Coastal Subsidence in Cape Canaveral, FL, and Surrounding Areas: Shallow Subsidence Induced by Natural and Anthropogenic Processes

Abstract

Cape Canaveral, home to critical space exploration infrastructure, is facing potential flooding hazards from land subsidence and sea-level rise. This study utilized three geodetic datasets, the Interferometric Synthetic Aperture Radar (InSAR), the Global Navigation Satellite System (GNSS), and precise leveling, to investigate the spatial and temporal patterns of vertical land motion (VLM) in Cape Canaveral and its surrounding areas. Our analysis revealed that Cape Canaveral experiences both long-term regional subsidence and localized subsiding areas, while Merritt Island and the Peninsular Mainland remain relatively stable. The long-term regional subsidence in Cape Canaveral is likely driven by the compaction of younger, unconsolidated siliciclastic sediments, with a small contribution from glacial isostatic adjustment (GIA). The three localized subsiding areas identified in Cape Canaveral are each driven by distinct mechanisms: wetland modification in the western area, runway infrastructure development in the central area, and the natural compaction of young siliciclastic sediments in the southeastern region. Historical leveling data indicated temporal variations in subsidence rates at Cape Canaveral, from 5 mm/yr during the 1950–70s to 2 mm/yr in the 2000s. These findings have significant implications for infrastructure resilience and flood hazard assessment, as the observed subsidence compounds with the projected accelerated sea-level rise in the region. Our results highlight the importance of integrating long-term datasets to better characterize VLM in the dynamic coastal region for effective planning and risk mitigation.

1. Introduction

Cape Canaveral, located on the east coast of Florida, is renowned for its significant role in space exploration and research. This iconic site, home to NASA’s Kennedy Space Center and Cape Canaveral Space Force Station, has been the launchpad for countless missions that have significantly advanced our understanding of space. The area hosts critical infrastructure including launch complexes, vehicle assembly buildings, and advanced tracking systems [1]. These assets represent billions of dollars in investments and irreplaceable scientific resources. However, the long-term stability of this critical area is increasingly at risk of flooding caused by eustatic sea-level rise and vertical land motion (VLM).

In this study, we define VLM with positive values indicating uplift (upward motion) and negative values indicating subsidence (downward motion). When specifically discussing subsidence or uplift rates, we present them as positive values. VLM, encompassing both uplift and subsidence, poses significant challenges for coastal regions worldwide, impacting infrastructure, economies, and ecosystems. The process is driven by a complex interaction between natural and anthropogenic factors. Natural causes include tectonic movements, glacial isostatic adjustment, and sediment compaction [2,3,4], while human-induced factors encompass groundwater extraction, hydrocarbon withdrawal, and urban development [5,6,7], with different human-induced factors often superimposed on natural ones [8]. In coastal areas like Cape Canaveral, these factors, in combination with rising sea levels, create a particularly vulnerable environment.

The methods for detecting and measuring VLM have evolved significantly over time, transitioning from traditional techniques to advanced space-based technologies. In the pre-space geodesy period, traditional methods such as precise leveling and the analysis of tide gauge data were the primary tools for monitoring VLM. Precise leveling, involving the use of optical instruments to measure height differences between points, provided accurate measurements but was limited by its labor-intensive nature and sparse spatial coverage [9]. Tide gauge analysis offered insights into relative sea-level changes but could not be used to differentiate between actual VLM and sea-level variations. However, these pre-space geodesy studies do offer critical insights into VLM on a regional scale. Early research using sea-level data analysis significantly contributed to understanding VLM rates and patterns along the U.S. East Coast [10]. Subsequent research integrated geodetic leveling with tide gauge data for a comprehensive assessment of VLM in the U.S. [9]. Additional works further refined our understanding by correlating VLM with geological structures along the East Coast [11,12]. These pre-space geodesy studies reported subsidence rates of 4–9 mm/year in the Cape Canaveral region, laying the groundwork for subsequent research. These studies also highlighted the limitations of traditional techniques in capturing fine-scale spatial patterns and the need for more detailed, localized studies.

The advent of space-based geodetic techniques, particularly the Interferometric Synthetic Aperture Radar (InSAR) and the Global Navigation Satellite System (GNSS), revolutionized VLM monitoring capabilities. The InSAR utilizes the phase difference between multiple Synthetic Aperture Radar (SAR) images acquired over time to detect a millimeter-per-year scale of VLM over vast areas [13]. This technology, often integrated with GNSS data, provides unprecedented spatial coverage and temporal resolution in VLM monitoring [14,15,16,17]. A recent research employed advanced InSAR techniques and GNSS data to reveal complex VLM patterns along the U.S. Atlantic coast, documenting subsidence rates exceeding 4 mm/year in Cape Canaveral and surrounding areas [17]. Although pre-InSAR and InSAR studies have significantly advanced our understanding of VLM in Cape Canaveral, critical research gaps persist, particularly regarding the underlying causal factors driving long-term subsidence. Recent studies on the Holocene sea-level database reported subsidence rates of 0.5–2 mm/yr from Maine to northern Florida that are attributed primarily to glacial isostatic adjustment (GIA) [18,19,20]. However, it is unclear whether the long-term subsidence detected in Cape Canaveral is also a part of this larger pattern [21] or is more strongly influenced by local factors unique to the area’s geological [22,23] and anthropogenic context [24,25,26], as diverse subsidence processes often interplay and are superimposed [27].

