The Landscape Fragmentation: Analysis of Land Cover Transformation in High Mountains Environment on the Example of Tatra Region (Southern Poland)

The research area is the Tatra and its northern foreland in the article, for easier understanding, referred to as Tatra region, which is located in the southernmost part of the Małopolskie Voivodeship in Poland (Fig. 1). The area of the Tatra region is 471.6 km², and the population is 67,719 people (Bank Danych Lokalnych 2020). The largest city and the only urban municipality in the region is Zakopane. In terms of the region’s geographic location, it is divided into six sections: Wysokie Tatry, Tatry Zachodnie, Tatry Reglowe, Bruzda Podtatrzańska, Magura Spiska and Pogórze Przedtatrzańskie (Solon et al. 2018).

Fig. 1.

With the inclusion of five different mesoregions, there is a great variety of terrain relief, which is a factor that makes the area exceptionally attractive for tourists and settlement. The region of Wysokie Tatry (High Tatras) in particular is characterised by unique natural values, and it is one of the best examples of glacial relief in the entire country. The region’s landscape has several glacial landforms, like U-shaped valleys, middle, side and bottom moraines, and glacier cirques. Most of the lakes are tarns located in the cavities of cirques (Klimaszewski 1988). The biggest and most recognisable ones are Morskie Oko and Wielki Staw, both located in the High Tatras.

In the southernmost point of the area, there is the highest mountain peak in Poland, Rysy (2499 m a.s.l.). The lowest point of the area is the valley of Dunajec by Nowy Targ, where the absolute height is 641 m a.s.l. As a result, the study area is subject to altitudinal zonation of its climate and flora (Hess 1965). This is the most important condition shaping LC in the Tatra region. It creates six vegetation layers: the foothills zone, lower montane zone, higher montane zone, subalpine zone, alpine zone and crag zone (Fabijanowski 1962). The research area is located in the mixed temperate climate zone but belongs to the alpine climatic region (Woś 2008). A characteristic feature of the regional climate is the so-called halny, which is a type of foehn wind. It is a dry and gusty wind that causes windfall and damage to forests.

The outstanding natural beauty of the High Tatras and their importance for the whole ecosystem resulted in the creation of the Tatra NP in 1955. Its area is 21,164 ha. Both the Tatra NP and the neighbouring Slovak national park, TANAP, have had the status of international UNESCO biosphere reserves since 1993 (UNESCO 2022).

The primary data source is the CLC database, which is a part of the European programme Coordination of Information on the Environment (CORINE), which is led by the European Environment Agency. It was mainly created through the use of a semi-automatic method of visual interpretation of high-quality satellite images (Landsat, RapidEye, IRS, SPOT-4). The CLC database is available for the years 1990, 2000, 2006, 2012 and 2018. Currently, the CLC programme covers most European countries, including Poland. The data used in the study was downloaded through the website of the Inspectorate of Environmental Protection (pol. Inspektorat Ochrony Środowiska) (CORINE Land Cover 2022).

As auxiliary research material, the authors used the current orthophoto map obtained from the National Geoportal (geoportal.gov.pl). It allowed the results to be verified, thus facilitating a more reliable analysis. Apart from the above, the study used plans and sketches made during the field study, as well as acquired photographic documentation.

The software used for the spatial analyses is the licensed geoinformation programme ArcGIS v.10.7.1. Landscape fragmentation metrics were calculated using a free V-late plug-in (Vectorbased Landscape Analysis Tools, v.2.0) for ArcGIS software. Calculations of the indicators of landscape changes were done in Microsoft Excel software.

For the purpose of analysing LCC and landscape fragmentation, the research procedure was divided into three stages (Fig. 2). The spatial phenomena were quantified with the use of landscape metrics based on measurements of the homogeneous elements of the landscape (patches).

