Abstract

Understanding the generation, transportation, and accumulation of aerosol particulate matter (PM) pollutants in the atmosphere needs an investigation into its geographical and temporal fluctuations. Due to the absence of air quality monitoring stations and insufficient background data about PM issues in developing countries, remote sensing data was used because it has been widely utilized as an alternative to studying PM concentrations. Measurements from satellites and ground measurements introduce a complete overview of particle pollution behaviour. The main aim of this study was to evaluate PM matter concentration for 14 locations over Sulaimani city- Kurdistan region/Iraq from July to November 2021. EPAM 5000 DUST-HAZ meter was used to measure the ground particulate matter concentrations with the aerodynamic diameter for PM1, PM2.5, and PM10. Also, Satellite data at 1-km-resolution by Moderate Resolution Imaging Spectroradiometer-Aerosol Optical Depth (MODIS-AOD) have been used at 0.550 µm wavelengths have been used. The linear regression model was applied to find the slope between satellite MODIS-AOD values and ground measurement. The Inverse Distance Weight (IDW) interpolation methods were used to map the particulate sizes of PM1, PM2.5, and PM10; their efficiencies were 0.92, 0.95, and 0.91, respectively. The results of ground measurements showed high variability in PM concentration values and ranged between; 7-253, 8-273 and 11-169 µg m^-3 for PM1, PM2.5 and PM10, respectively, among the studied locations during the study periods. MODIS predicted value for PM1, PM2.5 and PM10 ranged between 7.62-59.35, 6.64-43.65, and 50.54-62.17 µg m^-3, respectively. Correlation coefficients between observed and predicted values for September were 0.80, 0.73, and 0.81 for PM1, PM2.5, and PM10, respectively. The main conclusion that can be identified from this study is; that remote sensing data can be used to study PM matter concentration; a dense ground monitoring network is required to improve monitoring and environmental services management in Sulaimani City.

1. Introduction

High-speed industrialization has significantly impacted atmospheric chemical composition and air quality over the past four decades; For example, urban air quality has become a popular issue in the press and among the general public due to its largely adverse effects on human health and the environment ¹. Studies on severe air pollution generated from the main sources of haze formation, particulate matter, vehicular exhaust, ozone formation, and greenhouse gases have received a lot of interest owing to their contribution to causing global warming, health risk, and economic loss ^{2, 3}. Globally, atmospheric visibility is ordinarily used at different ground meteorological stations as an air quality indicator ⁴. Tuygun et al. ⁵ defined aerosol particle matter as the primary component of air quality monitoring networks. At the same time, an aerosol is defined by Mishra et al. ⁶ as a complex combination of solid and liquid particles floating in the air. Aerosol properties such as; size, chemical ingredients, and other physical and biological features change with time and place.

Moreover, aerosol alters the Earth-atmosphere system's radiative balance and can indirectly influence the environment through aerosol-cloud interactions. As yet, aerosol particle matter affects aerosol-cloud interactions and continues as one of the largest uncertainties in climate change studies ⁷. Ambient particulate pollutants can either originate from natural sources, such as forest fires, volcanic eruptions and resuspended dust, or from anthropogenic activities, such as driving automobiles and operating industrial or power plants ^{8, 9, 10}. Accordingly, differences in pollution sources cause variations in pollutant types.

Ground-based monitoring systems for aerosol measurements can offer spatially and temporally dispersed concentrations in metropolitan areas ¹¹. The collected data about high-frequency ground-based aerosol optical properties can enhance global understanding of aerosols' sources, transport, and diurnal variation. It also provides extensive data to prove aerosol impact on the ecological system and human health ¹. World ground stations such as; AERONET (Aerosol Robotic Network), EARLINET (European aerosol Lidar network), GAW-PFR Network (Global Atmosphere Watch Programmer-Precision Filter Rad, and both CAR-NET (China Aerosol Remote Sensing Network) and CSH-NET (Chinese Sun Haze meter Network) were established to obtain information on aerosol optical characteristics ^{12, 13, 14, 15, 16}. Due to the significant fluctuation of aerosol particulate matter concentration in space and time, ground measurements do not provide adequate information on the spatial distribution of atmospheric particulates ¹⁷. Imani ¹⁸ has demonstrated that the on-site measurement may face many difficulties and requirements, especially expenses and technical personnel as well as maintenance and operation, so they cannot be typically selected unless the requirement is available. However, satellite-based aerosol optical depth (AOD) has played a vital role in reducing the constraints of local measurements in areas where monitoring is few or nonexistent ^{19, 20}. The AOD measurement, based on measures of light absorption by atmospheric particles, was created using satellite remote sensing. As AODs have a strong relationship with ground-level PM2.5/PM10 concentrations, they've been utilized to track global and regional pollution problems. Assessment of satellite AOD against equivalent ground AOD data has demonstrated the worldwide trustworthiness of satellite AOD data ^{21, 22}. Also, many researchers have used satellite data to assess PM in the Southeast Asia region for evaluating AOD [23-28] ²³. Unfortunately, there is no significant study to discuss and assess PM matter, particularly in Iraqi-Kurdistan.

