The seed bank of weeds in the soil is remarkable. Some plants are incredibly prolific generating seeds and the seeds remain viable in the soil for years. I’ve got a weed called “poa annua” in our lawn that is like that. Very difficult to eradicate.
"Seed coat ecology" is now a scientific field, and it's complicated. Microbial components modify cycles of dormancy for example.
Our property adjoins what was originally an Indian trade route that then became an El Camino between Spanish missions. That road became the first wagon road across the Santa Cruz Mountains in 1851. It was terraced for an apple orchard in the early 20th Century with that exotic seed bank harrowed in every year. When that orchard was abandoned in the mid 1930s it was invaded by trees. Effectively, it was "preserved" for the next 50 years. By then, it was down to only 60 species of plants. 80% of native plants, documented to be in this immediate area, were no longer reproducing. Now, after 35 years' restoration work, we are up to nearly 400.
So I had to learn a thing or two about the seed bank. It's not what they think. It was "weed bank" more than seed bank. After over 200 years of abuse, our native seed bank, for the most part, was nearly dead.
Yet at the same time there have been big surprises. When I started it was thought grass seed for example only remained viable for a few years, and perhaps that is true of commercial seed in bags. Yet I have found grasses germinating after at least 50 years of cover, and in one case, perhaps much longer than that. So there is a lot yet to learn about seed dormancy. One thing I have learned is the role of karrakinolides in smoke in breaking seed dormancy. It has become a very useful tool in using pre-emergence herbicides to cull much of that exotic weed bank.