Our invasives seem to be becoming resistent to our normal cut and drip techniques.
Our property contains plants from three major bioregions: North Coast, Central Coast, and an island of inland species usually found 150 miles away in the Sierra Foothills. As such, it's an interesting place to study. We have identified and photo-documented nearly 200 native plant species on the property and 91 exotics. Our weed control methods include, spot spray, hand weeding, mulching, light control (by means of graduated overstory removal), solarization, cut and treat, carbo depletion, heavy equipment... you name it, not to mention how we deploy natives to help us out.
The ability to make an early identification is our most valuable process knowledge. Things happen fast in meadows, as you know. So, if we can identify weed seedlings as early as possible then there's more time to treat them before they seed and more time to plan coordinated attacks on multiple species. Sometimes I have to do successive processes because herbicide injury to a species that will surivive the hit can kill my opportunity to get it later with something more effective. For example, hit catsear with a 2,4-D/dicamba/mecoprop mix (which is supposed to kill it but doesn't) and it turns brown all righty, and then sends up a nearly invisible shoots with tiny flowers on them. At that stage the plant lacks the green material to adsorb and translocate glyphostate, so it will blow seed no matter what you do. See below.
See the brown leaves at the bottom? Normally you would see a rosette about three inches across on the one on the right. Herbicides failed to kill it when treated as specified. Catsear is the dandelion from hell.
Given that the catsear hides under (and sometimes within) bunch grasses, I still have to get the "overstory weeds" (such as cranesbill) early enough for the catsear to recover and leaf out while it's still too cold for the weed to bolt. That means I have to know where I will be spraying, for example cranesbill, when the cotyledons are all that is visible.
Then there are the grass weeds. I have to distinguish weeds such as Bromus hordeaceus, diandrus, secalinus, madritensis, or commutatus versus natives such as Bromus carinatus and vulgaris, long before boot stage. That means I either have to know a visible difference, or have such a detailed knowledge of what weeds are where that I can make a prescriptive decision. When these annuals are hiding out inside native bunch grasses, removal has to be done very carefully by hand for several years. Early treatment with pre-emergence herbicides would preclude germination of the groundcovers between bunches. I therefore have to decide whether to weed it once early, and then apply pre-emergents or to rely upon the longer dormancy of most native seeds and sterilize between the grasses with pre-emergents while supplimenting their nutrition artificially. I plan to be experimenting with spot application of pre-emergents directly in the bunch grasses early in the season to save time and improve yields. Most broadleaf weeds that I have seen before on the property I can distinguish from their cotyledons, which means that I can get away with a lot less herbicide if I can spot spray them early. Early isn't necessarily the best time to weed as too often the taproot breaks at that stage, better to get them at early bolt. That of course means that the spraying merely reduced the numbers to the point that hand operations aren't fruitless.
Together, these tradeoffs mean that I am usuall spraying very early, in narrow time slots in December and January as temperature permits, so that by February and early March I can clean up the problem weeds with spot applications of glyphostate. Handwork starts in February and March and progresses until mid June.
Reliable information is hard to come by. I have found the claims of most herbicide manufacturers too often suspect. Their testing regimens are directed to turfgrass or farming, which have little in common with what we are doing. The botany keys (such as the Jepson Manual I use) are overwhelmingly set up for the convenience of biologists making fine distinctions, not for people looking to identify whether or not to kill a patch of unknown seedlings. Distinguishing plants by seeds may be easy, but it's too late for me to begin developing a control process.
Ultimately, establishing healthy natives is the best way to gain control, but one has to get the situation sufficiently in hand to even contemplate attempting it. There is a huge body of process knowledge involved and nobody in the business is apt to share it, particularly with respect to propagation techniques. To complicate things further, there are apparently succession processes in soil that make certain native pioneer species particularly effective at grabbing a niche before the weeds do. For example: Just after thinning an oak forest, as the duff rots, the purslane species such as miners lettuce or red maids are good at keeping down chickweed or annual blue grass (at least around here they do). As the process progresses (and the as mycorhizae get going) we install Satureja douglasiana, lotus species, and other groundcovers to preclude the re-emergence of (for example) French Broom. Needless to say, we have dozens of strategies that vary according to light, soil chemistry, moisture, wind exposure, seed dormancy, and the characteristic mode(s) of seed transmission, whether weed or native.
All of that was, of course, grossly oversimplified. I think habitat restoration is one of the least understood bodies of knowledge found in agriculture. The tools and equipment are terrible (something I intend to do something about). The propagation processes and popular planting techniques are unaffordable. The best thing that could happen is to get the damned government out of it. They bring me more weeds with their road maintenance equipment than anybody else.
I hope that gives you a feel for what we are doing. It's a lot more involved than that, as I am sure you know.
