Planets around other stars are too faint to be imaged directly, and although claims have been made for such detections (2M1207b is a case in point), it’s safe to say that our current techniques need significant upgrading to achieve reliable images of such distant worlds. But studying terrestrial planets is a long-term objective and numerous studies have gone into concepts like Terrestrial Planet Finder and Darwin. One day and with some instrument we will indeed be looking at an exoplanet as small as the Earth, working with estimates of surface temperatures and checking its atmosphere for biomarkers that flag the presence of life.

So let’s suppose that in fifteen years or so we’re looking at actual reflected light from a terrestrial world. What else can we learn about the place? The brightness of a planet like this can be affected by many things, including the presence of deserts on the surface or bright clouds above it. An active weather pattern would indicate the presence of a hydrological cycle like the one we see here on Earth. Such changes in brightness would be difficult to detect on planets that rotate in a few days or less, but the presence of oceans on these worlds may well be apparent. Imagine being able to look at starlight reflected off an alien sea.

Darren Williams (Penn State Erie) and Eric Gaidos (University of Hawaii), who address the question in a new paper, say it’s possible. And as we improve our instruments, potentially measurable variations might even include the seasonal blooming of land plants or oceanic algae or the coming and going of snow. All such fluctuations will vary depending on the planet’s obliquity and orbital inclination with respect to the observer. The staggering thing is the amount of potentially recoverable information from a source so dim in relation to its parent star that we cannot see it today.

In terms of liquid surfaces, the chances of detection look quite interesting. From the paper:

A reflected light curve also contains information about the scattering properties of the surface, independent of any seasonal changes. Planets with water will reflect light toward the observer more efficiently in crescent phase than in gibbous phase because of the higher reflectance at low incident angles. This glint from water will make a planet appear anomalously bright in crescent phase compared to diffuse-scattering surfaces observed in the same geometry. Light reflected from water will also impart some polarization to the disk-averaged signal, which might be measurable under idealized (i.e., optically-thin, cloud-free) atmospheric conditions.

The paper goes on to study how starlight reflected off water might be detected, and examines the light curves from a variety of surfaces in the visible and near-infrared spectrum. Half of all extrasolar planets should have the kind of orbital inclinations that would make the glint of existing oceans apparent. But the reflection of starlight from an ocean surface begins to dominate only under certain circumstances:

Specular reflection of starlight from an ocean surface occurs at all phase angles, but only begins to dominate the wholedisk signal when a planet is nearest its star as a thin crescent. Observations at such phase angles can be obtained of planets around G and F stars where they have adequate angular separation and orbit within the habitable zone.

The authors’ figures show that detections within 0.66 AU of a parent star are unlikely. To use these methods, we’ll need to stick with G and F stars, where the angular separation should make such observations possible. Indeed, planets with a surface of continents and oceans, like ours, should polarize the reflected signal by as much as 30-70 percent. Of all the tricky measurements a terrestrial planet finder instrument of this class might make, the identification of reflected water may well be the easiest.

The paper is Williams and Gaidos, “Detecting the Glint of Starlight on the Oceans of Distant Planets,” in press at Icarus and available online.