3.1.1. Emission Data

Due to the location of the studied area (direct neighborhood of Poland, Czechia and Germany), and the resulting problems with the unification of emission databases, data from three sources described below were used for more comprehensive analyses.

In the area of Poland, data from a detailed emission inventory were used for annual air quality assessments performed up to 2017 by the Voivodship Inspectorate for Environmental Protection in Wrocław (currently the Regional Department of Environmental Monitoring of the Chief Inspectorate for Environmental Protection) [28,29]. This database contains data on all types of emissions from the voivodship, including emissions from residential heating (SNAP0202), transport (SNAP07), agriculture (SNAP10) and industry (SNAP0201, SNAP01, SNAP03, SNAP04). Inventories took into account the activity of sources. Due to the fact that the above database has been created continuously since 2012, data contained therein can be considered of a good quality. The database was additionally validated with annual model calculations [30].

In the calculations, an indispensable element is capturing the temporal and spatial variability of the emission field, because such changes together with the variability of meteorological conditions practically determine the final distribution of pollution concentrations in the studied area. Considering the temporal variability of emissions can also capture the cases of high concentration episodes and link them to the types of emissions responsible for poor air quality. Therefore, the concentration calculations were implemented in an emission model, which includes primarily variations dependent on changes in meteorological conditions (e.g., temperature for emissions from residential heating and precipitation for emissions from transport, agriculture) or the mode of source operation. This methodology was used, among others, in modeling air quality for the needs of annual air quality assessments in the Lower Silesian Voivodship [29]. The results of the calculations obtained in the context of the balance of specific sources in PM2.5 and PM10 emissions in Bogatynia commune are presented in Figure 4.

**Figure 4.** Particulate matter emissions for different emission sources inventoried in Bogatynia commune (source: own study).

The performed analyses show that Turów power complex, particularly the mine, has the largest share in the total PM10 emission in Bogatynia commune. In the case of PM2.5 emissions, the communal and household sector is clearly the main source of emissions in the studied area. While emissions from household sources, or from transport, can be considered a seasonal variable (heating and post-heating season, daily and weekly transport cycle) and of a quasi-uniform nature, the emission associated with industrial sources may be characterized by short-term changes and be dependent on a number of factors resulting from, e.g., technology, specificity of the sources or a specific mode of operation of an industrial installation. Therefore, considering the impact of industrial sources on air quality, it was necessary to build an emission model dedicated to the examined facility (in this case, ELT), which was presented in the report on the environmental impact of the KWBT [23].

In the case of the ELT power plant, organized emissions resulting mainly from the boiler operating conditions were taken into account. In addition, in the case of ELT, fugitive emission sources were considered, i.e., combustion waste buffer area, where ash from boilers is temporarily stored before being mixed with overburden in an open pit. This happens when it cannot be collected by the mine (KWBT). Emissions for this source were determined on the basis of United States Environmental Protection Agency (US-EPA) emission factors adapted to the characteristics of the source in terms of quantity and quality of the stored material, including machine operation [23]. Examples of emission values included in the calculation scenarios are presented in Table 2.


**Table 2.** PM10 and PM2.5 emission from ELT in 2018.

In the case of KWBT, four basic groups of emission sources were identified: emissions related to the mining process, storage and reloading of coal, transport and entrainment of material from exposed surfaces. The emissions related to the mining process covered the exploitation area (northern part of the open pit) and the backfilling area (southern part of the open pit). It was assumed that the volume of emissions from the exploitation area compared to the backfilling area would be significantly lower due to the quarried material, which is very moist and heavy. In the backfilling area, the quarried material is much more volatile because it consists mainly of overburden loam and ash from power plant boilers. Emissions related to storage and reloading concerned a coal bin located in the north of the facility. In the area of the bin, there are two trenches in which higher quality coal is deposited and two coal sales points—wholesale and retail. In addition to emissions from coal loading and unloading, emissions associated with wind entrainment are an important factor in this case. The last group is formed by emissions related to transport, both through belt conveyors and through vehicles and railways (wholesale and retail). The emission model for the mine was based on indicators determined with reverse modeling, US-EPA methodologies, considering the specificity of meteorological conditions and mineral material mined and backfilled in the open pit, and the time of operation of the machines [23,30]. Examples of emission values included in the calculation scenarios are presented in Table 3.

**Table 3.** PM10 and PM2.5 emission from ELT in 2018 by sector (on the basis of [23,30]).


The last group of data is information on emissions from Czechia and Germany from the emission inventory posted on the public websites of the Czech Hydrometeorological Institute and the German Ministry of the Environment [31,32]. In Germany, the inventory included sources related to residential heating (SNAP 0202) and SNAP 07 road transport (also SNAP 08 other transport), and the total emission of PM10 and PM2.5 was 44.6 Mg. For Czechia, the data concerned industry (SNAP01, SNAP03, SNAP04), residential heating (SNAP0202) and road transport (SNAP07). Total emissions in the studied area are around 182 Mg for PM10 and 169 Mg for PM2.5.

#### 3.1.2. Validation of Modeling Results in the Base Year

The obtained calculation results were validated using the available measurement data from measuring stations located in Poland, Czechia and Germany (Figure 3, Table 2). The stations were selected based on the availability of results in publicly available databases (e.g., AIRBASE, EEA European Air Quality Portal, JPOAT via the Air Quality Portal of the Chief Inspectorate for Environmental Protection). It was also assumed that the completeness of the measurement series must meet the requirements of the CAFE directive [1]. Ultimately, the analysis included measurements from six stations: four in Poland (Wyszków, Jasna Góra, Bogatynia, Działoszyn), one in Czechia (Frýdlant) and one in Germany (Zittau), the location of which is shown in Figure 5. Three of the

selected stations located in Poland belong to the measurement network of ELT, and the station in Działoszyn is included in the network of the State Environmental Monitoring (SEM) run by the Chief Inspectorate of Environmental Protection under the national code DSDzialoszyn. The stations in Zittau and Frýdlant are national stations of the German and Czech networks.

