Rock glaciers in the Western and High Tatra Mountains, Western Carpathians

ABSTRACT A detailed map of rock glaciers at a scale of 1 : 40 000 is produced for the Western and High Tatra Mts., Western Carpathians, based on remotely sensed mapping. We inventoried a total of 383 rock glaciers, covering a total area of 13.84 km2. Most rock glaciers (85 %) are considered relict (without permafrost). These landforms have an average lower limit of 1684 m asl and occupy a total area of 12.50 km2. In contrast, intact rock glaciers (containing permafrost) cover a total area of 1.34 km2 and their average lower limit is located at 1986 m asl, which is 56 m above the previously suggested lower limit of discontinuous permafrost. The inventory adds new information about rock-glacier occurrence in the European high-mountain areas and improves the understanding of present and past environmental conditions in the region.

Despite a considerable effort resulting in the above extensive list of literature, there still remain large areas both in the Western and High Tatra Mts., where rock glaciers have not been mapped yet. Furthermore, many previous investigations are only descriptive, incomplete and do not enable to determine the rockglacier occurrence precisely or to extract their morphological attributes. Finally, none of the above rock-glacier studies have focused on the entire area of the Western and High Tatra Mts. Accordingly, the aim of our study is to complete the information on rock-glacier occurrence based on remotely sensed mapping, validated by field and literature data, and to introduce the first detailed map of rock glaciers covering the entire area of the Western and High Tatra Mts.

Study area
The Western and High Tatra Mts. are c. 42 km and c. 26 km, respectively, long mountain ranges, located in the northernmost part of the Carpathian arc and stretching longitudinally along the Slovak-Polish border ( Figure 1). They reach their maximum elevations of 2248 m asl and 2655 m asl at the top of Bystrá Peak and Gerlachovský štít Peak, respectively. Geology of the mountain ranges is dominated by Carboniferous-Permian granitic rocks, enriched with Palaeozoic methamorphic rocks, such as gneisses, migmatites or amplibolites in the Western Tatra Mts. On northern sides of the mountain ranges, the crystalline core is overlaid by thrust sheets of Mesozoic sedimentary rocks, particularly limestones and dolomites. The foothills have an extensive cover of Cretaceous-Palaeogene flysch rocks, such as claystones, siltstones, sandstones, conglomerates and breccias (Jurewicz, 2007;Nemčok et al., 1994;Figure 2).
Extensive glaciers recurrently evolved during the cold phases of the Pleistocene (Figure 3), which resulted in typical alpine-type topography with numerous cirques, U-shaped valleys and distinct moraine ridges in both mountain ranges (e.g. Křížek & Mida, 2013;Makos, Nitychoruk, & Zreda, 2013a, 2013bZasadni & Kłapyta, 2014), but all pronouncedly more developed in the High Tatra Mts. Currently, the mountains host no glaciers, but perennial snowfields and firn-ice patches can be found in suitable locations (Gądek, 2014). In the period 1981-2010, the mean annual air temperature (MAAT) at the top of Lomnický štít Peak, at 2635 m asl, was -3.4°C and mean annual precipitation amounted 1653 mm (Slovak Hydrometerological Institute). The present-day climatic snowline has been calculated at c. 2500-2600 m asl and c. 2700-2800 m asl on northern and southern slopes, respectively (Zasadni & Kłapyta, 2009), and the lower limit of discontinuous permafrost has been estimated at 1930 ± 150-200 m asl based on the analysis of air thawing and freezing indices and complementary geophysical soundings and measurements of bottom temperature of snow (Dobiński, 1997a(Dobiński, , 2004(Dobiński, , 2005.

