On the spectral resolution of the MILES stellar library

Empirical stellar libraries are extensively used to extract stellar kinematics in galaxies and to build stellar population models. An accurate knowledge of the spectral resolution of these libraries is critical to avoid propagation errors and uncertain estimates of the intrinsic stellar velocity dispersion of galaxies. In this research note we re-assess the spectral resolution of the MILES stellar library and of the stellar population models based on it. This exercise was performed, because of a recent controversy over the exact MILES resolution. We perform our test through the comparison of MILES stellar spectra with three different sets of higher-resolution templates, one fully theoretical - the MARCS library - and two empirical ones, namely the Indo-U.S. and ELODIE v3.1 libraries. The theoretical template has a well-defined very high (R=20000) resolution. Hence errors on this theoretical value do not affect our conclusions. Our approach based on the MARCS library was crucial to constrain the values of the resolution also for the other two empirical templates. We find that the MILES resolution has previously been slightly overestimated. We derive a new spectral resolution of 2.54 A FWHM, instead of the nominal 2.3 A. The reason for this difference is due to an overestimation of the resolution for the Indo-U.S. library that was previously used for estimates of the MILES resolution. For the Indo-U.S. we obtain a new value of 1.35 A FWHM. Most importantly, the results derived from the MARCS and ELODIE libraries are in very good agreement. These results are important for users of the MILES spectra library and for further development of stellar population models aimed to obtain accurate stellar kinematics in galaxies.