How do I interpret the visibility in the METAR correctly? Let's say the METAR reports 9000 meters, I was under the assumption, that I could see 9km in either direction.

I had to go around at the minimums a few times lately. Whenever the METAR is not 9999, the fog appears quite dense. I made a habit of checking the perceived visibility when it is not 9999 and came to the assumption, that MSFS interprets the visibility as the diameter instead of the radius. In other words, the visibility is halved.

This screenshot was taken at GCTS this morning, the runway is 3.2km long for reference. The visibility here looks much closer to 4.5km instead of 9km to me. Does anyone know which interpretation of the METAR is correct?

You're right, you should be able to see 9KM in any direction - and agree, that's nowhere near 9km, but can't help you with why?

Eugene

I'm wondering whether it might be a units mixup <?>. 9km would give you 4.85nm.

The visibility reported in a METAR is the prevailing visibility, defined as the minimum viewable distance over at least one-half (180 deg) of the horizon.  So it's entirely legit to have a situation where 9km PV might give you 9km looking one direction, yet much less in another.

If RVR or RVV are reported, they reflect the measured visibility at the approach end of the runway looking down the runway (RVR) or at some point other than the approach end (RVV).

A typical morning fog situation on an eastbound runway 09 (looking into the sun) might be 3SM R09/1200FT, which means 3 miles visibility over at least half the horizontal plane, but only 1200 ft when looking down runway 09 from the approach end.  If you're flying a Cat I ILS with a typical 200 ft HAT in that weather, you're probably gonna end up going missed approach, since you need more like 2400ft vis to see the runway at DH.

Also keep in mind that the METAR might be up to an hour or more more old...visibility can and often does change quickly, which is why ATIS/ASOS or ATC reports are used for operations and METARs for planning.

I've never seen an ATIS change meteorological data without a corresponding METAR being generated first though, so if you're hitting refresh every couple minutes on WSI you're gonna see a new METAR before a new ATIS.

I've checked the MSFS visibility depiction during very low vis and found it to be very good.  I use runway lights or centerline lights for reference; runway edge lights are 200ft apart, and centerline lights are 50ft apart (in the US, unsure if it's a meter-based value elsewhere.). So if you're sitting on a runway in 600rvr, you should be able to see 3 runway lights during the day, and 12 centerline lights in front of you.  It's usually pretty darn close.

In higher vis, you have to take into account the fact that the sim is blending vis at your airport with vis at the next airport in the direction you're looking.  Like Bob said, this is not an unrealistic situation, to have different visibilities in different directions.  

Thanks guys for your input, this clarifies a few things.

I still find it strange, that anything less than 9999 means an automatic go-around for any non precision approaches. As I said, I kept an eye on this for the past weeks and I'm sure its not airport blending. Southern Tenerife was in a dense fog all the way to the sea. Here is another example at VNKT, the runway is 4km long:

The METAR was stable at that time, there are no other airports nearby and the whole valley was filled with fog, this is not a low cloud. I hope that Ambient Visibility in the debug menu would give a clue to what is going on, but it seems kind of unrelated to the perceived visibility.

Does any one else have this problem with the visibility? If so, I will file a Zendesk ticket.

My understanding is that MSFS generates the weather based on METAR reports. It has no way of knowing what the actual instantaneous weather looks like at different points and directions at any airport.