Prior studies provide regional subsidence rates (e.g., ≥4 mm/yr) for Cape Canaveral; however, they offer little insight into how local geological processes (e.g., sediment compaction) and anthropogenic influences (e.g., land-use changes) contribute to subsidence. Previous research also lacks both a detailed area-specific analysis of the causal factors driving VLM and a mapping of the spatial variability of localized subsidence in the region. This study addresses these gaps by (1) identifying and evaluating the primary drivers of long-term subsidence in Cape Canaveral, and (2) analyzing the spatial distribution and causes of localized VLM across the region. Given Cape Canaveral’s critical importance to national interests and its vulnerable coastal location, this comprehensive investigation into its VLM provides vital insights for the coastal planning and long-term sustainability of this strategic area.

2. Geological Setting

This study area is located on the east central Florida coast and is part of the Atlantic Coastal Complex [28]. It consists of three predominant geomorphic features each separated from the other by the Indian River Lagoon (IRL) and Banana River (BR), respectively: (1) Peninsular Mainland, (2) Merritt Island, and (3) Cape Canaveral (Figure 1).

The Atlantic Coastal Ridge is present along the peninsula’s mainland shoreline. It is composed of a mixture of Pleistocene quartz sand and shell (unit Qa—Figure 2a), locally consolidated as coquina and part of the Anastasia Formation that is present along most of the east coast of Florida (Figure 2). The formation is about 5–20 m thick, with a local topographic relief of as much as 10 m at the Indian River Lagoon shoreline. The Anastasia grades westward into undifferentiated Quaternary quartz sand and shell (unit TQsu—Figure 2a) that is locally overlain by Holocene sediment (unit Qh—Figure 2a) associated with the St. Johns River basin. Seaward of the mainland is a cuspate foreland consisting of Pleistocene (Merritt Island) and Holocene (Cape Canaveral) beach-ridge complexes. Both are sea-level high-stand deposits composed of unconsolidated mixtures of quartz sand and shell that are about 5–10 m thick and exhibit a local relief of ~5 m. A thin layer (~1 m) of organic-rich, fine-grained sediment is present in the sub-tidal areas of the cuspate foreland and at locations where elevations are near the present sea level (e.g., inter-ridge troughs). The origin of Merritt Island and Cape Canaveral has been the subject of considerable scientific debate. Early hypotheses suggested that the formation was controlled by antecedent topographic relief created by a resistant bedrock formation [29] or an underlying structural feature [30,31]. Subsequent research suggested that the Holocene beach-ridge complex was created by converging littoral drift cells [32]. Most recent research has suggested that the cuspate foreland is an abandon paleo-delta of the St. Johns River [22]. However, to date, the debate over its origin has yet to be resolved.

The subsurface geology of the study area consists primarily of shallow marine Tertiary carbonates (e.g., limestone and dolostone), locally interbedded with siliciclastic sands [28]. The Tertiary sequence is subdivided into four formations, each bounded by an unconformable contact generated during intervals of subaerial exposure associated with sea-level low stands (Figure 2).

3. Data

Our study integrates three geodetic datasets, SAR, precise leveling, and continuous GNSS, to investigate the spatiotemporal pattern of VLM in Cape Canaveral and the surrounding areas. The SAR data used in this study are Sentinel-1 SAR scenes obtained from the Alaska Satellite Facility. We used 182 Sentinel-1 single-look complex (SLC) images along path 48 and frame 89 (Figure 1), acquired in ascending mode from July 2016 to June 2024. The data are archived at NASA Distributed Active Archive Center at the Alaska Satellite Facility [35].

The GNSS data are from four continuously operating sites with records spanning more than five years. Two of these stations (CCV5 and CCV6) are in Cape Canaveral, whereas the other two (COKO and TTVL) are situated in the Peninsular Mainland (Figure S1). Both GNSS stations in Cape Canaveral are ground-mounted and acquired data during 1998–2019. In the Peninsular Mainland, station TTVL is ground-mounted and acquired data during 2017–2024, while station COKO is mounted on a building and acquired data during 2004–2009. The GNSS data were processed by the Nevada Geodetic Lab (NGL) at the University of Nevada [36] using the MIDAS algorithm [37]. The GNSS data products are in reference to the International Terrestrial Reference Frame 2014 (ITRF14) [38]. In this study, we used the vertical component of the processed daily solutions.

The National Geodetic Survey (NGS) provided precise leveling data in Geographic Information System (GIS) shapefile format, spanning observations from 1932 to 2008 across Cape Canaveral, Merritt Island, and the Peninsular Mainland [39]. The dataset includes elevation measurements between successive benchmark stations, all in reference to the North American Vertical Datum of 1988 (NAVD 88). Our analysis focused on benchmark stations with multiple repeat measurements across different years. Comprehensive survey lines and their temporal sequences are illustrated in the Supplementary Materials (Figure S2). In addition, we analyzed aerial and spatial imagery to investigate the potential causes of the observed subsidence. The aerial/spatial imagery was sourced from the Florida Department of Transportation (FDOT)/Google Earth. The analysis of aerial imagery spanning was conducted through systematic visual inspection to identify temporal changes in land cover and infrastructure development. The quantitative GIS-based analysis was not feasible due to the lack of metadata (e.g., spatial resolution, ground control points, and sensor specifications) across the historical imagery collection. Nevertheless, the visual inspection approach proved effective in identifying major land-use changes that correlate with observed subsidence patterns.

