PM10 and PM2.5 concentration changes in Poland
Neither any individual urban area’s, nor any voivodeship’s air quality complies with the WHO air quality guidelines in either annual mean PM10 or PM2.5 concentrations . Moreover, in the Cracow and Opole urban areas the PM10 concentrations were about twice as high as recommended, while the PM2.5 concentrations were about three times higher (Figure 1, Figure 2). The same situation is observed in the whole of the Małopolskie and Śląskie Voivodeships, where these urban areas are located. On the other hand, in other urban areas, for example Białystok, Tricity and Szczecin, as well as their corresponding voivodeships, the mean annual atmospheric PM10 concentration was just slightly higher than advised by the guidelines. However, the PM2.5 concentrations in these locations were still about 50% higher than recommended by the WHO (Figure 3, Figure 4).
For the most polluted regions, the data from the Chief Inspectorate of Environmental Protection reports is in line with independently collected data in other studies [16, 17]. Moreover, these and other studies  show that in these regions submicron PM1 concentrations are also elevated and that the smaller the particle, the more toxic it is .
When it comes to seasonal air pollution distribution, the WHO-recommended annual PM10 concentrations were met in June, July, and August (Figure 5, Figure 6). On the other hand, in the November-March period the WHO guidelines were greatly exceeded. Annual PM2.5 requirements were not satisfied in any month, but in June, July and August these values were only slightly exceeded. Again, a similar pattern was observed by other researchers [16–18].
Air quality in Poland follows spatial trends, as, in general, the southern regions of the country are more polluted than the northern regions . This gradual improvement in air quality may be caused by increased proximity to the Baltic Sea and the predominantly lowland character of the northern part of the country. These factors allow air to be seamlessly and continuously exchanged . On the other hand, southern regions of Poland are made up of upland, mountainous regions  and have heavily coal-based economy and energy production . Also, larger cities in the south are located in basins and valleys. Such environmental conditions reduce air movement, cause its trapping, and result in the formation of smog and decreased air quality .
A seasonal pattern of particulate matter concentration changes is also observed. Air quality is much worse in winter than in summer which is confirmed by other studies [16, 17]. This happens due to temperature inversion episodes that occur throughout the whole year, but they are more frequent and intensive in the winter season [23, 24].
Poles’ interest in air pollution
Whether people are aware of elevated pollution levels in their local area or not is an important issue. Evidence for this interest can be seen in the high mean Google Trends SVI values found in the Cracow or Opole urban areas, and in the corresponding voivodeships. Moreover, a strong linear correlation  between the annual level of air pollution in a given area, and the frequency of air-pollution related search queries is observed (Figure 1 to Figure 4). This means that the higher the mean yearly concentration of PM10 and/or PM2.5 concentration in a given area, the more frequently people are looking for their local air quality indexes, for example searching online for “air quality” or “air pollution”. It was also found that the Google Trends SVIs have higher correlations with PM concentrations for urban areas’ than for voivodeships’. Thus, it can be said that people who live in cities are more aware of the problem. Of all calculated R2 coefficients of determination the highest one was observed for the correlation between PM2.5 concentration and urban areas’ SVI. On the other hand, the weakest correlation was observed for PM2.5 concentration and voivodeship’s SVI. The correlations observed for PM10 concentration and voivodeship’s SVI as well as urban areas’ SVI are placed in-between the PM2.5-related correlations. Thus, it cannot be definitely stated whether changes in PM2.5 or PM10 concentration are more connected with changes in Poles’ air pollution-related information-seeking behaviour. Also, the Wrocław (Dolnośląskie Voivodeship) and Cracow (Małopolskie Voivodeship) datapoints are located far above the trendline. Therefore, residents in these locations search more for information on air pollution in comparison to communities in other areas with similar PM10 and PM2.5 concentrations. On the other hand, Łódź (Łódzkie Voivodeship) and Poznań (Wielkopolskie Voivodeship) residents pay less than average attention to pollution.
