Introduction

Air quality is one environmental concern that has steadily risen to the fore as a result of thriving economic growth and swift industrial advancement. According to World Health Organization (WHO) (2022) report, one of the greatest hazards to global health is air pollution. In 2019, it was estimated that ambient (outdoor) air pollution contributed to 4.2 million premature deaths globally. Particulate matter (PM), sulfur dioxide (SO2), nitrogen dioxide (NO2), ozone (O3), carbon monoxide (CO), and carbon dioxide (CO2) are the most common air pollutants that people experience daily (Chen et al. 2007). The extent and intensity of air pollution in metropolitan areas are determined by various factors, including the characteristics of the land surface, weather patterns, and the types of air pollutants and pollution sources. Natural factors like air pressure, temperature, wind direction, and speed affect these characteristics; however, human causes like domestic coal combustion, vehicle exhaust emissions, and industrial waste gas emissions have a greater effect on urban air quality (Filonchyk et al. 2018; Özdemir et al. 2018; Park et al. 2018; Liu et al. 2019; Yavuz 2023).

In this respect, continuous and accurate monitoring of air quality is necessitated for accurate prediction and decision-making to mitigate the effects of air pollution. For extended periods, it is not feasible to conduct physical ground-based monitoring over large areas with either permanent stations or mobile sensors. Some of the drawbacks of these ground-based monitoring stations and sensors include increased maintenance costs and data point collections that are restricted to small areas or just a few meters surrounding the station, which leads to the incorporation of discontinuous data to provide interpolated results (Amanollahi et al. 2013). In contrast to conventional methods, remote sensing and geographic information system (GIS) techniques are able to gather data about the Earth, such as land and sea surface temperature, vegetation cover, and air quality. Also, it provides supplementary information for unmonitored areas and investigates historical land use/land cover (LULC) and air quality (Nacef et al. 2016; Voors et al. 2017; Chaikaew, 2019; Nagy et al. 2022).

Multiple studies utilized the moderate-resolution imaging spector-radiometer (MODIS) instrument to obtain and measure aerosol optical depth (AOD) concertation and derive ground-level PM2.5 concentrations from columnar AOD measurements for various places across the world. The majority of them were applied in the USA and China, and only a few were carried out in Europe, including France (Kacenelenbogen et al. 2006), the Netherlands (Schaap et al. 2009), Italy (Di Nicolantonio and Cacciari 2011) and the UK (Beloconi et al. 2016). Zhang et al. (2021) recently released a review article that included the most current methodological advancements for PM2.5 derivation from AOD.

In addition, numerous studies conducted since the launch of Sentinel-5p have demonstrated the effectiveness of the TROPOMI instrument. For example, Theys et al. (2017) investigated volcanic SO2, Borsdorff et al. (2018a, b) gave an overview of the preliminary CO measurement using TROPOMI data over China, Mamić (2021) measured CO levels by the device TROPOMI of the Sentinel-5P satellite, additionally, Grzybowski et al. (2023) conducted a study to verify the potential of the Sentinel-5P Tropospheric NO2 Column Number Density to support air pollution monitoring in Poland.

On the other hand, the change in LULC has an essential role in reducing biodiversity and ultimately affects the climate in an area at both local and regional levels. Large-scale changes in the LULC pattern are caused by rapidly expanding human activities, such as population expansion, industrialization, and urbanization. In order to achieve a sustainable environment, it is imperative to monitor and detect LULC changes because they are essential for understanding how human actions interact with the environment (El-Zeiny and Effat 2017; Yohannes et al. 2020).

For viewing and evaluating changes in land cover, medium spatial resolution satellite imagery, such as Landsat data, is most frequently utilized. These images were employed for monitoring LULC which is significantly impacted by human activity on a variety of levels. Such studies were conducted by Fan et al. (2007), Junfeng et al. (2011), Al-doski et al. (2013), Gad and El-Zeiny (2016), Tariq et al. (2021) and El-Zeiny et al. (2023).

One of the governorates in the Central and East Delta region with the greatest agricultural area reaching 880,670 acres is Sharkia. Due to urban growth, its loss of agricultural land is estimated to be more than 25% during the last two decades. Additionally, the governorate’s air quality has gotten worse due to rising urban populations and levels of industrialization (Belal and Moghanm 2011; Shalaby 2012; Khalil et al. 2014). Consequently, there is an urgent need for precision and up-to-date LULC and air quality data to develop corrective measures and resolve spatial problems.