Areal extent and coordinate system
The map focuses on the Western and High Tatra Mts. in a scale of 1 : 40 000 and covers the area of 838.34 km 2 between c. 49°06 ′ -49°17 ′ N and c. 19°39 ′ -20°15 ′ E ( Figure 1). The map is projected in the coordinate system WGS 1984 UTM Zone 34N. Accordingly, some layers, originally acquired in the coordinate system S-JTSK (based on Bessel ellipsoid), had to be transformed into the map projection. Considering the medium map scale, we converted the data using the transformation tools implemented in ESRI ArcMap 10.3. We employed the most accurate transformation algorithm (c. 1 m error) according to our own testing of the conversion accuracy between the above two coordinate systems (Křížek, Uxa, & Mida, 2016).

Digital elevation model
We used the existing digital elevation model (DEM) with a horizontal resolution of 10 m produced by stereo photogrammetry from aerial images (Tatra National Park, 2009) as a primary topographic base for the Main Map and as a supplementary source of information for mapping of rock glaciers, delineation of their contributing areas and extraction of elevation attributes. The model represents the most detailed and most widely utilized DEM covering the entire study area to date. Nonetheless, it contains elevation errors and other defects, such as straight break lines, which are mostly located in lower elevated areas unoccupied by rock glaciers, particularly along the alpine treeline and adjacent forest belt located on southern slopes (c.f. Zasadni & Kłapyta, 2014). Therefore, the model was adjusted prior to final map compilation so that the defective parts and their immediate surroundings (accounting for c. 11 % of the DEM area) were replaced by a supportive co-registrated 10 m DEM constructed from contour data with equidistance of 10 m provided by the Geodetic and Cartographic Institute Bratislava (GCIB). Both models were merged together and filtered by averaging within a moving window of 5 × 5 cells to ensure smooth transition at the contact of the models. Subsequently, shaded relief and contours with equidistance of 50 m were generated based on this updated and smoothed DEM and thereafter used in the final map.

Other topographic features
Beside the above DEM-based elements, we also added a set of elevation points into the map to provide complementary elevation information. In addition, other topographic features, such as watercourses, lakes, roads, railways, cable cars, built-up areas and state boundary, were integrated into the map for better orientation.
The elevation points and cable cars were obtained by digitizing the national topographic map series in a scale of 1 : 25 000 and 1 : 50 000 accessible on geoportals of the GCIB (geoportal.sk) and the Head Office of Geodesy and Cartography (HOGC; geoportal.gov.pl). In the Slovak territory of the study region, watercourses and lakes were taken from the database SVM 50, which is based on the Basemap of Slovak Republic in a scale of 1 : 50 000. In the Polish territory, these topographic features were acquired by manual digitization of the topographic map accessible at the geoportal of HOGC. Roads, railways and built-up areas for the entire area were extracted from the freely available OpenStreetMap database (download.geofabrik.de).
The topographic layers were checked for quality and particularly their mutual consistency and, if necessary, they were manually adjusted, for example, watercourse directions were slightly changed to appropriately fit the contour lines.

