Acta geographica Slovenica, 61-1, 2021, 57–74 SPATIAL DIVERSITY OF ECOLOGICAL STABILITY IN DIFFERENT TYPES OF SPATIAL UNITS: CASE STUDY OF POLAND Jolanta Jóźwik, Dorota Dymek Spring landscape of Roztocze, Poland. J O L A N T A J Ó Ź W IK 61-1_acta49-1.qxd 28.7.2021 8:06 Page 57 Jolanta Jóźwik, Dorota Dymek, Spatial diversity of ecological stability in different types of spatial units: Case study of Poland DOI: https://doi.org/10.3986/AGS.8779 UDC: 913(438):502.131.1 COBISS: 1.01 Jolanta Jóźwik 1 , Dorota Dymek 1 Spatial diversity of ecological stability in different types of spatial units: Case study of Poland ABSTRACT: The study estimates and compares the spatial distribution of ecological stability within admin- istrative units in Poland. Its method permitted the value of the coefficient of ecological importance parameter to be determined, and enabled the design of a spatial unit typology. The units originally analyzed were municipalities (Pol. gminy). In this variant, areas with low and average ecological stability were evident- ly dominant. Verifying the results obtained involved extending the study, and using of a square with sides of 1 km as the basic unit of assessment. This approach yielded dominance of areas extreme in terms of eco- logical stability. The spatial analyses also allowed for the spatial dependence of the phenomenon to be identified and illustrated spatially. KEYWORDS: Coefficient of Ecological Importance, spatial autocorrelation, spatial planning, land cover, landscape, Poland Prostorska raznolikost ekološkega ravnovesja v različnih tipih prostorskih enot: primer Poljske POVZETEK: V raziskavi avtorici proučujeta in primerjata prostorsko porazdelitev ekološkega ravnovesja v upravnih enotah na Poljskem. Z izbrano metodologijo sta določili vrednost koeficienta ekološkega pome- na in izdelali tipologijo prostorskih enot. Osnovna prostorska enota, ki sta jo najprej analizirali, je bila občina (pol. gminy). Rezultati so razkrili, da v njih prevladujejo območja z nizkim in povprečnim ekološkim rav- novesjem. Da bi avtorici preverili dobljene rezultate, sta raziskavo razširili in za osnovno enoto tokrat uporabili kvadrat s stranico 1 km, za katero so rezultati pokazali prevlado območij z ekstremnimi vrednostmi ekolo- škega ravnovesja. S prostorskimi analizami sta avtorici lahko določili tudi prostorsko odvisnost proučevanega pojava in jo prikazali v prostoru. KLJUČNE BESEDE: koeficient ekološkega pomena, prostorska avtokorelacija, prostorsko načrtovanje, pokrovnost tal, pokrajina, Poljska The paper was submitted for publication on June 30 th , 2020. Uredništvo je prejelo prispevek 30. junija 2020. 58 1 Maria Curie-Skłodowska University in Lublin, Lublin, Poland jolanta.jozwik@umcs.pl (https://orcid.org/0000-0001-7041-3781) dorota.dymek@umcs.pl (https://orcid.org/0000-0002-8902-9373) 61-1_acta49-1.qxd 28.7.2021 8:06 Page 58 1 Introduction The cultural landscape is constantly evolving to meet the ever-changing needs of present and future gen- erations. One of the main factors currently influencing significant changes in the landscape structure is human activity (Verburg et al. 1999; Verburg et al. 2002; Dotterweich 2008; Geri et al. 2010; Baran-Zgłobicka and Zgłobicki 2012; Ribeiro and Šmid Hribar 2019). New anthropogenic elements introduced into the nat- ural environment contribute to landscape transformation. Their intensity has an impact on the ecological stability of the landscape. Considerable accumulation of such elements may lead to gradual degradation of the natural environment and disturbances in the area’s ecological stability (Richling and Solon 1994; Król and Gałaś 2008). Ecological stability is defined as the ecosystem’s ability to return to equilibrium, or to its »normal« direc- tion of development, via its own internal mechanisms. The sooner the ecosystem returns to its original balance, the more stable it is (Holling 1973; Vološčuk and Míchal 1991). Forman and Godron (1986) define landscape stability as the landscape’s resistance to disturbances and its ability to regenerate after they occur. Bičik et al. (2015, 9) define it as »a condition that is inversely related to the amount of energy, material, and labor invested by the society so that the landscape remains in a balanced condition.« Over time, land- scapes and ecosystems undergo natural transformations (Widacki 1979). As a result, the forms, functions, and significance of landscapes also change (Urbanc et al. 2004). Therefore, the stability of the natural envi- ronment is dynamic. Considering all of this, when a disruptive factor is introduced, the natural environment is not able to return to its exact original state, even though it can achieve an approximation of it (Balon 2006). Zaušková and Midriak (2007) also point to the dynamic ability of ecosystems to maintain and restore the conditions of their existence through self-regulatory mechanisms. This is reflected in their stability and resistance to natural and anthropogenic factors. Two main trends are designated in landscape stability research (Balon 2006; Gigon and Grimm 2014). The first refers to natural areas capable of functioning owing to internal mechanisms, without human inter- vention: the »natural« approach (e.g., Gigon 1983; Geng et al. 2019). The second refers to stability assuming the presence of anthropogenic activities and economic uses of the natural environment: the »utilitarian« approach (e.g., Messerli 1983; Winiger 1983; Fuentes 1984; Zhang et al. 2017). a mixed approach also exists that combines both of these (e.g., Kienholz et al. 1984; Ganjurjav et al. 2019). Pinpointing and evaluating the ecological stability of a landscape is a complex process, in which the level of ecological stability of a given area may reflect a coefficient of ecological importance. It can be expressed numerically, whereas the result is only an approximation of the reality behind the model (Bastian and Schreiber 1999). Mandatory large-scale landscape studies of this type have been conducted in the Czech Republic and Slovakia as part of their Territorial Systems of Ecological Stability (Moyzeová and Kenderessy 2015; Kočická et al. 2018). Meanwhile, a number of approaches to measuring ecological stability have been developed over the years, presented in works by Turner et al. (1993), Y ang et al. (2016), and (Kazakov 2019), and others. Poland’s attempts to estimate ecological stability levels to date have concerned the local level (the area of a municipality and the area around