Likewise, Taheri et al.17 also assigned most weight to bedrock lithology (34%) compared to other predisposing factors like distance to faults, groundwater withdrawal, distance to deep wells, the thickness of alluvium etc. Further, we compare two widely used methods of sinkholes mapping- AHP based and LR based approaches. Brinkmann, R., Parise, M. & Dye, D. Sinkhole distribution in a rapidly developing urban environment: Hillsborough County, Tampa Bay area, Florida. to the east. The large sinkhole that opened up along Baseline Road in Marion County on Tuesday night is the second one to happen in three days. Geol. Logistic regression approaches to susceptibility mapping have been widely used across different hazard-prone regions to reliably predict natural hazards like landslides27,29,30,74 and sinkholes8,24. 9a,b). While the AHP based approach tended to allocate more area to very high class and less area to very low class (2.7%), the LR based approach allocated more area to very low class (42%) and less area to very high class. Singh, K. B. OCALA, FL October 2017 – A sinkhole opened at the intersection of Rolling Greens Drive and Peebles Beach Road. Other natural attractions in and around Marion include the Ocala National Forest and the Florida Trail. Gao, Y., Alexander, E. C. & Barnes, R. J. Karst database implementation in Minnesota: Analysis of sinkhole distribution. Floridia land use, cover and forms classification system. Distance from active mines (m), (d) Surficial geology, and (e) Distance from closed topographic depressions (m). J. Cave Karst Stud. Geosci. Normalization of principle eigen vector associated with the pairwise comparison matrix for predisposing factors resulted in the highest weight to surficial geology (0.426) followed by closeness to topographic depression (0.322), soil permeability (0.147), distance to flow channels (0.063), and distance to active mines (0.042). volume 9, Article number: 7140 (2019) The LR based approach showed a low presence of sinkholes in the very low susceptibility class and low absence of sinkholes in the very high susceptibility class. Model with the higher area under the ROC curve was considered a better performing model. Kim, Y. J. Comparison of these techniques in a common setting provides important inferences on the benefits and pitfalls of these modeling approaches. While it is a common practice in literature to use a balanced presence and absence data in logistic regression modeling of rare events60, numerous logistic regression based natural hazard susceptibility mapping studies have used unbalanced data24,61,62,63,64. Geomorphology 134, 157–170 (2011). 1). 3. Stratigraphic geology vector data file (last updated in 2017) from the Florida Geological Survey (2001) available online at: http://publicfiles.dep.state.fl.us/OTIS/GIS/data/GEOLOGY_STRATIGRAPHY.zip was used to extract surficial geology polygons for the study area. Sinkholes of Marion County, Florida , 2008 This map was created by FCIT and represents reported sinkhole events in Marion County based on data gathered by the Florida Geological Survey (FGS) and the Florida Department of Environmental Protection (FDEP). " Prevalence of circular sinkholes in the study area highlights that the karst topography is relatively young68. Catena 72, 1–12 (2008). 2c). Soil permeability raster data was originally prepared from the soil survey geographic data (SSURGO) maintained by the USDA-Natural Resources Conservation Service. Yalcin, A. GIS-based landslide susceptibility mapping using analytical hierarchy process and bivariate statistics in Ardesen (Turkey): Comparisons of results and confirmations. Salvati, R. & Sasowsky, I. D. Development of collapse sinkholes in areas of groundwater discharge. The distribution of existing sinkholes is primarily concentrated in the Ocala area and along major highways. Marion County was formed in 1844 and named in honor of General Francis Marion of South Carolina who fought in the American Revolutionary War. Recently, logistic regression (LR) approach has been popular among susceptibility modelers mainly because it works well with both categorical and continuous variables23,29,30,31. Spring Hill, located in Hernando County, made national news in 2014 when a sinkhole swallowed up a neighborhood street. 1987–1998 (AUC: AHP = 0.831, LR = 0.860), (c). The maximum and the median diameter of circular sinkholes were respectively 24.8 m and 1.5 m. The largest elongated sinkhole reported had the maximum length of 39.6 m. The median length of the elongated sinkhole was 3 m for this study. 