HydroGraphNet¶

Overview¶

hydrographnet is the PyHazards entrypoint for flood forecasting on irregular meshes with graph-structured hydrologic state updates.

At a Glance¶

Hazard Family

Flood

Public catalog grouping used for this model.

Maturity

Implemented

Catalog maturity label used on the index page.

Tasks

Streamflow

Benchmark Family

Flood Benchmark

Primary benchmark-family link used for compatible evaluation coverage.

Description¶

hydrographnet is the PyHazards entrypoint for flood forecasting on irregular meshes with graph-structured hydrologic state updates.

In PyHazards, this model is typically paired with the ERA5-based hydrograph adapter load_hydrograph_data for end-to-end smoke validation.

Benchmark Compatibility¶

Primary benchmark family: Flood Benchmark

Mapped benchmark ecosystems: HydroBench

External References¶

Paper: Interpretable physics-informed graph neural networks for flood forecasting

Registry Name¶

Primary entrypoint: hydrographnet

Supported Tasks¶

Streamflow

Programmatic Use¶

import torch
from pyhazards.models import build_model

model = build_model(
    name="hydrographnet",
    task="regression",
    node_in_dim=2,
    edge_in_dim=3,
    out_dim=1,
)

batch = {
    "x": torch.randn(1, 3, 6, 2),
    "adj": torch.eye(6).unsqueeze(0),
    "coords": torch.randn(6, 2),
}
preds = model(batch)
print(preds.shape)

Notes¶

The smoke test uses a synthetic graph batch so it stays CPU-safe in CI.