Heading east towards Chicago on certain winter mornings, the bank of clouds over the lake almost look like a mountain range, with clear blue sky above them and to the west. Pretty neat.
I95 and it’s location that follows the low coastal areas up the east coast is another case of easy identification. It is rather remarkable, I live barely a mile west of 95 and the change across the highway is very noticeable but we are right where the terrain in rising.
Narragansett Bay certainly makes a difference though. It tends to split a nor’easter passing over it. It’s quite clear on a radar map with snow storms. My house is about 10 miles due north of the bay and we’ve been saved from heavy snow that came down just a couple of miles to the east, west, and north.
Speaking as an amateur meteorologist, and Chicago-area resident, I think that this is absolutely what’s going on. Roads and rivers are just too small, in the grand scheme, to have any appreciable effect on the weather, or to serve – in and of themselves – as a “dividing line” between two different weather conditions.
But, valleys (which rivers, and some roads, tend to follow) can have some effect, as others have said. And, roads and rivers can happen to serve as easy-to visualize markers for the boundary line between two different microclimates. For example, as @Dinsdale notes, Interstate 65 in northwest Indiana is often used by meteorologists to define, for the public, the rough dividing line where significant lake-effect snow is predicted to fall. This has nothing to do with any impact of I-65 itself, and more that it’s a rough marker for where wind out of the northwest will have had enough distance traveling over Lake Michigan to generate snow: if you’re substantially west of I-65, lake effect snow is pretty uncommon, whereas if you are on or east of I-65, you frequently get dumped on by snow.