4. Methods

4.1. Multi-Temporal InSAR Data Analysis

In this study, multi-temporal InSAR analysis was applied to Sentinel-1 data to detect land subsidence. The ISCE-2 (version 2.6.1) topsStack processor facilitated the co-registration of SAR scenes and the creation of interferograms stack with 5 nearest neighbor connections [40]. The application of a multi-look factor of 18 in the slant range and 6 in the azimuth direction yielded a ground pixel resolution of 90 m by 90 m. The Copernicus GLO30 digital elevation model (DEM) was used to remove the topographic phase component from the interferograms [41]. The Miami InSAR time series software in Python (MintPy) (version 1.6.1) was used to apply the small baseline subset (SBAS) approach on the interferogram network. The routine processing workflow for InSAR time series analysis, as presented in [42], was followed. Bridging and phase closure methods were implemented for correcting phase unwrapping errors. A weighted least-squares (WLS) inversion was performed to derive the line-of-sight (LOS) velocity for each pixel. The stratified tropospheric delay was corrected using the ERA5 reanalysis model from the European Centre for Medium-Range Weather Forecasts [43] with the PyAPS software (version 0.3.5) [44]. We validated the effectiveness of ERA5-based tropospheric correction by comparing interferograms before and after correction (Figure S2). Additionally, topographic residuals caused by DEM errors were mitigated [45], and noisy acquisitions were excluded during the estimation of the average LOS velocity by applying a residual phase root mean square cutoff value of three times the median absolute deviation. A temporal coherence threshold of 0.80 was applied in the resulting velocity map to mask unreliable pixels.

The LOS velocity ($V_{LOS}$) obtained by the InSAR analysis is a combination of velocities in the east–west ($V_{east}$), north–south ($V_{north}$), and vertical directions ($V_{vertical}$).

$$V_{LOS}=\left[V_{north}{V}_{east}{V}_{vertical}\right]\left[\begin{array}{c}\mathit{cos\; \alpha }.\mathit{sin}\beta \\ \mathit{sin}\alpha .\mathit{sin}\beta \\ \mathit{cos}\beta \end{array}\right]$$

(1)

where $\alpha$ and $\beta$ are the azimuth angle and the incidence angle of the satellite sensors, respectively. Since the study area is part of a passive continental margin, no differential horizontal tectonic movements are expected. Thus, we assume that the velocities are entirely in the vertical direction. We computed the vertical velocities ($V_{vertical}$) from LOS velocities ($V_{LOS}$) using the below equation:

$$V_{vertical}={\displaystyle \frac{V_{LOS}}{\mathit{cos}\beta }}$$

(2)

Multiple InSAR time series analyses were conducted by shifting the reference points to five distinct locations within the study area. Five reference points were selected at stable buildings in the northern and southern regions of the Peninsular Mainland, the northern and southern regions of Merritt Island, and the western region of Cape Canaveral. These locations were chosen based on their high temporal coherence and minimal expected ground motion. The identified localized subsiding areas (i.e., an area experiencing significantly higher rates of VLM compared to the surrounding region) remained consistent across all analyses, though areas with dense vegetation and wetlands exhibited noise due to phase unwrapping errors. This study presents results with the reference point located at a stable building structure in the southern Peninsular Mainland, as this location minimized noise from unwrapping errors due to a higher density of connected components in the region. The selection of a reference point within an area with densely connected components proved effective in reducing noise from unwrapping errors and providing more reliable VLM measurements.

4.2. Post-Processing of the GNSS Data

The GNSS daily solutions produced by the NGL enable us to calculate the site velocities of the four sites based on the vertical displacement time series. We post-processed the vertical displacement data to remove statistical outliers and correct for step discontinuities. Outliers were identified using a standard statistical approach based on the mean and standard deviation of the data, where values falling outside three standard deviations from the mean were removed iteratively. Step discontinuities, which can result from equipment changes, environmental factors, or seismic events, were addressed using a two-phase correction process. First, we identified the known steps using station metadata files. Second, we detected undocumented steps through the statistical analysis of the displacement time series. A step was identified when two criteria were met: the absolute difference between consecutive measurements exceeded 30 mm, and the difference in mean displacement before and after the suspected step (calculated using 100-day windows on either side) exceeded 20 mm. For each identified step, we computed the offset between pre- and post-step mean values and adjusted the post-step measurements to align with the pre-step mean. Vertical velocity was estimated using least-squares linear regression on the corrected time series, where the slope of the best-fit line represents the vertical velocity. The uncertainty in the vertical velocity was derived from the covariance matrix of the regression.

