Clean air is important for a long and healthy life. In order to achieve the aim of clean air, immission limit values for air pollutants were set in Germany based on the limit and guideline values set by the EU Directive 2008/50/EC in the 39. BImSchV (39. Federal Immission Control Act). Immission limit values define the impact of harmful substances on humans and the environment [1
]. Besides immission limits for gaseous substances (e.g., NO2
, CO, etc.), immission limits for particulate matter are defined for PM10 and PM2.5 categories. While PM10 is regulated based on a daily and annual average, the PM2.5 immission limit is currently only defined as an annual average. The WHO (World Health Organization) publishes guideline values for air pollutants regularly. A new air quality guideline was released in 2021, in which the guideline values for PM2.5, PM10, and NO2
are tightened in comparison to the last guideline of 2005 [2
]. Since this paper focuses on PM2.5 immissions, the current limit values for PM2.5 of the 39. BImSchV as well as the guidelines from the WHO from 2005 [2
] and 2021 [3
] are listed in Table 1
Compared to other countries, as well as to the guideline values of the WHO, there are significant differences in the particulate matter immission limits. The WHO recommends a new annual guideline value for PM2.5 of 5 µg/m3
, while the European annual average of 25 µg/m3
is five times as high [2
]. The 39. BImSchV does not specify a limit value for daily mean values of PM2.5 [3
According to current studies, particulate matter is a harmful component and can lead to cardiovascular and lung diseases in the case of long-term exposure. It has also been shown that health is negatively affected even at low PM2.5 concentrations (e.g., 8.7 µg/m3
in the study of Crouse et al., (2012) [5
]. Even if fine dust concentrations in Germany are declining, efforts should be made to further reduce exposure [1
]. The evaluation of the annual mean values of PM2.5 at selected measuring stations in Germany are illustrated in Figure 1
The European PM2.5 annual average of 25 µg/m3
has not been exceeded at the measuring stations measuring PM2.5 in the last decade. However, the annual WHO guideline value of 2005 (10 µg/m3
) was undercut for the first time at the measuring stations in rural areas in 2015. At this time, the stations in the suburban background and near traffic were 1.3 times (suburban background) and 1.5 times (near traffic) above the specified annual guideline value of the WHO (2005). The annual mean value of 10 µg/m3
(WHO 2005) was undercut in all three areas (rural, suburban, and near traffic) for the first time in 2020. In this context, it is important to mention that, in 2020, the COVID-19 pandemic in Germany led to some lockdowns, which may also have contributed to the decrease in immissions [9
]. However, due to the current tightening of the WHO guideline, the measured annual mean values are again above the guidelines of the WHO Air Quality Guideline (2021). The difference in PM2.5 concentrations between rural and urban areas (see Figure 1
, near traffic and suburban background) has converged in recent years.
The presented annual mean values (from Figure 1
) refer to measuring stations in Germany. Table 2
lists the classifications of the monitoring stations in Baden-Württemberg measuring PM2.5 and the overages in 2019 and 2020.
