Direction 01 · Eco-Evolutionary Macroecology & Population Genetics
How much does evolutionary history tell us about a species' environmental niche?
Related species often occupy similar environments, yet similarity can reflect conserved physiological responses, shared accessible geography (M), or spatially uneven sampling. With Jorge Soberón, I co-develop arfun to model how ellipsoidal niche position and covariance evolve under Brownian Motion and Ornstein–Uhlenbeck processes along dated phylogenies. Using controlled simulations and withheld-record experiments on well-sampled clades, we quantify when relatives' information genuinely improves predictions for sparsely sampled species and when it misleads.
Within species, I developed the Suitability Prevalence Area (SPA) theoretical framework (Ecography, 2025; co-first author) to reconstruct where environmental suitability persisted across the Late Quaternary and link that temporal stability to standing population-genetic diversity in fragmented montane populations (Sciurus aberti). Developing infrastructure in ancemb (with Henry Secaira-Morocho) and my independent phyloseek / genoaligner phylogenomic pipeline extends these comparative tests across deep clades.
Direction 02 · Stochastic Population Dynamics & Coexistence
What do differences in environmental response mean for persistence and coexistence?
Standard distribution models rely on long-term climate averages, yet population trajectories respond nonlinearly to year-to-year environmental sequences. With Daniel Reuman and collaborators, I co-developed XSDM (bioRxiv), embedding Lewontin–Cohen stochastic demography (pt+1 = λtpt) and long-term stochastic growth rates into geographic range modeling. Because nonlinear averaging across concave performance curves depresses long-run growth even where mean climate appears benign, accounting for interannual variability reduced projected distributions by 22% on average and up to 45% across focal species.
Ongoing work develops Bayesian XSDM (xsdmbayes) informed by experimental thermal and physiological performance curves to quantify parameter and forecast uncertainty. Because the baseline model estimates single-species density-independent growth from temporal means of log-multipliers, my next theoretical and empirical steps extend the framework to temporal ordering and autocorrelation, stage-structured demography, and density-dependent low-density invasion rates — testing when response-breadth and asymmetry differences promote or erode multi-species coexistence under shared climate fluctuations.
Direction 03 · Community Assembly, Ecological Networks & Ecoinformatics
When does evolutionary ancestry explain specialization in ecological networks?
In ecological networks, narrow partner breadth can arise from conserved trait compatibility or simply from which species co-occur and are sampled together. Across wildlife host–pathogen systems (PNAS, 2022), ectoparasite assemblages (Parasitology, 2025), and parasitic-plant networks (Phoradendron; Am. J. Bot., 2025), I showed that phylogenetic, environmental, and geographic distances carry system-specific explanatory weight — treating relatedness as a testable mechanistic hypothesis rather than a universal proxy.
Evaluating network assembly at macroecological scales requires traceable primary data and inference methods that treat unobserved pairs as uncertain rather than true negatives. I designed and maintain EcoTagger, which normalizes taxa against NCBI Taxonomy and the GBIF Backbone and accepts an interaction only when anchored to a verbatim sentence or table row (>100,000 records across host–parasite, mutualistic, and trophic systems, audited via a blind multi-reviewer portal). On this foundation, I develop positive–unlabeled learning methods (phylopred, xplus, xpam, interseekr) and prospective spillover-surveillance prototypes (WildSeek, in development) to test how climate variability rewires encounter opportunities and context-dependent interactions.