Therefore, the current study combines remotely sensed data from various satellites (Landsat, Aqua, Terra, and Sentinel-5P) with GIS techniques to analyze land and air characteristics in Sharkia Governorate. It contributes to assessing LULC changes in the past 20 years. In addition, it aids in evaluating air pollutants (i.e., PM2.5, CO, NO2, SO2) to produce seasonal maps over 2022 and define the most deteriorated areas in the governorate. The study stands out for its innovation regarding the integration of images from several satellites; Aqua, Terra, and Sentinel-5P . With the aid of this integration, the method can be enhanced to analyze air pollutants. This method can serve as a foundation for further research in the environmental assessment of urban air quality on a worldwide level.

Materials and methods

Study area

Sharkia Governorate is one of the East Delta governorates and is situated in Egypt’s Eastern Nile Delta. It extends between latitudes 29° 54′ and 31° 12′ N and longitudes 31°20′ and 32° 15′ E (Fig. 1). Manzala Lake and Dakahlia Governorate are located to the north of Sharkia Governorate, whereas Kalubia is to the west of the governorate. Ismailia and Port Said, two Suez Canal Governorates, are situated to the east of the Sharkia Governorate. The governorate extends over a total area of 4911 km2, equivalent to 1.072.470 million acres. In terms of topography, the region is elevated at a height of around 13 m above mean sea level (A.M.S.L). The main water supply for Sharkia Governorate is the Dameitta branch of the Nile, which is located near the western border of the governorate. Several significant irrigation canals, including the Manayef Canal in the north, the Bahr Moyes and Bahr Faqous in the middle, and the Ismailia Canal in the south traverse the governorate. Additionally, the governorate is crossed by important agricultural drains including the Mahsama and Bahr El Baqar Drain (ESIAF 2010). According to Central Agency for Public Mobilization and Statistics (CAPMAS) (2021), following the governorates of Cairo and Giza, Sharkia is the third most populous at the republican level, with an estimated 8 million people, including expatriates, making up 23% of the population in urban areas and 77% in rural areas.

Fig. 1
figure 1

GIS-based map showing the geographical location of the study area, Sharkia, in Egypt (2022)

Full size image

Landsat data analyses for land variables retrieval

Landsat imagery acquisition and preparation

In this study, three Landsat 7 and 8 with Enhanced Thematic Mapper (ETM) and Operational Land Imager (OLI) images (path 176, row 039) with a spatial resolution of 30 \(\times\) 30 m were acquired in July 2002, 2012, and 2022 (Site: https://glovis.usgs.gov/). Spectral data may be affected by multiple factors like sensor-target-illumination geometry, atmospheric absorption and scattering, sensor calibration, and satellite data processing procedures. These processes also tend to change over time. Radiometric correction had to be carried out to monitor actual landscape changes as shown by fluctuations in surface reflectance from satellite imagery. For calibrating Landsat data, mostly radiometric and atmospheric corrections were applied using ENVI 5.3 software (Table 1) (El-Zeiny and El-Kafrawy 2017).

Table 1 Technical and spectral specifications of Landsat 7 and Landsat 8 satellites
Full size table

Image classification

In this study, per-pixel supervised classification was applied. It grouped spectral reflectance features from satellite image pixels with the same or similar properties into the same information categories. Multi-spectral images’ pixel components could be automatically grouped into labeled categories including water, vegetation, urban areas, and bare lands. As a result, a pixel can be identified by its spectral signature, which is based on the relative reflectance of the target area in each band. The maximum likelihood classifier (MLC) is one of the important and accurate classifiers used in supervised classification. It provided the maximum accuracy for the generated LULC maps in the current study. This classifier is frequently employed to monitor and evaluate LULC changes (Campbell 2002).

