I do not know IBD and IBR. If Mantel tests show no significant spatial autocorrelation, this indicates no coherent geographical pattern in gene frequencies.

I can think of at least two explanations.

There are selectively induced differences among populations due to small-scale and repeated patterns of local environmental differences, so that much variation occurs even among populations close to one another. Correlation between gene frequencies and environmental variables might give you a clue.

Alternatively, if the populations are widespread, but actually are small and have rather little gene flow between them, then there may be much variation as a result of drift or a set of founder effects.

In both cases, if you subdivide your populations into geographic blocks, you will find that variation within each accounts for most of the variation overall (Fst statistics)

Continously distributed species (means individuals or population). This condition rarely happened due to the species nitch for resident species or habitat route in migrant species. In addition to gene flow caused by migrant individuals while this condition didnt happened easily because population would reject the strangers particularly in reproduction process. So, as I indicate to Dr. Cameron answer, I expect always there are genetic variation among (isolated) population.

It would be helpful to think about the situation in context to the study system and your analysis. Are you comparing genetic differences among populations or individuals? In other words, in your IDB tests are you grouping individuals into suspected populations and estimating genetic distance between these groups or estimating genetic distance between the actual individuals themselves? What type of genetic data do you have? How many populations/individuals are there in the dataset? Most importantly, does it make sense given the biology of the species that it could be totally panmictic? It is entirely possibly that there really is no genetic structure if gene flow is that high and/or the markers are not providing enough resolution?

Here are a couple of options to consider. First, you could use the program Migrate to explicit test whether your system displays more of an IDB or panmictic structure. Second, you can conduct a Mantel correlogram to assess autocorrelation between genetic and geographic distance over various distance classes. Third, there are several methods out there that can measure non-linear relationships between genetic and geographic distance. Check out the sPCA analysis in adegenet (https://www.nature.com/articles/hdy200834). Also check out this Bayesian kriging method (Article Nonstationary patterns of isolation-by-distance: inferring m...