From Aerosol Optical Depth to Risk Assessment: A Novel Framework for Environmental Impact Statistics of Air Quality Using AERONET

Abstract

The implementation of European Union policies contributed to substantial air pollution reductions in recent years, but atmospheric aerosols remain a key pollutant class with environmental and public health risks. This study develops a novel method for assessing environmental impact and the risk associated with urban atmospheric aerosols. The integrated approach for air quality evaluation and prediction of the effects and risk of certain pollutants is based on Aerosol Optical Depth (AOD) analysis, considering the Aerosol Robotic Network (AERONET) database. To validate the method, it was applied using monitored air quality data for two cities in Romania, with 13 years (from 2011 to 2023) in one case and 12 years (from 2012 to 2023) in the other. The results demonstrated that an AOD risk index can be developed and utilized for air quality evaluation and prediction, enabling estimation of impacts and risks. In this case, aerosols measured by AERONET (Aerosol Robotic Network) over Cluj-Napoca (2011–2023) were dominated (46%) by a mixture of elemental (EC) and organic carbon (OC), while measurements over Iasi (2012–2023) showed 55% of the EC/OC mixture. The impacts and risks, as calculated by the AOD index for EC, show few significant ones, with an AOD range of 0.88 to 1.05 for Iasi and 0.73 to 0.88 for Cluj-Napoca.

1. Introduction

Air quality plays a vital role in human and ecosystem health, with poor air quality causing numerous health problems [1,2,3,4]. According to the 2024 State of Global Air report, exposure to air pollution contributed to 8.1 million deaths in 2021 [5]. This underscores the necessity for clear strategies and policies to address air pollution. Globally, countries have adopted legislative measures to reduce pollution and establish limits for key pollutants [6,7,8]. While the European Green Deal primarily targets greenhouse gas emissions, many of its measures—such as promoting clean transport and reducing fossil fuel combustion—bring co-benefits in lower air pollutant emissions, including fine particulate matter (PM_2.5) [9]. In line with EU requirements, Romania has adopted national strategies such as the Integrated National Energy and Climate Plan (INECP) and the National Air Pollution Control Programme (NAPCP), which include objectives for reducing key air pollutants and aligning with EU Directive (EU) 2016/2284 on national emissions ceilings [10].

A global analysis of 13,160 urban areas with populations >50,000, found that approximately 65% experienced an increase in PM_2.5 concentrations between 2000 and 2019 [11]. This pollutant originates from both anthropogenic and natural sources, including transport, industry, biomass burning, and dust, and it contributes to environmental degradation and public health issues. Elemental carbon (EC) and organic carbon (OC) are key particulate matter components with distinct sources and behaviors. Biomass burning is a major contributor to both, generally producing a higher proportion of OC, resulting in a low EC/OC ratio. Desert dust (DD), primarily composed of minerals, can interact with anthropogenic emissions, incorporating EC and OC into complex aerosol mixtures. In such cases, the EC/OC ratio serves as an indicator of source contribution and mixing processes, with implications for both air quality assessment and source apportionment [12,13,14]. These aerosol types (EC, DD, and mixtures) are linked to a wide range of health effects due to varying chemical composition, particle sizes, and toxicological properties, as summarized in

Table 1. Health effects of aerosol exposure.

Aerosol Type	Main Health Effects
EC	Pulmonary toxicity, respiratory and cardiovascular diseases, chronic obstructive pulmonary disease, infections, cancer, and premature death.
Mixtures (Fine and Coarse Particles)	Influenza, tuberculosis, skin reactions, inflammation, oxidative stress, and cardiorespiratory risks.
DD	Asthma, coughing, wheezing, bronchitis, pneumonia, allergic rhinitis, high blood pressure, and heart issues.

. Aerosols rich in EC—often originating from combustion processes—are associated with elevated health risks due to their containing polycyclic aromatic hydrocarbons, black carbon, and other toxic compounds. These substances can penetrate deep into the lungs, triggering systemic inflammatory responses, and in the case of ultrafine particles, may translocate into the bloodstream. Although DD can affect health, especially in vulnerable people (children, the elderly, the chronically ill), it is generally considered less hazardous than aerosols from biomass burning, which are strongly linked to chronic disease and mortality [15,16,17,18,19,20,21,22,23,24,25,26,27]. Sources of EC in Romania include both local (vehicular traffic) and long-range transport (Eastern European High and regional influences, as well as agricultural activities) [28,29,30,31].

