Whitmee & Orme (2013) showed that mammalian natal dispersal distances can be predicted cross-species from home range area, geographic range size, and body mass, with strong phylogenetic signal, but flagged four unresolved issues that bear on making trait-based estimates of mammalian dispersal trustworthy at species and population level. Substantial adjacent progress exists — a mechanistic bioenergetic model of maximum dispersal capacity, a global movement database, genetic parentage and deep-learning dispersal estimators, and bird studies demonstrating systematic limits of trait-based prediction — but there is still no macro-scale integration of direct and genetic estimates for mammals, no characterization of the true long-distance tail of the mammalian dispersal kernel, no causal resolution of the trait–dispersal relationship, and no validation of species-level predictions against local population kernels. The problem remains open.

Background

Natal dispersal is the one-way movement of an individual animal from its birth site to the site where it first attempts to breed. In mammals it governs gene flow between populations, the connectivity of fragmented landscapes, and the capacity of species to shift their geographic ranges under climate change. The dispersal kernel is the probability distribution of dispersal distances for a population or species; its far tail describes rare long-distance dispersal (LDD) events, which contribute disproportionately to range expansion and the genetic rescue of isolated populations. Direct observation (radiotelemetry, mark–recapture, ringing) systematically truncates this tail because study areas are finite: individuals that move beyond the study boundary are simply lost, so recorded maximum distances are censored rather than true maxima.

Because empirical dispersal data are scarce for most species, comparative trait-based approaches estimate dispersal from species attributes. The key priors are that median natal dispersal distance scales with body mass and home range size across terrestrial birds and mammals (Sutherland et al. 2000), and that mammalian dispersal distance is roughly proportional to home range size (Bowman et al. 2002). Indirect genetic estimates offer an alternative window: under isolation by distance (the increase of genetic differentiation with geographic distance, whose steepness reflects ongoing gene flow), the spatial distribution of genetic variants encodes the dispersal distances of past generations, and genetic parentage assignment (identifying parent–offspring pairs genetically to reconstruct realized movements) recovers individual dispersal events. Direct and indirect estimates, however, often disagree, because they average over different timescales and suffer different biases (Thompson & Goodman 1997).

The source work, Whitmee & Orme (2013), compiled a cross-species dataset of mammalian natal dispersal distances and analyzed it with phylogenetic comparative methods — statistical models that correct for the non-independence of related species (phylogenetic signal being the tendency of closely related species to resemble each other). They found strong phylogenetic signal in both maximum and median dispersal distance, identified home range area, geographic range size, and body mass as the most important predictors, and showed by cross-validation that dispersal distance per generation can be estimated for mammals from comparatively few traits, suggesting the models could be extended to species whose dispersal is unknown.

That claim of extendability is exactly what the authors themselves left unresolved, in four linked respects. First, the relationship between empirical dispersal distances and the true tail of the dispersal kernel is unknown, because direct studies miss rare LDD events and no macro-scale datasets combining direct observations with indirect genetic estimates (for example, isolation-by-distance or parentage-based) existed in 2013. Second, some species are consistently over- or under-predicted by the trait models, and it is unsettled whether these deviations are biological (genuine mismatch between a species' vagility, i.e. its movement capacity, and its life-history traits), methodological (biases in the underlying empirical data such as study-area limits or landscape fragmentation), or caused by the omission of landscape-level factors; the quality and quantity of data needed for accurate species-specific predictions in a landscape context were not assessed. Third, the link between traits and dispersal is purely correlative: whether home range size causally determines dispersal distance, or both reflect a common underlying factor such as energy requirements, and more broadly how 'dispersive traits' co-evolve, was beyond the study's scope and unresolved. Fourth, the strong species-level phylogenetic signal coexists with evidence that dispersal-kernel shapes are labile within species and responsive to landscape heterogeneity, so whether species-level predictions can be extrapolated to local populations under different environmental contexts required — and still requires — validation.

Problem Statement

For terrestrial mammals, as posed by Whitmee & Orme (2013) and still unanswered: establish trustworthy, species- and population-level estimates of natal dispersal distance, including rules for predicting dispersal distance for species where it is unknown. Since the dispersal kernel is defined for a population or species, the kernels to be characterized — including the far tail in facet (1) — are those of mammalian populations and species rather than a single pooled distribution. This involves four linked facets: (1) determining the true tail of the mammalian natal dispersal kernel — the frequency and distance of rare long-distance dispersal events — given that direct observational records systematically under-report such events because finite study areas censor observed distances at their limits; (2) explaining why trait-based models consistently over- or under-predict dispersal distance for particular species — that is, decomposing prediction error into genuine biological mismatch between vagility and life history, methodological bias in the underlying empirical data (study-area limits, landscape fragmentation), and omission of landscape-level factors — and thereby assessing the quality and quantity of data needed for accurate species-specific predictions in a landscape context; (3) replacing the correlative trait–dispersal relationship with a causal account of the pathways linking life-history and ecological traits to dispersal distance, including whether home range size determines dispersal distance or both reflect a common underlying factor such as energy requirements, and how dispersive traits co-evolve; and (4) testing whether species-level predictions, which carry strong phylogenetic signal, can be extrapolated to local populations whose dispersal-kernel shapes are labile and responsive to landscape heterogeneity. These four facets are linked components of a single problem: each is independently attackable and a defensible result on any one is a valid partial contribution, but a full resolution of the problem addresses all four.

