Discussion

This study demonstrates that sugar‐flotation combined with metabarcoding is an effective approach for recovering aquatic invertebrate biodiversity data from dry lake sediments in ephemerally inundated environments and records more taxa than conventional methods (rehydration trials). We also showed that increasing the initial mass of sediment for the sugar‐flotation treatment significantly increases the number of aquatic macroinvertebrate species recovered, although there are diminishing returns at higher sediment weights.

Benefits of Combining Sugar‐Flotation and DNA Metabarcoding

In Experiment 1, the sugar‐flotation method consistently detected aquatic invertebrate taxa, with nearly all sugar‐floated samples detecting at least one target taxon, whereas only a single standard‐extraction sample resulted in a positive detection. The results of the standard extraction method are consistent with previous studies which have found that unprocessed sediment samples are an inefficient method of detecting macroinvertebrates (Pawlowski et al. 2022). Two features of the sugar‐flotation workflow likely contribute to this difference. The first is the composition of the material extracted. Although DNA in sediments can be both extracellular and intracellular (Ellegaard et al. 2020), the significant difference in the performance of the sugar‐floated versus standard samples is likely in part due to the differential efficiency in recovering DNA from macroinvertebrate resting stages (although other sources of DNA cannot be excluded, see below). Elutriation increases the proportion of aquatic macroinvertebrate DNA recovered from a sample and reduces interference from non‐target DNA sources in the sediment matrix (Macher et al. 2018), which can otherwise dominate the DNA pool and obscure detection of rare taxa. The second is the mass of sediment each workflow can process, which is a property of the methods themselves rather than an incidental difference between treatments. The column‐based kit used here accepts a maximum of 250 mg of sediment, and even large input formats designed for high biomass samples accept no more than approximately 10–15 g (e.g., DNeasy PowerMax Soil Kit, Qiagen) while the sugar‐flotation is less restricted by larger sediment fractions (see below).

Effect of Sediment Weight on Taxa Detection

As expected, species richness increased with the amount of sediment treated with sugar‐floatation. This increase was most notable from 50 to 250 g with no significant difference in species richness among the 250, 500 and 1000 g weight classes, although the 1000 g samples did detect the most taxa overall. These findings are consistent with previous studies using metabarcoding on sediment samples from non‐lake environments targeting metazoans and non‐metazoan eukaryotes (Nascimento et al. 2018), bacteria (Ellingsøe and Johnsen 2002), fungi (Penton et al. 2016) and soil macroinvertebrates (Kirse et al. 2021), which found that increasing the amount of source material for DNA extraction also provides a more comprehensive assessment of the biodiversity present. However, increasing sediment weight within a site was not correlated with a significant reduction in within‐treatment variation, though some sites did show a trend toward increasing the similarity of the communities recovered by replicates at greater sediment weight. This was contrary to our expectations that within a site larger sediment treatments would have lower variation in the community composition recovered between replicates because processing more sediment would capture more of the inherent spatial variability of resting eggs within sediments (Rahman, Chaplin, Lawrie, and Pinder 2025). This may reflect insufficient biological replication to detect this effect, suggesting that increased sediment weight alone is unlikely to compensate for limited sampling effort when characterising spatially heterogeneous egg banks. Pooling separate DNA extracts from a site prior to PCR has been shown to reduce within‐site heterogeneity in marine sediments (Hestetun et al. 2021) and could be potentially used to improve the consistency of the results generated here. Together our results suggest that metabarcoding biodiversity surveys of aquatic macroinvertebrates in dry lake sediments should sugar‐float 1000 g of sediment where logistically feasible. The 1000 g treatment recovered both the highest mean species richness and the greatest total number of taxa across sites, and although it could not be statistically separated from the 250 and 500 g treatments, there is no indication that returns had ceased by 1000 g. Where processing capacity is limited, 250 g represents a defensible minimum, since this was the mass above which mean richness no longer increased significantly and below which detection declined markedly. Additional biological replicates should be prioritised alongside sediment mass, given that within‐site variation was not detectably reduced by increasing sediment weight alone.

