San Francisco Estuary and Watershed Science

Current ecological conditions in the San Francisco Estuary are considered inhospitable to many native estuarine species, and have placed the endemic Delta Smelt (Hypomesus transpacificus) at serious risk. Programmatic monitoring regimes conducted by government agencies are insufficient for associating Delta Smelt occurrence with relevant habitat attributes, which limits inference about the relationships between putative habitat, restoration activities, and population response. Indirect observation of macro-organisms via detection of environmental DNA (eDNA) has proved a compelling alternative monitoring approach, particularly for rare and/or protected species. Yet, factors that influence eDNA detection in estuarine habitats remain poorly characterized, which hinders refinement of sampling methods. This study employed a fixed sampling array to explore how tidal phases affect the detection of Delta Smelt eDNA. Our primary objective was to estimate the effects of covariate metrics on concentration of eDNA (calculated as ln[eDNA]) observed in the tidal environment. Secondary objectives included comparing the effect of distance on eDNA concentration in the tidal system vs. a unidirectional system and estimating how time since species absence affects observed eDNA concentration. Model predictors that consistently affected eDNA concentration were distance from source, eddy diffusivity, time since species absence, tidal direction, and species. Distance had a consistently negative effect on eDNA concentration across both systems, though non-detections were more frequent in the tidal system than expected in a unidirectional one. Environmental DNA detections decreased over time, which suggests that positive eDNA detections from estuarine water sampling are more likely to co-occur contemporaneously with the actual presence of individuals. These findings improve the capacity to design sampling strategies that will detect target species within an estimated probability, if individuals are present within a specified distance. Enhanced detections using eDNA sampling approaches would refine determinations of if and when Delta Smelt are present at a location, which in turn lessens known information gaps related to species occurrence and distribution.