Geomorphology of Abandoned Underground Slate Mines in Czechia

We conducted one of the first complex geomorphological investigations of underground abandoned slate mines in Czechia, Central Europe. Sixty types of subterranean geomorphic forms were described from 114 abandoned underground slate mines in the Nízký Jeseník Upland. The forms, evaluated according to their type, scale, frequency, and fragility, are independent of each other or embedded. Their cascades occur in large multilevel mines affected by fluvial and gravitational processes. Analysis of the microforms revealed mine pearls enriched with titanium, calcareous soda-straw stalactites, amorphous calcareous-sulphurous piles, and quartz veins. Forty-seven of the identified mines were affected by gravitational failures. Four-year measurements of cracks revealed both negligible oscillations and distinct displacement of rock blocks up to 4 mm. The range of described geomorphic forms together with identified cascades and measured crack displacement point to rich geodiversity and high dynamics within the abandoned mines. Compared to other subterranean geosystems, the geomorphology of abandoned underground slate mines remains poorly investigated, although knowledge of subterranean geomorphic systems is crucial for future natural evaluation, risk assessment, remediation, or tourist management of abandoned mining sites. The recent results are applicable to other types of abandoned mines in the world. However, we should expect new findings and investigation of as yet undescribed geomorphic forms.


Introduction
During recent decades, some segments of world landscapes have been deeply investigated and described from a geomorphological point of view. Recent geomorphic research has often been driven by the growing conflicts between social activities and the natural functioning of geosystems. This is the case, for example, of the river systems-broadly devastated by land exploitation (Wohl 2014;Galia et al. 2016).
An important part of the landscape is its subterranean component, i.e., cavities of both natural and artificial origin. These are historically the most explored in karst areas (Kaufmann 2009;De Waele et al. 2009;Gutiérrez et al. 2014;Malagò et al. 2016;Parise 2022), which follows the need to understand the complexity and specifics of the landscapes covering an estimated 20% of the world's land surface and 20% of the world's drinking water sources (Ford and Williams 2007;Goldscheider et al. 2020;Olarinoye et al. 2020). The geomorphology of nonkarst caves has been less explored and has accelerated in recent years. For example, Mörner and Sjöberg (2018) introduced the concept of fracture caves occurring in crystalline bedrock and their evolution in Sweden. Urban and Margielewski (2013) proposed the classification of gravitationally induced nonkarst caves based on examples from Poland. Lenart (2015) described the morphological patterns of crevice-type caves in Czechia. Similar research has been conducted on glacier (Holmgren et al. 2017) and lava caves (Kempe 2012).
A specific but widespread phenomenon of the subterranean component of landscapes is abandoned underground mines, which are mines for which mining leases or titles no longer exist and for which the responsibility for their rehabilitation cannot be allocated to any individual, company, or organization that has undertaken mining activities (Unger 1 3 2017). These mines belong to typology E in the classification of artificial cavities adopted by the Commission on Artificial Cavities of the International Union of Speleology Parise 2015a).
According to UNEP and COCHILCO (2001), there are more than one million abandoned mine lands worldwide. Nevertheless, abandoned underground mines have been rarely studied from the geomorphological point of view thus far and pose a scientific challenge (Schuchová and Lenart 2020). They have been established mainly outside human settlements, but later urban expansion brought houses and infrastructure above them (Parise 2010(Parise , 2015aLollino et al. 2013), and many are still preserved within different types of landscapes worldwide (López-García et al. 2011), evolving partly or completely under natural conditions. While they were undergoing continuous processes of renaturalization, including weathering, evolution of hydrological regimes, or collapsing, human activities have expanded to occupy the formerly wild landscape and have experienced conflicts due to abandoned mines (Mossa and James 2013;Lenart 2016;Schuchová and Lenart 2020). From the perspective of increasingly overcrowded landscapes, abandoned underground mines are becoming an important geomorphic phenomenon.
This paper presents the results of one of the first complex geomorphological investigations of underground abandoned slate mines. It proposes an overview of the respective geomorphic forms, describes their context in terms of geological and geomorphological processes, and classifies their origin and evolution from both spatial and temporal standpoints. For this study, the area of the Nízký Jeseník Upland in Czechia (central Europe) was chosen due to the good preservation and natural character of the underground mines.

