Application of GPR and ERT methods for recognizing of gypsum deposits in urban areas
- 251 Downloads
This paper presents the selected results of GPR (ground penetrating radar) and ERT (electrical resistivity tomography) surveys carried out on the sites in Poland where shallow karst forms were found in gypsum deposits. The aim of the surveys was the noninvasive detection of karst forms as well as weathered and fractured bedrock which may threaten the stability of the surface and, consequently, may cause damage to buildings, as well as overground and underground infrastructure. The geophysical surveys were conducted at a depth of only a few meters, i.e., to the depth of buildings foundations. GPR surveys were carried out in short-offset reflection profiling mode with standard orientation of the antennae set; however, on one site, different orientations of antennae were tested. During ERT surveys, different measurement arrays were applied in order to analyze which array was optimal for the detection of karst forms as well as weathered and fractured bedrock. Complex interpretation of geophysical surveys resulted in reduced ambiguity and revealed some regions, dangerous for surface stability. Due to the fact that gypsum deposits were investigated to the depth of maximum 10 m; therefore, hydrological processes were analyzed in the paper instead of hydrogeological processes.
KeywordsGeophysical methods GPR ERT Gypsum karst Karst hydrology
Karst phenomena occur in limestone and evaporate like salt and gypsum deposits. The limestone karst occurring in Poland is well recognized and described in geological and geophysical literature. Typically, authors focused on gypsum karst which has been considerably less studied than limestone karst.
The increased spatial development of cities and villages has meant that more and more new buildings are located in areas where karst phenomena occur. The presence of shallow karst forms as well as weathered and fractured bedrock often causes damage to buildings and other overground and underground infrastructure. In order to ensure the safety of people and property, it is necessary to investigate the geological medium using invasive geological and geotechnical techniques and/or noninvasive geophysical surveys.
The application of methods of engineering geophysics for the investigation of gypsum deposits is novel in Polish geophysics and, to date, only a few works concerning this problem have been published, notably Łój et al. (2014), Porzucek et al. (2018), and Rudzki (2015). Among the wide range of geophysical techniques for high-resolution imaging of karst forms occurring up to a depth of approximately 10 m (i.e., to the depth of buildings foundations), the GPR (ground penetrating radar) technique seems to be the most appropriate. This technique allows for fast and inexpensive investigation of the geological medium with a resolution of a few centimeters. In addition, the electrical resistivity tomography (ERT) methods were used, resulting in thorough investigations which reduced interpretation ambiguity.
Geophysical literature contains applications of the above-mentioned geophysical methods for the examination of gypsum deposits and gypsum karst. In a paper by Manoutsoglou et al. (2010), the ERT method in combination with a detailed geological investigation was tested as tool for the mapping of gypsum/anhydrite transition zones. Derobert and Abraham (2000) present a combination of the GPR reflection method and seismic imaging (tomography) performed in a gypsum quarry. In a paper by Ulugergerli and Akca (2006), the existence of cavities and weak zones in gypsum was explored using ERT, GPR and seismic refraction methods. Prokhorenko et al. (2006) explored the GPR reflection technique for the detection of caves and galleries in gypsum deposits as well as for searching of discontinuities and cracks in rocks. McCormack et al. (2017) demonstrated the use of ERT and discrete conduit network modelling to characterize the hydrogeology of the catchment by determining flow pathways and their likely hydraulic mechanisms. Chalikakis et al. (2011) presented a profound and thorough overview of the application of different geophysical methods to karst system exploration. In this work, the author utilized GPR, ERT, CSAMT (controlled source audio magneto-telluric method), VLF (very low frequency method), seismic tomography, MASW (multichannel analysis of surface waves method) and microgravimetry methods for the examination of karst forms.
The shallow karst forms as well as weathered and fractured bedrock analyzed in this paper were created by weathering processes and mainly by infiltration of surface water to the rock mass; therefore, the authors focused on karst hydrology instead of karst hydrogeology. Karst hydrology was a subject of interest of several authors; e.g., in a paper by Silverii et al. (2019), GPS time series and hydrological data from a tectonically active region in the Apennines, hosting large karst aquifers, were analyzed. Stokes et al. (2010) analyzed cases when allogenic water, which can be very aggressive (i.e., acidic), infiltrated into the rock mass from non-karst wetlands or bogs; when such water encounters carbonate bedrock, it can result in more intensive karst development. In their paper, Ługowski et al. (2016) described unique relief features, occurring near Lviv (Ukraine), called “hydration domes,” which were formed due to weathering (hydration) of anhydrite and its transformation into gypsum. Watlet et al. (2018) clam that water infiltration and recharge processes in karst systems are complex and difficult to measure with conventional hydrological methods; therefore, the authors proposed the analyses of time-lapse 2D ERT surveys carried out over a period of 3 years.