4.3. Leveling Data Analysis

Processing the precise leveling data was challenging because the data were acquired along different traverses in different years and rarely repeated measurements along the same traverse of benchmark stations (Figure S3). Nevertheless, we identified pairs of benchmark stations that were surveyed multiple times (Figure S4). Such pairs allowed us to calculate localized subsidence, in which one benchmark was used as a reference point. The relative height of a benchmark A with respect to a nearby benchmark B (reference benchmark) between two acquisition times ($t_1$ and $t_2$) is

$${∆h}_{A-B/1-2}=\left(h_{A2}-h_{B2}\right)-\left(h_{A1}-h_{B1}\right)$$

(3)

where $h_{A1}$, $h_{A2}$, $h_{B1},$ and $h_{B2}$ are the measured heights at benchmarks A and B at times $t_1$ and $t_2$. We assumed that elevation changes were all negative (subsidence) and accordingly selected the reference benchmark as the point with the least amount of elevation change. For each pair of benchmarks, we calculated the elevation changes over time to determine the relative velocity, which represents the average rate of elevation change. The uncertainty in relative velocity was computed using the formula outlined in [11]. The results of such an analysis provide only partial information on height changes, mainly on the amplitude of changes between the acquisition times.

5. Results

5.1. InSAR

InSAR processing of the July 2016 to June 2024 Sentinel-1 data revealed a patchy and heterogeneous subsidence pattern in the Cape Canaveral area (Figure 3). The patchy pattern reflects our choice to present only reliable results, determined by a temporal coherence of the time series results using a threshold of 0.80. The highly temporal coherence pixels (i.e., ≥0.80) occurred mostly over strong scattering land cover, such as unvegetated and built environments. SAR scattering over vegetated areas is highly affected by interferometric decorrelation and, hence, does not yield reliable results. The vertical velocity field was calculated with respect to a reference point building located in the southern part of the Peninsular Mainland (marked by a plus sign in Figure 3). Consequently, the results obtained from the SBAS InSAR processing express differential movements of the measured points relative to this reference point. The negative values indicate subsidence, and the positive values indicate an uplift. The mean standard deviations of all the vertical velocities are smaller than 1 mm/year.

In general, the Cape Canaveral infrastructure (e.g., administrative buildings, hangars, and support facilities) exhibited no significant vertical movement. However, our InSAR analysis identified three localized subsiding areas located in the western, central, and southeastern parts of the study domain (marked by white triangles in Figure 4). The western area, which contains solid and liquid fuel storage facilities, exhibits vertical velocities between −6 and −8 mm/yr (indicated by the white triangle (b) in Figure 4). The central area, located along the runway at Cape Canaveral Space Force Station, shows vertical velocities ranging between −4 and −7 mm/yr (indicated by the white triangle (c) in Figure 4). The southeastern area, encompassing infrastructure facilities, wetlands, and natural coastal features, demonstrates vertical velocities ranging from −4 to −9 mm/yr (indicated by the white triangle (d) in Figure 4). The InSAR analysis did not detect any significant vertical movement or localized subsiding area in Meritt Island. The Peninsular Mainland was also found to be generally stable with a localized subsidence detected at the Space Coast Regional Airport located in Titusville (Figure S5).

5.2. GNSS and Precise Leveling

An analysis of continuous GNSS and historical leveling data provided important insights into the long-term VLM patterns across the study area. The continuous GNSS data analysis revealed the distinct regional patterns of VLM. In Cape Canaveral, stations CCV5 and CCV6 indicated ongoing subsidence with vertical velocity rates of −2.49 ± 0.32 mm/yr and −2.68 ± 0.78 mm/yr, respectively (Figure 5a,b). In contrast, stations in the Peninsular Mainland exhibited minimal to negligible vertical motion: station TTVL showed 0.10 ± 0.15 mm/yr, while station COKO recorded −0.81 ± 0.10 mm/yr (Figure 5c,d).

Historical leveling analysis revealed temporally variable subsidence patterns. In Cape Canaveral, measurements showed high subsidence rates (~5 mm/yr) during the 1950s–1960s, which decreased to lower rates (~2 mm/yr) during the 2000s (Figure S6). In Merritt Island and the Peninsular Mainland, definitive vertical velocity patterns could not be established due to limited temporal coverage and insufficient repeated measurements along common benchmarks (Figure S4).

6. Discussion

We utilized SAR, continuous GNSS, and precise leveling geodetic datasets to comprehensively assess the spatial and temporal patterns of VLM in Cape Canaveral and its surrounding areas. The results from all three measurement techniques are summarized in

Table 1. Summary of vertical land motion (VLM) rates (mm/yr) and observation periods across study areas based on InSAR, GNSS, and leveling measurements.

Years	1950–1970	1999–2006	1998–2019	2016–2024
Method	Leveling	Leveling	GNSS	InSAR
Cape Canaveral	−5 to −6	−1 to −2	−2.5 to −2.7	−9 to 3
Merritt Island				−3 to 3
Peninsular Mainland			−0.8 to 0.1 (2004–2024)	−5 to 3

, revealing distinct spatial and temporal variations in subsidence patterns.

The InSAR analysis (2016–2024) identified three localized subsiding areas in Cape Canaveral, while most areas in Merritt Island and the Peninsular Mainland remained relatively stable. The GNSS station measurements confirmed subsidence in Cape Canaveral, with VLM rates ranging from −2.49 to −2.68 mm/yr. In contrast, stations in the Peninsular Mainland exhibited low VLM rates: station TTVL showed 0.10 mm/yr, while station COKO recorded −0.81 mm/yr.

