One of the reasons, I've been told, that Nelson's Sharp-tailed Sparrow is so rare in Michigan, is likely that we are just not in the prime corridor of their migrations. To explore this possibility, I took the range map from the Birds of North America accounts and used all of the sources I own to look at the geography of this bird's migration. Here is the range map:The three disjunct breeding populatios are as follows: the nominate northern Plains population, Ammodramus nelsoni nelsoni, the James Bay population A. n. alterus, and the Maritime breeders, A. n. subvirgatus. Pyle 1997 (Identification Guide to NA Birds, Pt. 1) states that nelsoni winters in "coastal TX-FL-SC", and alterus in coastal "LA-FL-NY", with subvirgatus remaining coastal year-round. Using this data the migration corridors might look like this:
Under this model, nelsoni would be much less likely than alterus to pass through Michigan. Of course, this assumes that my two-dimensional lines correlate with actual migration routes on a three-dimensional surface (and that birds do not take circuitous routes), and doesn't consider that alterus might be outnumbered considerably by nelsoni (the only estimate I know of is the Partners in Flight estimate for the entire species of 510,000, lacking subspecific breakdowns). If either of these assumptions were broken, my conclusions would be invalidated.
Rising and Beadle 2002 (Sparrows of the U.S. and Canada: Photographic Guide) state that "most of [alterus and subvirgatus] apparently winter on the S. Atlantic coast whereas most of [nelsoni] winter on the Gulf Coast." This is slightly different from what Pyle says. This scenario might look like this:This has alterus being the only likely possibility for Michigan. Unfortunately, alterus and nelsoni are not separable in the field (and often not in the hand). Still, Granlund, McPeek, and Adams 1994 (The Birds of Michigan) states that "specimens indicate that [A. c. (sic) nelsoni] ... primarily occurs in Michigan. It is not clear what criteria were used to identify the specimens to subspecies.
One final source of information is the range maps given in Sibley (2000) and subsequent editions. Sibley 2000 shows migration corridors (in yellow) occuring thinly through the central Great Plains, SE Wisc., Ill., Missouri (presumably for nelsoni), and Pennsylvania (presumably for alterus). To what extent this represents where birders are finding and reporting the birds rather than their actual distribution is unclear. Unsurprisingly, my guess will be that the actual corridors used are far wider than this.