The first stage of the research was a statistical analysis of the increases and decreases of the identified LC forms in the entire period of study covering the years 2000–2018. For this purpose, the authors used the index of landscape change (ICh), which expresses the percentage value of changes in a given LC form (Vavrouchová, Toman 2013). The greater it is, the greater the change of a given LC form is. A value of 0 means that no change occurred. The index was calculated for each of the LC forms, for the entire period of 2000–2018, as well as for the three sub-periods: 2000–2006, 2006–2012 and 2012–2018. The index of landscape change is expressed using the following formula: ICh(XY)=∑i=1n|rax−ray|2c×100$$IC{h_{(XY)}} = {{\sum\nolimits_{i = 1}^n {|{r_{ax}} - {r_{ay}}|} } \over {2c}} \times 100$$

Fig. 2.

where: –

ICh – Index of landscape change (%); for the same LC form in the period between years X and Y,

In the second part of the research, the authors analysed the spatial structure of LCC. For this purpose, a cartogram presenting the percentage share of LCC was created for each sub-period. The class intervals for the cartogram were based on relative data (area of changes in ha km^-2) and a square was chosen as the base field. Thanks to cartographic representations of the phenomena, it was possible to indicate where the most changes in the form of LC took place.

The third research stage was dedicated to the analysis of landscape fragmentation in the years 2000–2018. To do this, the authors computed a set of six simple landscape fragmentation metrics. The metrics, along with their mathematical formulas and abbreviated forms used in the following chapters, are presented in Table 1. Apart from the metrics-based analysis, this stage was extended using a graphical representation of the Landscape Division Index (LDI) for each of the LC forms in the study area. The results obtained from the analysis and the available published works will allow the last phase of research to formulate regularities about the causes of the ongoing fragmentation of the landscape.

When comparing the values of the landscape change index of individual LC forms, it can be concluded that both urban and agricultural areas, as well as forest and semi-natural areas, have been subject to significant changes. It is possible to observe a wide range in the value of this indicator (Table 2). The overall ICh was 13.16%, which points to high dynamics of LCC in the research area.

In the category of urban areas (CLC level I), each of the LC classes has increased its share of the area. One of the strongest increases in the share of the area occurred in urban fabric (112). An increase was observed in the case of sports facilities (142). Over time, green urban areas (141) have appeared, but the share of the area occupied by this LC form remains insignificant.

In the category of agricultural areas (CLC level II), an increase was only observed in the case of arable land (211). On the other hand, there was a notable shrinkage in pastures (231). A more significant decrease occurred in the case of complex cultivation (242) and land principally occupied by agriculture (243).

When it comes to the category of forest and semi-natural areas (CLC level III), a slight increase in its share was observed in broad-leaved forests (311) and grasslands (321), as well as sparsely vegetated areas (333). The class which increased its area the most is the transitional woodland shrub (324). It has enlarged its share several times (ICh = 2.61%). However, this goes together with a decline in the area occupied by coniferous forests (312) and rocks (332). Shrinking of mixed forests (313) and moors (322) was also observed, but to a smaller extent.

In terms of LC dynamics in the three sub-periods, the results point to a significant disparity in the values of overall ICh. In the first sub-period, the changes remained insignificant (ICh_06–12 = 1.38%), while in the second and third sub-period, they were much greater (ICh_12–18 = 30.19%; ICh_12–18 = 25.81%). Thus, LCC are becoming more and more rapid in the Tatra region. However, this is mainly caused by extraordinarily high dynamics in only three categories: arable land (211), pastures (231) and woodland shrub (324).

The elaborated cartograms presenting the spatial distribution of changes confirm the tendency described in the previous sub-chapter (Fig. 4), and the graph shows the ICh values for the sub-periods studied (Fig. 3).

Fig. 3.

Fig. 4.

In the area of the Tatra National Park, although the landscape is changing, they are twice as weak as in the surrounding area. This regularity applies to both the second and third periods of analysis. This means that the high value of the indicator for the entire research area results mainly from changes taking place in the surroundings of the national park, not the park itself.

In the first period, the total area of LCC is relatively small, amounting to 622.45 ha (Fig. 4), and it is mainly concentrated in the central part of the area, in Zakopane and Poronin. Some changes also occurred in the Tatra NP. In the eastern and northern parts, the LCC are negligible, and in most of the areas, there are no changes recorded.

In the next sub-period, the changes amount to 8214.52 ha, so they are several times larger. In this case, LCC are concentrated mainly in the north-western part of the area, in the municipalities of Biały Dunajec and Poronin (Fig. 4). Referring the figures of the LC structure, it is possible to observe that some significant changes (1961 ha) are due to the transformation of pastures into arable land.

The third period (2012–2018) is also characterised by a very large total area of LCC (10,417.52 ha). The most important similarity to the previous period is the concentration of changes in the north-western part of the area and reoccurring transformation of arable land into pastures (Fig. 4). However, unlike the previous period, in the years 2012–2018, there were also radical landscape transformations within the Tatra NP caused by the transformation of coniferous forests into woodland shrubs.

