Can Sinkholes Be Predicted? Florida Sinkhole Risk, Detection and Prediction

Let’s be honest. Living in Florida has never been on my radar. However, becoming an older American makes me think about areas that are more friendly toward retirees. Don’t get me wrong. I am not against living in Florida. More affordable housing (compared to where I live) and winters that don’t make your face hurt for starters. Being a life-long northeastern resident, the idea of warm weather and daily epic thunderstorms in the summer (I’m a huge fan of dramatic weather) makes this seem worthy of considering. Of course, you gotta take the good with the bad and in Florida, that means hurricanes, cockroaches big enough to mug you, and sinkholes.

Many of Florida’s lakes and ponds are created by the same process that is creating new sinkholes in modern times. Sinkholes in Florida can be attributed to its increase in population and the demand for groundwater that the increased population creates. Due to the bedrock underneath most of Florida being made from a water-soluble mineral called Carbonate, human activity has created a perfect environment for sinkholes to form.

One of the biggest factors in sinkhole formation is the lowering of the water table. As the water table drops, it creates a cavity that is susceptible to pressure from the overlying soil layers (similar to Texas, where sinkhole development is widely caused by oil extraction). This happens in areas where there is a lot of agriculture and a high concentration of community development. Rainwater can also be a factor as it introduces water that seeps into the underground aquifers, dissolving Carbonate on its way down. Some sinkholes can be directly associated with hurricane activity.

The Real Estate sector is one of the hardest-hit areas of the economy where sinkholes are concerned. And this isn’t just about property values, this is also about the lives of people whose homes get swallowed up by a hole in the ground, sometimes without warning. This type of sinkhole is called a “cover collapse”. Given the attraction of living in a climate that rarely sees snow, this brings an unwelcome element of unpredictability to potential and existing homeowners.

I went to school to study Geographic Information Systems (GIS) and learned how the technology of putting data into maps is used across several disciplines where the characteristics of a location are being quantified. Aside from GIS, the business of sinkhole study and prediction utilizes a lot of tools and fields of expertise. Geology, Chemistry, Statistics, and others are used in discovering the conditions where sinkholes develop. GIS is how it gets put onto a map. Following is a list of how some of these pieces fit together to form a picture of a favorable sinkhole environment.

Geology – At its most elemental, Geology is the study of the rocks and the ground beneath our feet. Being an entire branch of science, it goes a lot deeper (pun not intended) than that. Geologists also study fossils, continental plate tectonics, and the formation of bedrock structures. Florida’s bedrock (Carbonate) is actually the remains of ancient mollusk shells (clams, barnacles, snails) whose primary element is calcium. Calcium dissolves readily when in contact with acidic water. This is the root of Florida’s sinkhole issue.

Chemistry – Elements and water acidity are squarely in the realm of Chemistry. This piece fits into the puzzle by describing how water and the elements that influence its characteristics (such as acidity) can contribute to the conditions that make sinkhole formation likely.

Statistics – If science is really the study of measurements, then statistics would be the study of probabilities. Since the earnest study of sinkholes began in the 1980’s it has been determined that there is no step-by-step recipe for predicting sinkhole formation. What can be done is to create a model that shows the probability of a sinkhole forming given known conditions.

GIS – Geographic Information Systems is the technology that is used to put data on a map. This is powerful scientific software that can map the location of physical objects, or map probabilities based on several inputs. It gives a user the power to tabulate results based on “where” they are. Speaking personally, I find maps much more compelling and deliver much easier to grasp information than spreadsheets alone. Together, it makes information and educated decisions much easier.

Through decades of research, it has been determined that there is no straightforward answer to where or when a sinkhole might be created. As mentioned earlier, the existence of carbonate-based bedrock  (in Florida’s case) is a major factor. Another is the depth of this bedrock. According to the USGS (United States Geological Survey), “the absence of carbonate bedrock eliminates the possibility of sinkhole occurrence.” The best thing about GIS is the amount of public data that is readily available to use in models and prediction studies. If I were interested in researching an area for sinkhole prediction, I would also include land-use data (another available GIS dataset in most states) and note where the highest concentration of sinkholes occurs. Later on, in this blog rant, I go into how this can be done.

