Intercouncil high replicate eDNA trial
How a nationwide council trial set the standard for river eDNA monitoring
In the summer of 2020/2021, New Zealand's regional councils joined forces on a nationwide eDNA trial across 53 well-known rivers and streams. The results set the six-replicate sampling protocol now widely used for riverine eDNA monitoring, and the data went on to train a new eDNA-based measure of river health.
Councils already had well-established biomonitoring programmes, but adding eDNA raised practical questions. How many samples are enough to capture the fish and invertebrates at a site? And does filtering more water make a difference? Without clear answers, results from different councils would be hard to compare.
Between December 2020 and May 2021, regional council staff each sampled one to five of their well-known monitoring sites, covering 53 rivers and streams and a wide range of habitats and flow conditions. At each site, they collected 16 replicate samples at a single point in time using Wilderlab's eDNA syringe kits. Eight were standard 1 L filtered samples and eight were "boosted" samples, each made from two 1 L samples pooled together.
Wilderlab processed all 848 samples with our comprehensive freshwater metabarcoding panel, creating one of the most detailed national eDNA datasets for New Zealand rivers.
The analysis found no significant difference between the 1 L and 2 L samples, so filtering extra water did not improve results. Species accumulation curves showed that for both fish and macroinvertebrates, the point of diminishing returns came at five to seven replicates per site. Six replicates captured on average 89.5% of fish species and 80.4% of macroinvertebrate species found at a site, compared with 64.0% and 46.7% from a single sample. These results held across small and large streams, and at both high and low elevations.
Councils and agencies discussed the findings at an intercouncil meeting in August 2021, which led to the standardised six-replicate, 1 L syringe-filtered protocol for characterising river fish and macroinvertebrate communities. The work brought together researchers from Hawke's Bay, Waikato and Bay of Plenty regional councils, the Department of Conservation, the Ministry for the Environment, the University of Waikato and Wilderlab.
The dataset has also had a second life. Most trial sites (40 of 53) had five-year median macroinvertebrate community index (MCI) data. We used results from eight metabarcoding assays across 16 samples per site (848 samples) to train a taxonomy-independent community index (TICI). Rather than relying on species names, TICI scores river health from the DNA sequences themselves, so it can use organisms that have not yet been identified or added to reference databases. TICI scores closely matched historical MCI scores (R² = 0.82) and varied very little between replicate samples. The results are published in PeerJ.