Oh hey! I didn't think I got any replies, but apparently the notification email didn't go through or ended up in my spam folder.
I will be planting most of the fruit trees and other perennials on contour, and intend for those to make up the bulk of my production. I've seen zero-irrigation gardening work well in my climate, but that's with decades of adding organic matter to the soil and heavy mulching (and saving seeds from crops that thrived in those conditions). I've also seen planting annuals under trees work effectively, though the trees are pruned very open to allow plenty of light through. If the annuals were suffering, I couldn't tell; if anything the trees seemed to help prevent evaporation and provided a net benefit to the annuals. It wasn't a thick canopy and the annuals didn't seem to be hurting for lack of sunlight. That was with fruit trees. I don't know if it would be as effective with nitrogen fixers.
I do tend to favor a messier polyculture design, but I need to scale up my production and keep it more organized for the sake of harvesting and seed saving. That's why I'm leaning towards a return to rows as opposed to polyculture on contour. I imagine that as I develop the property, a lot of the annuals will end up being planted throughout the forest garden (where available light allows) and the actual beds will be reserved for calorie dense/staple foods.
As far as location of the garden... I'm not sure how happy I am with my available options. I think when they brought the excavator in to prepare the ground for building, they really did a number on things; I think they've complicated my plans for garden access and water catchment. I'm in charge of food, not building, so a lot of time they don't ask for my input on things that affect my ability to produce food. We'll see. And since the garden is going in before the house does, I have to try to magically make present and future access tenable. I think I'll only be putting in about a third of the garden to begin with, and I can adjust the other two thirds based on how accessable things are from the house. I think they're finally done with the excavator, so now I can go back up to the site and reevaluate.
I'm still debating including trees in my design. I might experiment with it. I can always take them back out, or at least thin them, if it seems to be negatively impacting production. Weeds/spent crops will be going to the chickens to supplement their feed, so while it will eventually end up back in the garden, it won't be around for use as mulch.
Also, the main reason for the sort of inverted hugelkuktur is that we have large slash piles from a logging operation that need to be used up/out of the way, so this simply turns that problem into a solution (if a solution that requires a fair amount of digging.) I at least want to be in a position to get garlic and favas in for this coming season, and then see what things look like come spring.