There are a few kinds of sinkholes. Some are responsible for gradual land subsidence (known to crack foundations of buildings and roads over a long period of time), and others abruptly collapse  (called cover-collapse) where the ceiling of an underground cavity created by the carbonate rock gets dissolved by water. Below is a geological map of Florida indicating the depth of bedrock in the state and the overlying soil composition. The Dept of Natural Resources has developed 4 classes of land that can define sinkhole formation as well as the type of sinkhole likely to form there. I georeferenced this map to compare where sinkholes have been observed and recorded.

This next map has data from the Florida DEP sinkhole repository (sinkhole data accessible here, indicated as dark blue dots) superimposed on the bedrock map shown above. There are 4,134 surveyed sinkholes in this layer. Sinkholes can happen almost everywhere in Florida, but as you can see, areas with the highest sinkhole concentration occur where the bedrock is 30 to 200 feet below the surface (in Area III). Surface soil layers also play a significant role in the type of sinkhole that can be formed. In the yellow areas, the overlying weight of the soil is more shallow than the blue areas because the bedrock is found closer to the surface. Sinkholes that form here tend to be gradual land subsidence as opposed to catastrophic cover-collapse that is more dominant in the blue areas. The rose-colored areas indicate where the bedrock is deepest. Sinkholes are least frequent in these areas but also tend to be of the largest, deepest, and terrifyingly of the cover-collapse type that is created abruptly.

The USGS also states that overuse of groundwater and wells is another significant impact on sinkhole development. As mentioned earlier, water acidity can be a factor as well. Water Acidity is influenced by the decomposition of organic matter (of which Florida obviously has an abundance of). Organic matter can come in the form of tree roots. There has been a correlation between proximity to mature trees and sinkhole development. Not only do dead tree roots decompose (and contribute to water acidity in doing so), but they are also very good at exploiting cracks in bedrock to find water sources when they are alive. As the roots grow, so do the cracks.  

Since these areas are going to be a part of the discussion moving forward, here are the definitions as stated on the image above:

Area I. Bare or Thinly Covered Limestone: Sinkholes are few, generally shallow and broad, and develop gradually. Solution sinkholes dominate

Area II. Cover is 30 to 200 Feet Thick: Consists of mainly in-cohesive and permeable sane. Sinkholes are few, shallow, of small diameter, and develop gradually. Cover-subsidence sinkholes dominate.

Area III. Cover is 30 to 200 Feet Thick: Consists mainly of cohesive clayey sediments of low permeability. Sinkholes are most numerous, or varying in size, and develop abruptly. Cover Collapse sinkholes dominate.

Area IV. Cover is more than 200 Feet Thick: Consists of cohesive sediments inter-layered with discontinuous carbonate beds. Sinkholes are very few, but large diameter, deep sinkholes occur. Cover-collapse Sinkholes dominate.

So looking at data on a map could give one the impression that sinkholes are everywhere and Florida must be a terrifying and unpredictable place to live. I do not think that is the case. 4,134 recorded sinkholes across a state that is 65,758 square miles give a lot of room to live safely. Each sinkhole would have to cover over 13 square miles to totally affect the state. Clearly, it is still a concern. If you narrow down the areas where sinkholes are statistically likely to form, those odds obviously increase a lot. Using GIS, I was able to tabulate the number of sinkholes that occur in each zone of bedrock depth. Those results are in the table below:

One of the more interesting insights gained from this data is that the land type where sinkholes are most prevalent also happens to cover the smallest area of Florida. Area III, with its 1885 recorded sinkholes has an area of 11,408.59 square miles. On average, you could travel 6.05 miles from one sinkhole to its nearest neighbor. On Average, that is. I’m certain there are ones that are closer in proximity.

