What is it about?

There are two primary objectives of this study: (1) the short-term estimation of local residential natural consumption (RNGC) by using time series methods (TSMs) based on meteorological factors, and (2) the prediction of future levels of PM 10 and SO 2 depending on the RNGC estimates implementing multivariate TSMs for the province of Düzce. To implement the models and perform statistical analysis, we have used the functions in RStudio with forecasting package and RExcel (R3.1.0) (www.r-project.org/) and Mathworks© Matlab. Factor analysis (FA) was applied before modelling to reveal the hidden correlations among RNGC, meteorological variables and APs by principal factors that also assist in the selection of the model variables. A time series data set was designed covering the concentrations of PM 10 and SO 2 , RNGC, meteorological factors, and some socioeconomic indicators for 2007 – 2013. In the modelling stage, TSMs including ARIMAX and SARIMAX, smoothing methods and multiple regression were examined to produce better estimations for future levels of RNGC and APs for 2014 – 2015, and estimated short-term concentration of air pollutants (APCs) were interpreted considering shifting to NG for domestic heating on temporal air quality.

Featured Image

Why is it important?

This is paper connects and explains two issue: natural gas consumption over a region considering meterological factors and its effect on air pollution reduction with diffferent scenarios. Air pollution due to fossil fuel is stil a major problem and reducing natural gas utilization in residential heating sector may be a solution. So, modeling of natural gas consumption and reduction in air pollutant concentrations of PM10 and SO2 by time series methos have been realized.

Perspectives

The present study investigated short-term RNGC modelling and the abatement of air pollution from PM 10 and SO 2 as a result of NG substitution, by using time series modelling techniques. The most signi fi cant parameter among others on the variations in RNGC was found to be AT that was fi tted by a seasonal model. In the estimation of RNGC, the AT dependent ARIMAX(1,1,2) model yielded the best scores representing historical data and the trend in increase. Based on the short-term AT and RNGC estimates for 2013 – 2015, the SARIMAX(0,1,1)(1,1,0) 12 and ARIMAX(1,1,0) models produced the most satisfactory scores for short-term estimates of PM10 and SO2 levels, respectively. The estimated PM 10 levels indicated the abatement of fossil fuel air pollution to a level approximately three times lower than the average of the past six years. Although short-term PM 10 levels meet the national air quality regulations, winter averages of PM10 remain high according to international standards. In contrast, the lower levels of SO2 estimates meet all air quality standards. The results revealed that the estimated PM 10 levels, rather than those of SO2 , can still pose health risks during the winter, despite increasing RNGC. The approach employed in this study resulted in a better understanding of short-term RNGC and its positive effect of air pollution by modelling, considering fuel shifting. However, further research is required to more accurately construct improved models for realistic estimates, and should include compilation of additional data such as daily fuel consump- tion statistics for the other fuel types in interest.

Dr Fatih Taşpınar
Duzce Universitesi

Read the Original

This page is a summary of: Time Series Models for Air Pollution Modelling Considering the Shift to Natural Gas in a Turkish City, CLEAN - Soil Air Water, February 2015, Wiley,
DOI: 10.1002/clen.201400461.
You can read the full text:

Read

Resources

Contributors

The following have contributed to this page