The major finding in the analysis of landscape fragmentation was a significant increase in the number of patches, 259 in 2000 and 268 in 2018. On a landscape level, the calculation of Shannon Diversity Index (SDI) indicated a growing diversity of LC forms in the district’s landscape over the years (from 1.869 to 2.070), which was caused by the appearance of green urban areas, sports facilities and water courses. Simultaneously, the change in Shannon Evenness Index (SEI) (from 0.690 to 0.731) showed that relative proportions of the various LC forms have become slightly more even.

On a class level, the analysis of Mean Patch Edge (MPE), Mean Shape Index (MSI) and DIVISION indexes draws attention to their high dynamics in several classes. In terms of the MPE values, the LC forms that increased their MPE are urban fabric (112), moors (322), grassland (321), woodland shrub (324), rock (332), sparsely veg. (333) and water courses (511, Table 3). At the same time, three forms of LC decreased their MPE in 2000–2018: arable land (211), coniferous forests (312) and mixed forests (313). There are also many LC forms that are not very dynamic: green urban areas (141), pastures (231), complex cultivation (242), land principally occupied by agriculture (243) and broad-leaved forests (311).

Analysing the MSI, the decreasing values of which indicate the simplification of geometry, there are also five forms of LC with a downward tendency: urban fabric (112), arable land (211), complex cultivation (242), grassland (321) and sparsely veg. (333, Table 4). There are seven LC forms with the opposite tendency: sports facilities (142), pastures (231), coniferous forests (312), mixed forests (313), moors (322), woodland shrubs (324) and rock (332). Their areas are getting more and more complex and fragmented. These are mostly forest and semi-natural areas, which are mainly present in the national park.

The most spectacular results concerned the changes in the LDI in the different LC forms (Table 5). First of all, the LC categories can be divided into three groups. The first are those characterised by constantly low fragmentation: green urban areas (141) and sports facilities (142). The second is those with constantly high fragmentation: pastures (231), complex cultivation (242), land principally occupied by agriculture (243) and woodland shrubs (324). There are also those whose degree of fragmentation has changed significantly: urban fabric (112), arable land (211), broad-leaved forests (311), coniferous forests (312), mixed forests (313), grassland (321), rock (332) and sparsely veg. (333). The spatiotemporal differentiation of LDI has been visualised in Figure 5. It is significant that agricultural LC forms, like arable land (211), pastures (231), complex cultivation (242) and land principally occupied by agriculture (243), had LDI values that were clearly higher than others. Most of them had very slight fluctuations in LDI in the following years. The complete opposite are those LC forms belonging to the category of forests and semi-natural ecosystems. Their fluctuations between 2000 and 2018 reached as much as several dozen per cent in both directions.

Fig. 5.

When analysing the cartograms showing the LDI for each of the studied years, completely different directions can be observed in the northern and southern parts. First of all, the southern part of the Tatra region, located within the park and dominated by forests, has much lower LDI values than the northern part for the first analysed years (2000; 2006). This regularity is not so distinct in the case of the last two analysed years (2012; 2018). Moreover, during this time, the fragmentation of this part increased dramatically. This is related to the influx of new LC forms into previously homogeneous areas, especially in the Tatra NP.

The northern part of the area is characterised by completely opposite values, and it is dominated by highly fragmented agricultural areas. The year 2012 stands out from the rest of the analysed years. Referring to the CLC maps, a large area in the northwest was briefly reclassified from complex cultivation (242) to arable land (211), and consequently, the LDI decreased temporarily. In the last year, the share of land with an average LDI value also increased in the north. Taking this into account, it can be concluded that the degree of LDI contrast between the areas in the south (protected) and the north (urbanised) decreased over the years 2000–2018, with a simultaneous increase in the share of areas with high LDI values.

