Geology topics

Michael Itzkin

Publications and source records attributed to Michael Itzkin.

5 recordsLinked to original sources

USGS2025

Modeling the impacts of sand placement strategies on barrier island evolution in a semi-enclosed bay system

This study assesses the impacts of five proposed restoration actions at Little Dauphin Island, a low-lying relic spit in a semi-enclosed bay system on the Alabama coast. A Delft3D model is developed to simulate annual scale (five-year) sediment transport and resulting bed level changes. The model is validated with observed water level and wave data, as well as sediment tracers that were deployed offshore of the island. An XBeach model is developed to simulate storm-driven morphologic change and is validated for hurricanes Ivan (2004), Katrina (2005)and Sally (2020). Together, the models are used to assess differences in the island's morphological response under a no-action (status quo) scenario representing a continuous island, tidal inlet realignment, a sand motor nourishment, beach and dune restoration and a dredged offshore borrow area. The no-action scenario revealed that the island breached at multiple locations including the location of the proposed inlet realignment during each storm. The realigned channel did not prevent breaching on the island, but reduced the magnitude of sand transported through the breaches. The sand motor provided some sheltering to leeward shorelines during storms but did not prevent breaching from occurring elsewhere. Fairweather waves and currents were not strong enough to transport sand outside of the vicinity of the feature to feed adjacent shorelines as intended. The beach and dune restoration reduced storm-driven overtopping along the nourished shoreline. For habitat purposes, strategically placed bayous provided low elevation points that allowed overwash depending on the direction of cross-barrier water level gradients.

Alabama↗

USGS2024

Wave runup and total water level observations from time series imagery at several sites with varying nearshore morphologies

Coastal imaging systems have been developed to measure wave runup and total water level (TWL) at the shoreline, which is a key metric for assessing coastal flooding and erosion. However, extracting quantitative measurements from coastal images has typically been done through the laborious task of hand-digitization of wave runup timestacks. Timestacks are images created by sampling a cross-shore array of pixels from an image through time as waves propagate towards and run up a beach. We utilize over 7000 hand-digitized timestacks from six diverse locations to train and validate machine learning models to automate the process of TWL extraction. Using these data, we evaluate two deep learning model architectures for the task of runup detection. One is based on a fully convolutional architecture trained from scratch, and the other is a transformer-based architecture trained using transfer learning. The deep learning models provide a probability of each pixel being either wet or dry. When contoured at the 50% level (equal chance of being wet or dry), the deep learning models more accurately identified TWL maxima than minima at all sites. This resulted in accurate predictions of 2% exceedance runup, but under predictions of significant swash and over predictions of wave setup. Improved agreement with the complete TWL time series was obtained through post-processing by utilizing the wet/dry probability of each pixel to weight the contouring toward lower dryness probabilities for runup minima (maxima agreed well with observations without tuning). Overall, a transformer-based model using transfer learning provided the best agreement with wave runup statistics, including a) the 2% exceedance runup, b) significant swash, and c) wave setup at the shoreline. For a random subset of images, the model was found to be within the uncertainty range of hand-digitization. The relative success of the transfer learning model suggests that fine-tuning a large model has advantages compared to training a smaller model from scratch. Models provide per-pixel probabilistic estimates in less than 10 s per timestack on a single computational unit, versus the more than 5 min required for hand-digitization. The model is therefore well-suited for near real-time applications, allowing for the development of early warning systems for difficult to forecast events. Real-time wave runup and total water level observations can also be incorporated into coastal hazards forecasts for data assimilation and continual model validation and improvement.

Coastal Engineering↗

USGS2024

Developing a decision tree model to forecast runup and assess uncertainty in empirical formulations

The coastal zone is a dynamic region that can change rapidly and significantly with respect to the morphology of the beach and incoming wave conditions. Runup forecasts may be improved by adapting a dynamic approach that allows for different runup models to be implemented in response to changes in beach state. Accurately forecasting wave runup is critical to characterize exposure to coastal hazards and provide an early warning against potential erosion and inundation. Here, we developed a decision tree model to produce a weighted ensemble of existing runup models to predict 1.25 years of runup at Duck, North Carolina, USA. We then applied the calibrated decision tree model to reproduce observed runup during the DUNEX experiment in Pea Island, North Carolina, USA. We found that the decision tree approach yielded a prediction that was comparable or greater in accuracy (i.e. higher r2, lower RMSE) than the individual runup models. We also interrogated the decision tree predictions to determine how the individual models perform relative to each other and why certain models perform better than others under the same observed wave and beach conditions. We found that the decision tree approach drew on the processes represented in the individual models in the ensemble to produce a forecast that is accurate and explainable without relying on prior knowledge of the study site(s) or requiring manual adjustments beyond the initial model training.

North Carolina↗

USGS2023

Modeling total water level and coastal change at Pea Island, North Carolina, USA

The DUring Nearshore Event eXperiment (DUNEX) was carried out on Pea Island, North Carolina, USA between September-October 2021. We use a coupled numerical model (Windsurf) to hindcast the evolution of the DUNEX transect and produce a time series of hourly water levels at the shoreline from the model output. In addition to assessing the ability of Windsurf to reproduce TWL, we use model output paired with an ensemble of empirical models to assess how TWL forecasts can be improved by incorporating dynamic morphology. The morphological hindcast achieved an RMSE of 0.10 m and a BSS of 0.47, while the total water level (TWL) time series from the model correlates with the U.S. Geological Survey forecast (r 2 = 0.61) for the study period but with a 0.47 m bias that is primarily due to the much steeper beach slope used to produce the forecast compared to that surveyed at the start of the study period. We find that using dynamic morphology provides a small but statistically significant (α = 0.05) improvement in predicting TWL versus applying the pre-storm beach slope to the whole time series.

North Carolina↗

USGS2022

Combining process-based and data-driven approaches to forecast beach and dune change

Producing accurate hindcasts and forecasts with coupled models is challenging due to complex parameterizations that are difficult to ground in observational data. We present a calibration workflow that utilizes a series of machine learning algorithms paired with Windsurf, a coupled beach-dune model (Aeolis, the Coastal Dune Model, and XBeach), to produce hindcasts and forecasts of morphologic change along Bogue Banks, North Carolina. Neural networks paired with genetic algorithms allow us to fine tune calibration parameters for the hindcast, and then a long short-term memory neural network, trained on the hindcast, produces a 4-year forecast. We compare our hindcasts to observations from 2016 to 2017 and find they successfully reproduce observed modes of dune and beach change except for seaward growth of the dune face. We compare our forecasts to observations from 2016 to 2020 and find that they produce reasonably accurate predictions of dune change except when there are significant instances of erosion during the forecast period.

North Carolina↗