Aqua and Terra data analyses for PM 2.5 retrieval

Aqua and Terra data acquisition and processing

MODIS, an instrument fixed on Aqua and Terra satellites, provides every 1–2 days global coverage, with a viewing swath of 2330 km. For this study, Level 2 MODIS aerosol product data (MCD19A2-V6) over 2022 were obtained from the Level-1 and Atmosphere Archive and Distribution System (LAADS DAAC). The MCD19A2 Version 6 data product is a MODIS Terra and Aqua combined Multi-angle Implementation of Atmospheric Correction (MAIAC) AOD gridded Level 2 product produced daily at 1 km pixel resolution (Table 2). The AOD data at 550 nm from the MCD19A2 dataset are used in this study, which is only retrieved for daytime, cloud-free, and snow/ice-free conditions with a ⁓20% confidence level in the uncertainty (Site: https://www.earthdata.nasa.gov/). MODIS conversion toolkit on ENVI 5.3 software was utilized to convert the obtained satellite data from hierarchical data format (HDF) into Tiff format and also to perform data projecting and georeferencing. The layers were consequently loaded on ArcMap to mask with Sharkia boundary shapefile.

Table 2 Technical and spectral specifications of Aqua and Terra satellites
Full size table

Calculation of PM 2.5

AOD is a measurement of the total attenuation of solar radiation that the entire atmosphere’s aerosols produce. In contrast, the ground monitoring site measurements of PM2.5 mass concentrations can only be used to assess the near-surface air quality. Based on the assumption of a linear relationship between AOD (unitless) and PM2.5 (μg/m3), Zheng et al. (2017) introduced PM2.5—AOD relationship to monitor PM2.5 concentration below 500 m, which can be defined as:

$${\text{PM}}2.5 =271.34 \times \mathrm{ AOD}+22.24.$$
(1)

After processing, the average, minimum, and maximum values of PM2.5 were statistically calculated for each season in 2022 over Sharkia Governorate.

Sentinel-5P data analyses for gases retrieval

On October 13, 2017, Sentinel-5 Precursor (Sentinel-5P), a satellite mission, was launched with TROPOMI aboard. The Sentinel-5P mission is the first in a line of atmospheric monitoring systems within Copernicus, the European Union’s Earth observation program, which is primarily concerned with observing the atmosphere’s chemical composition. (Borsdorff et al. 2018a, b; Kleipool et al. 2018). The TROPOspheric Monitoring Instrument (TROPOMI) instrument covers wavelength bands between the ultraviolet and the shortwave infrared, with a spectral resolution ranging between 0.25–0.54 nm and a minimum spatial resolution of 7 by 3.5 km. It evaluates the levels of ozone (O3), nitrogen dioxide (NO2), sulfur dioxide (SO2), bromate (BrO3−), formaldehyde (HCHO), water vapor (H2O), carbon oxide (CO), and methane (CH4). The data can be used directly without further pre-processing (Veefkind et al. 2017) (Table 3).

Table 3 TROPOMI spectral bands and their key features
Full size table

Sentinel-5P data acquisition

Through a number of web services, the Sentinel-5P data is freely downloaded (Site: https://www.sentinel-hub.com/). It is classified by time into NRT (near real-time), OFFL (offline), and RPRO (reprocessed) data. For the purposes of this paper, Sentinel-5P/TROPOMI product type level 2 and offline (OFFL) satellite imageries from January 1, 2022, to December 31, 2022, were obtained from EO Browser (Table 3).

Processing and georeferencing data

ArcMap software was utilized for data processing. Images that did not show proper sequences for the analysis made up the missing sample dates for all of the periods examined in 2022. By loading the data on ArcMap 10.5 software, it was noticed that all data were probably georeferenced corresponding to World Geodetic System 1984 (WGS84) coordinate system. Using ArcMap toolbox, all layers for each gas were resampled reaching 1 km spatial resolution then masked using the governorate shape file. After processing, the average, minimum, and maximum values of the air pollutants (CO, NO2, and SO2) were statistically calculated for each season in 2022 over Sharkia Governorate. Similar methods were conducted by Mamić (2021) and Savenets et al. (2022). To compare gas values with the Egyptian threshold limits, the obtained satellite unit (mol/m2) should be converted into μg/m3. This was achieved using the equation supposed by Savenets (2021) which is:

$$C= \frac{{C}_{{\text{col}}}}{H} \times M \times A,$$
(2)

where \({C}_{col}\) is the pollutant column content (mol/m2), \(M\) is the molar mass (g/mol), \(A\) is a constant equaling 1000, for conversion from g/m3 to mg/m3, \(H\) is the expressed in m. The adopted methodology for this study is introduced in Fig. 2.

