Project

The principal project objectives can be listed as follows

• O1. Developing procedures and protocols for continuous and sequential data acquisition and processing
• O2. Quantifying landform change
• O3. Predicting the landform behavior, short- and long-term (using different climate change projections)
• O4. Targeting the critical landforms and proposing preliminary design for preventive or stabilizing measures
• O5. Sustainability – providing the base for further monitoring (beyond the project’s life cycle)
• O6. Promoting the Devils’ town site for a better touristic visibility and economic improvement of the local society, by attracting UNESCO, scientific community and general public attention.
• O7. Networking with domestic and abroad research groups interested in the subject

All these objectives (O1-O7) are measurable by Key Performance Indicators – KPI indicated hereinafter (e.g., by the number of intermediate surface models, predictive models, published articles, external researchers involved, number of critical cases), and achievable/realistic in the project lifetime, which has been proven by the 1-year pilot project.

The subject of the DEMONITOR project is a rare geological and geomorphological phenomenon at the “Devils’ town” site, near Prolom Spa in southern Serbia. The mystique interplay of erosional forces and volcanic rock left a multitude of remarkable landforms – tall rock pillars, colloquially called “the Devils” in literature known as the “badlands” (Fig. 1). The host rock is a relic of an ancient pyroclastic flow, containing loose and frail, erodible material. Ages and ages of weathering and erosion carved about 200 pillar-like forms, up to 15 m tall and 6 m in diameter.

The Devils’ town site is just a piece in a larger puzzle that is Lece volcanic complex in the south of Serbia (Fig. 3). A vast area, roughly 700 km2, covered in andesitic rock and its pyroclastite, is a reminiscence of a large volcanic complex (equivalent of present-day Pinatubo) formed during many effusive and extrusive episodes, which lasted from upper Oligocene to possibly end of Pliocene (30-3 Ma ago). Abundance of hot springs in the area today (Prolom Spa, Kuršumlija Spa, Sijarinjska Spa) witness of a very intensive and long volcanic activity. Magma production was staggering since three different calderas were formed across the complex, one of which is central to the Devils’ town (Fig. 3). Today, the calderas are invisible, as millions of years of various erosional processes reshaped the volcanic landscape entirely, but geological evidence remains. Such a large stratovolcano alternated between violent (Plinian) and more peaceful eruptions depending on the acidity and amounts of volatiles in the magma chamber at the time of eruption. The violent eruptions producing lahars and pyroclastic flows are of particular interest for this story. When the eruption is violent, enormous amount of ash is emitted. If conditions are right, meaning, if the eruption is shortly followed by an acidic rain, there is a chance for generation of lahars and pyroclastic flows, which create tephra deposits. Lahars – overheated debris flows, are furthermore interesting since they hold a key to the authenticity of the Devils’ town site.
What makes the Devils’ town unique and worthy of research?
There are many examples of badlands, which typically leave bizarre pyramidal landforms ending with sharp peak or a large caprock. Rittner and Steinegger in Italy, Euseigne and Auvergne in France, Cappadocia in Turkey, Bryce-Canyon in USA, Kuklice in North Macedonia, to name few. But all of these did not have a unique combination of genesis and erosion that is typical for the Devils’ town. Most of them are made of moraine materials with irregular gradation of the constitutive fragments. Some of them have cap made of volcanic rock (as a result of polyphase effusion and alternation of harder and softer layers) which is harder than the underlaying tuffs (Cappadocia). However, the caprocks that sit atop of the Devil’s town pillars came about in a different manner. Fly-ash and rock boulders during one of many eruptions have formed a tephra layer that started flowing down the caldera slopes. It contained just right density and water content and had just right velocity and size of the boulders, that the inverse gradation could take place. Namely, instead of plunging to the bottom of the flow, large and heavy volcanic boulders, weighting over a couple of metric tons, were pushed upwards by static buoyancy forces, which were extreme due to the high bulk density of the tephra matrix (imagine a flowing cement mortar as an equivalent). Further reduction of the boulder’s weight is caused by suspension force and selective buoyancy, which are introduced by Ter-Stepanian in 2002 . These effects are based on the presence of a deflected flowline around a large submerged boulder floating in fluid of equal or higher density. Flowlines tend to increase the frictional stresses from the underside and reduce the stresses above the boulder, forcing it to rise upward. Once repositioned by such mechanism to the surface or near it, the boulders were cemented (as flow seized and cooled). The erosion succeeded and did the rest of the work by slowly chiseling friable tephritic rock around the boulders (Fig. 2). How slowly? This could be additional objective of DEMONITOR project, in the case it lasts long enough to establish the annual erosion rate, allowing the back-calculation of the current landforms age. Only capped pillars have a long lifecycle and can reach large size. Eventually, the rock gives way to erosion and pillars collapse under their own weight once the sloping angle (which is constantly increased by vertical erosion) becomes too steep for the rock to handle (Fig. 2).

