Comparative Evaluation of Measured and Predicted Air Pollutant Concentrations Using AERMOD Around the Fallujah Cement Plant

The construction sector in Iraq has experienced rapid growth in recent years, with numerous projects currently underway, including multi-story residential buildings and large-scale infrastructure developments [1], [2], [3], [4]. Facilitated by the availability of raw materials for cement production and government measures to encourage industrial investment, several new cement plants are being planned to meet the growing domestic demand [5], [6]. Cement, the primary binding component in concrete—the most widely used construction material worldwide after water—plays a critical role in modern infrastructure by binding sand and gravel into a hardened matrix [7].

As cement production expands, emissions of atmospheric pollutants inevitably increase. Extensive evidence in the scientific literature has demonstrated that ambient air pollution adversely affects human health to varying degrees [8], [9], [10]. Children are particularly vulnerable due to incomplete physiological defenses and their higher inhalation rates relative to body weight [11], [12]. Exposure to fine particulate matter (PM) has been associated with increased sensitivity to allergens, aggravated asthma symptoms, and reduced pulmonary function [13]. Air pollution has also been linked to congenital disabilities, low birth weight, preterm birth, and elevated rates of cardiovascular diseases, with long-term exposure significantly increasing the risk of morbidity and mortality from heart and vascular disorders [14]. Elevated concentrations of nitrogen dioxide (NO$_2$), carbon monoxide (CO), ozone (O$_3$), and PM have been correlated with specific respiratory complications [15], [16]. According to the World Health Organization (WHO), ambient air pollution is causally associated with several severe health problems, including acute lower respiratory infections, lung cancer, chronic obstructive pulmonary disease, stroke, and ischemic heart disease [17], [18].

Cement manufacturing is one of the oldest and most strategic industries in Iraq, representing a cornerstone of national economic development due to its direct connection with the construction sector and its essential role in supporting urban growth, housing development, and infrastructure projects [19]. The industry contributes to local job creation, reduces reliance on imports, and enhances self-sufficiency in construction materials [20]. In 2024, the total production capacity of cement plants in Iraq reached approximately 32 million tons per year, with forecasts suggesting a rise to more than 52 million tons in the coming years due to the introduction of new production lines and additional facilities.

Currently, Iraq hosts around 18 state-owned cement plants operated by the Iraqi General Cement Company, alongside 13 major private-sector plants serving the growing market. Al-Anbar Province has emerged as one of the key production centers due to the abundance of raw materials such as limestone and gypsum [21]. In recent years, the Iraqi government has focused on expanding overall production capacity through policies designed to attract further industrial investment [20].

Globally, the cement industry is classified as a high-emission sector due to its complex thermal and chemical processes involved in clinker production, particularly sintering and calcination. These processes release several major pollutants, including sulfur dioxide (SO$_2$), NO$_2$, CO, carbon dioxide (CO$_2$), total suspended particulates (TSP), particulate matter $\leq$10 $\mu$m (PM$_{10}$), and fine particulate matter $\leq$ 2.5 $\mu$m (PM$_{2.5}$) [22], [23]. SO$_2$ emissions arise primarily from the combustion of sulfur-containing fossil fuels and sulfur compounds in raw materials, while nitrogen oxides are formed at the extremely high combustion temperatures of rotary kilns [24]. CO$_2$ originates from two dominant sources: the thermal decomposition of calcium carbonate and the combustion of fossil fuels used to power the process [25], [26]. These risks highlight the urgency of assessing the environmental and health burdens posed by cement production, particularly in regions of rapid industrial growth such as Al-Anbar Province.

The present study aims to evaluate the impacts of cement production on ambient air quality at three receptor sites surrounding the Fallujah Cement Plant in Iraq. Specifically, the study seeks to:

1. Quantify the concentrations of seven major pollutants (SO$_2$, NO$_2$, CO, CO$_2$, PM$_{10}$, PM$_{2.5}$, and TSP).

2. Simulate the spatial dispersion of these pollutants using the AERMOD model.

3. Compare model predictions with field measurements and assess statistical performance indicators.

4. Benchmark the results against Iraqi national air quality standards and WHO guidelines.

Through this integrated approach, the study provides a comprehensive assessment of the environmental and public health risks associated with emissions from the Fallujah Cement Plant.

