Assessment and Prediction of Air Pollution Trends in Kuwait Using Machine Learning: An Analysis of PM10 , CO, and SO2 and Their Environmental Health Implications

Assessment and Prediction of Air Pollution Trends in Kuwait Using Machine Learning: An Analysis of PM10 , CO, and SO2 and Their Environmental Health Implications | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Research Article Assessment and Prediction of Air Pollution Trends in Kuwait Using Machine Learning: An Analysis of PM 10 , CO, and SO 2 and Their Environmental Health Implications Raslan Alenezi, Naeema Al-Darmaki, Fahed Javed, Sami Walid Azzam This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-8776307/v1 This work is licensed under a CC BY 4.0 License Status: Posted Version 1 posted You are reading this latest preprint version Abstract Kuwait is known for its hot, arid desert climate, frequent dust storms, and intensive oil and gas-related activities, all of which directly affect air quality. The current study show a comprehensive assessment of air pollution in Kuwait by using 10 fixed Air quality monitoring stations operated by the Kuwait Environment Public Authority. Air pollutant levels were analysed form 2020, while long-term trends in particulate matter (PM 10 ) were examined over the period Jan-2020 to Dec-2024. The results show persistently elevated particulate matter concentrations across Kuwait. Mean PM 10 levels ranged from approximately 60 to over 180 µg/m 3 , sometimes exceeding the World Health Organization (WHO) annual guideline of 15 µg/m 3 due to mainly weather conditions. Gaseous pollutants such as carbon monoxide (CO) and Sulfur dioxide (SO 2 ) displayed moderate but spatially variable concentrations. CO annual means ranged from about 486 to 1,572 µg/m 3 , while SO 2 averaged between 12 and 33 µg/m 3 , with higher levels observed near industrial and refinery areas. Seasonal analysis revealed strong PM 10 peaks during spring and summer due to dust storms and shamal winds, whereas CO and SO 2 exhibited weaker seasonal patterns linked mainly to traffic and energy production. Overall, Kuwait’s air quality reflects the combined influence of natural dust loading and localized anthropogenic emissions, posing significant environmental and public health concerns and underscoring the need for targeted mitigation strategies in arid regions. Air quality assessment Particular matter Kuwait ARIMA LSTM Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Figure 6 1 Introduction Air quality deterioration has emerged as one of the most pressing environmental challenges worldwide. Rapid population growth, urbanization, industrialization, and increasing energy consumption have contributed to elevated air pollution levels, particularly in urban regions (Bouhamra and Abdul-Wahab, 1999 ; Raslan A Alenezi and Aldaihani, 2023 ). Although natural phenomena such as wildfires and volcanic eruptions emit pollutants, anthropogenic activities remain the dominant sources of environmental air pollution (Raslan A Alenezi, 2022 ). Pollutants such as sulfur dioxide (SO 2 ), nitrogen dioxide (NO 2 ), carbon monoxide (CO), carbon dioxide (CO 2 ), ozone (O 3 ), and particulate matter (PM 10 and PM 2.5 ) significantly degrade air quality, affecting both human health and ecosystems. Numerous studies have demonstrated strong associations between air pollution exposure and increased risks of respiratory and cardiovascular diseases (Hwang and Lee, 2010 ; Shang et al., 2013 ; He et al., 2011 ). Kuwait represents a particularly critical case for studying air pollution due to its unique combination of environmental, industrial, and socioeconomic factors. As one of the world’s largest oil producers and a rapidly developing nation, Kuwait’s air quality faces significant challenges from both anthropogenic and natural sources. The country’s expanding population, intensive energy consumption, and rapid industrialization, especially in the petrochemical and oil refining sectors, have made it one of the most polluted countries in Southwest Asia (Barkley et al., 2017 ; Martínez Vallejo et al., 2021 ). Moreover, the country’s harsh desert climate and frequent dust storms further exacerbate airborne particulate concentrations, with more than 270 tons of dust deposited per km² annually in Kuwait City, the highest in the world (Al-Hemoud et al., 2018 ). These natural and human-induced conditions combine to create persistent air quality challenges, making Kuwait an ideal environment for assessing pollutant dynamics and their potential impacts on public health. Among the major contributors to Kuwait’s deteriorating air quality are vehicular emissions, industrial discharges, and dust events. The transport sector plays a dual role in Kuwait’s economy, supporting rapid development while simultaneously producing large volumes of pollutants such as CO, CO 2 , NO x , and volatile organic compounds (VOCs) (Ettouney et al., 2010 ; Raslan Alenezi et al., 2012 ). Between 2001 and 2010, the number of vehicles increased by approximately 6% annually, far outpacing the 2% expansion in road networks. This imbalance has resulted in severe traffic congestion and elevated vehicular emissions. Combined with heavily subsidized fuel prices and limited public transport, these factors have intensified air pollution and increased greenhouse gas emissions (Aldaihani and Alenezi, 2017 ; Al-Khulaifi et al., 2014 ). Despite technological inspections and modern vehicles, Kuwait’s transport sector remains one of the primary sources of CO 2 and CO emissions, contributing substantially to climate change and local air degradation. Furthermore, Kuwait’s geographic location and climatic conditions make it especially vulnerable to dust-related pollution. Frequent dust storms elevate levels of fine and coarse particulate matter (PM 2.5 and PM 10 ), often exceeding World Health Organization (WHO) air quality guidelines (Alkharafi, 2025 ). Such elevated concentrations are associated with various respiratory and cardiovascular health effects (Li et al., 2020 ; Al-Enezi et al., 2024 ). These challenges highlight the urgency of understanding the spatial and temporal behavior of key air pollutants in Kuwait to inform environmental management and health protection strategies. Hence, the present study focuses on assessing the levels and temporal variations of PM 10 , CO, and SO 2 across Kuwait. Data was collected from ten air quality monitoring stations over a four-year period to analyze pollutant trends and seasonal fluctuations. Using machine learning techniques, this research models and predicts future pollutant concentrations and evaluates their potential environmental and health impacts. The findings aim to support policymakers in developing evidence-based strategies for improving air quality and safeguarding public health in Kuwait and other arid regions facing similar environmental challenges. 2 Material and methods 2.1 Data collection The country under study is Kuwait, located in the northeastern corner of the Arabian Peninsula. Its surface elevation ranges from approximately 280 m above sea level in the southwest to sea level along the eastern Gulf Coast. Kuwait shares land borders with Saudi Arabia to the south and Iraq to the north and west, and a maritime boundary with Iran across the Arabian Gulf (Fig. 1 ). The nation’s economy is heavily dependent on the oil and gas industry, which plays a central role in both economic growth and environmental challenges. To monitor air quality across the country, the Kuwait Environment Public Authority (KU-EPA) operates a network of fixed Air Quality Monitoring Stations (AQMS) that continuously measure concentrations of key atmospheric pollutants. The spatial distribution of these AQMS is illustrated on the map (Figs. 2 ), where every station is represented by a colored dot indicating its location. These stations provide critical data for evaluating spatial and temporal trends in air pollution levels across Kuwait from which we select 10 station to get overall data of Kuwait. 2.2 Time-Series Modelling Methodology This study employs time-series modelling to analyze and forecast monthly variations in selected air pollutants using aggregated monitoring data. The methodological framework focuses on constructing consistent monthly time series, addressing missing values and extreme observations, and applying statistical and machine-learning models suitable for short to medium-term forecasting, as shown in Fig. 3 . 2.2.1 Construction of Monthly Averaged Time Series Monthly air quality data for CO, PM 10 , and SO₂ were obtained from multiple monitoring stations. For each pollutant, only concentration values reported in µg/m³ were retained to ensure unit consistency. The original datasets were re-organized into a unified format containing station identifiers, year and month, and pollutant concentrations. Monthly pollutant concentrations were then spatially averaged across all available stations to produce a single regional-scale time series per pollutant. This approach reduces station-specific noise and yields 64 monthly observations for each pollutant, which is appropriate for global forecasting models such as ARIMA and LSTM. 2.2.2 Treatment of Missing Values and Outliers Missing values were handled using time-based linear interpolation applied separately to each station-level monthly series before spatial averaging. This preserves temporal continuity while avoiding artificial discontinuities. CO and SO₂ series contained no missing values, whereas PM₁₀ included a limited number of missing months that were successfully interpolated. Exploratory analysis revealed the presence of extreme PM₁₀ values that strongly influenced model behaviour. To assess model robustness, two versions of the PM₁₀ series were considered i.e. an unclipped series, retaining all observed values, and a clipped series, where station-level PM₁₀ concentrations exceeding 800 µg/m 3 were truncated before recomputing monthly spatial averages. This threshold was chosen to reduce the influence of a small number of extreme observations while retaining the overall structure of high-pollution periods. 2.2.3 ARIMA Model Specification Seasonal Autoregressive Integrated Moving Average (SARIMA) models were used as a classical statistical benchmark for forecasting monthly pollutant concentrations. Given the monthly resolution of the data, a seasonal period of 12 months was adopted. Model identification was based on an Akaike Information Criterion (AIC) grid search over plausible combinations of autoregressive, differencing, and moving-average orders. The optimal configuration for each pollutant was selected as the model with the lowest AIC value. All ARIMA models were univariate and did not include exogenous predictors, relying solely on past pollutant values to generate forecasts. 2.2.4 LSTM Neural Network Model To capture potential nonlinear and long-range temporal dependencies, a univariate Long Short-Term Memory (LSTM) neural network was implemented for each pollutant using the spatially averaged monthly series. Prior to modelling, each series was scaled to the [0,1] range using min–max normalisation. The time series was transformed into supervised learning samples using a six-month sliding window, where the previous six months were used to predict the subsequent month. The LSTM architecture consisted of a single LSTM layer followed by fully connected layers, trained using the Adam optimizer and mean squared error loss. Early stopping was applied to limit overfitting. For PM₁₀, separate LSTM models were trained on the unclipped and clipped datasets to explicitly examine the effect of extreme values on model performance. 2.2.5 Model Evaluation Metrics Model performance was assessed using standard error metrics commonly applied in air-quality forecasting studies i.e. Mean Absolute Percentage Error (MAPE), Mean Absolute Error (MAE), Root Mean Squared Error (RMSE). For ARIMA models, these metrics were computed using a 12-month hold-out validation window, with one-step-ahead forecasts compared against observed values. LSTM performance was primarily evaluated using in-sample RMSE, reflecting the exploratory nature of the neural-network implementation. 2.2.6 Comparative performance analysis Model comparison was conducted at multiple levels to provide a comprehensive assessment of forecast performance. First, a within-pollutant comparison was performed by contrasting the predictive accuracy of ARIMA and LSTM models for each individual pollutant. Second, an across-pollutant comparison examined differences in forecastability among CO, PM 10 , and SO 2 , acknowledging that each pollutant exhibits distinct variability and temporal characteristics. Finally, sensitivity to extreme values was evaluated by comparing PM 10 model results with and without outlier clipping. This structured comparison framework enables a balanced evaluation of the relative strengths and limitations of classical statistical models and machine-learning approaches under realistic data constraints. 3 Results and discussion 3.1 Variability analysis of different stations in Kuwait Table 1 presents the statistical summary of CO, SO 2 , and PM 10 concentrations recorded across ten monitoring stations in Kuwait from Jan-2020 to Dec-2024, including minimum, maximum, mean, and standard deviation values. The results indicate marked spatial variability among stations, reflecting the diverse environmental and emission characteristics across the country. The variability patterns highlight the influence of industrial activities, roadway traffic emissions, and natural dust events on local air quality. The analysis of CO concentrations reveals substantial differences among stations. Al-Mutla exhibits the highest mean CO concentration (1,572.04 µg/m 3 ), accompanied by a relatively high standard deviation, indicating persistently elevated levels with frequent fluctuations. Similarly, Al-Shuaiba and Al-Ahmadi show markedly high CO means (1,472.61 µg/m 3 and 1,159.19 µg/m 3 , respectively), suggesting the presence of consistent emission sources, likely tied to industrial operations and heavy vehicular movement. In contrast, stations such as Al-Jahra and Al-Rumaithiya recorded much lower CO means (486.27 µg/m 3 and 563.10 µg/m 3 ), with comparatively smaller ranges between minimum and maximum values, reflecting more stable conditions and fewer intense emission episodes. The exceptionally high maximum values observed at several sites (e.g., 3,395.38 µg/m 3 at Al-Shuaiba) point toward episodic industrial plumes or localized high-emission events. SO 2 levels across the monitoring network display lower overall variability compared to CO, yet notable differences persist among stations. Al-Ahmadi stands out with the highest mean SO 2 concentration (33.20 µg/m 3 ), which may be indicative of combustion-related activities or refinery emissions in the surrounding area. Other stations, such as Al-Fahaheel and Saad Al-Abdullah, recorded moderate SO 2 means (24.29 and 24.37 µg/m 3 , respectively), while stations including Al-Mutla and Al-Rumaithiya show relatively low mean concentrations (12.46 and 19.07 µg/m³). The narrow ranges and modest standard deviations for SO 2 at many locations suggest that sulfur-related emissions are more spatially concentrated and less temporally volatile than CO or PM 10 . . Table 1 Air pollutants statistical analysis by different stations in Kuwait from Jan-2020 to Dec-2024. Sr.no. Station name CO (µg/m 3 ) SO 2 (µg/m 3 ) PM 10 (µg/m 3 ) Min Max Mean SD Min Max Mean SD Min Max Mean SD 1 Al-Ahmadi 410.55 3280.03 1159.19 436.63 14.64 76.95 33.20 10.80 38.55 362.28 114.01 62.66 2 Al-Fahaheel 504.05 1416.69 1015 185.15 3.77 47.29 24.29 9.25 53.16 4525.39 206.48 582.36 3 Al-Jahra 112.36 2081.50 486.27 386.94 6.99 73.55 19.99 9.86 60.98 387.04 144.63 71.34 4 Al-Mutla 844.56 2395.91 1572.04 367.48 5.42 24.45 12.46 3.82 74.08 1258.56 210.49 168.58 5 Al-Rumaithiya 117.99 1750.76 563.10 329.45 6.65 40.28 19.07 7.17 110.04 43.48 421.60 62.43 6 Al-Salem 427.11 2784.61 1022.08 463.19 5.42 43.27 27.09 7.25 43.65 378.86 114.48 63.67 7 Al-Shuaiba 267.84 3395.38 1472.61 952.56 13.56 49.69 23.68 23.68 55.74 377.50 125.48 56.66 8 Al-Shuwaikh 478.52 1754.44 916.75 315.03 10.21 40.37 22.26 6.86 52.14 225.44 109.49 35.86 9 Al-Subah Al Salem 1 624.11 2219.62 943.11 277.40 7.42 31.82 16.22 5.27 86.16 474.06 185.89 81.22 10 Saad Al-Abdullah 398.48 3244.04 853.73 386.14 8.27 45.16 24.37 8.41 67.20 424.08 158.08 75.05 PM 10 concentrations exhibit the most pronounced spatial and temporal variability among the three pollutants. Several stations recorded exceptionally high maximum PM 10 values, such as Al-Fahaheel (4,525.39 µg/m 3 ) and Al-Mutla (1,258.56 µg/m 3 ), reflecting the strong influence of dust storms and resuspended desert particles, which are characteristic of Kuwait’s arid climate. Mean PM 10 levels were highest at stations such as Al-Subah Al-Salem 1 (185.89 µg/m 3 ) and Saad Al-Abdullah (158.08 µg/m 3 ), indicating persistent exposure to coarse particulate matter. Conversely, stations like Al-Rumaithiya and Al-Salem registered considerably lower PM 10 means (62.43 and 114.48 µg/m 3 , respectively), likely reflecting their relative distance from industrial zones or major dust-entrainment areas. The large standard deviations associated with PM 10 at numerous stations underscore the episodic nature of particulate pollution, driven by both natural events and anthropogenic activities Overall, this station variability analysis reveals clear differences in pollutant behavior across Kuwait’s monitoring network. Stations situated near industrial complexes or major roadway corridors consistently exhibit elevated mean concentrations of CO and PM 10 , while SO 2 levels remain more localized and comparatively stable. The pronounced fluctuations in PM₁₀ further emphasize the combined effect of natural dust activity and human-induced emissions on Kuwait’s air quality. These station-specific patterns underscore the need for targeted air quality management strategies that consider local emission sources, land-use characteristics, and regional meteorological factors that shape pollutant dispersion and accumulation. 