The precise leveling data revealed significant temporal variations in Cape Canaveral’s subsidence pattern, with subsidence rates (~5 mm/yr) during the 1950s–1970s decreasing to lower rates (~2 mm/yr) during the 2000s. The variation in the VLM range across different methods is due to the inherent characteristics of each method, such as the non-overlapping observation periods, the different reference frames used by each technique (localized for InSAR, IGS 14 for the GNSS, localized for leveling), and the distinct measurement characteristics of the InSAR (spatially averaged and surface-sensitive), the GNSS (point-specific, deep-anchored), and leveling (traverse specific and surface-sensitive). The InSAR detects average elevation changes across multiple scatterers within a pixel (approximately 90 m in this study) and is particularly sensitive to shallow subsurface changes. In contrast, GNSS measurements reflect point-specific motion influenced by the depth of monument installation. Similar leveling (in this study) reflects land motion between benchmark stations. However, the combination of these geodetic datasets provides a comprehensive overview of VLM in the study area. While the Peninsular Mainland and Merritt Island are predominantly stable, Cape Canaveral is experiencing persistent subsidence from the late 1950s to the present day.

6.1. Long-Term Subsidence in Cape Canaveral

Our multi-dataset analysis highlights persistent long-term subsidence in Cape Canaveral from the late 1950s to the present day, in contrast to the relatively stable conditions observed in the Peninsular Mainland and Merritt Island. To understand the mechanisms driving long-term subsidence in Cape Canaveral, we conducted a comprehensive review of recent and historical studies focusing on VLM along the U.S. East Coast and geological/geomorphological assessments specific to Cape Canaveral. Studies on the Holocene sea-level database have reported subsidence rates of 0.5–2 mm/yr along the coast from Maine to northern Florida, primarily due to GIA [18,19,20]. The glacio-isostatic corrections using ¹⁴C data from tide gauges near Daytona Beach and Miami Beach have reported subsidence rates of 1.86 mm/yr and 0.69 mm/yr, respectively [46]. Given that the tide gauge near Daytona Beach is located approximately 85 km north of Cape Canaveral, and the one near Miami Beach is about 300 km to the south, these subsidence rates suggest that GIA has likely influenced the Cape Canaveral region as well. Further evidence comes from a global GNSS dataset specifically developed for GIA applications, which reported a GIA-induced VLM rate of −0.59 ± 0.16 mm/yr at station SG05, located approximately 25 km south of Cape Canaveral [47]. These observations collectively suggest that GIA contributes a baseline subsidence of approximately 0.5–0.7 mm/yr across the Cape Canaveral region. This GIA estimate is supported by direct empirical evidence from GNSS station COKO in the Peninsular Mainland, which shows a VLM rate of −0.81 ± 0.10 mm/yr. This in situ station measurement validates our regional GIA subsidence estimates. However, GIA represents only one component of the subsidence budget.

The geological context provides additional insights into subsidence patterns. Sediment compaction rates are known to vary considerably depending on factors such as sediment age, thickness, composition, and overburden pressure [48]. Additionally, the subsurface architecture of the coastal plain plays a crucial role, as deeper channel incisions can lead to more pronounced differential subsidence compared to flatter areas with more uniform sediment deposition [49]. Cape Canaveral consists of late Holocene coastal sand deposits (Figure S7) organized into 16 distinct beach-ridge sets that decrease in age toward the Atlantic coastline [23]. Hence, the southeastern region, characterized by younger coastal sediments relative to areas further west (Figure S3), is likely compacting faster and thereby exhibiting higher subsidence rates. Overall, the long-term subsidence observed in Cape Canaveral is likely influenced by a small component of GIA but is predominantly driven by the compaction of younger, unconsolidated siliciclastic sediments.

6.2. Localized Subsiding Areas and Apparent Stable Infrastructures—Observations and Explanations

The InSAR time series analysis of Sentinel-1 SAR data revealed a patchy spatial pattern of subsidence. Our results identified three localized subsiding areas in Cape Canaveral and one in Peninsular Mainland. It is important to note that the InSAR observations of localized subsidence primarily reflect the near-surface movements of scatterers on the ground. This is because the radar signal interacts with surface features, such as buildings, roads, and vegetation, which act as dominant scatterers. The phase shift measured by the InSAR corresponds to the cumulative displacement of these surface reflectors relative to the satellite’s line of sight (LOS). Therefore, the detected subsidence represents near-surface deformation rather than deeper subsurface processes unless deep-seated structures (e.g., compaction of thick geological layers or fault displacements) cause surface displacements. In regions with strong, stable scatterers, such as buildings with deep foundations, the InSAR may underestimate or fail to detect vertical movement if deeper processes do not translate to observable surface changes. This limitation is particularly relevant for our study area, where we must consider both surface and potential deeper deformation mechanisms. In this section, we describe the subsidence detected in the three localized subsiding areas and use geographic (land cover change) and geologic (stratigraphy) context to explain the observed subsidence.

The localized subsiding area, in the western part of Cape Canaveral, shows a 6–8 mm/yr subsidence near fuel storage facilities. Historical aerial imagery reveals that this site, originally a wetland in 1951, was transformed into a developed area by 1969 to support early space infrastructure (Figure 6). Construction activities, including the clearing of wetland vegetation, likely initiated shallow subsidence by destabilizing the organic-rich soils through root collapse and promoting oxidation upon exposure to air [50]. Comparable studies, such as those in wetland-drained regions [51,52], attribute similar subsidence to the rapid decomposition of organic matter and the loss of soil structure, often yielding high subsidence rates that diminish over decades [53]. Given that the wetland alteration here occurred between 1951 and 1969, the subsidence observed from 2017 to 2024 likely reflects a residual, slower phase of this process, consistent with the depletion of compressible organic layers over time. The steady subsidence trend in the InSAR time series (Figure 4b) and its spatial alignment with the former wetland footprint (Figure 6) support this interpretation.

