This repository contains data and code related to the estimation of subpixel percent impervious surface area (%ISA) across the contiguous United States (CONUS).
The dataset will be made publicly available upon paper publication.
The current repository is mainly for peer review purposes.
The repository contains the scripts, auxiliary data and statistical results related to the CONUS %ISA dataset generation and ISA analysis.
subpixel-isa-us/
├── data/ # Auxiliary datasets for the production generation
├── pythoncode/ # Core scripts for data processing, analysis and visualization
├── results/ # Statistical results for accuracy assessment, analysis and visualization
├── LICENSE # License file
└── README.md # Project documentation
The pythoncode folder contains the main code for data processing, analysis and visualization, written with Python 3.10 and Google Earth Engine.
- high_resolution_land_cover_process: Process high-resolution land cover datasets as training sample
- model_training: Train the U-Net model for %ISA estimation
- conus_isa_production: Generate the CONUS %ISA product on HPC environment
- post_processing: Post-processing the original %ISA estimation results to generate the final %ISA product
- accuracy_assessment: Accuracy assessment of %ISA estimation and IS change detection
- conus_isa_analysis: Analyze the spatial and temporal patterns of %ISA and its changes
- conus_isa_financial_crisis_resilience: Analyze the reduction and recovery of %ISA
- conus_isa_centroid: Analyze the centroid shift of %ISA
- conus_isa_socio_economic: Analyze the %ISA changes with socio-economic metrics changes
- util_function: Utility functions used across different scripts
- gee_app: Google Earth Engine app for visualizing the %ISA product: https://gers.users.earthengine.app/view/conus-isa
For questions or further information, please contact:
- Falu Hong (faluhong@uconn.edu)
- Zhe Zhu (zhe@uconn.edu)