Slate and Local Mining History
Slate has been traditionally excavated worldwide, mainly as a roofing material. The European Committee of Standardization (CEN 1997) published a definition for roofing slates as rocks that are easily split into thin sheets along planes of slate cleavage, caused by very low to low-grade metamorphism due to tectonic compression. The mineral composition includes quartz, mica, and chlorites, assembled in a lepidoblastic texture (Ward and Gómez-Fernández 2004). The typical dark-grey colour is caused by the organic matter content (Borradaile et al. 1991). The quality of slate is influenced by several factors: (i) the mineral composition, (ii) grade of metamorphism, and (iii) weathering (Lombardero et al. 2002;Cárdenes et al. 2014). The aesthetic quality is typically reduced by iron sulphide oxidation (Cárdenes Van den Eynde et al. 2009). Gypsification leads to a softening and destruction of the slate and decreases the resistance to weathering (Cárdenes et al. 2012).
The underground excavation of cleavable slate of sedimentary origin requires a special system of underground quarrying methods (Jirásek et al. 2019). There is a considerable lack of scientific publications investigating slate quarries and mines, even though slate roofing is the bearer of a long tradition, as it has been mined in Europe since the Roman period (Cárdenes et al. 2014). Focusing on central Europe, the Nízký Jeseník Upland reaches an area of ~ 2800 km 2 , situated on the northeastern margin of the Bohemian Massif ( Fig. 1; Demek et al. 2006).
The Nízký Jeseník Upland belongs to central European Variscides and is interpreted as an accretionary wedge composed of deep marine flysch sediments represented predominantly by Lower Carboniferous slate and greywacke (Kalvoda et al. 2008;Botor et al. 2017). The four rock formations become younger towards the east (Fig. 1). During the Carboniferous-Cretaceous period, the territory underwent deep weathering and planation, resulting in the formation of a flattened plateau with a mean altitude of ~ 480 m a.s.l. The upland is bounded and dissected by distinct fault lines in the NW-SE and NE-SW directions (Kalvoda et al. 2008) with deeply incised valleys along them. During the Pleistocene, the area was affected by periglacial processes at the forefront of Scandinavian glaciers (Lenart 2016).
The upland was colonized during the twelfth and thirteenth centuries mostly by Germans (Vencl and Strohalm 2005). Apart from agriculture, forestry, and milling, silver ore mining was established at the beginning of the thirteenth century. Vlach (1967) mentioned that Pb(Zn)-and Cu-Pb-Ag-rich metals were mined from greywackes (Novák 1978). Kretschmer (1917) was interested in iron ore resources within the area.
The oldest written mention of slate mining comes from 1776 (Němec 1949). The extensive mining of slate for roofing purposes began after King Joseph II introduced the fire patent, which prevented the use of wood for such purposes (Malach 2016).
Because slate has never been registered locally as a reserved mineral deposit under the mining law, anyone was allowed to excavate it. At the very beginning, the mines (described also as underground quarries by, e.g. Lollino et al. 2013;Jirásek et al. 2019) were owned by local farmers and later became the property of German immigrants who were professionally engaged in slate mining (Jahn 1917;Strohalm 2003). In 1880, fifty-two companies processing slate were operating in the area (Wagner and Staněk 1991). During the nineteenth century, more than a hundred quarries and mines were in operation (Jahn 1917). Hrušecký (1946) documented seventy-two small mines before the First World War.
During the Industrial Revolution of 1860-1914, the slate mining boom in Nízký Jeseník was able to supply the whole Austro-Hungarian Empire (Strohalm 2003). The postwar disintegration of the empire led to a mining decline, and consequently, slate from England and France took over the market (Strohalm 2003). The renaissance of mining came during the 1930s, when the area was part of the Czechoslovak Republic. An almost final stroke for the slate mining industry within the region came with the expulsion of Germans after the Second World War. The countryside lost generations of owners, mining experts, and qualified workers. Many skills, methods, and documents were lost along with them, and many mines remained abandoned. Those in operation were confiscated during the process of nationalization that took place in the 1950s (Labus et al. 2020). Slate mining never reached its former glory, and after the democratic revolution in 1989, it finally ceased.

Typical Structure of Slate Mines
Historically, the main excavating method in the broader Nízký Jeseník slate mining region was the room-pillar method with backfilling, known as the Rhenish mining method, which resulted in a system of underground rooms established parallel to the heading adit. Between two adjacent mining rooms, the protective pillar was left. Although appropriate cleavable layers were found under various inclinations, the most suitable for excavation was the nearly vertical layer. Under such conditions, the dimensions of one mining room were up to 25 m in length, 12 m in width, and high (Grygárek 2014;Wichert 2020).
Tailings were deposited on extensive spoil heaps outside the mines or, predominantly, were stored inside the excavated rooms, forming vast dry-walled backfill. These were important for several purposes: (i) they played a stabilization role. In the case of rooms excavated in subhorizontal slate layers, the backfill supported the ceiling and prevented collapse of the rooms and areas below passing headings or cross-cuts. (ii) As the excavation progressed upwards along the rock layers, it enabled miners to climb to their workplace up ladders through the chimneys preserved within the backfill mass and, standing on the top of backfilled material, to continue excavating upwards. (iii) It enabled preservation of adits with multiple functions (transport, drainage, ventilation) below them and chimneys inside them by enclosing the space with backfilled material. (iv) As the spoil heap  Kalvoda et al. 2008) and their approximate distribution according to topographical settings outside the mine was too far from mining rooms, storing of the material underground enabled easier and faster work. (v) It enabled continuous work throughout the year, even during severe weather conditions.
In general, the abandoned underground slate mines (AUSMs) of the Nízký Jeseník Upland are situated at shallow depths up to ~ 100 m below the surface, and rooms and adits were mined under an open and unsecured ceiling. Thus, fractures or tectonics may induce the collapses (Grygárek 2014). Many mines are preserved only as short exploratory adits or small mines, with one mining room reaching an overall length of less than a dozen metres (Lenart 2016). The arrangement of medium and large mines tends to be more complicated. Many of them have more than one entrance combining adits and shafts opened at different elevations. Such mines are usually branched into two connected levels reaching overall lengths from hundreds of metres to kilometres and a depth of dozens of metres below the surface (Lenart 2016). In the case of rare multilevel mines, adits in the lowest level are often covered fully within the backfill with arched ceilings or represent the remaining gap between inclined slate layers and the supporting backfilled walls (Fig. 2).