Fig. 2
figure 2

Flowchart showing methodology adopted in the study

Full size image

Results and discussions

Detection of LULC changes in Sharkia governorate

For this study, LULC for Sharkia Governorate was created to assess the anthropogenic activities and the use of land for different purposes, such as agriculture, urban areas, water bodies, and barelands for change detection analysis. The main purpose of LULC in the present study is to evaluate if the variable classes contribute to worsened air quality. LULC changes were monitored in the past 20 years (2002–2022) in two different periods: 2002–2012 and 2012–2022. A total of four classes were produced for each of the three satellite images (bare land, urban, vegetation, and water). It was noticed that the governorate was continuously subjected to spatiotemporal changes throughout the whole study period. It is mainly occupied by agricultural areas which represent the most dominant class in LULC maps with a total area of 3711.64 km2, 3851.05 km2, and 3794.59 km2 in 2002, 2012, and 2022, respectively (Fig. 3). However, urban areas revealed a continuous increase as follows 201.11 km2 (2002) < 279.02 km2 (2012) < 499.16 km2 (2022). On the contrary, bare lands and water bodies showed a different pattern of decrease as illustrated in Table 4. Radwan et al. (2019) reported similar pattern of LULC changes in a study conducted in Sharkia Governorate during the period from 1992 to 2015.

Fig. 3
figure 3

LULC maps of  Sharkia Governorate during 2002 a, 2012 b, and 2022 c

Full size image
Table 4 Monitoring changes in areas (km2) and % of LULC classes
Full size table

In order to analyze the changes that occurred in each period separately, the annual change in LULC was also determined (Table 5). During 2002–2012, a significant annual increase in vegetation (139.41 km2) and urban classes (77.91 km2) mostly occurred on account of the bare land (–183.29 km2) and water classes (–34.05 km2), as presented (Table 5 and Fig. 4). This can be attributed to the effects of the governorate’s development programs, particularly urban development and desert reclamation, which significantly altered the land cover during this time. On the other hand, an obvious annual increase in the urban class (220.14 km2) occurred at the expense of three other classes: bare land (− 146.68 km2), vegetation (− 56.46 km2), and water (− 16.99 km2) during 2012–2022. It was noticed the urban expansion, in Sharkia Governorate, occurred on behalf of the agricultural and bare lands in this period. This explains the deteriorating situation because numerous locations were in distress. In another study carried out by Elagouz et al. (2020), it was found that Sharkia Governorate undergoes a very severe LULC change due to urban sprawl as result of rapid population growth during the period of 1987–2015.

Table 5 Annual change (km2) in LULC classes during the periods of study
Full size table
Fig. 4
figure 4

Annual rate of decrease/increase (km2) in LULC classes during the periods of study

Full size image

Bare lands decreased in the governorate over the whole study period as shown in Table 5 and Fig. 5. The total bare land area was decreased by − 183.29 and − 146.68 km2 from 2002 to 2012 and from 2012 to 2022, respectively. This may be caused mostly by an expansion in agricultural land and land reclamation, as well as an increase in urban sprawl. Similarly, water bodies have been decreased from 234.92 to 200.87, to 183.88 km2 in 2002, 2012, and 2022, respectively. Water bodies were represented by the main irrigation and drainage canals, fishponds, and part of Manzala Lake on the northern border, of Sharkia Governorate. Human activities and development projects may result in reduced agricultural irrigation and drainage water in the study area and consequently change land use, as reported by Shalaby (2012).

Fig. 5
figure 5

Spatial distribution maps for seasonal mean of PM2.5 in Sharkia Governorate during summer a, autumn b, winter c and spring d (2022)

Full size image

A previous study on Sharkia was conducted by Fahmy et al. (2023) to investigate LULC from 2001 to 2022 using remote sensing data. The study reported that the urban regions rose considerably from 10.22% to 29.12%. In contrast, vegetation reduced substantially from 66.15% to 53.39% and bare land decreased by 5.63%. It has been noticed that the variability pattern is consistent with the investigated pattern of the current study.