The localized subsiding area, detected in the central part of Cape Canaveral, shows a 4–7 mm/yr subsidence along the runway. Google Earth imagery reveals significant runway expansion activities conducted between 2015 and 2018 (Figure 7). Such expansion activities typically involve extensive ground modification processes that can trigger subsidence through multiple mechanisms extending beyond the immediate construction area. The primary mechanisms include soil consolidation, where new structural loads from runway expansion and infrastructure development compress the underlying sediments, and construction-related dewatering, which lowers the groundwater table and reduces the pore pressure support in the soil [15,54,55,56]. Additionally, the introduction of impervious runway surfaces alters natural drainage patterns, preventing water infiltration and potentially triggering subsidence as the water content of the underlying sediments decreases. These combined mechanisms with the continuous load from runway operations and infrastructure aging explain the observed subsidence during 2017–2024. Additionally, localized subsidence detected near the Space Coast Regional Airport in the Peninsular Mainland is similarly linked to runway expansion and infrastructure development (Figure S4).

The localized subsiding area identified in the southeastern part of Cape Canaveral exhibits subsidence rates of 4–9 mm/yr. The InSAR scatterers associated with this hotspot are primarily roads and small buildings that have been present since 2000 or even before (Figure 8). Unlike the major infrastructure in this area, which is supported by deep pile foundations, these structures were built on shallow foundations and thus directly reflect the shallow ground deformation of relatively young, unconsolidated siliciclastic sediments [23] (Figure S7), which are more susceptible to natural compaction than older sequences located to the west.

The spatial pattern of subsidence in Cape Canaveral reflects the complex interaction between near-surface geology and anthropogenic modifications. The observed subsidence is primarily associated with the mechanical compaction of near-surface (<10 m) unconsolidated Holocene and Pleistocene sedimentary sequences, particularly affecting loosely compacted Holocene sands and organic-rich deposits [23]. While older, deeper sediment sequences are present beneath the study area, they are less susceptible to deformation from localized surface loading due to their greater consolidation.

The apparent stability of building infrastructure in Cape Canaveral and Kennedy Space Center can be explained by two factors. First, in dense urban settings, InSAR measurements primarily reflect building movements due to the double-bounce backscattering effect. Second, and more importantly, the major infrastructure in these areas is supported by pile foundations extending up to 164 feet deep, reaching more stable, consolidated strata below the compressible near-surface layers [57]. The overall pattern suggests that subsidence in Cape Canaveral results from a combination of natural processes (compaction of young, unconsolidated sediments) and anthropogenic activities (infrastructure development and land-use changes), primarily affecting the shallow subsurface layers.

6.3. Temporal Variations Based on the GNSS and Precise Leveling

The integration of the GNSS and precise leveling data reveals a significant temporal evolution in Cape Canaveral’s ground deformation patterns, providing insights into both long-term trends and historical development impacts. The consistent linear trend observed in the GNSS time series from stations CCV5 and CCV6 (−2.5 to −2.7 mm/yr, 1998–2019) suggests the establishment of a steady-state deformation regime in recent decades. This modern deformation rate likely represents the combined effects of natural processes such as sediment compaction and GIA, rather than anthropogenic influences. However, historical leveling data reveal a more dynamic period of ground deformation, with rates decreasing from approximately 5 mm/yr during the 1950s–1970s to around 2 mm/yr in the 2000s. This temporal variation appears to track Cape Canaveral’s development history—the higher rates coincide with the period of intensive infrastructure construction for the space program. The subsequent decrease to rates more closely matching modern GNSS measurements suggests a transition from construction-induced rapid subsidence to a more stable regime dominated by natural processes. The convergence of rates between historical leveling and modern GNSS measurements in the 2000s not only validates both datasets but also indicates that Cape Canaveral’s ground deformation has largely stabilized following its initial development period. This temporal evolution pattern provides crucial context for understanding the area’s long-term stability and has important implications for infrastructure management and future development planning.

6.4. InSAR and GNSS Comparison

Our InSAR vertical velocity estimates for pixels near the GNSS stations in Cape Canaveral and the Peninsular Mainland showed good agreement with GNSS-derived vertical velocity rates. The comparison of VLM rates from the GNSS station and the InSAR is listed in

Table 2. Comparison of vertical land motion (VLM) rates (mm/yr) for GNSS and InSAR measurements.

Station	GNSS		InSAR
	VLM (mm/yr)	Years	VLM (mm/yr)	Years
CCV5	−2.49 ± 0.32	August 1998–September 2019	−3.86 ± 0.33	July 2016–June 2024
CCV6	−2.68 ± 0.78	August 1998–September 2019	−3.86 ± 0.33	July 2016–June 2024
TTVL	0.10 ± 0.15	July 2017–September 2024	−1.15 ± 0.44	July 2016–June 2024
COKO	−0.81 ± 0.10	February 2004–December 2009	−1.58 ± 0.45	July 2016–June 2024

.