Methodology
The field survey was performed from 2016 to 2020. In the area of the Nízký Jeseník Upland, 114 sites with AUSMs were investigated ( Table 1). Most of the mines were located using the Czech Geological Survey database, and thirtythree were located following descriptions by the locals. The speleological maps of the mines were collected from the Czech Speleological Society, local miners, and members of informal Czech tramping communities, who have visited and mapped the mines since the 1940s as part of their adventurous journeys (Kupka and Pohunek 2017). Each of the evidenced mines was localized, and its interior was inspected. Existing maps were revised and corrected. In addition, more than a hundred new maps were created using methods of speleological mapping with compass and tape or a system based on a Leica Disto A3 range finder with a Disto-X component and PocketTopo software (Heeb 2008). The maps were printed, and the approximate length of the AUSMs was measured using a ruler. According to their overall length, the AUSMs of the study area can be divided into three groups: (i) small mines (< 100 m), (ii) medium mines (101-500 m), and (iii) large mines (> 500 m).
Within the mines, the following data were registered in the field: the number of entrances and levels, occurrence of gravitational failures (Table 1), and structural data and diverse underground geomorphic forms with related processes that led to their establishment. The latter were divided into three categories according to their temporal relation to the particular historical mining event: pregenetic, syngenetic, and postgenetic. Moreover, the geomorphic forms were sorted into macro-, meso-, and microforms according to their relative size.
To record the possible recent gravitational processes in five AUSMs (the Woodboys, Peklo, Soudek, Klokočov, and Raab mines), the MGM 250 Deformometer tool (Diagnostic Research Company, Italy) was used to measure short-term movements on rock discontinuities. We installed twenty-one measuring points along vertical cracks, gaps between shifted sedimentary layers, and between presumably gravitationally unaffected walls and tilted rock blocks. The measurements were performed repeatedly on 18 April 2017, 18 April 2019, and 30 July 2020. The maximum measured distance was 255 mm, with 0.01 mm accuracy. The resulting distance was calculated as: where AB determines the displacement between two measuring points, 250 is a nominal deformation metre measurement base, Xm is the measurement base reading, and Xa is a calibration bar reading.
The accuracy of the measured distance between measuring points was determined with respect to the principle of the law of accumulation of errors. Because it was not possible to carry out independent control measurements, a standard deviation σ was used to evaluate the accuracy. The mean displacement error σ v was estimated according to the following formula: A three-fold value was used to determine the 99% confidence level. Because all of the measured values far exceed the confidence level, the displacements are proven to have a 1% risk.  To reveal the properties and possible genesis of the identified microforms, we used X-ray fluorescence (XRF) analysis. The concentrations of Mg, Al, Si, S, Ca, Ti, Mn, and Fe elements within the rock were determined using a VANTA VCR C-series (Olympus) spectrometer with a silicon drift detector and a 4-W X-ray tube with an 8-50 kV silver anode. The SoilGeoChem protocol was used. Each sample measurement was divided into two phases: 60 s measurement at 50.0 kV and 30 s at 10.0 kV. The results were averaged from three measurements of each sample. The XRF analysis was applied to investigate collected samples of mineral veins and speleothems from eight AUSMs: two different samples of mine pearls from the Karol mine (MP1-Karol, MP2-Karol), two pieces of needle-shaped and dense bundles of fragile acicular/grainy crystals from both the Soudek (CR1-Soudek) and the Modrá (CR2-Modrá) mines, a group of soda-straw stalactites from the Klokočov mine (ST1-Klokočov), two samples of amorphous material from the Carl (AM1-Carl) and Libor mines (AM2-Libor), and three samples of exposed veins obtained from the Flaschar (VN1-Flaschar), Gotická štola (VN2-Gothic), and Karol (VN3-Karol) mines. The samples were ground to a powder. The mine pearls were cut lengthwise, and the surface and core were analysed separately.

Geomorphic Forms and Processes
In Fig. 3, we present a list of the detected subterranean geomorphic forms classified according to their genesis, temporal relation to a historical mining event, frequency, and size scale. The scale was set relative to the underground space extent to microforms up to 0.1 m in diameter, mesoforms from 0.1 to 1.0 m, and macroforms with more than ~ 1 m diameter. Detailed descriptions of the forms with typical or extreme examples recognized during the investigation are listed in Supplementary Material-Atlas of geomorphic forms.

Excavation Forms
As the main bearers of the shape of the AUSMs, the excavation forms have been established during active mining as a result of human will. As they emerged at the decisive time of the active mining period, all of them are of syngenetic origin, with possible preservation after mine abandonment. Due to erosion, they gradually vanish. According to their differences in size, they may be combined with each other, e.g. room, side, and chisel marks. Adits, shafts, or rooms are primarily of rock excavation origin. Alternatively, they are partly or fully enclosed by backfill material, which, in this case, substitutes the mature rock and plays the role of human-deposited sedimentary material.
The excavation forms are listed in Fig. 3, and examples are presented in Fig. 4. For such forms, we used traditional mining terminology. An adit is a horizontal mine corridor, whereas a shaft is a vertical one. Depending on the geological conditions, inclined shafts connect the different levels. The vertical staple shaft is not open to daylight. After extensive excavation of slate material, rooms were preserved with supporting pillars between them. Rooms of smaller and rather irregular dimensions may be named stopes. For various purposes (tool storage, religious), niches were dug into rock. Raises were established to better reach the bearing outcrop, and chutes were excavated as inclined openings for slate transport. Tanks of various shapes were created for storage of the water seeping from rock joints. Some of the excavation forms emerged as a result of irregular slate extraction. Thus, rock steps (some of them of possible gravitational origin), excavation rock arches, or even rock windows were observed, representing irregularities within faces, sides, or floors of adits or rooms. Various types of construction holes, gaps for sleepers, bores, or chisel marks belong to microforms of that type.