a water reservoir), as in Król and Gałaś (2008), Gałaś and Gałaś (2009), the regional level (Subcarpathia Province and Holy Cross Province) by Salata et al. (2016), Ciupa and Suligowski (2018), and the country level (based on land-use structure data from the Central Statistical Office for each province) by Harasim (2015). These studies appear insufficient, and do not pro- vide necessary information on the ecological stability of Poland overall. The studies cited here primarily focus on selected areas of Poland, with analysis of various types of single spatial units, such as drainage basins, municipalities, or provinces. As a result, the results obtained are not comparable. Moreover, the local character of the research does not permit conclusions to be drawn regarding the level of ecological stability throughout the entire country. Some of them are also based only on statistical data or individu- ally vectorized objects from base maps. Particularly in the second case, this carries the risk of generalization and subjectivity. It should be emphasized that the statistical data used in Harasim’s work (2015) refer to the provincial level, and provide only a very general view of ecological stability, with no differentiation between them. The methodology adopted in this paper is the first attempt at comprehensive research on ecological stability carried out at the level of Poland’s administrative units based on spatial data. This paper is an important contribution to help fill this research gap, and such an in-depth analysis of ecological sta- bility is likely to be a useful tool for shaping the broadly defined spatial policy. Acta geographica Slovenica, 61-1, 2021 59 61-1_acta49-1.qxd 28.7.2021 8:06 Page 59 Jolanta Jóźwik, Dorota Dymek, Spatial diversity of ecological stability in different types of spatial units: Case study of Poland The objective of this study is to compare the spatial diversity of Polish municipalities’ ecological stabil- ity and to calculate the spatial autocorrelation of this phenomenon in order to determine spatial dependencies. An additional aim is to verify the results by comparing them with those of an artificial geometric division; that is, a square with sides of 1 km. 2 Material and methods The source material for this paper was the CORINE Land Cover (CLC) 2018 database maintained by the European Environment Agency. This database contains data on current land cover across all of European territory. The data contained in the database are hierarchically grouped at three levels. The first level con- sists of five main land-cover classes: anthropogenic land, agricultural land, forests and semi-natural ecosystems, wetlands, and water areas. The second and third levels provide further details within more pre- cise categories (Heymann et al. 1994). Poland’ s data were provided by the Chief Inspectorate of Environmental Protection. In our study, estimating the degree of ecological stability in administrative units employed the Bičik’s (1995) method of classifying and verifying areas. For the purpose of this analysis, the typology of land- cover classes included in CLC 2018 (Table 1) was replicated to enable appropriate weights of ecological importance (cei), as proposed by Bičik et al. (2015), to be assigned to particular classes of land cover. This treatment allowed a classification to be obtained that can be used in similar area studies at the local, region- al, national, and international scale. An additional advantage of the division proposed in this paper is the fact that the CLC database is widely available, free of charge, and identical for many European countries. It offers access to unified data, reinforcing the spatial compatibility of the dataset, and permitting com- parison of different areas. A number of methods exist for determining a site’s level of ecological stability. The majority are based on assigning a numerical value to the ecological stability indicator, allowing for a qualitative assessment of the area under investigation. The most basic methods are those proposed by Míchal (1982), Löw (1984), and Miklós (1986). Míchal’s method is the simplest. It pertains to the relationship between the surface area of areas defined as stable (e.g., forests, waters, meadows, and pastures) and the surface area of unstable areas (e.g., arable and built-up land). This approach was modified by Löw to assign individual landscape elements to five degrees of stability that were given constants reflecting their importance. The process devel- oped by Miklós does not divide landscape elements into stable and unstable ones, but introduces numerical coefficients that differentiate their ecological stability. This method reflects the ecological stability of the spatial composition of the area studied by determining the relation between the sum of products of areas occupied by individual landscape elements and their corresponding weights of ecological stability to the total area of the terrain in question. This approach was the starting point for Bičik et al. (2015) in their pro- cedure for assessing complex ecological stability used in this paper. In line with the adopted methodology (Bičik et al. 2015), the Coefficient of Ecological Importance (CEI) was used to estimate the degree of eco- logical stability. It is the sum of the products of the appropriate weights of ecological stability and the percentage of the area of each basic unit of assessment (BUA) that is covered by the classes of the features mentioned above. Graphically, it represents a projection of the degree of ecological stability of these BUAs. The CEI for an individual spatial unit is expressed as the following formula (Bičik et al., 2015): (1) where: CEI i = coefficient of ecological importance in BUA i , cei c = weight of ecological importance of land- cover class c, P ci = percentage of the area of the BUA i covered by land-cover class c, n = number of land-cover classes, and i = individual BUA. The values of individual weights of ecological importance (cei) reflect the ecological stability of indi- vidual landscape elements, and fall within a range of 0 to 1, where the value »0« represents anthropogenic areas (heavily transformed by human activity), and »1« represents valuable natural areas (scarcely trans- formed by human activity). Similarly, values of the synthetic coefficient of ecological importance (CEI) are in a range from 0 to 1, with the value »0« standing for ecologically insignificant areas, and »1« eco- logically