If the measuring stations for PM2.5 of the LUBW (Landesanstalt für Umwelt Baden Württemberg) are considered, it is noticeable that these are mainly located in urban backgrounds and near traffic. This is due to the regulation of the 39. BImSchV Annex 5 in Germany. This ordinance specifies the minimum number of sampling points for stationary measurements in metropolitan areas. The number of measuring stations is related to the population size. In Karlsruhe (population on 31.12.2020: 303 970 [13
]), for example, three measuring stations may be set up for PM measurements. One measuring station is to be set up in the urban background and another at a traffic location [4
]. However, this regulation does not address residential areas. The highest exceedances in 2019 (Table 2
) are measured at the stations near traffic and with urban background (most exceedances in 2020 are at the stations near traffic), but due to the few rural stations, no statements can be made about the air quality in residential areas in Baden-Württemberg. For this reason, the LUBW carried out fine dust measurements (PM10 and PM2.5) in three communities in the Black Forest (Rastatt, Forbach; Lörrach, Kleines Wiesental and Ortenaukreis Schuttertal in Germany) in order to evaluate the impact of wood combustion on fine dust pollution [14
]. The annual mean values of the measurements in the Black Forest (Forbach, Kleines Wiesental, Schuttertal) and, as a comparison, the annual mean values of two measuring stations in Karlsruhe (Reinhold-Frank-Street, Nordwest) in 2019 are presented in Figure 2
In this case, the measurement results confirm that the fine dust limits for PM2.5 are observed, and even the stricter guideline values of the WHO of 2005 are just exceeded at one test station (Schuttertal) for the PM2.5 concentration. However, the measured values are comparable to the stations in Karlsruhe close to traffic and are slightly below the Reinhold-Frank-Street. In the winter months, the particulate matter is dominated by wood-fired heating systems, and it is partially higher than at the urban background. The determination of the contribution of wood combustion was based on the levoglucosan concentration in the particulate fraction PM10. Levoglucosan is an anhydro-sugar that is formed during the combustion of carbohydrates (e.g., starch or the cellulose contained in wood) [14
]. The presented study clarifies that the limit value of the BImSchV is undercut by all stations. However, the guideline value of the WHO 2005 is partially exceeded and the tightened guideline value from the year 2021 was exceeded by all stations. The study of the LUBW reveals that measuring stations in rural areas are also important in order to check air quality and compliance with limit values there as well.
Other studies from Northern Europe show comparable results. For example, Glasius et al. (2006) and Krecl et al. (2008) already showed years ago that PM concentrations in rural areas without traffic were at a similar level as in large cities, where air pollution mainly resulted from street traffic [15
]. This ratio has increased in recent years [7
]. A study by Kukkonen et al. (2020) compared the impact of residential wood combustion in four Nordic cities. It was shown that, in Helsinki and Copenhagen, the highest levels of particulate matter were measured outside the city center in suburban areas [18
]. Nevertheless, air quality is mainly measured in large cities, where the amount of wood combustion for energy production is lower.
High immissions of particulate matter can also be a result of chimneys that are installed too low. Chimneys built before the amendment of the 1. BImSchV (13 October 2021) did not have to be erected near the highest point of the roof (horizontal distance from the ridge is smaller than its horizontal distance from the eaves) with an overhang of at least 40 cm [19
]. In VDI3781 part 4, it was shown how an improperly installed chimney causes the exhaust gas to remain in a recirculation zone, preventing the exhaust gas from flowing away [21
]. This leads to an increase in PM pollution and reveals that the BImSchV represents the state of the art just for newly built chimneys. During inversion weather conditions, even a higher chimney and, therefore, the optimal discharge of exhaust gases from wood-burning stoves is not guaranteed. Due to the different temperatures of the air layer, the particulate matter cannot be mixed and the pollutants concentrate in the layers close to the ground.
In this study, PM2.5 pollution will be measured simultaneously and temporally at three different locations, urban (near-traffic), suburban (near-traffic), and residential (non-traffic) background. The focus is on a residential area, as current studies either measure air pollution in a city (traffic) or rural areas with a lower population density than residential areas outside metropolitan regions. Because of the regulations in Germany (39. BImSchV), measuring stations are rarely installed in residential areas to measure the fine dust pollution with respirable PM2.5. Here, however, the breathing air quality can be worse than on high-traffic streets. Due to the trend of wood combustion as an alternative heating method, PM2.5 concentrations during different months, as well as extreme events (PM2.5 concentrations > 40 µg/m3), will be observed more closely in this study. The aim is to investigate and compare the short-term (peaks), medium-term, and long-term particulate matter levels and to assign them to sources by observation, considering wind direction simultaneously at three locations in the area of Karlsruhe.