Additional map content
We included a schematic geological map ( Figure 2) based on Nemčok et al. (1994) and Bezák et al. (2013) as a background layer in the Main Map in order to infer bedrock lithology of rock glaciers, with division into five geological units: Palaeozoic gneiss rocks, migmatite and amphibolite; Carboniferous-Permian granitic rocks; Mesozoic limestone, dolomite, sandstone, shale and quartzite; Cretaceous-Palaeogene claystone, siltstone, sandstone, conglomerate and breccia; and Palaeogene sandstone, limestone, conglomerate and breccia. Furthermore, the above layer was complemented with both proved and assumed faults based on the same data sources.
We also added the Last Glacial Maximum glacier extent in the Tatra Mts. (Figure 3) after Zasadni and Kłapyta (2014) to provide information about the relationship between rock glaciers and former glaciation limit.
Finally, as rock glaciers are widely accepted indicators of the lower limit of discontinuous mountain permafrost (Barsch, 1996), we show the level of 1930 m asl in the Main Map, which corresponds to the average minimum elevation of discontinuous permafrost occurrence in the Tatra Mts. suggested by Dobiński (1997aDobiński ( , 2004Dobiński ( , 2005.  Nemčok et al. (1994) and Bezák et al. (2013). Legend: 1gneiss rocks, migmatite, amphibolite (Palaeozoic); 2granitic rocks (Carboniferous-Permian); 3limestone, dolomite, sandstone, shale, quartzite (Mesozoic); 4claystone, siltstone, sandstone, conglomerate, breccia (Cretaceous-Palaeogene); 5sandstone, limestone, conglomerate, breccia (Palaeogene); 6fault (proved); 7fault (assumed). Red dashed lines indicate geographic subdivision of the Tatra Mts. or transverse ridges and furrows. In areas with closed or semi-closed vegetation cover, we followed vegetation patterns associated with the rock-glacier surface morphology as well (Barsch, 1996). The mapping was also supported by field inspections in 2014, 2015 and 2016, during which c. 22 % of the mapped landforms, mostly located in the High Tatra Mts., were visited and their occurrence and delineation was verified. The survey was supported by previous literature reports and maps as well (Dzierżek & Nitychoruk, 1986;Kłapyta, 2008Kłapyta, , 2009Kłapyta, , 2010Kłapyta, , 2011Kłapyta, , 2012Kłapyta, , 2013aKłapyta, , 2013bKłapyta, , 2015Kłapyta & Kołaczek, 2009;Kotarba, 1988Kotarba, , 1991Kotarba, -1992Kotarba, , 2007Nemčok & Mahr, 1974;Zasadni & Kłapyta, 2016) that enable to extract precise information about the location of rock glaciers. However, we did not take over the literature information completely in all cases because the above mappings have emerged over several decades and had inconsistent data sources and mapping methodologies. Consequently, there is occasionally no consensus among researchers regarding the landform identification or classification. An excellent example of this inconsistency is the debris accumulation located near the Zadni Staw Polski Lake in the Pięciu Stawów Polskich Valley in the Polish High Tatra Mts., which was classified as a rock glacier (Dzierżek et al., 1987), a moraine ridge (Makos et al., 2013a) as well as a rock avalanche (Zasadni & Kłapyta, 2016). Therefore, we carefully considered the integration of each individual previously mapped landform into the database mainly based on its appearance on aerial imagery in order to ensure the consistency of the inventory in terms of landform identification, classification and delineation throughout the whole region.

Rock-glacier mapping
Each rock glacier was manually delineated along the foot of its frontal and lateral slopes towards the rooting zone ( Figure 5) based on aerial photographs (c.f. Baroni et al., 2004;Colucci et al., 2016;Kellerer-Pirklbauer et al., 2012;Krainer & Ribis, 2012;Nyenhuis et al., 2005;Onaca et al., 2017;Scotti et al., 2013). In the lower part, the interface between the rock glacier and its surroundings was in most cases clearly discernible. On the contrary, a distinction between the rock-glacier rooting zone and the above-lying contributing area was often much more difficult to differentiate (c.f. Colucci et al., 2016;Krainer & Ribis, 2012). In such cases, the boundary was identified by both visual inspection of aerial images and searching for a sudden change in the DEM-derived slope inclination. Places with gradient exceeding the angle of repose of c. 35° (Barsch, 1996) were implicitly considered as parts of the contributing area, probably corresponding to debris-free rock slopes or rock walls (sensu Gądek, Grabiec, Kędzia, & Rączkowska, 2016). The layers of slope inclination and shaded relief were also employed to refine the delineation of some large rock-glacier outlines. In cases where more rock-glacier bodies merge to form socalled multipart rock glacier (Barsch, 1996), we delineated each of these distinguishable parts individually (c.f. Falaschi, Castro, Masiokas, Tadono, & Ahumada, 2014;Kellerer-Pirklbauer et al., 2012).

