Glyphosate is not soil active, so there are no "trees that can grow in glyphosate-doused soil."
The primary reason for broad herbicide treatment as part of site prep is to avoid low-value, or ecologically opportunist species that thrive in disturbed soil/land, and prevent either the target species from growing, or create an environment which lacks the diversity necessary for the region. For example, sweetgum, huisache, black locust, chinese tallow (as examples from specific regions in the US), will all take over and completely dominate a deforested section and prevent oaks, pines, etc. and appropriate forb for wildlife without consistent, ongoing burns.
FWIW, there are no "trees which are GMOd to live with glyphosate application" - you're thinking non-tree crops. Nearly every softwood and hardwood tree is susceptible to damage from Glyphosate.
Only one of the trees I listed was invasive, the others are opportunistic natives to their regions that will outgrow everything else.
The nice "diverse" forest you're thinking of in your mind took a long time to become that way, the normal state of nature is to not create a perfect balance out of the gate, but for constant competition and regularly have to cycle through multiple iterations of configuration which are, by all means, not as productive or valuable for wildlife/nature as their final states. None of that means that using a herbicide is sufficient, but without, you're looking at potentially hundreds of years to get back a usable environment for wildlife that is well-balanced vs 10's of years.
Outside of a few soil-active herbicides, most of what they use is one-and-done and can be applied selectively to only problem plants with minimal unintended consequences.
Where I live in germany, there used to be extensive spruce monocultures forests everywhere.
They are mostly chopped down now and replaced with a mixed young forest, all without herbicides. (But with some planted trees, cleansing and fences to protect the young forest from deers) So after 15 years they surely are not comparible to old grown forests, but they are very diverse and alive. So I strongly question the assumption that herbicides are necessary or beneficial to create a diverse forest.
Most of the dominating species in the first years will be (were) replaced by something else eventually.
To be clear: herbicides are not essential in every area, but in some areas it is completely cost-ineffective to promote appropriate diversity and wildlife habitat without some use of it.
FWIW, in the region I'm currently managing a 100-acre habitat that was previously a pine plantation, it would be sacrosanct to "fence out deer." Early stage re-growth is wonderful deer habitat, lots of sunlight generates lots of forbs. However, in the same region I am in, any area left to its own devices becomes quickly overgrown to the point of making poor habitat for wildlife (no viable food, no viable cover, even though it's "thick" it is not useful to species such as deer, rabbits, quail, turkeys, etc. lacking the right kinds of food and cover).
Mechanical and fire (prescribed burns) are our primary tool we use, along with appropriate canopy thinning. However, when dealing with opportunistic species (the most aggressive here being sweetgum and chinese tallow), these methods are not effective. As each of these species re-sprout and spread via roots as well as seed, mechanical and fire only top-kill, resulting in them coming back thicker again within months. Repeated mechanical control presents significant issues both for valuable forb and impacts on land (a skid-steer is very heavy and results in significant compaction of soil, for example) and is incredibly expensive at about $1,000/acre when following proper selective practices.
We were also very much against the use of herbicides, but after numerous conversations with local biologists and forestry management professionals (our state provides them as a service), we finally realized that we were in a losing battle and selective application was the way to go. With basal spraying for larger stems of unwanted species and selective foliar for seedlings, we've reduced our costs to a fraction, reduced the damage to land and erosion, and we're seeing higher value (ecologically, not monetary) habitat with a faster turn time. Our approach is to eliminate all non-native, invasive species, develop the mix of pine savannah and hardwood bottoms our region has historically represented, and we're seeing the returns we expected much quicker than mechanical methods were providing us.
None of the "chemicals" we use are soil-active, and all of them have a half-life measured in days. We don't use them where girdling or sawing are sufficient to open canopy or create snags, and we don't broadly apply them.
I'm glad you live in a region where there are no opportunistic trees and shrubs which will crowd out other species, and where mechanical control is sufficient to restore traditional diversity, but alas, it still doesn't have the same reward everywhere. Anything left to its own devices in this region will rapidly, I mean within 5 years, become what we call the "pine curtain," useless to both wildlife and man. For centuries even the indigenous tribes had to practice regular controlled burning to fight this.
Well, we do have invasive species as well (and other problems, like too much acid from the spruce in the soil), but fortunately not your kind, which seem indeed tricky to deal with.
This I found interesting (on wikipedia):
"Herbivores and insects have a conditioned behavioral avoidance to eating the leaves of Chinese tallow tree, and this, rather than plant toxins, may be a reason for the success of the plant as an invasive"
So the main problem is, for whatever reasons, animals could eat the chinese tallow tree, but don't? That is a problem indeed, but I think one that evolution would sort out eventually. Might take more time, though.
"restore traditional diversity"
And this is a common debate here as well, but I don't think it makes sense to try to restore the "pristine condition". Things have changed too much and they will continue to change. So yes also here we have invasive species that achieved local domination. But it won't last. If there is a monoculture, then other species will evolve (or find their way towards it) to make use of that food and space, as there is so much of it. But if you want fast results, well, you have to do more than waiting, I agree on that.
To elaborate on this great answer, the technical term is "ecological succession", defined on Wikipedia as "the process of change in the species that make up an ecological community over time."
Plants do not just fill their niche, they alter the environment over time, which in aggregate alters the ecosystem as a whole. Animals and microbes also play a role in this process. E.g. the way rodents and birds disperse seeds, or how pests can destroy a species, or even how elephants can uproot whole trees.
> the normal state of nature is to not create a perfect balance out of the gate
That is true. Additionally, a balance will never be achieved no matter how long you wait, either. That's the state of nature; some things are always replacing other things.
Ecological stability refers to a concept that is never realized. You can pretend it exists by ignoring the variation that you feel is unimportant. Over time, that variation will make its way into the areas that you thought were stable.
The reason I wrote my remark is the relatively recent uptick in "in nature everything always changed regardless of whether we exist, therefore we're absolved from everything that's happening with the biosphere"-style arguments.
Obviously on the face of it the premise is true, but time scales, operationalisation and causes/responsibilities matter critically there.
Islands of stability exist (like our homeostatic bodies, species that have survived for millions of years, old growth forests, etc) in nature. But whether one cares about (not unnecessarily, prematurely ending) any of this is a very different matter.
The primary reason for broad herbicide treatment as part of site prep is to avoid low-value, or ecologically opportunist species that thrive in disturbed soil/land, and prevent either the target species from growing, or create an environment which lacks the diversity necessary for the region. For example, sweetgum, huisache, black locust, chinese tallow (as examples from specific regions in the US), will all take over and completely dominate a deforested section and prevent oaks, pines, etc. and appropriate forb for wildlife without consistent, ongoing burns.
FWIW, there are no "trees which are GMOd to live with glyphosate application" - you're thinking non-tree crops. Nearly every softwood and hardwood tree is susceptible to damage from Glyphosate.