significant areas. The level of ecological stability of a study area increased with an increase in the 60 CEI i =∑ (c=1) n cei c · P ci 61-1_acta49-1.qxd 28.7.2021 8:06 Page 60 Acta geographica Slovenica, 61-1, 2021 61 Table 1: Reclassification of CORINE Land Cover 2018 land-cover classes with assigned cei weights. Land-cover class Level I CLC Level II CLC Level III CLC CLC Code cei weight 1. Forest and Forest and Forest Broad-leaved forest 311 1.00 semi-natural semi-natural Coniferous forest 312 areas areas Mixed forest 313 Scrub and/or Natural grassland 321 herbaceous Moors and heathland 322 vegetation Sclerophyllous vegetation* 323 associations Transitional woodland/shrub 324 Open spaces with Beaches, dunes, sands 331 little or no Bare rock 332 vegetation Sparsely vegetated areas 333 Burnt areas 334 Glaciers and perpetual snow* 335 2. Wetlands and Wetlands Inland wetlands Inland marshes 411 0.79 water areas Peat bogs 412 Coastal wetlands* Salt marshes* 421 Salines* 422 Intertidal flats* 423 Water bodies Inland waters Water courses 511 Water bodies 512 Marine waters Coastal lagoons 521 Estuaries* 522 Sea and ocean 523 3. Permanent Agricultural areas Pastures Pastures 231 0.64 grasslands 4. Permanent Agricultural areas Permanent crops Vineyards* 221 0.34 crops Fruit trees and berry plantations 222 Olive groves* 223 5. Other Agricultural areas Arable land Non-irrigated arable land 211 0.14 agricultural Permanently irrigated land* 212 areas Rice fields* 213 Heterogeneous Annual crops associated with 241 agricultural areas permanent crops* Complex cultivation patterns 242 Land principally occupied by agriculture, 243 with significant areas of natural vegetation Agro-forestry areas* 244 6. Other areas Artificial surfaces Mine, dump and Mineral extraction sites 131 0.14 construction sites Dump sites 132 Construction sites 133 Artificial, non-agricul- Green urban areas 141 tural vegetated areas Sport and leisure facilities 142 7. Built-up areas Artificial surfaces Urban fabric Continuous urban fabric 111 0.00 Discontinuous urban fabric 112 Industrial, Industrial or commercial units 121 commercial Road and rail networks and associated land 122 and transport Port areas 123 units Airports 124 * Classes that do not occur in Poland. They are included in the table to ensure comparability with other potential studies in other EU countries. 61-1_acta49-1.qxd 28.7.2021 8:06 Page 61 Jolanta Jóźwik, Dorota Dymek, Spatial diversity of ecological stability in different types of spatial units: Case study of Poland index value. It is worth emphasizing that within the same methodology involved in the approach – the concept of planning used to optimize spatial organization, protection, and utilization of the landscape, called Landscape Ecological Planning (LANDEP; Miklós et al. 2019) – some authors use different ranges for the degree of ecological stability values (e.g., Reháčková and Pauditšová 2007; Igondova et al. 2016; Miklós and Špinerová 2019). Two variants were used to estimate the degree of ecological stability. The first one used gminy – admin- istrative units that roughly correspond to municipalities – as the basic unit of assessment. The average area of a municipality was 126 km². The main advantage of using this type of unit is that the message of the results is clear, and data for comparisons are commonly available from various offices and agencies. Unfortunately, a very serious disadvantage of accepting administrative units as BUAs is their internal het- erogeneity that does not fully reflect the spatial distribution of the phenomenon surveyed. The heterogeneity of the area makes it difficult to compare individual units’ results (Balon and Krąż 2013). In order to ver- ify and improve the precision and level of detail of the results in this paper, a second variant was used, based on an artificial geometric division of the area studied, wherein a square with sides of 1 km was adopted as the BUA. The consistency in the area of each field – in this case 1 km² – facilitates statistical calcula- tions, making individual results easily comparable with one another. Moreover, a BUA with a smaller area presents differences in the spatial distribution of ecological stability better and more precisely. Regardless of the variant adopted, the procedure for estimating ecological stability was conducted in the same way. CLC 2018 land-cover classes were trimmed to the borders of the BUA. Then, each newly created parcel within a given BUA was assigned an appropriate weight (cei, according to land-cover class), and its share in the total area of the BUA (P) was calculated. Based on the above, the municipalities were ordered based on the numerical value of their ecological stability coefficient. The calculated values permitted spa- tial units to be classified into five equal classes corresponding to different degrees of ecological stability. Because the method used in this paper does not have fixed threshold values for individual classes, the classification proposed by Petrovič (2005) and used in works such as Mederly et al. (2006), Boltiziar and Olah (2009), Salata et al. (2016), and Krivosudský (2012) was applied. The following classes were distinguished: A: very low eco- logical stability (CEI 0.00–0.20); B: low ecological stability (CEI 0.20–0.40); C: average ecological stability (CEI 0.40–0.60); D: high ecological stability (CEI 0.60–0.80); E: very high ecological stability (CEI 0.80–1.00). Determination the spatial dependence of the phenomenon studied involved performing an analysis of spatial autocorrelation. Spatial autocorrelation permits estimation of the relationship between the value of the examined variable in a given location and the value of this variable in another location. Spatial auto- correlation is referred to when a given phenomenon occurring in a particular location increases or decreas- es the probability of its occurrence in the neighborhood (Bivand 1980). This paper employs the global Moran’s I index, one of the best-known autocorrelation coefficients. It is expressed by the following for- mula (Moran 1950): (2) where: z i = deviation of an attribute for feature (BUA) i from its mean (x i – X ˆ ), z j = deviation of an attribute for feature j from its mean (x j – X ˆ ), w i,j = spatial weight between feature i and j, n = total number of fea- tures, S 0 = aggregate of spatial weights. The Global Moran’s I index permits detection of the strength and character of spatial dependence in the area studied. The statistical