Predicting natural cavities in chalk

Abstract

IntroductionChalk is a soluble carbonate rock with extensive karst development. Natural cavity occurrence initially appears to be random. In an area where the degree of influence of all cavity formational factors is similar, but dissolution is focused on one set of joints rather than another, then solution feature occurrence is perhaps random. This might be termed the ‘microscale’ view, measured at a scale of metres. However, if the pattern of natural cavity occurrence is considered at a ‘macroscale’ level, say measured in hundreds of metres or kilometres, then spatial patterns emerge suggesting dissolution is not as random as it might first appear.Spatial characteristics of natural cavity occurrence on the ChalkIn order to analyse the spatial characteristics of natural cavity occurrence it was first necessary to collect as many records of solution feature occurrence as possible from published and unpublished sources including local authorities, Construction Industry Research and Information Association (CIRIA), Transport Research Laboratory (TRL), Building Research Establishment (BRE), National House Building Council (NHBC), the Environment Agency, water companies, and site investigation reports. Particular emphasis was placed on fieldwork to record new features revealed in a wide range of engineering works, road and motorway construction, and on visits to large numbers of working/disused chalk quarries and aggregate workings. A database of 2226 natural cavities was compiled, composed mainly of solution pipes, sinkholes (dolines) and swallow holes.The spatially related database (each cavity location being recorded by National Grid Reference) was carefully scrutinized to determine

Evaluating and comparing methods of sinkhole susceptibility mapping in the Ebro Valley evaporite karst (NE Spain)

“…These indirect methods can be divided into (1) heuristic and (2) statistical or probabilistic (Hansen, 1984;Carrara et al, 1995). Heuristic models base susceptibility assessments on the establishment, in a rather subjective way, of threshold values (Gao and Alexander, 2003) or a scoring system to a group of conditioning factors (Ogden, 1984;Thorp and Brook, 1984;Brook and Allison, 1986;van Rooy, 1989;Buttrick, 1992;Forth et al, 1999;Edmonds, 2001;Zisman, 2001;Kaufmann and Quinif, 2002;Zhou et al, 2003;Lei et al, 2005;Tolmachev and Leonenko, 2005;Dai et al, 2008;Koutepov et al, 2008). Probabilistic methodologies derive the susceptibility models from the analysis of statistical relationships between the known sinkholes and a group of conditioning factors.…”

“…Although the goodness of fit (fitting-rate) between the distribution of sinkholes used in the analysis and that of the susceptibility zones of the prediction models have been calculated by several authors (Edmonds, 2001;Yilmaz, 2007;Dai et al, 2008;Kaufmann, 2008), the evaluation of the predictive capability of the models using independent sinkhole populations has only been performed by Galve et al (2008a,b) and Lamelas et al (2008). Consequently, the reliability of most of the susceptibility models presented in the literature remains unknown.…”

Subsidence hazard in Berkshire in areas underlain by chalk karst

“…The mapping technique was applied following the principles of subsidence hazard mapping developed by Edmonds (1987Edmonds ( , 2001. The mapped boundaries of terrain units have not been set out by accurate survey, but have relied on topographic base maps of various scales and observation by walkover survey.…”

“…The subsidence hazard mapping techniques are explained in detail in Edmonds (2001). The surface morphology, subsurface geology, geomorphological development and the hydrogeology of the study area have been utilized to prepare a subsidence hazard map.…”

Collapse dolines susceptibility mapping in Doukkala Abda (Western Morocco) by using GIS matrix method (GMM)

“…Furthermore, any susceptibility model or map must be validated by two methods (Guzzetti et al 2006); i.e., the validation of a collapse doline susceptibility model can be ascertained using the same collapse doline data used to obtain the susceptibility assessment (Edmonds 2001;Kaufmann and Quinif 2002;Yilmaz 2007;Dai et al 2008), or by using independent collapse dolines information (Galve et al 2008;Lamelas et al 2008) not available to build the model of collapse dolines susceptibility map. In the anterior works which were carried out in the study area (Mazéas 1967;Theilen-Willige et al 2014), the dolines susceptibility models were generated but not tested.…”