Scissor-tailed Flycatcher is an iconic and readily-identified bird of the Great Plains. It breeds from Texas to Kansas, west to New Mexico and east to Louisiana. In winter it moves south along the Gulf Coast and winters in the Yucatan Peninsula and from southern Mexico to Costa Rica.
Scissor-taileds arrive rather suddenly across much of Texas in early March and by mid-April they have made their way well north into Kansas. Scissor-taileds hang around longer than many other kingbirds (it is a kingbird after all, since it is in the genus Tyrannus). Several weeks after Eastern and Western Kingbirds have vacated these same prairies, one can still find Scissor-tailed Flycatchers. Notice also that their southward movement in October leads to concentrations along the Gulf Coast, especially in Texas. The last few stragglers are still moving through south Texas in early November. These birds presumably move south along the coast from here and in Mexico it is not unusual to see hundreds flying south along the coast in places like Veracruz during October and November. Although many flycatchers are thought to migrate at night, kingbirds are an exception and can often be seen in active migration during the daytime.
The STEM predictions show Scissor-taileds well into Nebraska. While the habitat seems appropriate the species is actually quite rare there. In this case there are probably two elements at work: sample size and scaling. Compared to much of the continent we have relatively few effort-based checklists of traveling counts less than five miles and stationary counts from the Great Plains. This limits how well the model can perform, particularly since much of the habitat seems fairly similar.
The scaling of the map is also off a bit so that the fringes of the map are a bit liberal: the actual range limits for Scissor-tailed Flycatcher are probably defined by the 0.05 probability contour, but the STEM map is scaled such that 0.03 predictions remain bright enough orange to be seen. A research challenge for us to to find ways to automatically scale these predictions for all species in North America so that the apparent limits on the maps actually match the known distribution.