Casablanca city ( Figure 1 ) is located on Morocco’s Atlantic coast, bordered by the ocean to the west and Settat and Ben Slimane provinces to the north, east, and south. The city covers 1140.54 km², with 18.8% (227.82 km²) being urbanized. Urban areas doubled since the 1980s, growing from 100 km².

Casablanca, the largest city in Morocco, serves as the country’s economic and industrial hub. With a population of over 3.3 million in the metropolitan area, Casablanca has experienced rapid urbanization and motorization, leading to significant air quality challenges ( Sekkat et al., 2012 ).

The landscape features plains, plateaus, scattered hills, and a 98 km coastline, extending 22 km at Mansouria. The soil varies, with Tirs in rural areas and sandy soil along the coast. Rivers are minor, with Oued El Malleh, Oued N’fifikh, and Oued Hassar being the main ones. The city has 4000 ha of forests, mainly in Bouskoura.

The climate of Casablanca is a semi-arid one, with irregular rainfall, temperatures from 8˚C to 26˚C, and humidity levels always above 60%. Winds average 9 m/s from the northeast.

Geologically, Casablanca sits on sandy tuff formations with a shallow layer of soil. Hydrogeologically, groundwater is sparse and not highly vulnerable.

Hydrology: Urban development dried up Oued Bouskoura’s lower course, but flooding remains a risk during heavy rains.

This study utilized several satellite datasets accessed through Google Earth Engine (GEE): 1) Sentinel-5P NO₂ data, 2) Sentinel-2 Land Use Data, and 3) ERA5.

Other additional datasets were integrated into the analysis: 4) road network, and 5) traffic data.

The Sentinel-5 Precursor mission instrument acquires data pertinent to air quality assessment. The TROPOMI instrument, a multispectral sensor, records reflectance of wavelengths crucial for measuring atmospheric concentrations of ozone, methane, formaldehyde, aerosol, carbon monoxide, nitrogen oxide, and sulfur dioxide, as well as cloud characteristics at a spatial resolution of 0.01 arc degrees ( https://developers.google.com/earth-engine/datasets/catalog/sentinel-5p ).

The Sentinel-5P satellite provides high-resolution measurements of atmospheric nitrogen dioxide (NO₂) concentrations, which are used as a proxy for vehicle emissions and urban air pollution. The NO₂ data was retrieved and preprocessed using GEE.

a. https://developers.google.com/earth-engine/datasets/catalog/COPERNICUS_S5P_OFFL_L3_NO2#bands .

This dataset provides offline high-resolution imagery of NO₂ concentrations, available from 2018-06-28T10:24:07Z to 2024-09-28T20:22:15Z. and the period used in this study is from 2019 January to 2023 December.

Nitrogen oxides (NO₂ and NO) are significant trace gases in the Earth’s atmosphere, present in both the troposphere and the stratosphere ( Table 1 ). These compounds enter the atmosphere through anthropogenic activities (primarily fossil fuel combustion and biomass burning) and natural processes (wildfires, lightning, and microbiological processes in soils). In this context, NO₂ is utilized to represent concentrations of collective nitrogen oxides because during daytime, i.e., in the presence of sunlight, a photochemical cycle involving ozone (O₃) converts NO into NO₂ and vice versa on a timescale of minutes ( https://developers.google.com/earth-engine/datasets/catalog/COPERNICUS_S5P_OFFL_L3_NO2#description ).

Sentinel-2 satellite imagery was used to derive urban development indicators, such as the Normalized Difference Built-up Index (NDBI), to analyze the relationship between urban features and air quality.

Sentinel-2, as shown in Table 2 , is a wide-swath, high-resolution, multi-spectral imaging mission supporting Copernicus Land Monitoring studies, including the monitoring of vegetation, soil and water cover, as well as observation of inland waterways and coastal areas ( https://developers.google.com/earth-engine/datasets/catalog/COPERNICUS_S2_SR_HARMONIZED#description ).

b. https://developers.google.com/earth-engine/datasets/catalog/COPERNICUS_S2_SR_HARMONIZED#bands .

ERA5-Land, as shown in Table 3 , is a reanalysis dataset that provides a consistent view of the evolution of land variables over several decades at an enhanced resolution compared with ERA5.

Hourly wind speed and temperature data were obtained from ERA5-Land covering the study period ( https://developers.google.com/earth-engine/datasets/catalog/ECMWF_ERA5_LAND_MONTHLY_BY_HOUR ).

c. https://developers.google.com/earth-engine/datasets/catalog/ECMWF_ERA5_LAND_MONTHLY_AGGR#band .

Road network was collected from Organizing Authority for Urban Transport of Greatest Casablanca Region. A total of 587 vehicle traffic counts have been incorporated into the road network, covering the entire study area. Of these, 370 counts distinguish between different vehicle types.

