Graduation Year

2023

Document Type

Dissertation

Degree

Ph.D.

Degree Name

Doctor of Philosophy (Ph.D.)

Degree Granting Department

School of Geosciences

Major Professor

Sarah Kruse, Ph.D.

Committee Member

Timothy Dixon, Ph.D.

Committee Member

Rocco Malservisi, Ph.D.

Committee Member

Shimon Wdowinski, Ph.D.

Committee Member

Boo Hyun Nam, Ph.D.

Keywords

geophysics, InSAR, sinkholes, subsidence, time series

Abstract

This dissertation consists of three studies that employ Persistent Scatterer Interferometry (PSI, also known as PSInSAR) to better understand how subsidence in west-central Florida relates to underlying geological processes. In the first study, near-surface geophysical methods (Ground Penetrating Radar (GPR) and Electrical Resistivity (ERT)), terrestrial remote sensing applications (Light Detection and Ranging (LiDAR) and Structure from Motion (SfM)), and PSI were used to monitor the spatial and temporal behaviors of a suspected growing sinkhole in the Sandhill Boyscout Reservation, Hernando County, Florida. The survey area was located within and around a topographic low assumed to be the surface of the suspected sinkhole. The GPR details the structural variabilities below this low, indicating the presence of small pockets of shifted soil. ERT datasets present the possibility of multiple water-filled voids below the survey area. LiDAR and SfM failed to identify surface changes over time on the grassy surface. PSI datasets from installed corner reflectors show faster subsidence outside of the central topographic low in the area, indicating that an increase in the spatial coverage of the application would be beneficial in pinpointing the complex subsidence pattern surrounding the suspected sinkhole. The results illustrate that sinkholes in Florida deform on multiple spatial time scales and have complex morphologies beyond a conical depression over a void.

In the second study, the post-processing of 75 TerraSAR-X PSI time series datasets was completed to test the limits of using InSAR-derived time series to pinpoint the timing and location of small-scale activity expected to be associated with sinkholes. Two time series analysis methods designed to capture non-linear subsidence were applied: a slope break analysis and change point analyses to identify changes in the mean. The results indicate that the timing of slope breaks and change points in the PSI time series correlate with periods of low groundwater levels and high rainfall, both of which are expected accelerators of sinkhole growth. The change points correlate best with structural changes related to sinkhole remediation and other construction activities. A differential subsidence analysis was also used to locate homes subsiding faster than their neighbors. This method shows the strongest correlation with buildings with reported sinkhole activity. (The term sinkhole activity is used in a generic sense and not following the statutory definition of the State of Florida.)

In the third study, we characterize subsidence in the Tampa Bay region using 172 Sentinel-1 images acquired over six years between 2016 and 2022. We correlate the spatial distribution of the generated PSI velocity against other factors related to subsidence, such as geological data, the distribution of permitted well use, the wellfield groundwater elevation, and known sinkhole reports for the area. Variable subsidence was observed, with agricultural and mining areas experiencing the most subsidence (<-2 mm/yr). By applying band-pass filtering to the PSI velocity, we identified subsidence patterns along waterways and from agricultural groundwater use and clusters of sinkholes. The results show that, unlike many coastal cities, the Tampa Bay area is not experiencing rapid regional subsidence due to the overexploitation of groundwater resources. Most subsidence in the bay area relates to small-scale and short-term processes like sinkhole formation and local groundwater well use.

Download

Share

COinS