Given these risks, accurately assessing the presence and distribution of aerosols is essential. A critical parameter for evaluating atmospheric aerosol loadings is Aerosol Optical Depth (AOD), which measures the attenuation of solar radiation due to scattering and absorption by particles suspended in the atmosphere. Although AOD does not directly represent surface-level PM_2.5 concentrations, it serves as a proxy when integrated with other datasets. The relationship between AOD and PM_2.5 is complex and context-dependent, affected by local meteorology, aerosol type, and geographical factors. Several studies demonstrate statistically significant correlations between AOD and surface-level PM_2.5 concentrations in urban environments with predominant anthropogenic emissions. These findings support the use of AOD data as an indirect proxy for pollution exposure, despite its limitation as a column-integrated measure. However, AOD interpretation can be affected by the vertical distribution of aerosols, humidity-related scattering effects, and seasonal variations in instrument calibration, which may limit its accuracy as a proxy for surface-level PM_2.5 [32,33]. Nonetheless, several studies demonstrate the utility of AOD for tracking pollution patterns and identifying aerosol sources, particularly when supported by the Ångström exponent (AE), which provides information on aerosol size distribution, precipitable water (PW), which accounts for atmospheric water vapor effects, and atmospheric models like HYSPLIT (HYbrid Single-Particle Lagrangian Integrated Trajectory), which enable the analysis of long-range transport and source trajectories [34,35,36,37,38,39]. These tools enhance aerosol source apportionment or PM_2.5 estimations but are not required in methodologies focused solely on evaluating aerosol-related impacts using AOD data.

Despite widespread use of AOD in atmospheric research, few studies integrate AOD data into standardized environmental risk assessment frameworks. Here, we introduce a new AOD-based impact prediction method that enables the classification and evaluation of aerosol-related environmental risks in urban areas. The new approach was applied based on monitoring air quality data for two cities in Romania (Iasi and Cluj-Napoca) for 13 and 12 years, respectively (for the period 2011–2023 and 2012–2023). City selection was based on a previous study by the authors, which examined the evolution of PM_2.5, arsenic, and carbon monoxide pollution over an 11-year period (2011–2021) in four Romanian cities—Brasov, Cluj-Napoca, Iasi, and Timisoara [40]. These cities were chosen for their similar size, development level, and urban characteristics. The study found that Iasi and Cluj-Napoca had the highest pollution levels among the four, justifying their inclusion in the present research. Additionally, both cities host AERONET stations and have been the focus of previous fieldwork by the authors, making them suitable for the AOD-based integrated assessment approach developed here. By integrating AOD into environmental impact and risk assessment, this research provides a novel toolset for monitoring pollution dynamics, identifying high-risk periods, and supporting policy making. The framework enhances traditional monitoring by enabling broader spatial analysis and improved forecasting capabilities. Furthermore, it emphasizes the importance of aerosol classification and source apportionment in understanding local air quality trends, ultimately contributing to more informed and effective mitigation strategies. Among the aerosol types identified through AOD-based classification—including OC dominated, EC/OC mixtures, Dust/EC mixtures, coated large particles, and other mixed profiles—this study focuses on EC and DD. These were selected because they are associated with well-characterized source profiles, clearly defined environmental and health effects, and are supported by a more extensive body of scientific literature compared to other, more complex or mixed aerosol types. The remaining categories are included for descriptive purposes only, to illustrate the broader aerosol composition and distribution in the study area. However, given the inherent complexity and relevance of mixed aerosol categories, we also briefly mention some of their potential effects in general terms, without attempting a detailed assessment that would exceed the scope of this study. In this research, “environmental impact” refers to inferred relative aerosol burden and potential risks derived from AOD measurements, representing aerosol loading in the atmospheric column rather than direct surface-level pollutant concentrations. While AOD alone is an indirect proxy, its integration with epidemiological and toxicological evidence supports its use in assessing urban aerosol impacts and risks. Moreover, AOD reflects the spatial variability of aerosol loading, providing an advantage over point-based measurements at ground level.