Metabarcoding vs. Rehydration Trials

Biodiversity assessments conducted using a combination of approaches usually provide a more comprehensive depiction of biotic communities by capitalising on the complementary strengths of different methods (Schenekar 2023). A combination of different sampling approaches is considered best practice for environmental monitoring in salt lakes (e.g., Campbell et al. 2023; Saccò et al. 2025) and freshwater ecosystems more broadly (Schenekar 2023). While ultimately, we suggest that conducting both rehydration trials and DNA metabarcoding on dry lake sediments produces highly complementary information, in practice environmental managers typically face resource constraints that force pragmatic trade‐offs to select method(s) that best align with their specific objectives. The evaluation of the performance and assumptions of DNA metabarcoding and rehydration trials below has been developed within this context.

Baseline surveys of biodiversity are a key component in conservation planning that provide foundational ecological context and identify the presence of rare or threatened taxa. Consistent with previous comparisons between DNA metabarcoding and conventional biodiversity survey methods (Iacaruso et al. 2025), DNA metabarcoding detected a greater overall number of taxa, including seven unique taxa, whereas only a single taxon was exclusively recovered through rehydration trials. These findings are not surprising, given that DNA metabarcoding approaches generally have a superior ability to detect rare species and identify taxa to lower taxonomic levels compared with morphological analyses (Andersson et al. 2023; Bush et al. 2019), especially in cryptically diverse macroinvertebrate groups (Pfrender et al. 2010). The disparity between the taxa recovered from the same sediment samples between both methods may indicate that the artificial conditions used in the rehydration trials do not meet the hatching requirements of some taxa. For example, Branchinella was not recorded in the rehydration trials despite producing resting eggs (Timms and Lindsay 2011), being consistently detected by DNA metabarcoding, and being present in sweep net samples collected from Lake Way during inundation (Bennelongia Environmental Consultants 2020). While the absence of viable eggs cannot be ruled out as an explanation for the non‐hatching of Branchinella and other taxa (see below), their detections demonstrate the utility of DNA metabarcoding in establishing baseline biodiversity information for dry lakes given our currently poor understanding of the ecological requirements of many salt lake invertebrates (Lawrie et al. 2021; Lawrie, Chaplin, Rahman, et al. 2023).

It is important to note that neither rehydration trials nor metabarcoding recorded all taxa known to occur in Lake Way during inundation (Bennelongia Environmental Consultants 2020), highlighting the importance of opportunistic sampling of salt lakes during flooding conditions. Missing taxa were mostly transient insects that do not deposit resting stages in dry lakes but re‐colonise during filling events and therefore will not hatch in rehydration trials and are unlikely to be detected in the metabarcoding of lake sediments. However, calanoid, harpacticoid, and cyclopoid copepods were detected either inconsistently or not at all across the metabarcoding data, despite being abundant under flooding conditions (Biologic unpublished data) and cladocerans were inconsistently detected but common in the rehydration trials. Although a local DNA reference library was generated from previous aquatic macroinvertebrate collections at Lake Way, which included representative harpacticoid, calanoid, cyclopoid, and cladoceran species collected from Lake Way, it is possible that their non‐detection may reflect gaps in the DNA reference library used here. In addition, we used multiple DNA assays as is typically recommended in metabarcoding studies to overcome primer biases and to compensate for phylogenetically diverse environments (Corse et al. 2019; Hajibabaei et al. 2019; Zhang et al. 2018). These assays did collectively increase the taxonomic coverage of this study; however, the inconsistent detection of copepods and cladocerans suggests that the metabarcoding results could have been improved with the inclusion of additional primer sets (see Bylemans et al. 2025).

As metabarcoding data cannot differentiate between DNA from living versus dead organisms (Pawlowski et al. 2022), a detection of a species in the metabarcoding data cannot necessarily be used as evidence of viable resting stages. For example, Tanytarsus barbitarsis was detected via metabarcoding in multiple sites and occurs in Lake Way under flooding conditions (Biologic unpublished data) but was not present in the rehydration trials. Although T. barbitarsis does deposit eggs with some desiccation resistant capabilities, experimental results of Kokkinn and Williams (1988) suggest that egg viability is limited to 5 days at 20°C. However, T. barbitarsis can generate very high population densities within salt lakes (e.g., 140,000 individuals/m2; Paterson and Walker 1974) suggesting that the metabarcoding detection of T. barbitarsis was most likely from residual DNA deposited in the sediment during previous flooding events. Therefore, rehydration trials remain a necessary component in the monitoring programmes of ephemeral aquatic environments where the viability of resting egg banks needs to be established.