SCIMAP Flood¶
Identifies potential source areas of flood peak water, to help spatially target nature-based solutions that slow or store water. Implements the algorithm from Reaney, S.M. (2022) "Spatial targeting of nature-based solutions for flood risk management within river catchments", Journal of Flood Risk Management, e12803.
Unlike SCIMAP Sediment/FIO, this tool takes pre-computed input rasters rather than a DEM and land cover — connectivity (e.g. from Network Index), a runoff/weights raster, one or more rainfall pattern rasters, and one or more overland flow distance rasters (e.g. from Overland Flow Distance to Point). It combines every rainfall × overland-flow-distance combination and reports the per-pixel mean and standard deviation across them.
Where to find it: Processing Toolbox → SCIMAP → SCIMAP Flood; toolbar
icon ("Run SCIMAP Flood"); &SCIMAP menu; or the
SCIMAP Panel's Flood tab, which also offers click-on-map
impact points and a shortcut to compute the overland flow distance rasters.
Screenshot:
The Processing dialog with rainfall and overland-flow-distance raster lists populated.
Parameters¶
| Parameter | Description |
|---|---|
| Connectivity Raster (pre-computed) | A connectivity raster, e.g. from Network Index. |
| Runoff / Land Cover Weights Raster (pre-computed) | A runoff-generation-potential raster. |
| Rainfall Pattern Rasters | One or more rainfall pattern rasters — every one supplied is used. |
| Overland Flow Distance Rasters (pre-computed) | One or more overland flow distance rasters. |
| Colour ramp | Applied to both output rasters. |
Outputs¶
| Output | Description |
|---|---|
| SCIMAP-Flood Mean | Per-pixel mean response across every rainfall × overland-flow-distance combination, normalised by its own global maximum. |
| SCIMAP-Flood Standard Deviation | Per-pixel standard deviation across the same combinations — high values mark cells whose flood response depends strongly on which rainfall pattern or impact point is considered. |
How it works¶
For each cell:
base = connectivity × runoff
rainfall_norm = (rainfall − raster_min) / (raster_max − raster_min)
overland_norm = peak-at-median stretch to [0, 1] (1.0 at the median overland
flow distance, falling to 0.0 at the min/max)
for every (rainfall, overland-flow-distance) combination:
product = base × rainfall_norm × overland_norm
mean = per-pixel mean(product) over valid combinations at that pixel
stdev = per-pixel population standard deviation of the same set
The mean output is finally divided by its own global maximum. Every input raster's own declared nodata value, and any negative value, is treated as invalid.
Notes¶
- All four rasters must be aligned to a common grid — this tool does not resample them.
- Runs window-by-window rather than loading whole rasters into memory, so peak memory use does not scale with raster size.
- The mean/standard-deviation pair is best read together: high mean with low standard deviation is a dependable flood-response hotspot; high mean with high standard deviation depends heavily on the specific rainfall pattern and impact point chosen.