3.2 Pollutant historic trend in Kuwait Figure 4 illustrates the annual mean concentrations of CO, SO 2 and PM 10 , averaged over the ten monitoring stations in Kuwait for the period Jan-2020 to Dec-2024. Overall, the series shows clear interannual variability rather than a simple monotonic increase or decrease, reflecting the combined influence of changing anthropogenic activities (traffic, power generation, and industrial operations) and natural factors such as dust storm frequency and regional meteorology (Alkharafi, 2025 ). Across all years, PM 10 annual means remain substantially higher than international health-based guidelines. According to the 2021 WHO Global Air Quality Guidelines, the recommended annual PM 10 guideline value is 15 µg/m³ (Organization, 2021 ). In contrast, the multi-year mean values in Kuwait’s stations (Table 1 ) lie roughly between 60 and > 180 µg/m³, indicating that PM 10 levels exceed the WHO annual guideline throughout the study period. This pattern is consistent with previous work in Kuwait showing that PM 10 annual averages in both urban and industrial areas surpass both national and WHO guideline values (Alahmad et al., 2021 ). The annual behavior of PM 10 is driven strongly by year-to-year differences in dust storm activity and synoptic weather. Years with higher annual PM 10 in Fig. 1 (c), can reasonably be linked to more frequent or more intense dust episodes and stronger shamal winds, which transport dust from exposed soil in Kuwait and neighboring countries (Iraq, Syria, Saudi Arabia) (Sabbah et al., 2018 ). Studies from Kuwait have shown that dust storms and suspended dust events are the dominant contributors to extreme PM 10 concentrations and are closely associated with hot, dry, and windy conditions (Al-Hemoud et al., 2018 ). Thus, peaks in the annual PM 10 pollutants are likely to reflect years with more dust storms and enhanced regional dust transport, while relatively lower annual values correspond to years with fewer such events. For CO, the annual mean concentrations at all stations are on the order of a few hundred to around 1,500 µg/m³ (0.5–1.6 mg/m³), well below the WHO short-term guideline of 4 mg/m³ as a 24 h and 10 mg/m³ as an 8 h (Organization, 2021 ). Although annual averages cannot be directly compared to short-term limits, the fact that long-term means are substantially lower than these thresholds suggests that CO is not the primary health driver in this dataset. Any variation from year to year is likely linked to changes in traffic volume and power generation demand, which can modify CO emissions from vehicles and thermal power plants. Time-series analysis in Kuwait has shown that pollutants tied to energy use and traffic often track changes in domestic energy demand and mobility patterns (Al-Hurban et al., 2021 ; Al-Bassam and Khan, 2004 ). For SO 2 , the station-averaged means (typically 12–33 µg/m³) remain below the WHO 24-h guideline of 40 µg/m³ in terms of annual averages (Organization, 2021 ). However, SO 2 is usually more localized in space and time, with short-term peaks near industrial stacks and power plants. Studies in Kuwait have reported that SO₂ levels in industrial zones can exceed national or guideline values at certain times, even if long-term means appear moderate (Al-Rashidi et al., 2005 ). Thus, interannual SO 2 differences in Fig. 1 (b) can be interpreted in terms of variations in industrial output, fuel sulphur content and effectiveness of emission control technologies. 3.3 Seasonal and quarterly trend analysis of Kuwait Figure 5 shows the seasonal evolution of CO, SO 2, and PM 10 across the monitoring network during Jan-2020 to Dec-2024, revealing a clear cyclic behavior driven primarily by Kuwait’s desert climate and regional meteorology. Among the three pollutants, PM 10 exhibits the strongest seasonal signal, with concentrations peaking during spring and summer and decreasing in autumn and winter. This pattern matches long-term observations in Kuwait and the wider Arabian Peninsula, where multiple satellite-based and ground-based studies have confirmed that aerosol loading and dust concentrations reach their highest levels during warm months due to shamal winds, extremely low soil moisture, and enhanced surface heating (Al-Hemoud et al., 2018 ; Omar et al., 2022 ). Remote-sensing analyses specifically show that aerosol optical thickness over Kuwait increases sharply from April to August, reflecting intense dust generation and transport (Uddin et al., 2023 ). In winter, cooler temperatures, calmer winds and slightly higher humidity suppress dust emission, leading to the lowest dust particles. The elevated spring–summer PM 10 levels observed in this study are also consistent with the broader Middle Eastern literature. In Kuwait, Al-Hemoud et al. ( 2018 ) showed that dust storms and suspended dust events are strongly associated with sharp increases in PM 10 and are the dominant contributors to extreme particle levels during hot, dry, and windy conditions (Al-Hemoud et al., 2018 ). In another study by Alahmad et al. ( 2021 ) demonstrated that regional dust storms and suspended local dust together account for the majority of PM mass in Kuwait, confirming the dependence of PM 10 on seasonal dust climatology (Alahmad et al., 2021 ). CO and SO 2 display more moderate but still noticeable seasonal variations. In Kuwait and similar arid cities, meteorology, particularly temperature, boundary-layer height, wind speed, and atmospheric stability, plays a central role in modulating combustion-related pollutants. Studies from Makkah (Munir et al., 2017 ; Habeebullah et al., 2015 ) provide evidence from environments with nearly identical climate characteristics, showing that CO and SO₂ vary systematically with heat-driven electricity demand, traffic intensity, and atmospheric mixing depth. In Kuwait, extremely high summer temperatures drive intense air-conditioning use, increasing electricity consumption and thereby fuel combustion in gas and oil-fired power plants, which contributes to seasonal rises in CO and SO₂. Although mixing heights are generally deeper in summer, persistent thermal lows and stagnant atmospheric conditions can still trap pollutants near the surface, contributing to the slight summer elevations visible in Fig. 5 . Short-term SO 2 behavior is also strongly influenced by the spatial clustering of major point sources such as refineries and power stations. Air-quality assessments around Kuwaiti oil and gas facilities have documented SO 2 enhancements downwind of industrial stacks, especially during warm seasons when dispersion conditions vary rapidly (Al-Rashidi et al., 2005 ; Al-Hurban et al., 2021 ). In winter, thermal inversions can form near the surface, reducing mixing and intermittently trapping SO 2 close to the ground even when emissions are lower. For this reason, the seasonal SO 2 cycle reflects a combined interplay of power-plant load, refinery activity, and seasonal atmospheric stability rather than dust-related processes. Overall, these findings show that PM 10 is overwhelmingly controlled by natural meteorological drivers and dust climatology, with spring and summer peaks linked to dust storms, shamal winds, regional dust transport, and strong surface heating. CO and SO₂, on the other hand, show more modest seasonal modulation linked to anthropogenic factors such as fuel combustion and electricity demand, combined with the influence of atmospheric stability and synoptic conditions typical of Kuwait’s hyper-arid climate. 3.4 Pearson Correlation analysis Pearson correlation coefficients were used to evaluate the linear associations between air pollutant concentrations and to assess whether pollutants exhibit common temporal behavior, as shown in Table 2 . The results show no strong inter-pollutant correlations, indicating limited temporal interdependence between most pollutant pairs. A weak negative correlation between CO and PM 10 suggests that these pollutants are likely influenced by different emission sources or atmospheric processes. Likewise, the negligible correlation between CO and SO 2 indicates largely independent behavior, reflecting distinct combustion or industrial origins and dispersion mechanisms. In contrast, a weak positive correlation between PM 10 and SO 2 may indicate partial overlap in emission sources; however, the low correlation strength suggests that common sources are not the dominant drivers of their variability. Overall, the lack of strong correlations highlights the complex, multi-source nature of air pollution in Kuwait. These findings imply that pollutants respond differently to emission patterns and environmental conditions, supporting the need for multi-pollutant and source-specific air quality management strategies. The current correlation analysis is consistent with previous studies, where weak correlations were linked to independent pollutant behavior (Ramadan et al., 2024 ). Table 2 Pearson correlation coefficients among air pollutant concentrations. Pollutants CO PM10 SO 2 CO 1.00 -0.19 -0.01 PM 10 -0.19 1.00 0.19 SO 2 -0.01 0.19 1.00 3.5 Forecast Behavior and Visual Assessment for Kuwait air pollutants The forecast behavior of the ARIMA and LSTM models was evaluated through visual inspection of the predicted trajectories in conjunction with the associated numerical performance metrics, as shown in Fig. 6 . For CO, the ARIMA model achieved a MAPE = 27.94%, MAE = 270.94 µg/m 3 , and RMSE = 335.46 µg/m 3 on the 12 month validation window. The corresponding forecast indicates that monthly spatially averaged CO concentrations are likely to remain within a similar range to recent years, with moderate seasonal oscillations. However, the widening confidence intervals beyond approximately 24 months highlight growing uncertainty in long-term projections. In contrast, the LSTM model achieved a substantially lower in-sample RMSE = 182.05 µg/m 3 , and its fitted curve closely follows the observed monthly mean, producing a relatively smooth 36-month forecast centered around 1100–1150 µg/m 3 . This suggests that the LSTM captures short-term variability in aggregated CO dynamics more effectively than the linear ARIMA framework. In case of SO 2 , the best ARIMA yielded a MAPE = 25.24%, MAE = 6.86 µg/m 3 , and RMSE = 9.28 µg/m 3 . Although absolute errors are small due to the low concentration levels, the relatively high MAPE reflects pronounced relative variability. The ARIMA forecast shows largely stable SO 2 levels with wide confidence intervals. The LSTM model, by comparison, achieved a lower in sample RMSE of 5.59 µg/m 3 , and visually reproduces both local peaks and troughs more closely. Its forecast exhibits a gentle upward drift with modest seasonality, remaining within plausible bounds. For PM 10 , forecast behavior depends strongly on outlier treatment. When the unclipped series is used, the ARIMA model becomes numerically unstable, producing implausibly large error metrics (MAPE and RMSE on the order of 10 35 ), indicating a breakdown of the linear seasonal model due to extreme monthly values. The LSTM remains operational on the same data, with an in sample RMSE of 82.06 µg/m 3 but still struggles to balance typical variability against extreme peaks. After applying clipping at 800 µg/m 3 , forecast behavior improves markedly for both models. The clipped-series ARIMA model achieves a MAPE of 29.68%, MAE of 32.72 µg/m 3 , and RMSE of 45.30 µg/m 3 , while the corresponding LSTM attains an in-sample RMSE of 47.18 µg/m 3 . The resulting forecasts are stable and realistic, although uncertainty remains substantial beyond one to two years due to the intrinsic volatility of PM 10 . This visual assessments, supported by numerical metrics, confirm that both models produce reasonable short to medium-term forecasts when data quality issues are addressed, while long-term projections remain uncertain. 3.6 Comparative Performance of Forecasting Models A comparative evaluation of ARIMA and LSTM performance was conducted across pollutants using error metrics and visual diagnostics. In case of CO, the LSTM model substantially outperforms ARIMA. The LSTM achieves an RMSE of 182.05 µg/m 3 , compared with the ARIMA validation RMSE of 335.46 µg/m 3 and MAPE of 27.94%, representing an approximate 45% reduction in RMSE. While ARIMA provides a reasonable and interpretable linear baseline, the lower error and closer visual fit of the LSTM indicate its superior ability to capture nonlinear temporal patterns in aggregated CO concentrations. In case of SO 2 , a similar advantage is observed. The ARIMA model yields an RMSE of 9.28 µg/m 3 and MAPE of 25.24%, whereas the LSTM reduces the RMSE to 5.59 µg/m³, corresponding to a reduction of roughly 40%. This suggests that even for pollutants with relatively smooth dynamics, nonlinear models can provide measurable gains in predictive accuracy. For PM₁₀, model performance is highly sensitive to outliers. On the unclipped series, ARIMA fails entirely due to numerical instability, while the LSTM remains usable with an RMSE of 82.06 µg/m³. After clipping extreme values at 800 µg/m³, both models improve substantially. In this cleaned setting, ARIMA achieves an RMSE of 45.30 µg/m³ (MAPE = 29.68%), slightly outperforming the LSTM (RMSE = 47.18 µg/m³). This result highlights that for highly variable pollutants, appropriate preprocessing can make classical statistical models competitive with more complex machine-learning approaches. 3.7 Health Related to Air Pollutants The elevated pollutant concentrations observed in this study, particularly the persistently high PM 10 levels and the moderate SO 2 and CO values, carry well-documented implications for public health. Long-term evidence shows that coarse particulate matter such as PM 10 can penetrate deep into the respiratory tract, aggravating asthma, chronic obstructive pulmonary disease, and lung infections, while also increasing the risk of cardiovascular events through systemic inflammation and blood-pressure alterations (Grzywa-Celińska et al., 2020 ; Zhao et al., 2017 ). Zhao et al. ( 2017 ) specifically demonstrate that each 10 µg/m 3 increase in PM 10 is associated with a significant rise in coronary heart disease mortality, illustrating how elevated particulate levels, such as those measured across Kuwait’s stations, can cause heightened cardiovascular risk (Zhao et al., 2017 ). Lipfert ( 2018 ) further study long term exposure of PM 10 further strengthens this understanding by showing that PM 10 has some of the most consistent long-term associations with respiratory morbidity, impaired lung function, and adverse birth outcomes among hundreds of epidemiological studies (Lipfert, 2018 ). In comparison, the SO 2 levels recorded in this study, although lower than PM10, remain relevant because chronic or repeated exposure to SO 2 irritates the airways, triggers bronchoconstriction, and exacerbates respiratory symptoms, especially among children, the elderly, and individuals with pre-existing lung disease (Grzywa-Celińska et al., 2020 ). Long-term evidence also links SO 2 exposure to increases in respiratory and cardiovascular hospital admissions, indicating that even moderate concentrations contribute cumulatively to health burdens over time (Lipfert, 2018 ). Meanwhile, the CO concentrations measured across Kuwait’s stations fall below acute toxicity thresholds but still present concerns in high-traffic or poorly ventilated settings. Chronic exposure, particularly among roadside workers and commuters, can reduce oxygen transport, intensify cardiovascular strain, and heighten vulnerability among those with existing heart conditions (Sitanggang et al., 2023 ; Manisalidis et al., 2020 ). Manisalidis et al. ( 2020 ) further emphasize that even low-level, long-term exposure to combustion-derived pollutants like CO diminishes cardiorespiratory resilience and increases overall susceptibility to disease (Manisalidis et al., 2020 ). Hence, these findings show that the pollutant levels identified in Kuwait, most notably the extremely high PM 10 concentrations and intermittent SO 2 and CO exposures, represent meaningful environmental health risks that align closely with established epidemiological evidence. The consistent pattern across the literature underscores that populations exposed to such pollutant mixtures face compounded risks over time, particularly vulnerable groups such as children, older adults, and outdoor workers. Overall, the evidence indicates that without targeted mitigation and sustained air-quality management, the pollutant levels documented in this study are likely to contribute to persistent and preventable respiratory and cardiovascular disease burdens in Kuwait. 