References

  1. Whitmee S, Orme CDL (2013) Predicting dispersal distance in mammals: a trait-based approach. Journal of Animal Ecology 82(1):211–221. doi:10.1111/j.1365-2656.2012.02030.x, https://doi.org/10.1111/j.1365-2656.2012.02030.x
  2. Thompson PM, Goodman S (1997) Direct and indirect estimates of dispersal distances. Trends in Ecology & Evolution 12(5):195–196. doi:10.1016/s0169-5347(97)84095-5, https://doi.org/10.1016/s0169-5347(97)84095-5
  3. Sutherland GD, Harestad AS, Price K, Lertzman K (2000) Scaling of natal dispersal distances in terrestrial birds and mammals. Conservation Ecology 4(1):art.16. doi:10.5751/es-00184-040116, https://doi.org/10.5751/es-00184-040116
  4. Bowman J, Jaeger JAG, Fahrig L (2002) Dispersal distance of mammals is proportional to home range size. Ecology 83(7):2049–2055. doi:10.1890/0012-9658(2002)083[2049:DDOMIP]2.0.CO;2, https://doi.org/10.1890/0012-9658(2002)083%5B2049:DDOMIP%5D2.0.CO;2
  5. Stevens VM, Trochet A, Blanchet S, Moulherat S, Clobert J, Baguette M (2013) Dispersal syndromes and the use of life-histories to predict dispersal. Evolutionary Applications 6(4):630–642. doi:10.1111/eva.12049, https://doi.org/10.1111/eva.12049
  6. Stevens VM, Whitmee S, Le Galliard J-F, Clobert J, Böhning-Gaese K, Bonte D, Brändle M, Dehling DM, Hof C, Trochet A, Baguette M (2014) A comparative analysis of dispersal syndromes in terrestrial and semi-terrestrial animals. Ecology Letters 17(8):1039–1052. doi:10.1111/ele.12303, https://doi.org/10.1111/ele.12303
  7. Bestion E, Clobert J, Cote J (2015) Dispersal response to climate change: scaling down to intraspecific variation. Ecology Letters 18(11):1226–1233. doi:10.1111/ele.12502, https://doi.org/10.1111/ele.12502
  8. Santini L, Cornulier T, Bullock JM, Palmer SCF, White SM, Hodgson JA, Bocedi G, Travis JMM (2016) A trait-based approach for predicting species responses to environmental change from sparse data: how well might terrestrial mammals track climate change? Global Change Biology 22(7):2415–2424. doi:10.1111/gcb.13271, https://doi.org/10.1111/gcb.13271
  9. Bode M, Williamson DH, Harrison HB, Outram N, Jones GP (2018) Estimating dispersal kernels using genetic parentage data. Methods in Ecology and Evolution 9(3):490–501. doi:10.1111/2041-210x.12922, https://doi.org/10.1111/2041-210x.12922
  10. Hall LA, Van Schmidt ND, Beissinger SR (2018) Validating dispersal distances inferred from autoregressive occupancy models with genetic parentage assignments. Journal of Animal Ecology 87(3):691–702. doi:10.1111/1365-2656.12811, https://doi.org/10.1111/1365-2656.12811
  11. Norman JA, Phillips SS, Blackmore CJ, Goldingay R, Christidis L (2019) Integrating measures of long-distance dispersal into vertebrate conservation planning: scaling relationships and parentage-based dispersal analysis in the koala. Conservation Genetics 20(5):1163–1174. doi:10.1007/s10592-019-01203-2, https://doi.org/10.1007/s10592-019-01203-2
  12. Smith CCR, Kern AD (2023) disperseNN2: a neural network for estimating dispersal distance from georeferenced polymorphism data. BMC Bioinformatics 24(1):385. doi:10.1186/s12859-023-05522-7, https://doi.org/10.1186/s12859-023-05522-7
  13. Logghe G, Batsleer F, Fandos G, Maes D, Bonte D (2025) Consistent covariation of dispersal, life history and thermal niche across terrestrial arthropods. bioRxiv preprint. doi:10.1101/2025.01.10.632369, https://doi.org/10.1101/2025.01.10.632369
  14. Wilkinson C., Brose U, Dyer A, Hirt MR, Ryser R (2025) A Mechanistic Approach to Animal Dispersal—Quantifying Energetics and Maximum Distances. Ecology Letters 28(2):e70085. doi:10.1111/ele.70085, https://doi.org/10.1111/ele.70085
  15. Wilkinson C, Berti E, Dyer A, Menz MHM, Meyer C, Tucker MA, Wootton KL, Broekman MJE, Hoeks S, Roth J, Ryser R, Hirt MR (2025) MacroMove: a global database of animal movement across taxa and movement types. bioRxiv preprint. doi:10.64898/2025.12.08.692918, https://doi.org/10.64898/2025.12.08.692918
  16. Fandos G, Talluto M, Fiedler W, Robinson RA, Thorup K, Zurell D (2021) Standardised empirical dispersal kernels emphasise the pervasiveness of long-distance dispersal in European birds. bioRxiv preprint. doi:10.1101/2021.11.08.467775, https://doi.org/10.1101/2021.11.08.467775
  17. Fandos G, Robinson RA, Zurell D (2026) Simple mechanistic traits outperform complex syndromes in predicting avian dispersal distances. Communications Biology 9(1):376. doi:10.1038/s42003-026-09676-x, https://doi.org/10.1038/s42003-026-09676-x