The answer is MS, or Mid-side. There are 2 types of "ambiance" situations that are very different. The first is dialogue recording with mono mics. Mono ambiances from those mics are needed to patch up issues and replace the talent mics at certain parts to cover other sounds. That is a critical function of mono sounds, but those are not technically field recordings nor ambiances.
In field recordings, we are looking to add ambiance as opposed to fixing some other "department's" problems. Fields recordings take time and ideally are recorded in a general sense with variations so that you can use them on future projects.
Stereo is probably the least interesting as compared with blumlein ( double figure-8 with great coverage, bad directivity) or ORTF (110 degree stereo). The most useful and maybe widely used professional technique is M-S.
M-S is not stereo. It is a sideways figure-8 recording mixed with a mono recording.
It gives a much more vivid representations of "stereo" than the previous formats in most occasions.
Despite taking up only 2 channels, the front and sides can be mixed AFTER the recording into a narrow stereo, or a very wide stereo.
It is perfectly phase coherent when down mixed to mono, or the sides can simply be dropped like a straight mono recording.
It can be spread easily to Left, Center and right in a 5.1 setup. This avoids the "center hole" problem sometimes encountered on 5.1 mixes.
2 MS Recordings can be combined to create a 5.1 recording.
If there is a sound source directly ahead, the sound in front and on the sides can be mixed louder or softer relative to each other, not left or right like stereo.
There are 3 output formats to consider at the end of the day. From a 5.1 mix to a stereo broadcast to a mono speaker, you always have to think of mono when you work in stereo, and you have to think of everything when working in 5.1. M-S is the most versatile for getting sounds that are well recorded for all formats.
In other words, MS is about as simple as stereo but can do an enormous amount more.
In the previous answer posted that was accepted, the person mentioned having a 414 and a 57. Well those are different types of microphones, but I have done MS with that combination and it worked well. With things that are more point-source, having the exact same capsule is less important. For prairie winds or long enveloping sounds, having the same capsule for all the mics in an M-S setup is much more important. In the city or indoors or when there is more action it is less important.
M-S is really the workhorse of ambiances. It needs to be recorded "not decoded". For example, the figure 8 on channel 1, and the mono mic on channel 2. In the mix, you split the figure 8 into 2 separate channels panned left and right for example, and flip the phase on one of them. Then you bring up the mono mic in the center and it all takes shape. You mix the sides and the front how it seems appropriate in the right environment, not in the field.
Also, if you pan it like that, when it falls to mono because you listen in the iphone the sides perfectly start to cancel out while the center mic, with the most important information, stays at the same volume. Phase coherent means it self-down mixes better.
This is a really simple technique. But getting familiar with Blumlein and ORTF is very good also. Figue-8 mics can be expensive but some aren't. Knowing how to build your own M-S setup with two mufti-purpose microphones is a good skill.
There are a number of cheap ms electret microphones that were sold for years for digital camcorders. They can be useful on a tight budget. Two things to watch out for is the noise floor and verifying if the m-s signal is "encoded M-S" or already decoded to stereo. If it is already decoded, which is likely is for a consumer product like that, it means that you can't adjust it at the mix as you could with encoded M-S. Encoded meaning one figure 8 channel and one mono channel that sound like perpendicular mics when played together. They aren't mixed together into a stereo image yet.
If you want to move up a whole new dimension, you can record with ambiophonic microphones, which are basically 3 figure 8's that record a 360 degree sphere that is completely phase coherent. The additional cost and considerations of recording in 5.1 or ambiophonic often make it less practical.
Second to M-S, I might say mono is the most important. You can't go wrong with mono. You won't have any space however. You can create the illusion by adding up mono tracks but now you just turned something coherent into something that likely either has phase issues, or with the sounds so unrelated that they create point sounds as opposed to ambiances. That is the best way I would describe it.
Here is some good info from UA explaining the same things in their own way. There is alot of good info from professional manufacturers. You won't hear much about manually decoding MS on prosumer sites.
When you get to the mix and you feel the need to change something, you can do so much by simply adjusting the center-sides matrix before getting into any kind of processing or anything. A bit like going back in time to the recording.