Impact of Climate Change and Air Pollution Forecasting Using Machine Learning Techniques in Bishkek

Abstract

During recent years, severe air-pollution problems have garnered worldwide attention due to their effects on human health and the environment. Air pollution in Bishkek, Kyrgyz Republic, is an ever-increasing problem with little research conducted on the impact of air pollutants on public health. We evaluate the performance of several machine learning algorithms applied to air quality and meteorology datasets and compare prediction accuracies of Bishkek air quality given its significant public importance. Data on 16 synoptic atmospheric process were collected by Kyrgyzhydromet from 2016 to 2018 and used to train and build a forecasting model. The model was then tested using data collected in 2020. Climate change in Bishkek and the impact on air pollution was assessed via the frequency of days characterized by daytime temperature inversions and air stagnation. Atmospheric stability increased from 2015 to 2020 with ongoing climate change leading to more temperature inversions. About 80%-90% of days with temperature inversions are associated with winter heating seasons and these numbers increased two-fold during the past 5 years. The impact of lockdown during May 2020) on air quality in Bishkek is also shown. During the lockdown period, CO, NO, NO2, SO 2 , and PM 2.5 decreased by 64%, 1.5%, 75%, 24%, and 54%, respectively, compared to concentrations of these pollutants in 2019. Where identified, emissions from vehicles make up a significant part of the air pollution.

Reconstruction of Hydrometeorological Data Using Dendrochronology and Machine Learning Approaches to Bias-Correct Climate Models in Northern Tien Shan, Kyrgyzstan

“…Once the best model was fully evaluated, it was applied to the full period of tree-ring data Water 2022, 14, 2297 6 of 16 to generate the reconstruction. Models were evaluated based on statistical methods for assessing errors between actual and reconstructed data, including root-mean-square error (RMSE) and mean absolute error (MAE) [14]. We compared values of RMSE and MAE with the various models (ANN, HPT, XgbR, RFR, KNN, DTR, LR, and LaR).…”

“…This is especially true for mountainous Kyrgyzstan, where 94% of the terrain is at least 1000 m a.s.l. [14].…”

“…Different statistical approaches have been applied in these hydrometeorological reconstructions using tree-ring data. Several studies employed Machine Learning (ML) methods to solve different earth and climate related problems [9,14,[43][44][45][46]. Our research is the first to use ML methods in Kyrgyzstan to reconstruct hydrometeorological data by comparing different statistical-regression and ML models.…”

Deep learning approach to forecast air pollution based on novel hourly index

“…The artificial neural network is proven to be the best classifier as well as a detector in many applications related to medical, scholarly literature and engineering [29][30][31][32][33]. There have been several studies that have utilized artificial intelligence (AI) to predict PM2.5 and PM10 levels in various locations, including Polish agglomerations [34], Santiago de Chile [35], Bishkek [36], Tehran [37] , and the study in Shanghai improved PM2.5 forecasts by employing machine learning (ML) technology in connect with the 'WRF-Chem model' simulations [38]. Iterative ANN (Artificial Neural Networks) in addition to Recurrent Neural Networks were used to estimate the value of PM10 and SO2 pollutants during the pandemic period of Covid 19 and RNN estimated pollutants with good accuracy [39] .…”

Bias correction of Sentinel-2 with unmanned aerial vehicle multispectral data for use in monitoring walnut fruit forest in western Tien Shan, Kyrgyzstan

“…Misregistration of Sentinel-2 imageries was addressed in the processing baseline (version 02.04) 21 and deployed by the European Space Agency on June 15, 2016. 29 However, when considering vegetation over complex terrain, such as mountainous Kyrgyzstan with 94% of the land occurring above 1000 m a.s.l., 30 RS becomes more challenging. Indeed, steep and complex orography leading to varying lighting conditions may deeply affect the computation of the overall spectral indices leading to a biased vegetation status assessment.…”