GPS combo to use when running in forest / dense area.

It depends of what pace you're talking about. If you're using the "Pace" field (described here) which is supposed to be "The current pace", but if you're using a HRM strap or anything that provides 'Running dynamics" and also depending on which activity you track, the watch will use the running dynamics data to give you instant pace, which GPS is pretty much never precise enough to give you.

So basically what happens, it calibrates what it perceives as your stride length depending on the dynamics, and apply those measured dynamics multiplied by your instant running cadence to give you what it thinks is your "instant pace". But of course, your running dynamics change a lot with terrain and that throws a spanner in the works when you cross a forest or run on a muddy path.

GPS on the other hand, would not be able to give you anything as "accurate" on the span of a few seconds. The best thing you can do if you want to use a more GPS measurement is to use lap pace and create a short lap.

The average pace for the lap or the activity relies on GPS and should be accurate.

I indeed use the Pace field when i'm running, which is the actual pace / tempo at that time. 

I don't use a strap or other add ons .

When talking about lap pace i assume you're referring to the pace (average) per kilometer (or other predefined distances). And there lies my problem. 

When i run at a pace (average) of 5:30 (which i a pretty constant and accurate pace fore me), and i enter a forest, that pace drops significantly to like 6:30 or even above 7:00. 

And there's no correction at all. a kilometer or more at pace 6:30 in the forest remains the same when i conclude my run. So there's no correction when i save the run. 

which i find strange as it should correct in my opinion because i ran a certain distance at a certain pace which is certainly above that 6:30 pace.

So I guess you're talking about perceived pace, what you probably mean is that your cadence and perceived effort remain constant, but because of the nature of the ground changes (less grip for example), your actual pace slows.

GPS losing precision should increase your measured speed not decrease it. Imagine running in a perfectly straight line to cross the forest, for 1 mile, in 10 minutes. Whether or not GPS is accurate or not, it's still going to measure 10 minute, but if it's inaccurate, your measured path will be wiggling around your real track, have a longer distance and therefore, for the device, you will have covered more ground in the same amount of time, therefore faster.

I have experienced that gps-based live pace may vary 1 min/km up and down, but over 1km this should balance to at most15sec up and down. Maybe your way contains that much changes of direction that the gps-track cuts short up to 30% - which would surprise me. 

So, what to do: 

- give GLONASS a try (or even GPS alone) 

- or better use a foot pod for acurate live pacing 

I’ve been thinking about this a lot as well, and just as you says where does the time go?

sometimes I run in 5min/km speed in dense forests, pace says 7-8min for a few minutes . And once out of the forest with no change in speed I’m back on 5 min. 
Something happens, because the watch apparently realize that I’ve traveled further than first calculated in that time. But still you see these dips in speed on the graph after wards.

but the distance seems ok.

but I have no idea where the minutes or distance goes.

With the galmon.eu coverage map you can check if/where there are coverage issues with the various systems is at the current time.