cosinekitty
commented
2 years ago This is an interesting problem. Reading your link to "The New Kinetic", I would use a similar idea. You need some way to calculate the latitude and longitude for every pixel on your map. It looks like you are using a Mercator projection. But this would work for any projection, so long as you have a way of calculating the geographic coordinates for a given pixel.
First of all, there are solar eclipses and lunar eclipses. Both will require different approaches. A lunar eclipse and a crescent moon observation will both be very similar, because they are asking "can you see the Moon from this place?" The solar eclipse is asking, "is the Moon blocking sunlight from reaching this place?"
I will have to think more about the solar eclipse case. The closest thing I have right now is SearchLocalSolarEclipse, which will tell you the beginning, middle, and end of a solar eclipse at a given location on the Earth. The problem is you would have to call it for every single pixel on your map, which would be extremely slow. Maybe I can think about this to find a better answer.
Lunar eclipses and crescent moons are more feasible with the existing Astronomy Engine code. You will want to find the moment in time you are interested in for observing the Moon. For example, perhaps you call SearchLunarEclipse to find a lunar eclipse. This will tell you when the lunar eclipse starts, peaks, and ends. Let's assume we want to make a map for a single moment in time, such as the peak of a lunar eclipse.
Calculating the Moon's position is expensive (the formulas are very complicated). So you will want to calculate its position just once for your entire map. The best way would be to call GeoMoon to obtain the Moon's geocentric position vector at whatever moment in time you are interested in.
Then, for each pixel on your map, calculate the geographic coordinates (latitude, longitude). Use those coordinates to create an Observer object that you can pass into a call to ObserverVector to calculate a geocentric vector of that point on the Earth's surface. Subtract the observer vector from the moon vector to obtain a vector from the observer toward the moon.
Then call Rotation_EQJ_HOR to calculate a rotation matrix that converts equatorial J2000 vectors into local horizon vectors, for that same observer location. (This part could use some optimization to avoid redundantly calculating the Earth's axis and rotation for every single pixel on your map. I will have to come back and think about that some more.)
Pass your vector from the observer toward the moon, and the rotation matrix, into a call to RotateVector. The result will be a horizon-based vector for the given location on the Earth. The x component points north, the y component points west, and the z component points up. If z is a positive number, it means the object is above the observer's horizon, and is thus visible.
This might be enough to get you started. Again, this is very interesting! I may consider adding some new code to Astronomy Engine to make map calculations like this more efficient, because I can see how the approach I outlined above will waste time by repeating a lot of calculations for every pixel on the map.
ebraminio
Hi! Wondering if is possible to see where an eclipse and/or crescent will be visible?
See for example https://www.moonsighting.com/visibility.html and https://svs.gsfc.nasa.gov/4906
Initially my first question was how to draw day and night map which the question led me to http://thenewkinetic.blogspot.com/2011/05/how-to-draw-daynight-world-map_7282.html
Thus I have the following very initial result on my foss app :)