The Fallujah Cement Plant is located in the district of Fallujah, Al-Anbar Province, western Iraq, at coordinates (N: 33.36863°, E: 43.85057°). It is regarded as one of the country’s most prominent industrial landmarks, serving as a cornerstone of the province’s infrastructure by meeting local demand for cement essential to construction and reconstruction activities.

The foundation stone of the plant was laid in 1978, and production commenced in 1984. However, following the events of 2003, the plant suffered extensive damage due to regional security conditions, resulting in years of inactivity. As part of Iraq’s post-conflict industrial revitalization efforts, the plant was reopened in 2021. It currently comprises three rotary kilns, of which two are operational while the third remains idle. Each kiln has a design capacity of 700 tons per day, with actual production reaching approximately 500 tons per kiln, yielding a total output of about 1000 tons per day. Heavy fuel oil serves as the primary energy source, while raw materials are supplied from Hit District, specifically Abu Tayban village.

Beyond its industrial function, the facility significantly contributes to the local economy by directly employing nearly 400 workers, thereby reducing unemployment and supporting economic stability in the city (Moqdad Taleb Jassim–Chief Senior Engineer–Al-Gharbiya Channel).

Climatic conditions in Al-Anbar Province are semi-arid, characterized by scarce rainfall, pronounced diurnal temperature variation, and low humidity. Summer temperatures exceed 52 ℃, while winter temperatures may drop to 9 ℃. Prevailing winds are predominantly northwesterly, with occasional southwesterly flows, reaching speeds of up to 21 m/s. Average annual rainfall is about 115 mm, mostly concentrated in winter. These meteorological conditions play a decisive role in controlling the dispersion of air pollutants from industrial activities, where wind direction, stability, and temperature directly influence pollutant transport, accumulation, and deposition patterns on surrounding residential and agricultural zones.

Although the Fallujah Cement Plant employs rotary kiln technology similar in principle to that used in U.S. cement plants, several differences exist. Most notably, Iraqi plants typically rely on heavy fuel oil rather than natural gas or coal, and their pollution-control systems (e.g., electrostatic precipitators (ESPs) or bag filters) are often limited or partially functional due to maintenance and resource constraints. These factors contribute to higher uncontrolled emissions relative to U.S. facilities.

Multiple processes within cement production contribute to atmospheric emissions. These include kiln stacks, raw material mixing and homogenization, conveyor transport, clinker cooling, raw and finish grinding, filter cleaning, as well as material loading and unloading operations and leakage from conveyors and filters.

Air quality sampling was conducted at three receptor sites around the Fallujah Cement Plant at varying distances and directions (Table 1). Pollutant concentrations were measured as follows:

SO$_2$ and NO$_2$: Multi-gas detector (S316).

CO: CO gas analyzer (GM8805).

CO$_2$: Infrared gas detector (Wintact WT8807).

Particulates (TSP, PM$_{10}$, PM$_{2.5}$): Laser-based particle counter (Y09-PM).

Instruments were calibrated according to manufacturer specifications before deployment. Measurements were taken at a height of 1.6 m, corresponding to the average human breathing zone, while minimizing interference from deposited dust. At each site, three consecutive readings were collected at 15-minute intervals, and mean values were used in the analysis.

Because gas analyzers report concentrations in parts per million (ppm), conversions to micrograms per cubic meter ($\mu$g/m$^3$) were performed using the ideal gas equation:

$ \text {Concentration}\left(\mathrm{\mu g / m^3}\right)=\text {Concentration}(\mathrm{p p m}) \times \frac{M W \times P}{R \times T} \times 1000 $

where,

• MW = molecular weight of the gas (g/mol),

• P = atmospheric pressure,

• R = universal gas constant (0.08206 L·atm·mol$^{-1}$·K$^{-1}$), and

• T = absolute temperature (K).

Meteorological parameters including wind speed, direction, relative humidity, and temperature were also recorded. In the absence of continuous on-site stack monitoring data, emission rates were estimated using standard emission factors published by the U.S. Environmental Protection Agency (USEPA) in AP-42, Chapter 11.6: Portland Cement Manufacturing, which are globally recognized and derived from extensive empirical emission datasets [27].