2. Materials and Methods

2.1. Site Description

Following European regulations, Romania monitors air quality through 148 stations: 30 traffic-type (assessing the impact of traffic), 58 industrial-type (monitoring industrial activity), 37 urban and 13 suburban background (assessing the influence of settlement and human activity), 7 regional reference stations, and 3 EMEP (European Monitoring and Evaluation Programme) stations (for monitoring cross-border pollution). To enhance aerosol property analysis, Romania also hosts 8 AERONET stations in Timisoara, Cluj-Napoca, Iasi, and Bucharest, which monitor local and long-range transported aerosol events. This study considers Iasi and Cluj-Napoca for integrated air quality assessment. Iasi (northeast Romania) faces major air pollution issues, with PM_2.5 levels 2.3 times higher than the legal annual limit of 25 µg/m³, as established by the European Union Directive 2008/50/EC on ambient air quality. In Cluj-Napoca (northwest Romania), PM_2.5 is also the main pollutant. In both cities, the primary sources of PM_2.5 pollution include industrial activity, traffic, and residential heating. Iasi, with about 271,692 residents and 60,000 students, is a major cultural and academic center. It has six air quality monitoring stations and one AERONET station (Iasi_LOASL), positioned near a main traffic artery and in proximity to residential neighborhoods that rely on biomass burning for heating, particularly in cold seasons (Figure 1). Cluj-Napoca, known as “the heart of Transylvania”, has about 286,598 inhabitants, six air quality monitoring stations, and one AERONET station, sited on the university campus in a mixed residential and traffic-influenced area (Cluj_UBB) [40,41,42,43,44]. The urban character of both sites influences local aerosol composition. The urban configurations are typical for medium-sized Eastern European cities and support the assumption that combustion-related sources, especially traffic and domestic heating, are major contributors to aerosol composition. Figure 1 provides an overview of the geographic context of the study, highlighting the location of Romania in Europe and the positions of the two selected urban areas, Iasi and Cluj-Napoca, along with the representative AERONET stations. Though point measurements cannot fully capture spatial heterogeneity within urban areas, these stations represent well-established AERONET sites for assessing urban aerosol optical properties and provide relevant integrated columnar measurements for this study’s impact assessment goals.

2.2. Experimental Data

In this study, aerosol optical properties were derived from Cimel CE318 sunphotometers, operated within the AERONET network—a global aerosol monitoring system operated by NASA and partners (agencies, universities, institutions, researchers, etc.). These instruments record solar radiation at wavelengths between 340 and 1020 nm, enabling retrieval of AOD and related parameters. Annual calibration, following AERONET protocols and using the Langley method at high-altitude sites, ensures data accuracy and consistency. Unlike models limited to ground-level or full-column data via remote sensing, this study focuses on optical aerosol parameters that indicate aerosol type, such as fine mode particles (typically associated with EC/OC mixtures or EC dominated aerosols) signaling biomass burning or soot; coated large particles that absorb solar radiation (black carbon, polluted dust); and mixtures like OC/dust, dust/EC, or dust dominated aerosols [28,45,46,47,48].

Adapting the Integrated Impact and Risk Assessment method to aerosol-type characterization via photometric data enables impact analysis on atmospheric components. AERONET data are categorized into three levels: unscreened (Level 1.0), cloud-screened (Level 1.5), and quality-assured cloud-screened (Level 2.0). AOD serves diverse applications in climate and weather modeling, visibility, aviation safety, and air quality due to its indication of total aerosol load [49,50,51,52]. Higher AOD values indicate a greater presence of aerosols, which negatively affects visibility and air quality [53,54]. AOD helps fill gaps where ground monitoring stations are lacking and correlates with PM_2.5, a major health-impacting pollutant. Studies show that PM_2.5 estimates improve when AOD is integrated into machine learning models [55,56,57].

For this study, AERONET photometric data from Iasi (LOASL: 47.19° N, 27.55° E) and Cluj-Napoca (UBB: 46.77° N, 23.55° E) stations were used. Both stations have improved understanding of local air quality by contributing aerosol optical property data and validating satellite-based observations [58,59,60].