For the geophysical surveys, three sites where gypsum outcrops occurred or the gypsum roof was covered by thin layer of Quaternary deposits were chosen (Fig. 1c), i.e., the Siesławice, Wiślica and Staszów sites.
Relative electrical permittivity εr [−] (*)
Electrical resistivity ρ [Ωm]
Fresh water (from precipitation and melted snow)
Dry quaternary deposits (mixture of sand and clay)
Terrain surveys were carried out with the use of Swedish georadar PROEX and Ukrainian georadar VIY. For the measurements, 70 MHz, 250 MHz and 500 MHz shielded, bistatic antennae were used. GPR surveys were carried out in 2D mode, using the short-offset reflection technique. Traces were recorded alongside the profiles every 0.02 m (for 500 MHz antennae), 0.05 m (for 250 MHz antennae) and 0.2 m (for 70 MHz antennae). Stacking equal to 8-times or 16-times (depending on location) was applied to increase the signal/noise ratio.
The radargrams presented in the paper were shown in two forms, i.e.: as two-dimensional distribution of reflections and three-dimensional distribution of instantaneous amplitudes counted from Hilbert transform. Instantaneous amplitudes may be treated as energy distributions of georadar signals reflected from karst forms existing in gypsum. All radargrams were presented in normalized form, i.e., with amplitudes (energies) normalized to the maximum amplitude (energy) of the direct air wave. For proper time-depth conversion of radargrams, procedure of adaptation of diffraction hyperbolae was performed (ReflexW manual 2018); this procedure was carried out on radargrams recorded in the Siesławice site.
Radargrams were digitally processed using the ReflexW software. The following procedures were applied: t0 and topographic corrections, removal of wowing effect, DC correction, Butterworth time-dependent filtration, 1D median filtering, gaining with the use of energy decay function, background removal, stacking, 2D average, spectral whitening and morphologic filtration. A detailed description of the applied procedures as well as a description of the parameters assumed for the aforementioned procedures can be found in publications by Annan (1999), Gołębiowski (2012) and ReflexW Manual (2018).
Basic parameters of measurement arrays in ERT technique (Loke 1999)
Depth range versus length of profile
Distribution of measurement points
Strength of signal
ERT data were processed in the 2D inversion method using Res-2D-Inv software produced by Geotomo Software, applying nonlinear optimization techniques (Loke 2018). At the inversion stage, the limits were applied to the model to eliminate the inverse model solution ambiguity. For this purpose, robust inversion (L1 norm) was selected, as it is preferable in the case of rapid changes in underground object boundaries (i.e., karst forms). The robust inversion includes an iterative algorithm that compares the calculated model with the apparent resistivity measurements and gradually improves the calculated model by the minimization of the sum of the absolute difference (Abs error) between them.
Results from the Siesławice site
Two-dimensional, reconnaissance GPR surveys in the Siesławice site were carried out as part of a BSc thesis (Ornacka 2014); at the same time preliminary microgravimetric measurements in 2D mode were taken. The results of the reconnaissance works were presented in a paper by Łój et al. (2014). The results of detailed microgravimetric surveys, carried out along five profiles (Fig. 4b), were presented in a paper by Porzucek et al. (2018).
In Fig. 6, it is easily noticed that gypsum on the whole investigation site is highly weathered, to the depth of approximately 0.8 ÷ 1 m. Locally, a weathered zone is developed vertically to the depth of approximately 1.5 m. The most dangerous area is located between x = 4 m and x = 11 m, where highly fractured and weather gypsum occurs around the cave. If a building would be located on such area, it is highly probable that it would be damaged due to instability of the geological medium. In addition, in an area of highly weathered gypsum, e.g., x = 0 ÷ 22 m along profiles no. 4, 5, 6, 7, the fissures might appear on the building due to instability of the near surface zone.