This dataset corresponds to the permanent traffic counts managed by The Regional Directorate of Equipment, Transport and Logistics (RDETL) of Casablanca. Unlike the earlier peak-hour counts, these figures represent Annual Average Daily Traffic (AADT), aggregated across both traffic directions. Each traffic count is georeferenced to a specific road network segment.

The dataset includes the following key fields:

The following table ( Table 4 ) presents additional data related to the structure of travel patterns used for the implementation of the model and provides detailed information on the data sources.

The following ( Figure 2 ) are the key components of the methodology used in this study: 1) Temporal Analysis; 2) Spatial Analysis; 3) Land Use Regression.

Time series decomposition was performed on the NO₂ data to identify seasonal patterns and long-term trends in air pollution levels.

In this proposed method, we begun by converting the unit mole/m² to micromol/m² for more representativity of the Total vertical column of NO₂ (ratio of the slant column density of NO₂ and the total air mass factor), after that, we used ee.Reducer.mean() function of GEE to generate temporal aggregation monthly average NO₂ for study period from 2019 to 2023.

Spatial interpolation and hotspot detection methods were applied to the NO₂ data to map the distribution of air pollutants and identify pollution hotspots across Casablanca.

ee.Reducer.mean() function of GEE was also used to generate the image series of the average of netrogene_monthly for the study period.

A regression modeling approach was used to quantify the influence of urban development indicators, traffic patterns, and meteorological factors like the impact of speed wind and temperature on NO₂ concentrations ( Hoek et al., 2008 ; Jiang et al., 2020 ).

However, the methodology of our approach, which is completely built on the GEE cloud computing platform, began firstly by selecting a collection of images of the S2 L2 sentinel: 5 images per year; from 2019 to 2023. For each image, the Normalized Difference Built-up Index (NDBI) ( Zha, Gao, & Ni, 2003 ) were calculated.

$NDBI=\frac{B11-B8}{B11+B8}$

where B8 is NIR, and B11 is SWIR 1. The NIR and SWIR bands are used to emphasize manufactured built-up areas, also this ratio is based to mitigate the effects of terrain illumination differences as well as atmospheric effects.

And use ee.Reducer.pearsonsCorrelation() function of GEE for correlation between NO₂ concentration and NDBI.

Secondly, use ERA5-Land on GEE to generate yearly charts of speed wind average and temperature average for each year, from 2019 to 2023, used ee.Reducer.mean() function of GEE, and ee.Reducer.pearsonsCorrelation() to correlate NO₂ concentration to speed wind, and correlate NO₂ concentration to temperature.

The temporal analysis of Sentinel-5P NO₂ data revealed distinct seasonal variations in air pollution levels across Casablanca. A time series decomposition showed consistent peaks in NO₂ concentrations during the winter months, potentially due to increased vehicle emissions and unfavorable meteorological conditions for pollutant dispersion.

From the charts ( Figure 3 ), between 2019 and 2023, nitrogen dioxide (NO₂) concentrations exhibited peaks during the winter months. The highest concentrations were recorded in 2021, 2022, and 2023, showing a marked increase. However, a significant decrease in NO₂ concentrations was observed, from 93 µmol/m² in April 2019 to 78 µmol/m² in April 2020, and from 109 µmol/m² in October 2019 to 95 µmol/m² in October 2020. This significant reduction in NO₂ levels was observed, attributed to the COVID-19 lockdowns, which resulted in decreased vehicular traffic.

The spatial analysis of NO₂ levels using GEE-based techniques identified several pollution hotspots within Casablanca. These hotspots were often located in areas with high traffic density, major transportation corridors, and industrial zones, highlighting the significant contribution of vehicular emissions and urban activities to air quality ( Beirle et al., 2011 ).

The spatial analysis of NO₂ concentration levels ( Figure 4 ) from 2019 to 2023 reveals that the highest concentrations, hotspots, were consistently recorded in the city center, an area characterized by the high traffic density of both private vehicles and heavy goods vehicles. This is primarily driven by the significant freight transport activity concentrated in this zone. Traffic density notably decreased in 2020 as a result of the COVID-19 pandemic but witnessed a rapid surge in 2023. This rebound reflects an increase in economic activity and vehicular traffic, as evidenced by the expansion of the vehicle fleet.

The land use regression analysis demonstrated a strong correlation between NO₂ concentrations and urban development indicators, such as the Normalized Difference Built-up Index (NDBI) ( Figure 5 ). Areas with higher urban density and proximity to major roads exhibited elevated NO₂ levels, underscoring the impact of transportation and land use patterns on air pollution in Casablanca.

The chart of urban growth ( Figure 6 ) depicts a substantial expansion of urban sprawl from 2002 to 2023, which has directly influenced increased mobility and is reflected in elevated air pollution levels, specifically the concentration of NO₂. This period exhibited a progressive increase in NO₂ concentrations extending towards the urban periphery. However, between 2019 and 2023, stabilization and subsequent reduction in urban growth levels were observed.