Data quality assurance follows AERONET standard procedures comprising cloud screening algorithms, retrieval quality flags, and calibration updates. Inversion products from Level 1.5 data (version 3) were used, selected using Level 2.0 criteria: solar zenith angle >50° and retrieval error <5% [61]. AOD at 440 nm was used to cover both polluted and unpolluted events. Level 1.5 was preferred due to its near-real-time availability—within one hour of measurement [62,63,64,65,66]—important for timely risk and impact assessment [67,68,69,70]. Note that Level 2.0 data require annual instrument calibration and verification. As a result of this multi-step filtering process, a total of 1209 AOD measurements were selected for Iasi and 1380 for Cluj-Napoca. These values represent individual retrievals, not daily averages, and were acquired under cloud-free, high-quality conditions. Although the overall number may seem low for a 12–13 year period, it reflects the strict data quality assurance applied according to AERONET procedures to ensure reliable inputs for impact classification [71,72].

Following established clustering methods, the absorption Ångström exponent (AAE) and scattering Ångström exponent (SAE) were calculated at 440 nm and 675 nm, respectively, to classify aerosol types [46,58,66,73]. Both AERONET stations regularly acquire multiple measurements per day under clear sky conditions. Temporal coverage across seasons ensures representative sampling of aerosol variability in the target urban areas.

2.3. Integrated Assessment Methodology

Integrated Impact and Risk Assessment is an approach for quantifying environmental impact. The first step consists of establishing the environmental components, which are given a degree of importance (from 0 to 1), depending on the specifics of the assessed activity and on expert-based weighting criteria defined within the standardized assessment methodology, which considers the nature of the activity, environmental context, and pollutant characteristics. Based on matrix calculations, normalized scores and importance units (IUs) are obtained, which are subsequently used in calculations. The next step is represented by environmental quality quantification (Q), as shown in Equation (1). Finally, considering the units of importance (IU) and environmental quality, environmental impact quantification (EI) and risk quantification (ER) are completed using Equations (2) and (3) [67,68,69,70]. In this case, the method was adjusted, and the importance of the environmental component under assessment (AOD) was considered at its maximum value (1). Thus, the IUs were 1000, from the conversion of the degree of importance (1 indicates maximum importance) using a standardized weighting matrix included in the methodology. The probability of pollution occurrence (p) was also considered at its maximum value (1). This conversion framework is designed to reduce subjectivity and ensure consistency across pollutants and environmental contexts. It results that ER is similar to EI, having the same values. The method remains adaptable and may incorporate adjusted p values or event-based differentiation where detailed probability data is accessible.

$$Q={\displaystyle \frac{{MAC}_i}{{MC}_i}}$$

(1)

Equation (1). Environmental quality quantification (Q)

• where

Q is the environmental quality;

MAC is the maximum allowed concentration (in the absence of clear maximum allowed concentrations established by legislation for aerosols, it was necessary to adapt the integrated assessment method; in this case, MAC was defined as the highest AOD value in the dataset);

MC is the measured concentration.

Since no legal threshold exists for AOD, the highest measured value in our dataset was used as a relative benchmark for applying the integrated assessment method. This adaptation enables its application to column-integrated aerosol data. While this data-driven definition of MAC ensures internal consistency, it may limit transferability of results across different datasets or regions. Future implementations could benefit from using health-based or regulatory thresholds, where available. Furthermore, a rescaling procedure was introduced to differentiate between aerosol types and to ensure that the impact scores better reflect their known environmental relevance.

$${EI}_i={\displaystyle \frac{IU}{Q}}$$

(2)

Equation (2). Impact on environmental quality quantification (EI)

• where

EIᵢ is the environmental impact considering the quality indicator i;

IU is the importance unit assigned to each environmental component;

Q is the quality of the environmental component considering the quality indicator i.

$${ER}_i={EI}_i\times p$$

(3)

Equation (3). Environmental risk quantification (ER)

• where

ER_i is the environmental risk;

EIᵢ is the environmental impact considering the quality indicator i;

P is the probability of impact/pollution occurrence.

After quantifying the environmental impact and risk, their classification follows, depending on the scenario in which the obtained EI values fall, according to

Table 2. Classification of environmental impact and risk [67,68,69,70].

