urban_pfr

Urban Pluvial Flood Risk Assessment

Python package conversion of the ArcGIS workflow from Urban Pluvial Flood Risk Mapping: A High-Resolution Assessment for the City of Hamburg (von Szombathely et al., 2025).

The package is designed to be reusable across cities through city-specific configuration files, for CS6 to transfer the Hamburg case study to target city.

Note: This version is under development and requires verification by CS3 (Hamburg team) before finalizing the analysis pipeline and visualization.

Installation & Configuration

Install

git clone https://github.com/FAIR2Adapt/urban_pfr_toolbox_hamburg.git
cd urban_pfr_package
pip install -e .

Configure

One YAML file per city. Copy config_template.yaml and fill in paths and field names.

Note: Configuration values such as TOPSIS weights and SVPF parameters require verification by CS3 (Hamburg team), as some are not explicitly defined in the original ArcGIS script files.

project:
  city_name: "Hamburg"
  crs: "EPSG:25832"  # projected CRS in meters, not geographic (lat/lon)
  paths:
    input_gdb: "path/to/data.gdb"
    flood_dir: "path/to/Floodlevels"
    output_dir: "./outputs/hamburg"
  ...

See hamburg_config.yaml for a complete example and config_template.yaml for a blank starting point.

Run

python run_analysis.py hamburg_config.yaml

Or from Python:

from urban_pfr.analyzer import PFRAnalyzer

analyzer = PFRAnalyzer('hamburg_config.yaml')
analyzer.run()

Required Input Data

Note: Required verification by CS3 (Hamburg team).

The GDB (or individual shapefiles/GeoPackage files) must contain:

Layer	Required Fields	Purpose
Buildings	ID, Floors, Building_type, geometry	Building footprints with structural attributes
Statistical units	StatisticalUnit, Residents, LivingArea, + sensitivity/coping fields, geometry	Census/population data at neighborhood level
Streets	geometry	Road network for evacuation route analysis (HMA)
Flood layers	geometry (one per depth level)	Pluvial flood extents at each depth threshold

CRS requirements: All layers must be in the same coordinate reference system. The package will attempt to detect and fix mismatches automatically, but consistent input CRS is recommended.