3.1. PM2.5 Concentration—Time Courses at the Measuring Stations
depicts the temporal development of the PM2.5 concentrations measured at the three measuring stations in January and May. In January 2021, the temperature in the area of Karlsruhe ranged between 4.4 and 13.8 °C, resulting in an average temperature of 3.1 °C. In May 2021, the minimum temperature was 3.8 °C and the maximum temperature was 29.8 °C, leading to an average temperature of 13.0 °C. Short-term data failures occurred due to disturbances in the data transmission and are presented as data gaps in the curve progressions. In January, the particulate matter concentrations at all stations ranged from approximately 2 to 60 µg/m3
. High concentration peaks up to 60 µg/m3
mainly occurred in the residential area. Short-term peaks also appeared at the other stations, but these reached comparatively lower concentrations. In general, more peaks were determined in January than in May. In May, the concentration was mainly between 2 and 40 µg/m3
—apart from 60 µg/m3
, detected once at the urban location. The two selected months are also comparable with other months, e.g., January is comparable with December and May with the available results of June.
In addition to the short-term peaks, the concentration courses at the three stations are superimposed. This suggests large-scale PM2.5 pollution in the Karlsruhe area. In May, the large-scale PM2.5 concentrations are about two times lower than in January at all stations.
The illustrated lower concentrations are also reflected in the monthly mean concentration, i.e., the background levels, at the measuring stations (Table 6
). Since the high concentration peaks especially in January, the monthly mean concentration is comparatively low. The largest PM2.5 exposure was determined in the suburban area with 7.9 µg/m3
. In the residential area and at the urban location, the monthly mean concentration fluctuates around a comparable value of 6.9 µg/m3
and 6.3 µg/m3
, in consideration of the respective standard deviation. In May, the deviations between the monthly mean values are smaller. Consequently, particulate matter levels do not vary as much between stations because there are fewer local events with high and prolonged exposure.
The comparatively low monthly PM2.5 concentrations, especially in January, imply that temporally limited and large particulate matter loads are only slightly considered in the balancing of the annual limit values.
In order to verify the temporal occurrence of PM2.5 pollution in January and May, the PM2.5 concentration is plotted as a function of cumulative time in Figure 8
Missing measuring values due to data failures were left out for the sake of clarity. Due to the missing data, the temporal exceedance of the limit value is underestimated.
As was already evident from the monthly averages, the concentration level in January is almost twice as high as in May. High short-term concentrations over 25 µg/m3 account for approximately 1% of the month in January and 0.5% in May. Three-quarters of the time, the concentration is in a range of 2 to 11 µg/m3 in January and, respectively, at 2 to 5 µg/m3 in May.
3.2. Comparison of the Three Stations
For a comparison of the three stations, the suburban as well as the residential station are considered concerning the measuring station close to traffic. For the sake of clarity, the evaluation is based on 4-h average values. The 4-h average concentrations in the suburban area and the residential area are compared to the traffic-dominated urban location. In Figure 9
, the ratio of the particle concentrations (values) is illustrated for January 2021 (a) and May 2021 (b).
Generally, the differences in PM2.5 concentration at the suburban measuring station and the residential area concerning the urban station are larger in January than in May. In January, the concentration level reached was often two to eight times higher in the overages than in the undercutting of the reference concentration. Especially, in the residential area, the concentration was up to eight times higher than at the urban area (near traffic) on two days. In May, the largest deviations occurred in the residential area. Here, up to 3.7 times higher PM2.5 exposures are determined over almost 12 h.
In order to show trends, the number of data points is related to the particular range (0.5 steps) of the ratio of PM2.5 concentration in Figure 10
. In January, as well as in May, most detected concentrations are one to two times higher at the suburban and the residential area than at the urban location. In January, 67–71% of the measured values are above the level measured at the urban area. The proportion of significantly higher concentration ratios > 3 also occurred more frequently in January. In May, 57–61% of the measured values are below the reference concentration. Furthermore, the reference concentration is exceeded only four times by factors > 3 in the residential area, in May. Accordingly, significantly higher levels of particulate matter occurred at the measuring stations further away from traffic than in the immediate vicinity of traffic, particularly in January.
