How much detail could the Hubble telescope see?

May 12, 2009 8:35 PM

What resolution would the Hubble Telescope be able to discern, if we pointed it at my back yard instead of into the stars? Would it be able to pick up individual blades of grass? Would it be able to see bacteria or anything not visible to the naked eye? Could it see the illuminated dial of my watch at night?

For the sake of argument, assume that there is zero cloud cover or atmospheric interference. Also assume the camera and mirrors etc. would even work pointed towards the earth instead of away. Bonus points if you can show how you calculated the resolution.

posted by crazyray to science & nature (19 comments total)

Hubble's angular resolution is 0.05 arcsecond. This is the "sharpness" of Hubble's vision. If you could see as well as Hubble, you could stand in New York City and distinguish two fireflies, 1 m (3.3 feet) apart, in San Francisco.
posted by crapmatic at 8:38 PM on May 12

Hubble orbits at 347 miles, so divide the 3.3 ft firefly distance by about 10 to get roughly 4 inches or so from orbit.
posted by crapmatic at 8:39 PM on May 12

From the ESA homepage for Hubble:

If Hubble looked at the Earth - from its orbit of approx. 600 km - this would in theory correspond to 0.3 meter or 30 cm. Quite impressive! But the 'blurring' of the atmosphere downwards would make the actual resolution worse. Unfortunately, Hubble will never be turned to Earth since a) the brightness of the Earth could be damaging for the telescope and the instruments, and b) there is not any particularly interesting astronomical research to be done there.

posted by jedicus at 8:41 PM on May 12

To compare this with other technology, the most accurate satellite imagery is more detailed than Hubble could obtain even through zero atmosphere. Civilian imagery is limited to 0.5 meters, 50 cm. Military spec imagery is thought to be measured in a centimeter or two. However, said satellites would be utterly worthless in astronomical research. It could be compared to racing a motorboat with a car. Not a lot of point, and if you put the motorboat on the highway, you're just going to break things. Pointing the Hubble at your back yard would likely result in a blurry image of your yard possibly a bit better than Google Earth, and a lot of really expensive machinery being damaged.
posted by Saydur at 8:49 PM on May 12

Might be worth pointing out that 0.3m resolution is slightly finer than the best commercially-available satellite imagery allowed on the market today, which is 0.5m. (The satellites do better than 50cm but the U.S. bars distribution of <5>
If you are interested, here is a page with downloadable samples at various resolutions, including one from WorldView-1, which does 0.5m panchromatic (monochrome). Or if you just want to see 0.5m, here is Sydney (image is hideously watermarked, unfortunately). Be sure to zoom in!

I am always impressed at how much you can discern, below what you'd imagine you can at a particular resolution level. E.g., at 50cm objects much smaller than a half meter are discernible, like the lines on parking lots.

Anyway, you would definitely not be picking out "blades of grass" with the Hubble. I think that the Hubble's strength is not necessarily its resolving power — which I suspect is well surpassed by the sensors on modern or even contemporary reconnaissance satellites — but its sensitivity to very weak signals across a wide swath of the EM spectrum. I'd be willing to bet that a lot of the engineering in the Hubble relates to pushing the noise floor down rather than getting the resolution up.
posted by Kadin2048 at 9:06 PM on May 12

Ugh; didn't preview and my less-than sign got eaten. Also Saydur beat me to the punch...

Might be worth pointing out that 0.3m resolution is slightly finer than the best commercially-available satellite imagery allowed on the market today, which is 0.5m. (The satellites do better than 50cm but the U.S. bars distribution of sub-50cm imagery without a DoD permit.) ...
posted by Kadin2048 at 9:09 PM on May 12

Hubble's resolution is 50 milliarcseconds. It turns out that the ESO's VLT has a resolution of 1 milliarcsecond when using its interferometer, "equivalent to the distance between the headlights on a car at the distance of the Moon".

(The Keck has a resolution of 5 milliarcseconds.)
posted by Chocolate Pickle at 9:14 PM on May 12

Everyone else is right, but to give you a little more of the math, Hubble's resolving power (the smallest angular separation it can measure) is limited by the diameter of the primary mirror. This is called the diffraction limit, since the images of two sources that are even closer together would be blurred into one blob by the diffraction of light as it enters Hubble's aperture. The diffraction limit is specified by the wavelength divided by the diameter of the mirror. Assume roughly 600nm light, and the mirror on Hubble is 2.4m, convert the units and multiply by the magic number 206265 (the number of arcseconds in a radian) to get 0.05 arcseconds. Take the height of the orbit (600km) and multiply by (600nm) / (2.4m) to get the size scale of 15 cm (close enough).

(On a side note, the blurring caused by the atmosphere as seen by a spy satellite (or hubble looking down) is far less significant than the atmospheric distortion of a telescope on the ground looking up, for purely geometrical reasons. Kind of fun to work out, but maybe that's just me.)

(Another side note: Hubble's resolving power really is the main allure for astronomers. The current set of detectors is not too different from what's available at ground-based observatories. But none of those observatories can get anywhere near the resolution that Hubble has. But it is true that the detectors took a tremendous amount of work to limit noise sources too.)