Therefore, this study aimed assessment of particulate pollutants via using MODIS satellite and ground-based measurement for the city of Sulaimani-Kurdistan Region of Iraq from July to November 2021 because MODIS instruments offer near-daily measures of global coverage. The study also aimed to create a general framework for future satellite-based ambient particulate pollutants concentration mapping techniques and to improve monitoring and environmental services management in the Kurdistan region. However, our study's main hindrances and complications were the lack of a monitoring station and the shortage of background data concerning aerosol particulate matter (PM).

2. Material and Methods

This research study was conducted at Sulaimani City, which is located in the east of Iraq's Kurdistan Region, not far from the Iran–Iraq border, and has the coordinates of Latitude and Longitude 35°33′26′′N 45°26′08′′E respectively, Figure 1. Sulaimani is characterized by its fertile plains of Sharazur and Bitwen, which give way to hills and the Zagros mountain range in the northeast. The elevation of the Sulaimani centre is about 830 meters above sea level ²⁹. Nonetheless, the fourteen research locations' elevations range between 633 to 1706 meters above sea level.

Sulaimani city has a climate typical to the region, with hot, dry summers and cooler winters. Sulaimani is wetter in winter and cooler in summer compared to neighbouring areas. Rainfall is restricted to the winter months ³⁰. As a whole, the climate of Sulaimani governorate is semi-arid continental, with scorching and dry summers from June to September and typical temperatures of 39 to 43°C. The winter months are cold and damp, especially in the high mountains; average winter high temperatures range from 7 to 13°C, with average low temperatures ranging from 2 to 7°C. The average temperature in the spring season is 13 to 32°C, while the average temperature in the fall season is 24 to 29°C ³¹. The mean annual precipitation is about 600 millimetres. The annual prevailing wind direction is SW, with an annual mean wind speed of 2.1 m s^{-1 32}. The governorate area is about 17,023 km², and the population at the current time is more than 1,783,270 ³⁰.

Spatial and temporal data on aerosol particle pollutant concentrations of different aerodynamic diameters were measured using environmental particulate air monitors (model EPAM-5000, PM size range 1–100µm) intended to measure ambient air particulate pollutants in real-time.

From July to November 2021, the concentration of PM matters was measured in fourteen locations for one hour each time per week and at each site. The sites are categorized into three groups depending on the location features of rural, common urban, and urban hot spot areas within Sulaimani city. Also, metrological parameters such as temperature, wind speed, relative humidity and pressure were measured by Kestrel -4000 Pocket Weather Meter for every single location and at each PM measurement. The average values of temperature, wind speed, relative humidity and pressure are presented in Table 1.

Moderate Resolution Imaging Spectroradiometer (MODIS) is a 36-channel Spectroradiometer with wavelengths spanning from 0.41 microns to 15 microns; MODIS is part of NASA's Terra (initially known as EOS AM-1) then Aqua (initially known as EOS PM-1) satellite missions. Since February 2000 and June 2002, both Terra Modis and Aqua Modis have been monitoring or viewing the entire Earth's surface every 1 to 2 days and acquiring data in 36 spectral bands or groups of wavelengths ^{33, 34}. In this study, the MODIS -AOD used readings at green band AOD 0.55 nm over the ground and at 1 km resolution. AOD is a daily dataset accessible via NASA's Earth Data Portal (https://earthdata.nasa.gov). That includes information acquired when a satellite passes over the imaged area; hence it contains varying amounts of data each day (depending on the number of satellite transits in a day). Then, from July to November 2021, satellite images were downloaded daily. Data from AOD was processed, and HDF files have converted to GeoTIFF format to acquire an AOD value for a specific latitude and longitude. These files were used to compute AOD data at our point's coordinates. MODIS files provide date and time information that can be used to construct an AOD and surface PM dataset. All AOD images were removed from dark backgrounds and corrected by uploaded file via MODIS satellite and calculated by raster calculator in ArcGIS10.8 to find the monthly average concentration in µg m^-3 for atmospheric particulate sizes of PM1.0, PM2.5, and PM10.