value is in a range from −1 to 1, where negative values indicate occurrence of different values of observations in the neighborhood, 0 indicates randomness of the distribution of obser- vation values (lack of autocorrelation), and positive values indicate similarity of values located in the neighborhood (Janc 2006). Moreover, the Local Indicator of Spatial Association (LISA) is used to identify systems occurring in space. It allows for estimation of the degree of similarity of individuals to their neighbors, and determi- nation of the statistical significance of these relationships (Anselin 1995). As a result, each spatial unit was classified as a high-value unit with neighbors of similar value (High-High Cluster), a low-value unit with neighbors of similar value (Low-Low Cluster), a high-value unit with low-value neighbors (High-Low Outlier), a low-value unit with high-value neighbors (Low-High Outlier), or a unit without significant statistical local autocorrelation (Janc 2006). In this paper, a local version of Moran’s I statistics (LISA) was used. Global and local Moran’s I statistics were determined in ArcGIS based on Spatial Statistics Tools. 62 I= n S 0 ∑ i=1 n z i 2 ∑ i=1 n ∑ j=1 n w i,j z i z j 61-1_acta49-1.qxd 28.7.2021 8:06 Page 62 Acta geographica Slovenica, 61-1, 2021 63 Content by: Jolanta Jóźwik Map by: Dorota Dymek Source: Corine Land Cover Data, 2018 © 2020, Maria Curie–Skłodowska University 0 25 50 km Legend Forest and semi natural areas Wetlands and water areas Permanent grasslands Permanent crops Other agricultural areas Other areas Built–up areas MZ WP LB ZP PL DŚ KP PK ŁD WM PM LS MP ŚW ŚL OP 0 100 km Figure 1: Spatial distribution of the land-cover classes identified. DŚ – Lower Silesia, KP – Kuyavia-Pomerania, LB – Lublin, LS – Lubusz, ŁD – Łódź, MP – Lesser Poland, MZ – Masovia, OP – Opole, PK – Subcarpathia, PL – Podlasie, PM – Pomerania, ŚL – Silesia, ŚW – Holy Cross, WM – Warmia- Masuria, WP – Greater Poland, ZP – West Pomerania. 3 Research area The preliminary research permitted the determination and presentation of the spatial distribution of seven main classes of land cover (Figure 1), as well as calculation of their share in the total area of Poland (Table 2). Table 2: Share of land-cover classes as a percentage of Poland’s surface area. Legend Land-cover class Share (%) 1 Forest and semi-natural areas 33.0 2 Wetlands and water areas 2.1 3 Permanent grasslands 9.0 4 Permanent crops 0.6 5 Other agricultural areas 49.2 6 Other areas 0.5 7 Built-up areas 5.6 61-1_acta49-1.qxd 28.7.2021 8:06 Page 63 Jolanta Jóźwik, Dorota Dymek, Spatial diversity of ecological stability in different types of spatial units: Case study of Poland 64 Over 80% of the country’s area is covered by two of the seven classes: other agricultural areas and forest and semi-natural areas; almost half of the country is covered by the former. The highest concentration of these areas occurs in a belt stretching from the north through the central part of the country towards the southwest. Other types of land cover form mosaic systems. The second most dominant class is forest and semi-natural areas, which occupy more than 30% of the area analyzed. They occur in a plane system, par- ticularly in the northwestern and southern part of the country, and in a mosaic system in variable proportions over the remaining area. The remaining classes of land cover occupy a much smaller area, and together account for less than 20% of the country’s terrain. Permanent grasslands, primarily located along river val- leys and in northeastern Poland, dominate among these classes. a small percentage of the country’s area is occupied by wetlands and water areas, as well as built-up (developed) areas. The first two are mainly locat- ed in the north of the country. Built-up areas correspond to the settlement network of the country. Larger concentrations of built-up areas occur in administrative capitals and larger cities. Permanent crops occu- py the lowest percentage of the area analyzed. They are concentrated in three main basins located in central and eastern Poland, where fruit crops are grown. 4 Results A detailed analysis of the land-cover classes identified was used to estimate the degree of ecological sta- bility in municipalities. Further in the study, the percentage of areas occupied by individual groups in regional terms and their spatial distribution were analyzed (Figures 2 and 3). Group a includes heavily urbanized municipalities. The landscape of these areas is not stable or con- sistent. The municipalities are dispersed, occupying a relatively small percentage of the country’s area (4.7%). They only merge into small clusters in several places in Poland. More than one-third of the country’ s area (35.7%) is occupied by areas of low ecological stability (group B), forming relatively extensive patches scattered throughout Poland. Group C occupies the largest area in the country (37.9%). Municipalities belonging to this type form relatively large clusters cutting across areas that primarily belong to group B. In total, groups B and C occu- py nearly three-quarters of the area of Poland. These groups also dominate in almost all provinces. Areas of high ecological stability (group D) cover almost one-fifth of the country’s area (18.6%). They are most highly concentrated in northern and western Poland. Small concentrations are also found in the southern and southeastern parts of the country. Group E occupies the smallest area (only 3.1% of Poland’s total area). This type emerges in ecological- ly stable areas with significant natural functions and little transformation by human activity. Municipalities included in group E are characterized by a high degree of spatial dispersion. They only form a band-shaped cluster along the southeastern border of the country. In the central part of Poland there are hardly any such areas. The Lubusz Province (LS) compares most favorably to the other provinces. No municipalities classi- fied as group a were recorded there, and more than half of the province’s area (70.4%) belongs to groups D and E. The most unfavorable situation occurs in the Łódź (ŁD) and Kuyavia-Pomerania (KP) Provinces. More than half of their area is occupied by groups a and B. Visual evaluation of the obtained results suggests the occurrence of spatial autocorrelation. To confirm this assumption, global Moran’s I statistics were used. The calculations employed the spatial weighting matrix resulting from linear standardization