weird rain flow patterns on roof

[lunarhighway]

when i got home today after driveing trough the rain i noticeds some strange flow patterns on the cars roof. these where very clearly defined, but unfortunately where wiped out by the rain before i could take some pictures

for the most part the car is covered in small doplets of water and relatively dry given the amount of rain (wich accumulates quicky when the car is stopped) but halfway the roof untill the seem of the rear hatch there where some distinctive water streaks that started earlier near the center than the sides and they formed continuous parallel lines from the center of the roof until the seem of the hatch. after the seem there where doplets again.

what could this indicate? is this an indication of attached flow or just the opposite?

what i also notice when it's raining or when the car is wet, is at relative high speeds the drops will stay on the bonnet but they will fly of the leading edge/frontal part of the car, and strangely, seem to shoot up and over the car.
at higher speeds the rear windshield will stay clear and receive no additional raindrops (wich made the rear wiper pretty much obsolete) but existing drops will stay in place for as far as i can see.
i noticed most panels on my car seem fish scaled with the leading panel ending a little higher than the trailing panel.
even though the rear hatch slops down very gently this step, combined with the big seem might be enough to have the air separate again at that point, thus leading to the end of the stripes

my car has a pretty low 0.29 Cd so i think most of the angles where pretty well considered, so what's going on?

is the front shaped such that most of the airflow is deflected over the bonnet and windshield and reattaches somewhere at the center of the roof? given that the roofs gently slopes down from that point it would create a pretty teardrop profile?


any thoughts on this?

for reference here's a profile picture of my car, the streaks started somewhere between the doors

[Cd]

I'm no expert. I'm just guessing - but to me it seems as if the airflow detatches at the area where the streaks stop. If there was attached flow, then I would think that the streaks would continue onward.
Since you have a really good .Cd, the flow must reattach itself further downwind.

Is the rear hatch flush with the roofline, or is there a slight hump ?

[lunarhighway]

the hatch is pretty much flush with the roofline although there is a very mild rubber stip over the window seem, but what surprised me, this strip sits in a a very shallow "valley" so that it doesn't projects beyond the overall roof/hath line. i don't know if this is purely aero motiveted, as the sedan version (wich my dad has) uses the same roof panel, but has a slightly steeper window, yet the fact part of this setup required the rear glass to bulge slightly at the top, gives me the impression this shape was intentional.

however i find it slightly strange the flow is detached at the first part of the roof, althoug the windshield/roof angle; is a little blunt. (this time i think practical reasons overruled perfect aero... people still tend to hit their shoulders on the top door sill getting in and out)

so perhaps vortex generators, or turbulator strips at the end of the windshield might prove beneficial

[Cd]

Have you considered doing a tuft test ?

[Shawn D.]

There's nothing mysterious about this. The seam (not "seem") can both trip the thin boundary layer itself as well as trap droplets that further trip the boundary layer. Putting vortex generators at the top of the windshield probably won't do anything except impart sideways flow, given how thin the boundary layer is.

Originally Posted by .Cd

Since you have a really good .Cd...

Sorry, .Cd, but there's no "." in "Cd."

[MetroMPG]

Another point to consider is that you can't directly infer airflow from water patterns. The water has a lot more mass and will not be able to follow the same path/contours as air. I'm thinking specifically of what you were seeing at the leading edge of your bonnet for example, where the flow coming "up" from the grille is flinging water upwards, but the air itself is more than likely bending over the transition and staying attached on the bonnet surface, windshield and roof.

As for the pattern on the roof, consider things would also have been complicated by your wipers during the drive, no?

I think you're right about the rear hatch seeing separated flow though. While it looks like a gentle angle, my bet is that it's simply too steep an angle, and the flow separates fairly close to the top of the glass.