≡ Menu

The Terrascope: Challenges Going Forward

Yesterday I renewed our acquaintance with the idea that large natural objects can stand in for technologies we have previously been engineering into existence. The progression is a natural one. The early telescope work of Hans Lippershey and Galileo Galilei began with small instruments, but both refractor and later reflector designs would grow to enormous size, so that today, even with the best adaptive optics and segmented mirrors working together, we are pushing hard on what can be done. Not to mention the fact that controversies over land use can come into play with gigantic observatory installations, as we’ve seen recently in Hawaii.

The fascination is that there is nothing in physical law to preclude ever increasing segmented mirror instruments, but we have to question their economic realities and their practicality. I think it’s a nod to the sheer ingenuity involved in linking seemingly disparate phenomena that David Kipping could turn work on the ‘green flash’ seen at sunrise and sunset into, first, the realization that the refractive flash would be a globe-circling ring as seen from space when the Earth occulted the Sun, and second, into the idea that it marked a potentially usable lensing phenomenon.

Here we’re letting nature do the work. Using the Earth as a lens would allow us to extend our powers remarkably, and at comparatively low cost. Working with a one-meter detector of the sort we already know how to fly (Hubble uses a 2.4-meter mirror), we could achieve results we would expect from a 150-meter space telescope. Assuming, of course, that Kipping’s recent analysis is correct, or more accurately, that the issues he identifies in the paper are solvable and our imaging technologies are up to the task.

For we start digging into the challenging bits almost as soon as the concept is explained. What Kipping wants to exploit is the fact that light from a star behind the Earth would undergo refractive lensing as it entered and then exited the atmosphere. Crucially, the focal point this would create as light rays bent around the Earth and converged on the other side would vary in distance depending on the altitude in the atmosphere that the light passed. This would give us a focal line, so that as we moved further away from the Earth, we would be experiencing the lensing of light at higher and higher altitudes.

The Perils of Extinction

I hope by now the point is clear that because we are talking about light in the atmosphere, we have to be aware of all the extinction events — loss of light — that could occur, from low-level light being obscured by topographical features to much higher weather causing disruption in our seeing. Extinction is dependent on the color, or wavelength, of light in question, and we also have to take scattering effects into consideration.

But Kipping believes that light lensed from about 14 kilometers would be relatively free of these effects, allowing a robust image to be potentially acquired. The focal line sampled at roughly the distance of the Hill Sphere (about four times Earth-Moon distance) is thus what we’re after. Moreover, the alignment doesn’t have to be exact. It could be offset by about an Earth diameter.

Have a look at the figure below, drawn from Kipping’s paper, which shows the extinction expected at various distances from the Earth, and you’ll see that the most robust acquisition of signal is going to be at the Hill sphere distance.

Image: This is Figure 11 from the paper. Caption: Amplification after extinction expected for a 1 metre diameter telescope at the Earth’s Hill radius (top), half the Hill radius (middle) and the Moon’s separation (bottom). Six atmosphere models are shown… [the color coding is provided in the paper, q.v.], which control temperature-pressure profiles (and thus refractivity profile) as well as the extinction computed using lowtran7 [a propagation model and computer code for predicting atmospheric transmittance and sky thermal and scattered radiance]. All models assume no clouds. Standard photometric filters highlighted in gray, except for L, M and N which are slightly offset to encompass the optimal regions. Credit: David Kipping.

As you see, these models assume no clouds, but we must take into effect the fact that at any given time, clouds cover approximately three-quarters of the Earth, meaning they can block out the desired signal. The Hill sphere distance is, in Kipping’s estimation, sufficient to bypass the bulk of these. His calculations show that light passing through the atmosphere at an altitude of 14 kilometers should display a net loss of about eight percent, small enough to make the concept of the Terrascope viable.