of the neighborhood matrix, where a common boundary expressed by linear or point contact was used as a criterion of neighborhood. The statistic value obtained is 0.542 (significantly different from 0). The positive sign means that the analyzed case shows a tendency to con- centrate units with similar CEI value in the neighborhood. Moreover, given the z-score of more than 2.58 and p-value < 0.001, the likelihood that this clustered pattern could be the result of random chance is less than 1%. The high values of global Moran’s I statistics are confirmed by the image obtained from the Local Indicators of Spatial Association (LISA) analysis. This analysis allowed to confirm the assumption of the occurrence of cluster systems in the spatial distribution structure of the CEI (Figure 4). In the second variant, the percentage of areas corresponding to particular classes of ecological stability changed quite significantly (Figures 5 and 6). Among all the distinguished classes, areas included in group a constitute by far the largest surface area of the country (41.0%). a similar dynamic is observed at the sub- 61-1_acta49-1.qxd 28.7.2021 8:06 Page 64 national (provincial) level. Almost all of the provinces are dominated by this class, and in several cases these areas occupy up to half of their area. The highest concentration of these areas occurs in central and western Poland, and in a belt stretching from the southwest to the southeast. Areas belonging to group B occupy a relatively small area of Poland (12.9%). They are characterized by a mosaic system and signifi- cant dispersion throughout the country. They are primarily located in the vicinity of areas classified under group A. Greater concentrations of these (B group) areas occur in the Masovia (MZ), Łódź (ŁD), Holy Cross (ŚW), Lublin (LB), and Podlasie (PL) Provinces. Group C shows similar dynamics. These areas occu- py the smallest area in the country (10.2%), and are characterized by significant, but uniform dispersion. Areas included in group C do not merge into larger clusters. The situation is slightly different for areas with high ecological stability (group D). These areas are considerably scattered throughout the country, and occupy a similar percentage of the area as groups B and C (11.1%). They merge and form several larg- er clusters, particularly in the northern part of the country. The differences in the share of groups B, C, and D across all provinces are not significant, and remain at a similar level. The second largest group in Acta geographica Slovenica, 61-1, 2021 65 Content by: Jolanta Jóźwik Map by: Dorota Dymek Source: Corine Land Cover Data, 2018 © 2020, Maria Curie–Skłodowska University 0 25 50 km Legend Very low ecological stability (A) Low ecological stability (B) Medium ecological stability (C) High ecological stability (D) Very high ecological stability (E) Figure 2: Spatial distribution of Poland’s ecological stability classes based on the CEI value (BUA: municipalities). 61-1_acta49-1.qxd 28.7.2021 8:06 Page 65 Jolanta Jóźwik, Dorota Dymek, Spatial diversity of ecological stability in different types of spatial units: Case study of Poland terms of surface area is group E (24.9%). The spatial distribution of this group is fairly diversified. Areas of this type occur in both mosaic and plane systems. The highest concentration of these areas is located in the northern, western, and southern parts of Poland. Like in the first variant, the most favorable situation in terms of ecological stability occurred in the Lubusz Province (LS), where groups D and E cover over 50.0% of the area. The worst situation occurred in the Łódź (ŁD) and Kuyavia-Pomerania (KP) Provinces, where groups a and B cover more than 60.0%. In this variant, the latter also included the Opole (OP) and Greater Poland (WP) Provinces. Like in the case of municipalities, the global Moran’s I statistics indicated the occurrence of spatial auto- correlation. The statistics value obtained was 0.716, suggesting a tendency to group units with similar CEI values. Given the z-score of more than 2.58 and p-value < 0.001, also in this case the likelihood that this clustered pattern could be the result of random chance is less than 1%. LISA analysis confirmed the occur- rence of clusters in the area analyzed (Figure 7). 66 ŚW ŚL PM PL PK OP MZ MP ŁD LS LB KP DŚ POL ZP WP WM Very low ecological stability (A) Low ecological stability (B) Medium ecological stability (C) High ecological stability (D) Very high ecological stability (E) 10 20 30 40 50 60 70 80 90 100 0 Figure 3: Share (%) of ecological stability classes at the national and provincial level (BUA: municipalities). POL – Poland, DŚ – Lower Silesia, KP – Kuyavia-Pomerania, LB – Lublin, LS – Lubusz, ŁD – Łódź, MP – Lesser Poland, MZ – Masovia, OP – Opole, PK – Subcarpathia, PL – Podlasie, PM – Pomerania, ŚL – Silesia, ŚW – Holy Cross, WM – Warmia-Masuria, WP – Greater Poland, ZP – West Pomerania. 61-1_acta49-1.qxd 28.7.2021 8:06 Page 66 5 Discussion This study estimates the degree of ecological stability at the level of administrative units in Poland. The assessment was performed in two variants. The units originally analyzed were municipalities (Pol. gminy). The verification of the results obtained involved extending the study. a square with sides of 1 km was used as the basic unit of assessment. The spatial analyses conducted also permitted the spatial depen- dence of the phenomenon to be identified and spatially illustrated. This research is an important contribution to Polish research on ecological stability. Owing to the use of the CLC 2018 unified database, the method is characterized by a relatively high level of detail and high degree of objectivity. The basic unit of assessment applied (an artificial geometric division: a square with sides of 1 km) permitted comparison of units with each other, which until now was not possible due to different types of spatial units used by other authors. Moreover, the analysis was carried out for the entire Acta geographica Slovenica, 61-1, 2021 67 Content by: Jolanta Jóźwik Map by: Dorota Dymek Source: Corine Land Cover Data, 2018 © 2020, Maria Curie–Skłodowska University Anselin Local Moran's I High–high cluster High–low outlier Low–high outlier Low–low cluster 0 25 50 km Figure 4: Distribution of cluster and outlier analysis (Anselin Local Moran’s I) for CEI in Poland (BUA: municipalities). 