In this specific case the dominant allele is the presence of a gene (RHD), and the recessive allele is (usually) the absence of that gene. This is why you will see the statement "there is no d gene". There does indeed appear to be an environmental factor which governs the frequency of the various genotypes - Toxoplasma gondii.
I find this topic quite a challenge. Here is a distillation of my research.
It is thought that the RHD gene may have arisen through a duplication of the adjacent RHCE gene. It seems that the possession of this gene (a D allele) confers a selective advantage in individuals suffering from toxoplasmosis (infection by the protozoan Toxoplasma gondii. Toxoplasmosis has an effect upon what is termed psychomotor performance, which is usually measured as reaction time. So, for example, it has been reported that there is an increased incidence of road traffic accidents amongst those suffering from (asymptomatic) toxoplasmosis.
Heterozygosity at RHD seems to reduce this effect of toxoplasmosis. So, the current theory is that the D allele appeared in humans because of the prevalence of toxoplasmosis, with the selective disadvantage of haemolytic disease of the newborn being balanced by the heterozygote advantage (similar to the malaria/sickle cell anaemia interaction). There is currently no biochemical explanation of why the RHD gene product has this effect.
In a region with a very low risk of toxoplasmosis there would be a strong selection for the (ancestral?) dd genotype since these individuals have the best reaction time of all (when not infected). The definitive hosts (i.e. where the parasite reproduces) of Toxoplasma gondii are cats, and these were rare in Europe before the domestication of cats. This could explain why the dd genotype is more frequent in this area.