SDStools

A toolbox for analyses of Satellite Derived Shorelines (SDS) generated from CoastSeg (CoastSat)

RoadMap

Currently there are some post-processing needs that users of coastseg have requested. Our team has grouped these needs into the following major categories.

Interpolation

• detect nonstationarity
• imputation techniques:
  1. linear interpolation
  2. regression (covariates)
  3. autoregression
  4. ML (covariates)
  5. Deep learning

Uncertainty

• SDS variability quantification and classification
• nonstationarity detection

Trend analysis

• linear trend
• nonlinear trend

Classification

• classify SDS time-series into pre-determined classes
• classify into custom (unsupervised) classes

Filtering

• Isolating Noise: Techniques to identify and isolate noise within the shoreline data.
• Quality: Tools to assess the quality of the shoreline data.
• Clouds: Methods to deal with cloudy data segments and improve accuracy.

Time Series/Seasonality Analysis

• monthy and yearly averages
• STL: Utilize STL (Seasonal-Trend decomposition using LOESS) for shoreline time series decomposition.
• Seasonal Shorelines: Analyze the periodic movement of shorelines.
• Time Series of Narrowest/Widest Beaches: Assess the evolution of beach widths over time.
• wavelets for nonstationary shoreline timeseries

Comparison/Evaluation

• Shoreline Location from Situ Surveys (First): Primary method for ascertaining shoreline location.
• Shoreline Location with Lidar: Use of Lidar data for a more granular assessment.

Visualization

• Animations: Dynamic representation of shoreline movements.
• Shorelines Color Coded by Time: Visualize temporal shoreline changes with color gradients.

Other Analysis

• Beach Width Time Series (Provide Non-Erodible): Analyze the temporal changes in beach width.
• Non-Erodible Line for the Beach:
• Shoreline Position Coefficient of Variation:
• Estimate beach volume?