Another way to look at the data would be to use Florida’s DEP-published land use GIS layer and perform the same operation of joining the land use type to each sinkhole. That activity resulted in this tabulation:

You can still see that Area III is statistically more likely to develop sinkholes, but now you can also see what type of land use they like to develop in. Even as I write this I have to question the robustness of the sinkhole dataset because residential neighborhoods are where they are most likely to be reported because this is where it is likely to cause the most problems. It is interesting to note that roadways (transportation) score pretty high on this table as well. What I’m getting at is that I wonder how many sinkholes go unreported because they aren’t causing a problem. Being completely honest, I am surprised that agricultural areas did not score higher, but then again maybe I shouldn’t be because there aren’t many people around to notice.

If I were looking into purchasing a home in Florida, my unprofessional but educated thought would be this: if you know a geologist, consult with them to find other information and resources. It is possible to have a survey taken of a property, although I have not researched the cost into this. I’d imagine it is possible that home insurance would cover some of this cost. Florida remains a very popular place to live, and services like sinkholemaps.com take some of the uncertainty out of the equation. And with some research on the land use and geology of an area, you can better understand the risk factor of sinkhole development.

The final table I want to present is the result of looking at the different land use types that exist in each bedrock area, and then dividing the square miles by the number of sinkholes that exist in each land use. It gets pretty complicated at first but due to the miracle of excel spreadsheets and copy/pasting formulas, it was not as unwieldy as it sounds.

Here is the same data in sortable form:

The way to read this table is to identify that there are two types of columns: an Area column, which gives the square miles of each land use type, and a Miles Per Sinkhole column which is the result of dividing the square miles by the number of sinkholes that occur in that type of land use. Try to find the lowest value in Miles Per Sinkhole, and that will give you the highest rate of sinkhole occurrence for that land use. There are some exceptions to this rule because some of the land-use types have very little area and are what statisticians would call “outliers”, meaning they mess with the data. For example, “Slough Waters” has a miles-to-sinkhole ratio of .31. This sounds exceptionally hazardous, right? It does unless you take into account that it has a total area of 2.88 square miles in all of Florida. The items that caught my attention are the Residential Medium Density in Area III with a rate of .98 and Transportation in Area III with a rate of .67. Again, this speaks to how often a sinkhole gets reported. If it’s on a road, I would imagine that it gets reported and dealt with very quickly!

Here is the same table as above in sortable form:

Ultimately, I hope this article gives you some insight as to where sinkholes are happening. Between these tables and the maps of sinkhole locations, you can make a more educated decision about sinkhole risks and weigh them against what kind of insurance you might need to protect your property from an encounter with one. For an extra layer of assurance, speak with an engineering firm that has a geologist and/or geochemist on staff to do a survey of land you are interested in.

References

https://pubs.er.usgs.gov/publication/cir968

https://blogs.missouristate.edu/mindseye/sinkholes/

• Author
• Recent Posts

Christopher Bride

Christopher Bride is a GIS Analyst in Cohoes, NY. He received his MS in Geographic Information Science from Lehman College (City University of New York) in the Bronx, NY., and a BA in Environmental Policy & Management from Hunter College (also City University of New York) in Manhattan, NY. Research was a primary duty for his degrees. He won the Miriam & Saul Cohen Prize for Excellence in Geographic Research for his undergraduate capstone “A Triple Bottom Line Analyses of Seven Food Policy Documents” in 2011. While a graduate student, he was hired as a GIS Data Researcher for the Lehman College-sponsored website https://www.lehman.edu/community-connect/links.php where he honed in on data research for GIS as a primary skill.

His professional interests are environmental management/conservation and working with environmental data. He was a GIS Technician on NYC’s 2015 tree census and QA’d half of the data himself (a dataset with over 660k datapoints). He has also been a NYC photographer for the real estate website Propertyshark.com since 2005.

When not working, he likes to kayak, fly drones, garden, cook, and raise two daughters with his wife in upstate New York. He also has a previous career as a musician and audio engineer and still loves to be involved, though in a less than professional capacity. He has been playing piano since age 5.

LinkedIn Profile

Latest posts by Christopher Bride (see all)

• Why Does Tennessee Have so Many Sinkholes - May 13, 2022
• Can Sinkholes Be Predicted? Florida Sinkhole Risk, Detection and Prediction - March 27, 2022
• Texas Sinkholes Research – Is Texas Sinking? - August 27, 2020