The results of the GPR surveys presented in Fig. 7 are similar to those presented in Fig. 6, i.e., highly weathered gypsum is observed to the depth of approximately 1.5 m; locally, weathered and probably additionally fractured rock mass is extended to the depth of approximately 2 m, i.e., from x = 14 m to x = 21 m (profile no. 8, 9, 10), from x = 0 m to x = 6 m (profile no. 11, 12, 13) and from x = 30 m to x = 45 m for all profiles. These areas are the most dangerous for building’s stability and for safety of overground and underground infrastructure.
Results from the Wiślica site
On the Wiślica site, the geological medium was recognized to the depth of 2 m by 14 boreholes (Sowiński 2009). Geological information from boreholes located near the profiles analyzed in this section shows (Fig. 8b) that under a soil and asphalt/pavement, anthropogenic material occurs; under this material, a roof of gypsum rock is located, and the top part of the gypsum is weathered. The GPR and ERT anomalies located in the near surface zone, to the depth of 1.5 ÷ 2.0 m were not analyzed, because it is a zone strongly changed by human activity (Fig. 8b).
Velocity of electromagnetic wave for typical gypsum (εr = 7—Table 1) is equal 0.113 m/ns; it was confirmed by velocity analysis (Fig. 5). Velocity for soil and Quaternary deposits (εr = 9—Table 1) is equal 0.1 m/ns. Velocity for anthropogenic material should be similar to velocity of soil and Quaternary deposits. Considering information presented above, for time-depth conversion of all radargrams, mean velocity equals 0.1 m/ns was assumed.
The results of ERT survey (Fig. 9b) show that the roof of gypsum is located at the depths between 1.7 and 4.0 m. Information from borehole B-7 (Fig. 8b) depicts that the roof of gypsum is located at the depth of 1.7 m, but the top part of rock, between depths 1.1 ÷ 1.7 m, is weathered. Borehole information correlates well with the result of ERT survey. In the gypsum bedrock, two subzones may be distinguished, i.e., lower resistivity zone (Fig. 9b—yellow color) and higher resistivity zone (Fig. 9b—brown color).
Higher resistivity zone correlates well with high-amplitude reflections, recorded between x = 19 ÷ 38 m; this zone may be interpreted as fractured gypsum with free spaces filled mainly with air. In higher resistivity zone, high resistivity anomaly (Fig. 9b, red and violet colors) is easily noticed; taking into account information from Table 1, resistivity for solid gypsum is equal c.a. 1000 Ωm, so high resistivity anomaly, where reflections have low amplitudes (Fig. 9a, between x = 26 ÷ 31 m), depict to the solid gypsum.
In the first part of GPR profile (Fig. 9a, between x = 0 ÷ 19 m), amplitudes of reflections were significantly lower which was caused by an increasing of attenuation and consequently by a decreasing of resistivity (Fig. 9b—blue, green and yellow colors). Recording of lower resistivity in gypsum bedrock (Fig. 9b—yellow color) was caused by an increase in clay minerals in fractured gypsum (Fig. 2b). Recording of low resistivity in anthropogenic material and in weathered gypsum (Fig. 9b—blue and green colors) was caused by presence of clay minerals and water in the near surface zone.
The results presented in Fig. 9 depict that the most dangerous region for building foundations appear in the northwestern part of the Salt Square.
Borehole B-5, which is located the nearest to the profiles GPR-2 and ERT-2 (Fig. 8a) delivers information that to the depth of 2 m, asphalt/pavement and anthropogenic material are located; neither weathered gypsum nor solid gypsum bedrock were revealed to the depth of 2 m (Fig. 8b). Considering information from borehole B-7 (Fig. 8b), which is located 3.3 m lower than borehole B-5 and taking into account, the result of ERT survey (Fig. 9b) roof of gypsum deposits, under profiles GPR-2 and ERT-2, should be placed at depths of c.a. 6 ÷ 7 m.
In the first zone, the absence of reflections is observed. In the result of ERT survey (Fig. 10b), for depths greater than c.a. 7 m, very low resistivity was recorded, and therefore, the attenuation of electromagnetic wave caused the reduction in amplitudes of reflections. Low-resistivity anomaly may be interpreted as fractured/weathered gypsum, where fractures and porous spaces were filled with water or were colmatated with wet clay.
