Re: "just north of Antarctica"
Surely most things are a certain colour due to 'reflected light'?
Yes, or emitted light. In some cases color is influenced by various perceptual quirks, and perceived color is not just a simple function of visible-light wavelength and amplitude,1 but to a first approximation, reflected and emitted light are color.
There's also a philosophical question of whether it makes sense to qualify the description of an object's color (as a sensory attribute) with "appears". Is the color of an object anything other than an appearance? We can say, for example, "we have measured the light reflected from this object given a sufficiently strong diffuse white light source containing all frequencies in the visible spectrum, and that reflected light contains the following frequencies blah blah blah", but that's clearly not what's meant by "color" in the common-language sense.
When people say "polar bears are white", they mean precisely "when I look at a polar bear, the overwhelming color quale I experience is of white" - nothing more or less (though they're unlikely to phrase it that way, unless they're, y'know, me). They don't give a crap (for the purposes of evaluating the bear's color) whether polar bear skin is black or white or stripped pink and orange, or how much the hairs of its fur reflect, transmit, or absorb and emit light.
1Perceived color is highly influence by the relative sensitivity of the cones to different frequencies, which is why the sky looks blue (when it doesn't look black, or grey, etc) rather than violet. And it's significantly influenced by color juxtapositions, both spacial and temporal. And, of course, color perception and sensitivity vary widely among the population; there's a sex-linked characteristic which gives a significant population of women four types of hue stimuli instead of the normal three, while on the other end there's a considerable number of people with limited color vision.