Rock-glacier classification
After completion of the rock-glacier mapping, we classified rock glaciers according to the main source of material, permafrost presence and activity, and bedrock lithology.
Two rock-glacier categories, talus rock glaciers and debris rock glaciers (Barsch, 1996; Figure 4), were adopted to differentiate the landforms based on the main source of material. As talus rock glaciers are regarded those landforms, which are unambiguously related to talus slopes situated above, delivering frostshattered rock fragments for rock-glacier formation, and do not significantly extend down to the bottom of the valley. By contrast, debris rock glaciers are usually more extensive debris deposits, mostly covering valley floors, which evolve particularly from morainederived materials. We also include the most distinct landforms referred to as protalus (pronival) ramparts into the first category for two main reasons. Firstly, these landforms are usually very difficult to differentiate from talus rock glaciers not only from aerial imagery, but also in the field (e.g. Hedding, 2016;Hedding & Sumner, 2013). Secondly, these geomorphic features are largely genetically related to talus rock glaciers and many researchers describe them as embryonic stage of rock glaciers, that is, a kind of developmental continuum, closely resembling them in morphology, internal structure, ice content and morphodynamics (e.g. Barsch, 1996;Haeberli, 1985;Scapozza, 2015).
Intact and relict classes were set to differentiate the rock glaciers according to the presence of permafrost and activity (Barsch, 1996). Intact rock glaciers (including both active and inactive landforms) host permafrost inside and may (active rock glaciers) or may not (inactive rock glaciers) move due to permafrost creep. In contrast, permafrost has completely thawed within relict rock glaciers. Borehole drilling, ground temperature measurements, geophysical soundings and/or monitoring of rock-glacier movements are particularly helpful to distinguish between the intact and relict rock glaciers (e.g. Haeberli et al., 2006). Some of the above methods have been employed in the High Tatra Mts., suggesting probable permafrost presence in several rock glaciers (Kędzia, 2014), but the data are available for a limited number of landforms (Dobiński, 1997b;Gądek & Kędzia, 2008, 2009Kędzia et al., 2004;. Consequently, the classification is largely based on morphological attributes discernible on aerial imagery that are widely accepted to be indicative of ground ice presence or absence and were adopted in many recent rock-glacier inventories as well (e.g. Colucci et al., 2016;Falaschi et al., 2014;Onaca et al., 2017;Scotti et al., 2013). Intact rock glaciers are characterized by steep frontal and lateral slopes, which often exceed the angle of repose of c. 35°, exhibit well-defined ridge-and-furrow topography, and completely lack or have sparse vegetation cover. In contrast, relict rock glaciers are typical of subdued topography with gentler slopes caused by permafrost degradation, and usually have extensive vegetation cover.
The bedrock lithology of individual rock glaciers was inferred from the schematic geological map (see section 3.2.3; Figure 2) based on rock-glacier outlines. In total, three geological units were identified for the mapped rock glaciers: Palaeozoic gneiss rocks, migmatite and amphibolite; Carboniferous-Permian granitic rocks; and Mesozoic limestone, dolomite, sandstone, shale and quartzite.

Rock-glacier contributing areas
Contributing area of each rock glacier, that is, the zone where source material for its formation is collected ( Figure 5), was computed using the Hydrology tools (e.g. Bolch & Gorbunov, 2014) implemented in ESRI ArcMap 10.3 based on the void-filled DEM (see section 3.2.1.). Because this debris-contributing area (sensu Janke & Frauenfelder, 2008) may not be identical to the hydrological catchment of a rock glacier (particularly in the case of debris rock glaciers), only those parts of the respective rock glaciers, which presumably received material directly from the source zone (i.e. not secondarily through the rock-glacier flow), were used as the lower modelling limit. Several lines of evidence, such as the orientation of ridges and furrows and the connection between the outer edge of a rock glacier and neighbouring slope, were considered to determine these rock-glacier parts. The delineated contributing areas are essentially based on calculations confined to locations outside the rock glaciers, and therefore no problems that could be expected on rugged rock-glacier topography, such as flow divergence or presence of sinks with undefined flow direction, have been encountered. The contributing areas of lower-lying rock glaciers forming parts of multipart rock glaciers were set to include both the area of all the above-lying and genetically related rock glaciers and their contributing areas.

