I chose the options 16 bit, RGB, PNG, etc, and the image saved on my desktop says in its details "48 bit" color depth, then when saved again in a paint program, it got lowered too to 32 bit and 1/4th its original size, which had me upset btw. But did it even save it in 48 bit in the first place?
There are lots of programs that refer to bit depths in different ways. The confusion comes from whether they are referring to the number of bits per pixel or the number of bits per color channel.
16 bit per color channel RGB is also referred as 48bit color by some programs ( each pixel will use 48 bits to store color information: 16 bits for the Red channel, 16 for the Green and 16 for the Blue: 16 x 3 = 48).
16 bit RGB with Alpha channel is also referred sometimes as 64bit (RGB + A or 16 x 3 + 16 = 64 )
What most programs label 32 bit images are really 8 bit per channel images plus an Alpha channel: (RGB + Alpha 8 x 3 + 8 = 32 bits per pixel.
Without the Alpha channel an 8bit RGB image would use 24bits per pixel.
To make things more confusing it is possible to use 32bit per channel when using OpenEXR or HDR files...
There are some programs (like versions of GIMP previous to 2.9 or other paint programs) that cannot deal with 16 bit images and will convert them to 8 bit which will always result in data loss.
8 bit images allow for only 256 different values per color channel, with a total of 16 million combinations.
Even though 16 million colors sounds like a lot and is close to the limit of what the eye can see, it presents problems when representing subtle gradations where you need many more values. The most common artifact that you'll see in 8 bit images is known as "banding". In this image banding is clearly visible in the sky:
Photo by geograph.org.uk
16 bit allows for 65,536 values per channel giving over a trillion possible combinations. Using a higher bit rate will solve most of the banding issues and will give you more flexibility in terms of color correcting or further processing of the images.