In Cape Canaveral, both the InSAR and GNSS showed subsidence, with GNSS stations CCV5 and CCV6 recording VLM rates of −2.49 ± 0.32 mm/yr and −2.68 ± 0.78 mm/yr, respectively, while nearby InSAR pixels showed slightly higher VLM rates of −3.86 ± 0.33 mm/yr. In the Peninsular Mainland, GNSS station TTVL recorded VLM rates of 0.10 ± 0.15 mm/yr with corresponding InSAR measurements showing VLM rates of −1.15 ± 0.44 mm/yr, while station COKO showed VLM rates of −0.81 ± 0.10 mm/yr with nearby InSAR pixels indicating VLM rates of −1.58 ± 0.45 mm/yr. These variations are well within the typical ±1–2 mm/yr agreement threshold generally accepted for InSAR-GNSS comparisons, especially considering InSAR’s inherent sensitivity to atmospheric noise, temporal decorrelation, and the challenges in maintaining phase coherence in vegetated areas.

The overall consistency between the InSAR and GNSS measurements validates our findings and strengthens our confidence in the identified deformation patterns. The small differences observed can be attributed to several factors: the non-overlapping observation periods, the different reference frames used by each technique, and the distinct measurement characteristics of the GNSS (point-specific, deep-anchored) versus the InSAR (spatially averaged, surface-sensitive). These complementary aspects of the two techniques, along with their general agreement within expected uncertainty ranges, provide robust validation of the observed deformation patterns in the study area.

6.5. Comparison with Previous Studies

Our analysis of VLM in Cape Canaveral complements and advances previous studies, revealing new insights into the spatial and temporal complexity of ground deformation. Recent work developed a comprehensive VLM dataset for the region by combining multiple data sources: Advanced Land Observing Satellite-1 (2007–2011), Sentinel-1 (2015–2020), and GNSS measurements through a joint inversion approach [17]. Their analysis in Cape Canaveral yielded reliable measurements primarily along the coastal boundary (Figure S8b). Our analysis revealed three previously undetected localized subsiding areas, providing new insights into the area’s deformation patterns. Furthermore, their findings of subsidence in Merritt Island and the Peninsular Mainland contrast with our results showing relative stability in these areas. A detailed quantitative comparison between these studies is provided in Table S1 of the Supplementary Materials. However, direct comparisons between the studies are complicated by differences in observation periods (2007–2020 vs. 2016–2024) and methodological approaches (joint inversion ALOS, Sentinel-1, and GNSS vs. SBAS Sentinel-1).

The temporal evolution of VLM rates revealed by our study provides critical context for understanding historical observations. Pre-space geodesy studies using precise leveling data referenced to benchmarks in Maine documented a regional subsidence rate of 4 mm/yr in Cape Canaveral [11,12]. Our analysis enhances this understanding by revealing temporal variability in the deformation pattern: from localized rates of approximately 5 mm/yr during the intensive infrastructure development of the 1950s–1970s to reduced rates of around 2 mm/yr by the 2000s. The consistency between our measured rates during the 1950s–1970s and these historical observations (5 mm/yr versus 4 mm/yr) validates both studies.

Overall, our study provides a detailed investigation of the specific mechanisms (both natural and anthropogenic factors) driving both long-term and localized subsidence in Cape Canaveral. We identify three distinct localized subsiding areas that were not reported in [17,58] and also analyze the temporal variability of VLM, showing that subsidence rates were higher (~5 mm/yr) during the intensive infrastructure development of the 1950s–1970s but decreased to ~2 mm/yr by the 2000s. Our findings establish a contextual understanding of the complex interplay between natural processes (sediment compaction and GIA) and anthropogenic factors (infrastructure development and land-use changes) that will be valuable for future predictive models incorporating both natural and human-induced subsidence processes.

6.6. Future Work

A reliable estimate of both the magnitude and spatial variability of VLM in Cape Canaveral and its surrounding areas is crucial due to the presence of critical infrastructure in this region. In this study, the InSAR vertical velocities are relative to a selected reference point (see Section 5.1 InSAR) and are not tied to a stable reference frame, such as IGS14. The sparse distribution of GNSS stations, only two in Cape Canaveral (located in close proximity to each other) and two in the Peninsular Mainland, with none on Merritt Island, along with their limited temporal overlap with Sentinel-1 data, posed challenges in establishing absolute reference frames for InSAR measurements. The leveling analysis was constrained by significant limitations: precise leveling data were acquired along varied traverses across different years, with rarely repeated measurements along the same benchmark stations. Additionally, the lack of temporal overlap between precise leveling data (1932–2008) and InSAR data (2016–2024) prevents the direct calibration of subsidence rates across these periods. We address this by using each dataset’s strengths to track long-term changes rather than comparing rates directly. Precise leveling data provide a historical baseline, showing subsidence rates of ~5 mm/yr in the 1950–70s and ~2 mm/yr in the 2000s, reflecting Cape Canaveral’s development phase. InSAR data from 2016 to 2024 offer high-resolution spatial patterns of recent subsidence, highlighting localized subsiding areas. GNSS data (1998–2019), overlapping with InSAR data from 2016 to 2019, act as a temporal bridge, validating InSAR rates near stations and matching the leveling’s later rates (~2 mm/yr in 2000s).

