Relating Geotechnical Sediment Properties and Erodibility at a Sandy Beach

Coastal areas are subject to increasingly difficult challenges due to climate change. Sea level rise and increases in the frequency and severity of coastal storms are expected to lead to increasing beach erosion, shoreline retreat, and flood events (Sallenger 2000; Zhang et al. 2004; Jones et al. 2009; Bender et al. 2010; Knutson et al. 2010). Central to improving the prediction and mitigation capabilities of these impacts is enhancing the understanding of sediment dynamics in coastal areas. Specifically, the role of sediment properties beyond grain size, for example, geotechnical sediment properties and their relationship to driving mechanisms of erosion, is still not well understood. Such relationships have the potential to improve the prediction of sediment dynamics and to help with mitigating climate-related coastal change and impacts on coastal communities.

Erosion is typically modeled as a fluid–particle interaction, and thus, grain size and particle density predominantly govern the prediction of erodibility (Shields 1936; Briaud et al. 2001). For fine-grained sediments, there have been numerous attempts to correlate a wide variety of sediment properties with erodibility (NASEM 2019). Properties such as water content, unit weight, plasticity index, undrained shear strength, void ratio, swell potential, and fines fraction, among others have been tested with various degrees of success (Cao et al. 2002; NASEM 2019). Developing empirical relationships between critical shear stress and the aforementioned soil properties from remolded laboratory samples represents the state of the art for determining the erosion behavior of fine-grained sediments, and these types of relationships are often site-specific, owing to the complex nature of the erosion problem and the influence of a large number of sediment-specific properties on erodibility (Briaud 2013). For coarse-grained sediments, sands, and gravels, grain size is accepted as the controlling parameter for erodibility relationships (Briaud 2013). For many beach environments, sand is the predominant sediment type, and thus, sediment dynamics are predicted mostly based on grain size, if sediment properties are considered at all. For example, Cheng et al. (2023) examined the relationship between beach morphology, sediment size, and sediment shape and found that sediment size and shape had an impact on sediment transport regimes and could indicate different types of beach morphology. However, geotechnical sediment properties such as moisture content, density, and shear strength have been shown to be important for a variety of related coastal processes such as dune retreat, tidal flat evolution, and Aeolian sediment transport (Davidson-Arnott et al. 2005; Erikson et al. 2007; Sassa and Watabe 2007; Zambrano-Cruzatty et al. 2019) and may be particularly important in the intertidal zone that is subject to saturation and drainage and variations in forcing conditions on different time scales (waves, tides, inundation events). While geomorphodynamics of coastal areas are driven by local hydrodynamics (Dalrymple 1992; Masselink and Short 1993; Dean and Dalrymple 2004; Masselink et al. 2006; van der Lugt et al. 2019; Hsu et al. 2023), it is hypothesized that including geotechnical sediment properties in erodibility relationships could further improve the prediction of coastal change by understanding sediment stabilizing forces better.

Challenges of conducting geotechnical measurements in energetic coastal environments have been approached in recent years. Tools such as portable free-fall penetrometers (PFFPs) have enabled conducting geotechnical surveys of energetic coastal areas (Stark and Kopf 2011; Stark et al. 2014; Albatal and Stark 2017; Albatal et al. 2019). When deployed in submerged coastal areas, PFFPs have been used for sediment characterization and for estimating the bearing strength, relative density of sands, undrained shear strength, and soil behavior under dynamic loading (Akal and Stoll 1995; Chow et al. 2018; Lucking et al. 2017; Albatal et al. 2020; Kiptoo et al. 2020). More recently, PFFPs have been tested for geotechnical surveying in emerged intertidal environments (Reeve et al. 2018; Brilli and Stark 2020). The ability of a single tool to collect geotechnical measurements seamlessly from the emerged to submerged coastal zones is essential to integrate geotechnical sediment properties in sediment transport prediction and simulation at beaches. The goal of this study was to relate geotechnical sediment properties, erodibility parameters, morphodynamics, and hydrodynamic regime in a sandy beach environment. To this end, a field survey was conducted at the United States Army Corp of Engineers Field Research Facility (USACE FRF) in October 2021 in the framework of the During Nearshore Event Experiment (DUNEX). The field survey included in situ testing and the collection of sediment samples for subsequent laboratory erodibility testing. Based on the collected field data and physical sample analysis, sediment properties and mobility were assessed and related to bed-elevation change.