Mapping results and interpretation
The inventory contains a total of 383 rock glaciers occupying the area of 13.84 km 2 , which are supplied by rock material from 51.81 km 2 of contributing areas (Table 1). Hence, the total rock-glacier-affected area in both mountain ranges constitutes 65.65 km 2 , which comprises c. 16 % of the area above 1375 m asl (∼lower limit of rock-glacier occurrence). The average rock-glacier density above this limit is 0.93 landforms per km 2 and the average specific rock-glacier area accounts for 3.34 ha per km 2 . Less rock glaciers, 183 (c. 48 %), occur in the Western Tatra Mts. than in the High Tatra Mts. where 200 rock glaciers (c. 52 %) are located. However, this does not translate into the difference in the total rock-glacier area, which is by 0.45 km 2 larger in the Western Tatra Mts. On the other hand, the total contributing area is by 2.60 km 2 more extensive in the High Tatra Mts. (Table 1).
Talus rock glaciers predominate in both mountain ranges, with c. 63 % and c. 74 % in the Western Tatra and High Tatra Mts., respectively, but their total area represents only c. 25 % and c. 42 % of all the rock glaciers. Accordingly, debris rock glaciers are substantially larger in size, which is on average around four to five times the average size of talus rock glaciers (Table 1). Debris rock glaciers also have substantially larger contributing areas (Table 1), which are capable to supply these voluminous landforms with a sufficient amount of material necessary for their development. Most rock glaciers are considered as relict; only seven landforms were classified as intact in the Western Tatra Mts. (c. 4 %) and other forty nine in the High Tatra Mts. (c. 25 %), representing c. 15 % of the total number of rock glaciers in the entire study region and covering the total area of 1.34 km 2 (Table 1). Relict landforms extend down to 1375 m asl and 1414 m asl in the Western and High Tatra Mts., respectively, with the average front elevation of 1644 ± 119 m asl and 1731 ± 143 m asl, respectively ( Figure 6). Lower limit of intact rock glaciers occurs at 1761 m asl and 1831 m asl in the Western and High Tatra Mts., respectively, and their front elevation averages 1812 ± 30 m asl and 2011 ± 92 m asl, respectively. The discrepancy of nearly 200 m between the mountain ranges between the mountain ranges in the latter parameter results from wider elevation range where intact rock glaciers occur in the High Tatra Mts., and also from the limited number of intact landforms in the Western Tatra Mts. Consistently with the distribution patterns observed elsewhere (see Barsch, 1996), both relict and intact rock glaciers have their lower limits located higher in southern aspects than in northerly exposed places. In the Western Tatra Mts., the average front elevation of relict and intact rock glaciers facing the NW-NE directions is at 1599 ± 107 m asl and 1823 ± 37 m asl, respectively, while landforms oriented to the SW-SE are at 1723 ± 103 m asl and 1819 ± 18 m asl, respectively. In the High Tatra Mts., the respective elevations are 1692 ± 159 m asl and 1943 ± 63 m asl for the NW-NE sectors and 1768 ± 116 m asl and 2018 ± 88 m asl for the SW-SE. Intact rock glaciers are mostly of talus type, with c. 86 % and c. 78 % in the Western and High Tatra Mts., respectively, and have on average around one and a half to two times smaller size than relict landforms (Table 1).
Rock glaciers occur in three different geological units (Table 2), which represent 100 % and c. 97 % of the area above 1375 m asl in the Western and High Tatra Mts., respectively. The total numbers and areas of rock glaciers built by different substrates well correlate with the areal proportion of these materials in the study region. Granitic rocks predominate here (Figure 2), and therefore c. 65 % and 96 % of rock glaciers, covering the total area of 4.84 km 2 and 6.46 km 2 , are formed within these substrates in the Western and High Tatra Mts., respectively (Table 2). More diverse geology of the Western Tatra Mts. promotes c. 27 % of rock glaciers (2.00 km 2 ) consisting of Palaeozoic metamorphic rocks (particularly gneisses and migmatites) and c. 8 % of the landforms (0.31 km 2 ) developed within Mesozoic   Dobiński (1997aDobiński ( , 2004Dobiński ( , 2005. limestones, dolomites, sandstones, shales and quartzites. On the contrary, only 1 % and 3 % of rock glaciers (0.07 km 2 and 0.17 km 2 ) in the High Tatra Mts. are formed by these materials (Table 2). In total, the nongranitic rock glaciers cover the area of 2.31 km 2 and 0.24 km 2 in the Western and High Tatra Mts., respectively. Rock glaciers consisting of Mesozoic limestones, dolomites, sandstones, shales and quartzites have the smallest average size and also show the lowest rock-glacier density as well as the lowest specific rock-glacier area in both mountain ranges (Table 2), which suggests that these materials are less favourable for rock-glacier formation in the study region. In contrast, rock glaciers built by granitic rocks and Palaeozoic metamorphic rocks are substantially larger and occur more frequently within the hosting geological units (Table 2). Their average sizes are almost identical in the individual mountain ranges, but in total the latter landforms are slightly larger, which is associated with both comparatively greater abundance of Palaeozoic metamorphic rocks in the Western Tatra Mts. and more extensive rock glaciers occurring there. On the other hand, the average rockglacier density in the entire study region is distinctly the highest for granitic rocks, 1.12 landforms per km 2 , as is the average specific rock-glacier area, which in these substrates equals 4.07 ha per km 2 .