The modeled 24-hour average concentrations obtained from AERMOD simulations are summarized as follows. SO$_2$ exhibits the highest 24-hour mean among the gaseous pollutants (488.1 $\mu$g/m$^3$), followed by CO (273 $\mu$g/m$^3$) and NO$_2$ (81.8 $\mu$g/m$^3$). Among the particulate fractions, PM$_{10}$ reached 115 $\mu$g/m$^3$, TSP 111 $\mu$g/m$^3$, and PM$_{2.5}$ 23.7 $\mu$g/m$^3$, while CO$_2$ averaged 429.40 $\mu$g/m$^3$. These values represent typical summer conditions during which measurements were conducted—characterized by high ambient temperatures and prevailing northwesterly winds in Al-Anbar Governorate. Such meteorological conditions enhance convective mixing and horizontal dispersion yet intensify combustion activity in the kilns. Consequently, the presented results reflect a realistic summer operational scenario. To capture potential seasonal variability, future work should incorporate additional monitoring during winter months when lower temperatures and greater atmospheric stability may increase pollutant accumulation.

Emission factors for major pollutants from cement production, based on USEPA AP-42, are listed in Table 2.

The total annual emissions of each pollutant were estimated using the following equation:

$ E=\left(\frac{C}{100}-1\right) \times E F \times A $

where,

• E (kg/year): total pollutant emissions,

• A: annual activity rate (tons of clinker produced),

• EF: emission factor (kg/ton), and

• C (%): pollution control efficiency.

The AERMOD atmospheric dispersion model was employed to simulate pollutant transport in the vicinity of the plant. AERMOD is a steady-state Gaussian plume model designed for near-field ($<$50 km) applications and is widely used to estimate ground-level pollutant concentrations from stationary industrial sources. Developed in 1995 and officially adopted by the USEPA, AERMOD incorporates complex terrain effects, meteorological variability, and plume behavior [28].

Input parameters (see Figure 1) included stack height, stack diameter, exit velocity and temperature of exhaust gases, emission rates (g/s), and local meteorological data (wind speed/direction, temperature, atmospheric pressure). The AERMAP terrain preprocessor was integrated with digital elevation model (DEM) data to generate a receptor grid accounting for topographic features such as slopes, valleys, and hills, which significantly influence pollutant dispersion.

The model outputs included concentration fields and contour maps for each pollutant across receptor networks, thereby enabling spatial analysis of pollutant plumes under site-specific conditions.

AERMOD was selected because it is the regulatory model recommended by the USEPA for industrial point-source simulations, particularly when site-specific meteorological data and terrain corrections are available. Although other models such as California PUFF Model (CALPUFF) and Community Multiscale Air Quality (CMAQ) could provide additional insights into long-range or chemical transformation processes, AERMOD was deemed most appropriate for the near-field ($\leq$5 km) analysis in this study.

A direct comparison between the highest pollutant concentrations predicted by the AERMOD model and those measured in the field reveals a marked discrepancy (Table 3 and Figure 2). The model performance can be classified into two distinct categories: relatively good performance for PM and weak performance for gaseous pollutants.

For gaseous pollutants, AERMOD systematically underestimated concentrations. As shown in Table 3 and illustrated in Figure 3, the scatter comparison demonstrates that nearly all gaseous pollutants fall well below the 1:1 reference line. The relative error for CO reached an extremely high value of +440.79%, while CO$_2$ showed a relative error of +221%. Similarly, SO$_2$ and NO$_2$ concentrations measured in the field exceeded model predictions by +104.91% and +78%, respectively.

In contrast, the model performed considerably better for particulates. As presented in Table 3 and highlighted in Figure 4, the relative errors were much lower: +30% for TSP, +20.73% for PM$_{2.5}$, and +13.53% for PM$_{10}$. This closer agreement indicates that AERMOD was more effective in simulating the dispersion and deposition of particulates compared with gaseous pollutants.