3.8 Strategies to Reduce Air Pollution The findings of this study highlight the urgent need for integrated and evidence-based strategies to reduce air pollution in Kuwait, where natural dust events and anthropogenic emissions jointly drive deteriorating air quality and related health risks. Persistently elevated concentrations of PM 10 , intensified by frequent dust storms, regularly exceed international safety thresholds and are strongly associated with respiratory and cardiovascular diseases (Achilleos et al., 2023 ; Anagnostopoulou et al., 2023 ). These challenges are further exacerbated by climate change, as projections indicate rising temperatures, prolonged drought, and increased desertification across the Middle East by 2050 (Alkharafi, 2025 ; Brown and Crawford, 2009 ). Addressing air pollution in Kuwait requires coordinated action at local, national, and regional scales. Local governments play a central role through strengthened air quality regulations, emission reduction policies, and early warning systems that forecast dust and pollution episodes, enabling behavioral adaptation such as limiting outdoor activities during high-exposure periods (Liu et al., 2023 ; Kouis et al., 2024 ). Public awareness initiatives are equally critical, ensuring communities understand air-quality warnings, associated health risks, and appropriate protective responses (Gross et al., 2018 ). Improving indoor air quality represents a key protective strategy, particularly during severe dust events. Enhancing building insulation, regular maintenance of HVAC systems, and installing effective filtration can significantly reduce indoor particulate infiltration. For outdoor and industrial workers, occupational protection measures including N95 respirators, scheduled work breaks, and access to clean-air shelters, can substantially reduce exposure. In parallel, wearable and fixed air-quality sensors offer valuable tools for real-time exposure assessment and targeted interventions (Al-Dousari et al., 2025 ). Long-term sustainability strategies must also address emission sources. Strengthening industrial emission regulations, expanding renewable energy initiatives such as the solar project, and increasing vegetation cover to stabilize soil and limit dust resuspension are essential measures (Alkharafi, 2025 ). Enhanced monitoring networks and regional cooperation, similar to the Saudi Sand and Dust Storm Warning Center, can further support evidence-based environmental governance (Luck, 2023 ). Together, these integrated strategies can reduce air pollution, protect public health, and advance long-term environmental sustainability in Kuwait. 4 Conclusions This study provides a comprehensive assessment of air pollution dynamics in Kuwait by integrating multi-station observational data with time-series forecasting techniques. Analysis of PM 10 , CO, and SO 2 concentrations from Jan-2020 to Dec-2024 reveals pronounced spatial, seasonal, and temporal variability driven by the combined influence of anthropogenic emissions and natural desert processes. Among the investigated pollutants, PM 10 poses the most critical environmental and public health concern, with concentrations consistently exceeding World Health Organization guideline values across all monitoring stations. These elevated levels are strongly linked to Kuwait’s arid climate, frequent dust storms, and regional dust transport, underscoring the dominance of natural factors in shaping particulate pollution. In contrast, CO and SO₂ exhibit comparatively moderate concentrations and weaker seasonal variability, reflecting their closer association with localized sources such as traffic, power generation, and industrial activities. While long-term averages of these pollutants generally remain below international guideline thresholds, episodic peaks near industrial and urban hotspots highlight the need for continued regulatory oversight and targeted mitigation strategies. Correlation analysis further confirms that the three pollutants largely behave independently, reinforcing the importance of pollutant-specific management approaches. From a methodological perspective, the comparative evaluation of ARIMA and LSTM models demonstrates that machine-learning approaches can outperform classical statistical models, particularly for pollutants with nonlinear dynamics such as CO and SO 2 . However, for highly variable PM 10 data, appropriate preprocessing of extreme values is essential, after which traditional models can remain competitive. Overall, the findings emphasize that sustained air quality challenges in Kuwait require integrated solutions combining robust forecasting tools, enhanced monitoring, emission controls, and adaptive public health strategies. These insights are not only relevant for Kuwait but also for other arid and dust-prone regions facing similar environmental pressures. Declarations Data availability The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request. Acknowledgement The authors would like to acknowledge the support of the Public Authority for Applied Education and Training (PAAET) for funding sabbatical leave, the Kuwait Environment Public Authority (KU-EPA) for providing data, and the United Arab Emirates University Climate Action Program grants (number 12N142) . Funding The research reported in this manuscript is funding by the Public Authority for Applied Education and Training (PAAET). Ethics declarations Ethics approval Not applicable Consent to participate Not applicable. Consent for publication Not applicable. Clinical trial number Not applicable Competing interests The authors declare no competing interests References Achilleos, S., Michanikou, A., Kouis, P., Papatheodorou, S. I., Panayiotou, A. G., Kinni, P., et al. (2023). Improved indoor air quality during desert dust storms: the impact of the MEDEA exposure-reduction strategies. Science of the Total Environment , 863, 160973. Al-Bassam, E., & Khan, A. (2004). Air pollution and road traffic in Kuwait. WIT Transactions on the Built Environment , 75. Al-Dousari, A., Al-Khalaifah, H., Abbas, S., Alahmad, B., Omar, A., Koutrakis, P., et al. (2025). Integrating Health and Economic Perspectives: A Comprehensive Review of Dust Mitigation Policies. Earth Systems and Environment , 1–20. Al-Enezi, D., Al-Enezi, A.-A., Al-Dousari, A., & Aldashti, H. (2024) 'Assessing the Impact of Sand and Dust Storms on Residents’ Health in the State of Kuwait' Proceedings of the International Conference on Sustainability: Developments and Innovations . Springer, pp. 151–157. Al-Hemoud, A., Al-Dousari, A., Al-Shatti, A., Al-Khayat, A., Behbehani, W., & Malak, M. (2018). Health impact assessment associated with exposure to PM10 and dust storms in Kuwait. Atmosphere , 9(1), 6. Al-Hurban, A., Khader, S., Alsaber, A., & Pan, J. (2021). Air Quality Assessment in the State of Kuwait during 2012 to 2017. Atmosphere , 12(6), 678. Al-Khulaifi, N. M., Al-Mudhaf, H. F., Alenezi, R., Abu-Shady, A.-S. I., & Selim, M. I. (2014). Seasonal and temporal variations in volatile organic compounds in indoor and outdoor air in Al-Jahra City, Kuwait. Journal of Environmental Protection , 2014. Al-Rashidi, M., Nassehi, V., & Wakeman, R. (2005). Air quality modelling of sulphur dioxide emission from power plants in Kuwait. WIT Transactions on Ecology and the Environment , 82. Alahmad, B., Al-Hemoud, A., Kang, C.-M., Almarri, F., Kommula, V., Wolfson, J. M., et al. (2021). A two-year assessment of particulate air pollution and sources in Kuwait. Environmental Pollution , 282, 117016. Aldaihani, N., & Alenezi, R. (2017). Estimation of CO2 emissions of the vehicles transport sector in the state of Kuwait. Acta Chemica Malaysia , 1(1), 08–12. Alenezi, R., Al-Anzi, B., Abusam, A., & Ashfaque, A. (2012). Seasonal influence on the ambient air quality in Al Jahra City for year 2010. Journal of Environmental Protection , 3(12), 1711–1718. Alenezi, R. A. (2022). BTEX and nitrogen oxides fume assessments in fuel dispense stations and residential areas. Journal of Engineering Research , 10(3A). Alenezi, R. A., & Aldaihani, N. (2023). Investigation of the atmospheric chemical properties over Kuwait. Water, Air, & Soil Pollution , 234(4), 216. Alkharafi, N. (2025). An examination of air quality in Kuwait: implications for environmental sustainability and public health. Kuwait Journal of Science , 100447. Anagnostopoulou, P., Kouis, P., Papatheodorou, S. I., Middleton, N., Papasavvas, I., Avraamides, P., et al. (2023). MEDEA randomised intervention study protocol in Cyprus, Greece and Israel for mitigation of desert dust health effects in adults with atrial fibrillation. BMJ open , 13(3), e069809. Barkley, M. P., González Abad, G., Kurosu, T. P., Spurr, R., Torbatian, S., & Lerot, C. (2017). OMI air-quality monitoring over the Middle East. Atmospheric Chemistry and Physics , 17(7), 4687–4709. Bouhamra, W., & Abdul-Wahab, S. (1999). Description of outdoor air quality in a typical residential area in Kuwait. Environmental pollution , 105(2), 221–229. Brown, O., & Crawford, A. (2009). Rising temperatures, Rising tensions. Climate change and the risk of violent conflict in the Middle East. Winnipeg: International Institute for Sustainable Development . Ettouney, R. S., Zaki, J. G., El-Rifai, M. A., & Ettouney, H. M. (2010). An assessment of the air pollution data from two monitoring stations in Kuwait. Toxicological and Environ Chemistry , 92(4), 655–668. Gross, J. E., Carlos, W. G., Dela Cruz, C. S., Harber, P., & Jamil, S. (2018). Sand and dust storms: Acute exposure and threats to respiratory health. American journal of respiratory and critical care medicine , 198(7), P13-P14. Grzywa-Celińska, A., Krusiński, A., & Milanowski, J. (2020). ‘Smoging kills’-effects of air pollution on human respiratory system. Annals of Agricultural and Environmental Medicine , 27(1), 1–5. Habeebullah, T. M., Munir, S., AA, A. H., Morsy, E. A., Seroji, A. R., & Mohammed, A. M. (2015). The interaction between air quality and meteorological factors in an arid environment of Makkah, Saudi Arabia. International Journal of Environmental Science and Development , 6(8), 576. He, F., Shaffer, M. L., Rodriguez-Colon, S., Yanosky, J. D., Bixler, E., Cascio, W. E., et al. (2011). Acute effects of fine particulate air pollution on cardiac arrhythmia: the APACR study. Environmental health perspectives , 119(7), 927–932. Hwang, B.-F., & Lee, Y. L. (2010). Air pollution and prevalence of bronchitic symptoms among children in Taiwan. Chest , 138(4), 956–964. Kouis, P., Galanakis, E., Michaelidou, E., Kinni, P., Michanikou, A., Pitsios, C., et al. (2024). Improved childhood asthma control after exposure reduction interventions for desert dust and anthropogenic air pollution: the MEDEA randomised controlled trial. thorax , 79(6), 495–507. Li, J., Garshick, E., Al-Hemoud, A., Huang, S., & Koutrakis, P. (2020). Impacts of meteorology and vegetation on surface dust concentrations in Middle Eastern countries. Science of the total environment , 712, 136597. Lipfert, F. W. (2018). Long-term associations of morbidity with air pollution: A catalog and synthesis. Journal of the Air & Waste Management Association , 68(1), 12–28. Liu, Z., Fang, C., Sun, B., & Liao, X. (2023). Governance matters: Urban expansion, environmental regulation, and PM2. 5 pollution. Science of The Total Environment , 876, 162788. Luck, T. (2023). Climate Priorities in the Middle East and North Africa: Examining Nationally Defined Contributions, Targets, and Gaps in Wealthy Versus Middle-Income States. Wilson Center . Manisalidis, I., Stavropoulou, E., Stavropoulos, A., & Bezirtzoglou, E. (2020). Environmental and health impacts of air pollution: a review. Frontiers in public health , 8, 14. Martínez Vallejo, L. A., Hernández Pardo, M. A., Benavides Piracón, J. A., Belalcázar Cerón, L. C., & Molina Achury, N. J. (2021). Exposure levels to PM2. 5 and black carbon for people with disabilities in rural homes of Colombia. Environmental Monitoring and Assessment , 193(1), 37. Munir, S., Habeebullah, T. M., Mohammed, A. M., Morsy, E. A., Rehan, M., & Ali, K. (2017). Analysing PM2. 5 and its association with PM10 and meteorology in the arid climate of Makkah, Saudi Arabia. Aerosol and Air Quality Research , 17(2), 453–464. Omar, A. H., Tackett, J., & Al-Dousari, A. (2022). CALIPSO observations of sand and dust storms and comparisons of source types near Kuwait City. Atmosphere , 13(12), 1946. Organization, W. H. (2021). WHO global air quality guidelines: particulate matter (PM2. 5 and PM10), ozone, nitrogen dioxide, sulfur dioxide and carbon monoxide. World Health Organization . Ramadan, M. S., Abuelgasim, A., & Al Hosani, N. (2024). Advancing air quality forecasting in Abu Dhabi, UAE using time series models. Frontiers in Environmental Science , 12, 1393878. Sabbah, I., Léon, J.-F., Sorribas, M., Guinot, B., Córdoba-Jabonero, C., de Souza, A., et al. (2018). Dust and dust storms over Kuwait: Ground-based and satellite observations. Journal of Atmospheric and Solar-Terrestrial Physics , 179, 105–113. Shang, Y., Sun, Z., Cao, J., Wang, X., Zhong, L., Bi, X., et al. (2013). Systematic review of Chinese studies of short-term exposure to air pollution and daily mortality. Environment international , 54, 100–111. Sitanggang, J. W., Sunarsih, E., & Hasyim, H. (2023). Literature Review: Analysis of Exposure of Vehicle Emission Gases (Co, No2, So2, Pm2. 5, and Pm10) to Public Health Risks. Journal of Social Research , 2(7), 2278–2287. Uddin, S., Habibi, N., Fowler, S. W., Behbehani, M., Gevao, B., Faizuddin, M., et al. (2023). Aerosols as vectors for contaminants: a perspective based on outdoor aerosol data from Kuwait. Atmosphere , 14(3), 470. Zhao, Y., Cheng, Z., Lu, Y., Chang, X., Chan, C., Bai, Y., et al. (2017). PM10 and PM2. 5 particles as main air pollutants contributing to rising risks of coronary heart disease: a systematic review. Environmental Technology Reviews , 6(1), 174–185. Additional Declarations No competing interests reported. Cite Share Download PDF Status: Posted Version 1 posted You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-8776307","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":600679537,"identity":"09ba8767-a5fd-4de4-8301-25c612dafb28","order_by":0,"name":"Raslan Alenezi","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAABA0lEQVRIiWNgGAWjYLCCByCCmfkAjG/AwHAAh1IYSABrYQNTEiRoYeAxIE6LefsBxg8JFXfk5Nt5vj34UWNTx8DevE2C4cxhnFpkziQwSySceWbM2My73bDnWJoEA8+xMgmGG7i1SABdJZHYdjixmZl3mwQP22EJBokcMwmGD7dxa+F/wPwj8d/h+jZmnmeSf/79l2CQf0NAi0QCm0Riw+EEHmYeNmnetgNAW3iAWm7g0/KwzSLh2GHDGcxs5sayfcmSbTxpxRYJZ/7jcVjy4Rsfag7Ly/cffvbwzTc7fn72wxtvfDiWhlMLAwNjA4zFhiAT8GhABmxEqhsFo2AUjIKRBgBmDU96gn0xAwAAAABJRU5ErkJggg==","orcid":"","institution":"Public Authority for Applied Education and Training","correspondingAuthor":true,"prefix":"","firstName":"Raslan","middleName":"","lastName":"Alenezi","suffix":""},{"id":600679540,"identity":"16653f11-f59b-4021-9a80-fb0768abe5e6","order_by":1,"name":"Naeema Al-Darmaki","email":"","orcid":"","institution":"United Arab Emirates University","correspondingAuthor":false,"prefix":"","firstName":"Naeema","middleName":"","lastName":"Al-Darmaki","suffix":""},{"id":600679541,"identity":"b9ed3485-941c-448d-91f1-f8999be732f6","order_by":2,"name":"Fahed Javed","email":"","orcid":"","institution":"United Arab Emirates University","correspondingAuthor":false,"prefix":"","firstName":"Fahed","middleName":"","lastName":"Javed","suffix":""},{"id":600679542,"identity":"dae953b9-a4d4-41d9-99f0-4c635bfbb01e","order_by":3,"name":"Sami Walid Azzam","email":"","orcid":"","institution":"United Arab Emirates University","correspondingAuthor":false,"prefix":"","firstName":"Sami","middleName":"Walid","lastName":"Azzam","suffix":""}],"badges":[],"createdAt":"2026-02-03 13:11:03","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-8776307/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-8776307/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":104245734,"identity":"1671f6f5-a30d-477b-9902-cad2bfc6198f","added_by":"auto","created_at":"2026-03-09 15:20:31","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":295055,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eLocation map of the selected monitoring stations (Al-Hurban et al., 2021).