Intact rock glaciers can be utilized to approximate the contemporary lower limit of discontinuous permafrost on a regional scale, and relict rock glaciers can indicate the variations in discontinuous permafrost extent in the past (e.g. Barsch, 1996;Haeberli et al., 2006). The average front elevation of intact rock glaciers calculated collectively for both mountain ranges is at 1986 ± 109 m asl, which on average represents the area of 60.45 km 2 above this level in the entire study region. The average value of 1986 m asl is 56 m above the previously proposed average discontinuous permafrost boundary of 1930 m asl (Dobiński, 1997a(Dobiński, , 2004(Dobiński, , 2005. Undoubtedly, the difference is principally affected by distinct methodologies. The former investigations (Dobiński, 1997a(Dobiński, , 2004(Dobiński, , 2005 assessed the permafrost occurrence in principle via the elevation of zero isotherm of air temperature, while the present study builds on the visual inspection of rock-glacier activity based on aerial imagery. Active rock glaciers typically exist in areas where MAAT is −2°C or less (Barsch, 1996). However, internal ice core may persist inside inactive rock glaciers even under positive MAAT because coarse blocky materials of rock glaciers tend to be substantially colder than outside air (e.g. Gorbunov, Marchenko, & Seversky, 2004;Harris & Pedersen, 1998). Because inactive rock glaciers dominate the intact category in the Western and High Tatra Mts. and the cooling effect has been extensively observed here as well , the average elevation of the intact rock-glacier front and the zero isotherm of air temperature should likely be reversed. This issue may be attributed to climate warming because the earlier estimates of permafrost distribution (Dobiński, 1997a(Dobiński, , 2004(Dobiński, , 2005 were mostly based on air temperature series two to three decades older (from 1985-1989 or 1985-1994) than our aerial images. The air temperature at Kasprowy Wierch (1991 m asl), located almost exactly in the middle of the study region, has been increasing on average by 0.02°C per year (Żmudzka, 2011) during 1966-2006 and has continued to rise at least until 2010 (Gądek & Leszkiewicz, 2012). Such a warming rate would elevate the zero isotherm level by 73-109 m in 20-30 years, assuming the average temperature lapse rate of 0.0055°C per m that is based on 1951-1970 data from the eight highest elevated stations in the Western and High Tatra Mts. (Niedzwiedz, 1992). In that case, the zero isotherm of air temperature would be at 2003-2039 m asl, which much better fits the presumed relation between air temperature and distribution of mostly inactive rock glaciers. In addition, the contemporary zero isotherm level may lie even higher because the warming apparently accelerated in the eighties of the last century (see e.g. Gądek & Leszkiewicz, 2012;Żmudzka, 2011).
Because the lowest fronts of relict rock glaciers descend to around 1400 m asl (Figure 6), the lower boundary of discontinuous permafrost in the Western and High Tatra Mts. in the Late Glacial or the early Holocene probably occurred around this level, provided that the rock glaciers could fully develop under the given climate conditions. In that case, discontinuous permafrost would occupy the area of c. 393 km 2 in the entire study region, that is, around six and a half times more than the estimated contemporary discontinuous permafrost extent, and the MAAT at 1400 m asl could be −2°C or less if we accept this value to be valid for active rock glaciers to develop (Barsch, 1996). The contemporary MAAT at this elevation is estimated to be +3.4°C based on the 1981-2010 temperature record from Lomnický štít Peak and the present average lapse rate of 0.0055°C per m. Consequently, the temperature at the time of the rock-glacier formation was probably at least 5.4°C lower than at present.
The spread between the average front elevation of intact and relict rock glaciers in the Western and High Tatra Mts. is well within the ranges observed in most surrounding regions, but local rock glaciers occur on average about 400-600 m lower than in the European Alps (Table 3). The decline in the front elevation is similar to that reported from some of the easternmost sub-regions of the European Alps and the Southern Carpathians (Table 3), which has been attributed to less precipitation towards the east (e.g. Onaca et al., 2017), causing thinner snow cover during winter, and thus lower ground temperatures (sensu Gruber & Haeberli, 2009). About 100 m lower average front elevation of the intact rock glaciers in the Western and High Tatra Mts. than in the Southern Carpathians is most likely due to latitudinal temperature decrease, and is close to the previously reported difference in the contemporary lower limit of discontinuous permafrost of 70 m between these two Carpathian regions (Dobiński, 2005).

