Assessment and Prediction of PM 2.5 in Delhi in View of Stubble Burn from Border States Using Collaborative Learning Mod
- PDF / 2,907,016 Bytes
- 12 Pages / 595.276 x 790.866 pts Page_size
- 110 Downloads / 171 Views
ORIGINAL PAPER
Assessment and Prediction of PM2.5 in Delhi in View of Stubble Burn from Border States Using Collaborative Learning Model S. R. Mani Sekhar1 · G. M. Siddesh1 · Sarthak Jain1 · Tilak Singh1 · Vinay Biradar1 · Umer Faruk1 Received: 11 July 2020 / Revised: 31 October 2020 / Accepted: 2 November 2020 © Institute of Earth Environment, Chinese Academy Sciences 2020
Abstract Harvest scum burning in post-monsoon season in the nearby states increases air pollution levels. A thick layer of smog blankets can be seen in Delhi, India in November, the primary season of stubble burning. PM2.5 can play a major role in analyzing the stubble burning, as it has a better living time in the air. Numerous machine-learning techniques have been adopted to forecast the air quality, but none of them focuses primarily on the issues of stubble burning and its effect on Delhi’s air. In this paper, the author attempts to estimate the value of P M2.5 in Delhi, mainly due to stubble burning in neighboring states. For this, the P M2.5, PM10, NO2, CO, and S O2 data is taken into consideration for 9 ground-based continuous air quality monitoring stations in the neighboring states of Delhi for the duration of 6 months and 15 days in the calendar year of 2019. Subsequently various meteorological parameters are considered like wind speed, temperature, and relative humidity. The model used here is based on collaborative learning (stacking regression) which is trained using the out-of-folds predictions based on the complete training set, later the meta-regression is trained based on the outputs of the different regression methods in the group. The presented model is validated using different machine-learning methods, statistical measures, and the realtime stubble data of Delhi. The results show that the proposed method performs well when compared with previous methods. Keywords Stubble burning · Delhi · PM2.5 · PM10 · NO2 · CO · SO2 · Stacking regression · Air quality · Regression · Ensemble learning
1 Introduction As winter starts to set in, Delhi’s air quality will again start getting polluted by industrial waste, crop stubble burning (Ravindra et al. 2019; Zhuang et al. 2018; pal Singh et al. 2015), is yet again severely polluted by a combination of factors: crop stubble burning, garbage incineration, industrial disposal, vehicular pollution and dust from construction sites out of these factors most of the factors are present all year around but one factor that is not common year around is stubble burning in Punjab and Haryana (Ghosh and Parida 2015; pal Singh et al. 2015). Hence considering stubble burning a model can be made to get the extent of pollution in Delhi. Over the few years, pollution in Delhi started increase rapidly during the month of October and November. In the year 2020, the air quality index measuring over 900 in some * S. R. Mani Sekhar [email protected] 1
Department of Information Science and Engineering, M.S. Ramaiah Institute of Technology, Bangalore, India
areas (WHO 2020), which is cate
Data Loading...