Using Pearsons Correlation r from 2019 to 2023 ( Table 5 ) demonstrates a weak negative correlation between NO₂ and NDBI in 2019. The negative sign indicates that as NDBI increases, indicating more built-up areas, and NO₂ concentrations levels tend to decrease. The magnitude of the correlation coefficient (−0.202) suggests a weak relationship. The p-value of 0 indicates that the correlation is statistically significant, meaning it is unlikely to be due to chance. These variables might show that while built-up areas (NDBI) are increasing, other factors like vehicular traffic could be contributing to the decrease in NO₂ concentrations levels. And in 2020 there is a very weak negative correlation between NO₂ concentrations and NDBI. The relationship is even weaker than in 2019. The p-value of 0 again indicates that the correlation is statistically significant. That reveal the reduction in NO₂ concentrations levels is more influenced by changes in traffic patterns rather than the built-up areas themselves. This could be due to lockdowns or reduced economic activities during the COVID-19 pandemic, which affected NO₂ emissions more than the built-up areas. However, in 2022, there is a very weak negative correlation between NO₂ concentrations and NDBI, similar to 2021 but slightly stronger, and the p-value of 0 indicates statistical significance. The slight increase in the correlation might indicate the changes in land use patterns or the introduction of new pollution control measures. These factors could be contributing to a more nuanced relationship between NDBI and NO₂ concentrations levels.

In 2023 there is a weak positive correlation between NO₂ concentrations and NDBI. This is a change from the previous years where the correlation was negative. The positive sign indicates that as NDBI increases, NO₂ levels tend to increase as well. The magnitude of the correlation coefficient (0.105) suggests a weak relationship. The p-value of 0 indicates that the correlation is statistically significant. The increase in NO₂ levels is due to a combination of factors such as increased traffic, industrial activities, and possibly changes in land use that favor higher NO₂ emissions. This suggests that the relationship between NDBI and NO₂ is becoming more complex and influenced by a variety of environmental and socio-economic factors ( Terry et al., 2024 ).

Figure 7 and Figure 8 clearly illustrate the impact of traffic in principal roads and sum of origine destination displacement in air quality. The origine destination displacement is the effect of urban growth. Additionally, they illustrate the elevated levels of concentration that were found along major highways, including the A3 and A5 motorways. Lower concentrations were also shown in coastal areas, likely due to sea breezes and fewer emission sources.

For more accurate representation of traffic’s impact on air pollution, the chart in Figure 9 shows the hourly traffic fluctuations and the pic of traffic were impact on air pollution.

The incorporation of meteorological data ( Figure 10 ) from the ERA5 reanalysis product revealed the influence of weather conditions on air quality in Casablanca. Factors such as temperature and wind patterns were found to play a significant role in the dispersion and accumulation of NO₂, contributing to the observed spatial and temporal variations.

The analysis of NO₂ concentrations levels in Casablanca, from 2019 to 2023 reveals significant correlations with meteorological variables such as temperature and wind. Including using Pearsons Correlation r ( Table 6 ), the analysis demonstrated a consistent negative correlation is observed between NO₂ and temperature, indicating that higher temperatures are associated with lower NO₂ levels. This relationship has strengthened over the years, suggesting increased dispersion and chemical breakdown of NO₂ at higher temperatures ( Jacob & Winner, 2009 ). The correlation between NO₂ and U-Wind (east-west wind component) is weak and positive, indicating that stronger east-west winds are associated with slightly higher NO₂ concentrations. This may be due to the transport of pollutants from urban or industrial areas ( Seinfeld & Pandis, 2016 ). The correlation between NO₂ and V-Wind (north-south wind component) is strong and positive, highlighting the significant influence of north-south winds on NO₂ concentrations. This suggests that stronger north-south winds transport pollutants from high-emission areas, leading to increased concentrations downwind ( Stull, 2012 ).

Table 6. Correlation Pearson’s r and P-value between NO₂ concentration and meteorological variables: temperature, U-Wind, and V-Wind, from 2019 to 2023.

The impact of automobile traffic on NO₂ concentrations levels can be inferred from the observed correlations. The strong positive correlation between NO₂ and V-Wind suggests that north-south winds may be transporting pollutants from high-traffic areas, in Casablanca, such as major roads and highways. This is supported by the findings that traffic flow and emissions are closely linked, and mitigating traffic congestion has a positive impact on air quality ( Behera & Balasubramanian, 2016 ).

Moreover, the weak positive correlation between NO₂ and U-Wind indicates that east-west winds may also play a role in the transport of pollutants from traffic-heavy areas. This is consistent with the observation that vehicular emissions are a major source of urban air pollution ( Behera & Balasubramanian, 2016 ). The increasing trend in the correlation between NO₂ and temperature over the years may also be influenced by changes in traffic patterns and urban development, which can affect the dispersion and chemical reactions of NO₂ ( Eaton, 2022 ).