In modern times, during the site’s touristic exploitation, it has been noticed that pillars are changing, i.e., eroding, collapsing, sinking, but also emerging. In light of the proliferating climate change, it remains intriguing how further erosional effects will impact their evolution. The site is currently protected as a natural monument by the Institute for Nature Conservation of Serbia (INCS) and was nominated for the UNESCO natural heritage site, but unfortunately, reached only the tentative list in 2002. There is, potential for renewing/furthering promotion in the UNESCO context. It is also worth mentioning that the entire area, including the Prolom Spa and the Radan Mt., is very attractive but still underappreciated destination in comparison to the other touristic centers in the country. As it is an important source of income for this underdeveloped region in Serbia, there is interest in promoting the touristic offer supported by the evidence from our research (how large/tall/wide are the pillars, how heavy are the caps, how old are the pillars, and when and why can they collapse, are all interesting pieces of information for visitors). The site is currently co-managed by the local authority – Municipality of Kuršumlija and governed by the local joint-stock company Planinka under the supervision of the INCS.

DEMONITOR project concept can be generalized into four frameworks or four designated Work Packages (WP):

• (i) Coordination WP1
• (ii) Acquisition WP2
• (iii) Modeling WP3
• (iv) Promotion WP4

which will all be described in detail later on, but herein, some basic and specific details will follow (Fig. 3b, 4).

WP1 The coordination involves the usual technical preparations for the project kick-off and subsequent operation (equipping, organizing meetings, visits, etc.).
WP2 The acquisition involves interdisciplinary approach, based on Geological, Engineering-geological, Geophysical and Geodetic knowledge combined in harmonized acquisition activities. These activities will allow execution of objective O1(Developing procedures and protocols), which can be indicated by KPI that can be expressed through the number of successfully realized field visits, amount and type of related data, processed and uploaded to the project data cloud (number of processed point clouds, satellite images, 3D surface models). It also concerns the O5 (Sustainability) which can be indicated by the level of completeness of the final data repository (at the end of the project) as KPI, e.g., did all field visits resulted in measurable change, that can guarantee that future researchers will be able to measure such change in erosion in foreseeable period of time (according to our acquisition dynamics). Finally, acquisition incorporates networking with other groups which is why it also addresses O7 (Networking), measurable by the number of achieved contacts, joint work on the field, number of joint papers (co-authored) with members of other research groups (e.g. in North Macedonia, Turkey, University of Priština group seated in Kuršumlija) as KPI.

The acquisition can be subdivided into:
A) Monitoring of the surface displacement:

It is primarily based on sequential monitoring of the landforms, which will be performed by using the state-of-the-art non-invasive acquisition techniques:

• Terrestrial Laser Scanning (TLS), using already available Leica ScanStation P20 device
• Unmanned Airborne Vehicles (UAV) photogrammetry, based on Structure-from-Motion (SfM) approach, combined with mobile Global Navigation Satellite System (GNSS), using already available drone DJI MATRICE 600 and GNSS rover
• Satellite Radar Interferometry (InSAR), based on ESA Sentinel-1 open data (and software which will be procured – currently unavailable).

