TF-BLPT reproducible code

This repository contains the Google Earth Engine and Python scripts used for the manuscript:

Tracing tidal-flat degradation under coastal development using tidal-behavior mapping and lagged conversion analysis

The code implements the Tidal Flat Behavior and Lagged Process Tracing framework (TF-BLPT) for annual tidal-flat mapping, threshold diagnostics, accuracy assessment, cross-shore metric calculation and lagged conversion analysis.

Workflow

1. Generate annual Landsat tidal-behavior feature images in Google Earth Engine.
2. Prepare masked land-cover and tidal-flat raster inputs.
3. Build cross-shore coastal statistical units.
4. Estimate constrained local thresholds for dry-state, wet-state and temporal-variability features.
5. Extract annual tidal-flat maps.
6. Evaluate feature separation and threshold sensitivity.
7. Assess mapping accuracy using reference samples.
8. Calculate cross-shore development and tidal-flat metrics.

Repository structure

gee/
  01_generate_landsat_tidal_behavior_features.js

python/
  01_mask_clcd_and_tidal_flat_rasters.py
  02_generate_cross_shore_units.py
  03_estimate_constrained_local_thresholds.py
  04_extract_annual_tidal_flats_parallel.py
  05_feature_distribution_sensitivity.py
  06_threshold_diagnostics_heatmap.py
  07_accuracy_assessment.py
  08_cross_shore_metrics.py

config/
  config_example.yaml

data/
  README.md

docs/
  workflow_overview.md

Data sources

The scripts use publicly available remote-sensing data and user-prepared vector inputs:

• Landsat Collection 2 Tier 1 Level-2 Surface Reflectance, accessed through Google Earth Engine.
• China Land Cover Dataset (CLCD), used to represent impervious-surface development footprints.
• User-defined study-area boundaries, cross-shore coastal units and validation samples.

Large raster files, intermediate products and local GIS assets are not stored in this repository. See data/README.md and config/config_example.yaml for expected input organization.

Running the scripts

The original scripts were written for a local project folder. Most file paths are relative to the repository root and can be adjusted in the parameter section at the top of each script. A typical run order is:

gee/01_generate_landsat_tidal_behavior_features.js
python/01_mask_clcd_and_tidal_flat_rasters.py
python/02_generate_cross_shore_units.py
python/03_estimate_constrained_local_thresholds.py
python/04_extract_annual_tidal_flats_parallel.py
python/05_feature_distribution_sensitivity.py
python/06_threshold_diagnostics_heatmap.py
python/07_accuracy_assessment.py
python/08_cross_shore_metrics.py

Environment

The Python scripts require common geospatial and scientific Python packages. A minimal dependency list is provided in requirements.txt; an optional conda environment file is provided in environment.yml.

Reproducibility note

The repository is intended to document the processing workflow and provide reusable scripts. Users should update local paths, Google Earth Engine assets and validation samples before running the workflow in another study area.