Thanks for your suggestions. In this specific case I decided to pull out the applause sections into a second track, and then amplify the song-track different from the applause-track, to avoid clipping and get close to an even loudness level. I used the split points as track labels and exported clips split by those markers. So whoever wants to listen just to the music can skip the applause parts, and those who want to listen to the full concert can just listen from start to end.
This left s few loudness steps between the end a song (when people started applauding too early) and the (attenuated) applause. I ignored this for this time, futhermore I ran into this problem with crossfades which kept me from smoothing out these transitions - will do next time.
In short: I learned a lot about how to treat parts of a recording differently and will be happy use more crossfading 