2023MNRAS.525.3486J

Photometric metallicity formulae of fundamental-mode RR Lyr (RRab) stars are presented using globular-cluster data exclusively. The aim is to check whether this selection may help increasing the overall accuracy of the fits and eliminating the systematic bias of the photometric results, namely that they tend to overestimate [Fe/H] of the most metal-poor variables. The G band time-series data available in the Gaia DR3 archive and a new compilation of the published spectroscopic globular cluster [Fe/H] values on a uniform solar reference metallicity scale are utilized. We have derived a new [Fe/H]_phot - P, φ₃₁ formula, and have diagnosed that no significant increase in the accuracy of the fit can be achieved using non-linear or multiparameter formulae. The best result is obtained when different formulae are applied for variables with Oosterhoff (Oo)-type I and II properties. However, even this solution cannot eliminate the systematic bias of the results completely. This separation of the variables has also led to the conclusion that the photometric estimates of the [Fe/H] are less reliable for the Oo-type II variables than for the Oo-type I sample. Published [Fe/H]_phot values and the results of the available photometric formulae in the Gaia G band are compared with the present results. It is found that each of the solutions yields very similar results, with similar accuracy and systematic biases. Major differences are detected only in the zero-points of the [Fe/H] scales, and these offsets are larger than differences in the accepted solar reference values would explain.