PM 2.5 assessment

Based on the combined MODIS AOD dataset, daily PM2.5 datasets were produced. As previously described, PM2.5 was driven and has been averaged seasonally. Results showed a significant fluctuation ranging from 29.49 to 104.93 µg/m3 through 2022. Table 6 illustrates PM2.5 (µg/m3) statistics in the study area over 2022. PM2.5 mean of each season followed this order: spring (77.59 µg/m3) > summer (50.76 µg/m3) > winter (49.29 µg/m3) > autumn (42.71 µg/m3). The same finding was reported by Handschuh et al. (2022), which assumed that the seasonal variation could be an indication of changing aerosol type. Generally, the obtained PM2.5 concentrations were lower than the maximum allowable concentration (i.e., 100 µg/m3 daily) that was given by the Egyptian Environmental Low as reported by Egyptian Gazette (No. 199 of August 28, 2011).

Table 6 Statistics of PM2.5 (µg/m3) over Sharkia Governorate in 2022
Full size table

Spatial distribution maps of PM2.5 for each season over 2022 are shown in Fig. 5. It was found that major parts of Sharkia Governorate experienced high levels of PM2.5 along the spring season, whereas the low levels dispersed in a wide range over the governorate during the autumn. Moreover, the distant northern and southern parts of the governorate were highly affected by PM2.5 all over the year. These parts were synchronized with part of Manzala Lake in the northern area and by bare lands and industrial zones in the south. Similar to the findings of the present study, Lu et al. (2019) and Yu et al. (2022) indicated that anthropogenic activities in addition to coastal areas (haze atmosphere) might be strongly affected by high PM2.5 concentrations.

According to NASA’s Mediterranean Dust Storm report, in late April 2022 (spring season), a reported dust storm over the Middle East was triggered by thunderstorms. The dust lowered vision, hampered flight, yellowed the skies, and lowered the quality of the air (https://earthobservatory.nasa.gov). The maximum concentration of PM2.5, exceeding 70 μg/m3, was recorded in spring 2022. Furthermore, compared to the other seasons, the spatial distribution in spring was more widely dispersed.

Gases assessment

Seasonal assessment of CO gas

CO, NO2, and SO2 gases are important atmospheric trace gases for better knowledge of tropospheric chemistry. In this study, Sentinel-5P is utilized to estimate the total vertical column of pre-mentioned gases, which provides their concentrations. The spatial distribution of tropospheric CO over Sharkia Governorate during 2022 is shown in Fig. 6. The selected period is from summer, autumn, winter, and spring. Table 7 shows that the CO means were as the following: winter (946.58 µg/m3) > spring (926.59 µg/m3) > summer (876.93 µg/m3) > autumn (787.42 µg/m3). The maximum concentrations were found in winter 2022. As was shown in another country (Rinsland et al. 2000), CO rapidly declines and OH radicals start to remove CO more intensely, as the air temperature rises. Koike et al. (2006) noted that in addition to the seasonal change in CO loss by interaction with OH, the observed seasonal variation was driven by a combination of seasonal fluctuations of several source contributions.

Fig. 6
figure 6

Spatial distribution maps for seasonal mean of CO in Sharkia Governorate during summer a, autumn b, winter c, and spring d (2022)

Full size image
Table 7 Statistics of CO (µg/m3) over Sharkia Governorate in 2022
Full size table

In Fig. 6, it was clearly found that the southern and western parts of the governorate are more likely to have a higher level of CO and that the areas of urban and bare land areas. The northern region stands out as the lower polluted geographical area over the study period. These findings are in line with other studies that have investigated the relationship between land cover and air quality and have reported that air pollution increases with urban and bare land areas, while declines with water and agriculture. Consumption of fossil fuels, burning biomass, and atmospheric oxidation of methane and other hydrocarbons are the main sources of CO (Mamić, 2021; Maurya et al. 2022).

Seasonal assessment of NO 2 gas

By observing both statistical results of NO2 (Table 8) and visual (Fig. 7), it can be noticed that the mean tropospheric vertical column of NO2 has a clear seasonality with a maximum 9.7 µg/m3 in winter, and a minimum in other seasons lower than 2.5 µg/m3. Savenets (2021) reported that the content of NO2 tends to be higher in the winter because of the longer NO2 lifetime. The extended lifetime causes NO2 accumulations throughout the atmospheric vertical column.