The pronounced discrepancies between AERMOD predictions and field measurements, particularly for CO and CO$_2$, likely arise from a combination of factors, including uncertain emission inventories, lack of direct stack monitoring, and the use of default emission factors derived from U.S. data that may not fully represent local combustion conditions. Additionally, heavy fuel oil combustion efficiency and fluctuating kiln operating loads can introduce temporal variability not captured by steady-state AERMOD simulations.

The statistical performance of AERMOD was evaluated using the Index of Agreement (IOA) and Mean Bias Error (MBE). Consistent with the variance analysis, the pollutants can be grouped into two categories: particulates, which showed strong model performance, and gaseous pollutants, which were represented less accurately.

For particulates, the model exhibited excellent agreement with field data. IOA values were very high—0.9613 for TSP, 0.9875 for PM$_{10}$, and 0.9783 for PM$_{2.5}$—indicating that AERMOD successfully reproduced the general dispersion trend, particularly the decline in concentrations with increasing distance from the source. The MBE values for particulates were small and negative (-212 $\mu$g/m$^3$ for TSP, -78.33 $\mu$g/m$^3$ for PM$_{10}$, and -30.33 $\mu$g/m$^3$ for PM$_{2.5}$), suggesting a consistent but relatively minor underestimation of observed concentrations.

In contrast, the performance of the model was weaker for gaseous pollutants. IOA values were considerably lower, with NO$_2$ achieving 0.7486 (good agreement), SO$_2$ at 0.661 (moderate agreement), CO at 0.4969 (acceptable agreement), and CO$_2$ at only 0.2428 (very poor agreement). This indicates that the model was less effective in capturing the temporal and spatial variability of gaseous emissions. The MBE values reinforced this observation, as they were large and negative across all gases: -1903 $\mu$g/m$^3$ for SO$_2$, -558 $\mu$g/m$^3$ for NO$_2$, -3948 $\mu$g/m$^3$ for CO, and -427000 $\mu$g/m$^3$ for CO$_2$. These results confirm that AERMOD systematically underestimated the concentrations of gaseous pollutants relative to field measurements

Analysis of AERMOD outputs revealed clear and consistent spatial patterns of pollutant dispersion. In all cases, emissions formed a plume oriented southeast of the plant, directly influenced by prevailing northwesterly winds. Detailed interpretations are as follows:

Both SO$_2$ and NO$_2$ exhibited elongated elliptical plumes extending southeastward (Figure 5 and Figure 6, respectively). Maximum SO$_2$ concentration reached 2873 $\mu$g/m$^3$ at the stack, decreasing to 2650 $\mu$g/m$^3$ at 300 m—far above the Iraqi national limit (169 $\mu$g/m$^3$) and WHO guideline (40 $\mu$g/m$^3$). Concentrations gradually decreased with distance, reaching 400 $\mu$g/m$^3$ at 3 km.

For NO$_2$, the maximum predicted value was 1182 $\mu$g/m$^3$ directly above the source, dropping to 1000 $\mu$g/m$^3$ at 300 m. These values significantly exceed Iraqi (100 $\mu$g/m$^3$) and WHO (25 $\mu$g/m$^3$) limits. At 3.5 km, concentrations decreased to ~80 $\mu$g/m$^3$.

Both CO and CO$_2$ plumes followed the southeast wind-driven direction (Figure 7 and Figure 8, respectively). The maximum predicted CO concentration was 1609 $\mu$g/m$^3$ at the source, decreasing to 1520 $\mu$g/m$^3$ at 300 m. While elevated, these values remain below Iraqi (11000 $\mu$g/m$^3$, 8 h) and WHO (4000 $\mu$g/m$^3$, 24 h) standards, suggesting lower acute risk compared with SO$_2$ and NO$_2$.

Predicted CO$_2$ concentrations reached 306804 $\mu$g/m$^3$ at the stack and 295000 $\mu$g/m$^3$ at 300 m. Although no WHO guideline exists for CO$_2$, these values are well below the Iraqi limit (720000 $\mu$g/m$^3$). Nonetheless, they reflect the plant’s significant combustion intensity.