415–422, https://doi.org/10.5038/9780991000951.1083 (National Cave and Karst Research Institute, Carlsbad, NM, 2015). 52, 45–50 (1999). Soil permeability raster map was reclassified manually into five classes; 0–23, 23–75, 75–119, 119–163, and 163–200 in/hr (Fig. Others who lean more on a country feel could find some ideal choices at Silver Springs Shores considered as a bedroom community of Ocala. 99, 169–184 (2008). Both modeling approaches performed well in identifying sinkhole susceptibility zones for the study area on evaluated time periods (Fig. Surficial geology and distance to flow networks were also important predisposing factors for sinkhole susceptibility for this study. Geol. Between the three time periods evaluated, the LR based approach performed better than the AHP based approach. Ciotoli, G. et al. http://publicfiles.dep.state.fl.us/otis/gis/data/STATEWIDE_LANDUSE.zip, http://publicfiles.dep.state.fl.us/otis/gis/data/FGS_SUBSIDENCE_INCIDENTS.zip, https://prd-tnm.s3.amazonaws.com/StagedProducts/Hydrography/NHD/State/HighResolution/Shape/NHD_H_Florida_State_Shape.zip, http://publicfiles.dep.state.fl.us/OTIS/GIS/data/GEOLOGY_STRATIGRAPHY.zip, http://publicfiles.dep.state.fl.us/otis/gis/data/LAND_SURFACE_ELEVATION_24.zip, https://doi.org/10.5038/9780991000951.1062, https://www.census.gov/quickfacts/fact/table/marioncountyflorida/POP060210#viewtop, https://doi.org/10.1061/9780784480441.048, https://doi.org/10.1061/9780784480717.023, https://doi.org/10.5038/9780991000951.1083, http://creativecommons.org/licenses/by/4.0/, https://doi.org/10.1038/s41598-019-43705-6. 1:9,244,648 Surficial geology was the predominant contributor to sinkhole formation (43% contribution) compared to other predisposing factors for the AHP model. We used closed topographic depression features available in this dataset for this study. Zhou, W., Beck, B. F. & Adams, A. L. Application of matrix analysis in delineating sinkhole risk areas along highway (I-70 near Frederick, Maryland). As a result, there has been a growing interest combining these approaches with GIS to map susceptibility of environmental hazards7,8,25,28,29,39,40,41. Eng. Given this, hurricanes do not affect this county as much as the other more coastal counties west and east. Res. Geomorphology 42, 213–228 (2002). Article  2b). We manually reclassified the distance from topographic depressions layer into seven classes as, 0–100, 100–200, 200–400, 400–600, 600–800, 800–1600, and >1600 m (Fig. That sinkhole grew to 125 feet wide and forced four homes to be evacuated. Report on Review of the 2010 Sinkhole Data Call (2010). Marion County’s inland location is centered between the Gulf of Mexico to the west and the Atlantic Ocean to the east. Environ. Xiao, H., Kim, Y. J., Nam, B. H. & Wang, D. Investigation of the impacts of local-scale hydrogeologic conditions on sinkhole occurrence in East-Central Florida, USA. Ciurleo, M., Cascini, L. & Calvello, M. A comparison of statistical and deterministic methods for shallow landslide susceptibility zoning in clayey soils. 9c,d). Geomorphology 266, 1–10 (2016). Given this, hurricanes do not affect this county as much as the other more coastal counties west and east. We verified sinkhole absence data points post selection via field visit for easily accessible sites (Aug 2016–Mar 2017) and for inaccessible sites via use of digital ortho-rectified aerial imageries from 2008 (0.3 m × 0.3 m resolution; Florida Department of Revenue, 2008), multi temporal high resolution satellite images (1999–2018) from Google Earth Pro (0.15 m × 0.15 m resolution; Google Inc., 2018), and 1.5 m resolution digital elevation model data available for the study area (prepared by the Marion County IT/GIS team- Marion County using LiDAR data collected from 2003 to 2004) to make sure they did not represent potential sinkholes or sudden depressions. Check the latest Florida Sinkhole News >>, Check our Florida Sinkhole Alley Guide >>. Saha, A. K., Gupta, R. P., Sarkar, I., Arora, M. K. & Csaplovics, E. An approach for GIS-based statistical landslide susceptibility zonation-with a case study in the Himalayas. The analytical hierarchy process. The model used for our LR based approach is: where p is the probability of occurrence of a dependent variable, b0 is the intercept, b1, …, bn are the coefficients of independent variables (X1, …, Xn). Procedia Earth Planet. Yalcin, A. Environ. Ercanoglu, M., Kasmer, O. For instance, both the percentage of sinkhole absence and frequency ratio were higher for the very low susceptibility classes than other classes. Water Resour. Marion County Sheriff’s Sgt. SINKHOLE INFORMATION FOR MARION COUNTY RESIDENTS. We obtained elevation and contours dataset shapefiles available at: http://publicfiles.dep.state.fl.us/otis/gis/data/LAND_SURFACE_ELEVATION_24.zip prepared by Florida Geological Survey (2015). 