Abstract
AbstractIt is widely recognized that biotic interactions may act as important mediators of species responses to climate change. However, collecting the abiotic and biotic covariates at the resolution and extent needed to reveal these interactions from species distribution models is often prohibitively expensive and labor‐intensive. Here we used crowd‐sourced environmental DNA sampling—the inference of species presence from genetic material in the environment—and high‐resolution habitat covariates across 630 sites over an area of nearly 10,000 km2 to build an accurate species distribution model (AUC = 0.96; prediction accuracy = 0.90) for bull trout in cold‐water habitats that incorporates fine‐scale, context‐dependent interactions with invasive brook trout. We then used this model to project possible climate change and brook trout invasion scenarios for bull trout forward in time. Our environmental DNA sampling results were concordant with traditional electrofishing samples in the basin and revealed species patterns that were consistent with previous studies: Bull trout were positively associated with larger stream sizes and negatively associated with high brook trout abundances. However, our modeling also revealed an important nuance: At high abundance, brook trout appear to exclude bull trout from small streams, even those below the thermal optima for brook trout. Climate projections suggest a loss of suitable bull trout habitat as streams warm and summer flows decrease, which could make deleterious interactions with brook trout more common in the future. Where brook trout are invading bull trout habitats, streams that are both large and cold are most likely to provide native bull trout with long‐term refuges.
Highlights
Predicting species distributions in response to climate change is key to conserving biodiversity this century (Guisan et al 2013)
Brook trout environmental DNA (eDNA) concentrations were low at most sites (76% of positive sites had
Sympatry with brook trout was associated with bull trout being found in even larger streams
Summary
Predicting species distributions in response to climate change is key to conserving biodiversity this century (Guisan et al 2013). Because similar sampling and analysis protocols have been developed to detect eDNA from bacteria, fungi, plants, insects, mollusks, fishes, amphibians, mammals, birds, and reptiles (e.g., Schmidt et al 2013, Padgett-Stewart et al 2015, Newton et al 2016, Valentini et al 2016), individual environmental samples can be used to evaluate occupancy of many species, even those that were not targets of the original sampling effort (Dysthe et al 2018) When those samples are collected at ecologically and demographically meaningful scales, patterns of species occurrence may reveal the influence of species interactions. EDNA sampling is unlike most traditional sampling methods in that sample collection is rapid and requires little expertise if robust protocols are followed (e.g., Carim et al 2016a) This facilitates crowd-sourced data collection and rapid inventories of large areas at relatively low cost (Biggs et al 2015, McKelvey et al 2016). These attributes make eDNA sampling an efficient tool for developing datasets used in SDMs that enable assessments of multiple species and their interactions
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
