Code and Derived Results for Phenology-Guided Mapping of Euryale ferox Using Sentinel-1 and Sentinel-2 Data

This repository contains the simplified Python workflow and derived results associated with the study “Phenology-Guided Mapping of Euryale ferox Using Sentinel-1 and Sentinel-2 Data.”The code covers the principal analytical procedures, including Google Earth Engine-based Sentinel-1 and Sentinel-2 data acquisition and feature construction, Jeffries–Matusita distance analysis for phenological-window identification, ReliefF feature selection, Random Forest and Support Vector Machine classification, convolutional neural network training with stratified 10-fold cross-validation, and SHAP-based model interpretation.The derived results include cross-validation summaries, class-specific accuracy metrics, J-M distance results, ReliefF feature rankings, SHAP feature contributions, annual Euryale ferox extent estimates, 2020–2024 land-cover transition results, trained CNN model weights, annual classification maps, and publication figures.Raw Sentinel-1 and Sentinel-2 imagery is not redistributed because it is publicly accessible through the Google Earth Engine platform. Reference samples, point coordinates, vector data, and private Google Earth Engine assets are not included in this repository. The provided scripts use generic configuration placeholders that must be replaced with assets available to the user.