Are cloud bases in new METAR system in AGL or MSL?

[WxMarc]

@glatze1980

I don’t understand that no one noticed that the clouds are generel too low before they put out SU7. I remember that I immediately saw that something could not be right.

I agree with this. I’ve seen cloud bases that were too low long before SU7. Something may have changed that makes it more likely now than before, but I don’t think this is a simple bug with only one cause. Based on my experience, here are a few factors that could be causing this:

1. Mismatch between MeteoBlue weather model terrain and MSFS terrain. From what I gather, MSFS is importing global model data from MeteoBlue in 30 km horizontal resolution. That means weather datapoints are spaced 30 km apart. The model will use an average altitude for each datapoint. If an airport is at a higher altitude than the nearest data point, there could be a data mismatch with clouds depicted below the ground in MSFS.

2. Global model data is always referenced to pressure, not altitude. I work with model data files routinely as a meteorologist. Model data is almost always reported on standardized pressure levels, and data is interpolated for below-ground areas. As such, the model will output data for the 1000 hPa pressure level, even in mountainous areas where the surface pressure is only 800 hPa. I’m not sure how Asobo handles that. It would be very possible to end up with clouds growing out of the ground without code that properly handles data points below ground level. According to the original weather video from Asobo last year (see link below), clouds in the sim are drawn based on 32 vertical layers of model data. In mountainous areas, some of those layers will be below ground.

Feature Discovery Series Episode 2: Weather

3. Excessive tendency of some global models to extend stratus clouds to ground level. As an example, the GFS model routinely depicts widespread fog in areas where the real-world conditions are low stratus with mist. I documented an example of that in this post:

Weather transition is not smooth after SU7

I’ve put together some images to illustrate the limitations of the GFS global model data that I was talking about yesterday. Today at 12:00 UTC, the weather across the United States looked like this: [Screenshot (15)] This is a nighttime microphysics satellite image. If the colors are unfamiliar to you, here’s a basic rule of thumb: dark red colors are ice clouds (cirrus, etc), and light blue is liquid water clouds (usually low-level stratus, etc). Yellow is somewhere in the middle – perhaps …

If MSFS is taking raw cloud data from a global model similar to the GFS and drawing clouds in the sim, too many clouds will be depicted at ground level due to the model’s excessive fog simulation.

I’m not ready to rule out the negative height adjustment, as mentioned by other posters above. It’s possible there is a bug in adjusting the displayed height after introducing negative heights on the slider. But I suspect the three factors above are also in play.