Skip to content

Releases: fmemuir/COASTGUARD

CoastLearn release

11 May 12:44

Choose a tag to compare

Third Release

  • This is the third release of the COASTGUARD repo. The repo has been extended to a new toolset CoastLearn for generating short-term data-driven predictions of coastal vegetation edge and waterline positions along cross-shore transects.
  • The goal with the CoastLearn release/toolset is to serve as a real-time early warning system for coastal change across the entire shoreface, trained and forced with solely satellite-derived coastal metrics.
  • This is the last update containing an entirely new toolset; releases after this will most likely be related to bugs and issues as they arise.

New Functionality

  • The transect-based data from CoasTrack is used to train, validate and test a Long-Short-Term Memory neural network;
  • Various functions have been written to interpolate over the datasets to regulate them;
  • A function to prep the data and compile the neural network allows for different hyperparameters to be fed in by the user;
  • Training takes place using TensorFlow, with the training histories able to be viewed in a browser with TensorBoard;
  • A feature importance function uses Integrated Gradients to assess the influence of each coastal metric in a transect dataframe;
  • Various plotting routines can be called to visualise all of the above (full list in CoastLearn_DriverTemplate and PredictionsPlotting)

Full Changelog: CoasTrack-v1.0...CoastLearn-v1.0

Acknowledgements

  • Daniel Tudor: original code snippets
  • @eduardogomezdelapena: assistance in implementing ShoreShop loss function

CoasTrack release

22 Jan 15:16

Choose a tag to compare

Second Release

  • This is the second release of the COASTGUARD repo. The repo has been extended to a new toolset CoasTrack for generating and analysing timeseries of coastal vegetation edges, waterlines, topographic swaths and wave conditions along cross-shore transects.
  • The goal with the CoasTrack release/toolset is to build a digital twin of the coast, using only satellite- or aerially-derived data.
  • Note: the tag date for this release is 2024-06-12, even though the release is 2025-01-22
  • More tools for predicting changes to vegetation edges and waterlines will be coming soon

New Functionality

  • Transects now intersect with and store:
    • Timeseries of VE positions, dates of capture, NDVI thresholds used to contour;
    • Timeseries of WL positions, dates of capture, NDWI thresholds used to contour;
    • Timeseries of beach widths between them;
    • Timeseries of tidal elevations at the time of each satellite image (from FES2014 and now FES2022 #26);
    • Rate of change from oldest to newest VE and WL in m/yr;
    • Rate of change from second newest to newest VE and WL in m/yr;
    • Timeseries of corrected WL positions based on tidal elevation and beach slope;
    • Dynamic beach slope value based on WL timeseries;
    • Timeseries of maximum slope values at each VE position, extracted from single DEM;
    • Timeseries of mean slope values at each VE position, extracted from single DEM;
    • Timeseries of widths of the vegetation transition zone;
    • Mean width of the vegetation transition zone;
    • Timeseries of significant wave height from requested Copernicus Marine Service product (full timeseries, not just satellite dates);
    • Timeseries of mean wave direction from requested Copernicus Marine Service product (full timeseries, not just satellite dates);
    • Timeseries of peak wave period from requested Copernicus Marine Service product (full timeseries, not just satellite dates);
    • Timeseries of wave alpha AKA difference between mean wave direction and shore angle (full timeseries, not just satellite dates);
    • Net wave diffusivity index;
    • Net wave stability index;
    • Wave runups calculated using chosen runup formulation (Senechal et al., 2011 or Stockdon et al., 2006);
    • Iribarren number based on extracted wave conditions
  • Updates to geemap implemented (#10)
  • New Google Earth Engine Landsat collections updated (#13)
  • Streamlining of GEE ImageCollection creation in a loop (#19 and #24)
  • Tidal correction functions extended, enhanced, and now use CoastSat.slope to calculate the dynamic beach slope (#20)
  • Plotting functions for analysis (Plotting.py and PlottingSeaborn.py):
    • VE timeseries, WL timeseries Transition Zone timeseries;
    • Seasonality assessment where seasonal (annual) signals are extracted from the timeseries and a Seasonal Strength Index is calculated;
    • Wave direction rose plots with wave height histograms on radius axis;
    • Multivariate matrix of all the single-value-per-transect metrics listed above, plotted as scatter plots with linear and LOESS lines of best fit;
    • Ability to cluster per-transect values in multivariate matrix by VE trend (eroding or accreting);
    • Perform Principal Component Analysis on all the single-value-per-transect metrics listed above, and generate a biplot of the first and second PCs with eigenvectors of each metric;
    • Timeseries of storms, tidal elevations, and successful vs. unsuccessful satellite images (new functionality also added to record success rates and reasons for images being ignored)

Full Changelog: VedgeSat-v1.0...CoasTrack-v1.0

Acknowledgments

  • @IdhamN: bug-finding, testing, TZ adjustments, plotting suggestions
  • @nguardo24: bug-finding, testing, and PlanetScope API suggestions
  • @cmac95: bug-finding, testing, datetime suggestions
  • @ccromwellu: bug-finding, testing
  • @DakDune: bug-finding, testing

VedgeSat release

30 Jan 12:07

Choose a tag to compare

First Release

  • This is the first official release of the COASTGUARD repo
  • This repo is being used primarily as a holder for the VedgeSat tool
  • More tools for monitoring and modelling coastal change using machine learning and satellite observations will be added in due course!

Contributions

New Contributors

Full Changelog: https://github.com/fmemuir/COASTGUARD/commits/v1.0