Carbon Stock Mapping of the Bahía Blanca Estuary: Python Notebooks and Datasets

This repository contains Python notebooks and datasets used for the generation of carbon stock maps of the Bahía Blanca Estuary using remote sensing data and machine learning. Notebooks FullbandAOI_ImageGeneration.ipynb This notebook generates the multi-band remote sensing imagery for the Area of Interest (AOI) used in the analysis.It performs the preprocessing and export of the full spectral band stack, which is later used for training and prediction of carbon stock models. In particular, it genetares the TIFF files FullImage0_190*, as well the csv files samplesAOIF_10_5000_2023-2024_training_test.csv and samplesAOIF_10_2500_2023-2024_validation.csv. RandomForestModel_Training_Validation.ipynb This notebook implements a Random Forest model for clases inference and carbon stock estimation.Main tasks include: Loading the spectral dataset and field samples Training the Random Forest model Model validation Generation of prediction outputs used for carbon stock mapping Datasets samplesAOIF_10_2500_2023-2024_validation.csv Validation dataset containing sample points and associated spectral variables extracted from the imagery.This dataset is used exclusively for model validation and accuracy assessment. samplesAOIF_10_5000_2023-2024_training_test.csv Dataset containing training and testing samples used to build the Random Forest model.It includes spectral variables derived from the multi-band imagery, and is used exclusively for model training and testing. FullImage0_190* Dataset containing the raw TIFF AOI images with the 10th, 25th, 50th, 75th, and 90th percentiles of all bands and indices used in this study Model rf_model_2023-2024 Random Forest Model trained from samplesAOIF_10_5000_2023-2024_training_test.csv datasets.