Image: This is Figure 13 from the paper. Estimated transmission through the Earth’s atmosphere due to clouds for a Terrascope detector at a distance L. At one Hill radius (1,500,000 km), lensed rays travel no deeper than 13.7 km and thus largely avoid clouds, thereby losing less than 10% of the lensed light. Credit: David Kipping.

Parts of the Earth which are illuminated are a problem, because they insert a background component to the desired signal. The author notes that “background suppression strategies, such as leveraging polarization, wavelength information, and temporal light curve variations” may be of value in reducing some of this loss, but he proceeds to account for the loss with a conservative halving of this original amplification figure to 22,500. An actual Terrascope may able to make this figure climb, though the paper does not explore the idea further.

On the matter of extinction and its problems due to clouds, it’s interesting to consider the Terrascope in other wavelengths. Kipping’s primary thrust here is in optical and infrared light, but the paper moves on to consider the interesting possibility of moving outside this range. From the paper:

Moving further out into the radio offers two major advantages… First, extinction due to clouds can be largely ignored, allowing for detectors much closer including on the lunar surface. Second, Solar scattering is far less problematic in the radio and indeed it is typical for radio telescopes to operate during daylight phases. The simple refraction model of this work was extended to the radio and indeed the amplification was estimated to be largely achromatic beyond a micron. Nevertheless, the model did not correctly account for the radio refractivity as a function of humidity, nor the impact of the ionosphere on lensed rays. Accordingly, a radio terrascope may be an excellent topic for further investigation.

A Path for Research

Limited by working with only those targets that are behind the Earth as seen from the spacecraft, we seem to be boxed in, but consider the possibility, which Kipping discusses in his video on the Terrascope, of creating a fleet of small detectors around the Hill sphere. Now the target list widens as we gain pointing control, and as we build out such a fleet, we can choose our early targets while anticipating the next high-priority items to be served by future detectors. The important first step is to find out whether a working Terrascope delivers on its promise.

Those with a long memory may recall my Centauri Dreams articles on Claudio Maccone’s FOCAL ideas in relation to creating not just a telescope but a radio bridge between distant targets for interstellar exploration [see, for example, The Gravitational Lens and Communications]. Could we do something similar to create a radio Terrascope? The paper gives a nod to the notion, and I asked Kipping if he could expand upon the idea, which he was kind enough to do yesterday.

“Any system like this can also be used in reverse, we can switch out the detector for a transmitter and suddenly we have an antenna with an amplification of 45,000 instead,” he wrote in an email. “In fact, it’s even easier because you don’t need an Earth-shade anymore. To realize this in the radio, there would need to be some further work on how the ionosphere affects the lensing but certainly the optical and infrared are already viable according to my study. One could imagine using this as the bedrock for an interplanetary high speed internet, enabling missions like Cassini and Juno to send back very high fidelity data.”

Given current limitations on data return from deep in the Solar System, the term ‘interplanetary high-speed internet’ seizes the imagination, which is why I’m going to make sure that Vinton Cerf, inventor of the TCP/IP protocols and a man who has done extensive work on extending these methods into forms adapted for Solar System distances, sees the Terrascope idea. Looking down the road, I can only imagine the effect on public support for robotic exploration of the outer system if we were able to offer high-speed data and virtual reality capabilities for our targets.

The benefit of exploring David Kipping’s ideas plays hand in glove with research into gravitational lensing. We’re talking in both cases about placing a detector at the right position to acquire data while minimizing background contamination, developing the needed occulter and creating the software to untangle what does get through from the lensed photons we want. But with the Terrascope, we are working close to home, and can experiment at relatively low cost with strategies for maximizing the acquisition of our imagery and the untangling of it. Much of this work would play directly into the development of the longer-term FOCAL mission.

We need to investigate these ideas. The collecting power of a 150-meter telescope, if it can be achieved as per this concept, takes us into the possibility of detecting topography on exoplanets, not to mention its potential for biosignatures as well as deep space astronomy. Moving further along will involve looking deeper into the question of atmospheric effects and occulter design, which is where Kipping believes the next research effort should be mounted.

