Hi all,
I've pointed this out previously, but one should be a little careful about these tremendously detailed animated maps. Sort of like the old adage about legislation and making sausage, the product is much prettier than what goes into them!
If you slow down the animated videos, you'll notice that many tiny, discontiguous areas in Alberta, which appear to have Song Sparrows in early August, blink out in unison as the weeks advance. These are generally areas where there are very few observers. In some cases, these changes might reflect disappearance of Song Sparrows from as few as one or two yards or observation sites, across an area 50 or 100 km wide.
eBird is using a bit of a trick to create these visualizations in areas where data and observers are sparse or absent. The assumption is that if you have a few observation points in a particular type of habitat (say, "spruce bog"), within a particular range of latitude/longitude, those observations can be extrapolated to guess what happens in all similar patches of habitat, in the same lat/long range.
The effect is visually more impressive than eBird's first efforts to produce this type of visualization, which used a spatial interpolation ("smoothing") method based mainly on distance to the nearest observation (rather than land cover types). That previous method produced some strange effects, like Swainson's Thrushes which seemed to migrate up into Manitoba in spring only to drip back into North Dakota, like a glob of colored wax in a lava lamp. But the underlying problem of sparse observations in many regions is still there.
The other thing to beware of in these visualizations is more psychological/perceptive than analytical. Our eyes and brains naturally infer a direction of movement from these animations, just as we mentally fill in the movement between stop-motion frames of old Mickey Mouse cartoons. It's kind of a fascinating phenomenon, how we're able to do this! But it's not real, when we apply it to these type of data.
All that the data really say is that a few observers in some area, say, 100 km north of Calgary, weren't seeing Song Sparrows in a given week where they saw them previously. Whether this represents actual migration or just post-breeding dispersal away from breeding habitat (something that we commonly observe with sparrows right here in the mid-valley), it's doubtful that eBird can discriminate.
As the saying goes, "Seeing is believing!" ... but sometimes these slick animations can fool our hunter-gatherer brains!
Joel