Spatio-Temporal Studies of Air Pollutants and Their Impacts on Karachi Region, Pakistan

Abstract

Air pollution is becoming a larger concern for the health of people, biodiversity, and the environment in big cities. In order to fully understand the spatiotemporal variability and primary causes of urban air quality (UAQ), it is crucial to monitor and model the UAQ in these locations. Due to Karachi's absence of a reference network, it was customarily impossible to create high-resolution AQ maps in metropolitan regions. It is now important to configure and develop high-resolution AQ maps using remote sensing retrievals corresponding to the four-year availability of federal and provincial air quality ground base. This is what this research study was aimed to accomplish by examining the suitability for urban AQ monitoring with modelling and how measurements can be improved using advanced validation techniques, ground base concentration measurements, based on multiple criteria and using sensors of various grades, employing various AQ modelling and mapping techniques, including geospatial interpolations, Pearson coefficient correlation (PCC) land use regression (LUR) and Weather Research and Forecast (WRF) modelling. Using univariant and multivariant linear regression and a generalized linear model, universal references were employed as the Pakistan standards to analyze the ground base contaminations. In order to map air quality, validate models, analyses spatiotemporal variability of contaminants and model performance, data from the designed GB data was employed. This research study was designed to cover four components, geospatial and PCC analysis of criteria pollutants, integrating and correlate significance of remote satellite origin (OMI NO2) and ground base NO2 data and application of LUR and WRF techniques to nine years (2013-2021) PM10-AOD data with climatic variables to predict the surface PM10 concentration. Geospatial analysis to generated database of criteria pollutants to account the pollutants trend and distribution over Karachi urban area. PCC applied to three Criteria pollutants databases, correlate to each variable and estimate the influence to each other with correlation and significance levels. LUR and WRF univariant and multivariant linear regression analysis to formulate quantitative models to predict and forecast PM10 concentration and estimate the near surface PM10 values with MODIS AOD retrievals. The spatial analysis of NO, NO2, SO2, CO, O3, PM2.5 and PM10 suggested that Karachi urban area is distributed in three regions on the basis of air quality trends, industrial, commercial with high traffic flows and elite residential localities. PCC statistically correlate the alldesignated criteria pollutants and estimated the significance among them, the results were acceptable and there was moderately correlation with high level of significance between the major pollutants, NOx and PMs. This study investigates the air quality in Karachi over the years 2015, 2017, and 2018, focusing on all criteria pollutants such as NO, NO2, SO2, CO, O3, PM10 and PM2.5. Using Pearson correlation analysis, the study highlights the strong and increasing correlations between NO, NO2, SO2, CO and PM2.5 pollutants, indicating common sources such as vehicular and industrial emissions. The concentration of various air pollutants in Karachi's atmosphere showed a significant positive correlation with each other except the ground-level Ozone (O3) which displayed a marked deviation in behavior from other pollutants. NO, NO2 has weak correlation coefficient with O3. O3 demonstrated high insignificant negative correlation with SO2, CO and PM10. Only PM2.5 has moderate correlation with O3 during 2018. O3 showed insignificant very weak correlations with other pollutants in Karachi region in all three study years, indicating it is due to unfavorable climatic conditions to restrict photochemical reactions to form this secondary pollutant. However, extremely polluted areas continue to be a source of bias and uncertainties due to meteorological variables, land use land cover and socioeconomic indicators. Monthly average OMI NO2 concentrations from space borne tools with GBNO2 observations were used to train and validate by PCC correlation. Comparing estimated and measured concentrations allowed for cross-validation of the models. Planet Boundary Layer Height (PBLH) had a significant effect on NO2 concentrations, whereas major roads intersections, commercial and industrial areas had a positive significant effect on NO2 growth in air Karachi. This technique estimated realistic (based on prior expectations) NO2 data measurements at GB sites are compared and correlated with Ozone Monitoring Instrument (KNMI TEMIS OMI NO2) retrievals over Karachi, together with validation against ground measurements. The approach proved effective in emphasizing the hotspots of NO2 concentrations in Karachi and illustrating the regional heterogeneity. In Karachi Metropolitan, LUR models were created for the first time incorporating a major variable (PM10) of criteria pollutants and climatic related factors. Here, univariant and multivariate regression procedures were also employed to build LUR and WRF models, which performed better than their linear versions, in contrast to earlier studies that mostly used linear techniques. In addition, the development and testing of an WRF forecasting model revealed that whereas point sources were mostly responsible for controlling PM10 concentrations, road traffic was primarily responsible for controlling PM10 concentrations. The study's key findings showed that over a nine-year period, the average values of observed PM10 and AOD increased by approximately 182% and 208%, respectively. However, during the COVID-19 lockdown period, these values experienced a significant reduction, ranging from approximately 28% to 35%. In Karachi, using the data fusion technique known as natural Neighbor (ArcMap), modelled and measured concentrations were fused (integrated) to produce high-resolution AQ maps. The key conclusions of this study are as follows: (a) It suggested a geospatial and temporal model for AP in ambient air quality. Although in-situ measurements are a relatively expensive source of AP data, they necessitate reliable outfield instrumentation. (b) Pearson Coefficient Correlation (PCC) two-dimensional generalization of temporal auto-correlation in which the correlation (r) and degree of significance (p) between multiple air quality variables. (c) insitu NO2 concentrations are integrate and correlate (PCC) with OMI NO2 retrievals, identified the emission and dispersion of NOx (NO+NO2), (d) statistical modelling estimations and observed concentrations were combined using data fusion techniques (LUR and WRF). These data fusion techniques are helpful tools for enhancing data quality and creating high-resolution predictions and predicting spatial missing data to fill in spatial missing PM10 data. The all studies are suggested that gray pollutants (PM) are the major portion of air pollutants in Karachi urban region and have adverse impacts on individual health and economic losses. There is a dearth of trace gaseous and particulate matter (fine and ultrafine particles) concentrations data continuously, therefore intensive work is needed on monitoring, modelling and management of NOx (NO+NO2) and PMs in Karachi region. To conclude, human activities have had impact on the air quality contamination in Karachi urban areas, in the meantime, the quality and quantity of fossil fuel as energy surging consumption may add significantly air pollution in Karachi s atmosphere. Based on multiple criteria air pollutants quantification studies concluded and proposed for future to structuring an AAQMN in Karachi urban settlements to record inventories continuously minutely, hourly, daily, monthly and yearly to helping in establishment of management policies on ground reality basis and enforce regulations properly. The future trajectory of air pollutant research in Karachi region needs the effective implementation of regulations and policies to reduce air pollution, addressing poverty, illiteracy and public awareness. It also discusses the development of better modeling and prediction techniques, improving satellite-based estimation, geostatistical hybrid modelling and exploring AI modeling approaches for future research.