Excavated Geological Features
During mining, former primary and secondary geological structures hidden inside the rock mass were excavated within the ceiling, face, side, or floor of adits, shafts, or rooms and remained preserved within the underground space. As they can be defined by their shape and dimensions, they act as geomorphic forms mostly at the meso-and microscales. Although excavated geological features developed during active mining, they can also be of postmining origin, as they can be exposed due to other successive processes (e.g. erosion, collapsing). Within the AUSMs, these forms are combined with other excavation forms, e.g. side with marks. The excavated geological features are listed in Fig. 3, and examples are presented in Fig. 4. For such forms, we used the traditional geological terminology. During the time of active mining, folds, beds, planes, joints or faults with slickensides, bending of beds, or tectonic breccia were found. Various types of veins, concretions, marks, and traces or even excavated fossils protruding from the rocks are rather rare geological features. Due to erosion, they gradually vanish, also because of their fragility.

Fillings
Among the syngenetic forms are those formed by the storage of backfills. Fillings are listed in Fig. 3, and examples are presented in Figs. 4 and 5. Common spoil piles are formed by disorganized spoil material within rooms or adits. The second documented form is dry-walled backfill, which occurs commonly in investigated mines. After 100 years of abandonment, formerly individual spoil clasts may have been cemented, forming modern breccia-like sedimentary rocks of anthropogenic origin. In some places, clasts are cemented at the surface by flowstones. Filling forms are usually well preserved, but they may be degraded by gravitational processes.

Fluvial Forms
Surprisingly, some of the typical fluvial geomorphic forms may be discovered within AUSMs. They are listed in Fig. 3, and examples are presented in Fig. 5.   Fig. 3 List and classification of geomorphic forms in slate mines Permanent or episodic autochthonous or allochthonous water streams play a crucial role as geomorphic agents in their presence. After dozens of years of concentrated runoff, typical channel morphology may evolve in both bedrock and fine sediments at the floor of the mine. Autochthonous or allochthonous sediment supply may result in alluvial plane or even in the formation of terraces. When episodically flooded muddy alluvium dries, mudcracks may open. As some of the adits were directly designed for drainage, they have higher inclination and may be terminated with a step within the side of the main adit. Such steps may be eroded by running water in the form of waterfall. Below it, or at positions with low transport capacity, alluvial fans may evolve. On the floor of  Lenart); F niche of filling type, Kunz mine, inclined shaft (photo J. Lenart); G tank, Nittmann mine (photo J. Lenart); H exposed fold, Flaschar mine (photo J. Wagner); I exposed marks on bedding plane; Pod Smilovem mine (photo J. Lenart) adits and rooms, water dripping from the ceiling forms shallow, rounded crater-like holes within fine sediments. Fluvial forms are postgenetic, although the initial channel morphology could have originated during the mine operation. Due to dynamic erosion, many fluvial forms are short-lived.

Fluvio-Gravitational Forms
Under extraordinary conditions, underground sinkholes may develop under the compound action of water flow erosion and gravitational collapsing. Such rare postgenetic features are among the most dynamic within the investigated mines ( Fig. 3; example in Fig. 5).

Gravitational Forms
Various types of catastrophic gravitational events may evolve in AUSMs after the termination of active mining. Thus, the resulting geomorphic forms are mainly postgenetic and are mostly of the macroscale. They are listed in Fig. 3, and examples are presented in Fig. 5. Small landslides accompanied by other typical features may evolve within the rock mass surrounding vast rooms excavated in landslideprone geological conditions. Whereas small mines are generally stable, collapses or falls of individual rock boulders or their piles commonly occur within the medium and large ones. Active landslides or collapses have the potential to completely change the geomorphic structure of the mine by (i) burying the whole mine or its portions, (ii) filling the Fig. 5 Examples of fillings, fluvio-gravitational, fluvial and gravitational forms: A backfilled adit, Lhotka mine (photo M. Kašing); B backfill, Flaschar mine (photo J. Wagner); C underground sinkhole with platform dimensions of 10 × 7 m, Soví kámen mine (photo J. Lenart); D holes after dripping with moonmilk, Gold mine (pencil for scale, photo K. Schuchová); E alluvial fan below waterfall, Nittmann mine (photo J. Lenart); F bulges, gold mine (photo K. Schuchová); G collapse, Pollak mine (photo J. Lenart) mine interior with collapsed material, and (iii) changing the direction of the underground mine water streams. In the case of vast gravitational accumulations, short talus caves may be preserved. Slow and successive gravitational deformations of the sides, ceiling, or floor may cause the opening of crevices or formation of bulges. At positions where adits or rooms are intersected by fault zones accompanied by disrupted rock or below chimneys, talus cones are formed after debris falls. Similarly, gravitationally affected dry-walled backfill may form spoil cones.

