Probably not much effect from iron in the ground.
HF communications are very sensitive to the ionosphere and other natural effects. Predictions are difficult, and only give a probability, never a guarantee. Real data are thin on the ground, and you know that the plural of anecdote is not data.
HF propagation happens mostly in the air, and by reflection off the ionosphere. The ground only gets involved at the antenna, and possibly if you are using ground wave propagation. Ground wave is more an MF thing, AM radio etc, above a few MHz it is only good for a few km.
The radiation pattern and efficiency of an antenna are affected by the ground constants, and the low-angle radiation, so important for long range HF, is particularly sensitive. No matter how far the signal will travel, the only ground that counts is the patch around each antenna. The size of the patch varies, but it's generally only some wavelengths in radius. In reasonable ground, all the RF conduction happens in the first few metres, so don't worry about what is being mined far below.
The two ground parameters, conductivity and permittivity, vary widely between locations, but in general, the "richer" the soil, the "better" the ground. So fertile farmland is much better than a desert. This effect is well studied.
The richness of the ground depends on its water content, the geological type of rock etc, and the minerals. Most ground is of course silicon dioxide, sand, which is an insulator. The size of the grains, and the other minerals, are what change its performance.
Iron oxide is not conductive, so along with the SiO2, AlO2 etc, would make up the inert part of the ground. Salts and water dominate that conductive parts of the ground.
To study this further, try to investigate the effect of various minerals on the ground permittivity and conductivity. This is where the hard data lies. Stories about good and bad reception, or even careful measurements of "how good the signals are", have too many confounding factors, and are not going to lead to useful conclusions.