Time Series Models for Air Pollution Modelling Considering the Shift to Natural Gas in a Turkish City
Yükleniyor...
Dosyalar
Tarih
2015
Yazarlar
Dergi Başlığı
Dergi ISSN
Cilt Başlığı
Yayıncı
Wiley
Erişim Hakkı
info:eu-repo/semantics/closedAccess
Özet
Fossil fuel utilization for residential heating is still a major source of fine particulate matter (PM10) and sulphur dioxide (SO2), despite the increasing consumption of natural gas in some cities in Turkey. In the present study, PM10 and SO2 air pollution and residential natural gas consumption (RNGC) were modelled by various multi-parameter time series modelling methods (TSMs). To estimate short-term pollution levels considering the future estimates of RNGC and meteorological factors, a time series dataset was designed for the years 2007-2013. Factor analysis was also performed to aid in the selection of variables for constructing TSMs. The error measures and coefficient of determination (R-2) were used to evaluate forecasting accuracy of the models constructed. In the short-term estimation of RNGC, PM10, and SO2 for 2014-2015, temperature dependent ARIMAX(1,1,2) (R-2 = 0.944) and RNGC and meteorological factors dependent SARIMAX(0,1,1)(1,1,0) 12 (R-2 = 0.761) and ARIMAX(1,1,0) (R-2 = 0.698) models, respectively, yielded the best fitting scores and accuracy measures. The models performed well in reflecting the time series data and thus, could be utilized in energy planning for sustainable development concerning environmental decision making and short-term air quality forecasting for public health.
Açıklama
Taspinar, Fatih/0000-0002-5908-8188
WOS: 000358509300003
WOS: 000358509300003
Anahtar Kelimeler
Air quality forecasting, Energy consumption, Factor analysis, PM10, SO2
Kaynak
Clean-Soil Air Water
WoS Q Değeri
Q2
Scopus Q Değeri
Q3
Cilt
43
Sayı
7