61-1_acta49-1.qxd 28.7.2021 8:06 Page 67 Jolanta Jóźwik, Dorota Dymek, Spatial diversity of ecological stability in different types of spatial units: Case study of Poland territory of Poland, allowing for some conclusions regarding Poland’s ecological stability to be drawn. The added value of the study is the extension of the statistical analysis to include a spatial analysis which, by demonstrating the spatial dependence of the ecological stability of the landscape, confirmed the occur- rence of spatial units with similar values in close neighborhoods (clusters). Adopting Poland’s principal administrative units (municipalities) as BUAs revealed a clear dominance of areas concentrated around low and medium ecological stability (groups B and C). Using the second vari- ant (artificial geometric divisions) showed the predominance of groups with extreme CEI values (group a and group E) which constituted a small percentage of the total in the first approach. Regardless of the variant applied, the most favorable situation in terms of ecological stability was observed in the following provinces: Lubusz (LS), Subcarpathia (PK), and West Pomerania (ZP). Moreover, in the case of the Subcarpathia Province (PK), the result is similar to the results of research for this area presented in 2016 by Salata et al., where a different database was used, one that is slightly more accurate than that used in 68 Content by: Jolanta Jóźwik Map by: Dorota Dymek Source: Corine Land Cover Data, 2018 © 2020, Maria Curie–Skłodowska University 0 25 50 km Legend Very low ecological stability (A) Low ecological stability (B) Medium ecological stability (C) High ecological stability (D) Very high ecological stability (E) Figure 5: Spatial distribution of Poland’s ecological stability classes based on CEI value (BUA: a square with sides of 1 km). 61-1_acta49-1.qxd 28.7.2021 8:06 Page 68 this work. This proves that the method described here can be a useful tool for comparisons between other European Union countries based on a database made according to uniform principles: the CLC database. The heterogeneity of the results derives from the type and size of the basic units of assessment that were used. As expected, the larger the BUA, the less accurate the obtained results. Both variants have their advantages and disadvantages. The first approach reflects the general character of the municipalities fair- ly well, and provides a relatively easy and clear message for non-specialists, especially decision-makers at various administrative levels. The second variant is much more precise, and reflects the actual state of the analyzed areas more accurately, while the equal area of each BUA facilitates comparison of results (Balon and Krąż 2013). This approach may facilitate the identification of sensitive areas where fluctuations in the level of ecological stability are likely to occur. It should be emphasized, however, that according to the logic of ecological fallacy, it is not appropriate to transfer conclusions for the examined elements to every sin- gle unit of area that makes up that element. Acta geographica Slovenica, 61-1, 2021 69 10 20 30 40 50 60 70 80 90 100 0 ZP WP WM ŚW ŚL PM PL PK OP MZ MP ŁD LS LB KP DŚ POL Very low ecological stability (A) Low ecological stability (B) Medium ecological stability (C) High ecological stability (D) Very high ecological stability (E) Figure 6: Share (%) of ecological stability classes at the national and provincial level (BUA: a square with sides of 1 km). POL – Poland, DŚ – Lower Silesia, KP – Kuyavia-Pomerania, LB – Lublin, LS – Lubusz, ŁD – Łódź, MP – Lesser Poland, MZ – Masovia, OP – Opole, PK – Subcarpathia, PL – Podlasie, PM – Pomerania, ŚL – Silesia, ŚW – Holy Cross, WM – Warmia-Masuria, WP – Greater Poland, ZP – West Pomerania. 61-1_acta49-1.qxd 28.7.2021 8:06 Page 69 Jolanta Jóźwik, Dorota Dymek, Spatial diversity of ecological stability in different types of spatial units: Case study of Poland The cartographic presentation of the studied phenomenon made it possible to distinguish two main systems of spatial distribution of ecological stability values, namely the plane system and the mosaic sys- tem. From the ecological point of view, the plane system is more advantageous to the extent that it is configured as a compact complex of areas that are easier to manage. The mosaic system is unfavorable, due to its sig- nificant dispersion and high internal heterogeneity of individual areas. These systems are characterized by relatively high volatility; that is, the susceptibility to transitioning rapidly to extreme states (Balon 2004; Gałaś and Gałaś 2009). Therefore, areas with mosaic systems should be given special attention to avoid further deterioration of their ecological stability. In the case of environmentally valuable areas, it is not desirable to have low or very low values of the ecological stability index in the neighborhood. It may lead to the weakening of their potential. High frag- mentation and dispersion of areas included in group E make it significantly more difficult – and, in extreme cases, impossible – to ensure that they remain undegraded. It is very difficult to take effective protective 70 Content by: Jolanta Jóźwik Map by: Dorota Dymek Source: Corine Land Cover Data, 2018 © 2020, Maria Curie–Skłodowska University 0 25 50 km Anselin Local Moran's I High–high cluster High–low outlier Low–high outlier Low–low cluster Figure 7: Distribution of cluster and outlier analysis (Anselin Local Moran’s I) for CEI in Poland (BUA: a square with sides of 1 km). 61-1_acta49-1.qxd 28.7.2021 8:07 Page 70 measures in such areas. Moreover, the risk of irrational economic management in these areas is high, which in turn may contribute to harmful changes in the way they are administered, and in extreme cases to a com- plete loss of ecological potential. Such areas require both specialized knowledge and well-thought-out actions. On the other hand, mosaic systems can contribute to a sustainable flow of ecosystem services, enrich the land- scape structure, and enhance the landscape’s aesthetic values (Waldhardt et al. 2004), prevent soil erosion (Boardman and Poesen 2006), and significantly reduce spatial tensions and conflicts between stakeholders. Research on the degree of ecological stability of a given area can be very useful for the implementa- tion of beneficial land cover or changes in land use. Area analyses of this type can be applied in practice both at the initial and final stages of spatial development planning (as an important element of environ- mental management), in addition to