Particulates exhibited southeastward plumes consistent with prevailing wind patterns but displayed distinct physical behaviors (Figure 9, Figure 10, Figure 11). TSP and PM$_{10}$ showed the highest concentrations in close proximity to the source due to their greater mass and rapid deposition rates. Maximum concentrations reached 1575 $\mu$g/m$^3$ for TSP and 937 $\mu$g/m$^3$ for PM$_{10}$ directly at the source, decreasing to 1350 $\mu$g/m$^3$ and 850 $\mu$g/m$^3$, respectively, at a distance of 300 m. Both pollutants exceeded national and international air quality standards. By contrast, PM$_{2.5}$ demonstrated a broader dispersion profile, with lower peak intensity near the source (275 $\mu$g/m$^3$ at 300 m) but greater atmospheric persistence. Its smaller aerodynamic diameter allows longer suspension times and farther transport, rendering it particularly hazardous for distant residential areas.

A comparison with environmental standards, as shown in Table 4, revealed significant exceedances for most pollutants. Under Iraqi national standards, the average SO$_2$ concentration was 4040 $\mu$g/m$^3$, surpassing the permissible limit of 169 $\mu$g/m$^3$ by approximately 23-fold. Similarly, NO$_2$ averaged 1390 $\mu$g/m$^3$, exceeding the national limit of 100 $\mu$g/m$^3$ by 13-fold. PM also showed pronounced exceedances, with PM$_{2.5}$ surpassing its standard by 12-fold, PM$_{10}$ by 9-fold, and TSP by 4-fold. In contrast, average concentrations of CO and CO$_2$ remained within Iraqi limits.

When compared against the stricter WHO guidelines, the magnitude of exceedance was even more severe. Average NO$_2$ levels were approximately 55 times higher than the WHO threshold, while SO$_2$ exceeded the guideline value by a factor of 100. Both PM$_{10}$ and PM$_{2.5}$ concentrations were nearly 20 times above the recommended limits, and CO exceeded the guideline by 21.7%.

By contrast, all pollutants were within the occupational exposure limits set by the Occupational Safety and Health Administration (OSHA). However, these standards are designed for healthy adult workers under controlled 8-hour exposures with personal protective equipment, and thus are not directly applicable to community-level exposure. Consequently, they provide limited relevance for assessing risks to the general population, including vulnerable groups such as children, the elderly, and individuals with pre-existing health conditions.

Health risk assessment was carried out for three distinct population groups. Plant workers located within 300 m of the facility were exposed to the highest pollutant concentrations. Although CO and CO$_2$ levels remained within OSHA occupational limits, chronic exposure to elevated SO$_2$, NO$_2$, and PM$_{2.5}$ levels poses substantial risks of respiratory and cardiovascular diseases.

Residents of housing complexes situated approximately 2 km south of the plant experienced severe exposure due to prevailing wind direction and their relative proximity to the source. At this location, SO$_2$ concentrations reached 1470 $\mu$g/m$^3$, exceeding the Iraqi national limit by 769%, while NO$_2$ concentrations averaged 332 $\mu$g/m$^3$, exceeding the limit by 232%. PM$_{2.5}$ also exceeded permissible values by 112%, with average concentrations of 53 $\mu$g/m$^3$. These findings underscore the inadequacy of current zoning practices, as industrial facilities of this type should be located at least 5 km from residential areas.

Finally, residents of Fallujah city, located 3 km west of the plant, experienced lower exposures due to their position upwind of prevailing emissions. Nonetheless, concentrations of several pollutants still exceeded WHO guidelines. SO$_2$ averaged 1185 $\mu$g/m$^3$, which is 601% above the Iraqi standard, while NO$_2$ averaged 275 $\mu$g/m$^3$, exceeding the limit by 175%. PM$_{2.5}$ reached 36 $\mu$g/m$^3$, 44% above the Iraqi limit. In contrast, CO, CO$_2$, and TSP levels remained within national and international standards. These results indicate that even at greater distances, emissions from the plant contribute significantly to degraded air quality and associated health risks

During site inspection, it was observed that the plant’s ESP system was operational but inefficient due to irregular maintenance and occasional bypass during peak production. This explains the elevated particulate emissions measured downwind.

