The Increase in the Proportion of Impervious Surfaces and Changes in Air Temperature, Relative Humidity and Cloud Cover in Poland

The increase in air temperature in the world is a fact. Starting from the 1960s, each subsequent decade has been warmer than the previous one. The main cause of contemporary warming is human activity, mainly greenhouse gas emissions (IPCC 2021) and the emission of anthropogenic heat caused by urbanisation and land cover change (Souleymane, Quesada 2020, Wang et al. 2021). The spatial development of urbanised areas causes continuous loss of natural surfaces and an increase in the proportion of the transformed impervious surfaces (ISs) that make up cities. The presence of ISs is evidence of human activity (Arnold, Gibbons 1996). The indication of the increase in the proportion of ISs in relation to pervious surfaces (PSs) makes it possible to determine the directions of human spatial expansion (Jat et al. 2008). Artificial urban surfaces (e.g. asphalt, concrete, brick, steel, aluminium, etc.) are characterised by an increased capacity and thermal conductivity compared to natural surfaces outside the city. An increase in the temperature in urbanised areas (urban heat island) favours the intensification of evaporation, but this natural cooling process is less effective than in undeveloped areas. Owing to the high proportion of ISs that limit the infiltration of water into the substrate, there are no humidity resources that could evaporate. The complex geometry of buildings in the city makes it difficult for energy to radiate efficiently through the streets and walls of buildings, because a significant part of it is absorbed by the surrounding buildings. The spatial development of cities causes a change in the values of the basic components of the heat and water balance, which have a direct impact on the formation of cloud cover over urban areas.

The effect of intense urbanisation on increases in air temperature in the context of the occurrence of urban heat islands, both surface urban heat islands and atmospheric urban heat islands, is well documented in the scientific literature (e.g. Oke 1973, Landsberg 1981, Arnfield 2003, Akbari, Rose 2008, Kaplan et al. 2018). There are also studies on the influence of the urban area on humidity, convection and the occurrence of precipitation (e.g. Zhu et al. 2017). The interrelationships between cloud cover and urbanisation, on the other hand, have been poorly researched, and the results are questionable. According to some authors, the city causes an increase in cloud cover (e.g. Abakumova et al. 1996), while, according to others, the agglomeration contributes to its decrease (e.g. Kuczmarski 1982). Kossowska-Cezak (1978) believes that the urban complex affects cloud cover in mechanical and thermal ways and that it usually increases cloud cover over this area, but under certain conditions, it may contribute to the reduction of cloud cover (in the evening, with generally high cloudiness). The explanation of this issue seems to be of particular importance in terms of the effects of contemporary climate change, i.e. heat, drought and flash floods, which particularly affect city dwellers.

The aim of the study is to analyse the characteristics of changes in the share of ISs in Poland and their impact on changes in air temperature, humidity and cloudiness. ISs include all areas that cannot be seeped into by water, i.e. man-made artificial areas identified on the basis of the analysis of satellite images. Low-level clouds are considered to be the visual, comprehensive indicator of changes taking place in the atmosphere, caused by thermodynamic processes related to the change in albedo and the latent heat of the active surface. They are most dependent on local conditions and have the strongest impact on the air temperature at the Earth's surface. By absorbing short-wave solar radiation and the Earth's long-wave radiation, clouds retain heat and cause an increase in temperature, which was called cloud greenhouse forcing.

This study is part of the current research on urban climatology explaining the causes of climate change, which uses modern methods to determine the magnitude of the changes in the Earth's surface transformed by man. The analysis of the increase in the proportion of ISs in Poland and their relationship with selected elements of the climate may help understand the mechanism of anthropogenic causes of contemporary warming.

The study uses meteorological data from ground stations and the results of satellite image classification containing information on the imperviousness index for the years 1991–2020. Only the data from the warm half of the year (April–September) were taken into account, because IS extraction is characterised by the highest accuracy in the period of increased vegetation, defined for instance based on the normalised difference vegetation index (NDVI) (Soltani, Sharifi 2017, Sun et al. 2017). Moreover, the highest air temperatures in Poland occur precisely at this time of the year (Ustrnul et al. 2021). Data from 36 weather stations (Fig. 1) were used to characterise the spatial diversity of selected climate elements for the whole of Poland, while the data from weather stations located in 10 large cities in Poland were selected for a detailed analysis of the relationship between the proportion of ISs and the maximum air temperature, relative humidity, and low-level cloudiness (Table 1); the analysis also included the small town of Lesko, which constituted the background for comparative research. The stations from which meteorological data (i.e. daily values of maximum air temperature, relative humidity and low-level cloudiness) were collected belong to the state meteorological service and operate as part of the Polish Institute of Meteorology and Water Management – National Research Institute (IMWM-NRI). The research included stations that met the criterion of continuity of the data series, without breaks or with slight gaps that could be supplemented on the basis of data from neighbouring stations.