3.3. Frequency of Short-Term Air Pollution with PM2.5—Peak Evaluation
In the following, to analyze the local peaks (short-term events, which have high particulate matter concentrations), only the maxima of the measured PM2.5 concentrations, which represent short-term pollution, are observed. In this case, a peak is defined as a local maximum, which has a PM2.5 concentration above 25 µg/m3
(local peaks below 25 µg/m3
are not considered). The limit concentration for the peak evaluation was selected based on the annual limit value of the 39. BImSchV [4
]. For this evaluation, the minute mean values of the measurements were used.
presents the comparison of the concentration peaks of the evaluated months: January (a) and May 2021 (b). The larger the point, the higher the pollution at the respective measuring station. The comparison of the two figures (Figure 11
a,b) shows that more peaks were measured in January than in May 2021. In January, the PM2.5 pollution in the residential area is significantly higher compared to the other two stations. At this station, 66% more peaks occurred and reached maximum values over 60 µg/m3
, which is more than twice the current annual average value of 25 µg/m3
. In January 2021, it is noticeable that the peak PM2.5 concentrations in the residential area mostly appeared in the afternoon and evenings as well as on Fridays. Concentration peaks in the night (0:00–5:00) and in the morning hours (5:00–9:00) were not measured in the residential area. In these periods, most of the concentration peaks were detected in the urban station and are distributed during the week, without regularity on a weekday, and have an average concentration of 35 µg/m3
(average of all peaks), which reduces only slightly in spring to 33.5 µg/m3
. Compared to the monitoring station close to traffic (urban), the residential area exceeded the average of all peaks with an average peak concentration of 40 µg/m3
in January 2021. Peaks at the suburban measuring station were rare in January and May 2021, and thus have no regularity in terms of day of the week and time of day.
A comparison of the number of peaks above 25 µg/m3
at the three stations is presented in Figure 12
. In January 2021, the residential area had three times higher numbers of peaks than the station close to traffic (urban). In May, the number of peaks decreased at the suburban station and the residential area. At the urban station, the number in January and May is comparable.
In Figure 13
, the PM2.5 concentration curves of two exemplary days in the residential area, 22 January and 6 March, are shown to illustrate the PM2.5 immission levels. This station was evaluated based on the previous measurement data (see, e.g., Figure 12
), since the loads in the residential area are significantly higher than the urban and suburban station in winter.
On 22 January 2021, a low PM2.5 concentration was measured in the residential area during the day. Starting at 5 p.m., the PM2.5 concentration increased steeply and reached a value of 38 µg/m3. Afterward, there is a peak about every half hour until the maximum peak of over 60 µg/m² is reached at 7:16 p.m. The PM pollution decreases until 9 p.m., with one more peak (23 µg/m3) at about 8 p.m.
On March 6, in contrast, the PM2.5 pollution was also significantly higher during the day and rising steadily from 8 a.m. until about 18 µg/m3 at 1:30 p.m. The measured peaks above 25 µg/m3 were recorded on this day at 6 p.m. and repeatedly recur every hour, with a maximum peak concentration of 42 µg/m3. The concentration curves shown here illustrate a common ambient air quality pattern in residential areas that PM2.5 immissions increase significantly in the evening hours.
Another example of PM2.5 immission measurement and source identification by optical observation is shown in Figure 14
. Here, Figure 14
a presents a schematic map of the location of the measuring station and the source of the PM2.5 pollution with the direction of propagation. The source of particulate matter is located north of the measuring station.