Moreover, the InSAR and GNSS measurements inherently capture different aspects of land motion. The InSAR detects average elevation changes across multiple scatterers within a pixel (approximately 90 m in this study) and is particularly sensitive to shallow subsurface changes. In contrast, the GNSS measurements reflect point-specific motion influenced by the depth of monument installation and are less sensitive to shallow subsurface changes. To address these limitations and better understand the true spatial variability of land subsidence, we recommend installing a dense network of GNSS stations at varied depths throughout Cape Canaveral and Merritt Island. Such a network would enable a more accurate integration of the InSAR and GNSS observations. Additionally, contextual knowledge of the ongoing natural and anthropogenic processes driving VLM in the region is vital to effectively integrate the InSAR and GNSS observations. Future studies should also incorporate phreatic groundwater levels and aquifer system dynamics to fully elucidate subsidence mechanisms, particularly where wetland drainage and construction-related dewatering contribute to shallow subsidence. Despite these limitations, InSAR analysis has proven highly effective in identifying localized subsiding areas that might otherwise remain undetected.

6.7. Implications for Flooding Hazard in Cape Canaveral

The identified localized subsiding areas in Cape Canaveral have significant implications for flood hazard assessment when considered alongside documented sea-level rise trends. The global mean sea level has increased by ~1.5 mm/yr since 1900 [59,60], with an accelerated rate exceeding 10 mm/yr since 2010 along the U.S. southeast coast [61]. In Cape Canaveral, since the installation of the Trident Pier tide station in 1994, local sea levels have increased by 0.5 feet, with projections indicating potential increases of 1.4 to 7.8 feet by the end of the century [62]. This progressive rise in sea levels intensifies Cape Canaveral’s vulnerability to flooding, particularly during high-tide events and storm surges. Furthermore, the compound effect of land subsidence and sea-level rise is particularly concerning for three reasons. First, the barrier island configuration of Cape Canaveral, situated between the Atlantic Ocean and the BR part of the IRL, makes it vulnerable to flooding from both sides. Notably, the BR, which borders Cape Canaveral and Merritt Island, experiences additional stress due to restricted hydrological connectivity and heightened water levels from tidal influxes. Second, the relative sea-level trend at Trident Pier Tide Station has accelerated from 6.42 ± 0.58 mm/yr (1993 to 2002) to 9.59 ± 1.64 mm/yr (2003 to 2022) [63]. Third, the topographic characteristics of the area, particularly in low-lying lagoon-adjacent zones, are already experiencing impacts through reduced drainage capacity and heightened groundwater elevations.

The localized subsiding area in the western part, near fuel storage facilities and the Space Force Station runway, respectively, are vulnerable to elevated flood risk due to their proximity to critical infrastructure (i.e., Satellite Processing and Storage Area). These areas may face an increased flood risk due to the combined effects of subsidence, sea-level rise, and potentially compromised stormwater drainage systems. Most of the area’s stormwater infrastructure was implemented between 1950 and 1980, when water levels were lower and development was less extensive [64,65]. Consequently, access roads linking the runway to nearby infrastructure could be impaired during storm surges due to inefficient drainage. The southeastern subsiding area, a low-lying coastal zone near the Atlantic, faces a heightened flood risk due to its elevation and subsidence rates. This area includes launch complexes 17, 18, 31, and 32, located between the central and southeastern subsiding zones. Recent projections indicate that the local sea level in Cape Canaveral could rise between 0.98 and 1.35 ft under Intermediate–Low and Intermediate–High scenarios by 2050 [66]. The combined impact of subsidence and projected sea-level rise poses a risk to launch operations by increasing the likelihood of flooding along access routes and critical support infrastructure. Overall, the compound effect of projected sea-level rise and our observed subsidence rates (−4 to −9 mm/yr in three localized subsiding areas) indicates an accelerating increase in flood vulnerability, posing severe operational challenges for Cape Canaveral’s infrastructure in the coming decades.

7. Conclusions

This study employed three geodetic datasets to comprehensively analyze the spatial and temporal patterns of VLM in Cape Canaveral and its surrounding areas. Our InSAR analysis revealed three localized subsiding areas in Cape Canaveral, with unique driving mechanisms: wetland soil oxidation in the western area (−6 to −8 mm/yr), infrastructure development impacts in the central area (−4 to −7 mm/yr), and natural sediment compaction in the southeastern region (−4 to −9 mm/yr). In contrast, Merritt Island and the Peninsular Mainland exhibited stability, with GNSS measurements in the mainland showing modest subsidence rates consistent with regional GIA estimates. Historical leveling data demonstrated significant temporal variations in subsidence rates, from high rates of ~5 mm/yr during the 1950s–1970s to a reduced rate of ~2 mm/yr in the 2000s, reflecting the dynamic nature of VLM in the region. The integration of multiple geodetic techniques provided robust evidence that while GIA contributes a baseline subsidence across the region, localized subsidence patterns are predominantly controlled by site-specific factors, including sediment compaction and land-use changes. These findings have critical implications for infrastructure management and coastal resilience planning, particularly given the compound effects of land subsidence and projected sea-level rise. Future research should prioritize continuous monitoring through enhanced GNSS networks and regular InSAR observations, especially in areas targeted for development. Additionally, developing comprehensive predictive models that incorporate both natural processes and anthropogenic factors will be crucial for ensuring the long-term stability and sustainability of this strategically important coastal region.