Speleothems
Within the AUSM interior, perfect conditions for supergene processes often lead to secondary mineral precipitation, forming diverse microforms with calcite, gypsum, or iron oxide contents. They are listed in Fig. 3, and examples are presented in Fig. 6. Flowstones, stalactites, stalagmites, or curtains are precipitated from Ca-, Fe-, or Mn-enriched solutions. In stream channels, rimstones may develop. In front of them or under the influence of dripping and spraying water, mine pearls are among the most spectacular microforms. Moonmilk or crusts develop under specific conditions. Gypsum or pyrite crystals precipitate on slate surfaces. Where water seeps from joints or cracks on sides or ceilings, biomineralization sheets can occur under the influence of iron-or sulphur-oxidizing bacteria. Due to weathering and erosion, many speleothems are very fragile.

Composition of Speleothems
The XRF analysis of microforms from eight AUSMs helped reveal their composition, which we describe in the "Discussion" section. Two groups of mine pearls were collected from different positions in the Karol mine (MP1-Karol, MP2-Karol). The MP1-Karol sample (Fig. 7) revealed pearls of ~ 0.2 to 1 cm in diameter, rarely up to 2.5 cm. Their shape was rather rounded and regular, and their colour was light grey to medium grey. These pearls were collected at the end of the adit, where they were lying freely on the damp floor. The core of the pearls is predominantly rich in Al, Si, S, Ca, Mn, and Fe ( Table 2). The ~ 0.5-mm-thin surface layer of these pearls was additionally enriched with Ti.
The MP2-Karol sample (Fig. 7) was composed of larger mine pearls of 1-2 cm in diameter, rarely 3 cm. Compared to the MP1-Karol sample, these pearls were uneven and irregular in shape. They appeared a dark grey colour and were collected in the middle part of the mine from the water pool with a constant seeping water supply. Similar to the MP1-Karol sample, the core and surface layers of the pearls were rich in Mg, Al, Si, S, Ca, and Fe. The ~ 1-mm-thick surface layer was also enriched with Ti. Both the MP1-Karol and MP2-Karol samples revealed high values of Ca and S.
The needle-shaped crystals from the Soudek mine (CR1-Soudek) and the dense bundles of fragile acicular/grainy crystals from the Modrá mine (CR2-Modrá) reached lengths of 0.2-2 cm (Fig. 7). The mm-wide brittle crystals are white to almost colourless. Ca and S are the main constituents of the samples (Table 2). To a lesser extent, they also contain Si, Al, and Fe.
Six soda-straw stalactites were taken from the ceiling of one of the terminating rooms of the Klokočov mine (ST1-Klokočov sample, Fig. 7). The stalactites are 1 to 3 cm long and 0.5-1.0 cm thick, and the water dripped actively from each of them. Their almost hollow inner structure and outer cover are fragile, and the colour is white to white-grey. Ca comprised the majority of their composition (Table 2). To a lesser extent, they also contained Si, S, and Fe.
Two samples of amorphous piles of lumpy, soft, and fragile whitish material from the Carl (AM1-Carl) and Libor mines (AM2-Libor) (Fig. 7) were collected from the adit floor and the side. The AM1-Carl sample revealed a high abundance of S, Ca, Al, and Fe and low abundance of Mn and Si. The AM2-Libor sample revealed higher Ca, S, Si, and Al content. Compared to AM1-Carl, the Al and Fe values were lower, but Si values were higher and there was a lack of Mn (Table 2).
Three samples of exposed hydrothermal veins from the Flaschar (VN1-Flaschar), Gotická štola (VN2-Gothic), and Karol (VN3-Karol) mines (Fig. 7) revealed high content of Si, indicating that they are of quartz composition ( Table 2). The VN1-Flaschar sample of coarse-grained whitish to pinkish material was up to 5 cm thick and rich in Si, Ca, and Ti and less abundant in Fe, Mn, and S. The VN2-Gothic sample revealed fine-grained white material taken from a ~ 10-cm-thick vein and was comprised entirely of Si. The VN3-Karol sample of fine-grained white material was taken from a 5-10-cm-thick vein and contained relatively high content of Si and lower Ti, Mn, and Fe content (Table 2).