Fig. 3 shows the measured significant wave height from a bottom-mounted acoustic wave and current sensor (AWAC) over the course of the week (US Army Corp of Engineers 2022). During a storm event on October 11, significant wave heights were > 2 m and gradually calmed over the next 5 days, reaching heights of <1 m by October 15 [Fig. 4(a)]. The peak period mostly fluctuated between 7 and 10 s with a decreasing trend toward the end of the week [Fig. 4(b)].

Fig. 5 displays morphological change over the course of the survey from ground-based LiDAR measurements (US Army Corp of Engineers 2022). Transect A underwent the most morphologic change over the course of the week, with deposition occurring from October 11 to October 13, with Station A4 seeing 60 cm of deposition, followed by slight erosion from October 13 to October 15 [Fig. 5(a)]. Transect B remained relatively stable, with erosion/deposition magnitudes on the order of <20 cm [Fig. 5(b)]. Transect C also remained stable in the upper reaches, with a notable deposition area on the order of 40 cm from October 13 to October 15 [Fig. 5(c)]. The diamonds on each transect represent the highest water level seen on the transect for that particular survey interval using total water level data available at the site (US Army Corp of Engineers 2022). Water levels were measured hourly, and the data points presented in Fig. 5 are the highest water levels observed in the 24 h between surveys. Any stations below the maximum water level were subjected to hydrodynamic forcing, and thus, only these stations were considered for further analysis. A possible contributor to the variation between transects in close proximity (−200 m spacing) is longshore geomorphodynamic features, mostly the presence of beach cusps (see Fig. 1 for a historical qualitative example) on the order of 5 m in the alongshore direction and the position of the transect in relation to the cusp. The presence of cusps was also apparent in the LiDAR scans taken during the survey period. Beach cusps are a common occurrence at this site, typically forming in the days following storm events, with a mean spacing of 12.4–38.8 m (Holland 1998). While longshore processes were not within the scope of this work, it is acknowledged that these processes affect morphology.

Fig. 6(a) depicts a histogram of the sample median grain sizes (d₅₀) for all samples taken over the course of the week. Samples were taken around low tide on all days. All of the samples were classified as poorly graded sand (SP) based on grain size analysis and the Unified Soil Classification System (USCS; ASTM D2487/D2487-17, ASTM 2017a). A majority of the samples had a d₅₀ classifying as fine sand via USCS, delineated by particles passing a No. 40 sieve (0.43 mm) and retained on a No. 200 sieve (0.074 mm) (dashed lines, Fig. 6) (ASTM D2487/D2487-17, ASTM 2017a). Fig. 6(b) shows the cross-shore variation of d₅₀. On the x-axis, zero indicates the dune toe, with the swash zone located near the 25–30 m station, depending on the low tide level on a given day. As seen from the figure, d₅₀ exhibited little variation in the cross shore until the lower intertidal zone where grain size and variability increased. Fig. 6(c) shows examples of grain size distributions in the swash zone of Transect C throughout the week, highlighting the variability of particle sizes.

Table 2 shows the results of the incipient motion analysis. Using grain sizes from the 51 stations that were shown to be inundated during the surveys, u_cr was calculated via the two methods presented in Section “Incipient Motion Analysis,” and the range of threshold velocities is presented. Additionally, u_cr was computed from measurements of τ_cr from the JET samples, and the range of these values is also presented. ADV data from sensors in the swash zone were available during the survey period to characterize the in situ flows. Typical values of flow velocity, u, ranged from 60 to 80 cm/s on October 11, decreasing to 30 cm/s on October 15, corresponding to the decrease in wave heights over the course of the week (Fig. 4). Maximum flows were on the order of 300 cm/s. As seen from Table 2, the measured flows greatly exceeded the threshold for motion for the entirety of the survey period.