Even stronger correlations were found between mean monthly SVI values and seasonal air quality variations. Similar to geographical distribution, the greater the seasonal PM10 and/or PM2.5 concentration, the greater Poles’ interest in air-pollution related search queries. Thus, Poles recognize the importance of checking their local daily air quality index in general, and especially in the winter season when smog is the most harmful [23, 24]. However, there is still space for improvement, as during March and April air pollution appears to be underestimated, as evidenced by relatively lower number of searches for air quality data despite still relatively high pollution levels (Figure 5, Figure 6). On the other hand, November and December data points are situated over the trend line. Therefore, in this period Poles pay more than average attention to air quality relative to the reported levels of air pollution. This may indicate general public expectation of smog before and during winter. Nonetheless, significant pollution levels are also present during the spring season.
Comparison with other studies
Our study demonstrates a strong positive relationship between reported levels of air pollution in Poland and air pollution related information seeking behaviour. This suggests that Poles are aware of risks posed by air pollution, and undertake steps to obtain specific data, and that this awareness is in line with actual levels of the environmental hazard (both seasonally and geographically). Awareness and staying up to date are the crucial first steps in the primary prevention. It is also vital that people respond to pollution alerts and engage in protective behaviour against the threat. This includes staying indoors, purifying indoor air or avoiding outdoor activity when and where air pollution levels are higher . These behaviours are primary prevention strategies too. Studies carried out in China and the US found that people reduced their outdoor activity when atmospheric PM2.5 concentration rose [26–28]. Other research showed that individuals limited their transportation-related physical activity and spent more time at home when particulate matter concentrations increased [27, 29]. These results show that people not only search for the information on their local air quality index but use it as a guideline to adapt their behavioural patterns. To date, no such study has been carried out in Poland . Nevertheless, based on internationally available data it can be assumed that Poles who check air pollution levels are likely to respond to them and take steps to protect themselves. Nonetheless, future studies investigating this problem are necessary.
Despite growing general awareness and specific information seeking and self-protection behaviours, an increased number of hospital admissions correlates with the worse air quality in Poland [31, 32]. A similar trend is observed worldwide [33–35]. Moreover, it was demonstrated that short-term exposure to polluted air correlates with increased mortality . This includes cardiovascular and respiratory diseases’ cause-specific mortality rates [31, 32, 37, 38].
Our study as well as others on the same subject show that people are aware of air pollution and execute primary prevention strategies  to protect themselves from potential harms. However, the general public may not link air pollution with its potential to increase the prevalence and/or exacerbation of cardiovascular or respiratory diseases, and therefore they may not recognize the symptoms in an early stage of disease exacerbation. Late admission to hospital increases the probability of death which results in greater mortality rates [31, 32, 37, 38]. Thus, general public may not be aware of secondary and tertiary prevention measures, which aim to reduce the impact of a disease that has already occurred . For example, some researchers propose that there is no link between the chronic inflammatory skin diseases symptoms and air pollution, which has been shown with the use of Google Trends SVI analyses [40, 41]. Also, increased air pollution levels may exacerbate the course of ongoing chronic diseases  and affected individuals may not know that. Therefore, starting with chronically ill patients and individuals who live in the most polluted regions worldwide, there is a need for health promotion programs focused on air pollution-related secondary and tertiary prevention strategies, including accurate symptom recognition. Such interventions would result in faster responses to rapid, acute exacerbation of chronic disease symptoms and would probably reduce the cause-specific mortalities related to high air pollution levels, both in the short and long-term.
Last but not least, the greatest contributors to the problem of air quality in Poland are coal-based economy and household heating systems . In order to reduce the hospital admission and mortality rates potentially caused by air pollution as well as to comply with WHO air quality guidelines, there seems to be a very urgent need for a shift towards a more ecological energy production sources.
Due to the use of databases, the data was not directly collected by the authors, and thus this study is subject to methodological limitations. The meteorological monitoring stations that measure PM10 and PM2.5 concentrations were pre-existing and arbitrarily positioned. This may have generated bias in the data, as the concentration of particulate matter is highly dependent on measurement location. However, similar trends in air quality observations were made by other studies which used their own measurement stations [16–18]. Also, as mentioned in the Google Trends methodological framework  sometimes Google Trends SVIs may differ, even though the keywords used are the same, for example because the search query was embedded in quotation marks. To reduce this bias, all keywords analysed in this study were always entered without quotation marks, plus signs or any other special characters. Only spaces were used to separate double-word keywords, allowing for the broadest tracking possible. Lastly, Google Trends SVI shows the relative, not the absolute number of search queries. However, correlation coefficients can be calculated from either absolute or relative values, thus the calculations in this study remain valid.