Displacement of Cracks
Among the most dynamic gravitational processes with strong potential to change the overall structure and natural functioning of AUSMs are gravitational processes. Of the 114 surveyed mines, 47 (42%) were gravitationally affected (Table 1). Past gravitational events were recorded mostly in medium and large mines, more precisely within their rooms and adits. We also identified extensive collapses of entrances.
To investigate the rate of gravitational activity in mines, the displacement of the dilation discontinuities selected in five underground mines was measured. Detailed maps of the mines with marked measurement points are shown in Fig. 8. Their coordinates are listed in Table 1. The displacement rates are summarized in Table 3.
The 240-m-long Soudek mine has two entrances and is a medium AUSM. It was excavated in one level ~ 10 m above the tectonically predisposed valley bottom within the steep southwestern slope. The heading adit is oriented perpendicularly to the direction of the valley. Along the heading adit, several side portions have undergone gravitational events. Some segments of backfill are disrupted by subvertical cracks (SOU3). The termination room was affected by the vast collapse of rock blocks several metres long. The measured displacement on various types of cracks revealed the second highest displacement values within all investigated mines. More than 3 mm of total displacement was revealed within the termination collapse (SOU4), where tilted and fallen rock blocks have gradually separated from the side and ceiling.
The Klokočov mine is 201 m long and is a medium AUSM. The interior of its only level was opened by two adits (one recently collapsed) and by an inclined shaft. All entrances are situated within the steep southern slope of the tectonically controlled tributary valley. The heading adit is oriented perpendicularly to the direction of the valley. The sides of the heading adit are gravitationally affected in several sections, revealing the highest displacement rates among the investigated mines. More than 4 mm of widening of the vertical bedding joint in the entrance section (KLO1) was probably affected by frost weathering, which occurs there due to strong ventilation during the cold season. Within the middle section of the mine, distinct movement resulted in the impossibility of a third measurement, as the measuring point was out of range after two years (KLO2). The KLO3 measuring point was fully destroyed in 2019. Within the termination part of the mine, the widening of the inclined crevice revealed toppling movement (KLO4).
The Peklo 2 mine is a small AUSM formed by a 13-m-long prospecting adit that is opened within the midsection of a steep and straight tectonically predisposed slope, ~ 70 m above the Odra River, which has been Table 2 The results of the XRF analysis (values in ppm) < LOD, low of detection The Raab mine is a medium AUSM composed of two levels connected through one chimney, with an overall length of ~ 150 m. The entrance cross-cut adit is opened ~ 20 m above the tributary valley bottom within the steep southeastern slope. The cross-cut adit is oriented perpendicular to the valley, whereas the room of the upper level is oriented parallel to it. The lower level is relatively stable, whereas the upper level is successively collapsed, as recognized from fresh rock falls and relaxed beds. The displacement measurement results reveal both the contracting and stretching of discontinuities. In the upper level, the RAA1 horizontal bedding joint contracted 0.3 mm, which was probably caused by irregular subsidence of overburden layers.
The Woodboys mine is a 470-m-long medium AUSM with one stepped level and two former entrances (one recently filled). The accessible entrance is situated ~ 15 m above the tectonically predisposed valley within the northeastern slope. The heading adit is oriented perpendicularly to the orientation of the valley. As one of the distinct gravitationally affected sections, the tectonically disrupted ceiling collapsed 20 m from the entrance. The mine structure now consists of eight rooms with lengths from 15 to 80 m. Their sides, formed by nearly vertically oriented layers, are affected by rock falls and toppling of blocks (W1, W2).

Geomorphic Forms in Underground Geosystems
Subterranean geomorphic forms were widely described in environments formed by natural processes without direct human influence. Unlike caves, abandoned underground mines have been studied primarily from the perspective of gravitational failures (Ferrero et al. 2010;Parise 2010Parise , 2013Parise , 2015bParise , 2016Sunwoo et al. 2010;Lollino et al. 2013;Luan et al. 2018 and others). Other published works have mostly focused on mature rock and hydrothermal vein composition analyses (Antunes et al. 2002;Stouraiti et al. 2019) or discoveries of new minerals (Holtstam et al. 2017;Pekov et al. 2018;Zaccarini et al. 2020). However, in recent years, abandoned underground mines have played a more important role in nature conservation as habitat for bats (López-González and Torres-Morales 2004; Kurta and Smith 2014), for preservation of cultural heritage (Mendes 2018) or tourist management (Różycki and Dryglas 2017). Thus, they remain insufficiently investigated landforms, which lead to an underestimation of their importance within the world's landscapes (Schuchová and Lenart 2020). Moreover, the complex Fig. 8 Maps of selected mines with positions of displacement measurements on cracks and structural parameters of rock layers spectrum of geomorphic forms of abandoned underground mines has never been described.
From the risk assessment point of view, abandoned underground mines may cause threats to human infrastructure (Canakci 2007;Parise 2015a;Martínez-Moreno et al. 2016), health (Coelho et al. 2007), and the environment (Longoni et al. 2016;Parise 2016). As geomorphic processes play an important role within the underground mine interior environment, knowledge of geomorphic forms is crucial for further evaluation of such mines in the future. We identified sixty underground geomorphic forms within 114 AUSMs, most of which were created in large multilevel mines. Their diversity is comparable with that of other underground geosystems, such as karst, gravitationally induced caves, or lava caves. The geomorphology of karst caves has been deeply investigated by many authors (e.g. Ford and Williams 2007;Parise and Lollino 2011). Researchers have mainly focused on describing karst cave forms (Hill and Forti 2007;Nader 2007;Onac and Forti 2011) or modelling underground geomorphology (Kaufmann 2009). De Waele et al. (2011) or Parise (2015a summarized natural hazards related to karst environment. The geomorphological research has also focused on noncarbonate caves. Various authors have investigated the problems of the evolution of specific geomorphic forms in arenitic caves of South American tepuis (Doerr 1999;Aubrecht et al. 2011Aubrecht et al. , 2017Sauro 2014), Spanish piping caves (Bartolomé et al. 2015), Swedish fracture caves (Mörner and Sjöberg 2018), or Italian quartzite caves .
Gravitationally induced caves have been intensively investigated in recent decades using modern scientific methods. After their initial description (Kunský 1957;Vítek 1980), Margielewski and Urban (2017) presented their modern morphogenetic classification. The morphology of crevice-type caves in flysch Carpathians was summarized by Lenart (2015). When such caves evolve within rather uniform geological conditions and occupy relatively small areas, the diversity of their underground forms decreases.
The geomorphology of various types of volcanic caves and their interiors has been described (e.g. Gadanyi 2008; Phuc et al. 2018) with the aim of possible further investigation of similar objects on the Moon (Hong et al. 2014) or Mars (Léveillé and Datta 2010;Sam et al. 2020).
The geomorphology of various types of artificial underground has been recently investigated, including underground quarries (Van Den Eeckhaut et al. 2007;Parise and Lollino 2011) and urban tunnels (Melelli et al. 2021).
From the geomorphic point of view, only the investigations of karst caves and volcanic cave geosystems are comparable to that of other highly surveyed superficial geosystems, such as fluvial geosystems (e.g. Stott 2010;Miklín and Galia 2017).