The corresponding image (Figure 14
b) to the measurement on 20 December (Figure 14
c) shows the transmission of the exhaust gas from the chimney in the direction of the measuring station. The exhaust gas from the combustion does not rise vertically but flows in a horizontal direction to the measuring point. This is attributable to the fact that the height of the chimney is too low and is not installed at the highest point at the single-pitch roof. The image was taken on 20 December at 18:22. On this day, the PM2.5 concentrations were about 20 µg/m3
in the time between 10 a.m. and 1 p.m. Due to problems with data communication, the measured PM2.5 concentrations were not recorded before 10 a.m. In the evening, the measured PM2.5 concentration increases steeply and reaches a maximum value of 144 µg/m3
at 6:20 p.m., shortly before the optical observation (Figure 14
b). Subsequently, further hourly peaks with concentrations of about 40 µg/m3
were measured. As already described above, a smoking chimney in the immediate vicinity of the measuring station was identified as the source for the short-term high particulate matter immissions. In this case, the visual observation was made between the measured values and the visible contamination caused by a combustion process.
The study confirms that the PM2.5 concentration at all three measuring stations is approx. 50% higher in January than in May, which other studies have also shown [16
]. In Karlsruhe, 31 heating days and 5 freeze days were counted due to the low temperatures in January (average temp., 3.09 °C) [27
]. January is in the heating season, which confirms the higher PM concentrations. However, the basic pollution by traffic remains comparable in summer and winter [16
The annual EU limit value of 25 µg/m3
is not exceeded in the considered two months in city areas as well as in rural areas. Since only two months of 2021 are treated in the evaluation, no statement for the annual mean value can be made from the data. Nevertheless, the data reflects the trend of recent years, as the PM2.5 concentration has been decreasing for years and the annual limit value has not been exceeded in Germany since 2010 [8
]. If the monthly mean values are compared with the value recommended by the WHO guideline of 5 µg/m3
], it can be seen that, in January, all three stations are above the guideline value. In May 2021, the behavior is different; here, the value recommended by the WHO guideline can be complied with. The results already reveal, based on the monthly mean values, that higher concentrations of fine particulate matter in winter are compensated for by lower concentrations in spring. Consequently, temporally limited high particulate matter pollution has no significant impact on the averaging process of the limit values, which means that the annual mean values are not affected by individual peaks.
Furthermore, the results indicate that PM2.5 concentrations varied at the three measurement locations. Comparing the mid-term PM2.5 immissions at the three stations in January, the concentrations (4-h averages) in the residential area were up to eight times higher than those at the urban station. The PM2.5 pollution was mostly up to 6 h above the concentration level at the urban station. Consequently, the air quality in the residential area in January is worse than near the main street, and accordingly, the health risk from respirable fine dust is also higher [1
]. Studies from Denmark and Sweden found similar results, as already mentioned in the introduction, by the comparison of rural and urban areas [15
High short-term exceedances (>25 µg/m3
) of the PM2.5 concentrations appeared more often in January than in May 2021. Temporally high concentrations are mainly measured in January in the second half of the day in the residential area and are significantly higher than at the station near traffic (urban) and the suburban station. Due to wood stove operations in January, significant amounts of particulate matter could be released. These wood stoves are mainly located in single-family houses, which are placed commonly in rural residential areas and rarely in city apartments and operated during the winter season [15
]. The use of wood as an additional or main energy source results in air pollution with particulate matter in residential areas [28
The local, very high, immissions in the residential area start mostly in the afternoon from about 2 p.m. as well as in the evening hours. During this period, most people are at home. The measured maximum values of the peaks (e.g., 144 µg/m3
on 20 December, 6:20 p.m., see Figure 14
c) resulted from small combustion furnaces. Similar results were also seen by Krecl et al. (2008) and Saffari et al. (2013), that the PM concentration by residential wood combustion is higher in the evening as well as on weekends [16
In Germany, the BImSchV should protect people as well as the environment. In the 1. BImSchV, it is specified, newly built chimneys (installation after 31 December 2021), that they must be placed close to the highest point of the roof and exceed it by at least 40 cm [19
]. However, all chimneys built previously are not affected by this new regulation, even if they do not comply with the directive. This regulation does not improve the situation shown in Figure 14
b and Figure 15
a. Figure 15
b, additionally, shows a schematic drawing of an incorrectly installed chimney and how the exhaust gas enters the recirculation zone. The people living in the immediate vicinity of such chimneys will be exposed to particulate matter and other gaseous components which are generated during the operation of wood combustion stoves that is harmful to their health over many hours on up to 11 days a month. This issue has already been noted in VDI3781 Part 4 [21
]. If all chimneys were rebuilt according to these guidelines, on many days, the pollution in the layers close to the ground would be significantly reduced. The measurements show that despite an adjusted state of the art in the guidelines, the situation in residential areas has not changed, as already built chimneys are not affected by the regulations. However, during inversion weather conditions, as seen in Figure 15
c, even a higher chimney and, therefore, the optimal discharge of exhaust gases from stoves, is not guaranteed. In this case, the exhaust gases from heating systems and other heat sources cannot be mixed due to the different temperatures of the air layers, and the pollutants concentrate in the layers close to the ground (see Figure 15
d). On these days, it would only help if no combustion took place.