being helpful in the preparation of landscape audits. They may be used to identify resources and evaluate their potential for further use. 6 Conclusions The objective of this study was to compare the spatial diversity of administrative units’ ecological stability, and to calculate the spatial autocorrelation of the phenomenon studied in order to study spatial dependencies. An additional goal was to verify the results obtained by comparing them with an artificial geometric divi- sion; that is, squares with sides of 1 km. The methods applied were sufficient for achieving the research objective. The results’ degree of detail mainly depends on the spatial unit used. Analyses of this type based on a geographical information system can be easily modified and adjusted depending on the purpose and area of analysis. Moreover, the applied method confirms that the CLC database can be successfully used to determine a site’s degree of ecological stability. It also permits continuous monitoring of changes in land cover or land-use structure, and can be a useful tool that supports sustainable development policies. The research showed that the use of different types of spatial units – administrative units (municipal- ities), and artificial geometric divisions (squares with sides of 1 km) – significantly affects the results: the larger the basic unit of assessment, the less accurate the results obtained. In the first variant (BUA: munic- ipalities), areas with low and average ecological stability were clearly dominant. It can be concluded that the ecological stability of Poland was close to the average level. The second approach (BUA: a square with sides of 1 km) yielded dominance of extreme areas in terms of ecological stability. Regardless of the adopt- ed variant, the most advantageous situation regarding ecological stability was determined in the Lubusz Province, and the most unfavorable in the Łódź and Kuyavia-Pomerania Provinces. In both cases, the results of the global Moran’s I and LISA analysis confirmed and illustrated the occurrence of spatial dependen- cies of the phenomenon studied. 7 References Anselin, L. 1995: Local indicators of spatial association – LISA. Geographical Analysis 27-2. DOI: https:/ /doi.org/ 10.1111/j.1538-4632.1995.tb00338.x Balon, J. 2004: Badania stabilności środowiska jako podstawa planowania ekologiczno-krajobrazowego w górach. Studia ekologiczno-krajobrazowe w programowaniu rozwoju zrównoważonego. Przegląd polskich doświadczeń u progu integracji z Unią Europejską. Gdańsk. Balon, J. 2006: Stability of the natural environment as a subject of geoecological research. Regionalne Studia Ekologiczno-Krajobrazowe, Problemy Ekologii Krajobrazu 16. Balon, J., Krąż, P . 2013: Ocena jakości krajobrazu  – dobór prawidłowych jednostek krajobrazowych. Identyfikacja i waloryzacja krajobrazów – wdrażanie Europejskiej Konwencji Krajobrazowej. Warszawa. Baran-Zgłobicka, B., Zgłobicki, W . 2012: Mosaic landscapes of SE Poland: should we preserve them? Agroforestry Systems 85. DOI: https://doi.org/10.1007/s10457-011-9436-x Bastian, O., Schreiber, K. F. 1999: Analyse und ökologische Bewertung der Landschaft. Berlin. Bičik, I. 1995: Analýza dat o využití půdy k hodnocení dlouhodobých změn krajiny. Geographia Slovaca 10. Bičik, I., Kupková, L., Jeleček, L., Kabrda, J., Štych, P ., Janoušek, Z., Winklerová, J. 2015: Land use changes in the Czech Republic 1845–2010: Socio-economic driving forces, Springer Geography. New Y ork. DOI: https://doi.org/10.1007/978-3-319-17671-0 Acta geographica Slovenica, 61-1, 2021 71 61-1_acta49-1.qxd 28.7.2021 8:07 Page 71 Jolanta Jóźwik, Dorota Dymek, Spatial diversity of ecological stability in different types of spatial units: Case study of Poland Bivand, R. 1980: Autokorelacja przestrzenna a metody analizy statystycznej w geografii. Analiza regresji w geografii. Poznań. Boardman, J., Poesen, J. 2006: Soil erosion in Europe: major processes, causes and consequences. Soil Erosion in Europe. Wiley. Chichester. DOI: https://doi.org/10.1002/0470859202.ch36 Boltiziar, M., Olah, B. 2009: Krajina a jej štruktúra (Mapovanie, zmeny a hodnotenie). Univerzita Konštantína Filozofa v Nitre. Nitra. Ciupa, T ., Suligowski, R. 2018: Użytkowanie ziemi a stabilność ekologiczna obszarów wiejskich województwa świętokrzyskiego. Roczniki Naukowe Ekonomii Rolnictwa I Rozwoju Obszarów Wiejskich 105-1. DOI: https://doi.org/10.22630/RNR.2018.105.1.2 Dotterweich, M. 2008: The history of soil erosion and fluvial deposits in small catchments of central Europe: deciphering the long-term interaction between humans and the environment. Geomorphology 101-1,2. DOI: https://doi.org/10.1016/j.geomorph.2008.05.023 Forman, R. T. T., Godron, M. 1986: Landscape ecology. New Y ork. Fuentes, E. R. 1984: Human impact and ecosystem resilience in the Southern Andes. Mountain Research and Development 4-1. DOI: https://doi.org/10.2307/3673168 Gałaś, S., Gałaś, A. 2009: Assessment of ecological stability of spatial and functional structure around Świnna Poręba water reservoir. Polish Journal of Environmental Studies 18-3A. Ganjurjav, H., Zhang, Y ., Gornish, E. S., Hu, G., Li, Y ., Wan, Y ., Gao, Q. 2019: Differential resistance and resilience of functional groups to livestock grazing maintain ecosystem stability in an alpine steppe on the Qinghai-Tibetan Plateau. Journal of Environmental Management 251. DOI: https://doi.org/ 10.1016/j.jenvman.2019.109579 Geng, S., Shi, P ., Song, M., Zong, N., Zu, J., Zhu, W . 2019: Diversity of vegetation composition enhances ecosystem stability along elevational gradients in the Taihang Mountains, China. Ecological Indicators 104. DOI: https://doi.org/10.1016/j.ecolind.2019.05.038 Geri, F ., Rocchini, D., Chiarucci, A. 2010: Landscape metrics and topographical determinants of large-scale forest dynamics in a Mediterranean landscape. Landscape and Urban Planning 95. DOI: https:/ /doi.org/ 10.1016/j.landurbplan.2009.12.001 Gigon, A. 1983: Typology and principles of ecological stability and instability. Mountain Research and Development 3-2. DOI: https://doi.org/10.2307/3672989 Gigon, A., Grimm, V . 