Embedded Forms and Cascades
Based on our investigation, the almost full spectrum of geomorphic forms listed in Fig. 3 is developed in medium and large mines. This amount is predisposed by (i) the high spatial extent of diverse underground excavation forms (adits, rooms), (ii) the high number of interconnected levels intersecting diverse geological conditions, (iii) differently situated entrances enabling ventilation, and (iv) the developed system of underground water drainage with subsequent sediment transport. Combinations of these factors lead to a high diversity of morphogenetic processes, namely, fluvial, gravitational, or frost weathering (see also Lollino et al. 2013;Parise 2015a). Both the morphogenetic processes and resulting forms are typically embedded and/or cascaded (Fig. 9).
Among the main elements of the mine structure are excavation forms, which are passively shaped by minor excavation forms or geological features. The passive fillings stand at the boundary between the geomorphic forms and human constructions; in the case of abandoned mines, they act as well-sorted and even cemented sediments. The other embedded forms are created dynamically and successively within the passive ones, forming a typical hierarchy of a hosting excavation form and gravitational or fluvial secondary forms (e.g. room -collapse -talus cave -stalactites; Fig. 9). These sequences are the result of the nonlinear evolution of subterranean geomorphic systems with dynamic instability (Phillips 2006). According to these principles, the stable nonchaotic system of excavation forms and excavated geological features passes under the influence of thresholds into an unstable chaotic system with embedded fluvial, gravitational forms, and speleothems.
Moreover, embedded subterranean forms are commonly arranged in the form of catena (Scheidegger 1973). In Fig. 9, a complex fluvial and gravitational subterranean catena with segments of slow-fast-slow motion is depicted.
If we consider the relationship of the mine geomorphic forms to the surrounding landscape, the most important section is the mine entrance (adit or shaft). This is usually the only part of the mine directly embedded within the other superficial landform (foot, slope, plateau; Fig. 1). The entrance portions of the AUSMs were strongly affected by the collision of both the external and internal air masses, leading to wetting, frost weathering, fast weathering, erosion, and gradual collapsing. It follows that the only connecting section between the superficial landforms and mine landforms is concurrently the least resistant. This corresponds with the theory of thresholds in nonlinear geomorphic systems (Phillips 2006), where "small changes may produce disproportionately large results" and "short disturbances may have disproportionately long-lived effects".

Minothems: Fragile Jewels
Speleothems are well investigated within abandoned underground mines. As they are formed under conditions similar to speleothems in natural cavities (Hill and Forti 1997) but in an artificial underground, Carbone et al. (2016) introduced the term minothems for them. In recent years, the number of publications focused on the development of minothems in abandoned or active mines has increased (Mackin and Coombs 1945;Jébrak et al. 1996;Rowe et al. 2007;Ziegler et al. 2009;Sawłowicz et al. 2014;Carbone et al. 2016;Cecchi et al. 2018;Broughton 2020;Falteisek et al. 2020;Manaka 2021;Jelenová et al. 2021;Melim and Spilde 2021).
Our research on minothems also revealed some consequences. Metal elements detected in the cores of mine pearls (MP1-Karol, MP2-Karol) may indicate that their initial grains are slate particles composed of quartz, feldspar, clay minerals (illite, muscovite, chloriteoids), or sulphides. Both samples revealed high amounts of Ca and S, indicating the presence of carbonate or gypsum, which substantially contributes to their composition. The Ti content in their surface layer may indicate the crystallization of secondary minerals such as ilmenite or rutile. The needle-shaped crystals (CR1-Soudek) and dense bundles of fragile acicular/grainy crystals (CR2-Modrá) are gypsum crystals precipitated from weathering of sulphides (Wichert 2020).

