origins

In the years since Inconstant Moon was first launched, one of the regular requests has been for more detailed maps of the Moon. I had definite ideas of what format such a set of maps should take, but what I had in mind required a very particular type of imagery which just did not seem to be available. Then in late 2000 NASA launched the Consolidated Lunar Atlas (CLA). Originally published by the US Air Force in 1967 as a set of photographic prints, this unparalleled resource offered a consistent set of high-resolution images covering the entire near side of the Moon. Having obtained confirmation that I could use the images, all that remained was to produce the Inconstant Moon atlas.

structure

The plan was to present detailed, high quality images of the Moon in a way which was as intuitive as an A-Z city street guide. There should be a high-level grid map allowing you to quickly select a low-level chart. The charts should allow easy movement from one to the next and have generous overlaps. There should be a full alphabetical list of features.

The interactive medium should also allow additional features not possible in the paper equivalent: the grid lines and feature labels should be switchable, allowing an unobstructed view of the images, and both images and labels could be capable of being mirrored or rotated, and so be equally useful to someone using binoculars in Boston and someone using a catadioptric in Canberra.

The framed arrangement of Inconstant Moon seemed to lend itself well to the task, allowing easy switching between navigation methods and one-click selection of any chart. The biggest departure was the heavy use of JavaScript to switch images rather than using many hundreds of separate HTML chart pages.

The next task was to select a scale and a grid layout. After some experimenting, the ideal arrangement appeared to be a 6×8 grid of charts. Both the CLA and Rükl's Atlas use 8 rows, allowing a degree of compatibility, and 6 columns produces well proportioned charts for a web page and happens to work rather well on the corner charts. The individual charts would be 360×270 pixels (plus a 35 pixel margin of overlap), giving a decent compromise between resolution, scale and download times.

images

A 2160×2160 pixel diagram of the Moon was drawn, with lines of latitude, longitude and chart boundaries, and the positions of hundreds of visible features carefully plotted. This master diagram was then sliced into individual chart templates.

For each chart, corresponding images were chosen from the CLA. Usually this would mean four or five images for a single chart, and in some cases as many as seven were needed. In one single case, the boundaries corresponded perfectly, and only one CLA photograph was needed for the chart!

The Moon's aspect is perpetually changing: its illumination varies with the spatial positions of Sun, Moon and Earth, and its orientation shifts due to libration. Because the Moon is spherical this can cause some complex perspective effects on the lunar topography. The chart templates were based on zero-libration, with the Moon's 0° meridian and equator crossing in the dead centre of the disc. Each CLA image had to be resized, rotated, pinched, skewed and stretched until it matched the template, then blended with its neighbours to produce a seamless mosaic.

The mosaics are not perfect. Sometimes it proved impossible to eliminate small distortions, and it was not always possible to choose similar illumination for all the photographs in a chart, so occasionally neighbouring craters are lit from opposite sides! However, nothing has been airbrushed in or out, and all features have the correct relationships of scale and topology. Selenographia is intended to be a quick and easy way of finding and recognising features, and learning or improving your lunar geography. If you need images with very high resolution and accuracy, or under a range of lighting conditions, you should return to the original Consolidated Lunar Atlas.

naming

In naming the atlas, my idea was to devise a pseudo-latin word with the correct resonance, as Microsoft have done with Encarta. The final selection adapted the word "selenography", the lunar equivalent of geography, into a proper noun. It turned out not to be an original idea: the astronomer and map maker Johannes Hevelius used the same name in 1647 for one of the first ever lunar atlases. This Selenographia is therefore humbly dedicated to one of the first great selenographers.