Increasing the Fidelity of Hyperlocal Simulations of Urban Pluvial Flooding Through Street Flooding Observations

Replication materials for Increasing the Fidelity of Hyperlocal Simulations of Urban Pluvial Flooding Through Street Flooding Observations by Stefano Annis, Maria Grazia Badas, and Giuseppe Mascaro, submitted in October 2025 for review. If you have questions about the code or find any errors/bugs, please contact Giuseppe Mascaro (corresponding author).

Overall folder structure

The folder includes:

• Three Python scripts required to generate the net precipitation grids: "NetP_0_Main.py" "NetP_1_Buildings.py" "NetP_2_Interception.py"

• "Data_NP_CB" and "Data_NP_VB": folders with the files required to generate the net precipitation grids in the calibration (CB) and validation (VB) basins, respectively, including: "BKNYRD.csv": time series of precipitation rates "LC_Basin.tif": Raster with land cover in the basin "LAI_StudyBasin.tif": Raster with LAI in the basin "BuildingsBasinRaster.tif": Raster with the buildings (binary) "Bacino.shp": Shapefile with the basin boundaries

• "Temp": folder for saving temporary files when running "NetP_0_Main.py"

• "LISFLOOD_FP_Setup_Annis_et_al_2025.zip": Compressed file with the setup of LISFLOOD-FP for the Out-SW scenario in the study basins. The modified "fv1.cpp" file is provided, which should be used to recompile the source code to incorporate the extraction of stormwater runoff at the inlet locations. The folder does not include the NetCDF file with the net precipitation data "rain_buildings_mask.nc" due to its large size. The user could generate that file by running the Python script "NetP_0_Main.py" (see below).

System requirements

The scripts were developed and executed using Python. The necessary libraries can be installed as: conda install conda-forge::numpy conda install conda-forge::rioxarray conda install conda-forge::matplotlib conda install conda-forge::netcdf4 conda install conda-forge::landlab conda install anaconda::scikit-image conda install conda-forge::geopandas

Usage of the Python codes

The generation of the net precipitation (P) grids can be conducted by running the "NetP_0_Main.py" Python script, which requires the "NetP_1_Buildings.py" and "NetP_2_Interception.py" as libraries. In particular, in "NetP_0_Main.py":

1. The user can create net precipitation (P) grids based on the options:
    - "option_rain" --> If 'constant', P is constant in space and time. If 'vartime_constspace', P varies in time and is constant in space.
    - "constant_drainage" --> If 0, no fixed P rate is removed. If 1, the P rate specified by "constant_drainage_rate" (in mm/h; e.g., 44 mm/h as in the paper) is removed

2. When "option_rain" is 'vartime_constspace', the function reads a CSV file with rainfall rates. Here, it is Data/BKNYRD.csv

3. The function requires three raster datasets with land cover, leaf area index (LAI), and building footprints in the same reference system. Here, they are "LC_Basin.tif", "LAI_StudyBasin.tif", and "BuildingsBasinRaster.tif" in the folders "Data_NP_CB" and "Data_NP_VB"

4. The script saves files named "P_LISFLOOD_TimeStep###.npy" in the folder "Temp", with ### being the time index. These files are then read to create the NetCDF file with the net precipitation grid for LISFLOOD-FP

5. The NetCDF file required by LISFLOOD-FP is generated in the last portion of the code by setting the option "save_NetCDF" to 1. The script currently creates a file named "rain_buildings_mask.nc"

6. The script includes the instructions to generate several optional figures to support debugging