4). Florida Department of Transportation Surveying and Mapping Office Geographic Mapping Section. J. We used the Spatial Analyst tool available in ArcGIS 10.3.1 (Environmental Systems Research Institute, Inc., Redlands, CA, 2015) for raster conversions. Geol. High susceptibility class covered almost 23% of the study area. 122, 371–388 (2013). Undifferentiated sediments (Pleistocene/Holocene) cover about 10% of the area and consist of silica-rich rocks, organics, and freshwater carbonates. The county also boasts of its Lake Weir ideal not only for boating, water skiing and other water-based leisure activities. Distribution of sinkhole dimension (length, width, or depth) had a log-linear distribution (Fig. 8a–c). Earth Syst. 8d). OCALA, FL June 2012 – A large sinkhole, spanning 100 feet, opened at the edge of a retention pond in the Fore Ranch Subdivision off State Road 200, forcing the evacuation of a townhome building. In contrast, the LR based approach resulted in very high susceptibility class in about 2.9% of the total area. Rev. One of the holes was 30 feet deep. B. Sinkhole subsidence due to mining. J. Spatial location of sinkhole susceptible zones in these areas are likely due to the pressure of urban water consumption on the aquifer or the presence of carbonate and dolomitic geology (Ocala formation)48,68. The regional map on the right depicts sinkhole locations that have been reported since 1954 by the Florida Geological Survey Series No. Map showing location of Silver Springs basin and potentiometric surface of the Upper Floridan aquifer, May 1989.....3 2. They also won’t be able to turn south from Dogwood Road. Parise, M. A present risk from past activities: sinkhole occurrence above underground quarries. A. We compared this model average to the LR model generated using full unbalanced dataset. Galve, J. P. et al. 53, 1007–1022 (2008). For example, the percentage of sinkholes falling on each class increased with its susceptibility to sinkhole occurrence for the AHP based approach. Stoch. Eng. & Engelbrecht, J. Soil permeability: Rawal (2016)51 showed a linear relationship between the time for soil surface collapse and soil permeability in a simulated sinkhole study. For the LR based approach, the percentage of sinkholes was highest in the high susceptibility class. Distance from topographic depressions: Closeness to topographic depression has been related to sinkhole occurrence in previous studies54. Sinkhole (training and validation) shapefiles were created from these data points for the Marion County. CAS  Get link; Facebook; Twitter; Pinterest; Email; Other Apps; By Mark. We presented the applicability of two common sinkhole susceptibility mapping approaches for an area with prominent karst topography in central Florida. Park, S., Choi, C., Kim, B. Ayalew, L. & Yamagishi, H. The application of GIS-based logistic regression for landslide susceptibility mapping in the Kakuda-Yahiko Mountains, Central Japan. Factors like karst topography17,39, surficial geology16,24, soil permeability46, proximity to the flow channels or drainage networks8,17, groundwater withdrawal47, depth to the water table9,48, mining activity49, the thickness of overburden20, recharge of aquifers50 etc., are thought to influence the formation of sinkholes. Google Scholar. CAS  Identifying and maintaining long-term sinkhole inventory is critical to modeling and validating sinkhole susceptibility models. The apparent sinkhole had developed following heavy rains extending eastward from Tropical Storm Cristobal. We suggest that future work on LR modeling of sinkholes should also focus on evaluating truly absent areas of sinkhole occurrence through the use of modern techniques like LiDAR or RADAR77,78,79. Environ. We used this 30 m × 30 m soil permeability raster digital data, available from the Florida Geological Survey, as soil permeability data layer for the study area (Florida Geological Survey 200652). See What Users Are Saying About Sinkholemaps.com >>, Marion County’s inland location is centered between the. Alin, A. Multicollinearity. Between 2000 and 2014, a conservative estimate of sinkhole damage costs for the United States was at least 300 million dollars per year with the actual cost being much higher2.