Fig. 7 shows the relationship between firmness factor, qsbc, and void ratio. In the plot, the shading of each data point differs with regard to the measured moisture content. Firmness factor [Fig. 7(a)] and qsbc [Fig. 7(b)] both correlate negatively with the void ratio. Additionally, samples with higher densities (low e), qsbc, and FF consistently had higher water contents. While this may appear counterintuitive, this can be explained by the location of these samples, which were found in and near the swash zone and, thus, were near or at full saturation (Fig. 7).

Changes in sediment properties were then compared to morphologic change. Fig. 8 depicts changes in the bed levels at inundated stations as a function of corresponding changes in void ratio [Fig. 8(a)], firmness factor [Fig. 8(b)], and qsbc [Fig. 8(c)]. Changes in properties and bed levels were computed over the 24 h interval between surveys. Of the 25 inundated stations where LiDAR data were available along with sediment and strength properties to make comparisons, 22 stations saw zero change or deposition, while only 3 stations appeared erosional. As seen in the figure, there are no readily apparent trends between the soil properties and morphologic changes. However, it can be observed that: (1) all sites where void ratio decreased the bed level increased, but negative changes in void ratio were overall smaller than the observed positive changes; and (2) an increase in bed level was more often accompanied by a decrease in beach surface strength or hardness (i.e., decrease in firmness factor and qsbc), while an increase in firmness factor or qsbc was mostly associated with little change in bed elevation.

Finally, sediment packing and strength properties, e, γ_total, FF, and qsbc were compared to the detachment rate coefficient, k_d, from the JET experiments (Fig. 9). As mentioned in the “Laboratory Testing” section, k_d was determined through linear regression of the raw erosion rate/shear stress data. Each data point represents a location where measurements were taken for the PFFP, a 25-cm JET sample, and a 10-cm density sample simultaneously. Scattering was low, with R² ranging from 0.80 to 0.99, with 10 of the 13 samples tested having an R² value > 0.97. For the directly measured sediment properties, increasing void ratio and decreasing total unit weight correlated well with increasing k_d [Figs. 9(a and b)]. Total unit weight was included here to capture the effects of water content. Both properties were similarly well-correlated with k_d, with R² = 0.87 for e and R² = 0.84 for γ_total. Penetrometer measurements, FF and qsbc, were both negatively correlated with k_d [Figs. 9(c and d)]. Although these relationships were generally less well-defined than the sediment properties, qsbc provided a stronger correlation than FF, with R² = 0.67 and R² = 0.43, respectively. These four properties were also compared to τ_c from the JET experiments, but no significant correlation was observed.

The modeled values of u_cr compared to the measured flows, u, show that motion would be initiated under the hydrodynamic forcing for all inundated stations sampled (Table 2). The range of threshold velocity, u_cr, from the JET was larger than the two grain-size-based methods, but it still predicted motion for all samples. This result is consistent with theory and the observations because sand is the easiest particle size to move on the Shields curve and geomorphological change is observed throughout the survey period (Shields 1936). Having the estimates of u_cr from the JET is a useful addition to the analysis because this test method captures the in situ properties and variability in each physical sample more thoroughly than grain-size-based estimates. The measured flows are consistent with the literature because swash bore velocities have been observed to travel at up to 3 m/s, matching the maximum velocities observed in this study (Nielsen 1992; Puleo et al. 2000). While the presented methodology provides a useful proof of initiation of motion, and thus, sediment mobility at this sandy beach, it offers little information on the actual geomorphodynamics since erosion, deposition, and fairly unchanged conditions have been observed at the different locations throughout the short survey period.

