No matter how much you read, you will make bonehead mistakes.
Off the SW corner of our lawn, a long time ago someone planted 4 or 5 rows of conifers (both pine and spruce) on some kind of plan, where the long line of the conifers is something like SE-NW. No water, no swales, too closely planted to have swales added after the fact.
What does one get for 40 years of conifers with no water? Not much. The tallest trees are the ones the most downwind. And I think the tallest might be pushing 20 feet.
Parallel to the east most row, is a ditch to try and keep water from running into the house in spring run off.
Between this ditch and the board fence which defines the "lawn", I planted two apple and two pear seedlings (all different varieties).
The tallest of the bunch is a pear which is about 4 feet tall. Might need to get shorter, as we had some -40C weather this winter which was unexpected. The others are in the 2-3 foot tall range.
Small trees, the dripline is at most 2 feet from the trunk.
A swale is on contour - constant elevation. Any swales I build here (by hand, it is completely fenced in) will be constant elevation nominally E-W; and I will try to extend them as far to the west to support this array of conifers as I can. But the swale will be designed to overflow to the west. The idea is to replace this SE-NW ditch, with a series of swales which all overflow to the west; at least in part.
These 2 apple and 2 pear seedlings are still all small, and only a couple of years in the ground. They are expected to be spending most of their energy into making roots. So, if the dripine is 2 feet from the trunk, how close does one come to the trunk with the swale?
Any roots for such a seedling should be small, and of limited extent. To cut one small root, should not be much of a problem. To cut all of the small roots is a problem.
There are some trees, which impose geometry between roots and branches. The geometry can get twisted. For example, you might find a tree which if you cut smaller roots on the east side of the tree, that the branches on the east side of the tree lose vigor. There is no "twist" there. You could find another tree if you cut roots on the east side, the branches on the south side could lose vigor. There is a twist in that relationship.
Maybe some trees have a helical arrangement, so loss of vigor could be west at 10 feet and east at 30 feet.
Some trees randomize things.
Some trees produce root systems which extend many times the diameter of the drip line.
There are 3 parts to a swale; viewed looking downhill:
How I would like to try building one, is with a 2 bottom plow. With a 14 in bottom, the ditch part would be about 28 inches wide. So, the first pass with the plow puts half the soil onto the berm, and half is still in the ditch. Perhaps a second pass puts all the soil up on the berm? The berm is supposed to be mostly non-compacted soil, now whether this means one has to till the berm in some regard, I don't know. As described, the ditch would be 28 inches wide and the berm would be 14 inches. I would be happier if it was 28 inches wide.. So, we could declare the approach to be 28 inches as well, and it would be symmetric.
It would be something like a square wave. I don't think this is good. In wood working, you could make a shaper. Cut a profile on rotating blades, and apply it to the wood. All civil construction seems to do, is to apply a flat blade to things. Rototillers are vaguely like a rotating flat blade. It may be that a person could treat a box blade like a shaper, if the soil was tilled. I don't think it would work for sod.
A tilt/angle rear blade, may be able to put in a linear (sloped) approach. It could possibly do this on the downhill side too; but you would probably end up needing to pull the two plowed rolls further downhill, so that you had room to work with when cutting the linear downslope. To do the "cut" on the approach, that dirt would end up in the ditch, and maybe the plow can move that to the berm? To slope the downhill side of the berm, I think you probably need to direct that cut to the ditch as well, if there is any kind of sod involved. I think a steeper slope on the downhill side would work better, especially since we are adding soil to the downhill side.
If we plow 6 inches deep, our 28 inch wide ditch is 1.1666666 cubic feet per foot of ditch. To cut a 28 inch wide slope on the approach is half that. To cut a 14 inch wide slope on the berm side generates 0.29166666 cubic feet per foot. Which means we are adding 2.0416666666666 cubic feet of soil to the berm area, per foot of swale.