Conclusions
The map includes a total of 383 rock glaciers, making it the most comprehensive rock-glacier inventory for the entire area of the Western and High Tatra Mts. to date. Relict rock glaciers account for 85 % of the database and cover the total area of 12.50 km 2 . These landforms have their lowest limit around 1400 m asl, and thus the lower boundary of discontinuous permafrost zone in the Late Glacial or the early Holocene probably lied at this level, and covered the total area of around c. 393 km 2 . Intact rock glaciers constitute 15 % of the inventory and cover the total area of 1.34 km 2 . Their average front elevation of 1986 ± 109 m asl delineates the contemporary lower limit of discontinuous permafrost zone on a regional scale that occupies the total area of 60.45 km 2 . The rock-glacier inventory adds to the current state of knowledge about the occurrence of these permafrost landforms in less investigated high-mountain regions located east of the European Alps. However, it must be emphasized that the inferred permafrost limits and extents should be understood as rather tentative in nature. Rock glaciers can provide a first-order evaluation of potential permafrost distribution, which generally tends to overestimate the permafrost extent at places without debris cover. Consequently, the present work is rather a starting point towards more thorough analyses of rock-glacier distribution and morphology and modelling of discontinuous permafrost distribution, which will substantially improve the understanding of present and past environmental conditions in the Western and High Tatra Mts.

Software
The mapping, digitizing, analyses of DEM and map compilation were all carried out using ESRI ArcMap 10.3. Final merger of vector and raster map elements was done in Adobe Acrobat Pro DC.

Data availability
The shapefiles of rock-glacier outlines and contributing areas are available on request by the authors. Data users are requested to inform the data owners about the planned activities and invite them to contribute to any work that would lead to a co-authorship.