The results presented in Fig. 8 indicate that changes in the void ratio and PFFP strength properties exhibit little to no direct relationship with morphologic change. The relatively limited number of data points available for comparison (i.e., 25) compared to the total number of soil measurements in the data set (i.e., 118) presents one of the challenges when developing more reliable relationships. This can be attributed to a variety of factors. First, not all stations were exposed to hydrodynamic forcing because they were above the high-water mark between survey intervals (Fig. 4), and thus, these stations were removed from the analysis, leaving 52 stations remaining. Not all hourly LiDAR coverage was available from the survey period because the high traffic of people and equipment on the beach rendered many of the scans unusable, requiring the available data to be averaged over the course of each day. Further, the limited available data were significantly biased toward deposition over erosion (22 depositional sites to 3 erosional sites). Results shown in Fig. 8 suggest that decreases in void ratio tended to predict deposition and that deposition was accompanied by a decrease in beach surface strength. However, these results must be considered biased by limited data points being associated with erosion and possible differences in the time since we are looking at net deposition or erosion over a certain period during which the specific site may have undergone different phases of erosion and deposition and variability in moisture contents affecting the strength measurements, among other possible impacts. Some of these factors may be addressed through changes in the survey strategy. Data collection could be improved in future work by collecting denser spatial measurements at locations below the high-water mark, increasing the frequency of sediment property measurements to capture more variability, increasing the temporal density of bed level measurements, specifically near survey times, and collecting a more representative data set encompassing both depositional and erosional behaviors. A further complicating factor in making direct comparisons between sediment properties and morphodynamics, as evidenced by Fig. 8, is the inherent variability and complexity of coastal and nearshore processes. Sediment properties are highly variable on spatiotemporal scales (wave-by-wave), and comparing changes in sediment properties from one snapshot in time to morphologic change over a 24-h period likely does not capture the variability at the necessary resolution required for making definitive associations between sediment properties and morphologic changes. Even though the three transects were relatively closely spaced, the differences in morphology and their evolution over the course of the survey (Fig. 4) highlight the complexity of sediment transport processes and their relation to local hydrodynamic forcing, even on a relatively small spatial scale. Differences in local grain size within survey periods or between data points could further complicate feedback between wave energy, local morphology, and sediment properties. Of the 112 sediment samples collected, 56 contained at least 20% medium sand (0.43–2 mm), 10 contained at least 5% coarse sand (2–4.75 mm), and 7 contained gravel size particles (>4.75 mm), highlighting the variability in the grain size distributions not captured by d₅₀. This variability is highlighted in Fig. 6(c), where a series of representative grain size distributions from the swash zone of Transect C (C6 and C7) show a large range of d₅₀ and also a wide range of particle sizes present in the samples. This variability in grain sizes has been documented at this site and identified as a potential driver of localized morphodynamics (Gallagher et al. 2016). Thus, while the authors conclude that available data are not complete enough to draw direct associations between sediment properties and morphologic changes, this data set offers an initial attempt and highlights pathways and challenges for future studies.

The results of JET testing on the in situ samples provided useful insights. Void ratio, total unit weight, and strength properties from the PFFP were related to k_d from the JET. Fig. 9 shows clear relationships between the detachment rate coefficient, k_d, and void ratio/total unit weight [Figs. 9(a and b)]. Additionally, k_d exhibited a negative correlation with FF and qsbc, albeit a weaker relationship than with the sediment properties. Lower k_d indicates less erodible sediment (in terms of erosion rate), and higher k_d indicates more erodible sediment (Hanson 1991). Denser, stronger sediments are less erodible than loose, weaker sediments (Fig. 9). While this is intuitively and theoretically expected, this data set represents a rare direct comparison made between in situ soil properties and direct measurements of erodibility, and the analysis offers a quantitative measure of how erodibility is affected by the void ratio, total unit weight, and sediment strength based on field measurements. It should be noted that the value of k_d depends on the chosen solution methodology, in this case, a linear regression of the erosion rate versus shear stress data, which is supported by the results of Wahl (2021). The outputs of this model will be used for comparison amongst results at the study site and not in relation to the magnitude of erodibility parameters compared to other sediments. While this study was conducted over a narrow range of grain sizes, void ratios, and total unit weights, specifically for beach sands, the results highlight the variability in erodibility present even over a small range of properties, which is often overlooked in studies relating soil properties over a larger range of soil types (NASEM 2019).

