.. This file is generated by scripts/render_dataset_docs.py. Do not edit by hand.

WaterBench
==========

Synthetic-backed streamflow benchmark adapter aligned to the WaterBench ecosystem.

Overview
--------

WaterBench is the public flood streamflow adapter used by the EA-LSTM path on the shared flood benchmark.

The current implementation is synthetic-backed, but it preserves the streamflow forecasting contract expected by the shared evaluator.

At a Glance
-----------

.. list-table::
   :widths: 28 72
   :stub-columns: 1

   * - Provider
     - WaterBench ecosystem surfaced through a PyHazards adapter
   * - Hazard Family
     - Flood
   * - Source Role
     - Streamflow Benchmark
   * - Coverage
     - Benchmark-aligned streamflow forecasting samples
   * - Geometry
     - Graph-temporal basin or node sequences
   * - Spatial Resolution
     - Basin or gauge nodes represented as graph elements
   * - Temporal Resolution
     - Rolling history windows for streamflow prediction
   * - Update Cadence
     - Generated locally for smoke and benchmark-alignment runs
   * - Period of Record
     - Synthetic-backed benchmark adapter
   * - Formats
     - PyTorch graph-temporal dataset objects via the dataset registry
   * - Registry Entry
     - ``waterbench_streamflow``

Data Characteristics
--------------------

- Graph-temporal sequences with node-level targets for next-step streamflow prediction.
- Registry-backed benchmark adapter rather than a raw WaterBench ingestion pipeline.
- Intended for benchmark-linked streamflow smoke runs.

Typical Use Cases
~~~~~~~~~~~~~~~~~

- EA-LSTM smoke tests.
- Shared flood benchmark runs with streamflow metrics.
- Regression checks for basin-scale forecasting experiments.

Access
------

Use the links below to access the upstream source or its public documentation.

- `WaterBench abstract <https://neurips.cc/virtual/2023/80632>`_
- `WaterBench repository <https://github.com/uihilab/WaterBench>`_

PyHazards Usage
---------------

Use this adapter when you want the public WaterBench-aligned streamflow surface exposed by the flood benchmark.

Registry Workflow
~~~~~~~~~~~~~~~~~

Primary dataset name: ``waterbench_streamflow``

.. code-block:: python

   from pyhazards.datasets import load_dataset

   data = load_dataset(
       "waterbench_streamflow",
       micro=True,
       history=4,
       nodes=6,
   ).load()

   train = data.get_split("train")
   print(len(train.inputs), train.inputs[0].x.shape)

Related Coverage
~~~~~~~~~~~~~~~~

**Benchmarks:** :doc:`Flood Benchmark </benchmarks/flood_benchmark>`, :doc:`WaterBench </benchmarks/waterbench>`

**Representative Models:** :doc:`EA-LSTM </modules/models_neuralhydrology_ealstm>`

Inspection Workflow
-------------------

This dataset is currently surfaced as a registry-backed benchmark adapter,
so there is no standalone inspection CLI documented for it.

Notes
-----

- This is a synthetic-backed benchmark adapter rather than a full WaterBench downloader.

Reference
---------

- `WaterBench: A Large-scale Benchmark Dataset for Data-driven Streamflow Forecasting <https://neurips.cc/virtual/2023/80632>`_ (`repo <https://github.com/uihilab/WaterBench>`__).