The latest research on landscape transformation has indicated in recent years that the speed of these changes is accelerating. They are being carried out at global as well as national and even regional scales. Such studies have also been conducted in Poland, in various regions of the country, both in rural areas (Krajewski et al. 2017a, Krajewski 2019) and in metropolitan ones (Pukowiec-Kurda, Vavrouchova 2020). Landscape changes in these areas were in the range of a few to several per cent of the ICh. This means that a percentage of the area changed from one type of cover to another over the time period studied. The results of this index obtained in the Tatra region indicate an extremely high degree of spatial change, which is several times higher than in other regions of Poland (Łowicki 2008, Krajewski 2019, Pukowiec-Kurda, Vavrouchová 2020). This indicates the existence of large causal forces responsible for land transformation. It also indicates the dynamics of these transformations. In the first studied period, 2000–2006, the rate of change was small. However, in the second period, 2006–2012, it underwent dynamic growth. This means that during this period, one-third of the area of the entire region changed from one type of LC to another. Comparing the results of this study with those of other authors (Krajewski et al. 2018), one can speak of the uniqueness of this area in terms of the strength and speed of landscape transformation. Moreover, the considerable variation in relief, and thus in climatic conditions, vegetation types, etc., in the study area is the natural basis for the functioning of different landscape processes. Thus, the same phenomenon operates with different dynamics in the foothills and high mountains, for example, overgrowth of fields in the foothills and natural succession of the forests after windthrow in the Tatras.

The reasons for these transformations can be traced back to a variety of sources. The first is the socio-economic and political transformation in Central Europe in the period after 2004 (Deslatte et al. 2022). The period studied, 2006–2012, is characterised by the highest transformation rate in any area studied in Poland. This period follows shortly after the accession of Poland and neighbouring Central European countries to the European Union (Bucała-Hrabia 2018). During this period, there was an economic acceleration, with many investments in both housing and services, the development of road networks and the creation of new industrial and technological spaces. It was also a period of increased work in the field of nature protection, e.g. the establishment of Natura 2000 areas. All this has influenced changes in LC. As a result of increasing urbanisation, new residential areas have been created, especially on the urban periphery, road density has increased, and shopping centres, service and storage areas have been created. These human-influenced transformations have also occurred in Zakopane and neighbouring tourist destinations. These changes have been particularly evident in this region due to the investment and rapid development of tourism. Large hotel complexes, restaurants, ski resorts and additional tourist attractions such as aquaparks have been built (Fidelus-Orzechowska et al. 2021), as well as second houses (Mika, Faracik 2008). At the same time, the local population began to change their lifestyle and shift their work from farming to tourism. This is evident in the conversion of agricultural land (arable fields, pastures, etc.) into wasteland or development for tourism and residential purposes. The role of tourism in the contemporary transformation of space in the Tatra Mountains is also highlighted by Rączkowska (2021). In turn, Bucała-Hrabia (2018) writes about the abandonment of agriculture in the Polish Carpathians after the fall of communism in 1989, which is evidence of the prevalence of this phenomenon in the study area.

The rapid changes in the Tatra region can also be attributed to the attractiveness of its location. It attracts those investing in the tourist industry on the one hand, and people wanting to live in beautiful surroundings in the neighbourhood of protected areas on the other (Zbierska 2022). Hence, the rapid development and building up of plots of land. The district’s attractiveness comes from its geographical location in a valley at the foot of the highest mountains in Central Europe, the closeness of the Tatra National Park, the possibility of living with a beautiful view of the Tatras, the clean environment and the close proximity of good tourist infrastructure, especially for skiing.

The above-described causes are mainly manmade. However, a large part of the region is under strict legal protection in the national park area. By analysing the transformation maps, it can be seen that these areas have also been subject to rapid transformation (Zbierska 2022). It is known that the cause of such large changes in the nature protection area could not be manmade. In the study period, 2000–2018, several extreme windfall events occurred in the Tatra area (Falt’an et al. 2020). Significant forested areas were devastated, transforming spruce forests into treeless areas. This change was recorded in the LC database and detected in the applied landscape change index. The force of this windthrow was so great that it is possible to speak of significant land transformation in the Tatra NP (Bodziarczyk et al. 2019). Moreover, in addition to the windthrow, there were other natural forces that also affected land use in the Tatras. These included debris flow or snow avalanches, as pointed out by Spyt et al. (2020) in their study. Thus, the second cause of change is the forces of nature. It is interesting to note that both factors responsible for the transformation of the landscape—humans and nature—interacted with great force at the same time. This combination of these two forces has resulted in spectacularly huge changes in the study area over the analysed time period.