Impact Scale	Description	Risk Scale	Description
<100	Natural environment, not affected by industrial/human activities	<100	Negligible/insignificant risks
100–350	Environment modified by industrial activities within admissible limits	100–200	Minor risks, and monitoring actions are required
350–500	Environment modified by industrial activities causing discomfort conditions	200–350	Moderate risks at an acceptable level, monitoring and prevention actions are required
500–700	Environment modified by industrial activities causing distress to life forms	350–700	Moderate risks at an unacceptable level, control and prevention measures are needed
700–1000	Environment modified by industrial activities, dangerous for life forms	700–1000	Major risks, remediation, control and prevention measures are needed
>1000	Degraded environment, not proper for life forms	>1000	Catastrophic risks, all activities should be stopped

. The integrated risk assessment method applied in this study uses a standardized classification scheme with predefined intervals for environmental risk categories. These thresholds are not derived from literature synthesis or subject to researcher-defined criteria.

2.4. Rescaling Method

The scores resulting from the initial impact and risk assessment method exceeded the threshold of 700, which, according to Table 2, indicates a severely affected environment, unsuitable for life. However, the measured AOD values do not reflect a critical or catastrophic risk. For this reason, it was necessary to develop a rescaling method that more accurately reflects the implications of different types of aerosols. The applied calculation formula is shown in Equation (4).

$$y={\displaystyle \frac{\left(x-x_{min}\right)·\left(y_{max}-y_{min}\right)}{\left(x_{max}-x_{min}\right)}}+y_{min}$$

(4)

Equation (4). Rescaling formula

• where

x is the value to be rescaled;

x_max, x_min are the limits of the original interval (EI calculated);

y_max, y_min are the limits of the new interval (this work stopped at 700 because no measurements were made under catastrophic conditions);

y is the rescaled value.

Based on the toxicological and epidemiological literature regarding the health effects of aerosols, the rescaling logic was refined to allow differentiated classification of aerosol-specific impacts. This rescaling does not reflect observed environmental outcomes, but rather introduces differentiated weighting based on aerosol type, supported by the literature on their relative environmental relevance. This approach ensures that the method is grounded in validated practice while reflecting the heterogeneity of urban aerosols behavior. Since the scientific literature demonstrates distinct health and environmental effects of different aerosol categories, these were considered only to justify the need for differentiated rescaling of environmental impact scores between EC and DD categories. These references are not used to infer environmental risks or to redefine the classification thresholds established by the assessment methodology. These two categories—EC and DD—were selected because they are well-documented in the literature, have clearly distinguishable environmental and health impacts, and allow for a reliable and methodologically consistent rescaling. In contrast, other categories such as OC dominated, EC/OC mixtures, coated large particles, or mixed aerosols present greater challenges in individual treatment due to overlapping sources, complex behavior, or insufficiently distinct toxicological profiles. Therefore, they were excluded from differentiated rescaling in this study. While biomass burning—the main source of EC—is associated with severe diseases and increased mortality rates [15,16,17,18,19,20,21,22,23,24,25,26,27], DD aerosols have less severe effects (Table 1). Environmental impacts of these two aerosol types also differ. Although DD pollution poses certain risks, especially in arid regions, studies consistently show that its environmental impact is substantially lower compared to that of EC from biomass burning. The latter releases complex toxic compounds and contributes more significantly to global warming, whereas DD increases suspended particle concentrations with more localized and less severe impacts [74,75,76,77,78,79]. Furthermore,

Table 3. Variation in the intensity of DD environmental and health effects depending on the distance from the source.

Distance from the Source	Health/Environmental Impact	Example Evidence
Close to the source	High PM, severe health effects, frequent exposure	Most epidemiological studies, systematic reviews
Far from source	Lower PM, milder or less frequent health effects, still detectable dust	WHO fact sheet, atmospheric transport studies

provides a conceptual illustration of how the intensity of DD effects tend to decrease with increasing distance from the emission source [15,16,17,18,19,20,21,22,23,24,25,26,27,74,75,76,77,78,79]. This attenuation is relevant for the Romanian context, which is located far from major Saharan sources, and supports the rationale for assigning lower impacts scores in the differentiated rescaling method applied here. Since aerosol mixtures originate from both natural and anthropogenic sources, including this category in the risk score assessment complicates the linkage between emission sources and policy development. Based on these differences and the risk categories presented in Table 2, the EI (environmental impact) values for EC were rescaled between 0 and 700, while those for DD were rescaled between 0 and 350, thus reflecting the different magnitude of the impact on the environment and health. Rescaling was applied starting from 0, as scenarios without significant pollution were also included, allowing for rescaling from the minimum value.

