FloodCast¶

Overview¶

floodcast is the first public inundation model in the staged PyHazards flood roadmap.

At a Glance¶

Hazard Family

Flood

Public catalog grouping used for this model.

Maturity

Implemented

Catalog maturity label used on the index page.

Tasks

Inundation

Benchmark Family

Flood Benchmark

Primary benchmark-family link used for compatible evaluation coverage.

Description¶

floodcast is the first public inundation model in the staged PyHazards flood roadmap.

The adapter uses shared raster tensors so it can be benchmarked through the flood.inundation evaluator without dataset-specific glue code.

Benchmark Compatibility¶

Primary benchmark family: Flood Benchmark

Mapped benchmark ecosystems: FloodCastBench

External References¶

Paper: Large-scale flood modeling and forecasting with FloodCast | Repo: Repository

Registry Name¶

Primary entrypoint: floodcast

Supported Tasks¶

Inundation

Programmatic Use¶

import torch
from pyhazards.models import build_model

model = build_model(name="floodcast", task="regression", in_channels=3, history=4)
preds = model(torch.randn(2, 4, 3, 16, 16))
print(preds.shape)

Notes¶

Outputs are next-horizon inundation depth rasters.