This study applied the inverse distance weighted (IDW) interpolation methods for interpolation data. IDW is a deterministic spatial interpolation approach to points near together and is more similar than those farther apart. As a result, a location's concentration may be anticipated by adding the weighted concentrations of other known places. The weights on nearby areas are the inverse of the squared distances ³⁵. The interpolation process validates six-point data randomly to ensure interpolation efficiency.

Different approaches like empirical-statistic methods (including statistical regression and machine learning), chemical transport models and other approaches (including semi-empirical model and vertical correction method) could be used for estimating ground-level PM concentrations by using satellite-derived aerosol products ³⁶. Thus creating a simple linear regression model is an approach and initial step for investigating the relationship between PM and aerosol optical depth (AOD). For each site, several data sets should be spatially and temporally integrated to obtain an AOD value for a specific location ^{37, 38, 39, 40, 41}. Additionally, a simple linear model (Eq1), as shown below, was used to extract three atmospheric PM aerodynamic diameter sizes (PM1.0, PM2.5, and PM10µm) according to the correlation between satellite (MODIS-AOD pixel ) and ground PM concentration data.

(1)

Where y= PM; a = intercept; b = slope and x = AOD value.

Validation was conducted using 85 MOD and 43 MYD PM2.5 images from July to November 2021. The PM predictions were derived from the created linear regression models over Sulaimani city from July to November 2021 and tested against the measured mentioned aerodynamic diameter of PM concentrations out of 6 points randomly at the end of the process to estimate the method’s accuracy.

3. Results and Discussions

National Ambient Air Quality Standards (NAAQS) stated by USEPA and CPCB have designated six criteria for air pollutants, and the criteria include particulate matter (PM2.5, PM10), Carbon monoxide (CO), ozone (O₃), nitrogen dioxide (NO₂), sulphur dioxide (SO₂) and lead (Pb). Thus, PM is among the six criteria of air quality standards, and it is the most important in terms of adverse effects on the environment, climate change and human health ⁴². For the study outcome, PM1 results in Table 2 showed high variability in concentration values (7 to 253 µg m^-3) among the studied locations and during the five months of the study. The minimum value was found with location 7 in July, while the maximum value was revealed with location October. Location 2 is considered one of the urban hot spots within Sulaimani city.

However, average PM1 levels ranged from 13.9 to 98.2 μg m^-3. In comparison to these results of the PM1 class, Majid ⁴³ found the average concentration between 13.1 to 170.1 µg m^-3 when the author measured the concentration of PM in ambient air seven (7) times during the measuring period from 31.9.2009 to 13.7.2010 and at 17 different outdoor sites in Sulaimani city. The variation trend of PM concentration in ambient air is due to many anthropogenic causes and sources such as; building and road construction, movement of trucks on unpaved roads, industrial, agricultural and commercial processes, then motor vehicle exhaust (vehicle's episodes and smoke haze owing to combustion of gasoline, oil, diesel fuel). Additionally, the world's natural causes and sources, such as trees and vegetation, pollen, sea salt, dust (airborne soil), and material from volcanic eruptions, are significantly associated with emitting aerosol particulate matter. PM may be either primary particles directly emitted from sources or formed as secondary particles in the atmosphere through chemical reactions of gases such as sulfur dioxide (SO₂), nitrogen oxides (NO_X), and certain organic compounds.

Although no specific standard limit value was proposed for PM1 level by international organizations and countries, increasing toxicity with decreasing aerodynamic diameter to submicron concentration was reported ^{44, 45}. For all these reasons, the fine particle could induce stronger adverse effects and was thus measured as more hazardous to human health than larger particles composed of the same material ⁴⁶. Furthermore, toxicological studies have suggested that the ultrafine fraction such as PM1 and PM2.5 contributed very little to the total mass concentration of particles but were the main component in a number of concentrations of particulate pollution ⁴⁷. Accordingly, research studies on PM2.5 air pollutants are critical for improving air quality and acquiring knowledge about the concentration and locations of air pollutant emissions like PM 2.5.