Moreover, to validate the rescaling method, the standard assessment method was applied again, this time assigning differentiated scores to aerosol categories. EC was considered of utmost importance (assigning a score of 1 in impact assessment), while for the DD category, scores ranging from 0.1 to 1 were assigned progressively.

Furthermore, to validate both the proposed rescaling method and the risk-level-based scoring system, two standard statistical indicators were calculated: Root Mean Square Error (RMSE) and Mean Absolute Error (MAE), which are presented in the following section.

2.5. Validation with Statistical Indicators—Error Calculation: Root Mean Square Error (RMSE) and Mean Absolute Error (MAE)

The performance of the proposed rescaling method and the risk-level-based scoring system was evaluated using Root Mean Square Error (RMSE, Equations (4) and (5)) and Mean Absolute Error (MAE, Equations (6) and (7)). These parameters allow quantifying differences between rescaled values and values obtained from applying the integrated method for EC and DD, thereby providing an objective basis for evaluating the method’s performance.

$$RMSE\left(absolute\right)=\sqrt{{\displaystyle \frac{1}{n}}\sum _{i=1}^n{\left({x_i-\widehat{x}}_i\right)}^2}$$

(5)

Equation (5). Absolute Root Mean Square Error

$$RMSE\left(\%\right)={\displaystyle \frac{RMSE\left(absolute\right)}{Amplitude of the rescaled EI value}}\times 100$$

(6)

Equation (6). Percentage Root Mean Square Error

$$MAE \left(absolute\right)={\displaystyle \frac{1}{n}}\sum _{i=1}^n\left|{x_i-\widehat{x}}_i\right|$$

(7)

Equation (7). Absolute Mean Absolute Error

$$MAE\left(\%\right)={\displaystyle \frac{MAE\left(absolute\right)}{Amplitude of the rescaled EI value}}\times 100$$

(8)

Equation (8). Percentage Mean Absolute Error

• where

n is the total number of data (EI in our case);

x_i is the rescaled EI value for observation “i”;

${\widehat{x}}_i$ is the EI value for the observation “i”;

Amplitude of the rescaled EI value: The range was established from 0 to 700 (corresponding to a value of 700 of the amplitude) for the EI considered in relation to EC, and from 0 to 350 (corresponding to a value of 350 of the amplitude) for the EI considered in relation to DD.

RMSE is sensitive to large deviations and penalizes them more heavily, making it useful for identifying score combinations that generate significant discrepancies—an essential aspect in establishing the most suitable pair of scores for EC and DD. At the same time, MAE indicates the average absolute deviation, offering a clear picture of the average error across the entire dataset. The use of both indicators allows for solid validation and a comprehensive understanding of the method: RMSE helps detect situations with large errors, while MAE provides a general benchmark of the method’s overall accuracy [80,81,82,83,84,85,86].

3. Results and Discussion

3.1. Overview of Aerosol Type Distribution (Iasi and Cluj-Napoca)

3.1.1. Iasi AERONET Monitoring Site

For Iasi, 1209 measurements were selected from 2012 to 2023, of which 975 correspond to the EC category and 234 to the DD category. Previous studies showed that DD events affecting Romania typically carry mineral aerosols rich in aluminum, calcium, potassium, and iron [87]. Aerosols measured by AERONET over Iasi, in general, were dominated by an EC and OC mixture at 55% (see “EC/OC mixture” in Figure 2). Fine mode particles, categorized as “EC dominated” in Figure 2, indicate that 18% of the aerosol mass came from combustion/biomass burning or soot particle types originating from combustion-related sources, including biomass burning. The aerosol categories “Dust/EC mix” (1%), “Mix” (2%), “OC dominated” (3%), “OC/Dust Mix” (6%) and “Coated large particles” (11%) are not discussed here, since they exhibit complex or mixed source profiles and less clearly defined environmental and health impacts such that a meaningful assessment of environmental relevance would require more complex and targeted studies beyond the scope of this paper. In contrast, the “Dust dominated” category was retained for analysis due to its clearly defined origin and well-documented relevance in the scientific literature.