The results indicate that exposure to harmful PM2.5 is partly higher in a residential area in the south of Germany than on a high-traffic street. These highest pollution levels occur mainly in the afternoon and evening when people are at home. This air pollution is not taken into account in current measurements, as the specifications do not provide for measuring points in areas that are not located directly on a main road or in the urban background [4
]. Additionally, the current guidelines do not help with already built chimneys that, due to their incorrect installation, allow the exhaust gas to flow into the recirculation zone.
The present study focused on the simultaneous measurement of PM2.5 immissions at three different spots in the region of Karlsruhe: a spot close to traffic, a suburban spot, and a location in a purely residential area in Stutensee. PM2.5 immissions were measured continuously over eight months from November 2020 to June 2021, applying fine dust sensors FDS 15 (Dr. Födisch Umweltmesstechnik AG). Two representative months, characterizing the immission situation during a month in the heating period (January 2021) and a month in spring with less heating, were analyzed concerning the PM2.5 immission in detail.
The results can be summarized as follows:
The FDS 15 PM2.5-signals correlate with those obtained by the reference systems, but underestimate the measured immission. The correlation coefficient is 0.92 on average.
The monthly average PM2.5 immissions ranged between 3.5 µg/m3 and 7.9 µg/m3 at the three spots and were below the currently valid PM2.5 annual immission limit value.
The existing annual average PM2.5 immission limit value (25 µg/m3) is inappropriate to identify and limit locally high immission events.
In a 4-h average, the PM2.5 immission in the residential area in January 2021 was up to eight times and, at the suburban station, up to six times higher than at the station close to traffic. In May 2021, the differences between the stations were smaller. Here, the measured PM2.5 immission was only up to 3.7-times higher in the residential area in comparison with the urban spot close to traffic.
The measurements show that the highest PM2.5 immission peaks occured in the residential area. Especially in January 2021, the number of peak immission events was three times higher than at the other two stations, while the situation in May 2021 was similar at the three spots.
By monitoring the close vicinity of the measuring station in the residential area, it could be shown that emissions of real-world wood stove operation clearly caused the high PM2.5 immission events.
The analysis revealed that improper exhaust discharge of wood stove emissions contributed to causing the high PM2.5 immissions in the residential area.
Visual observation proved that exhaust gases are discharged into the recirculation zone of buildings, leading to a downward exhaust gas flow.
The study demonstrates that current exhaust discharge regulations for existing wood stoves, as outlined in 1. BImSchV, are insufficient to eliminate the pollution of ambient air. The regulation needs to be updated to cover existing units.
Even tighter regulations (as in VDI 3781 Part 4) can cause high immissions, especially during inversion weather conditions or local structural conditions of building density.
In further investigations, additional residential areas should be evaluated to determine the air quality impacts of wood stove operation in these regions. Furthermore, the chemical composition of particulate matter has to be studied to quantify the amount of PM generated by wood burning. The analysis of other components, e.g., soot and gaseous components (NOx, VOC, and PAH) can also be performed.