The paper is Kipping, “The “Terrascope”: On the Possibility of Using the Earth as an Atmospheric Lens,” accepted at Proceedings of the Astronomical Society of the Pacific (abstract).


Comments on this entry are closed.

  • Geoffrey Hillend August 13, 2019, 17:38

    This idea of using the Earth as a telescope has two flaws. First, Sunlight is not magnified by its refraction through our atmosphere. Even if is was, the telescope would not be as large as the Earth since the thick part of our atmosphere is only 30 miles wide on each side of the Earth so the telescope lens would be only sixty miles wide. The Earth is 7926 miles in diameter which could not be included in the size of the telescope lens diameter.

    There is a theoretical limit how large an optical telescope can magnify an object in space due to the wavelength of the visible spectrum of light. We cannot look at something that is too much smaller than the wavelength we use to look at it according to quantum theory. https://phys.org/news/2015-11-universe-resolution-limitwhy-view-distant.html

    • jonW August 13, 2019, 22:39

      Geoffrey, your objection is nonsensical, look again at the diagram in Paul’s previous post to see how light is deflected to a focal point on the other side of the planet. It is not a question of using the atmosphere to magnify, but using it to refract light on either side of the planet towards a focal point, distant on the order of magnitude of 4 times the earth-moon separation.

  • Geoffrey Hillend August 13, 2019, 17:43

    It would be nice to have 7926 miles of light gathering power or even one mile of light gathering power. We might have to use many smaller mirrors put together to do that though whether in space or on Earth.

  • Geoffrey Hillend August 13, 2019, 17:50

    To be more precise, the part of the atmosphere which refracts sunlight is within only 10 miles above the ground, the troposphere. The thin upper atmosphere, the stratosphere and above it does not cause much refraction since the air there is much less dense.

  • Geoffrey Hillend August 13, 2019, 18:16

    The idea that the size of what we look at is limited to wavelength applies to the large as well as the small so we can only magnify something so much before that limitation is reached and of course, larger telescopes are needed because the resolution is based on light gathering power. Larger telescopes can see further into space than smaller ones.

    We can imagine that in the distant future larger space and land telescopes will be built than today.

    • Robin Datta August 13, 2019, 22:59

      Resolution is related to aperture diameter. Light gathering power is related to aperture area. A terrascope’s resolution should correspond to that of a telescope with similar aperture diameter.

  • Brett August 13, 2019, 20:34

    I’m all for this. It’s something we could easily try out in the next five years.

  • Randall August 13, 2019, 23:19

    This article reminded me of the Dittoscope concept from several years ago. Whatever came of that?

  • Alex Tolley August 14, 2019, 0:38

    @Geoffrey Hillend – I think your argument would equally apply to gravitational lensing and your argument appears to be analogous. Would you argue that gravitational lensing is wrong too? If not, can you create a diagram to show the differences between the 2 lensing approaches – one that works and one that doesn’t?

  • Antonio August 14, 2019, 4:19

    I read somewhere about the idea of combining the Terrascope with a laser in the opposite side of the Earth for calibration, a kind of adaptive optics for this concept.

  • Brett Bellmore August 14, 2019, 6:17

    So, how would this work for other planets in the Solar system? (Obviously, the ones that have atmosphere!)

    Is Earth at a suitable distance for any of them?

    • Patrick Underwood August 14, 2019, 10:32

      Yes. For example, what are the numbers for Jupiter?

    • Rogério Penna August 17, 2019, 9:20

      In the video on Youtube Dr Kipping released the concept to the public, after he submitted the paper, he mentions that in theory it would be possible to use the concept with other planets as well. But would require further studies, due to the individual atmospheric characteristics of each one.