Results of PM2.5 presented in Table 2 also exhibited a high variation in concentration among the studied locations and for the same measurement period of PM1. The concentration values ranged between 8.0-273.0 µg m^-3. The lowest value was also found at location 7 in July, but the maximum value was found at location 11 in November. Also, location 11 is regarded as one of the urban hot spots areas within Sulaimani because it is a block of a roadway with heavy traffic. Nonetheless, average values of PM2.5 ranged between 15.6-112.5 µg m^-3. Thus, the emission of the traffic vehicles caused considerable PM pollution throughout the street. On the other hand, Majid ⁴³ obtained a broader and higher average range for PM2.5 concentration(18.7-181.7 µg m^-3), as he investigated PM2.5 in the outdoor air seven (7) times from 31.9.2009 to 13.7.2010 and at 17 different outdoor sites in Sulaimani city.

Moreover, Health Effects Institute ⁴⁸ considered fine particulate matter (PM2.5) as one of the air pollutants with the most extensive health impact from the perspective of financial considerations because more than half of the world population is exposed to annual-average ambient concentrations of PM2.5. What's more, it exceeds the guideline for standard or target of the World Health Organization (WHO), as well as the National Ambient Air Quality Standards (NAAQS) by EPA, which are 15.0 and 35.0 μg m^-3 for the annual (arithmetic average) and averaging time of 24-hour respectively ⁴³. Diesel exhaust emission in Sulaimani city has significantly contributed to a large source of particulate matter (PM) produced by diesel power stations and diesel engine vehicles since many diesel-electric generators are in work.

Diesel particles seriously threaten human health and adversely impact our environment ³¹. Likewise, a large variation in the temporal and spatial concentration of PM10 particles in the air was observed among the studied locations, and the concentrations varied between11to 169 μg m^-3 (Table) 2. A high variance of PM concentration indicates that the location feature is characterized by rural, common urban, and urban hot spots within Sulaimani. The lower limit of concentration was found in location 2, which had the highest PM1concentration. Still, the highest level of PM10 concentration was also revealed in location 11, which is characterized by the highest level of PM2.5. High PM levels can be attributed to the fact that there are different sources of PM emissions and human activities.

Overall, the average concentration values for PM10 ranged between 29.4 to 113.5 μg m^-3, and they were discovered in locations 10 and 11, respectively. According to the National Ambient Air Quality Standards (NAAQS) guideline by EPA ⁴⁹, the current value for PM10 is a 24-hour level of 150 µg m^-3. Nevertheless, according to the European Union Directive on Particulate Matter (1999/30/EC), PM10 readings cannot be more than 35 times per year over the daily average of 50 µg m^-3 or the annual average of 40 µg m^{-3 50}. As the result of this study compared with the standard limit of the European Commission (EC), it approves that the PM10 level in the hotspot locations has exceeded the permissible limit of a daily based average of 50 µg m^-3 of European Commission. Currently, European Union (EU) Clean Air Directive is one of the firmest and strictest acts of legislation worldwide regarding PM10 emission and air pollution. The most observable result of the new PM10 regulation has been the rapid adoption of “Low Emission Zones” (LEZs), which define areas that vehicles may come in only if they are categorized as low PM10 emitting vehicles. Moreover, Under the European Green Deal's Zero Pollution Action Plan, the European Commission (EC) established the 2030 goal of reducing the number of premature deaths due to fine particulate matter (PM2.5) by at least 55% compared with 2005 levels ⁵¹.

It is worth noting that the researcher Majid ⁴³ obtained an average concentration range of 40.4 to 374 µg m^-3 for PM10 levels in ambient air during the study period from 31.9.2009 to 13.7. 2010 in 17 different outdoor sites in Sulaimani city. The highest value might be due to a dust storm event.

In conclusion, Table 3 shows the descending order of the locations according to PM concentration (µg m^-3). Generally, no specific trend was observed among the different aerodynamic diameters of PM1, PM2.5 and PM10 as well as the locations, and this could be attributed to the fact that the atmosphere is a dynamic system, and Its behaviour, structure and composition could interact and impact many natural and anthropogenic factors.