In the second zone, high-amplitude reflections were recorded to depth of approximately 8 ÷ 9 m. High-amplitude reflections and higher resistivity in the zone between x = 36 ÷ 51 m and z = 7 ÷ 10 m depict the presence of fractured/weathered gypsum with voids filled mainly with air.
At the depths from 2 m to 6 ÷ 7 m, high-amplitudes reflections were recorded (Fig. 10a). In this strip, three ERT anomalies with higher resistivity are easily noticed (Fig. 10b). The Basilica is located on the anthropogenic earthwork which height is c.a. 3 m. The surrounding areas of the Basilica have been subjected to strong anthropopression from the medieval ages to XX century. Borehole B-3 (Fig. 8b) delivers information that roof of weathered gypsum is located at the depth of 1.0 m and roof of solid gypsum bedrock at the depth of 1.8 m; as it was mentioned before, roof of weathered gypsum in borehole B-7 (Fig. 8b) is located at the depth of 1.1 m and roof of solid gypsum bedrock at the depth of 1.7 m. Taking into account all information presented above, it is difficult to interpret unequivocally GPR and ERT anomalies, but the fact that in borehole B-5 neither weathered gypsum nor solid gypsum bedrock were revealed to the depth of 2 m, allows to assume that GPR and ERT anomalies at the depths between 2 ÷ 7 m have anthropogenic nature.
Results from the Staszów site
As mentioned in the introduction, geophysical investigations had to be carried out to a maximum depth of 10 m. On the Staszów site, the surveys were conducted to the depths of 20 m (GPR) and 13.5 m (ERT), because in the low-values times (in GPR method), so-called air reflections from trees and buildings were expected; such reflections can make the interpretation difficult or even impossible. GPR measurements were taken with the use of 70 MHz antennae, and different orientations of antennae (Fig. 3) were tested during terrain surveys. Using the ERT method, different measurement arrays were applied (Table 2) in order to analyze which array is optimal for the detection of karst forms as well as fractured and weathered gypsum bedrock occurring in this region.
Additionally, in Fig. 12b a gain of amplitudes in low-value times is not visible, i.e., between x = 24 ÷ 60 m, where the superposition of reflections from underground and overground objects appeared; consequently, a radargram is easier for visual interpretation and “air reflections” do not disturb markedly the interpretation.
The least valuable results of the GPR surveys were obtained along profile named GPR using of the end-fire orientation antennae (Fig. 12c) but the main anomalies are still visible.
As it was mentioned formerly, the reflection coefficient depends on the contrast of electrical properties and orientation of antennae. Depending on site (i.e., geological conditions and level of disturbances), type of underground object and its parameters and geometry, usually only one, selected orientation of GPR antennae delivers satisfied result; sometimes it is a co-pole orientation, in other site it is an end-fire orientation and in the third site a cross-pole orientation may deliver the better results. Fractures and karst forms have stochastical distribution in the geological medium, they have usually complicated shape and their petrophysical parameters may change markedly in distance of a few meters depending of medium filling of free spaces. Therefore, there is not one universal orientation of GPR antenna and different orientations of antennae should be tested during detection of fractured zones and karst forms. In the Staszów site, the most suitable orientation of GPR antennae for fractures and karst forms detection seems to be a cross-pole orientation.
In general, the detailed interpretation of the results of geophysical surveys is as follows: (a) Gypsum is probably divided into horizontal blocks or the horizontal fracturing occurred in the rock mass; (b) the low resistivity medium is either fractured/weathered gypsum filled with Quaternary material or a mixture of Quaternary and anthropogenic materials; (c) the very low resistivity “Anomaly I” is a zone filled with water; (d) the most unsafe area for the buildings is the region of “Anomaly I” due to presence of water in weathered/fractured/loose geological medium where suffosion processes may be developed.
The results of the research presented in this paper show that selected geophysical methods can be successfully applied to the detection of karst forms occurring in gypsum. Due to the spatial and stochastical distribution of karst forms as well as fractured and weathered bedrock, measurements should be carried out in 3D mode, at it was shown for the Siesławice site. In terms of building development, it is important to detect karst forms and fractured zones developed in the vicinity to a depth of a few meters, i.e., around and directly under the foundations of buildings.