  • AlexTru August 14, 2019, 7:20

    1. Internet as it is implemented right now , i.e. using TCP/IP protocols – is useless for interplanetary network… TCP/IP protocol use request acknowledge (REQ/ACK) sequence for duplex communication, but due to huge distances between solar system objects REQ/ACK exchange timeout will be very long – i.e. very low speed duplex (dual direction) communication, nonsense, when one direction (simplex) communication channel can have very high speed.
    For example, let suppose – you have possibility to make internet connection between the Moon and Earth, ping from the Moon to the Earth cannot be shorter than 2 seconds… , still somehow usable for patient users :-), but if you are try the same from Mars …
    So it should be implemented using alternative protocols… Or very modified TCP/IP.

    • Paul Gilster August 14, 2019, 7:24

      My article points out that the work Vint Cerf and his team have done on modifying Internet protocols for interplanetary networking use is ongoing. So yes, TCP/IP in this environment is in need of considerable modification, which is why the effort started years ago. No one is suggesting that the Internet as currently implemented can be effective for this purpose.

  • AlexTru August 14, 2019, 7:36

    I am sure there is very simple and not expansive way to test “terrascope” concept:
    1. build downsized idealistic model that will include some diffracting sphere (with “core + clouds” inside)
    2. Invent optimal photosensor for this system
    3. Put this sensor somewhere in the focal line
    4. Try to catch (recover) image from this optical system.
    My prognose – it will not work as described in “the terrascope” proposition.
    If sphere shape will have some erratic changes (as the Earth atmosphere) – situation will became much more complicated…

  • ljk August 14, 2019, 9:11


    Earth as an Exoplanet: A Two-dimensional Alien Map

    Siteng Fan, Cheng Li, Jia-Zheng Li, Stuart Bartlett, Jonathan H. Jiang, Vijay Natraj, David Crisp, Yuk L. Yung

    (Submitted on 12 Aug 2019)

    Resolving spatially-varying exoplanet features from single-point light curves is essential for determining whether Earth-like worlds harbor geological features and/or climate systems that influence habitability.

    To evaluate the feasibility and requirements of this spatial feature resolving problem, we present an analysis of multi-wavelength single-point light curves of Earth, where it plays the role of a proxy exoplanet.

    Here, ~10,000 DSCOVR/EPIC frames collected over a two-year period were integrated over the Earth’s disk to yield a spectrally-dependent point source and analyzed using singular value decomposition.

    We found that, between the two dominant principal components (PCs), the second PC contains surface-related features of the planet, while the first PC mainly includes cloud information.

    We present the first two-dimensional (2D) surface map of Earth reconstructed from light curve observations without any assumptions of its spectral properties. This study serves as a baseline for reconstructing the surface features of Earth-like exoplanets in the future.

    Comments: Accepted by ApJL, 7 pages, 4 figures

    Subjects: Earth and Planetary Astrophysics (astro-ph.EP)

    Cite as: arXiv:1908.04350 [astro-ph.EP]
    (or arXiv:1908.04350v1 [astro-ph.EP] for this version)

    Submission history

    From: Siteng Fan [view email]

    [v1] Mon, 12 Aug 2019 19:29:19 UTC (1,672 KB)


    • Alex Tolley August 14, 2019, 10:30

      Interesting approach, but I am not clear how they are going to make this work for real exoplanets. The example given seems to assume images taken during the fully lit hemisphere, not partials. Exoplanet signals will be partial hemispheres depending on the orbital position. The tiny signal from an exoplanet is also going to be very noisy in comparison to the data they use.

      If I am wrong, then it is a useful approach and seems to give better spatial resolution than the Hubble telescope images of Pluto.

  • Geoffrey Hillend August 14, 2019, 16:29

    @Alex Tolley, this Earth sized telescope idea is not analogous to gravitational microlensing which uses the gravity of the exoplanet planet as it passes in front of it’s star to bend an focus the star’s light making it up to one thousand times brighter. This could be seen thousands of light years away and last weeks but the alignment of the planets orbit has to be exactly right, so the event will happen only once.