\u003c/strong\u003e\u003c/p\u003e","description":"","filename":"1.png","url":"https://assets-eu.researchsquare.com/files/rs-8776307/v1/1a90c0b36fcbfffc4ce87c59.png"},{"id":104405092,"identity":"4e63e64c-d569-446b-a388-8f17358a14c0","added_by":"auto","created_at":"2026-03-11 12:21:44","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":308095,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eThe location of different air quality monitoring stations is specified by a colored bullet over the map distributed of the northern part, the urban area, and the southern part of Kuwait.\u003c/strong\u003e\u003c/p\u003e","description":"","filename":"2.png","url":"https://assets-eu.researchsquare.com/files/rs-8776307/v1/dc363d50f9904d9e661bcd8e.png"},{"id":104405164,"identity":"a4ab1ac1-4ad8-4851-aaaf-9f3835a749c1","added_by":"auto","created_at":"2026-03-11 12:21:58","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":58168,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003esystematic flow diagram of time-series modeling and evaluation.\u003c/strong\u003e\u003c/p\u003e","description":"","filename":"3.png","url":"https://assets-eu.researchsquare.com/files/rs-8776307/v1/c296e4af3ccf25ec01fdb6c1.png"},{"id":104245731,"identity":"eb72eae9-6374-42a3-8af5-d07a53a1113c","added_by":"auto","created_at":"2026-03-09 15:20:31","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":87680,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eYearly average data from monitoring stations of Kuwait.\u003c/strong\u003e\u003c/p\u003e","description":"","filename":"4.png","url":"https://assets-eu.researchsquare.com/files/rs-8776307/v1/2529921bef5f5d2bd143073e.png"},{"id":104404201,"identity":"084cdb39-b8ff-4477-95e9-84bd02b5c4cd","added_by":"auto","created_at":"2026-03-11 12:19:49","extension":"png","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":83842,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eSeasonal data of all the stations over the years.\u003c/strong\u003e\u003c/p\u003e","description":"","filename":"5.png","url":"https://assets-eu.researchsquare.com/files/rs-8776307/v1/c221dfb6a8b001f171a5785a.png"},{"id":104245730,"identity":"8af30e03-ee1a-4623-b247-62ea4a50e881","added_by":"auto","created_at":"2026-03-09 15:20:31","extension":"png","order_by":6,"title":"Figure 6","display":"","copyAsset":false,"role":"figure","size":124713,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eForecasting behavior for air pollutants using LSTM and ARIMA.\u003c/strong\u003e\u003c/p\u003e","description":"","filename":"6.png","url":"https://assets-eu.researchsquare.com/files/rs-8776307/v1/2ed115d7fb1524c25c4371d8.png"},{"id":104779758,"identity":"4d0a0b83-13c2-4aee-850f-b9a70dcadf26","added_by":"auto","created_at":"2026-03-17 07:45:52","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":2102306,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-8776307/v1/d5ba2722-fa72-46c0-b897-c586e97be8fd.pdf"}],"financialInterests":"No competing interests reported.","formattedTitle":"\u003cp\u003eAssessment and Prediction of Air Pollution Trends in Kuwait Using Machine Learning: An Analysis of PM\u003csub\u003e10\u003c/sub\u003e , CO, and SO\u003csub\u003e2\u003c/sub\u003e and Their Environmental Health Implications\u003c/p\u003e","fulltext":[{"header":"1 Introduction","content":"\u003cp\u003eAir quality deterioration has emerged as one of the most pressing environmental challenges worldwide. Rapid population growth, urbanization, industrialization, and increasing energy consumption have contributed to elevated air pollution levels, particularly in urban regions (Bouhamra and Abdul-Wahab, \u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e1999\u003c/span\u003e; Raslan A Alenezi and Aldaihani, \u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e2023\u003c/span\u003e). Although natural phenomena such as wildfires and volcanic eruptions emit pollutants, anthropogenic activities remain the dominant sources of environmental air pollution (Raslan A Alenezi, \u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e2022\u003c/span\u003e). Pollutants such as sulfur dioxide (SO\u003csub\u003e2\u003c/sub\u003e), nitrogen dioxide (NO\u003csub\u003e2\u003c/sub\u003e), carbon monoxide (CO), carbon dioxide (CO\u003csub\u003e2\u003c/sub\u003e), ozone (O\u003csub\u003e3\u003c/sub\u003e), and particulate matter (PM\u003csub\u003e10\u003c/sub\u003e and PM\u003csub\u003e2.5\u003c/sub\u003e) significantly degrade air quality, affecting both human health and ecosystems. Numerous studies have demonstrated strong associations between air pollution exposure and increased risks of respiratory and cardiovascular diseases (Hwang and Lee, \u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e2010\u003c/span\u003e; Shang et al., \u003cspan citationid=\"CR37\" class=\"CitationRef\"\u003e2013\u003c/span\u003e; He et al., \u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e2011\u003c/span\u003e). Kuwait represents a particularly critical case for studying air pollution due to its unique combination of environmental, industrial, and socioeconomic factors. As one of the world\u0026rsquo;s largest oil producers and a rapidly developing nation, Kuwait\u0026rsquo;s air quality faces significant challenges from both anthropogenic and natural sources. The country\u0026rsquo;s expanding population, intensive energy consumption, and rapid industrialization, especially in the petrochemical and oil refining sectors, have made it one of the most polluted countries in Southwest Asia (Barkley et al., \u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e2017\u003c/span\u003e; Mart\u0026iacute;nez Vallejo et al., \u003cspan citationid=\"CR31\" class=\"CitationRef\"\u003e2021\u003c/span\u003e). Moreover, the country\u0026rsquo;s harsh desert climate and frequent dust storms further exacerbate airborne particulate concentrations, with more than 270 tons of dust deposited per km\u0026sup2; annually in Kuwait City, the highest in the world (Al-Hemoud et al., \u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e2018\u003c/span\u003e). These natural and human-induced conditions combine to create persistent air quality challenges, making Kuwait an ideal environment for assessing pollutant dynamics and their potential impacts on public health.\u003c/p\u003e \u003cp\u003eAmong the major contributors to Kuwait\u0026rsquo;s deteriorating air quality are vehicular emissions, industrial discharges, and dust events. The transport sector plays a dual role in Kuwait\u0026rsquo;s economy, supporting rapid development while simultaneously producing large volumes of pollutants such as CO, CO\u003csub\u003e2\u003c/sub\u003e, NO\u003csub\u003ex\u003c/sub\u003e, and volatile organic compounds (VOCs) (Ettouney et al., \u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e2010\u003c/span\u003e; Raslan Alenezi et al., \u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e2012\u003c/span\u003e). Between 2001 and 2010, the number of vehicles increased by approximately 6% annually, far outpacing the 2% expansion in road networks. This imbalance has resulted in severe traffic congestion and elevated vehicular emissions. Combined with heavily subsidized fuel prices and limited public transport, these factors have intensified air pollution and increased greenhouse gas emissions (Aldaihani and Alenezi, \u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e2017\u003c/span\u003e; Al-Khulaifi et al., \u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e2014\u003c/span\u003e). Despite technological inspections and modern vehicles, Kuwait\u0026rsquo;s transport sector remains one of the primary sources of CO\u003csub\u003e2\u003c/sub\u003e and CO emissions, contributing substantially to climate change and local air degradation.\u003c/p\u003e \u003cp\u003eFurthermore, Kuwait\u0026rsquo;s geographic location and climatic conditions make it especially vulnerable to dust-related pollution. Frequent dust storms elevate levels of fine and coarse particulate matter (PM\u003csub\u003e2.5\u003c/sub\u003e and PM\u003csub\u003e10\u003c/sub\u003e), often exceeding World Health Organization (WHO) air quality guidelines (Alkharafi, \u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e2025\u003c/span\u003e). Such elevated concentrations are associated with various respiratory and cardiovascular health effects (Li et al., \u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e2020\u003c/span\u003e; Al-Enezi et al., \u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e2024\u003c/span\u003e). These challenges highlight the urgency of understanding the spatial and temporal behavior of key air pollutants in Kuwait to inform environmental management and health protection strategies.\u003c/p\u003e \u003cp\u003eHence, the present study focuses on assessing the levels and temporal variations of PM\u003csub\u003e10\u003c/sub\u003e, CO, and SO\u003csub\u003e2\u003c/sub\u003e across Kuwait. Data was collected from ten air quality monitoring stations over a four-year period to analyze pollutant trends and seasonal fluctuations. Using machine learning techniques, this research models and predicts future pollutant concentrations and evaluates their potential environmental and health impacts. The findings aim to support policymakers in developing evidence-based strategies for improving air quality and safeguarding public health in Kuwait and other arid regions facing similar environmental challenges.\u003c/p\u003e"},{"header":"2 Material and methods","content":"\u003cdiv id=\"Sec3\" class=\"Section2\"\u003e \u003ch2\u003e2.1 Data collection\u003c/h2\u003e \u003cp\u003eThe country under study is Kuwait, located in the northeastern corner of the Arabian Peninsula. Its surface elevation ranges from approximately 280 m above sea level in the southwest to sea level along the eastern Gulf Coast. Kuwait shares land borders with Saudi Arabia to the south and Iraq to the north and west, and a maritime boundary with Iran across the Arabian Gulf (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e). The nation\u0026rsquo;s economy is heavily dependent on the oil and gas industry, which plays a central role in both economic growth and environmental challenges.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eTo monitor air quality across the country, the Kuwait Environment Public Authority (KU-EPA) operates a network of fixed Air Quality Monitoring Stations (AQMS) that continuously measure concentrations of key atmospheric pollutants. The spatial distribution of these AQMS is illustrated on the map (Figs.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003e), where every station is represented by a colored dot indicating its location. These stations provide critical data for evaluating spatial and temporal trends in air pollution levels across Kuwait from which we select 10 station to get overall data of Kuwait.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec4\" class=\"Section2\"\u003e \u003ch2\u003e2.2 Time-Series Modelling Methodology\u003c/h2\u003e \u003cp\u003eThis study employs time-series modelling to analyze and forecast monthly variations in selected air pollutants using aggregated monitoring data. The methodological framework focuses on constructing consistent monthly time series, addressing missing values and extreme observations, and applying statistical and machine-learning models suitable for short to medium-term forecasting, as shown in Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003e.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cdiv id=\"Sec5\" class=\"Section3\"\u003e \u003ch2\u003e2.2.1 Construction of Monthly Averaged Time Series\u003c/h2\u003e \u003cp\u003eMonthly air quality data for CO, PM\u003csub\u003e10\u003c/sub\u003e, and SO₂ were obtained from multiple monitoring stations. For each pollutant, only concentration values reported in \u0026micro;g/m\u0026sup3; were retained to ensure unit consistency. The original datasets were re-organized into a unified format containing station identifiers, year and month, and pollutant concentrations. Monthly pollutant concentrations were then spatially averaged across all available stations to produce a single regional-scale time series per pollutant. This approach reduces station-specific noise and yields 64 monthly observations for each pollutant, which is appropriate for global forecasting models such as ARIMA and LSTM.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec6\" class=\"Section3\"\u003e \u003ch2\u003e2.2.2 Treatment of Missing Values and Outliers\u003c/h2\u003e \u003cp\u003eMissing values were handled using time-based linear interpolation applied separately to each station-level monthly series before spatial averaging. This preserves temporal continuity while avoiding artificial discontinuities. CO and SO₂ series contained no missing values, whereas PM₁₀ included a limited number of missing months that were successfully interpolated. Exploratory analysis revealed the presence of extreme PM₁₀ values that strongly influenced model behaviour. To assess model robustness, two versions of the PM₁₀ series were considered i.e. an unclipped series, retaining all observed values, and a clipped series, where station-level PM₁₀ concentrations exceeding 800 \u0026micro;g/m\u003csup\u003e3\u003c/sup\u003e were truncated before recomputing monthly spatial averages. This threshold was chosen to reduce the influence of a small number of extreme observations while retaining the overall structure of high-pollution periods.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec7\" class=\"Section3\"\u003e \u003ch2\u003e2.2.3 ARIMA Model Specification\u003c/h2\u003e \u003cp\u003eSeasonal Autoregressive Integrated Moving Average (SARIMA) models were used as a classical statistical benchmark for forecasting monthly pollutant concentrations. Given the monthly resolution of the data, a seasonal period of 12 months was adopted. Model identification was based on an Akaike Information Criterion (AIC) grid search over plausible combinations of autoregressive, differencing, and moving-average orders. The optimal configuration for each pollutant was selected as the model with the lowest AIC value. All ARIMA models were univariate and did not include exogenous predictors, relying solely on past pollutant values to generate forecasts.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec8\" class=\"Section3\"\u003e \u003ch2\u003e2.2.4 LSTM Neural Network Model\u003c/h2\u003e \u003cp\u003eTo capture potential nonlinear and long-range temporal dependencies, a univariate Long Short-Term Memory (LSTM) neural network was implemented for each pollutant using the spatially averaged monthly series. Prior to modelling, each series was scaled to the [0,1] range using min\u0026ndash;max normalisation. The time series was transformed into supervised learning samples using a six-month sliding window, where the previous six months were used to predict the subsequent month. The LSTM architecture consisted of a single LSTM layer followed by fully connected layers, trained using the Adam optimizer and mean squared error loss. Early stopping was applied to limit overfitting. For PM₁₀, separate LSTM models were trained on the unclipped and clipped datasets to explicitly examine the effect of extreme values on model performance.