If the differing interfaces all have WDM drivers wouldn't it be possible to use those and aggregate the interfaces rather than using ASIO (which, with few exceptions, restricts you to one interface)? You wouldn't be able to monitor using Sonar's inout echo, but I guess you might not need to anyway.
If the band can get a reasonable balance of instruments, vocals etc why not just record them the old-fashioned way using a stereo pair of mics or even a single mic? It worked pretty well for the first 50ish years of the recording industry.
Another option might be to use the interfaces as stand-alone mic pres with their mic inputs routed directly to their outputs which then feed the ZED10's stereo channels. It depends on the internal routing in the interfaces, but if they can handle direct monitoring of their inputs straight to output without going through a DAW that might work as a way to grab a few more mic channels.
No possibility of later overdubs due to mic bleed of course, but if overdubs or punch-ins to fix bits are likely to be needed you'd do better to multi-track from the start with each instrument being individually recorded one at a time. Which means a loss of interplay between band members and longer recording sessions as well as more work....