    Any magnification of the green flash has a focal point on the surface of the planet, so the atmosphere and green flash would appear to be a thousand times smaller from space. One would need a telescope with a magnification of 1000 times to see the green flash from space the same size as it appears to someone standing on the surface of the Earth looking at the sunset. I don’t see how the focal point could be extended to a point which is a far away distance in space as no one can see the green flash with the naked eye from space which would be impossible since it would appear to be too small or not at all. Consequently, a telescope cannot be constructed in space using our atmosphere and any magnification of the Sun by our atmosphere as seen from Earth is very small and only at the horizon. It can’t be used as a telescope lens.

    • Alex Tolley August 15, 2019, 15:36

      Crudely, the magnifying power of a telescope is dependent on the focal lengths of the objective lens (Earth) and eyepiece (imaging satellite). The resolving power for any wavelength is dependent on aperture diameter.

      For both the Terrascope and a gravity lends, the path of any light beam is dependent on the distance from the Earth or Sun. In the case of the Terrascope, this is due to the thickness of the atmosphere. Therefore the higher in the atmosphere the beam passes, the less the refraction angle and the longer the focal length. For the gravity lens using the Sun, the bending of the light path due to gravity lessens the further from the sun, and again, the focal length.

      In both cases, the light bending around the body will be focused on. This ensures that magnification can occur. Because the light paths are separated by the width of the bodies, you also get increased resolution, which is what you want. This is no different that using multiple, widely separated telescopes and interferometry to increase the resolution.

      That is why I say that both approaches should be analogous in principle.

      The issues with the Terrascope are much more to do with the heterogeneity of the atmosphere around the globe. I suspect that any raw image will be much fuzzier than desired and may require considerable image processing to create a good image. The hetergeneity might even be a benefit as different wavelengths fracted from different layers in the atmosphere may reach the same focal point, allowing multispectral images to be created. Even if this does not work, multiple imagers situated to capture different wavelengths could be employed.

      While there are undoubtedly issues with image quality, it certainly seems like an experiment worth trying if the cost is moderate. A small satellite with either an ion engine or solar sail for propulsion might be a good way to place the imager in the correct orbit. This might be a good mission to test this potential in a solar sail and therefore partially fund the mission.

  • Ronald August 15, 2019, 5:15

    First of all, I hope you had a very refreshing summer vacation, Paul, so well-deserved.
    And, if I am not mistaken, CD just had its 15th (!) birthday yesterday or so. Congratulations, Paul, on your awesome website, and you tireless efforts to keep us informed about all interesting new studies, developments and discoveries with regard to stellar planetary systems and interstellar travel.
    Through the years CD has been my pied-a-terre for this topic.

    • Paul Gilster August 15, 2019, 7:17

      Ronald, kind words indeed. Let me also point out that you’ve been around Centauri Dreams for almost its entire run, and you’re right, we just hit the 15th anniversary. The site has always relied on reader insights and suggestions and I continue to appreciate yours.

  • Alex Tolley August 16, 2019, 5:50

    Experimental test idea: It occurs to me that a test of the Earth’s atmosphere as a refracting lens could be carried out by observations from the ISS or a small satellite (CubeSat?). Careful measurements of the position of stars observed through the upper atmosphere and then minutes later through vacuum could be used to confirm the degree and quality of image refraction. This is an experiment that can be done continuously for a number of stars, repeating each measurement for a specific star every orbit (about 90 minutes). With precise information on the location of the ISS or satellite, the refraction angles for different wavelengths and the quality of the star’s image (e.g. twinkling) can be assessed. Using this data the Terrascope concept could have issues tested before moving forward with a telescope that must be designed, launched, and delivered to the desired cis-lunar orbit.

    If these tests could be done with a CubeSat, and launched for free on a piggy-back mission, the cost could be comparatively very low. It might even be doable as a student project at a major university.

    If the tests proved positive, this might well support the proposal for a Terrascope mission, rather than it being a conceptual idea only. Conversely, if tests showed the idea was unworkable, then the idea could be abandoned. Experiment rather than speculation.