**Figure 5.** Location of measuring stations (source: own study).

The relative error rate of the modeling result in relation to the measurement was the basic measure of the correctness of the results of model calculations. According to the CAFE directive, this parameter for particulate matter (including PM10 and PM2.5) should not exceed 50% for annual mean values [1]. Negative values indicate an underestimation of concentrations.

The values obtained as a result of the comparison allowed us to determine the uncertainty of the model (relative errors) and are listed in Table 4. In most cases, the uncertainty of the model did not exceed 13%. The highest compliance was obtained at the station in Bogatynia (for both PM10 and PM2.5) and at stations in Zittau and Frýdlant, where the relative errors did not exceed 5%. The highest 32% underestimation of results was obtained at the station in Działoszyn. At the same time, it can be stated that a match of results was much better for PM2.5 concentrations than for PM10.


**Table 4.** Annual mean concentrations of PM10 and PM2.5 modeling results and measurement.

#### 3.1.3. Emission Change Scenarios

The assessment of the e ffectiveness of planned long-term actions or ad hoc measures limiting the emission of pollutants is carried out, among others, through changes in the emission introduced into the model, activity of sources and through a possible modification of the original assumptions regarding, e.g., operating mode or frequency of preventive measures used. These changes may concern both the volume of emission loads and the modification of time variations. Then, based on the new emission values, re-calculations were made, assuming no changes for the remaining emission data and/or meteorological parameters. This study discusses three scenarios for emission changes. The first scenario is related to changes in the volume of emissions from the mine, indicated in the report on the mine's Environmental Impact Assessment [23]. The second scenario evaluates the e ffectiveness of implementing the anti-smog resolution in force in the Lower Silesian Voivodship, which will result in a deep modification in the profile of residential heat sources [11]. The last scenario is an assessment of the implementation of both previously mentioned scenarios. The data included in the calculations for the three assumed scenarios are summarized in Table 5.


**Table 5.** Emission for individual sources included in the calculations for individual scenarios.

As a result of the analyses presented in the report on the environmental impact of the mine [23], in order to avoid a significant impact of the object on the neighboring areas, it was necessary to indicate possible additional mitigation measures (Table 6).

In addition, in 2019, a new ash conveyor was created directly from the power plant to the open pit, which will eliminate the impact of the storage area. It was estimated that, as a result of the implemented measures, the total emissions from the mine will be reduced from 935.4 Mg by approximately 44% for PM10 and from 363.6 Mg by approximately 43% for PM2.5 and will be 521.5 Mg and 207.9 Mg, respectively.

Another important change in the emission characteristics of the commune is the implementation of the anti-smog resolution, according to which, by 2028, boilers of a class lower than 5 will not be operating in the entire Lower Silesian Voivodship. Within the voivodship, it is allowed to burn solid fuels in devices from which "particulate matter emissions do not exceed the emission threshold values set out in Commission Regulation (EU) 2015/1189 of 28 April 2015 implementing Directive 2009/125/EC of the European Parliament and of the Council with regard to ecodesign requirements for solid fuel boilers" [12]. The implementation of the resolution is estimated to result in around a 95% reduction in dust pollution. Total PM10 emissions are to be reduced from 452.8 to around 19 Mg and PM2.5 from 394.3 to 18.1 Mg. The values of emission totals and the scale of emission reduction were estimated on the basis of an emission database from a detailed inventory of emissions prepared for annual air quality assessments carried out until 2017 by the Voivodship Inspectorate for Environmental Protection in Wrocław (currently the Regional Department of Environmental Monitoring of CIEP). This database contains, among others, information on surfaces heated using old-type boilers fed with coal and wood.



#### *3.2. Health Risk Assessment*

The health risk assessment associated with long-term (annual) PM2.5 exposure for the analyzed scenarios was performed using the dose–response function and relative risk index (RR (95% CI) = 1.062 (1.040–1.083) for every 10 μg/m3) [33,34].

The crucial element of health risk assessment for the analyzed area was accurate mapping of the exposure of the population at risk. For this reason, the health assessment was performed for each of the separated areas adopted as the air quality model grid, based on obtained results of the level of annual mean PM2.5 concentration and demographic indicators assigned to the model grid (Figure 6). The number of exposed people and the number of deaths classified by causes were determined on the basis of data from official statistics with a division of the population into age groups [35]. For more detailed analysis, data from spatial distribution were used, considering the current administrative division of the country according to Nomenclature of Territorial Units for Statistics (NUTS) level 5 and the state register of administrative borders [36–39]. For this purpose, among others, the dasymetric map of population density developed by the European Environment Agency—raster layer "Population density grid of EU-27+, version 4 and 5"—was used. The dasymetric map and population density were corrected in line with the current land use information. To verify the population in the model grid, the National Official Register of the Territorial Division of the Country was used with the distinction of towns with district rights, urban, rural and urban-rural communes as well as towns and rural areas in urban-rural communes.

The analyses were performed for the baseline condition without considering emission reduction scenarios and for each scenario separately. The results are presented both in the form of the estimated number of premature deaths related to the exposure of the general population to PM2.5 and relative changes in the impact assessment for each scenario.

**Figure 6.** Map of the population density distribution (source: own study).
