METAR as a corrective layer on top of model forecasts feels wildly underused in consumer weather. Building a multi-model + terrain-aware stack and finding it's one of the most valuable input. Anyone else doing this?

I've been building a weather app for landscape photographers for about a half year now, and the deeper I get into the model stack the more convinced I am that METAR is the most underutilised input in consumer weather forecasting. I'd love to get this sub's thoughts on whether oth