Fig. 1

Landsat 5–8 satellite imagery provided by the United States Geological Survey (USGS) was used to determine the changes in the proportion of ISs in Poland. In the research, analysis was performed for three moments in time, which facilitated the comparison of changes in the imperviousness index in 3 decades, i.e. the years 1991–2000, 2001–2010 and 2011–2020. For the same periods, the trends of maximum air temperature, relative humidity and the degree of sky coverage by low-level clouds were calculated for the aforementioned cities in Poland.

In the case of satellite data, they were characterised by the same spatial resolution of 30 m for red, green and blue (RGB) bands, as well as near-infrared (NIR) and short-wave IR (SWIR) bands. For each time period, several satellite images from the warm half of the year (April–September) were acquired. The data obtained from the Copernicus programme were also helpful in the analysis. The resources of High Resolution Layers Imperviousness were used to generate the classification pattern of the Support Vector Machine (SVM) algorithm (Langanke 2016).

The analysis of satellite images was carried out in several stages. The first was data pre-processing, in which the areas covered with clouds were removed from the analysis. The next step was to normalise the data by processing images into 16-bit rasters. Performing this operation made it possible to reduce the deviations resulting from the penetration of the atmosphere by electromagnetic radiation (Ban 2016). The last element of pre-processing was the preparation of four-band composites (three bands of RGB + NIR), which—in the next stage—were subjected to the operation of the SVM algorithm.

The second stage of the IS extraction was to calculate the spectral indices and to exclude, on their basis, areas that can be considered pervious with high probability. First, the NDVI (Friedl et al. 1995) was calculated, as a result of which biologically active areas were excluded. In the next step, the normalised difference water index (NDWI) was calculated and thresholded. This index makes it possible to extract and eliminate water features from the analysis (Gao 1996). The last-calculated index was the dry bare-soil index (DBSI), which made it possible to exclude large areas of exposed soil from the analysis (Rasul et al. 2018). For the classes of the areas excluded from further analyses, i.e. biologically active areas, water features and exposed soils, the classification accuracy was 100%, i.e. each pixel classified as one of the above classes was correctly marked.

The third stage of satellite image analysis was to separate PSs from ISs. Several classification algorithms were tested in the study. The best results were obtained for the SVM, which made it possible to determine the following classes: 1) ISs, 2) arable land, 3) meadows and other vegetation and 4) exposed soils. Classes such as forests and water features were successfully eliminated at an earlier stage of the analysis. The designated areas were aggregated into two main classes to form the final IS/PS map. The obtained IS/PS map was subjected to further analysis. First, the data were aggregated to the administrative borders of the cities. The purpose of this operation was to calculate the area and percentage of IS in each administrative unit, thus obtaining the value of the total imperviousness index of a given urban area.

In the last stage of the analysis, the results presenting the proportion of ISs for three time periods were compared, which made it possible to prepare maps of changes in the value of the area's imperviousness index in individual decades.

High-resolution digital orthophotomaps were used as reference data for the analysis of the accuracy of the IS classification made. In the case of absence of data in the GUGiK's resources, images published via the Google Earth service were used.

There is no doubt that the development of civilisation has changed the coverage of the Earth's surface. According to Hurtt et al. (2006), in the years 1700–2000, as a result of human activity, from 42% to 68% of the Earth's surface was transformed. The strongest changes occurred in the moderate latitudes of the Northern Hemisphere. They caused the modification of the radiation balance by affecting the albedo changes and the latent heat flux. They also affected the increase in air temperature, the water cycle, river runoff, surface roughness, dust emission, snow cover and more. The phenomenon of taking up soil for housing development, production of plants, public transport networks, etc. and covering it with impervious material is described in detail in the reports of the European Commission (Prokop et al. 2011). The results of the studies by Gardi et al. (2014) suggest that sealing the substrate is almost irreversible. In already heavily urbanised European countries, such as the Netherlands and Germany, the rate of loss of natural surfaces is high, and there is little room for further urbanisation. In Central and Eastern Europe, the accelerated growth of built-up areas was a consequence of the political and economic changes of the late 1980s, and the expansion of the European Union and the integration of new countries into the common market led to the development of transport infrastructure, urban expansion and further sealing of the surface (Gardi et al. 2014).

