NGWPC NWM Regionalization, Hindcast, and Coastal Tools 2026-06-09

[cmaynard-ngwpc]

Integrated EWTS logging support and updated logging behavior across components. Updated state serialization, deserialization, reset-time, and restart handling. Updated hydrologic model outputs, model coupling, and model-specific runtime behavior. Improved validation, warning, exception, and error-reporting behavior. Corrected unit conversions and size/type reporting.

Additions

• Integrated EWTS logging support across affected components, including module-specific logging updates, optional logging behavior, Docker/build integration, and unit-test coverage where applicable.
• Added state saving, state restoration, reset-time, restart, and hindcast workflow capabilities, including related time-management updates and model restart handling.
• Implemented and expanded serialization support, including serialized size metadata, message constants, integer-based transfer structures, deserialization bounds handling, and state-transfer messaging.
• Expanded automated testing coverage, expected-result data, unit-test support, tolerance-based comparisons, serialization checks, and CI-exercised validation paths.

Removals

• Removed additional release updates covering removal of object reference to model in the parameter.

Changes

• Corrected unit conversions and size/type reporting across model calculations, serialized buffers, time-step handling, and BMI-facing data exchanges.
• Refactored BMI model interfaces and input-forcing behavior, including model-class selection, getter/setter behavior, bounds handling, and BMI calculation fixes.
• Improved validation, warning, exception, traceback, permission, and user-facing error-reporting behavior.
• Reworked serialization behavior, message metadata, reported state types, size calculations, buffer validation, and transfer logic for state save and restore workflows.
• Updated hydrologic model outputs and model-specific runtime behavior, including formulation outputs, GIUH/channel-inflow values, ponded-depth variables, and model-specific state/runtime handling.
• Updated cache, dataset, temporary-file, scratch-directory, and data-access behavior, including lazy loading, explicit dataset closure, retry handling, and local/S3 access refinements.