When the forecast is wrong

Forecasting is a judgement call, legally speaking

The United States can be sued for the negligence of its employees, with exceptions, and one exception does most of the work here: claims based on the performance of a discretionary function are excluded. If an official had to exercise judgement, the decision is generally not actionable even if the judgement was bad.

A federal appeals court applied this to weather forecasting in 1986, in a case brought by lobstermen whose gear was destroyed after a marine forecast missed a storm. The court described a forecast as "a classic example of a prediction of indeterminate reliability" and "a place peculiarly open to debatable decisions", and the claim failed.

The agency says so itself

The disclaimer on weather.gov is short and worth reading once, because the same reasoning governs every weather product you will ever use, including this one.

The user assumes the entire risk related to its use of information
on NWS Web pages. NWS is provides such information "as is," and NWS
disclaims any and all warranties, whether express or implied,
including (without limitation) any implied warranties of
merchantability or fitness for a particular purpose.

Platforms and what other people say

A separate federal rule governs a service that publishes what its users write. Its core sentence is one of the most consequential in internet law and it is one line long.

No provider or user of an interactive computer service shall be
treated as the publisher or speaker of any information provided by
another information content provider.

In plain terms: a platform is generally not liable for what its users post, and it does not lose that protection merely by moderating, editing for length, or deciding what to show. That is what lets any user-generated service exist at all.

Which is why a report here is never a warning

This is the design consequence, and it is the reason the presentation rules on this platform are what they are. The legal reasoning and the warning-system reasoning point the same way, which is usually a sign that both are right.

What the platform does instead is the plausibility tiering every module since the fifth has described: a report is checked against radar, presented with an honest confidence, and never automatically discarded. Module 9 lesson 8 explained why auto-rejection would be wrong on the science. This lesson is why over-claiming would be wrong on the law.