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec9\" class=\"Section3\"\u003e \u003ch2\u003e2.2.5 Model Evaluation Metrics\u003c/h2\u003e \u003cp\u003eModel performance was assessed using standard error metrics commonly applied in air-quality forecasting studies i.e. Mean Absolute Percentage Error (MAPE), Mean Absolute Error (MAE), Root Mean Squared Error (RMSE). For ARIMA models, these metrics were computed using a 12-month hold-out validation window, with one-step-ahead forecasts compared against observed values. LSTM performance was primarily evaluated using in-sample RMSE, reflecting the exploratory nature of the neural-network implementation.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec10\" class=\"Section3\"\u003e \u003ch2\u003e2.2.6 Comparative performance analysis\u003c/h2\u003e \u003cp\u003eModel comparison was conducted at multiple levels to provide a comprehensive assessment of forecast performance. First, a within-pollutant comparison was performed by contrasting the predictive accuracy of ARIMA and LSTM models for each individual pollutant. Second, an across-pollutant comparison examined differences in forecastability among CO, PM\u003csub\u003e10\u003c/sub\u003e, and SO\u003csub\u003e2\u003c/sub\u003e, acknowledging that each pollutant exhibits distinct variability and temporal characteristics. Finally, sensitivity to extreme values was evaluated by comparing PM\u003csub\u003e10\u003c/sub\u003e model results with and without outlier clipping. This structured comparison framework enables a balanced evaluation of the relative strengths and limitations of classical statistical models and machine-learning approaches under realistic data constraints.\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e"},{"header":"3 Results and discussion","content":"\u003cdiv id=\"Sec12\" class=\"Section2\"\u003e \u003ch2\u003e3.1 Variability analysis of different stations in Kuwait\u003c/h2\u003e \u003cp\u003eTable\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e presents the statistical summary of CO, SO\u003csub\u003e2\u003c/sub\u003e, and PM\u003csub\u003e10\u003c/sub\u003e concentrations recorded across ten monitoring stations in Kuwait from Jan-2020 to Dec-2024, including minimum, maximum, mean, and standard deviation values. The results indicate marked spatial variability among stations, reflecting the diverse environmental and emission characteristics across the country. The variability patterns highlight the influence of industrial activities, roadway traffic emissions, and natural dust events on local air quality. The analysis of CO concentrations reveals substantial differences among stations. Al-Mutla exhibits the highest mean CO concentration (1,572.04 \u0026micro;g/m\u003csup\u003e3\u003c/sup\u003e), accompanied by a relatively high standard deviation, indicating persistently elevated levels with frequent fluctuations. Similarly, Al-Shuaiba and Al-Ahmadi show markedly high CO means (1,472.61 \u0026micro;g/m\u003csup\u003e3\u003c/sup\u003e and 1,159.19 \u0026micro;g/m\u003csup\u003e3\u003c/sup\u003e, respectively), suggesting the presence of consistent emission sources, likely tied to industrial operations and heavy vehicular movement. In contrast, stations such as Al-Jahra and Al-Rumaithiya recorded much lower CO means (486.27 \u0026micro;g/m\u003csup\u003e3\u003c/sup\u003e and 563.10 \u0026micro;g/m\u003csup\u003e3\u003c/sup\u003e), with comparatively smaller ranges between minimum and maximum values, reflecting more stable conditions and fewer intense emission episodes. The exceptionally high maximum values observed at several sites (e.g., 3,395.38 \u0026micro;g/m\u003csup\u003e3\u003c/sup\u003e at Al-Shuaiba) point toward episodic industrial plumes or localized high-emission events.\u003c/p\u003e \u003cp\u003eSO\u003csub\u003e2\u003c/sub\u003e levels across the monitoring network display lower overall variability compared to CO, yet notable differences persist among stations. Al-Ahmadi stands out with the highest mean SO\u003csub\u003e2\u003c/sub\u003e concentration (33.20 \u0026micro;g/m\u003csup\u003e3\u003c/sup\u003e), which may be indicative of combustion-related activities or refinery emissions in the surrounding area. Other stations, such as Al-Fahaheel and Saad Al-Abdullah, recorded moderate SO\u003csub\u003e2\u003c/sub\u003e means (24.29 and 24.37 \u0026micro;g/m\u003csup\u003e3\u003c/sup\u003e, respectively), while stations including Al-Mutla and Al-Rumaithiya show relatively low mean concentrations (12.46 and 19.07 \u0026micro;g/m\u0026sup3;). The narrow ranges and modest standard deviations for SO\u003csub\u003e2\u003c/sub\u003e at many locations suggest that sulfur-related emissions are more spatially concentrated and less temporally volatile than CO or PM\u003csub\u003e10\u003c/sub\u003e.\u003c/p\u003e \u003cp\u003e.\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab1\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 1\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eAir pollutants statistical analysis by different stations in Kuwait from Jan-2020 to Dec-2024.\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"14\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c6\" colnum=\"6\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c7\" colnum=\"7\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c8\" colnum=\"8\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c9\" colnum=\"9\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c10\" colnum=\"10\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c11\" colnum=\"11\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c12\" colnum=\"12\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c13\" colnum=\"13\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c14\" colnum=\"14\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003eSr.no.\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003eStation name\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colspan=\"4\" nameend=\"c6\" namest=\"c3\"\u003e \u003cp\u003eCO (\u0026micro;g/m\u003csup\u003e3\u003c/sup\u003e)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colspan=\"4\" nameend=\"c10\" namest=\"c7\"\u003e \u003cp\u003eSO\u003csub\u003e2\u003c/sub\u003e (\u0026micro;g/m\u003csup\u003e3\u003c/sup\u003e)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colspan=\"4\" nameend=\"c14\" namest=\"c11\"\u003e \u003cp\u003ePM\u003csub\u003e10\u003c/sub\u003e (\u0026micro;g/m\u003csup\u003e3\u003c/sup\u003e)\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eMin\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003eMax\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c5\"\u003e \u003cp\u003eMean\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c6\"\u003e \u003cp\u003eSD\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c7\"\u003e \u003cp\u003eMin\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c8\"\u003e \u003cp\u003eMax\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c9\"\u003e \u003cp\u003eMean\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c10\"\u003e \u003cp\u003eSD\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c11\"\u003e \u003cp\u003eMin\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c12\"\u003e \u003cp\u003eMax\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c13\"\u003e \u003cp\u003eMean\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c14\"\u003e \u003cp\u003eSD\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e\u003cb\u003eAl-Ahmadi\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e410.55\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e3280.03\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e1159.19\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e436.63\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e14.64\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e76.95\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e \u003cp\u003e33.20\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c10\"\u003e \u003cp\u003e10.80\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c11\"\u003e \u003cp\u003e38.55\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c12\"\u003e \u003cp\u003e362.28\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c13\"\u003e \u003cp\u003e114.01\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c14\"\u003e \u003cp\u003e62.66\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e\u003cb\u003eAl-Fahaheel\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e504.05\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e1416.69\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e1015\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e185.15\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e3.77\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e47.29\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e \u003cp\u003e24.29\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c10\"\u003e \u003cp\u003e9.25\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c11\"\u003e \u003cp\u003e53.16\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c12\"\u003e \u003cp\u003e4525.39\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c13\"\u003e \u003cp\u003e206.48\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c14\"\u003e \u003cp\u003e582.36\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e3\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e\u003cb\u003eAl-Jahra\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e112.36\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e2081.50\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e486.27\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e386.94\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e6.99\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e73.55\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e \u003cp\u003e19.99\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c10\"\u003e \u003cp\u003e9.86\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c11\"\u003e \u003cp\u003e60.98\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c12\"\u003e \u003cp\u003e387.04\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c13\"\u003e \u003cp\u003e144.63\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c14\"\u003e \u003cp\u003e71.34\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e4\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e\u003cb\u003eAl-Mutla\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e844.56\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e2395.91\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e1572.04\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e367.48\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e5.42\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e24.45\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e \u003cp\u003e12.46\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c10\"\u003e \u003cp\u003e3.82\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c11\"\u003e \u003cp\u003e74.08\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c12\"\u003e \u003cp\u003e1258.56\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c13\"\u003e \u003cp\u003e210.49\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c14\"\u003e \u003cp\u003e168.58\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e5\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e\u003cb\u003eAl-Rumaithiya\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e117.99\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e1750.76\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e563.10\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e329.45\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e6.65\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e40.28\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e \u003cp\u003e19.07\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c10\"\u003e \u003cp\u003e7.17\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c11\"\u003e \u003cp\u003e110.04\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c12\"\u003e \u003cp\u003e43.48\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c13\"\u003e \u003cp\u003e421.60\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c14\"\u003e \u003cp\u003e62.43\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e6\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e\u003cb\u003eAl-Salem\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e427.11\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e2784.61\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e1022.08\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e463.19\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e5.42\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e43.27\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e \u003cp\u003e27.09\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c10\"\u003e \u003cp\u003e7.25\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c11\"\u003e \u003cp\u003e43.65\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c12\"\u003e \u003cp\u003e378.86\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c13\"\u003e \u003cp\u003e114.48\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c14\"\u003e \u003cp\u003e63.67\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e7\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e\u003cb\u003eAl-Shuaiba\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e267.84\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e3395.38\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e1472.61\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e952.56\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e13.56\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e49.69\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e \u003cp\u003e23.68\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c10\"\u003e \u003cp\u003e23.68\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c11\"\u003e \u003cp\u003e55.74\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c12\"\u003e \u003cp\u003e377.50\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c13\"\u003e \u003cp\u003e125.48\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c14\"\u003e \u003cp\u003e56.66\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e8\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e\u003cb\u003eAl-Shuwaikh\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e478.52\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e1754.44\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e916.75\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e315.03\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e10.21\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e40.37\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e \u003cp\u003e22.26\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c10\"\u003e \u003cp\u003e6.86\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c11\"\u003e \u003cp\u003e52.14\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c12\"\u003e \u003cp\u003e225.44\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c13\"\u003e \u003cp\u003e109.49\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c14\"\u003e \u003cp\u003e35.86\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e9\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e\u003cb\u003eAl-Subah Al Salem 1\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e624.11\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e2219.62\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e943.11\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e277.40\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e7.42\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e31.82\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e \u003cp\u003e16.22\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c10\"\u003e \u003cp\u003e5.27\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c11\"\u003e \u003cp\u003e86.16\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c12\"\u003e \u003cp\u003e474.06\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c13\"\u003e \u003cp\u003e185.89\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c14\"\u003e \u003cp\u003e81.22\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e10\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e\u003cb\u003eSaad Al-Abdullah\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e398.48\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e3244.04\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e853.73\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e386.14\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e8.27\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e45.16\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e \u003cp\u003e24.37\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c10\"\u003e \u003cp\u003e8.41\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c11\"\u003e \u003cp\u003e67.20\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c12\"\u003e \u003cp\u003e424.08\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c13\"\u003e \u003cp\u003e158.08\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c14\"\u003e \u003cp\u003e75.05\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003ePM\u003csub\u003e10\u003c/sub\u003e concentrations exhibit the most pronounced spatial and temporal variability among the three pollutants. Several stations recorded exceptionally high maximum PM\u003csub\u003e10\u003c/sub\u003e values, such as Al-Fahaheel (4,525.39 \u0026micro;g/m\u003csup\u003e3\u003c/sup\u003e) and Al-Mutla (1,258.56 \u0026micro;g/m\u003csup\u003e3\u003c/sup\u003e), reflecting the strong influence of dust storms and resuspended desert particles, which are characteristic of Kuwait\u0026rsquo;s arid climate. Mean PM\u003csub\u003e10\u003c/sub\u003e levels were highest at stations such as Al-Subah Al-Salem 1 (185.89 \u0026micro;g/m\u003csup\u003e3\u003c/sup\u003e) and Saad Al-Abdullah (158.08 \u0026micro;g/m\u003csup\u003e3\u003c/sup\u003e), indicating persistent exposure to coarse particulate matter. Conversely, stations like Al-Rumaithiya and Al-Salem registered considerably lower PM\u003csub\u003e10\u003c/sub\u003e means (62.43 and 114.48 \u0026micro;g/m\u003csup\u003e3\u003c/sup\u003e, respectively), likely reflecting their relative distance from industrial zones or major dust-entrainment areas. The large standard deviations associated with PM\u003csub\u003e10\u003c/sub\u003e at numerous stations underscore the episodic nature of particulate pollution, driven by both natural events and anthropogenic activities\u003c/p\u003e \u003cp\u003eOverall, this station variability analysis reveals clear differences in pollutant behavior across Kuwait\u0026rsquo;s monitoring network. Stations situated near industrial complexes or major roadway corridors consistently exhibit elevated mean concentrations of CO and PM\u003csub\u003e10\u003c/sub\u003e, while SO\u003csub\u003e2\u003c/sub\u003e levels remain more localized and comparatively stable. The pronounced fluctuations in PM₁₀ further emphasize the combined effect of natural dust activity and human-induced emissions on Kuwait\u0026rsquo;s air quality. These station-specific patterns underscore the need for targeted air quality management strategies that consider local emission sources, land-use characteristics, and regional meteorological factors that shape pollutant dispersion and accumulation.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec13\" class=\"Section2\"\u003e \u003ch2\u003e3.2 Pollutant historic trend in Kuwait\u003c/h2\u003e \u003cp\u003eFigure \u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003e illustrates the annual mean concentrations of CO, SO\u003csub\u003e2\u003c/sub\u003e and PM\u003csub\u003e10\u003c/sub\u003e, averaged over the ten monitoring stations in Kuwait for the period Jan-2020 to Dec-2024. Overall, the series shows clear interannual variability rather than a simple monotonic increase or decrease, reflecting the combined influence of changing anthropogenic activities (traffic, power generation, and industrial operations) and natural factors such as dust storm frequency and regional meteorology (Alkharafi, \u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e2025\u003c/span\u003e). Across all years, PM\u003csub\u003e10\u003c/sub\u003e annual means remain substantially higher than international health-based guidelines. According to the 2021 WHO Global Air Quality Guidelines, the recommended annual PM\u003csub\u003e10\u003c/sub\u003e guideline value is 15 \u0026micro;g/m\u0026sup3; (Organization, \u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e2021\u003c/span\u003e). In contrast, the multi-year mean values in Kuwait\u0026rsquo;s stations (Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e) lie roughly between 60 and \u0026gt;\u0026thinsp;180 \u0026micro;g/m\u0026sup3;, indicating that PM\u003csub\u003e10\u003c/sub\u003e levels exceed the WHO annual guideline throughout the study period. This pattern is consistent with previous work in Kuwait showing that PM\u003csub\u003e10\u003c/sub\u003e annual averages in both urban and industrial areas surpass both national and WHO guideline values (Alahmad et al., \u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e2021\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eThe annual behavior of PM\u003csub\u003e10\u003c/sub\u003e is driven strongly by year-to-year differences in dust storm activity and synoptic weather. Years with higher annual PM\u003csub\u003e10\u003c/sub\u003e in Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e(c), can reasonably be linked to more frequent or more intense dust episodes and stronger shamal winds, which transport dust from exposed soil in Kuwait and neighboring countries (Iraq, Syria, Saudi Arabia) (Sabbah et al., \u003cspan citationid=\"CR36\" class=\"CitationRef\"\u003e2018\u003c/span\u003e). Studies from Kuwait have shown that dust storms and suspended dust events are the dominant contributors to extreme PM\u003csub\u003e10\u003c/sub\u003e concentrations and are closely associated with hot, dry, and windy conditions (Al-Hemoud et al., \u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e2018\u003c/span\u003e). Thus, peaks in the annual PM\u003csub\u003e10\u003c/sub\u003e pollutants are likely to reflect years with more dust storms and enhanced regional dust transport, while relatively lower annual values correspond to years with fewer such events.\u003c/p\u003e \u003cp\u003eFor CO, the annual mean concentrations at all stations are on the order of a few hundred to around 1,500 \u0026micro;g/m\u0026sup3; (0.5\u0026ndash;1.6 mg/m\u0026sup3;), well below the WHO short-term guideline of 4 mg/m\u0026sup3; as a 24 h and 10 mg/m\u0026sup3; as an 8 h (Organization, \u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e2021\u003c/span\u003e). Although annual averages cannot be directly compared to short-term limits, the fact that long-term means are substantially lower than these thresholds suggests that CO is not the primary health driver in this dataset. Any variation from year to year is likely linked to changes in traffic volume and power generation demand, which can modify CO emissions from vehicles and thermal power plants. Time-series analysis in Kuwait has shown that pollutants tied to energy use and traffic often track changes in domestic energy demand and mobility patterns (Al-Hurban et al., \u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e2021\u003c/span\u003e; Al-Bassam and Khan, \u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2004\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eFor SO\u003csub\u003e2\u003c/sub\u003e, the station-averaged means (typically 12\u0026ndash;33 \u0026micro;g/m\u0026sup3;) remain below the WHO 24-h guideline of 40 \u0026micro;g/m\u0026sup3; in terms of annual averages (Organization, \u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e2021\u003c/span\u003e). However, SO\u003csub\u003e2\u003c/sub\u003e is usually more localized in space and time, with short-term peaks near industrial stacks and power plants. Studies in Kuwait have reported that SO₂ levels in industrial zones can exceed national or guideline values at certain times, even if long-term means appear moderate (Al-Rashidi et al., \u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e2005\u003c/span\u003e). Thus, interannual SO\u003csub\u003e2\u003c/sub\u003e differences in Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e(b) can be interpreted in terms of variations in industrial output, fuel sulphur content and effectiveness of emission control technologies.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec14\" class=\"Section2\"\u003e \u003ch2\u003e3.3 Seasonal and quarterly trend analysis of Kuwait\u003c/h2\u003e \u003cp\u003eFigure \u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003e shows the seasonal evolution of CO, SO\u003csub\u003e2,\u003c/sub\u003e and PM\u003csub\u003e10\u003c/sub\u003e across the monitoring network during Jan-2020 to Dec-2024, revealing a clear cyclic behavior driven primarily by Kuwait\u0026rsquo;s desert climate and regional meteorology. Among the three pollutants, PM\u003csub\u003e10\u003c/sub\u003e exhibits the strongest seasonal signal, with concentrations peaking during spring and summer and decreasing in autumn and winter. This pattern matches long-term observations in Kuwait and the wider Arabian Peninsula, where multiple satellite-based and ground-based studies have confirmed that aerosol loading and dust concentrations reach their highest levels during warm months due to shamal winds, extremely low soil moisture, and enhanced surface heating (Al-Hemoud et al., \u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e2018\u003c/span\u003e; Omar et al., \u003cspan citationid=\"CR33\" class=\"CitationRef\"\u003e2022\u003c/span\u003e). Remote-sensing analyses specifically show that aerosol optical thickness over Kuwait increases sharply from April to August, reflecting intense dust generation and transport (Uddin et al., \u003cspan citationid=\"CR39\" class=\"CitationRef\"\u003e2023\u003c/span\u003e). In winter, cooler temperatures, calmer winds and slightly higher humidity suppress dust emission, leading to the lowest dust particles.\u003c/p\u003e \u003cp\u003eThe elevated spring\u0026ndash;summer PM\u003csub\u003e10\u003c/sub\u003e levels observed in this study are also consistent with the broader Middle Eastern literature. In Kuwait, Al-Hemoud et al. (\u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e2018\u003c/span\u003e) showed that dust storms and suspended dust events are strongly associated with sharp increases in PM\u003csub\u003e10\u003c/sub\u003e and are the dominant contributors to extreme particle levels during hot, dry, and windy conditions (Al-Hemoud et al., \u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e2018\u003c/span\u003e). In another study by Alahmad et al. (\u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e2021\u003c/span\u003e) demonstrated that regional dust storms and suspended local dust together account for the majority of PM mass in Kuwait, confirming the dependence of PM\u003csub\u003e10\u003c/sub\u003e on seasonal dust climatology (Alahmad et al., \u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e2021\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eCO and SO\u003csub\u003e2\u003c/sub\u003e display more moderate but still noticeable seasonal variations. In Kuwait and similar arid cities, meteorology, particularly temperature, boundary-layer height, wind speed, and atmospheric stability, plays a central role in modulating combustion-related pollutants. Studies from Makkah (Munir et al., \u003cspan citationid=\"CR32\" class=\"CitationRef\"\u003e2017\u003c/span\u003e; Habeebullah et al., \u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e2015\u003c/span\u003e) provide evidence from environments with nearly identical climate characteristics, showing that CO and SO₂ vary systematically with heat-driven electricity demand, traffic intensity, and atmospheric mixing depth. In Kuwait, extremely high summer temperatures drive intense air-conditioning use, increasing electricity consumption and thereby fuel combustion in gas and oil-fired power plants, which contributes to seasonal rises in CO and SO₂. Although mixing heights are generally deeper in summer, persistent thermal lows and stagnant atmospheric conditions can still trap pollutants near the surface, contributing to the slight summer elevations visible in Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003e.\u003c/p\u003e \u003cp\u003eShort-term SO\u003csub\u003e2\u003c/sub\u003e behavior is also strongly influenced by the spatial clustering of major point sources such as refineries and power stations. Air-quality assessments around Kuwaiti oil and gas facilities have documented SO\u003csub\u003e2\u003c/sub\u003e enhancements downwind of industrial stacks, especially during warm seasons when dispersion conditions vary rapidly (Al-Rashidi et al., \u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e2005\u003c/span\u003e; Al-Hurban et al., \u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e2021\u003c/span\u003e). In winter, thermal inversions can form near the surface, reducing mixing and intermittently trapping SO\u003csub\u003e2\u003c/sub\u003e close to the ground even when emissions are lower. For this reason, the seasonal SO\u003csub\u003e2\u003c/sub\u003e cycle reflects a combined interplay of power-plant load, refinery activity, and seasonal atmospheric stability rather than dust-related processes.\u003c/p\u003e \u003cp\u003eOverall, these findings show that PM\u003csub\u003e10\u003c/sub\u003e is overwhelmingly controlled by natural meteorological drivers and dust climatology, with spring and summer peaks linked to dust storms, shamal winds, regional dust transport, and strong surface heating. CO and SO₂, on the other hand, show more modest seasonal modulation linked to anthropogenic factors such as fuel combustion and electricity demand, combined with the influence of atmospheric stability and synoptic conditions typical of Kuwait\u0026rsquo;s hyper-arid climate.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec15\" class=\"Section2\"\u003e \u003ch2\u003e3.4 Pearson Correlation analysis\u003c/h2\u003e \u003cp\u003ePearson correlation coefficients were used to evaluate the linear associations between air pollutant concentrations and to assess whether pollutants exhibit common temporal behavior, as shown in Table\u0026nbsp;\u003cspan refid=\"Tab2\" class=\"InternalRef\"\u003e2\u003c/span\u003e. The results show no strong inter-pollutant correlations, indicating limited temporal interdependence between most pollutant pairs. A weak negative correlation between CO and PM\u003csub\u003e10\u003c/sub\u003e suggests that these pollutants are likely influenced by different emission sources or atmospheric processes. Likewise, the negligible correlation between CO and SO\u003csub\u003e2\u003c/sub\u003e indicates largely independent behavior, reflecting distinct combustion or industrial origins and dispersion mechanisms. In contrast, a weak positive correlation between PM\u003csub\u003e10\u003c/sub\u003e and SO\u003csub\u003e2\u003c/sub\u003e may indicate partial overlap in emission sources; however, the low correlation strength suggests that common sources are not the dominant drivers of their variability.\u003c/p\u003e \u003cp\u003eOverall, the lack of strong correlations highlights the complex, multi-source nature of air pollution in Kuwait. These findings imply that pollutants respond differently to emission patterns and environmental conditions, supporting the need for multi-pollutant and source-specific air quality management strategies. The current correlation analysis is consistent with previous studies, where weak correlations were linked to independent pollutant behavior (Ramadan et al., \u003cspan citationid=\"CR35\" class=\"CitationRef\"\u003e2024\u003c/span\u003e).\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab2\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 2\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003ePearson correlation coefficients among air pollutant concentrations.