The inverse distance weighting (IDW) interpolation approach, also referred to as inverse distance-based weighted interpolation, is deterministic or mathematical, supposing nearer values are more connected than further values with its function, or it is the estimation of the value (z) at location x and is regarded as a weighted mean of nearby observations ⁵². The proposed method was applied for ground aerosol particulate matter concentration calculation in this section, and each month separately with a scatter plot between ground observations and satellite-predicted measured value to explore their relation and accuracy. As shown in Table (4), the weight values of IDWr² and IDW-slope ranged between 0.91 to 0.95 and 0.776 to 1.061, respectively, and the upper values of IDW r² appeared by size. In contrast, PM2.5 size particles revealed the maximum values of IDW-slope. Normally, IDW gives larger weights to points closest to the prediction location, and the weights reduce as a distance function; hence the name inverse distance is weighted ⁵³. Thus, the weights are proportional to the proximity of the sampled points to the un-sampled sites and can be specified by the IDW power coefficient.

Even though there are several air pollutants in remarkable concentration in the atmosphere of the study area, the region is also exposed to the recurrence of dust and sand storms from global, regional and local sources. The authors Boloorani, et al. ⁵⁴ confirmed that dust storms are one of the main scientific struggles for environmental pollution and protection in the regions of West Asia because the phenomenon has significant detrimental impacts on the countries of the area, including Iraq, Saudi Arabia Iran, UAE, Syria, Kuwait, Qatar, and other countries of the region. The frequency of recent dust storm occurrences made it necessary to find appropriate solutions to struggle against disturbing events. Therefore, investigating the spatial and temporal variations of the ground measurement of atmospheric PM to the predicted data extracted from the pixel value is the most helpful way to assess atmospheric PM pollutants.

Table 5 reveals the ranges of concentrations for PM1, PM2.5, and PM10 recorded from ground measurements, IDW interpolation of in-situ measurements, and satellite-derived levels of PM employing MODIS-AOD images during the five months study period. The wider ranges of ground measurements for PM1 and PM10 occurred in October, while the broader ranges of PM2.5 was observed in November. These observations show the clear month cycle experienced in the study area, with notably high PM ranges during the more prolonged period of dryness and low during the shorter period of dryness, and this finding is in agreement with Atuhaire, et al. ⁵⁵ result of PM2.5 in Kampala District, Uganda.

As mentioned previously, ground-based measurements were combined with satellite remote sensing to estimate ground-level PM. In general, narrow-range trends were revealed by satellite-derived levels as compared with ground-based measurements for all PM sizes and for all months of the study. Moreover, a systematic trend in interval lowering differences was observed among the PM sizes and appeared as follows PM1 > PM2.5> PM10, and the intervals ranged between 32.81-46.79; 23.47-35.81 and 7.15-11.04, respectively.

In general, no apparent compatibility was observed between the ground and satellite measurements regarding the appearance sites for the range's limits of the different PM sizes during the different months of the study, and this may be attributed to the fact that Satellite images detect PM pollutants in the whole atmosphere. The realized PM level in the satellite image might be kilometres above the ground surface. Therefore, it is essential to compare the ground measurements to the satellite measurements in order to find out that the observed PM by the satellite images is on the ground surface. Furthermore, World Bank ⁵⁶ has demonstrated that satellite-extracted measurements cannot substitute appropriately well-operated and maintained ground-level observing nets for measuring the concentrations of particularly fine PM, such as PM2.5, that human beings are usually exposed to it daily. Accordingly, low and middle-income Countries (LMICs) must fortify support for establishing ground-level monitoring station networks to measure air pollutants, particularly PM2.5 and less, because of the cause of mortality in LMICs, Sub-Saharan Africa and other regions ⁵⁷. Conversely, Sorek-Hamer, et al. ⁵⁸ demonstrated that satellite data have a spatial benefit more than point sources; so far, it is essential to understand the restrictions and pitfalls of this type of data when investigating AOD-PM immediate associations. The key limits of satellite data are; existing under cloud-free conditions, incorporating the entire atmospheric column, instantaneous snapshot, i.e. satellite overpass time, and the relative humidity condition of the column.