Fig. 9 shows a weaker relationship between FF or qsbc and k_d, with the trend showing a tendency for lower FF (within the sandy sediments) to indicate less erodible sediment and vice versa. This result can be understood more clearly when considered in the context of Fig. 7. In Fig. 7, there is a clear relationship between FF, qsbc, and void ratio, with the general trend being increasing density with increasing FF and qsbc. However, there is a significant spread in the data. As seen from the color gradient, the result is also highly sensitive to changes in water content, with an increase in water content corresponding to an increase in FF and qsbc. However, this is likely due to increasing water content causing increased viscous effects due to high-strain rate, artificially increasing a_max (Martin et al. 2009). The modified strain rate factor [Eq. (3)] represents a first step toward accounting for these effects, but the relationship breaks down at water contents > 25% by volume due to the aforementioned viscous effects, which could explain why the spread in the data was greater at these high water contents (Fig. 7; Brilli 2023). Improvements in understanding partially saturated strain-rate effects on PFFP measurements, specifically at high water contents, could help reduce the spread of these data. If void ratios and unit weights of beach sands can be quantitatively obtained from PFFP measurements, it would follow that PFFP data could be used to estimate erodibility and predict morphodynamics as discussed previously in this section. Overall, the results show relationships between geotechnical sediment properties and erodibility and support the hypothesis that these properties should be included in future models of beach evolution.

The goal of this study is to measure geotechnical sediment properties in a sandy intertidal beach environment and investigate the relationships between those properties, erodibility, and local geomorphodynamics. These relationships are tested using a field data set collected at the USACE FRF in Duck, North Carolina, in October 2021 and JET erosion testing of obtained soil samples. Measurements include moisture content, grain size, density, and sediment strength via PFFP deployments. Morphological changes are recorded from terrestrial LiDAR, and hydrodynamics are measured using an offshore bottom-mounted AWAC and ADVs measuring in the swash zone. Theoretical sediment mobility modeling and laboratory erodibility testing are also performed. It should be highlighted that variations in sediment properties are tested within mostly one sediment type, fine sand with d₅₀ ranging from 0.3 to 0.43 mm (with some outliers in the range of 0.9–1 mm) and total unit weights and void ratios ranging from 14 to 22 kN/m³ and from 0.39 to 0.96, respectively, exploring small-scale variations in the void ratio/unit weight and associated strength and the relationship with erodibility and geomorphodynamic changes. The key takeaways are listed as follows:

Overall, this paper presents a rare field data set to compare in situ soil properties to direct measurements of erodibility, specifically in the context of coastal sediment transport, and the results show that sediment properties, namely, void ratio, change with and affect sediment erodibility.

All data, models, or codes generated or used during the study are available from the corresponding author upon reasonable request. Data are also available from an online repository in accordance with funder data retention policies (Brilli et al. 2023).

The authors thank Julie Paprocki, Matthew Florence, Saurav Shrestha, Jonathan Moore, and the scientists and staff at the USACE FRF for their support in data collection efforts. Special thanks go to Dr. Britt Raubenhiemer for providing ADV data. These data were collected as part of the During Nearshore Event Experiment (DUNEX), which was facilitated by the U.S. Coastal Research Program (USCRP). The authors thank USCRP for supporting this effort through funding for logistics and coordination. This material is based upon work supported by the National Science Foundation under Grant CMMI-1751463. Any opinions, findings, conclusions, or recommendations expressed in this material are those of the author(s) and do not necessarily reflect the views of the National Science Foundation.