> In the studies I've seen the time shift required is on the order of seasons and the capacity required is cost prohibitive.

Another option is too build some kind of overcapacity with the renewable so that you can avoid using the battery and recharge it even when the whether is not optimal. It doesn't work if the weather isn't stable enough[1], but for Hawaii I would be too surprised if it was viable.

[1]: that's why solar + wind in northern Europe is a dead end like what we're seeing with Germany: in winter here we have very little sun and weeks long periods with practically no wind, so you'd need to have something like 10x solar if you wanted the overcapacity strategy to work, which also make things prohibitively expensive.