Navigating between global robustness and fine-scale resolution: the potential of demographic inference based on identical-by-descent segments.

Understanding the demographic processes shaping population connectivity is essential for conservation. Yet disentangling the effects of dispersal and local effective density remains challenging because traditional genetic approaches integrate signals over evolutionary timescales. Identity-by-descent (IBD) segments provide opportunities to reconstruct recent demographic connectivity over timescales relevant to contemporary conservation challenges, but their reliability under realistic ecological conditions remains poorly evaluated. Here, we assess the ability of two promising spatiall IBD-based methods, IBD-Analysis and MAPS, to estimate effective density and dispersal variance. Using a simulation-inference framework combining spatially explicit simulations, IBD segment extraction, and demographic inference, we evaluated both methods across a wide range of demographic, ecological and methodological scenarios. Both approaches reliably recovered demographic parameters under homogeneous landscapes, with IBD-Analysis providing accurate estimates of global effective density and dispersal across a broad range of conditions. Spatial heterogeneity emerged as the main driver of inference performance. When demographic parameters varied across space, IBD-Analysis integrated local variation into biased global estimates, whereas MAPS captured spatial patterns but showed substantial variability and generated spurious spatial structure in homogeneous landscapes. Methodological factors had comparatively weaker effects than ecological complexity, highlighting the importance of accounting for landscape structure when applying IBD-based demographic inference. These results highlight the potential of IBD-based approaches to move beyond descriptive patterns of genetic connectivity toward mechanistic inference of demographic processes. By distinguishing the contributions of dispersal and density, these methods can improve conservation decisions by linking genomic patterns to key demographic parameters relevant for conservation.