The addition of the slope on the downhill side has made our ditch 42 inches wide, and our spoils to the berm is almost twice what the square wave ditch is. If we did this in a square wave manner, the berm would rise almost a foot over where the level of the land was before. With smoothing of corners, the peak of the berm will rise higher if you want to keep the berm at 28 inches wide.
We moved 2.04... cubic feet of dirt per foot of swale, So, if we discarded the soil, we would be storing a little over 15 (USA) gallons of water per foot of swale. Keeping the soil in a square wave type berm, our water storage should now be something like 41.5 (USA) gallons per foot. To me, this sounds like too big a number.
All of the calculations performed to this point are linear. So, if we go from a plowing depth of 6 inches to a plowing depth of 2 inches, our swale storage capacity should drop to something like 13.8 gallons
The problem with my farm, is the sod is 40 years old (mostly fescue), and to plow 2 inches deep probably only brings up grass and roots and not much soil. So, to make swales the first time, you probably need to do a few practice swales, just to find out what to do.
I gather a lot of fruit trees tend to have roots near the surface, where grass grows. And on my farm; the people who started things had no clue about grass. Lay down geotech fabric,and everything will be wonderful. Not! Trying to adjust an area where geotech fabric is present is annoying. To remove geotech fabric from an area can be very difficult.
If we had of planted our tree in the trailing edge of the berm (still above normal ground level), I will suggest the peak of the berm is at about 1/3 if the berm width (about 9 inches for a 28 inch wide berm), and that the tree would be planted about half way down the downslope (about 18 inches back). The tree would probably grow roots to the downhill side just the same way it would grow roots if planted on a level ground with respect to the local soil level. Most trees don't want to have roots smothered in water (there are exceptions). So, I can see a tree planted into a swale berm, to push roots up into the "peak" of the berm above the level that the trunk is planted at. Some trees have deeper roots, and I don't know how those kinds of roots would behave in a swale. Are they going to try to get to the uphill side, going under the swale? At least in part, the problem with roots being "submerged" (wet because of water filling all the soil pores) is a lack of oxygen. For trees that can fix nitrogen, it is also a lack of nitrogen.
If I put a swale uphill of a pre-existing tree (even a small one), it is possible that some of the "surface" roots of that tree will fall under the built up berm area. For older trees, there could be other kinds of roots which go beyond this point as well.
If I do nothing, I would expect the surface roots that are under this now built up berm to send out new roots, going upwards (to be closer to the surface). The problem is, that in a wet period, the elevated water table in the vicinity of the berm will keep some roots under water for too long, and they will die. It is possible, that at the point of dying, that later on they produce new roots which now follow the "proper" path to stay under the surface of the now elevated berm.
Another "solution" which comes to mind, is when one installs a swale "after the fact", that at the line where added soil to the berm ends, that a person use a shovel (or something) vertically, to cut all the surface roots. This should result in the root system generating new roots, which will try to move up into the elevated berm in part, and doesn't involve the "insult" of having some roots be submerged.
Cutting surface roots I don't believe is much of a problem. Cutting roots which are tiny probably isn't a problem (unless you are talking seedlings). Cutting big roots probably means problems.
There are trees which have root systems which far exceed the dripline of the canopy.
My guess, is for an established tree; you want to put the trailing edge of the berm (the transition back to normal soil levels) either at the dripline or slightly inside.
Hopefully someone who is more expert than I am; will comment. I suspect it might be best to cut roots (with a shovel) on that (berm) line, but to only go down 6 inches or so. Again, someone may correct me on this.
My land is mostly 40 year old fescue sod. My feeling, is that the first task in any swale work, is to run a single bottom subsoiler at an appropriate depth at least every foot, possibly every 6 inches across the path you need between the beginning of your approach to the end of your berm. The appropriate depth is how far you will be cutting with a plow or blade. Except for the trailing edge of the berm. I think that is the 6 inches or so one needs, to break all the roots.
Because my land is boreal forest consisting mostly of clonal aspens; I can expect to run into big aspen roots almost anywhere.
Your land could be different.