2014: Stabilitätskonzepte in der Ökologie: Typologie und Checkliste für die Anwendung. Handbuch der Umweltwissenschaften: Grundlagen und Anwendungen der Ökosystemforschung, New Y ork. DOI: https://doi.org/10.1002/9783527678525.hbuw1997030 Harasim, A. 2015: Użytkowanie powierzchni ziemi w Polsce w aspekcie stabilności ekologicznej. Roczniki Naukowe Stowarzyszenia Ekonomistów Rolnictwa i Agrobiznesu 17-1. Heymann, Y ., Steenmans, C., Croissille, G., Bossard, M. 1994: CORINE land cover: Technical guide. Luxembourg. Holling, C. S. 1973: Resilience and stability of ecological systems. Annual Review of Ecology and Systematics 4. DOI: https://doi.org/10.1146/annurev.es.04.110173.000245 Igondova E., Pavlickova K., Majzlan O. 2016: The ecological impact assessment of a proposed road devel- opment (the Slovak approach). Environmental Impact Assessment Review 59. DOI: https://doi.org/ 10.1016/j.eiar.2016.03.006 Janc, K. 2006: Zjawisko autokorelacji przestrzennej na przykładzie statystyki I Morana oraz lokalnych wskaźników zależności przestrzennej (LISA) – wybrane zagadnienia metodyczne. Idee i praktyczny uniwersalizm geografii. Warszawa. Kazakov, L. K. 2019: Landscape ecology culture and some principles of sustainable nature use. Current Trends in Landscape Research. Cham. DOI: https://doi.org/10.1007/978-3-030-30069-2_3 Kienholz, H., Hafner, H., Schneider, G. 1984: Stability, instability and conditional instability mountain ecosys- tem concepts based on a field survey of the Kakani area in the Middle Hills of Nepal. Mountain Research and Development 4-1. DOI: https://doi.org/10.2307/3673170 Kočická, E., Diviaková, A., Kočický, D., Belaňová, E. 2018: Territorial system of ecological stability as a part of land consolidations (cadastral territory of Galanta – Hody, Slovak Republic). Ekológia (Bratislava) 37-2. DOI: https://doi.org/10.2478/eko-2018-0015 72 61-1_acta49-1.qxd 28.7.2021 8:07 Page 72 Krivosudský, R. 2012: The evaluation of land use changes in land surrounding watercourses as one of the key elements for assessing the ecological stability in rural areas. Proceedings of International Ph.D. Students Conference MendelNet 2012. Brno. Król, E., Gałaś, S. 2008: Ocena stabilności ekologicznej krajobrazu gminy uzdrowiskowej Busko-Zdrój. Problemy Ekologii Krajobrazu 22. Löw, J. 1984: Zásady pro vymezování a navrhovaní územních systému ekologické stability v územne-pláno- vací praxis. Brno. Mederly, P ., Halada, L., Kartusek, V ., Benčo, J., Krautschneider, J., Vlčková, T. 2006: Projekt pozemkových úprav Dlhá nad Váhom. Nitra. Messerli, B. 1983: Stability and instability of mountain ecosystems: Introduction to a workshop sponsored by the United Nations University. Mountain Research and Development 3-2. DOI: https://doi.org/ 10.2307/3672988 Míchal, I. 1982: Principy krajinářského hodnocení území. Architektúra a urbanizmus 16. Miklós, L., Diviaková, A., Izakovičová, Z. (eds.) 2019: Conclusion. Ecological Networks and T erritorial Systems of Ecological Stability. Cham. DOI: https://doi.org/10.1007/978-3-319-94018-2_5 Miklós, L., Špinerová, A. 2019: Landscape-ecological planning LANDEP . Cham. DOI: https:/ /doi.org/10.1007/ 978-3-319-94021-2 Miklós. L. 1986: Stabilita krajiny v ekologickom genereli SSR. Životné Prostredie 20-2. Moran, P . A. P . 1950: Notes on continuous stochastic phenomena. Biometrika 37-1,2. DOI: https:/ /doi.org/ 10.1093/biomet/37.1-2.17 Moyzeová, M., Kenderessy, P . 2015: Territorial systems of ecological stability in land consolidation pro- jects (Example of proposal for the LSES of Klasov village, Slovak Republic). Ekológia (Bratislava) 34-4. DOI: https://doi.org/10.1515/eko-2015-0032 Petrovič, F . 2005: Vývoj krajiny voblasti štálového osídlenia Pohronského Inovca a Tribeča, Ústav krajinnej ekológie SAV . Nitra. Reháčková, T., Pauditšová, E. 2007: Metodický postup stanovenia koeficientu ekologickej stability krajiny. Acta Environmentalica Universitatis Comenianae (Bratislava) 15-1. Ribeiro, D., Šmid Hribar, M. 2019: Assessment of land-use changes and their impacts on ecosystem ser- vices in two Slovenian rural landscapes. Acta geographica Slovenica 59-1. DOI: https:/ /doi.org/10.3986/ AGS.6636 Richling, A., Solon, J. 1994: Ekologia krajobrazu. Warszwa. Salata, T., Cegielska, K., Gawroński, K., Czesak, B. 2016: Zróżnicowanie przestrzenne wskaźnika istotności ekologicznej w województwie podkarpackim. Inżynieria Ekologiczna 50. DOI: https:/ /doi.org/10.12912/ 23920629/65505 Turner, M. G., Romme, W . H., Gardner, R. H., O’Neill, R. V ., Kratz, T. K. 1993: a revised concept of landscape equilibrium: disturbance and stability on scaled landscapes. Landscape Ecology 8. DOI: https:/ /doi.org/ 10.1007/BF00125352 Urbanc, M., Printsmann, A., Palang, H., Skowronek, E., W oloszyn, W ., Konkoly-Gyuró, É. 2004: Comprehension of rapidly transforming landscapes of Central and Eastern Europe in the 20th century. Acta geographica Slovenica 44-2. DOI: https://doi.org/10.3986/AGS44204 Verburg, P . H., de Koning, G. H. J., Kok, K., Veldkamp, A., Bouma, J. 1999: a spatial explicit allocation pro- cedure for modelling the pattern of land use change based upon actual land use. Ecological Modelling 116-1. DOI: https://doi.org/10.1016/S0304-3800(98)00156-2 Verburg, P . H., Soepboer, W ., Veldkamp, A., Limpiada, R., Espaldon, V ., Mastura, S. 2002: Modeling the spatial dynamics of regional land use: The CLUE-S model. Environmental Management 30. DOI: https://doi.org/10.1007/s00267-002-2630-x Vološčuk, I., Míchal, I. 1991: Rozhovory o ekológii a ochrane prírody. Martin. Waldhardt, R., Simmering, D., Otte, A. 2004: Estimation and prediction of plant species richness in a mosa- ic landscape. Landscape Ecology 19. DOI: https://doi.org/10.1023/B:LAND.0000021722.08588.58 Widacki, W . 1979: Relacja człowiek-środowisko jako zagadnienia sterowania. Przegląd Geograficzny 51-4. Winiger, M. 1983: Stability and instability of mountain ecosystems definitions for evaluation of human systems. Mountain Research and Development 3-2. DOI: https://doi.org/10.2307/3672990 Acta geographica Slovenica, 61-1, 2021 73 61-1_acta49-1.qxd 28.7.2021 8:07 Page 73 Jolanta Jóźwik, Dorota Dymek, Spatial diversity of ecological stability in different types of spatial units: Case study of Poland Y ang, Y ., Liao, L., Y an, L., Hu, X., Huang, H., Xiao, S. 2016: The big data analysis of land use evolution and its ecological security responses in Silver Beach of China by the clustering of spatial patterns. Cluster Computing 19. DOI: https://doi.org/10.1007/s10586-016-0659-5 Zaušková, Ľ., Midriak, R. 2007: Únosnosť a využívanie krajiny. Banská Bystrica. Zhang, T., Meng, M., Cao, Y ., Jiang, Y ., Liu, J. 2017: Research on effect of construction of Huitengxile wind farm to grassland ecosystem with transitions of land use and landscape pattern. Procedia Engineering 174. DOI: https://doi.org/10.1016/j.proeng.2017.01.222 74 61-1_acta49-1.qxd 28.7.2021 8:07 Page 74