How Long Will Slate Mines Survive?
In underground environments, various authors have investigated gravitational failures (Slaker et al. 2017;Li et al. 2019). However, abandoned underground mines are sometimes considered preserved, untouched, and long-term isolated (e.g. Isaia et al. 2011). With the exception of precursors of the future collapse (Parise 2013(Parise , 2015aSwedzicki 2001), they usually do not reveal any apparent changes within their structure. However, we classified forty-five mines out of 114 recorded as gravitationally affected (Table 1). Short prospecting adits were mostly stable, whereas large multilevel mines with vast rooms were found to be partly collapsed or at least gravitationally affected in many places.
Large and multilevel mines are often established within rock environments with inclined layers or even the occurrence of folds, which increases the risk of gravitational events (e.g. flexural toppling; Li et al. 2019). The rectangular system of joints and fractures (faults) further disturbs the stability of the excavation forms. This appears similar to gravitationally (Iovine  (Šebela et al. 2021) affected karst caves or crevice-type caves (Margielewski and Urban 2017).
In our study, the displacement rates detected by 4-yearlong measurements in five mines (Table 2) revealed three types of rock mass behaviour: (i) oscillating movements with slight widening and narrowing of cracks (up to ~ 0.2 mm in total), which occur close to the entrance and are thus affected by freeze-thaw cycles (Cárdenes et al. 2012;Deprez et al. 2020), or shallow creep on steep slopes (Chang et al. 2015). Analogous oscillations were measured by Lekhuleni and Zvarivadza (2017) in simulated mine conditions. The exceptional 4-mm widening of the KLO1 crack is probably caused by the slow dropout of the rock block affected by freeze-thaw cycles within the mine entrance. As the annual air temperature of the area is 6-9 °C (Tolasz et al. 2007), we assume their frequent repetition. (ii) Distinct displacements (up to several mm in total) on cracks situated deeply within the mines in positions independent of temperature changes and freeze-thaw cycles. These displacements are reactions to various gravitational releases of rock blocks with subsequent types of movement (spreading, toppling, back rotation). Continuous movement led to the impossibility of further measurement (KLO2) and the destruction of a measuring point (KLO3). The exceptional narrowing of the RAA1 crack is caused by subsidence of the above situated and gravitationally released rock layer. Similar small-scale gravitational shifts of rock blocks were detected in crevice-type caves or on gravitationally affected rock structures (Lenart and Šesták 2016). (iii) Oscillating movements with slight widening and narrowing of cracks (up to ~ 0.2 mm in total) situated deeply within the mines in positions independent of temperature changes and freeze-thaw cycles. These indistinct shifts most likely reflect the tectonic strain changes within the rocks (Klimeš et al. 2012;Stemberk et al. 2017). It is likely that these slow preparatory movements may evolve into catastrophic gravitational failures (e.g. Parise 2015a; Loew et al. 2016).
It follows from the results that the AUSMs represent a highly dynamic underground environment. It is not surprising that 42% of the detected mines are affected by distinct falls or collapses. Therefore, based on the relatively high amount of revealed gravitational geomorphic forms within the AUSMs and the surprisingly high rate of measured displacement on discontinuities, we consider slate mines to be highly dynamic from the perspective of geomorphic evolution. Looking into the future, smaller mining works will most likely persist. In large mines, continuous succession will lead to the gradual collapse of vast rooms, relaxed sections of adits (Diederichs and Kaiser 1999), and entrance portions stressed by freeze-thaw cycles. This will further affect the overall natural processes within the mines, including flooding, cessation of ventilation, and changes in weathering and precipitation (see also Parise 2016;Lenart et al. 2022). During the successional destruction of excavation macroforms of the mines, the wide range of associated meso-and microforms will be created and then gradually vanish. After tens or hundreds of years, the existence of the AUSMs will probably end with overall filling of the underground space with collapsed material, resembling the overall infilling of cave passages by sediments (Auler 2009), leaving only sporadic surficial manifestation in the form of sinkholes (Lenart et al. 2017).

Conclusion
We identified and described 60 subterranean geomorphic forms in 114 abandoned underground slate mines situated in the Nízký Jeseník Upland (Czechia, Central Europe). Forms were studied in 74 small mines, 20 medium mines, and 10 large multilevel mines. According to their genesis, the forms were classified into seven types.
Based on the temporal relation to a historical mining event, excavation forms, excavated geological features, and fillings are pre-or syngenetic, and the fluvial, fluviogravitational, and gravitational forms and speleothems are mostly postgenetic. Most of the forms are macroscale (only excavated geological features are of mesoscale), and the speleothems are microscale size. One-third of them were found to be fragile and prone to gradual erosion. Moreover, almost half of them were classified as rare forms with limited occurrence. The forms occur independently of each other or are embedded. Cascades of forms occur in large multilevel mines affected by fluvial and gravitational geomorphic processes.
The microscopic and XRF analysis of microform samples revealed their composition. The groups of mine pearls contain surface layers enriched with Ti and gypsum crystals, and the calcite soda-straw stalactites are common minothems. The amorphous material with high content of S and Ca corresponds to gypsum. The exposed hydrothermal veins are of quartz composition.
Four-year displacement measurements on cracks were performed at 21 sites in five mines. Except for slight freeze-thaw oscillations or shallow creep with widening/ narrowing cycles, the total displacements of more than 4 mm reveal the underground slate mines to be gravitationally sensitive and highly dynamic. This is in accordance with the identification of 42% of mines as gravitationally affected. Although collapses or falls still occur rather locally, after the successional destruction of excavation macroforms lasting hundreds of years, the underground abandoned slate mines will probably experience overall filling with collapsed rock material.
This research contributes to the knowledge of the geomorphology of abandoned underground mines. Nevertheless, they remain insufficiently investigated landforms, which may lead to underestimation of their importance within the world's landscapes and ecosystems. Although the results are applicable to other types of abandoned mines in the world, we should expect new findings of as yet undescribed geomorphic forms.
Acknowledgements The research was supported by SGS01/PřF/2022 project.
Funding Open access publishing supported by the National Technical Library in Prague.

Data Availability
The original research data are available upon reasonable request.

Competing Interests
The authors declare no competing interests.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http:// creat iveco mmons. org/ licen ses/ by/4. 0/.