\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"4\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePollutants\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eCO\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003ePM10\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003eSO\u003csub\u003e2\u003c/sub\u003e\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eCO\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e1.00\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e-0.19\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e-0.01\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003ePM\u003c/b\u003e\u003csub\u003e\u003cb\u003e10\u003c/b\u003e\u003c/sub\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e-0.19\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e1.00\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.19\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eSO\u003c/b\u003e\u003csub\u003e\u003cb\u003e2\u003c/b\u003e\u003c/sub\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e-0.01\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.19\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e1.00\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec16\" class=\"Section2\"\u003e \u003ch2\u003e3.5 Forecast Behavior and Visual Assessment for Kuwait air pollutants\u003c/h2\u003e \u003cp\u003eThe forecast behavior of the ARIMA and LSTM models was evaluated through visual inspection of the predicted trajectories in conjunction with the associated numerical performance metrics, as shown in Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003e. For CO, the ARIMA model achieved a MAPE\u0026thinsp;=\u0026thinsp;27.94%, MAE\u0026thinsp;=\u0026thinsp;270.94 \u0026micro;g/m\u003csup\u003e3\u003c/sup\u003e, and RMSE\u0026thinsp;=\u0026thinsp;335.46 \u0026micro;g/m\u003csup\u003e3\u003c/sup\u003e on the 12 month validation window. The corresponding forecast indicates that monthly spatially averaged CO concentrations are likely to remain within a similar range to recent years, with moderate seasonal oscillations. However, the widening confidence intervals beyond approximately 24 months highlight growing uncertainty in long-term projections. In contrast, the LSTM model achieved a substantially lower in-sample RMSE\u0026thinsp;=\u0026thinsp;182.05 \u0026micro;g/m\u003csup\u003e3\u003c/sup\u003e, and its fitted curve closely follows the observed monthly mean, producing a relatively smooth 36-month forecast centered around 1100\u0026ndash;1150 \u0026micro;g/m\u003csup\u003e3\u003c/sup\u003e. This suggests that the LSTM captures short-term variability in aggregated CO dynamics more effectively than the linear ARIMA framework.\u003c/p\u003e \u003cp\u003eIn case of SO\u003csub\u003e2\u003c/sub\u003e, the best ARIMA yielded a MAPE\u0026thinsp;=\u0026thinsp;25.24%, MAE\u0026thinsp;=\u0026thinsp;6.86 \u0026micro;g/m\u003csup\u003e3\u003c/sup\u003e, and RMSE\u0026thinsp;=\u0026thinsp;9.28 \u0026micro;g/m\u003csup\u003e3\u003c/sup\u003e. Although absolute errors are small due to the low concentration levels, the relatively high MAPE reflects pronounced relative variability. The ARIMA forecast shows largely stable SO\u003csub\u003e2\u003c/sub\u003e levels with wide confidence intervals. The LSTM model, by comparison, achieved a lower in sample RMSE of 5.59 \u0026micro;g/m\u003csup\u003e3\u003c/sup\u003e, and visually reproduces both local peaks and troughs more closely. Its forecast exhibits a gentle upward drift with modest seasonality, remaining within plausible bounds.\u003c/p\u003e \u003cp\u003eFor PM\u003csub\u003e10\u003c/sub\u003e, forecast behavior depends strongly on outlier treatment. When the unclipped series is used, the ARIMA model becomes numerically unstable, producing implausibly large error metrics (MAPE and RMSE on the order of 10\u003csup\u003e35\u003c/sup\u003e), indicating a breakdown of the linear seasonal model due to extreme monthly values. The LSTM remains operational on the same data, with an in sample RMSE of 82.06 \u0026micro;g/m\u003csup\u003e3\u003c/sup\u003e but still struggles to balance typical variability against extreme peaks. After applying clipping at 800 \u0026micro;g/m\u003csup\u003e3\u003c/sup\u003e, forecast behavior improves markedly for both models. The clipped-series ARIMA model achieves a MAPE of 29.68%, MAE of 32.72 \u0026micro;g/m\u003csup\u003e3\u003c/sup\u003e, and RMSE of 45.30 \u0026micro;g/m\u003csup\u003e3\u003c/sup\u003e, while the corresponding LSTM attains an in-sample RMSE of 47.18 \u0026micro;g/m\u003csup\u003e3\u003c/sup\u003e. The resulting forecasts are stable and realistic, although uncertainty remains substantial beyond one to two years due to the intrinsic volatility of PM\u003csub\u003e10\u003c/sub\u003e. This visual assessments, supported by numerical metrics, confirm that both models produce reasonable short to medium-term forecasts when data quality issues are addressed, while long-term projections remain uncertain.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec17\" class=\"Section2\"\u003e \u003ch2\u003e3.6 Comparative Performance of Forecasting Models\u003c/h2\u003e \u003cp\u003eA comparative evaluation of ARIMA and LSTM performance was conducted across pollutants using error metrics and visual diagnostics. In case of CO, the LSTM model substantially outperforms ARIMA. The LSTM achieves an RMSE of 182.05 \u0026micro;g/m\u003csup\u003e3\u003c/sup\u003e, compared with the ARIMA validation RMSE of 335.46 \u0026micro;g/m\u003csup\u003e3\u003c/sup\u003e and MAPE of 27.94%, representing an approximate 45% reduction in RMSE. While ARIMA provides a reasonable and interpretable linear baseline, the lower error and closer visual fit of the LSTM indicate its superior ability to capture nonlinear temporal patterns in aggregated CO concentrations. In case of SO\u003csub\u003e2\u003c/sub\u003e, a similar advantage is observed. The ARIMA model yields an RMSE of 9.28 \u0026micro;g/m\u003csup\u003e3\u003c/sup\u003e and MAPE of 25.24%, whereas the LSTM reduces the RMSE to 5.59 \u0026micro;g/m\u0026sup3;, corresponding to a reduction of roughly 40%. This suggests that even for pollutants with relatively smooth dynamics, nonlinear models can provide measurable gains in predictive accuracy.\u003c/p\u003e \u003cp\u003eFor PM₁₀, model performance is highly sensitive to outliers. On the unclipped series, ARIMA fails entirely due to numerical instability, while the LSTM remains usable with an RMSE of 82.06 \u0026micro;g/m\u0026sup3;. After clipping extreme values at 800 \u0026micro;g/m\u0026sup3;, both models improve substantially. In this cleaned setting, ARIMA achieves an RMSE of 45.30 \u0026micro;g/m\u0026sup3; (MAPE\u0026thinsp;=\u0026thinsp;29.68%), slightly outperforming the LSTM (RMSE\u0026thinsp;=\u0026thinsp;47.18 \u0026micro;g/m\u0026sup3;). This result highlights that for highly variable pollutants, appropriate preprocessing can make classical statistical models competitive with more complex machine-learning approaches.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec18\" class=\"Section2\"\u003e \u003ch2\u003e3.7 Health Related to Air Pollutants\u003c/h2\u003e \u003cp\u003eThe elevated pollutant concentrations observed in this study, particularly the persistently high PM\u003csub\u003e10\u003c/sub\u003e levels and the moderate SO\u003csub\u003e2\u003c/sub\u003e and CO values, carry well-documented implications for public health. Long-term evidence shows that coarse particulate matter such as PM\u003csub\u003e10\u003c/sub\u003e can penetrate deep into the respiratory tract, aggravating asthma, chronic obstructive pulmonary disease, and lung infections, while also increasing the risk of cardiovascular events through systemic inflammation and blood-pressure alterations (Grzywa-Celińska et al., \u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e2020\u003c/span\u003e; Zhao et al., \u003cspan citationid=\"CR40\" class=\"CitationRef\"\u003e2017\u003c/span\u003e). Zhao et al. (\u003cspan citationid=\"CR40\" class=\"CitationRef\"\u003e2017\u003c/span\u003e) specifically demonstrate that each 10 \u0026micro;g/m\u003csup\u003e3\u003c/sup\u003e increase in PM\u003csub\u003e10\u003c/sub\u003e is associated with a significant rise in coronary heart disease mortality, illustrating how elevated particulate levels, such as those measured across Kuwait\u0026rsquo;s stations, can cause heightened cardiovascular risk (Zhao et al., \u003cspan citationid=\"CR40\" class=\"CitationRef\"\u003e2017\u003c/span\u003e). Lipfert (\u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e2018\u003c/span\u003e) further study long term exposure of PM\u003csub\u003e10\u003c/sub\u003e further strengthens this understanding by showing that PM\u003csub\u003e10\u003c/sub\u003e has some of the most consistent long-term associations with respiratory morbidity, impaired lung function, and adverse birth outcomes among hundreds of epidemiological studies (Lipfert, \u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e2018\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eIn comparison, the SO\u003csub\u003e2\u003c/sub\u003e levels recorded in this study, although lower than PM10, remain relevant because chronic or repeated exposure to SO\u003csub\u003e2\u003c/sub\u003e irritates the airways, triggers bronchoconstriction, and exacerbates respiratory symptoms, especially among children, the elderly, and individuals with pre-existing lung disease (Grzywa-Celińska et al., \u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e2020\u003c/span\u003e). Long-term evidence also links SO\u003csub\u003e2\u003c/sub\u003e exposure to increases in respiratory and cardiovascular hospital admissions, indicating that even moderate concentrations contribute cumulatively to health burdens over time (Lipfert, \u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e2018\u003c/span\u003e). Meanwhile, the CO concentrations measured across Kuwait\u0026rsquo;s stations fall below acute toxicity thresholds but still present concerns in high-traffic or poorly ventilated settings. Chronic exposure, particularly among roadside workers and commuters, can reduce oxygen transport, intensify cardiovascular strain, and heighten vulnerability among those with existing heart conditions (Sitanggang et al., \u003cspan citationid=\"CR38\" class=\"CitationRef\"\u003e2023\u003c/span\u003e; Manisalidis et al., \u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e2020\u003c/span\u003e). Manisalidis et al. (\u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e2020\u003c/span\u003e) further emphasize that even low-level, long-term exposure to combustion-derived pollutants like CO diminishes cardiorespiratory resilience and increases overall susceptibility to disease (Manisalidis et al., \u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e2020\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eHence, these findings show that the pollutant levels identified in Kuwait, most notably the extremely high PM\u003csub\u003e10\u003c/sub\u003e concentrations and intermittent SO\u003csub\u003e2\u003c/sub\u003e and CO exposures, represent meaningful environmental health risks that align closely with established epidemiological evidence. The consistent pattern across the literature underscores that populations exposed to such pollutant mixtures face compounded risks over time, particularly vulnerable groups such as children, older adults, and outdoor workers. Overall, the evidence indicates that without targeted mitigation and sustained air-quality management, the pollutant levels documented in this study are likely to contribute to persistent and preventable respiratory and cardiovascular disease burdens in Kuwait.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec19\" class=\"Section2\"\u003e \u003ch2\u003e3.8 Strategies to Reduce Air Pollution\u003c/h2\u003e \u003cp\u003eThe findings of this study highlight the urgent need for integrated and evidence-based strategies to reduce air pollution in Kuwait, where natural dust events and anthropogenic emissions jointly drive deteriorating air quality and related health risks. Persistently elevated concentrations of PM\u003csub\u003e10\u003c/sub\u003e, intensified by frequent dust storms, regularly exceed international safety thresholds and are strongly associated with respiratory and cardiovascular diseases (Achilleos et al., \u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e2023\u003c/span\u003e; Anagnostopoulou et al., \u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e2023\u003c/span\u003e). These challenges are further exacerbated by climate change, as projections indicate rising temperatures, prolonged drought, and increased desertification across the Middle East by 2050 (Alkharafi, \u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e2025\u003c/span\u003e; Brown and Crawford, \u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e2009\u003c/span\u003e). Addressing air pollution in Kuwait requires coordinated action at local, national, and regional scales. Local governments play a central role through strengthened air quality regulations, emission reduction policies, and early warning systems that forecast dust and pollution episodes, enabling behavioral adaptation such as limiting outdoor activities during high-exposure periods (Liu et al., \u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e2023\u003c/span\u003e; Kouis et al., \u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e2024\u003c/span\u003e). Public awareness initiatives are equally critical, ensuring communities understand air-quality warnings, associated health risks, and appropriate protective responses (Gross et al., \u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e2018\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eImproving indoor air quality represents a key protective strategy, particularly during severe dust events. Enhancing building insulation, regular maintenance of HVAC systems, and installing effective filtration can significantly reduce indoor particulate infiltration. For outdoor and industrial workers, occupational protection measures including N95 respirators, scheduled work breaks, and access to clean-air shelters, can substantially reduce exposure. In parallel, wearable and fixed air-quality sensors offer valuable tools for real-time exposure assessment and targeted interventions (Al-Dousari et al., \u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e2025\u003c/span\u003e). Long-term sustainability strategies must also address emission sources. Strengthening industrial emission regulations, expanding renewable energy initiatives such as the solar project, and increasing vegetation cover to stabilize soil and limit dust resuspension are essential measures (Alkharafi, \u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e2025\u003c/span\u003e). Enhanced monitoring networks and regional cooperation, similar to the Saudi Sand and Dust Storm Warning Center, can further support evidence-based environmental governance (Luck, \u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e2023\u003c/span\u003e). Together, these integrated strategies can reduce air pollution, protect public health, and advance long-term environmental sustainability in Kuwait.\u003c/p\u003e \u003c/div\u003e"},{"header":"4 Conclusions","content":"\u003cp\u003eThis study provides a comprehensive assessment of air pollution dynamics in Kuwait by integrating multi-station observational data with time-series forecasting techniques. Analysis of PM\u003csub\u003e10\u003c/sub\u003e, CO, and SO\u003csub\u003e2\u003c/sub\u003e concentrations from Jan-2020 to Dec-2024 reveals pronounced spatial, seasonal, and temporal variability driven by the combined influence of anthropogenic emissions and natural desert processes. Among the investigated pollutants, PM\u003csub\u003e10\u003c/sub\u003e poses the most critical environmental and public health concern, with concentrations consistently exceeding World Health Organization guideline values across all monitoring stations. These elevated levels are strongly linked to Kuwait\u0026rsquo;s arid climate, frequent dust storms, and regional dust transport, underscoring the dominance of natural factors in shaping particulate pollution. In contrast, CO and SO₂ exhibit comparatively moderate concentrations and weaker seasonal variability, reflecting their closer association with localized sources such as traffic, power generation, and industrial activities. While long-term averages of these pollutants generally remain below international guideline thresholds, episodic peaks near industrial and urban hotspots highlight the need for continued regulatory oversight and targeted mitigation strategies. Correlation analysis further confirms that the three pollutants largely behave independently, reinforcing the importance of pollutant-specific management approaches. From a methodological perspective, the comparative evaluation of ARIMA and LSTM models demonstrates that machine-learning approaches can outperform classical statistical models, particularly for pollutants with nonlinear dynamics such as CO and SO\u003csub\u003e2\u003c/sub\u003e. However, for highly variable PM\u003csub\u003e10\u003c/sub\u003e data, appropriate preprocessing of extreme values is essential, after which traditional models can remain competitive. Overall, the findings emphasize that sustained air quality challenges in Kuwait require integrated solutions combining robust forecasting tools, enhanced monitoring, emission controls, and adaptive public health strategies. These insights are not only relevant for Kuwait but also for other arid and dust-prone regions facing similar environmental pressures.