REsampling Large Inputs/Outputs (relio) : a model to resample low resolution hydrological rasters to high resolution
Résumé
relio is a set of scripts that build up a model to resample raster images featuring hydrological data.
The model is designed to take a low resolution (typically 1e3m x 1e3m) raster file of GCM-generated hydrological data (typically water discharge) and convert it to a high resolution raster file (typically 1e1m x 1e1m).
It is a statistical-physical mixture model that prevents running sophisticated LCMs to get higher resolution data but builds its satistical resampling technic upon a routed flow constraint (conversely to conventionnal GIS resempling algorithms, which are built upon the hypothesis of continuous domains).
Across 19 scripts, it implements 3 main algorithms:
1. A derasterization algorithm used to breakdown the source grid cells
2. A convolution algorithm used to spread hydrological data accross the destination grid cells
3. A (optionnal) rasterization algorithm used to map hydrological point data to the destination grid cells