The location (site) features might be specified according to rural, common urban, and urban hot spots based on the frequent appearances of the minimum and maximum levels. Generally, the frequency of minimum levels was mostly by L7, L8 and L14, with 4, 4, and 3 recurrences, respectively. While the frequency of maximum levels was, to some extent, by L14, and it was five recurrences (Table 5). However, the minimum and ultimate trends were not so frequent and systematic in accordance with the features of rural, common urban, and urban hot spots, and this can be attributed to the fact that the atmosphere has a dynamic nature; thus, the temporal, as well as the spatial concentration, are continually changing due to the natural, anthropogenic and climate factors.

In the case of the satellite using images for determining the spatial and temporal variations of PM1, PM2.5, and PM10, Figures marked 2, 3, 4, 5, and 6 represents three fragments of information about estimating ground and satellite data in July, August, September, October and November through A, B, and C which denotes a ground interpolation map employing inverse distance weighting (IDW). The interpolation maps for PM1, PM2.5 and PM10 values are symbolized by A, B, and C, respectively. In this study, the maps' small letters (a, b, c) stand for satellite aerosol particulate pollutants conducted through MODIS-AOD by applying the regression method and denoting PM1, PM2.5 and PM10, respectively.

Regarding the range of both satellite and ground measurements, the data for all PM sizes are also shown in Table 5 and maps a, b, and c for PM1, PM2.5 and PM10, respectively.

With regard to the range of both satellite and ground measurements, the data for all PM sizes are also shown in Table 5 and maps a, b, and c for PM1, PM2.5 and PM10, respectively. A scatter plot is drawn by the linear regression method to find the nearest similarity between ground-measured data measured by EPAM-5000 for each month of the study in 2021. AOD values are found in satellite data for each (PM1, PM2.5, PM10) size; besides, for each particulate size, a linear equation and R-squared value has been shown over the months of study in Table 6.

Table 6 shows the correlation coefficient (R-squared, R²) values between ground observation and satellite-predicted values according to the months of study. Clearly, the satellite-derived PM showed a good model performance with a coefficient of correlation (R²) values and ranged between 0.47-0.80, 0.47-0.80 and 0.71-0.90 for PM1, PM2.5 and PM10, respectively, hence the highest range was recorded by the largest PM10 size. Hoff and Christopher ⁵⁹ reviewed over thirty research papers to explore the correlations between ground PM and satellite AOD measurements by using single-variant of least-squares linear regression or a simple linear regression. Their finding showed a substantial variation with correlation coefficients (R²) and found values between 0 and 0.85 depending on the location, or region, of comparison and season. When we compare the strength of the correlation coefficient (R²) for linear regression between ground and satellite-observed AOD (from MODIS) measurements of PM2.5 in our study with the researchers Strawa, et al. ⁶⁰, for a study in six sites in the area of California San Joaquin Valley (SJV), that has a profile of poor air quality and high levels of PM pollutants, their results indicated a weak correlation coefficient of about 0.17 in the region.

In contrast, a stronger correlation coefficient in a range of 0.47-0.80 was revealed in our study. This might be a result of the very limited sites (only six sites) for their investigation by the researchers. Besides, Gupta, et al. ⁶¹ found a range of 0.11 to 0.85 to examine the linear correlation coefficient between MODIS AOD and ground measurements of PM2.5 in 25 urban areas. This result agrees with what we have found regarding the correlation coefficient of PM2.5.

Yet it is noteworthy that the researcher Yang, et al. ⁶² found in their study that relative humidity and boundary layer height (BLH) significantly affect the relationship between AOD and PM2.5 concentration at ground level. They tried to assess the improvement provided by modified regression models in order to correct MODIS AOD and then recover the correlation between MODIS AOD and PM2.5 because BLH can efficiently improve the AOD-PM2.5 correlation by converting AOD into near-surface aerosol extinction coefficient at the height of the aerosol scale.

Researchers rely on satellite data in many parts of the world, particularly in areas where few surface-based PM observations have been collected. Also, unfortunately, there were no sufficient experimental data on surface-based measurement in this research study area. The absence of this kind of in-site measurements and ground monitoring station causes a lake of information about surface background information on one side and, on the other side lake of satellite data, using cause difficulty for our analysis and way of aerosol particulate prediction.

