The coexistence of odd and even parity magnetic fields in disc galaxies

Aims. Naive dynamo models predict that large-scale magnetic fields generated in flattened disc-like structures will be steady and symmetric with respect to the equatorial plane, whereas fields generated in quasi-spherical volumes will be oscillatory and anti-symmetric. Spiral galaxies consist of a flattened disc and a quasi-spherical halo. We thus investigate to what extent this naive understanding of symmetry properties is realised in composite disc/halo models for galactic magnetic fields. Methods. We consider generation of galactic magnetic fields in the framework of galactic mean field dynamo theory, based on the effects of differential rotation and helical turbulent motions (the alpha-effect), using conventional profiles for both generators of magnetic field in the disc and halo. The halo and disc regions are mostly separated by a substantial contrast between their turbulent diffusivities, respectively eta(d) and halo eta(h). We solve the corresponding equations of mean field electrodynamics numerically, using contrasts up to eta(h)/eta(d) = 5, while realizing that it might be realistic to consider significantly larger values. Results. In contrast to our naive expectations coexisting steady symmetric (quadrupole-like) magnetic structures in the disc and oscillating antisymmetric (dipole-like) structures in the halo were not found. Usually one component of the dynamo system enslaves the other: a more dynamo-active disc creates a symmetric field in the halo as well as in the disc or, conversely, a more dynamo-active halo generates antisymmetric magnetic fields that pervade both halo and disc. Our most interesting models are mixed parity solutions at the transition between the two regimes. Conclusions. We consider the results obtained as presenting a challenge for the contemporary theory of galactic magnetic fields. We note that there is some recent observational evidence for a difference in symmetry properties between disc and halo. We see three possible resolutions of the problem. The contrast eta(h)/eta(d) used may have been too low (values up to about 50 can be argued for but implementation of such values is rather demanding numerically). Another option is that galactic magnetic structures may be separated in two distinct classes, namely objects with disc-like and halo-like magnetic field structures. To settle this issue observationally seems to require much higher resolution than is possible with telescopes now available. Finally, of course, can not be excluded that something fundamental is lacking from our understanding of the origin of galactic magnetic fields.