APOD: Lunations (2018 Sep 12)

[geckzilla]

...
Anyway, it's totally possible to put the eclipse into the simulation. One need only include a properly placed Earth...

Yes, but it would be a simulation needing additional image data to be accurate. I can see a telescopic image from Earth taken at an animated image frame time, and then overlaid on the modeled 2D mosaic. LRO imagery alone won't provide enough information to accurately render an eclipse.

I disagree. Accurate coordinates for each object (Sun, Earth, Moon) and a physics-based renderer would not be any less "accurate" than what is already rendered.

[alter-ego]

...
Anyway, it's totally possible to put the eclipse into the simulation. One need only include a properly placed Earth...

Yes, but it would be a simulation needing additional image data to be accurate. I can see a telescopic image from Earth taken at an animated image frame time, and then overlaid on the modeled 2D mosaic. LRO imagery alone won't provide enough information to accurately render an eclipse.

I disagree. Accurate coordinates for each object (Sun, Earth, Moon) and a physics-based renderer would not be any less "accurate" than what is already rendered.

I'm not sure if our lines are getting crossed.
No question the LRO mosaics are presented as accurate as they can be. What isn't rendered in the video is the total eclipse. Physics and ephemerides aren't the limiting factors to simulate a total eclipse (I treat "simulate" and "reproduce" the same). I'll say that present prediction / analysis capabilities could reproduce good partial eclipse simulations, but not for total eclipses. To generate accurate color reproduction for total eclipse requires data we don't know well enough (yet). We know that total eclipse color and brightness strongly depends on the integrated atmospheric conditions around the Earth's perimeter. They include clouds, aerosols and moisture. To my knowledge, these parameters cannot be modeled well enough to accurately predict / simulate color and brightness distribution on the fully eclipsed moon. This reasoning is behind my comment that augmenting the LRO mosaic with a full-disk color image of the moon (colors digitized with selenographic coordinates) may yield an accurate reproduction.

With that said, the LRO WAC can capture one good color 60-km swath across the eclipsed surface during it's 2-hr orbit. Maybe that data could "nudge" a predicted color/brightness distribution closer to reality, but I doubt it would be that good.

[geckzilla]

Yes, but it would be a simulation needing additional image data to be accurate. I can see a telescopic image from Earth taken at an animated image frame time, and then overlaid on the modeled 2D mosaic. LRO imagery alone won't provide enough information to accurately render an eclipse.

I disagree. Accurate coordinates for each object (Sun, Earth, Moon) and a physics-based renderer would not be any less "accurate" than what is already rendered.

I'm not sure if our lines are getting crossed.
No question the LRO mosaics are presented as accurate as they can be. What isn't rendered in the video is the total eclipse. Physics and ephemerides aren't the limiting factors to simulate a total eclipse (I treat "simulate" and "reproduce" the same). I'll say that present prediction / analysis capabilities could reproduce good partial eclipse simulations, but not for total eclipses. To generate accurate color reproduction for total eclipse requires data we don't know well enough (yet). We know that total eclipse color and brightness strongly depends on the integrated atmospheric conditions around the Earth's perimeter. They include clouds, aerosols and moisture. To my knowledge, these parameters cannot be modeled well enough to accurately predict / simulate color and brightness distribution on the fully eclipsed moon. This reasoning is behind my comment that augmenting the LRO mosaic with a full-disk color image of the moon (colors digitized with selenographic coordinates) may yield an accurate reproduction.

With that said, the LRO WAC can capture one good color 60-km swath across the eclipsed surface during it's 2-hr orbit. Maybe that data could "nudge" a predicted color/brightness distribution closer to reality, but I doubt it would be that good.

An eclipse within a few percent margin of error isn't really any different than what's already being presented.