Air pollution and high concentrations of fine particulate matter (PM) are extremely harmful to human health, ecosystems and the climate in many countries around the world. To solve this type of air quality problems, a huge number of studies has been devoted. It is important to note that pollution sources have a local character for each urban region and depend on a large number of factors such as meteorological, manufacturing, domestic, transport and other. In the field of mathematical modeling and processing of accumulated measured data for PM with aerodynamic size up to 10 microns (PM10), along with classical statistical methods, powerful, flexible data mining techniques and approaches such as neural networks, fuzzy logic, regression trees, multivariate adaptive regression splines, random forest (RF), and more are adapted. This study explores the possibilities of the random forest method for modeling the concentrations of PM10 in Blagoevgrad, Bulgaria. Average daily data over 9 years (2009-2018) and a large number of input variables as predictors are used – one lagged PM10, meteorological and temporal. The forecasting process is performed by a multistep procedure repeated five times with 3-day horizon ahead using data not included in the construction of the model. The constructed models show high-performance both in fitting and forecasting the measured data.
Numerous time series methods are used in environmental sciences allowing the detailed investigation of air pollution processes. The goal of this study is to present the empirical analysis of various aspects of stochastic modeling and in particular the ARIMA/SARIMA methods. The subject of investigation is air pollution in the town of Kardzhali, Bulgaria with 2 problematic pollutants – sulfur dioxide (SO2) and particulate matter (PM10). Various SARIMA Transfer Function models are built taking into account meteorological factors, data transformations and the use of different horizons selected to predict future levels of concentrations of the pollutants.
Fine particulate matter (PM10) air pollution is a serious problem affecting the health of the population in many Bulgarian cities. As an example, the object of this study is the pollution with PM10 of the town of Pleven, Northern Bulgaria. The measured concentrations of this air pollutant for this city consistently exceeded the permissible limits set by European and national legislation. Based on data for the last 6 years (2011-2016), the analysis shows that this applies both to the daily limit of 50 micrograms per cubic meter and the allowable number of daily concentration exceedances to 35 per year. Also, the average annual concentration of PM10 exceeded the prescribed norm of no more than 40 micrograms per cubic meter. The aim of this work is to build high performance mathematical models for effective prediction and forecasting the level of PM10 pollution. The study was conducted with the powerful flexible data mining technique Classification and Regression Trees (CART). The values of PM10 were fitted with respect to meteorological data such as maximum and minimum air temperature, relative humidity, wind speed and direction and others, as well as with time and autoregressive variables. As a result the obtained CART models demonstrate high predictive ability and fit the actual data with up to 80%. The best models were applied for forecasting the level pollution for 3 to 7 days ahead. An interpretation of the modeling results is presented.
Air pollution is one of the key problems in urban areas and its investigation is vital both for people's health and for the environment as a whole. In particular, ground ozone is a secondary air pollutant with concentrations dependent mainly on changes in the levels of other pollutants and meteorological conditions within a given region. This paper presents a statistical study based on multivariate analysis of hourly data on 9 air pollutants and 6 meteorological variables in the town of Dimitrovgrad, Bulgaria over a period of 7 years and 3 months. Yeo-Johnson power transformation is applied to each air pollutant variable to improve normality of the time series. The dominant patterns in the considered data are examined with the help of Principal Component Analysis (PCA) and factor analysis. Furthermore, particular focus is given for determining the concentration levels of ozone in relation to the other air pollutants and/or 6 meteorological time series using principal component regression (PCR). The good fitting of the obtained models with coefficients of determination R2 over 78% is obtained. An example of using the model to forecast the concentrations of ozone for 24 hours ahead is given. The obtained results could be used as an assessment in all analyses of the air quality of the town Dimitrovgrad, including the official reports of the Environmental Agency and also as an independent alternative to the official alerting systems.