An empirically developed cpt-based assessment method for characterization of sinkhole vulnerability in Florida Karst

Document Type

Conference Proceeding

Publication Date



Florida's unique geology poses multiple engineering concerns when characterizing the subsurface conditions in both previously and newly developed landscapes. Florida's covered-karst landscape has ushered the seemingly never-ending-battle between the propagating population and karst's consequential sinkhole geohazard. The cone penetration test (CPT) has been widely used for subsurface investigations in covered-karst landscapes due to the test's ability to provide a nearly continuous profile of soil resistance, allowing engineers to precisely determine the depths in which internal soil erosion may have taken place. However, once these conditions are identified, there still lacks a viable assessment technique which is both reproducible site-to-site and verified to quantify the location's vulnerability to future sinkhole collapse. In this paper, the authors review a database of 26 project sites located in Florida where CPTs were performed to investigate the potential of sinkhole activity. Within the 26 project sites, a total of 155 CPTs were performed in and around the identified sinkhole occurrence (either collapse or depression). Using statistical correlation and categorical regression of the location of CPT to the proximity to sinkhole feature, an empirical function was developed relating the location's vulnerability for sinkhole development to various CPT-obtained subsurface condition variables. The resulting empirical function can then be used as a relative estimation of sinkhole vulnerability when performing CPTs within karst project sites in Florida. When multiple CPTs are performed within an identified karst anomaly, the proposed assessment index can be calculated for each and compared against a proposed criterion for better discernment of mitigation strategies.

Publication Title

Geotechnical Special Publication




GSP 325

First Page


Last Page


Digital Object Identifier (DOI)