You need to check that the directories exist before mounting. And after chrooting, check that it looks like you are inside the correct environment (should look like normal arch linux filesystem in this case). I recommend ssh:ing to the device and configure mounting and chrooting everything that way (much more easy to handle terminal from a PC). Of course starting xwayland and xfce must be done on the device itself, so you need to also to chroot on the device before doing that (mounting is only required once).

The easiest thing to do is to put the image where you want and then correct the paths in the script. There is no reason in this case to keep the exact paths as written in the original script. The only reason I see for that in some other case, is if you have some external software managing chroots, which at least I don't have on the Jolla device. Or if the images are specially built (not like copies of a normal linux system).

DrYak's instructions may apply when you have custom chroot images (maybe built from the host system), but in this case we have a normal Arch filesystem inside a tar. It would be the same as when you have a linux system on a hard disk that is currently unbootable and you use a rescue image (USB/CD) to boot and then mount the hard disk in a temporary directory, bind mount proc, sys, dev, etc. and then chroot into it and do the resuce operations. I've done this many, many times. Once in a chroot in such an image, it works just like a normal linux environment and no need to mess around with any libraries etc.