In painting scale models there is always a danger of reverse engineering a personal choice of paint colour on the model to have relevance to the real paint colour whereas it is perhaps better to use the real paint colour as the basis for the personal choice - having regard to factors like paint brand and type, scale, weathering, etc. That's not a criticism of personal choice influencing belief btw, merely an observation, but it is very prevalent amongst modellers, myself included.
Colour in paint coatings is a two-way street. The paint surface absorbs and scatters light but the light reflected back to the observer's eyes depends on his or her receptors in order to see it. It's like a radio broadcast where despite the excellent quality of the transmission it will still depend on the quality of the receiver how the music sounds. Call it "subjective individual colour perception" which is the main cause of most online arguments about how the "colour music" sounds rather than what is being transmitted.
Reflectivity depends on both illumination and angle which is why colour science theory depends on determining that in order to fix and define a specific colour. The light wavelengths reflected from the paint surface (and actually from within it too) depend upon pigment and particle size which determine their "colour". Metamerism inherent in a colour space is different from issues of "scale". The same paint colour on a full-size aircraft and model will have the same metameric qualities because of the pigment type and particle size. The difference will be in plane (as in surface) size, angle and illumination. These contribute to the diffuse reflectivity by which the overall colour is recognised. In other words the full size of an aircraft fuselage or wing reflecting light compared to the (comparably) small and multi-faceted model. However, beyond these perceptive differences and in order to communicate and identify it effectively colour has to be defined in measurable terms. Paint colour might look different dependent upon these various factors but will still have a definable, measurable "fixed" value in colour science terms. Otherwise it would be impossible to produce and manage paint colour to a consistency that is reasonable (even so variance is a factor almost impossible to eradicate and it is only relatively recently that manufacturers have managed to successfully integrate measurement and visual assessment to improve consistency). Modellers tend to focus on the reception bit rather than the transmission bit which of course means there can be a significant margin of error in perception and therefore replication.
In reality hobby paint colour seldom matches the original paint in terms of pigment and almost certainly not in particle size. It seeks only to replicate or imitate the original appearance because there is always more than one way to skin a cat. Some hobby paints have sought to build in "scale" or toned down colour with various degrees of success. Modellers often do the same and there are various propositions out there for achieving the effect, some highly complex whilst others are straightforward and simple. Therefore it is probable that the metamerism of paint on a model might be different to that on the original subject, even where paint of similar appearance is applied. In some cases it is apparent that attempts to "scale" colour, either by hobby paint companies or individual modellers, result in colour which is significantly different in hue to the original and does not replicate or even imitate it. The balance between artifice that looks right and accuracy that looks wrong is fragile!
Apart from metamerism inherent in the paint colour there is metameric observer failure which is covered in the Zero e-guide on page 6. But there is another form of failure relevant to models which is called "field-size metameric failure". It occurs because the relative proportions of the three cone types in the retina vary from the centre of the visual field to the periphery, so that colours that match when viewed as very small, centrally fixated areas may appear different when presented as large colour areas. In many industrial applications, large field colour matches are used to define colour tolerances. The main effect in models is where full-size paint colour swatches, relatively "small, centrally fixated areas" are replicated precisely resulting in models which often appear too dark. This is more relevant to issues of "scale" than any atmospheric distortion of light over distance.
It is often surprising when viewing a colour swatch some distance from the painted object - a wall or door, say - and thinking "too dark", to find that when laid directly on the paint surface the swatch is identical. The reverse effect on a model is to match full-size, paint the model and then wonder why it appears darker looking than the real aircraft. That is perhaps the essence of "scale colour". Endeavouring to replicate the appearance of a "large colour area" on the "small, centrally fixated area" of a scale model. By matching two "small, centrally fixated areas" the result will always tend to look darker than it does in real life.