Fog Detection from the Surface of Titan: New Findings From Old Data

More than a decade ago, the Cassini-Huygens mission to the Saturn system experienced a highlight-reel moment as the Huygens probe hurtled through Titan’s dense nitrogen-rich atmosphere. After hours of controlled atmospheric descent, the probe finally came to rest on the moon’s surface, making it the first successful landing in the Outer Solar System. Following touchdown, Huygens continued to take data from the surface and relayed it back to Cassini for approximately an hour.

The data from the Huygens mission has been thoroughly analyzed by mission scientists and non-mission scientists alike. But, in planetary science, it’s always possible that asking a new question will reveal something that was there all along. The accessibility of the data to the public and other scientists for future reference and analysis is immensely important. It is what allows the greatest amount of information to be extracted from a mission, and the greatest number of conclusions to be drawn. Scientists beyond those on the mission are able to utilize new and different analysis techniques for years to come, offering up added potential for discovery. For this reason, scientists are consistently able to publish “new” results from “old” data—old mission data sets are gifts that keep on giving!

Two years ago our group at York University began to look at the data from the Side Looking Imager (SLI), which is one of three cameras that are part the Descent Imager and Spectral Radiometer (DISR) instrument aboard Huygens. The purpose of these cameras was to photograph the landscape and atmosphere during Huygens’ descent from three different angles. The data taken could then be stitched together to produce composite panoramas of the landscape.

The images of the descent were full of topographical details that came into focus as the probe’s altitude decreased. Once the probe landed, the first images from the surface were captivating, but they rapidly lost their charm and contained few obvious changes from frame to frame. Despite this, Huygens had been pre-programmed such that the three cameras continued taking images. The probe dutifully carried out its instructions, acquiring postcards of the same scene for approximately an hour after landing, relaying the data back to Cassini before it lost contact.  Not long after Cassini departed over the local horizon, Huygens’ batteries would have failed and the probe went silent, slowly cooling to the temperature of its surroundings.

Our story would have ended there were it not for a scientific development at Mars in the years to follow. In 2008, John Moores (then of the University of Arizona’s Lunar and Planetary Laboratory) adapted and developed a technique for the Phoenix Lander that had previously been pioneered by Mark Lemmon of Texas A&M University for use with the Mars Exploration Rovers. This technique was able to discern fine details in pictures of the sky, not visible to the naked eye, by acquiring several images over several minutes.

The technique, dubbed “mean frame subtraction” relies on the high quality of the raw images returned by spacecraft. While often rendered as JPEGs, there are more than just 256 shades of grey in these images. By removing the parts of the image that stay the same over the entire sequence of frames, it is possible to observe only those features which change from frame to frame—including dust devils, cosmic-ray strikes, and especially very thin clouds.

This was a bonanza for the Phoenix Mission, showing thin cloud features even before the first thick clouds were identified on sol 94 (for upward pointing movies, see here; for movies aimed at the horizon, see here). Later the same technique was successfully applied to the clouds of Gale Crater, which are exceptionally thin (see the complete set of movies here).

Being planetary scientists interested in atmospheres across the solar system, we saw the potential to observe changes in the sky over that hour or so that Huygens continued taking images and wondered what might be lurking in this old data. We thought there was, perhaps, an opportunity to extract lower level clouds from the surface, complementing the stunning imagery of clouds that Cassini was obtaining from orbit. While no change was obvious to the naked eye, we thought we could try to extract more from the pixels above the horizon using our mean frame subtraction technique which had been so successful on Mars.

The raw Huygens data received on Earth was different from what we had seen on Mars. The frame size was smaller, there were more anomalous pixels due to Cassini-Huygens’ long transit to get to Titan, and the data had been severely compressed to allow it to be relayed via Cassini in the short time available to Huygens before the orbiter set over the horizon. But these were problems that could be tackled. Ultimately, we chose to apply a smoothing function to our frames to tone down the visual artifacts and allow the larger patterns of a cloud or atmospheric feature to come through.  We assembled the subtracted frames in an animated GIF file, which allowed us to observe a “movie” of the sky over the imaging period and perceive any interesting motions that we might be able to relate to clouds or atmospheric phenomena.

What did that movie show? To our surprise, we found a feature of interest that seemed to fluctuate, and was present for a total of 6 images of the 82 that were analyzed in total. We were able to confirm that the feature was most likely the product of a real-world process and not the product of noise or a camera artifact. We quickly realized however, that the motion and elevation of this feature could not be properly explained by it being a high-altitude cloud.

Upon further analysis and investigation, we determined that the most likely source of the feature in the images was a fog bank that rose and fell over the observed period of time. Fog had been hypothesized to exist on Titan and was observed at the South Pole by M.E. Brown in 2009 utilizing data from Cassini. Furthermore, the geometry and speed of the change was appropriate for a fog bank. The linear appearance of the feature could also be explained by it being the upper edge of a fog bank during movement, as the constant bulk of the fog bank would have been removed by the mean frame subtraction process. We had previously observed on Mars that the mean-frame subtraction technique was particularly effective at observing feature edges, and such a result was not unanticipated at Titan.

When we first set out to take a look at this decade-old data, we had no idea what we might find, if anything. We hope our story reinforces the value of having publicly-accessible archives like NASA's Planetary Data System (PDS), and encourages academics and amateurs to analyze and utilize these data resources to their fullest extent.

For more details on our findings, see our paper:
Possible ground fog detection from SLI imagery of Titan, Icarus, doi:10.1016/j.icarus.2016.02.002.