Environmental DNA sampling detects between-habitat variation in soil arthropod communities, but is a poor indicator of fine-scale spatial and seasonal variation

Abstract

Arthropods have long been appreciated as useful and important ecological indicators of beneficial or detrimental changes occurring in their environment, especially those driven by anthropogenic activity. However, morphological identification, especially of key groups such as soil arthropods, is laborious and requires high-level expertise, limiting the spatiotemporal scales of such monitoring. Molecular methods based on environmental DNA (eDNA) sampling have potential for scaling up sampling of arthropod biodiversity for environmental monitoring, due to the high-throughput capabilities of such methods. However, we still do not have a clear understanding of how well molecular methods detect variation in arthropod biodiversity over space and time, particularly at fine grains. We therefore conducted a study employing a standard eDNA metabarcoding approach to monitor the composition of subterranean arthropod communities in a homogeneous forest (pine plantation) and a heterogeneous forest (regenerating native woody vegetation). We sampled the two habitats at a range of spatial scales (cm to m) and with temporal replicates over ten months to capture seasonal variation. Our analysis of almost 800 samples showed that arthropod community composition differed significantly between the two forest types and habitat explained 16.7% of the variation observed overall. However, we did not observe strong temporal or seasonal change in either richness or composition. We attribute the perceived lack of fine-grained spatiotemporal pattern to current limitations of eDNA metabarcoding, which captured inadequate biodiversity to detect changes at these fine grains. We show that increased sampling and replicate DNA extractions result in more biodiversity being captured and reveal amplification biases and a lack of taxonomic assignments. Identifying these weaknesses highlights where efforts to improve eDNA metabarcoding should be focused. The many benefits that eDNA metabarcoding can bring to biodiversity based ecological monitoring mean the efforts still required to improve the reliability and reproducibility of these methods are undoubtedly worthwhile priorities.