Supporting datasets for "Geospatial machine-learning framework for urban flood susceptibility mapping: a semi-supervised approach applied to Paragominas, Brazilian Amazon"
收藏资源简介:
This record contains the large supporting datasets for the manuscript "Geospatial machine-learning framework for urban flood susceptibility mapping: a semi-supervised approach applied to Paragominas, Brazilian Amazon". All processing scripts and the smaller processed datasets are available in the companion GitHub repository: https://github.com/npca-ufra-pgm/flood-ml-paragominas (see DATA.md there for the full file inventory). Files: spatial_data_raster_flood.tif — stacked predictor composite, 8 bands (distance, elevation, slope, ksat, hand, twi, landcover_classification_2023, classes), over the urban area of Paragominas, Pará, Brazil. twi_paragominas.tif — Topographic Wetness Index raster (SRTM-derived). separability_evaluation_data.csv — separability evaluation point set (387,097 points) with model outputs (classification, classification_hand, cluster) and predictor values. heat_map_flood_risk_postClassif.tif — FSIVI continuous flood-susceptibility map of the urban area. Band and column names are in English; values are unchanged from the original exports.