3.1.2. Cluj-Napoca AERONET Monitoring Site

For Cluj-Napoca, 1380 measurements were selected from 2011 to 2023, of which 1222 correspond to the EC category and 158 to the DD category. Aerosols measured by AERONET over Cluj-Napoca, in general, were also dominated by an EC and OC mixture at 46% (see “EC/OC mixture” in Figure 3). The “EC dominated” in Figure 3 indicates that 22% of the aerosol came from combustion/biomass burning or soot particle types. Other aerosol types, including “Mix” (2%), “Dust/EC mix” (3%), “OC/Dust Mix” (8%), “Coated large particles” (8%) and “OC dominated” (10%), are not discussed here.

3.2. Performance Evaluation Using RMSE and MAE

To validate the rescaling method and identify the most appropriate pair of scores for EC and DD, we based our analysis on error assessment (Figure 4).

Results show that, for both cities, the lowest values of both error metrics (RMSE and MAE), located in the percentage range 0–5% were achieved with importance scores of 0.4 and 0.5 for DD and 1 for EC (Figure 4). Thus, for Iasi, it can be observed that when the importance score for DD is set to 0.5, the RMSE values are 2.0% for EC and 4.0% for DD, while the MAE values are 1.8% for EC and 3.7% for DD. When the importance score for DD is set to 0.4, the RMSE is 3.2% for EC and 3.3% for DD, and the MAE reaches 3.2% for EC and 2.4% for DD. In the case of Cluj-Napoca, with a DD importance score of 0.5, the RMSE values are 2.8% for EC and 4.6% for DD, and the MAE values are 2.6% for EC and 4% for DD. When the importance score is set to 0.4, the RMSE rises to 4.2% for EC and 5% for DD, while the MAE reaches 4.1% for EC and remains at 4% for DD. Therefore, the error analysis confirms the validity of the selected rescaling thresholds and importance scores, demonstrating their effectiveness in accurately capturing the real differences between aerosol categories.

3.3. Environmental Impact Quantification Calculated Using AERONET Data

Figure 5, Figure 6, Figure 7 and Figure 8 show the rescaled EI, calculated using AOD values from the Iasi and Cluj-Napoca monitoring stations, considering only the “dust dominated” and “EC dominated” categories. The “EC dominated” category falls within the subset of fine particulate matter associated with pollution from anthropogenic sources, such as fossil fuel combustion, industrial emissions, and road traffic.

The categories “EC/OC mixture” and “OC dominated” are also specific to fine particles; however, in terms of health effects, EC is considered more toxic than OC [29]. The “EC/OC mixture” combines natural and anthropogenic sources, making it complex to consider this category when calculating a risk score and correlating pollution sources with policy making.

The distribution of the rescaled environmental impact values for the EC category in Iasi (Figure 5) shows a predominance in the minor to moderate risk range. More than half of the recorded values (53.3%) fall between 100 and 350, suggesting an acceptable level of impact that still requires monitoring. Only 0.2% of the values exceed the threshold of 500, indicating isolated episodes of severe pollution, likely attributed to meteorological conditions or intense anthropogenic sources. The AOD associated with these episodes ranges from 0.88 to 1.05, reflecting a significant aerosol load on the atmospheric column. Therefore, EC remains a major pollutant of concern for local public health and environmental strategies.

For the DD category in Iasi, the impact distribution is significantly more attenuated (Figure 6). The majority of rescaled values (65.4%) fall below the threshold of 100, reflecting minor or even negligible risks. This trend is consistent with the hypothesis that DD effects diminish with increasing distance from the emissions source. Even during episodes with higher AOD, the rescaled impact rarely exceeds 300, supporting the validity of the differentiated impact assessment method. Thus, DD’s contribution to air quality degradation in Iasi appears to be marginal compared to that of EC.

