Sediment DNA reveals how introduced trout reshaped a New Zealand lake foodweb

More than a century after brown trout were released into New Zealand waterways, scientists have reconstructed the ecological fallout using DNA locked deep in lake sediments. By sequencing genetic traces preserved in a sediment core from Lake Paringa on the West Coast of the South Island, researchers have pieced together a picture of a native fish community that collapsed after trout arrived, alongside shifts in algae and other organisms that ripple through the lake's foodweb. The challenge the researchers faced is a familiar one in ecology.

Understanding the impact of an introduced species requires knowing what the ecosystem looked like before it arrived, but most lake monitoring records are temporally limited and rarely extend back before introductions occurred. In Aotearoa-New Zealand, this problem is acute. From around 1860 CE onwards, acclimatisation societies shipped and released animals and plants from Britain during early European colonisation, transforming freshwater ecosystems before anyone thought to document them.

Brown trout (Salmo trutta) and rainbow trout (Oncorhynchus mykiss) were released into New Zealand rivers in the late 1800s with such success that brown trout are now the most widespread and often most abundant fish species in the country's waterbodies. With the exception of remote lakes in national parks, alpine lakes, or those lacking tributary connections, most New Zealand lakes now contain non-native fish, making contemporary comparisons between lakes with and without introductions nearly impossible. Lake Paringa offered a rare opportunity.

The lake is moderately sized, with a maximum depth of 58 metres and an area of 460 hectares, and sits in a catchment of temperate rainforest that has been largely undisturbed by human activity. It is oligotrophic, meaning nutrient-poor, and receives between 4,000 and 6,000 millimetres of rain per year, making the West Coast the wettest region in the country. The lake is culturally significant for Māori: seasonal settlements were established near the lake around 1300 CE, and it remains a noted tuna (eel) fishery and important spawning and breeding ground.

Crucially, while much of New Zealand was being deforested for pastoral agriculture during European settlement, the Lake Paringa catchment remained forested, minimising confounding land-use changes during the critical period when trout were likely introduced around 1870 CE. The team retrieved a six-metre-long sediment core from the deepest point of the lake's Windbag basin using a Mackereth corer. The sedimentology of Lake Paringa is well characterised, with layers influenced by the earthquake cycle of the Alpine Fault, but the researchers targeted layered silts deposited after 1800 CE that were uninfluenced by seismic deposition.

The chronology was anchored by correlating the new core with an earlier master core at sub-centimetre scale, using a radiocarbon-based age model built from 22 terrestrial macrofossils. Forty-two subsamples were taken for DNA analysis, with the top few millimetres of each layer scraped away to prevent contamination from the splitting process. The molecular approach targeted four genetic regions: the 18S ribosomal RNA gene to capture eukaryotes broadly, the mitochondrial 12S rRNA gene for fish, and two regions of the cytochrome c oxidase subunit 1 (COI) mitochondrial gene for eukaryotes, particularly macroinvertebrates.

Two of the COI primers were adapted specifically for this study to improve recovery of New Zealand zooplankton species, with in-silico testing against 99 freshwater zooplankton species from the BOLD database confirming the modifications added substantial base-pair matches. DNA was extracted from roughly three to five grams of sediment per sample, amplified by PCR, and sequenced on an Illumina MiSeq platform, with sequences processed into amplicon sequence variants using the DADA2 pipeline and taxonomy assigned against PR2, NCBI, and BOLD-derived databases. The fish results were striking.

Nine fish taxa were detected across the core using the 12S marker, seven identified to species level, six of them native: giant kōkopu (Galaxias argenteus), banded kōkopu (Galaxias fasciatus), common smelt (Retropinna retropinna), shortfin eel (Anguilla australis), longfin eel (Anguilla dieffenbachii), and upland bully (Gobiomorphus breviceps). Brown trout DNA first appeared in the sediment around 1906 CE (with a 95% probability range of 1892–1919) using the 18S marker, and around 1933 (1923–1943) using the 12S marker. Chinook salmon DNA appeared in a single sample dated to 1964.