The TLS and photogrammetry techniques will provide high-resolution local surface models, while InSAR will be used to monitor displacements of fixed ground control points (corner reflectors, and possibly strong natural reflectors, such as caprock), ensuring that the global stability of the site is satisfied, i.e., that the changes are relative to the pillars and not the wider terrain deformations, such as regional subsidence for instance (local surface models cannot cover this aspect) . The team has experience in working with similar concepts applied to rock-slopes and landslides . There will be at least two TLS and UAV acquisition sequences, i.e., two site visits, per year. Both visits will be scheduled in the non-vegetative part of the year, which coincides with the touristic low season (late autumn/early spring), in order to minimize possible noise introduced by vegetation and visitors. Each sequence of scanning/imaging will produce current surface model (3D point-cloud) of the major landforms. The principal idea is to compare these surface models in appropriate software package (CloudCompare, 3D Reshaper), which will allow us to interpret, i.e., to detect and quantify the 3D surface change between consecutive sequences (or any other combination of sequences). In plain words, we will be able to visualize which parts of the pillars have changed between two visits and for how much. Rough flight plan scanning stations and other systems are outlined in Fig. 4. The TLS and UAV acquisition techniques are capable of capturing sub-cm changes and indicate locations that are prone to weathering and collapsing. Therefore, it is expected that pre-failure deformations can be also registered and linked to the subsequent stability models. It has been shown by the members of the team that slight deformations can be indicative precursors to local failures .
B) Continuous near-real time monitoring of general conditions:

Visual monitoring of apparent conditions at the site will be performed through a simple time-lapse recording by using fixed optical video surveillance camera (OVSC). Its purpose is to provide more accurate timing of potential failures that occur between two site visits (TLS and UAV are limited to temporal resolution twice a year, while OVSC can be set to record daily or hourly). This knowledge can be crucial for further linking of failure with environmental conditions (primarily weather data). It is important to mention that this piece of equipment (unlike UAV and TLS) is not available and needs to be procured and installed at earliest convenience, while the data will be stored locally and logged during each field visit. Standard power outlet is available at the site.

C) Sequential monitoring of weathering:

Determination of tephra deposits thickness and its potential change using two geophysical, non-invasive methods, geoelectric method – electrical resistivity tomography (ERT) and seismic method – seismic refraction tomography (SRT). Seismic acquisition system RAS-24, a modular 24-channel high-resolution signal enhancement seismograph will be used. Signal triggering, by hammer, will be conducted away from pillars, at the stable part of the terrain and it will not endanger them. The geoeletrical survey will be conducted along the same profile lines as seismic (Fig. 4), but also along additional profiles, placed between pillars in order to obtain the most precise subsurface information in their immediate vicinity. The equipment that should be used for ERT is WGMD-3A 60channels 5m Space (Multi-function Digital DC Resistivity/IP Meter) which is currently not at disposal (it will be procured through the Project). ERT and SRT techniques use powerful inversion algorithms to achieve high resolution subsurface inversion models for resolving subsurface characteristics and geological conditions. There will be at least two ERT and SRT acquisition sequences. First to establish the model of subsurface, and the second to detect structural changes in rock mass nearby under the pillars. Visits will be scheduled once a year in accordance with TLS and UAV acquisition period (either autumn or spring visit).
D) Continuous weather monitoring: Weather data, primarily the temperature and precipitation, will be collected from the national hydrometeorological survey open databases (http://www.hidmet.gov.rs/) on daily level using the closest weather stations – Kuršumlija (15 km), Leskovac (45 km), Niš (55 km), but there is potential to cooperate with domestic research group from University of Priština, which has installed weather monitoring equipment at the site.
E) Determining rock characteristics: Rock data are needed for establishing parameters for stability analyses. We propose to test surrounding rock masses adjacent to the pillars (because pillars themselves are too fragile for sampling) using Point Load Test – PLT, Needle Penetrometer SH-NP, and Schmidt Hammer – SH (to have an impression of the andesitic intact rock strength, all instruments are already available at our host institution FMGUB), bulk density (appropriate digital level is available in FMGUB premises) and Slake Durability apparatus (needs to be procured at earliest convenience), which is used for simulating the weathering rate under atmospheric conditions, and estimate the weathering lifetime of the exposed rock. In addition, samples will be tested to acidity endurance (to simulate acid rain effects on the host rock) using simple chemical laboratory layout (simple lab dishes and appropriate already available chemicals at FMGUB). Results from the SRT will also be used to determine input elastic parameters (dynamic moduli).