For Cluj-Napoca, the impact distribution for the EC category (Figure 7) is similar to that observed in Iasi, with 55.9% of values falling between 100 and 350, indicating moderate but acceptable risks. As in Iasi, only 0.2% of the values exceed the threshold of 500, suggesting a low incidence of critical pollution events. The maximum recorded AOD value (0.88) is slightly lower than in Iasi, which may reflect either lower combustion source intensity or more efficient pollutant dispersion. These results confirm that EC remains a dominant factor in atmospheric impact within the urban environment.

For DD in Cluj-Napoca (Figure 8), the impact distribution follows the same pattern as in Iasi, with generally low values; however, the majority (58.9%) falls within the 100–350 range, characteristic of minor risks where monitoring actions are required. The AOD associated with this interval ranges from 0.24 to 0.67. This behavior confirms the secondary role of DD in the pollution profile of Cluj-Napoca.

Because the traditional approach to air quality assessment does not correlate with ground data, this study’s results underscore the urgent need to utilize an AOD risk index to effectively highlight the impact of pollution on the entire atmospheric column. Thus, in addition to the climate change already highlighted by scientists, the radiative budget (influenced by aerosol intrusion) can also negatively impact life quality and may be of interest.

On the other hand, the effects of other pollutant categories are complex; therefore, understanding the impact of each aerosol type is important, and future comprehensive research is needed to determine these effects. The visibility index could also be explored as a supplementary metric in future research [88].

4. Conclusions

This research demonstrates the applicability and relevance of an innovative integrated approach for air quality evaluation and the assessment of impact and risk related to atmospheric pollutants, using AOD as a key indicator. Environmental impact and risk scores were calculated and classified according to the methodological framework applied, while the health-related literature served solely as background for contextual differentiation in the proposed rescaling method. The study was conducted for two Romanian cities—Iasi and Cluj-Napoca—using AERONET data collected over extended periods (12 and 13 years, respectively), ensuring the reliability of the findings.

Results reveal a predominance of EC and OC mixtures in both cities, primarily sourced from traffic emissions, biomass burning, and industrial activities. Specifically, Cluj-Napoca exhibited an EC/OC mixture of 46%, while Iasi had 55%. These values are in line with expectations for urban environments characterized by traffic-related emissions and residential biomass burning, both of which are major sources of EC. The slightly higher EC share in Iasi likely reflects the influence of more intense traffic flows and increased reliance on wood-based heating during colder months, compared to Cluj-Napoca. The impacts and risks calculated as the AOD index for EC show a few major ones, with an AOD index range from 0.88 to 1.05 for Iasi city and 0.73–0.88 for Cluj-Napoca.

By adapting the Integrated Impact and Risk Assessment method to the specific context of aerosol characterization, the study offers a differentiated classification of aerosol types, emphasizing the higher risks associated with EC compared to DD. The proposed method was statistically validated using RMSE and MAE indicators, achieving excellent performance (<5%).

Therefore, this study offers an effective and replicable tool to support environmental authorities in monitoring, forecasting, and decision-making processes regarding air quality—especially in urban areas where PM_2.5 pollution is persistent and often insufficiently quantified by ground-based monitoring networks.

This study focuses on the integrated assessment of aerosol impacts and risks using AOD data and does not explicitly consider impacts on specific ecosystems, infrastructure or population distribution. It also does not evaluate aerosol deposition on vegetation or built surfaces, nor does it address processes such as pollutant uptake through plant tissues or material degradation. Direct validation of AOD data against ground-based PM_2.5, BC/EC or OC concentrations from European monitoring networks was not performed. Future research will aim to integrate AOD data with in situ measurements to improve spatial accuracy and quantification of surface pollution, as well as to include detailed ecosystem and urban/population maps, which could improve future analyses of the link between aerosol deposition and environmental and societal impacts. Although AOD does not directly quantify surface-level concentrations, its spatial variability remains a useful indicator for comparative assessments of aerosol loading in urban areas. We also recognize that the lack of legally defined thresholds for AOD required the use of a dataset-derived MAC, which may limit the generalizability of the results. The methodology remains adaptable to the inclusion of standard regulatory benchmarks in future applications. With that said, given the strong performance of the developed method, another future research direction involves its application at a more detailed temporal scale—enabling the analysis of aerosol pollution risks based on seasonality, specific days, or annual variations—as well as extending the study to other urban or rural regions to allow for comparative assessments and the development of regional pollution control strategies.