Monday, October 19, 2009

Perception Is Not Reality: Debunking the Myth of Gravity Hill

A month or so ago, my local NBC affiliate ran a silly puff piece about Gravity Hill Road in Sylacauga, a little town about an hour up the road. Check it out:

We'll forgive them for botching Newton's Law of Gravity. There's no such law as "What goes up must come down." Newton's Law of Universal Gravitation actually states that all objects in the universe attract one another with force proportional to the masses of the objects and inversely proportional to the square of the distance between them. But they would have actually had to fire up Google to get that one right. Let's not ask too much of the talking heads, huh? It's not really crucial to the story, anyway.

The first really egregious mistake the (ahem) "reporter" makes is presenting two opposing opinions as equally worthy of attention without investigating whether one position might be more valid than the other. Person A thinks it's spooky! Person B thinks it's an optical illusion! We report; you decide. Blah blah blippity blah.

Actually, it doesn't take long for the "reporter" to abandon the skeptical position completely and side with the spooky explanation. He relates that the GPS in his truck shows him climbing nine feet. Then, apparently, the GPS does all sorts of weird things. Never mind the fact that most automobile navigation systems are only accurate to between 20 and 50 feet; his initial reading was noise, nothing more.

"It's just a mystery..." "Scientists can't even figure it out..." Well, perhaps scientists haven't figure it out yet (umm, mostly because they haven't tried, maybe?), but two Alabama boys set out Saturday morning to do just that.

Before I became a writer, I worked with my dad as a land surveyor and engineer. Pop still runs his own surveying business and therefore has all the gear necessary to investigate a mystery such as this one. So I called him and asked if he wanted to spend a Saturday morning  surveying the area to see if we could get to the bottom of it all.

We arrived around 10:30 in the a.m., turned onto Gravity Hill, put the truck into neutral, and immediately started rolling in the direction that both of us would have sworn was uphill. It's freaky, there's no doubt about it. Or perhaps disorienting is a better description. Every fiber of your being swears that the road climbs away from the highway, and yet, that's the direction even Pop's big Ford F-150 wants to roll.

I hopped out to take pictures while Pop set up his GPS equipment and dialed into the local base station. We considered using a laser-based total station to take our measurements, but the accuracy of elevation measured by a total station is dependent on its being completely level from the get-go, and if there really were some mysterious gravitational force in the area that could conceivably pull on the level bubble (not sure what that might be, but hey, benefit of the doubt and all), we wanted to ensure that it didn't affect our readings.

GPS surveying gear is not only based entirely on satellite readings, it's also a lot more accurate than navigation systems. Instead of 20- to 50-foot accuracy, we were aiming for (and achieved) ±0.03-foot accuracy. We also took each elevation reading ten times so we could eliminate any noisy data.

The results speak for themselves. The horizontal number beside each  plus mark on the centerline of the road is the elevation of that point above sea level in feet. As you can see in the overhead plot of our work, we surveyed 421 feet of the road ─ enough to take us well past what seems like the crest of Gravity Hill, which is somewhere between Stations 2+00 and 2+50, just in front of where the truck is parked in the photograph above. (Apologies for the civil engineering jargon. To explain, when plotting a road centerline, you mark it in Stations. The distance between each full station is 100 feet. Increments smaller than 100 feet are noted after the +. So the distance between Station 0+00 and Station 2+25 is 225 feet. If that doesn't make sense, the apparent crest of the road falls somewhere between the points measuring 628.72 and 630.31 feet above sea level.)

Below is a profile plot of the centerline, which might require a bit more explanation. A profile is simply a way of showing the change in elevation of the centerline a road over a certain distance. The X and Y axes aren't to the same scale. Each of the dashed "squares" in this graph is 50 feet wide, but only 5 feet tall, so the height is exaggerated to make differences in elevation easier to see. On the right of the graph, at Station 4+21, is the intersection of Gravity Hill Lane (or Road; there doesn't seem to be much consistency in the naming) and Highway 280. On the ground, it appears that the elevation climbs from there back to the "crest"  between Stations 2+00 and 2+50, but you can clearly see that it drops the entire time. What looks like a climb is actually a drop of about five feet in a little less than 200 feet.

So, what the heck is going on here? For one thing, our eyes simply aren't very reliable. How, exactly, do we normally determine uphill from downhill, assuming the grade isn't terribly steep? We look to the horizon as a baseline. At Gravity Hill, though, there isn't much horizon to lock onto. On one side of the road you have to strain your neck to see sky. On the other side, you're looking down at the horizon, which is  unevenly dotted by distant hilltops.

Not to mention the fact that the intersecting road, Highway 280, is on a pretty steady grade itself and superelevated. So there's no really solid visual point of reference.

It's the same principle that used to make the House of Gravity at Six Flags so much fun to explore.

Also, when we see a break in a grade, we tend to assume that the break point is the crest of a hill. At Gravity Hill, though, the road falls away from the intersection at a rather gentle slope, then becomes much steeper past a certain point. We want that point to be a hilltop, since that's what we're used to seeing. So the brain, with no reliable frame of reference, no way to accurately estimate the zenith, perceives that section of road as something like this:

It's really as simple as that. No spooky magnetic force. No wacky gravitation anomalies. Just a really cool optical illusion that everyone within driving distance should experience at least once. It's wicked cool, even if it isn't a mystery. What makes this little stretch of pavement so interesting is that it gives us a bit of insight into the way the brain works. And it didn't even take scientists to figure it out ─ just a curious father/son team with $30,000 worth of precision instrumentation, the know-how to work it, and a thirst for understanding.

After we took a quick look at the raw data and hopped back into the truck to head home, Pop chuckled and said, "I reckon we debunked the shit outta that one. Reckon?"
Me: I think we did, Pop. Can't wait to see the plots!

Pop: Wanna put 'er in neutral and roll back'rds again before we head home?

Me: Sure! It doesn't get old, does it?

Pop: I figured knowin' how it works would suck all the fun right out of it, but it's still a neat trick, ain't it?

Me: It is! Like Richard Feynman said, understanding only adds to the excitement and awe of it all; I don't get how anyone can think it subtracts.

Pop: Is that that feller that told about how the space shuttle blowed up?

Me: That's him! Good for you, Pop! I'm proud of you for remembering who Richard Feynman was!

Pop: He needed a haircut.