Cement plants face the dual challenge of controlling both stack emissions and fugitive emissions from operational activities. Effective mitigation therefore requires a combination of administrative measures and advanced engineering controls.

From an administrative and operational perspective, measures such as paving internal roads, enclosing conveyors, covering crushers and storage piles, and applying water sprays or chemical stabilizers (e.g., latex) can substantially reduce dust emissions. Based on findings from previous studies in arid and semi-arid environments, the implementation of vegetative greenbelts around industrial facilities has been shown to reduce near-ground PM concentrations by approximately 10%–25%, depending on factors such as vegetation structure, width, and density [29], [30]. A greenbelt design incorporating a width of 50–100 meters and a planting density of 1000–1500 trees per hectare may offer meaningful reductions in airborne dust and particulates, particularly when using drought- and dust-tolerant species such as Eucalyptus camaldulensis and Tamarix aphylla, which are well-suited to the climatic and soil conditions of the Fallujah region [31], [32].

In terms of engineering interventions, several pollutant-specific strategies are available. For particulates, fabric filters are considered the best available technology (BAT), with removal efficiencies exceeding 99.9%, including for fine particulates such as PM$_{2.5}$. ESPs are also widely used in cement kilns and clinker coolers, where they operate by charging dust particles and collecting them on oppositely charged plates. To control nitrogen oxides (NO$_x$), techniques such as low-NO$_x$ burners, staged combustion, and selective non-catalytic reduction (SNCR) with injected urea or ammonia can achieve reductions of up to 80%. SO$_2$ emissions may be reduced through the use of low-sulfur fuels or substitution with natural gas, as well as by adding alkaline sorbents such as calcium oxide (CaO) to absorb SO$_2$ during calcination. CO emissions can be minimized by optimizing the air–fuel ratio to ensure complete combustion, with natural gas offering a cleaner alternative fuel. Finally, reductions in CO$_2$ emissions can be achieved by adopting alternative fuels such as biomass or waste, improving kiln efficiency, and substituting clinker with supplementary materials in blended cements.

While advanced mitigation measures such as low-NO$_x$ burners and natural gas substitution are technically effective, their implementation in Iraq faces logistical and economic barriers. Natural gas supply to Al-Anbar Province remains limited, and high-sulfur heavy fuel oil is the dominant available fuel. Therefore, short-term mitigation should prioritize filter maintenance and dust control, while long-term strategies focus on energy diversification and gradual infrastructure upgrades.

This study evaluated pollutant concentrations in the vicinity of the Fallujah Cement Plant in Iraq and analyzed the associated health risks, combining field measurements with AERMOD dispersion modeling. The results demonstrated that for certain pollutants—particularly TSP, PM$_{10}$, PM$_{2.5}$, and NO$_2$—AERMOD predictions were reasonably consistent with measured values. However, for other pollutants, especially CO$_2$, substantial discrepancies were observed between measured and modeled concentrations.

When compared with regulatory thresholds, it was found that all pollutants, with the exception of CO, CO$_2$, and TSP, exceeded both Iraqi national air quality standards and WHO guidelines within a 2000 m radius of the emission source. At 3000 m west of the plant, most pollutant concentrations had declined to levels approaching the permissible limits.

Overall, the findings suggest that AERMOD is an effective tool for simulating concentrations of certain pollutants, particularly particulates and NO$_2$, while its performance is less robust for gases such as CO$_2$ due to the nature of its multiple sources and emission dynamics. In general, the application of AERMOD provides a reliable framework for assessing emission-control scenarios and supporting the development of environmental policies aimed at reducing air pollution and mitigating its health impacts.

This study was limited by short-term sampling rather than continuous monitoring, reliance on emission factors instead of direct stack measurements, and data from only three receptor sites. In addition, AERMOD showed weaker performance for some gaseous pollutants, reflecting uncertainties in emission sources and model assumptions. Future studies should employ long-term monitoring and direct stack measurements to improve accuracy. Expanding receptor coverage and testing alternative dispersion models (e.g., CALPUFF, CMAQ) could enhance spatial and temporal resolution. Integrating health impact assessments with local medical data would also strengthen the link between exposure and health outcomes.