Noel, it looks as though you have some basic confusions about swales that don't seem to have been addressed by the other responses.
1. A swale is a ditch, plus berm on downhill side, that is on contour across the slope. (If you aren't clear on what "on contour" means, definitely ask!)
2. The purpose of the swale is to slow water flow down the slope to (A) enable more water to soak into the soil rather than run off, and (B) reduce the soil erosion that is associated with water runoff.
Swales are particularly useful in arid environments, where every drop of water counts -- especially on a slope, since slopes tend to dry out faster than level ground. But even in humid, temperate locales they can make a noticeably positive impact on plant growth, as I've seen during a visit to the Whole Systems Research Farm in the Mad River Valley in Vermont,
http://www.wholesystemsdesign.com/wsrf/. There I saw side by side plantings of black locust, some planted directly into the slope and some on the berms of swales. Though the trees on the swales were 1 year younger (if I remember correctly, 3 vs. 4 years old), they were clearly larger and more vigorous.
Because swales encourage water to enter the soil rather than quickly run off, your plants can get some benefit even if they are a ways downhill from the swales, as you draw in your second picture. However, greatest benefit occurs for plants that are planted onto the berm or only slightly downhill from it. These have ready access to the lens of water that forms in the soil at the swale and below it.
Over time, sediments and falling leaf litter will fill a swale, converting it to a naturalistic terrace. This is A-OK and you do not need to trench it back out. By the time this happens, your soil should have improved and root systems will have established well enough that water will more easily infiltrate into the soil even without an explicit swale, and the plants will be less dependent on the water gains provided by the explicit swale. (A terrace of this sort will also slow water and help it infiltrate, even if not quite as thoroughly as a proper swale.)
Regarding you worry about pooling of water and such, swales should be sized so that they capture some, but not too much water at any one contour line. If your spring rains are significant, the solution is to construct more swales, each smaller in size, and closer to one another up and down the slope. That way no one swale is handling too much water. Under normal (large'ish rain) circumstances, the water that collects in a swale should fully soak into the soil within a few hours, or day or two at most. Swales should have spillways built into them to allow overflow an appropriate exit, thus avoiding "catastrophic" breakdown of the berms.
By far, the best discussion of swales that I have found is in Brad Lancaster's book
Rainwater Harvesting for Drylands and Beyond, Volume 2, pages 66 and following. Lancaster calls them "berms 'n basins," in part to avoid confusion over the term swale, which in permaculture is limited to a ditch on contour but in other realms means a diversion ditch -- a ditch with slope to it that diverts water to some other location. (For similar reasons, in
Edible Forest Gardens, Volume 2 Jacke and Toensmeier call them "infiltration swales" to make their meaning clear.) The following PDF regurgitates some of basics from Lancaster's book, but I recommend reading his full take on swales before you proceed with digging anything:
http://www.riverlink.org/documents/CH-3-1BermsandSwales.pdf.
Unfortunately, even Lancaster's explanations leave me with some unanswered questions, in particular w/r/t determining the size and spacing of swales, but I'm a lot better off having read his stuff on the topic than I was before.
Note: per Lancaster, swales are
the wrong choice if your slope has rise/run ratio steeper than 1:4, aka 25% slope, aka slope of 14 degrees. Anything steeper than that requires a more substantive earthworks to avoid risk of mudslides.
Best of luck!
Jonathan