Last comment: It is misleading to compare the resolution of interferometers and actual imagers. The VLT is great for resolving point sources, but interferometers really can't "image" extended sources in the same way that Hubble can. There are subtle yet exceedingly complex issues with interferometry.
posted by kiltedtaco at 9:36 PM on May 12 [5 favorites has favorites]

The resolution of the Hubble Space Telescope is definitely one of its selling points. Telescopes on the ground have to look through the atmosphere. Technology called "adaptive optics", which uses lasers to study exactly how the atmosphere is continuously distorting a star's image, and then allows the telescope to correct for this blurring, has been improving the resolution of telescopes on the ground since the launch of Hubble.

A telescope (or the eye--any opening) has a theoretical resolving power limit, called the diffraction limit. It depends on the wavelength of the light. If L is the wavelength of the light and D is the opening diameter of the telescope, the resolving power is L/D. For green light, say, with a wavelength 5000 Angstroms or 5000e-8 cm, and the Hubble's diameter of 2.4 meters, the resolving power (converted from radians to arc sec) should be 0.04 arc sec.

One way to convert from resolving power in terms of angle to resolving power in terms of size at a certain distance: say the angle is a. Say the distance is R. A circle around Hubble at distance R has a circumference 2 Pi R. The angle a is a fraction of that circle given by a/360. So the linear size of something resolved at an angle a at a distance D should be 2 Pi R a / 360. An arc second is 1/3600 of a degree (60 arc sec in an arc min, 60 arc min in a degree); 0.05 arcsec means the size of something you can see would be 2 Pi (600 km) (1/360)(1/3600) 0.05 = (3600 km)(1/3600)(1/360)(1/20) if 2 Pi is about 6. This is 1/7200 of a km, or 1/7 or 0.14 m, about 14 cm.

Being able to see Ultraviolet and Infrared is also important. I've used the Ultraviolet spectrographs on the Hubble over the years a number of times.
posted by Schmucko at 9:43 PM on May 12 [2 favorites has favorites]

It is misleading to compare the resolution of interferometers and actual imagers. The VLT is great for resolving point sources, but interferometers really can't "image" extended sources in the same way that Hubble can.

That's true. but the scopes at Paranal, or the individual Kecks, can also be used for imaging. The four main scopes at Paranal have 8.2 meter mirrors, and the two Kecks are 10 meters, which means their theoretical resolution should be considerably better than the Hubble's 2.4 meter mirror, even when not using interferometry.
posted by Chocolate Pickle at 10:11 PM on May 12

I recall when Hubble was first being built, there was an article on its construction. The primary mirror is ground so precisely that, were the mirror enlarged to the size of the Gulf of Mexico, the variations--the dips and valleys from the grinding process--would be smaller than a nickel. If you've ever ground your own primary, you'll know how insane that is.
posted by Civil_Disobedient at 4:12 AM on May 13

Interesting stuff, with the unfortunate consequence of reminding me that my liberal ahts degree is fairly useless. And also that I'm not that smaht.
posted by eggman at 6:18 AM on May 13 [1 favorite has favorites]

When Hubble observes the Earth.

Incidentally, I highly recommend this book on Hubble.
posted by edd at 8:09 AM on May 13

The story I've always heard was that Hubble was enabled partly by the existence of surveillance satellites with similar specs. Supposedly HST is roughly equivalent to a KH-11 spy satellite; those have been in orbit since the late 70s or thereabouts. Presumably there are far more sophisticated things up there now.

This is just "astronomy lore," not firsthand knowledge -- I'm an astronomer, but the development of Hubble was before my time.
posted by chalkbored at 8:21 AM on May 13

There was an awesome post more or less on point on Phil Plait's awesome Bad Astronomy today. It says "But y'know, the company that made Hubble's mirror had an awful lot of those same sized mirrors lying around, and there are no other astronomical telescopes (you know, telescopes that point away from the Earth) with that same mirror. So what could those mirrors have been for?" (source). It also addresses Hubble cannot see the Apollo artifacts on the Moon and generally has lots of cool stuff you didn't know about Hubble.
posted by Lame_username at 10:02 AM on May 13

Just wanted to point out that the post above "Unfortunately, Hubble will never be turned to Earth" is not true. There was actually a good post at Bad Astronomy today about the Hubble. As Phil points out, it observes the Earth farily frequently and has been used to observe other bright objects such as the moon.

One of the cameras on the hubble has actually been pointed directly at the sun in order to blast it with UV and reset it sensors (this was done via a mirror while the main viewport was pointed directly away from the sun).
posted by Riemann at 11:42 AM on May 13

I was so beat to that post.
posted by Riemann at 11:42 AM on May 13

Chocolate Pickle: Keck doesn't have a resolution of 5 milliarcseconds. That's an order of magnitude off. It's at best 0.05 arcseconds (when using adaptive optics), more typically 0.5 arcseconds, and it can get substantially worse than that. Hubble's resolution is diffraction limited, but Keck's is atmosphere-limited (which can be alleviated somewhat via adaptive optics).

Both are great telescopes, of course! Let's hope the Hubble servicing mission is successful!
posted by pizzazz at 8:11 PM on May 13

From what I remember of the book I linked, there's certainly times when the military (or DoD or whoever it would be) basically told Hubble engineers "You don't want to do it like that.", nodded sagely, or just giggled to themselves at how behind the times the Hubble lot were. But they could rarely explain why.
posted by edd at 2:37 AM on May 14