Color temperature and effective temperature of a star

[alialice]

Is there someone how could explain me why is color temperature (Tcol) always greater than the effective temperature (Teff) of a star?
I know that Teff is the temperature of the last interaction surface, where photons are emitted.
In the case that the scattering opacity is the dominant one, photons have a Tcol higher than Teff. What we can observe with instruments is Tcol. But the temperature that gives luminosity is Teff. Why don't use Tcol? Please explain me the exact difference between the two.
Thanks

 

[Drakkith]

I think Teff refers to the total radiative power per square unit, while Tcol refers to the spectrum emitted if the object were a perfect blackbody. For example, according to wiki the Sun has a Teff of 5780k but a Tcol of 5900k. This would be because the Sun is not a perfect blackbody and its color will be slightly different.

A more extreme example can be found here: http://en.wikipedia.org/wiki/Color_temperature#Color_temperature_in_astronomy
The picture below is from that article. It is the spectrum of A0V star, similar to Vega. As you can see, the blue graph is the measured spectrum while the 15,000k dashed line is the Tcol. This 15,000k Tcol matches the visible output much better than the 9,500k "blackbody" dotted line, even though the effective temperature is actually 9,500k.

The picture's description in the article: The spectral power distribution of a typical A0V star (like e.g. Vega) from the HILIB stellar spectra library in comparison with a Planckian blackbody spectrum of the same effective temperature of 9500 K (dotted line) and visual color temperature of 15000 K (dashed line). Intensity B is in arbitrary units, and the curves are normalised to match at lambda = 555 nm. Both temperatures differ mainly due to the Balmer absoption bands in the ultraviolet region