For this purpose, the high-resolution, fast, low time- and labor-consuming GPR method appears to be optimal but other auxiliary methods have to be applied for reduction in interpretation ambiguity. In GPR method, it is very important to test different antennae orientations to obtain the best measurement results for further interpretation. The GPR method can be limited by the presence of materials with low resistivity (and consequently a high attenuation) in the geological medium, e.g., clays, loams, mineralized water. In urbanized areas, an additional limitation of the GPR method can also be interference from overground objects and underground infrastructure. The application of shielded antennae only partly protects against such interference.
In such situations, the GPR method should be replaced with the much more time- and labor-consuming ERT technique, and however, this should provide positive results in geological media with low resistivity and is not so sensitive to anthropogenic interferences. The ERT method also has its limitations, e.g., in urbanized areas, terrain surface is often covered with asphalt, concrete or pavement which makes it impossible to insert electrodes into the ground.
If GPR and ERT methods cannot be used, other geophysical techniques should be applied for the noninvasive detection of karst forms, e.g., microgravimetry, conductometry or CCR (Capacitively Coupled Resistivity) method. In order to reduce interpretation ambiguity, at least two of the above-mentioned methods should be used simultaneously.
This paper was financed by Cracow University of Technology, amongst others, as part of the Projects: Ś-2/372/2013/DS and Ś-2/299/2016/DS (the Wiślica site) and Ś-2/335/2017/DS (the Staszów site).
- Annan AP (1999) Practical processing of GPR data. Sensor & Software Inc., CanadaGoogle Scholar
- Annan AP (2001) Ground penetrating radar—workshop notes. Sensor & Software Inc., CanadaGoogle Scholar
- Bąbel M (1999) History of sedimentation of the Nida gypsum deposits (middle Miocene, Carpathian foredeep, southern Poland). Geol Q 43(4):429–447Google Scholar
- Bąbel M, Olszewska-Nejbert D, Nejbert K, Ługowski D (2015) The Badenian evaporative stage of the Polish Carpathian Foredeep: sedimentary facies and depositional environment of the selenitic Nida Gypsum succession. In: Proceedings of 31st IAS meeting of sedimentology, Cracow, PolandGoogle Scholar
- Cichy K (2014) Kompleksowe badania geofizyczne na stanowiskach archeologicznych. Praca magisterska, promotor T. Gołębiowski, WGGiOŚ AGH, Kraków (unpublished) (in Polish) Google Scholar
- Daniels JJ, Wielopolski L, Radzevicius S, Bookshar J (2003) 3D GPR polarization analysis for imaging complex objects. In: Proceedings of 16th SAGEEP conference, San Antonio, USAGoogle Scholar
- Documentation (2018) Wykonania niedestrukcyjnych badań archeologicznych przeprowadzonych z użyciem przynajmniej dwóch nieinwazyjnych metod pomiarowych terenu placu przykościelnego w Wiślicy (działka nr 437) w ramach projektu “Modernizacja Muzeum Archeologicznego w Wiślicy jako oddziału Muzeum Narodowego w Kielcach wraz z otoczeniem w celu zabezpieczenia i ochrony unikatowych obiektów dziedzictwa narodowego”. Opracowanie: firma MET-GEO, Trzebinia (unpublished), (in Polish) Google Scholar
- Drożdzak R, Twardowski K (2010) Przenikalność dielektryczna ośrodków porowych—czynniki wpływające na jej zmienność. Wiertnictwo – Nafta—Gaz 27(1–2):111–120 (in Polish) Google Scholar
- Fils J (1954) Kras gipsowy Niecki Nidziańskiej. Prace geograficzne nr 1, Instytut Geografii PAN, PWN, Warszawa (in Polish) Google Scholar
- Gołębiowski T (2008) Changeable-offset GPR profiling for loose zones detection in the levees. In: Proceedings of 14th European meeting of environmental and engineering geophysics, Cracow, PolandGoogle Scholar