In this study, the pixel value of atmospheric particulate was also measured by a raster calculator and MODIS-AOD layer in ArcMap by various analytical methods, including the linear regression method, based on previous studies [63-72].

In our locations for the research study, the elevation differences between the measurement sites were 427m above sea level. This issue should be considered with ground data because it plays a role in the metrological parameter's functions and then in the degree of atmospheric PM pollution, as applied by the researcher Yadav, et al. ⁷³ in Pune, city-India, from 2011–to 2012.

There was a considerable difference between the ground and satellite measurements, particularly for coarse particles in October, due to the notable changes in meteorological parameters between the seasons. This could be attributed to the fact that meteorological factors play a significant role in PM diffusion, aggregation, and spread. Consequently, it impacts PM concentrations when the local emission remains stable ⁷⁴. Moreover, the researcher also found through their correlation analysis that PM2.5 concentration has a positive relationship with pressure (r = 0.507), while they found a negative association with each of RH, rainfall, temperature, and wind speed with (r = - 0.237; r =-0.524; r =-0.512, and r = -0.284 respectively). Therefore, and based on this scientific fact, in studying atmospheric aerosol particles, the effect of metrological factors of RH, wind, rain, and temperature should be taken into account. Besides, the author Tai ⁷⁵ indicated that the daily variation in meteorological parameters could change up to 50% of the atmospheric PM2.5 concentrations.

Furthermore, a study by Lou, et al. ⁷⁶ explained that PM10 levels and RH exhibited an inverted V-shaped curve. On the other hand, the planetary boundary layer (PBL) high is another crucial factor in clarifying the differentiation between ground and satellite data. With high, the percentage of relative humidity decreases and dilution with separation of particulate increases. The lack of PBL height data could be another key factor contributing to temporal under- or overestimation in this study ⁷⁷. Although PBL is an important place for human actions and the major existence of air pollutants, it is about 1-2 km from the ground surface. It determines the atmospheric capacity for the diffusion, mixing, and transport of pollutants released into the PBL ⁷⁸; PBL remains uncertain, partially because heights have different definitions and are attained by numerous instruments for measuring PBL ⁷⁹.

On the other hand, numerous epidemiological research studies showed that a specific relationship exists between atmospheric particular matter pollution and cardiovascular and respiratory diseases as well as the mortality rate ^{80, 81}.

Yet it is noteworthy that the estimated ground measurement for most periods of the study months needs more ground measurements at all locations for future validation. The time trend of PM concentration measurements from satellite MODIS-AOD covers the whole study area and may more appropriately represent regional fine particulate concentrations; future research will investigate these correlations.

Similarly, fine particulate matter, like coarse particles, is sensitive to metrological data, especially at the height boundary layer, RH, and wind. The relationship between PM 2.5 levels and AOD is generally lower as far as the humidity is high. Although a high PBL allows particles to suspend in a greater vertical region, PM2.5 concentration measurement devices are often situated on the ground surface, allowing only surface PM2.5 to be recorded. Wind speed lowers the PM2.5 and PM1 to move long distances and remain in the atmosphere more than a coarse particle. The scatter plot shows the difference between observed and predicted particles in Figures( 2,3,4,5, and 6). Our study was driven by changes in metrological data and differences in topography, which led to a wide range percentage 46 to 77% in R² values. Accordingly, such a study needs a long period and a wide variety of research studies to indicate the real reasons for the high variation in PM concentrations.

PM1 investigation has caused us the same problems we faced with the other PM sizes. Moreover, there are limited published research studies worldwide on PM1, and that might be because it has a shorter lifetime in the atmosphere because it changes in many processes or routes to larger aerosol particle sizes. Nevertheless, PM1 is broadly thought to offer better information on the anthropogenic portion of PM pollution than PM2.5. However, data on PM1 are still inadequate in Europe, as well as insufficient information about its chemical composition and emission source; thus, this gap is more evident in the pollution hotspots and still continuing in Europe, such as the Po Valley in Northern Italy ⁸².

