer) and an ensemble of regional air quality models MACC-II (MACC-Interim Implementation) are used. The evaluation is carried out for three background stations located in Malta, Cyprus and Crete for a period of five months (May - September) for years 2011-2012. The results pointed out that although both modelling systems underestimate ozone concentrations systematically, they can capture to a certain extend the impact of the Etesian on the ozone zonal distribution with the ensemble of regional models better reproducing ozone levels than global model. In the second part of this study, the performance of a high spatial and temporal resolution photochemical model, i.e., the CAMx (Comprehensive Air Quality Model with Extensions) which is coupled with the WRF (Weather Research and Forecasting) meteorological model is examined. Simulations of gaseous pollutants (O3, NO2, NO, CO, SO2) and particulate matter are compared with air quality data from 28 monitoring stations over Greece. The analysis is done for one month of every season of the year 2012 (i.e., January, April, July and October). The results suggest that model performance is satisfactory for particulate matter and gaseous pollutants for almost all stations. However, the model performance is less satisfactory for SO2. The model overestimates ozone concentrations at several stations resulting in a slight overall overestimation with the exception of July. WRF-CAMx simulations are able to reproduce ozone diurnal cycle quite well. For the other gas phase pollutants, the model shows a small underestimation of concentration levels. Statistical metrics for PM10 show a slight overestimation during January and April and an underestimation during the other two months. Similar results are calculated for PM2.5. The aim of the last study is to improve the performance of the WRF-CAMx modelling system in simulating pollution levels in Greece. For this purpose, a post-processing filter was developed in the Laboratory of Atmospheric Physics, Department of Physics, University of Patras that uses the Analog Ensemble (AnEn) methodology. The filter is applied on WRF-CAMx ozone and particulate matter simulations. AnEn searches for analogs in past forecasts to correct the present forecast. The results are promising, suggesting the successful development and implementation of the filter. The model performance is considerably improved and the corrected forecasts are closer to observations one. More specifically, there is a clear improvement in the correlation between corrected forecasts and observations.
Degree
thesis:*- Grantor dc:publisher
- Aristotle University Of Thessaloniki (AUTH)
- Year dc:date
- 2021
Author and committee
dc:creator, dc:contributor.*- Authors dc:creator
-
- Solomou, Ermioni
- Σολωμού, Ερμιόνη
Subjects
dc:subject × 16- Αριθμητικό μοντέλο ρύπανσης
- Αξιολόγηση μοντέλου
- Numerical air pollution model
- Model evaluation
- Air pollution forecasting
- Filter
- Χημεία
- Φυσικές Επιστήμες
- Χημεία, γενικά
- Γεωεπιστήμες και Επιστήμες Περιβάλλοντος
- Περιβαλλοντική χημεία
- Chemical Sciences
- Natural Sciences
- Chemistry, general
- Earth and Related Environmental Sciences
- Environmental Chemistry
Rights
dc:rights- Statement dc:rights
-
- BY_NC_ND
- Language dc:language
- eng
Identifiers
dc:identifier.*- Identifier
- 10.12681/eadd/49120