This 15 page article I believe is at Wiley. It might be available freely.
Vegetation–microclimate feedbacks in woodland–grassland ecotones
Stephan F. J. De Wekker
Jose D. Fuentes
First published: 25 October 2012
To me, this is mostly a handwaving (theoretical) paper, which covers a number of grassland/not-grassland interfaces:
1. woodland/grassland arctic
2. woodland/grassland alpine
3. mangrove/salt marsh coast
4. grassland/shrubland desert
The authors present evidence which supports the idea that "woodlands" in these situations are altering microclimate so as to better allow for woody species to survive. The arctic situation is the one I am closest to. The woodland has lower daytime maximums, higher nighttime minimums and absorbs more solar energy in winter (no leaves on deciduous trees).
I went to high school here (Dawson Creek) in the late 1970s. Back then, there were lots of farms to the west (towards the Rocky Mountains). A year or so ago, I went to visit a tree nursery at Moberly Lake, and seen how many farms had reverted back to aspen (farmers walked away). Back then, I believe this was climate (hardiness) zone 2 (now probably 3b transitioning to 4a). The area around Dawson Creek has only been in cultivation for about 100 years.
By removing the aspen forest (or at least thinking they had), farmers here forced the microclimate to become colder and have larger daily extremes. The region has little water nearby to moderate temperatures, so we could swing from -40 to -55C in winter, to +30 to +35C in summer. The above article talks a bit about how frost and temperatures combine to damage plants (woody and herbaceous).
I am considering honey locust here fr now. I don't believe honey locust would have grown here 40+ years ago. Alternatives might include Russian olive (often thought invasive) and some species of alder (which fixes nitrogen with Frankia bacteria). One point I am considering with honey locust, is that has sparse shading, you can grow a pasture under it. At some point, I would like to try and crop crops like cereals or oil seeds under it. You need trees tall enough, that (with pruning) you can drive your farming equipment under. You want the trees close enough together, that winds are reduced at ground level and that the root system of the trees extend almost everywhere. With much reduced surface winds, lodging problems with the crops would be reduced (microbursts could probably still cause some lodging), and the warmer microclimate would tend to shorten the time to harvest and reduce frost problems in spring and fall. I don't have a good idea as to how it might affect freak snowstorms that we can get mid-summer. But last year we had a freak frost event on Aug 8 or Aug 9; that was our first frost of the harvest, and for many area farmers it was a killing frost.