The processes of taking up land in other parts of the world are much faster than in Europe, especially in emerging economies. For example, 5.1% of all Chinese territory was converted into land for manufacturing and communal activity between 1996 and 2003 (Chen 2007), and in the Beijing–Tianjin–Hebei region, urban area increased by 71% between 1990 and 2000 (Tan et al. 2005). Similar growth rates were also recorded in India and other rapidly developing countries (Fazal 2000). Worldwide, urban areas are increasing their surface area at twice the rate of their populations (Angel et al. 2011).

In Poland, as in many other countries, intensive urbanisation is taking place and the proportion of ISs is increasing. In 2020, the ISs in this country covered 9954 km², which constituted 3.2% of the area. In the years 1991–2020, the total area of ISs increased by approximately 30%. Such a large change in the coverage of the earth's surface in a relatively short time (30 years) and the progressive warming have motivated us to undertake research on the impact of the increase in the share of ISs on thermal conditions, humidity and cloud cover in Poland. Our research showed that in the warm half of the year, in all of the 10 analysed cities, the increase in ISs in recent decades coincided with the decrease in low-level cloudiness. On a local scale, both of these phenomena should strengthen the effect of the albedo decrease and contribute to the heating of the active surface and, thus, to an increase in temperature and a decrease in air humidity. Indeed, the highest rates of air temperature increase and relative air humidity decrease occurred in the years 2001–2010, when the highest IS increase was recorded in most cities. Another important observation is that the significant decrease in low-level cloudiness was also accompanied by a change in the cloud structure. The frequency of low- and mid-level stratiform clouds (Stratus, Nimbostratus, and Altostratus) decreased, while the frequency of high-level Altocumulus and Cirrus clouds increased. The decrease in the frequency of stratiform clouds may have been affected by processes taking place in recent years on a large spatial scale, i.e. the intensification of air dryness caused by an increase in air temperature in the entire vertical profile of the troposphere over Poland. However, the higher frequency of Altocumulus and Cirrus clouds could be related to the reduction of low-level cloud cover, which made it possible to see the clouds in the middle and upper troposphere.

This study is, therefore, part of the research on the causes of modern climate warming. According to many scientists (including Tett 2002, Meehl et al. 2004, Lean, Rind 2008, Gillett et al. 2012, Huber, Knutti 2012, Jones et al. 2013, Wigley, Santer 2013), natural factors on a global scale have not had a significant impact on the increase in air temperature in the past few decades and could even cause its decrease (e.g. the weakening of the Sun's activity, volcanic eruptions). Furthermore, the results of the research conducted in Poland (including Kulesza 2020, Bartoszek et al. 2021, Bartoszek, Matuszko 2021) indicate that radiation and circulation factors do not play a major role in the recorded increase in air temperature in this country. On the other hand, there are many arguments that the climate is warming as a result of human activity caused by the increase in greenhouse gas concentrations and the change in the Earth's surface cover as a result of urbanisation, soil transformation, deforestation and changes in the Earth's albedo (IPCC 2021).

Research by Stephens et al. (2015) and Goode et al. (2021) confirmed a decrease in the albedo of the Earth's surface over the past 20 years, which means that our planet absorbs much more solar radiation than it did before the beginning of the 21st century. According to Goode et al. (2021), the main reason for the decline in the Earth's solar reflectivity may be the lower amount of low-level clouds associated with the warming of the Eastern Pacific. The results of modelling studies on the hydrometeorological effects of urbanisation remain ambiguous (Lowry 1998, Shepherd 2005, Mahmood et al. 2014). The major challenges limiting our understanding of urban modification of convection is the difficulty in separating the urban signature from those of topography, land–sea or land–lake boundaries, and the large-scale atmospheric forcing in field experiments or in real-case modelling studies. As a result, studies conducted at different locations, in different seasons, and/or under different synoptic conditions often produce different findings (Zhu et al. 2017).

Unfortunately, the climatological literature lacks current studies on the impact of urbanised areas on the long-term variability of cloud cover and the occurrence of individual types of clouds. The problem of IS's influence on cloud cover was discussed only in the study by Theeuwes et al. (2019), which highlights the increase in cloud cover in the afternoon and evening hours over urban areas in Europe, i.e. in London and Paris. Owing to the still small number of scientific publications in this field, the results of our research constitute an important contribution to the discussion on the consequences of the increase in ISs in cities and contribute to further analyses on a larger spatial scale.