Are the spatial and temporal scales biologically meaningful?
Paz-Vinas et al. focus their critique on our time series analyses. Our analysis does conflate space and time to a certain extent, although we used the smallest possible spatial scale (1000 km) with sufficient data to test our hypothesis. We explain this (page 58) and make careful statements not to over-emphasize these results (i.e., no consistent effect of humans on animal IGD).
Whether 1000 km exceeds species’ dispersal capabilities is species-specific. Geographic and anthropogenic barriers can further fragment species’ ranges. Because there is no ‘right’ resolution, we introduced a multi-scale approach to assess IGD patterns at the global extent. Ideally, species traits, ranges and geological features should be considered in future studies.
Paz-Vinas et al. selected fish in their comment, which in general have unique dispersal abilities, and are the most ‘noisy’ taxa in our study as a result of watershed constraints. We did not account for watershed structure which is a shortcoming of our analysis, although fish populations constitute only 11.4% of our temporal dataset. The purely spatial analysis, that is the focus of our paper, assesses IGD patterns at smaller spatial scales (sequences < 10 km apart) where the risk of pooling sequences from isolated populations with divergent land use is reduced. As we have discussed in our Appendix, there is little to no spatial variation among sequences in the majority of populations: “the mean great circle distance among sequences was 0 for ~50% of species in our dataset, and usually < 100 km for other populations (Fig. S1) ”. Thus, the “daisy-chaining” issue is irrelevant for the majority of our analyses. We would have liked Paz-Vinas et al. to acknowledge that the issue they raise only applies to our time series analysis. Future macrogenetic studies should be aware of the consequences of aggregating sequences into populations.
Regardless of the marker(s) chosen, or method of forming biologically meaningful populations, Paz-Vinas et al.’s comment does not indicate either way (positive or negative) how we purportedly over-interpret and misinform readers on the effects of humans on IGD. Theory predicts positive and negative effects of humans on IGD, so inconsistent trends are not surprising.
We agree that macrogenetics will progress faster when researchers align behind a set of best practices for data, analyses, and the communication of uncertainties in our understanding of the patterns arising from data limitations. We look forward to seeing future analyses as the field tackles the problem of assessing human impacts on genetic diversity.