Partitioning environmental and philopatric drivers of nest site selection in an estuary-endemic turtle (Malaclemys terrapin)

Iwo Gross , Logan Register, Matthew Wolak

Abstract
1. Oviposition site selection is a critical maternal effect with direct implications for population-level adaptation to environmental change. This process is driven by a complex interplay of environmental, density-dependent, and social factors whose relative contributions to site selection are rarely quantified simultaneously. Natal philopatry, for example, is a maternal effect that can direct oviposition site selection independently of local conditions, thus confounding estimates of adaptive habitat selection.
2. We conducted a spatially-explicit habitat selection analysis of oviposition sites in the diamond-backed terrapin (Malaclemys terrapin), a species in which nesting habitat selection and nesting site philopatry have been documented yet never evaluated within a unified statistical framework. We hypothesized that females select sites based on vegetation and topographic features that maximize offspring survival. We predicted these patterns would exhibit scale-dependence, and that selected conditions would promote hatching success.
3. We integrated resource selection functions with distance-based Moran's eigenvector maps and drone-derived three-dimensional habitat maps to analyze terrapin nest-habitat associations and partition variance in nest site selection among environmental features and residual spatial structure across multiple scales. This framework controls for bias due to spatial autocorrelation in nest occurrence data, quantifies scale-dependence of habitat selection, and partitions residual spatial variance consistent with latent social processes.
4. We found that terrapins selected vegetation complexity and elevation at broad spatial scales (1773 to 1927m), and terrain aspect at moderate scales (1192 to 1773m). The strength and direction of selection varied spatially across several habitat features, suggesting latent philopatric processes influence nest distributions independently of measured environmental gradients. Furthermore, maternal selection for nest sites with vegetation and elevation values that predict reduced hatching success suggest a philopatric ecological trap in which maternally inherited site preferences have become decoupled from current habitat quality.
5. We demonstrate that environmental habitat quality alone cannot fully explain terrapin nest distributions, and that spatially-explicit variance partitioning can reveal latent philopatric signals in point-occurrence behavioral datasets when direct social observation is unavailable. These results highlight the importance of simultaneously modeling environmental and social drivers of maternal habitat selection, particularly as environmental change might decouple historically adaptive site use from environmental conditions.