Quantitative Assessment of In-situ Salt Karstification Using Shear Wave Velocity, Dead Sea

Highlights

• Shear Wave Velocity is applied to estimate Dead Sea salt karstification

• Hydraulic conductivity K versus Vs relationships were generated for salt

• Inter-relation can be used to estimate actual salt properties for in-situ conditions.

• Vs varies between limits of 750 m/s to more than 1650 m/s

• K varies between slightly higher than 10^− 4 m/s to slightly above 10^− 8 m/s

Abstract

The Dead Sea (DS) coastal areas have been dramatically affected by sinkhole formation since around 1990. Such sinkholes along both Israeli and Jordanian shores are linked to karst cavities that form through slow salt dissolution. A quantitative estimate of such in-situ salt karstification would be an important indicator of sinkhole hazard. One of the indications of salt karstification is its increased hydraulic conductivity, caused by the development of dissolution cavities forming conducting channels within the salt layer. We measured the hydraulic conductivity (K) versus shear-wave velocity (Vs) of DS salt in situ for estimating the actual salt karstification in areas of sinkhole development. These parameters were measured with the Magnetic Resonance Sounding (MRS) and Multichannel Analysis of Surface Waves (MASW) methods, respectively. Understanding of the field relationships was augmented by similar inter-relations obtained in the laboratory on samples of DS salt. In-situ salt velocities Vs vary from 750 m/s to over 1650 m/s, while hydraulic conductivity (K) in the same zones varies between about 10^− 4 m/s to slightly over 10^− 8 m/s. Both field and laboratory K and Vs values fit the exponential function ln(K) = − 0.0045 ∗ V_s − 5.416 with a determination coefficient (R²) of 0.88. A classification based on Vs and K was generated for salt conditions and the corresponding degrees of sinkhole hazard, which was verified in the Mineral Beach sinkhole development area. The mapping of sinkhole sites shows that they form within highly conductive zones with K ≥ 5.5 ∗ 10^− 5. It is suggested that this methodology, with some modification, can be used for evaluating the conductive properties of karstified rock and associated sinkhole hazards.

Introduction

The Dead Sea (DS) coastal areas of Israel and Jordan have been dramatically affected by sinkhole formation since around 1990 (Fig. 1). This development continues today, with an obvious potential for further collapse beneath roads, dwellings and in the potash works. Most researchers accept that such sinkholes near the shorelines are caused by salt dissolution (karstification) by under-saturated groundwater (Frumkin and Raz, 2001, Abelson et al., 2006, Yechieli et al., 2006, Legchenko et al., 2008a). Our hypothesis proposes that sinkholes form along a dissolution front over the edge of a buried salt layer (western edge in Israel and eastern edge in Jordan). This hypothesis is based on seismic refraction surveys carried out mainly along the western DS shore (Ezersky, 2006, Ezersky et al., 2010) and in Ghor Al-Haditha (El-Isa et al., 1995, Abueladas and Al-Zoubi, 2004).

Following the dissolution front salt edge concept, we have suggested a general model of the salt layer deposited around the DS. This model (Fig. 2) suggests that there is a 10–30 m-thick buried salt belt around much of the DS shore, controlling the sinkhole hazard (Legchenko et al., 2008b, Ezersky and Frumkin, 2013). All known cases of collapses and accidents including the alarming catastrophe at pond 18 in Jordan (Closson, 2005) have occurred within this salt belt. The model presented in Fig. 2 enables us to conclude that: (1) the dissolution front follows an ancient shoreline which existed at the stage of salt unit deposition (10.2–10.8 ka) – modern sinkholes are formed along this dissolution front; and (2) the buried salt layer extends from the modern Dead Sea shoreline landward, towards the dissolution front, permitting its investigation from the surface.

The working hypothesis is that circulating groundwater dissolves salt, causing a gradual increase of salt porosity and permeability. This increase, connected with the evolution of pore space during salt dissolution, was suggested by Bernabé et al. (2003). Shalev et al. (2006) applied this mechanism to DS salt dissolution, demonstrating that porosity and permeability are important factors in both salt dissolution and sinkhole development.

In our previous works, a geophysical methodology was proposed for sinkhole-hazard evaluation based on (1) salt-edge delineation by seismic refraction, and (2) the spatial distribution of bulk resistivities (ρ_x) and shear-wave velocities in salt (Vs). The mapping of ρ_x is based on the Transient Electromagnetic (TEM) method (Frumkin et al., 2011), whereas Vs is mapped by the Multichannel Analysis of Surface Wave (MASW) method (Ezersky et al., 2013a). Both studies showed that geophysical parameters are minimal in the 60–100 m zone (dissolution front) next to the salt edge, but that they gradually increase toward the DS reaching a stable value after some 100–200 m from the salt edge (Frumkin et al., 2011, Ezersky and Frumkin, 2013, Ezersky et al., 2013a). This shows that salt karstification is greatest closest to the salt edge and that it gradually decreases toward the DS. Another objective is to correlate the geophysical parameters with a quantitative evaluation of salt porosity and hydraulic conductivity, thus providing a reliable indication of sinkhole hazard. Hereafter, we will consider only the in-situ relationships between shear-wave velocity (Vs) and hydraulic conductivity (K).

It is known that shear-wave velocity is closely related to the porosity and permeability of rocks (Nelson, 1994, Schoen, 2004). Earlier work by Ezersky and Goretsky (2013) (presently manuscript is under review for possible publication in Engineering Geology) found correlative relationships between porosity and permeability, and velocities for DS salt samples saturated with brine. However, there is a problem in extrapolating laboratory-based relationships to field conditions. Comparison of ultrasonic velocities (Vp, Vs) measured on borehole samples in the laboratory with data from ultrasonic logging in the same boreholes, has shown that in-situ salt is characterized by significantly lower velocities than salt samples in the laboratory selected from the same depth. Both porosity and permeability values are thus higher in the borehole. It is suggested that hydraulic conductivity values in a salt layer will also be greater, because Vs velocities are significantly lower than those in laboratory salt specimens. To extrapolate the Vs and K parameters measured in samples to the field scale (100 m² lateral size), we used the field parameters measured on the Nahal Hever South (NHS) site (Site II in Fig. 1) using MASW and Magnetic resonance Sounding (MRS), also known as Surface Nuclear Magnetic Resonance (Legchenko et al., 2011, Legchenko, 2013, Song, 2013), which is a geophysical application of the Nuclear Magnetic Resonance (NMR) phenomenon.

The MASW method is known as an efficient tool for measuring Vs (Miller et al., 1999, Park et al., 1999, Xia et al., 1999). It is especially effective when the object is located below the groundwater table, because Vs is slightly sensitive to groundwater filling the pores in rock. It is simple to undertake a MASW survey in the field and it is relatively cheap. The MRS method allows in-situ measurements of hydraulic conductivity (K) values for comparison with those measured by pumping tests (Legchenko et al., 2008a, Nielsen et al., 2011, Knight et al., 2012, Vouillamoz et al., 2012a). However, the method is very sensitive to electromagnetic noise and is most suitable for measuring relatively high hydraulic conductivity values (Costable and Yaramanci, 2011).

The main goal of the present study is to extrapolate relationships between the hydraulic conductivity (K) and shear wave velocities (Vs) measured on salt specimens in the laboratory, to the same parameters measured in salt under field conditions. A classification of the salt conditions is suggested, based on Vs values. Finally, the obtained relationships are applied for generating a hydraulic conductivity (K) map of the Mineral Beach sinkhole development sites, in order to evaluate sinkhole hazard.