Persistence of environmental DNA in marine systems

Abstract

As environmental DNA (eDNA) becomes an increasingly valuable resource for marine ecosystem monitoring, understanding variation in its persistence across contrasting environments is critical. Here, we quantify the breakdown of macrobial eDNA over a spatio-temporal axis of locally extreme conditions, varying from ocean-influenced offshore to urban-inshore, and between winter and summer. We report that eDNA degrades 1.6 times faster in the inshore environment than the offshore environment, but contrary to expectation we find no difference over season. Analysis of environmental covariables show a spatial gradient of salinity and a temporal gradient of pH, with salinity-or the biotic correlates thereof-most important. Based on our estimated inshore eDNA half-life and naturally occurring eDNA concentrations, we estimate that eDNA may be detected for around 48 h, offering potential to collect ecological community data of high local fidelity. We conclude by placing these results in the context of previously published eDNA decay rates.

Fig. 1

Exponential eDNA decay. Environmental DNA decay over 192 h, two seasons (summer and winter), two species (shanny and common shore crab assays) and five experimental water treatments simulating an environmental gradient. Response variable is eDNA concentration in copies per litre of treatment water. Zero hour data at t = 0 are included. Trend lines show an exponential decay model

Fig. 2

Rates of eDNA decay. Environmental DNA decay over 192 h, two seasons (summer and winter), two species (shanny and common shore crab assays) and four experimental water treatments simulating an environmental gradient. The response variable is natural log_e transformed eDNA concentration normalised as a proportion of starting concentration, i.e. the value at time t = x divided by the value at time t = 0. Zero hour data at t = 0 were subsequently excluded after proportions were calculated. Trend lines show fitted linear regression values from the optimal linear mixed-effects model

Fig. 3

Half-life of eDNA. Environmental DNA half-lives (hours) for each water treatment and season–species combination. Half-lives were calculated from rate constants estimated from an optimal linear mixed-effects model using the emtrends function in emmeans. Dots represent point estimates derived from the model, with bars showing 95% confidence intervals also estimated by the model