The maximum ground value of PM1 was 253 µg m^-3 and occurred in location 2 in October; this value was abnormal because it was too high as compared to other values of PM1 in the other months of the study, while the minimum level of PM1 appeared in location 7 (Table 1). As referred by the authors' Tang, et al. ⁸³ and Titos, et al. ⁸⁴, the primary sources of PM1 pollutants are mobile (vehicles and motorcycles). PM1, as a submicron particulate matter, is a substantial component of PM-related research studies owing to its typically anthropogenic origin and its prevalent mass fraction in atmospheric PM and due to insufficient research studies about its chemical composition ⁸⁵. Furthermore, the relationship between exposure to submicron particles like PM1 and health hazards, their impact on other environmental constituents, and even climate change have been widely discussed in the literature and established ^{86, 87}.

The minimum ground measurement average of PM1 (21 µg m^-3) during the five months of the study appeared with location 14, a botanical garden of thousands of square meters and far from residential areas. This indicates that green spaces significantly impact reducing air pollutants, particularly PM, and the air components and particles in the atmosphere of the local area may be limited to the components emitted from the plants in the garden. And this can be supported by researcher Srbinovska, et al. ⁸⁸, who studied the effect of green areas as well as small green walls on decreasing PM concentration in an open area. Their findings approved that green spaces involving green walls significantly reduce PM concentrations in green areas and their vicinity areas.

In general, there are ambiguity and uncertainty about the AOD-PM correlation for finding the surface concentration of PM. However, some researchers consider that AOD-PM has a strong relationship and believe that measures and processes have to consider to improve the correlation between AOD-PM and obtain accurate data for surface PM concentration ^{33, 89, 90, 91}. However, Schaap, et al. ⁹² found weak relation regarding AOD-PM concentration. Also, Voss and Evan ⁹³ found inadequate admiration for fine particulate matter. Contrastingly, many researchers believe that without considering some factors, we cannot predict actual surface concentration ^{62, 94, 95, 96}.

Finally, it is worth noting that the ground-observed AOD and the MODIS-AOD were valuable and suitable for extracting the surface PM2.5 concentration; however, aerosol particulate measurements in the region were used to validate linear regression models, particularly in the high aerosol-generating zone. After analyzing data, interpolating ground measurement, and evaluating predicted data by AOD, more stations to cover all study areas are vital for validating satellite data because substantial uncertainties are associated with both AOD and PM data in these conditions.

4. Conclusions

Particulate matter cycles are a significant process of earth system governance and need a routine as well as satellite remote sensing technology for monitoring. Therefore, investigation and verifying the presence of PM aerosols in different ways may facilitate the matter of knowing the actual concentration of PM, in addition to determining the level of air pollution and types of pollutants, especially PM of different sizes, then their harmful impact on the environment and health consequences.

MODIS Satellite Images and ground measurements are performed for the first time to monitor the spatial and temporal variation of different PM sizes in Sulaimani City-Iraq from July to November 2021.

The anticipated PM findings were comparable with ground measurements, indicating that satellite measurements can be modelled accurately to represent ground observations. It was also observed that an increase in spatial resolution might potentially increase the estimation accuracy of satellite-derived PM concentrations.

The appropriate and approved measurement for increasing the possibility of spatial and temporal accuracy requires the presence of the proper number of ground stations in line with the reality of the area concerned to measure air pollutants, especially atmospheric particles and climatic elements, so that we can have the proper correlation between ground measurements and satellite images.

When initiating future studies on the accurate and reliable measurement of air pollutants, especially aerosol PM, through the ground station and satellite images connection, climatic factors such as temperature, relative humidity, precipitation, wind and pressure, in addition to planetary boundary layer (PBL) high must be taken into consideration because they are dynamic factors that impact measurements of the atmospheric pollutants.

Additionally, it is essential to consider the biases that come with satellite passes occurring at specific times of day and being limited to “snapshots” of only a few times throughout the day. More technological and scientific advancements in this field hold a lot of potential for bettering our understanding of local air quality across the area.

Funding

This study has received no external funding.

Acknowledgements

The authors are grateful to the University of Sulaimani, Northern Iraq, for help and technical support; “The Terra/MODIS Aerosol Cloud Water Vapor Ozone Daily L3 Global 1 Deg. CMG dataset[s] was/[were] acquired from the Level-1 and Atmosphere Archive & Distribution System (LAADS) Distributed Active Archive Center (DAAC), located in the Goddard Space Flight Center in Greenbelt, Maryland (https://ladsweb.nascom.nasa.gov/).”

Conflicts of Interest

The authors declare no conflict of interest.

References