Table 8 Statistics of NO2 (µg/m3) over Sharkia Governorate in 2022
Full size table
Fig. 7
figure 7

Spatial distribution maps for seasonal mean of NO2 in Sharkia Governorate during summer a, autumn b, winter c, and spring d  (2022)

Full size image

The spatial distribution of tropospheric NO2 showed that the governorate was widely covered with high levels of gas during the winter season. However, the distribution patterns of NO2 were similar in summer, autumn, and spring (Fig. 7a, b, and d) in which the highest values are concentrated in the southern part which is occupied by bare lands and industrial zones. Numerous emission sources, including fossil fuel burning, vehicle emissions, and natural sources, are related to the amount of NO2 in the atmosphere. This is strengthened by previous studies conducted by Kaplan and Yigit Avdan (2020) and Maurya et al. (2022) that reported similar findings. Furthermore, as was shown in other studies (Sharifi and Felegari 2022; Faisal and Jaelani 2023), since the coronavirus pandemic (COVID-19) in 2020 forced quarantine measures in place of normal conditions in 2019, the density of tropospheric NO2 columns has reduced. This was due to temporary restrictions on industrial activity, traffic, and population density.

Seasonal assessment of SO 2 gas

The spatial distribution of tropospheric SO2 over Sharkia Governorate during 2022 is presented in Table 9 and Fig. 8. It can be noticed that the winter season had the highest mean value (13.07 µg/m3), while the summer season had the lowest mean value (4.53 µg/m3). By observing the SO2 spatial distribution map, it was obvious that the governorate was dominated by low levels during the study period. Nevertheless, several regions experienced significantly higher values as shown in Fig. 8. Increased activity of heating facilities using solid fuels may be a reason for higher SO2 concentration values during the cold season (Savenets et al. 2022). Similar findings were reported by Maurya et al. (2022) who noticed that the maximum concentration of SO2 was seen during the winter.

Table 9 Statistics of SO2 (µg/m3) over Sharkia Governorate in 2022
Full size table
Fig. 8
figure 8

Spatial distribution maps for seasonal mean of SO2 over Sharkia Governorate during summer a, autumn b, winter c and spring d (2022)

Full size image

Overall, based on the threshold limits of air pollutants from WHO (2023), it can be clearly found that the investigated gases (i.e., CO, NO2, and SO2) in the current study do not exceed the allowable limits (www.eea.europa.eu).

Conclusion

Monitoring land and air changes is crucial for maintaining environmental sustainability. Remotely sensed data from several satellites (Landsat, Aqua, Terra, and Sentinel-5P) were integrated with GIS to evaluate land and air characteristics of Sharkia Governorate. It was found that Sharkia Governorate displayed significant variations in LULC over the two past decades. Some areas in the governorate are losing agricultural lands because of land degradation and urban sprawl. On the other hand, a good example of planned urban and agricultural developments can be seen in some desert regions, where their agricultural and urban lands are increasing at the expense of the bare lands. Results of PM2.5 for each season over 2022 revealed that the distant northern and southern parts of the governorate were highly affected by PM2.5 all over the year. Higher levels of PM2.5 were recorded in the spring season. However, the winter season has the highest values of CO, NO2 and SO2. Overall, it was obvious that deteriorating regions in Sharkia Governorate were related to the regions with anthropogenic contributions limited to the southern parts. Although the satellite-derived air pollutants monitoring method offers a higher spatial coverage than ground monitoring stations, the satellite’s temporal coverage is primarily limited by unfavorable observation conditions like clouds and fog. In the current study, there were 250 product-valid days for PM2.5 and gases during 2022. Therefore, it is necessary to develop alternative monitoring strategies with broader geographic coverage and less weather restrictions. In order to manage fragile ecosystems sustainably, this study recommends the creation of efficient land and air management information systems. Although the current study provided precise information about the governorate climatic and land characteristics status, additional work should be done to accurately elucidate the primary sources and factors affecting the air quality and clarify to what extent they affect public health. Therefore, in future studies, it was recommended to evaluate further pollutants, incorporate more ground measurements, and assess additional variables including building density, wind direction, and other variables that can affect air pollution.