Confluence [v5] Validation: Small Change

[nikki-t]

Scope: Small Change  ·  Auxiliary · Bug fix / correctness

The change: With no SVS dataset, Validation quietly falls back to the priors’ gauges and scores on those — logged, but it does not fail. Make that fallback explicit.

Input

• No change

Output

• No change

Open questions

• Dev-set gauges overlap only ~1/3 with SVS and cover the wrong time periods — how should this be surfaced?

Tasks

• Scope the SVS-fallback issue and plan changes so it is clear when Validation scores on the priors’ gauges
• Implement the changes
• Track and implement the downstream changes for QQ, neoBAM, MOI, SAD and ManningNet
• Complete an offline run to test changes and evaluate results