Rapid changes in LC were reflected in the occurrence of landscape fragmentation in the studied area. Its appearance was influenced by another factor, which is the lack of spatial planning and chaotic land development (Duarte et al. 2018). At the investment planning stage, no analyses of spatial design were carried out. New buildings were built chaotically, sometimes on random plots of land (Ćwik, Hrehorowich-Gaber 2021). This resulted in the fragmentation of landscape patches (LC patches) revealed in the research. The area of individual patches decreased, while their number and the length of their borders increased. In the case of the action of the second driving force—nature—it is difficult to talk about planning. Due to their nature, extreme events produce chaotic and unplanned consequences. Therefore, the fragmentation phenomenon also applies to the area of the national park (Kubacka et al. 2022) with the area of the forest felled by winds. The shapes of the windfall areas are also irregular; they occur in several places and are random and incidental in space.

The phenomenon of landscape fragmentation is common and characteristic in urbanised areas, as well as being subject to mainly human influence and its constant investment (Jaeger et al. 2016). In the case of the development of agricultural areas or naturally valuable areas, e.g. forests, meadows and wetlands, we can talk about the loss of ecosystems or at least disturbances in their existence. The development of such types of cover also involves the loss of the ecosystem services they provide (Zambrano et al. 2019). Another effect of anthropogenic changes is the creation of new barriers in space, e.g. in the form of fences, transmission lines, roads and large concrete areas. These are often an insurmountable barrier for animals. In this case, ecological corridors can also be threatened, both as a result of the fragmentation of the landscape (fragmentation, separation of patches into smaller ones) and the formation of barriers. All these ecological effects are detrimental to nature in urbanised areas (Atasoy 2018). However, their effects are much more visible in naturally valuable areas. The fall of tree cover disrupts the operation of the forest ecosystem. It breaks the continuity of the ecological corridor that the forest offers or causes perforations in it. As a result of windfall, valuable forest ecosystems, along with all their functions, such as water retention, slope stabilisation and habitat for plants and animals, are lost (Jaworek-Jakubska et al. 2020). This is particularly significant in the Tatra NP area since it is inhabited by rare and protected species, such as the brown bear. Once again, the action of a double changing force, in this case, fragmenting the landscape of the Tatra region, is revealed—man and nature. Man, mainly changes the areas of the city and its surroundings, and the forces of nature destroy the areas of the national park.

Landscape metrics have been used for years to determine various landscape parameters (McGarigal, Marks 1995, Urbański 2008). Some of the most common are analyses of landscape change, including landscape fragmentation, both using the landscape change index and other more sophisticated mathematical and statistical tools. One of the most substantial advantages of using landscape metrics is that the landscape can be monitored by determining a set of indices in successive years (Peano, Cassatella 2011). They give measurable and comparable results revealing changes in the landscape. Furthermore, using statistical models, predictions of future landscape changes can be made. Measurability, ease of use and universality are among the greatest advantages of landscape indices.

Another remarkable advantage of using landscape indicators as tools to determine its changes is their universality. They are used by researchers from many disciplines on examples from different regions (Cassatella, Peano 2011, Li et al. 2021). Therefore, it is possible to compare results, even on a global scale. Reference can be made to indicator values obtained in geographically similar areas and areas that are quite different.

However, the use of landscape indicators, like any method based on the use of a cartographic database, has its limitations. Some of these relate to the materials used and others to the metrics themselves. In order to avoid errors of interpretation, the set of indicators should be appropriately selected for the purpose and the scale of the study. Depending on the degree of precision of the research, different metrics can be used as they relate to different degrees of detail, e.g. some only to patches of a landscape, others to the entire landscape (Inkoom et al. 2018). The results of indicators are often dependent on the size of the study area and its landscape diversity. The best results are given by large-scale analyses at a smaller scale, e.g. regional, than by small-scale analyses, e.g. within a municipality or parcels. Limitations relating to the materials used most often relate to their validity and the degree of accuracy, generalisation and the way in which the satellite image was processed into a vector model (Śleszyński et al. 2020). In the latter, interpretation errors often occur, resulting, for example, from automation of the process. One of these is the misreading of satellite images and the incorrect classification of areas. Such an error is unfortunately present in the CLC database, which is revealed in the classification of the same area into two different types in successive analysed years. Nevertheless, the database is still useful due to its undergoing regular updates, its coverage of the entire study area and its universality and accessibility for the whole of Europe (Cieślak et al. 2020).