- Gołębiowski T (2012) Zastosowanie metody georadarowej do detekcji i monitoringu obiektów o stochastycznym rozkładzie w ośrodku geologicznym, Rozprawy – Monografie AGH, Kraków (in Polish) Google Scholar
- Gołębiowski T (2014) Dokumentacja badań georadarowych przeprowadzonych na działce nr 923 w Wiślicy. Ekspertyza dla UM w Wiślicy (unpublished), (in Polish) Google Scholar
- Gołębiowski T, Tomecka-Suchoń S (2012) GPR fractures detection using changeable antennae orientation. In: Proceedings of 5th Saint Petersburg international conference and exhibition, Saint Petersburg, RussiaGoogle Scholar
- Gołębiowski T, Pasierb B, Kiełtyka-Sołtysiak G (2018b) Metody geofizyczne w ochronie dziedzictwa archeologicznego Wiślicy. In: Materiały konferencji “Metody geofizyczne w archeologii polskiej”, Toruń (in Polish) Google Scholar
- Łój M, Gołębiowski T, Porzucek S (2014) Geophysical surveys and modelling for recognizing of gypsum karst. Geoinform Pol 13:83–97Google Scholar
- Loke MH (1999) Electrical imaging surveys for environmental and engineering studies—a practical guide to 2-D and 3-D surveys. Geometrics Ltd, ScottsdaleGoogle Scholar
- Loke MH (2018) Tutorial: 2D and 3D electrical imaging surveys. Geotomo Software, PenangGoogle Scholar
- Ługowski D, Jarzyna A, Bąbel M, Nejbert K (2016) Metody dokumentowania zastosowane w badaniach terenowych stanowiska wietrzejących anhydrytów w Piskach k. Lwowa. Biuletyn Państwowego Instytutu Geologicznego 466:201–214 (in Polish) Google Scholar
- Manoutsoglou E, Vachlas G, Panagopoulos G, Hamdan H (2010) Delineation of gypsum/anhydrite transition zone using electrical tomography. A case study in an active open pit, Altsi, Crete, Greece. J Balk Geophys Soc 13(2):21–28Google Scholar
- Marcak H, Gołębiowski T (2010) Analiza możliwości detekcyjnych metody GPR dla zmiennej geometrii układu pomiarowego. In: Materiały konferencji “Geofizyka w Geologii i Górnictwie”, Sosnowiec–Zawiercie (in Polish) Google Scholar
- Ornacka S (2014) Georadarowe obrazowanie 3D zjawisk krasowych. Praca inżynierska, promotor T. Gołębiowski, WGGiOŚ AGH, Kraków (unpublished), (in Polish) Google Scholar
- Pasierb B, Gołębiowski T, Łój M, Porzucek S, Cichy K, Pacek B (2014) Czy pod placem solnym w Wiślicy składowano sól? In: Materiały konferencji “Metody geofizyczne w archeologii polskiej”, Kraków (in Polish) Google Scholar
- Plewa M, Plewa S (1992). Petrofizyka. Wydawnictwa Geologiczne, Warszawa (in Polish) Google Scholar
- Porzucek S, Łój M, Matwij K, Matwij W (2018) Modern geodetic measurement techniques in gravimetric studies on the example of Gypsum Karst in the Siesławice Region. In: Proceeding of conference “POL-VIET 2017”, E3S Web of Conferences, no. 35Google Scholar
- Prokhorenko V, Ivashchuk V, Korsun S, Stefanyshyn I (2006) Ground penetrating radar survey in Podil’lya Karst Area—Ternopil Region, Ukraine. In: Proceeding of 11th international conference on GPR, Columbus, Ohio, USAGoogle Scholar
- ReflexW Manual (2018) SandeierGeo firm. Karlsruhe, GermanyGoogle Scholar
- Rudzki M (2015) Zastosowanie metody tomografii elektrooporowej do lokalizacji struktur krasowych. Geofizyka Toruń Company (in Polish) Google Scholar
- Sowiński J (2009) Opinia geotechniczna pod przebudowę Rynku i budowę Pawilonu Archeologicznego w Wiślicy. Sowiński—usługi geologiczne, Kielce (in Polish) Google Scholar
- Stokes T, Griffiths P, Ramsey C (2010) Karst geomorphology, hydrology and management. In: Compendium of forest hydrology and geomorphology in British Columbia, Land management handbook. vol 66, pp 373–400Google Scholar
- Ulugergerli E, Akca I (2006) Detection of cavities in gypsum. J Balk Geophys Soc 9(1):8–19Google Scholar
Open AccessThis article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.