Maybe you know this already, but maybe others do not and may find this rudimentary GNSS explanation helpful/interesting.
To get a position fix, what is being measured is the time it takes for a particular part of a coded radio signal to arrive from space, in medium earth orbit about 20,000 km away in space. Once you have the travel time from the satellite you have distance from it to the receiver since you know the precise speed of radio waves. Once you have the distance of three sats (4 is better), you can triangulate where those 3D distances intersect using good 'ol high school Pythagorus and... poof... you know where you are on the surface of the earth. I'm simplifying this of course... There is some adjustments calculus going on too.
The main problem is this: if there is the slightest error in measuring the arrival time then the calculated distance is way, way, way off. Since the radio waves are traveling at the speed of light, then millisecond errors in timing errors mean many meters of error (I'm too lazy to do the math tonight). This is why there are extremely accurate, expensive and reliable atomic clocks on each satellite... very precise stuff. So the main causes of position errors are:
- refraction of radio waves coming through the earth's ionosphere, giving a delayed timing error (now modeled well though)
- bad satellite geometry - many sats spread out over the sky from horizon to horizon is waaaay better then getting a position fix in an urban canyon or in the mountains where you have limited sky. GNSS radio waves do not travel through buildings or mountains very well at all.
- radio signals bouncing around at the earth's surface and being "delayed" at the receiver end (phantom)... more timing errors. This is called "multipathing" and it is worse in cities where we have lots of hard, reflective surfaces, windows, cars, etc. Also on water. Yes, there is probably some minor error when traveling through clouds but this is also modeled pretty well and corrected. I wouldn't worry about clouds.
Bottom line: the more sats the better, so GPS and GLOSSNAS working together not only give you more satellites, they are likely giving you better satellite geometry then when they are used as an individual system. So a P3 standard (no GLONASS), operating in an urban environment is at a distinct disadvantage to a P3A, P3P, P4 when it comes to knowing where it is, especially if you are navigating in confined spots. Of course, I'm only referring to outdoors... All bets are off indoors since the signals are so weak the position calculations are quite bad, if at all possible. But this is where the visual positioning system takes over on the P3P, P3A and P4.
Hope this helps.