Visualizing the Lunar Mass Anomaly found in the South Pole - Aitken Basin

Recently, news was announced that the largest impact basin in the solar system, the South Pole - Aitken Basin on the Moon, was found to have a mass anomaly buried beneath it. This mass anomaly is believed to be the remnants/core of the object that impacted the area billions of years ago. The news provides some maps showing the approximate location of the anomaly. I was curious if I could reproduce this visualization and I came up with the following 2 variants.

By default, the ReliefMap in the Wolfram Language includes lighting effects to emphasize the topography a bit better.

With[{data = 
   Entity["SolarSystemFeature", 
     "SouthPoleAitkenBasinMoon"][{"Position", "Size"}]}, 
 GeoGraphics[{White, Thickness[.007], 
   GeoCircle[data[[1]], data[[2]]/2], Dashing[.015], 
   GeoCircle[
    GeoPosition[{-62., 200.}, Entity["PlanetaryMoon", "Moon"]], 
    data[[2]]/7]}, GeoBackground -> "ReliefMap", 
  GeoProjection -> {"Orthographic", "Centering" -> data[[1]]}, 
  GeoRange -> All, GeoZoomLevel -> 2, Background -> Black, 
  GeoGridLines -> Quantity[5, "AngularDegrees"], 
  GeoGridLinesStyle -> GrayLevel[{0.8, 0.6}]]]

But the published versions lack this lighting effect and go for a flatter appearance which can be achieved by changing the GeoBackground option a bit.

With[{data = 
   Entity["SolarSystemFeature", 
     "SouthPoleAitkenBasinMoon"][{"Position", "Size"}]}, 
 GeoGraphics[{White, Thickness[.007], 
   GeoCircle[data[[1]], data[[2]]/2], Dashing[.015], 
   GeoCircle[
    GeoPosition[{-62., 200.}, Entity["PlanetaryMoon", "Moon"]], 
    data[[2]]/7]}, 
  GeoBackground -> GeoStyling["ReliefMap", LightingAngle -> None], 
  GeoProjection -> {"Orthographic", "Centering" -> data[[1]]}, 
  GeoRange -> All, GeoZoomLevel -> 2, Background -> Black, 
  GeoGridLines -> Quantity[5, "AngularDegrees"], 
  GeoGridLinesStyle -> GrayLevel[{0.8, 0.6}]]]

In both cases, I have added an additional circle showing the outer boundary of the Aitken Basin. The size of the circle representing the anomaly is determined based on visual inspection of the original graphic.

Without the relief highlighted, photographic satellite views make the Aitken Basin harder to see:

With[{data = 
   Entity["SolarSystemFeature", 
     "SouthPoleAitkenBasinMoon"][{"Position", "Size"}]}, 
 GeoGraphics[{White, Thickness[.007], 
   GeoCircle[data[[1]], data[[2]]/2], Dashing[.015], 
   GeoCircle[
    GeoPosition[{-62., 200.}, Entity["PlanetaryMoon", "Moon"]], 
    data[[2]]/7]}, 
  GeoProjection -> {"Orthographic", "Centering" -> data[[1]]}, 
  GeoRange -> All, GeoZoomLevel -> 2, Background -> Black, 
  GeoGridLines -> Quantity[5, "AngularDegrees"], 
  GeoGridLinesStyle -> GrayLevel[{0.8, 0.2}]]]

11 Likes

2015-11-02_10-07-11.png

I think the impact is at the small light beige area in the first photo, and the object was moving towards the location where the large mass came to rest. I edited one of your photos (I hope you don’t mind) to show where I think the impact took place, and an approximate direction of travel for the mass. Do you think that Mathematica 12 could model the impact, and see if the scenario I suggested is plausible ?

The “light beige area” you refer to is a visual illusion caused by the confluence of all the grid lines. If you turn off the grid lines, I don’t think you will still see it.

1 Like