\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eData availability\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e\u0026nbsp;Acknowledgement\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe authors would like to acknowledge the support of the Public Authority for Applied Education and Training (PAAET) for funding sabbatical leave, the Kuwait Environment Public Authority (KU-EPA) for providing data, and the United Arab Emirates University Climate Action Program grants (number 12N142)\u003cstrong\u003e.\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFunding\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe research reported in this manuscript is funding by the\u0026nbsp;Public Authority for Applied Education and Training (PAAET).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eEthics declarations\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eEthics approval\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eNot applicable\u0026nbsp; \u0026nbsp; \u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConsent to participate\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eNot applicable.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConsent for publication\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eNot applicable.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eClinical trial number\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eNot applicable\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eCompeting interests\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe authors declare no competing interests\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\u003cli\u003e\u003cspan\u003eAchilleos, S., Michanikou, A., Kouis, P., Papatheodorou, S. I., Panayiotou, A. G., Kinni, P., et al. (2023). Improved indoor air quality during desert dust storms: the impact of the MEDEA exposure-reduction strategies. \u003cem\u003eScience of the Total Environment\u003c/em\u003e, 863, 160973.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eAl-Bassam, E., \u0026amp; Khan, A. (2004). Air pollution and road traffic in Kuwait. \u003cem\u003eWIT Transactions on the Built Environment\u003c/em\u003e, 75.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eAl-Dousari, A., Al-Khalaifah, H., Abbas, S., Alahmad, B., Omar, A., Koutrakis, P., et al. (2025). Integrating Health and Economic Perspectives: A Comprehensive Review of Dust Mitigation Policies. \u003cem\u003eEarth Systems and Environment\u003c/em\u003e, 1\u0026ndash;20.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eAl-Enezi, D., Al-Enezi, A.-A., Al-Dousari, A., \u0026amp; Aldashti, H. (2024) 'Assessing the Impact of Sand and Dust Storms on Residents\u0026rsquo; Health in the State of Kuwait' \u003cem\u003eProceedings of the International Conference on Sustainability: Developments and Innovations\u003c/em\u003e. Springer, pp. 151\u0026ndash;157.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eAl-Hemoud, A., Al-Dousari, A., Al-Shatti, A., Al-Khayat, A., Behbehani, W., \u0026amp; Malak, M. (2018). Health impact assessment associated with exposure to PM10 and dust storms in Kuwait. \u003cem\u003eAtmosphere\u003c/em\u003e, 9(1), 6.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eAl-Hurban, A., Khader, S., Alsaber, A., \u0026amp; Pan, J. (2021). Air Quality Assessment in the State of Kuwait during 2012 to 2017. \u003cem\u003eAtmosphere\u003c/em\u003e, 12(6), 678.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eAl-Khulaifi, N. M., Al-Mudhaf, H. F., Alenezi, R., Abu-Shady, A.-S. I., \u0026amp; Selim, M. I. (2014). Seasonal and temporal variations in volatile organic compounds in indoor and outdoor air in Al-Jahra City, Kuwait. \u003cem\u003eJournal of Environmental Protection\u003c/em\u003e, 2014.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eAl-Rashidi, M., Nassehi, V., \u0026amp; Wakeman, R. (2005). Air quality modelling of sulphur dioxide emission from power plants in Kuwait. \u003cem\u003eWIT Transactions on Ecology and the Environment\u003c/em\u003e, 82.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eAlahmad, B., Al-Hemoud, A., Kang, C.-M., Almarri, F., Kommula, V., Wolfson, J. M., et al. (2021). A two-year assessment of particulate air pollution and sources in Kuwait. \u003cem\u003eEnvironmental Pollution\u003c/em\u003e, 282, 117016.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eAldaihani, N., \u0026amp; Alenezi, R. (2017). Estimation of CO2 emissions of the vehicles transport sector in the state of Kuwait. \u003cem\u003eActa Chemica Malaysia\u003c/em\u003e, 1(1), 08\u0026ndash;12.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eAlenezi, R., Al-Anzi, B., Abusam, A., \u0026amp; Ashfaque, A. (2012). Seasonal influence on the ambient air quality in Al Jahra City for year 2010. \u003cem\u003eJournal of Environmental Protection\u003c/em\u003e, 3(12), 1711\u0026ndash;1718.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eAlenezi, R. A. (2022). BTEX and nitrogen oxides fume assessments in fuel dispense stations and residential areas. \u003cem\u003eJournal of Engineering Research\u003c/em\u003e, 10(3A).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eAlenezi, R. A., \u0026amp; Aldaihani, N. (2023). Investigation of the atmospheric chemical properties over Kuwait. \u003cem\u003eWater, Air, \u0026amp; Soil Pollution\u003c/em\u003e, 234(4), 216.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eAlkharafi, N. (2025). An examination of air quality in Kuwait: implications for environmental sustainability and public health. \u003cem\u003eKuwait Journal of Science\u003c/em\u003e, 100447.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eAnagnostopoulou, P., Kouis, P., Papatheodorou, S. I., Middleton, N., Papasavvas, I., Avraamides, P., et al. (2023). MEDEA randomised intervention study protocol in Cyprus, Greece and Israel for mitigation of desert dust health effects in adults with atrial fibrillation. \u003cem\u003eBMJ open\u003c/em\u003e, 13(3), e069809.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBarkley, M. P., Gonz\u0026aacute;lez Abad, G., Kurosu, T. P., Spurr, R., Torbatian, S., \u0026amp; Lerot, C. (2017). OMI air-quality monitoring over the Middle East. \u003cem\u003eAtmospheric Chemistry and Physics\u003c/em\u003e, 17(7), 4687\u0026ndash;4709.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBouhamra, W., \u0026amp; Abdul-Wahab, S. (1999). Description of outdoor air quality in a typical residential area in Kuwait. \u003cem\u003eEnvironmental pollution\u003c/em\u003e, 105(2), 221\u0026ndash;229.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBrown, O., \u0026amp; Crawford, A. (2009). Rising temperatures, Rising tensions. \u003cem\u003eClimate change and the risk of violent conflict in the Middle East. Winnipeg: International Institute for Sustainable Development\u003c/em\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eEttouney, R. S., Zaki, J. G., El-Rifai, M. A., \u0026amp; Ettouney, H. M. (2010). An assessment of the air pollution data from two monitoring stations in Kuwait. \u003cem\u003eToxicological and Environ Chemistry\u003c/em\u003e, 92(4), 655\u0026ndash;668.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGross, J. E., Carlos, W. G., Dela Cruz, C. S., Harber, P., \u0026amp; Jamil, S. (2018). Sand and dust storms: Acute exposure and threats to respiratory health. \u003cem\u003eAmerican journal of respiratory and critical care medicine\u003c/em\u003e, 198(7), P13-P14.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGrzywa-Celińska, A., Krusiński, A., \u0026amp; Milanowski, J. (2020). \u0026lsquo;Smoging kills\u0026rsquo;-effects of air pollution on human respiratory system. \u003cem\u003eAnnals of Agricultural and Environmental Medicine\u003c/em\u003e, 27(1), 1\u0026ndash;5.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHabeebullah, T. M., Munir, S., AA, A. H., Morsy, E. A., Seroji, A. R., \u0026amp; Mohammed, A. M. (2015). The interaction between air quality and meteorological factors in an arid environment of Makkah, Saudi Arabia. \u003cem\u003eInternational Journal of Environmental Science and Development\u003c/em\u003e, 6(8), 576.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHe, F., Shaffer, M. L., Rodriguez-Colon, S., Yanosky, J. D., Bixler, E., Cascio, W. E., et al. (2011). Acute effects of fine particulate air pollution on cardiac arrhythmia: the APACR study. \u003cem\u003eEnvironmental health perspectives\u003c/em\u003e, 119(7), 927\u0026ndash;932.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHwang, B.-F., \u0026amp; Lee, Y. L. (2010). Air pollution and prevalence of bronchitic symptoms among children in Taiwan. \u003cem\u003eChest\u003c/em\u003e, 138(4), 956\u0026ndash;964.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKouis, P., Galanakis, E., Michaelidou, E., Kinni, P., Michanikou, A., Pitsios, C., et al. (2024). Improved childhood asthma control after exposure reduction interventions for desert dust and anthropogenic air pollution: the MEDEA randomised controlled trial. \u003cem\u003ethorax\u003c/em\u003e, 79(6), 495\u0026ndash;507.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLi, J., Garshick, E., Al-Hemoud, A., Huang, S., \u0026amp; Koutrakis, P. (2020). Impacts of meteorology and vegetation on surface dust concentrations in Middle Eastern countries. \u003cem\u003eScience of the total environment\u003c/em\u003e, 712, 136597.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLipfert, F. W. (2018). Long-term associations of morbidity with air pollution: A catalog and synthesis. \u003cem\u003eJournal of the Air \u0026amp; Waste Management Association\u003c/em\u003e, 68(1), 12\u0026ndash;28.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLiu, Z., Fang, C., Sun, B., \u0026amp; Liao, X. (2023). Governance matters: Urban expansion, environmental regulation, and PM2. 5 pollution. \u003cem\u003eScience of The Total Environment\u003c/em\u003e, 876, 162788.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLuck, T. (2023). Climate Priorities in the Middle East and North Africa: Examining Nationally Defined Contributions, Targets, and Gaps in Wealthy Versus Middle-Income States. \u003cem\u003eWilson Center\u003c/em\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eManisalidis, I., Stavropoulou, E., Stavropoulos, A., \u0026amp; Bezirtzoglou, E. (2020). Environmental and health impacts of air pollution: a review. \u003cem\u003eFrontiers in public health\u003c/em\u003e, 8, 14.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMart\u0026iacute;nez Vallejo, L. A., Hern\u0026aacute;ndez Pardo, M. A., Benavides Pirac\u0026oacute;n, J. A., Belalc\u0026aacute;zar Cer\u0026oacute;n, L. C., \u0026amp; Molina Achury, N. J. (2021). Exposure levels to PM2. 5 and black carbon for people with disabilities in rural homes of Colombia. \u003cem\u003eEnvironmental Monitoring and Assessment\u003c/em\u003e, 193(1), 37.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMunir, S., Habeebullah, T. M., Mohammed, A. M., Morsy, E. A., Rehan, M., \u0026amp; Ali, K. (2017). Analysing PM2. 5 and its association with PM10 and meteorology in the arid climate of Makkah, Saudi Arabia. \u003cem\u003eAerosol and Air Quality Research\u003c/em\u003e, 17(2), 453\u0026ndash;464.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eOmar, A. H., Tackett, J., \u0026amp; Al-Dousari, A. (2022). CALIPSO observations of sand and dust storms and comparisons of source types near Kuwait City. \u003cem\u003eAtmosphere\u003c/em\u003e, 13(12), 1946.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eOrganization, W. H. (2021). WHO global air quality guidelines: particulate matter (PM2. 5 and PM10), ozone, nitrogen dioxide, sulfur dioxide and carbon monoxide. \u003cem\u003eWorld Health Organization\u003c/em\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eRamadan, M. S., Abuelgasim, A., \u0026amp; Al Hosani, N. (2024). Advancing air quality forecasting in Abu Dhabi, UAE using time series models. \u003cem\u003eFrontiers in Environmental Science\u003c/em\u003e, 12, 1393878.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSabbah, I., L\u0026eacute;on, J.-F., Sorribas, M., Guinot, B., C\u0026oacute;rdoba-Jabonero, C., de Souza, A., et al. (2018). Dust and dust storms over Kuwait: Ground-based and satellite observations. \u003cem\u003eJournal of Atmospheric and Solar-Terrestrial Physics\u003c/em\u003e, 179, 105\u0026ndash;113.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eShang, Y., Sun, Z., Cao, J., Wang, X., Zhong, L., Bi, X., et al. (2013). Systematic review of Chinese studies of short-term exposure to air pollution and daily mortality. \u003cem\u003eEnvironment international\u003c/em\u003e, 54, 100\u0026ndash;111.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSitanggang, J. W., Sunarsih, E., \u0026amp; Hasyim, H. (2023). Literature Review: Analysis of Exposure of Vehicle Emission Gases (Co, No2, So2, Pm2. 5, and Pm10) to Public Health Risks. \u003cem\u003eJournal of Social Research\u003c/em\u003e, 2(7), 2278\u0026ndash;2287.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eUddin, S., Habibi, N., Fowler, S. W., Behbehani, M., Gevao, B., Faizuddin, M., et al. (2023). Aerosols as vectors for contaminants: a perspective based on outdoor aerosol data from Kuwait. \u003cem\u003eAtmosphere\u003c/em\u003e, 14(3), 470.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZhao, Y., Cheng, Z., Lu, Y., Chang, X., Chan, C., Bai, Y., et al. (2017). PM10 and PM2. 5 particles as main air pollutants contributing to rising risks of coronary heart disease: a systematic review. \u003cem\u003eEnvironmental Technology Reviews\u003c/em\u003e, 6(1), 174\u0026ndash;185.\u003c/span\u003e\u003c/li\u003e\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":true,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true},"keywords":"Air quality assessment, Particular matter, Kuwait, ARIMA, LSTM","lastPublishedDoi":"10.21203/rs.3.rs-8776307/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-8776307/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eKuwait is known for its hot, arid desert climate, frequent dust storms, and intensive oil and gas-related activities, all of which directly affect air quality. The current study show a comprehensive assessment of air pollution in Kuwait by using 10 fixed Air quality monitoring stations operated by the Kuwait Environment Public Authority. Air pollutant levels were analysed form 2020, while long-term trends in particulate matter (PM\u003csub\u003e10\u003c/sub\u003e) were examined over the period Jan-2020 to Dec-2024. The results show persistently elevated particulate matter concentrations across Kuwait. Mean PM\u003csub\u003e10\u003c/sub\u003e levels ranged from approximately 60 to over 180 \u0026micro;g/m\u003csup\u003e3\u003c/sup\u003e, sometimes exceeding the World Health Organization (WHO) annual guideline of 15 \u0026micro;g/m\u003csup\u003e3\u003c/sup\u003e due to mainly weather conditions. Gaseous pollutants such as carbon monoxide (CO) and Sulfur dioxide (SO\u003csub\u003e2\u003c/sub\u003e) displayed moderate but spatially variable concentrations. CO annual means ranged from about 486 to 1,572 \u0026micro;g/m\u003csup\u003e3\u003c/sup\u003e, while SO\u003csub\u003e2\u003c/sub\u003e averaged between 12 and 33 \u0026micro;g/m\u003csup\u003e3\u003c/sup\u003e, with higher levels observed near industrial and refinery areas. Seasonal analysis revealed strong PM\u003csub\u003e10\u003c/sub\u003e peaks during spring and summer due to dust storms and shamal winds, whereas CO and SO\u003csub\u003e2\u003c/sub\u003e exhibited weaker seasonal patterns linked mainly to traffic and energy production. Overall, Kuwait\u0026rsquo;s air quality reflects the combined influence of natural dust loading and localized anthropogenic emissions, posing significant environmental and public health concerns and underscoring the need for targeted mitigation strategies in arid regions.\u003c/p\u003e","manuscriptTitle":"Assessment and Prediction of Air Pollution Trends in Kuwait Using Machine Learning: An Analysis of PM10 , CO, and SO2 and Their Environmental Health Implications","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2026-03-09 15:20:22","doi":"10.21203/rs.3.rs-8776307/v1","editorialEvents":[{"type":"communityComments","content":0}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"2105ea5b-4070-4eaf-b7af-6f22438e087a","owner":[],"postedDate":"March 9th, 2026","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[],"tags":[],"updatedAt":"2026-03-10T14:24:01+00:00","versionOfRecord":[],"versionCreatedAt":"2026-03-09 15:20:22","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-8776307","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-8776307","identity":"rs-8776307","version":["v1"]},"buildId":"XKTyCvWXoU3ODBz1xrDgd","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}