Critically, both kōkopu species were detected only in the pre-trout phase of the core, declining just before trout DNA appeared and disappearing thereafter, a pattern consistent with predation and competition by salmonids that has been documented in streams and rivers across New Zealand, where only one of 30 known galaxiid species is currently classified as 'not threatened'. The broader eukaryotic community also shifted. The most significant assemblage change across the roughly 200-year record occurred between 1863 and 1876 CE, encompassing the probable timing of trout introduction, and was detected by all three eukaryotic markers.

The shifts involved different taxa depending on the gene: green algae including Mychonastes species drove changes in the COI datasets, while unclassified eukaryotes, chytrid fungi, and alveolates dominated the 18S signal. A generalised additive mixed model corroborated a significant period of change in the 18S dataset spanning 1810 to 1920 CE. In the algal community specifically, a small coccoid picoplankton species of Mychonastes surged to as much as 90% of the algal community around 1886, a pattern the researchers suggest could indicate a relaxation of grazing pressure at the zooplankton level following trout arrival, though this remains speculative.

Notably, the classic trophic cascade often associated with salmonid introductions did not clearly materialise. Algal pigment analysis by high performance liquid chromatography showed no significant difference in total pigment flux before and after trout arrival, with a t-test returning a p-value of 0.7. The researchers suggest this may reflect the oligotrophic nature of the lake, where low pre-existing algal abundance limited any cascade, or differences from other studies where algal shifts accompanied nutrient enrichment.

A further complication was the near-complete absence of zooplankton DNA across all markers, despite primers being specifically adapted to recover it. Because zooplankton are key components of trout diets and central players in trophic cascades, their absence means potential shifts in this vital foodweb layer may have been missed entirely, highlighting a methodological gap that the authors say requires further work on DNA preservation and recovery from lake sediments. The study also exposes the practical challenges of sedimentary DNA work.

Recovery of fish DNA was patchy, with native eels that certainly inhabited the lake appearing in only five samples collectively, and many COI-recovered taxa found only sporadically, meaning community shifts were sometimes driven by rare or inconsistently recovered species. The overwhelming majority of sequence variants could not be classified below order level, reflecting the limited representation of New Zealand species in international sequence databases. The data come from a single core in a single lake, limiting extrapolation, and the authors recommend multiple cores, greater replication, targeted quantitative assays such as qPCR for dating introductions, and sediment-spiking experiments to validate recovery.

They also note that shifts after 1950 may reflect anthropogenic pressures, particularly the construction of the Haast Pass Highway in 1965, rather than trout effects. Despite these caveats, the findings demonstrate that sedimentary environmental DNA can recover signals from organisms invisible to traditional paleolimnological techniques, including fungi and soft-bodied taxa that leave no preservable fossils. The evidence points to substantial changes across the Lake Paringa foodweb between roughly 1856 and 1891 CE, coinciding with the probable arrival of brown trout, whose DNA is first detectable in 1906.

The native galaxiid-dominated fish community gave way to a trout-dominated one, algae restructured, and the wider eukaryotic community shifted, all without a corresponding change in total algal productivity. While the precise ecological mechanisms, particularly top-down predation pressure, could not be confidently confirmed, the study shows that introduced trout can substantially affect lake ecology even in relatively pristine catchments, and that the mud beneath such lakes preserves a record of those transformations waiting to be read.

Subject of Research: Ecological impacts of brown trout introduction in an oligotrophic New Zealand lake reconstructed using sedimentary environmental DNA

Article Title: Insights into the ecological impact of trout introduction in an oligotrophic lake using sedimentary environmental DNA

Article References: Schallenberg, L. A., Thomson-Laing, G., Kelly, D., Pearman, J. K., Howarth, J. D., Vandergoes, M. J., Puddick, J., Fitzsimons, S., Rees, A., & Wood, S. A. (2023). Insights into the ecological impact of trout introduction in an oligotrophic lake using sedimentary environmental DNA. Metabarcoding and Metagenomics, 7, Article e111467. https://doi.org/10.3897/mbmg.7.111467

Image Credits: AI Generated

DOI: 10.3897/mbmg.7.111467

Keywords: sedimentary environmental DNA, brown trout, invasive species, Lake Paringa, New Zealand, paleolimnology, galaxiids, trophic cascades, lake foodweb, metabarcoding, algal pigments, conservation