European freshwater macroinvertebrate richness and abundance: alternative analyses and new findings

Studies at local to national extent have documented a recovery in macroinvertebrate taxonomic richness following improvements in water quality. The study by Haase et al . (2023) published in Nature claimed that the overall recovery came to a halt across Europe by 2010. However, the lack of monitoring design, the varying lengths in time series and heterogeneous taxonomic resolution (species, genus and families), along with insuficient information on data handling prior to statistical analyses are raising questions about the reliability of the findings.

Here I use the open access raw data of the original study to calculate the proportion of taxa identified to the targeted taxonomic resolution (species, genus or family), which revealed a lack of taxonomic consistency within some of the time-series. I then devised a simple taxonomic correction to remove potential biases in the richness trend estimates through the modelling procedures using linear models.

In order to make the data more comparable across studies and over time, I calculated an anomaly in taxonomic richness relative to a five-year reference period within 1990-2020, so all time series (≥15 years long, ≥8 samples) overlapped. The concept is borrowed from the familiar temperature anomaly in climate research to track deviations from a norm. I ran non-linear trend analyses to reveal changes in the anomaly in taxonomic richness during the period 1990-2020.

European taxonomic richness using 1816 sites in 47 studies (full dataset) increased linearly by about 0.29±0.09 taxa per year when using all taxonomic ranks (species, genus, family), compared to the average 0.20 taxa per year in the original study, but dropped to 0.15±0.04 taxa per year at family level. The same results were produced after geographical thinning to 687 sites separated by at least 20 km from each other’s. Further data analyses revealed the extent of discrepancies in taxonomic resolution (proportion of taxa identified to species or genus level) within time-series and its impact on trend estimates.

The linear increase in abundance over time was marginal (1 individual / year or 0.12% of average abundance) in the full dataset and not significant within 1990-2020 period, contrary to published findings (1.17%) due to a calculation error in the original study.

The linear analyses of species richness were run on centred years and did not allow the study of the temporal dynamics in taxonomic richness. Non-linear analyses using the anomaly in taxonomic richness for the period 1990-2020 revealed no change in taxonomic richness apart from a post millennium small and short rise using all taxonomic ranks (1120 sites, 27 studies), possibly due to a concurrent increase in sampling efort (abundance) across sites.

Coarsening the taxonomic resolution to family level did not alter the dynamic of the anomaly in taxonomic richness over time, possibly a result from poor sample sampling efort. The average ‘species’ richness (762 sites) was about 30 taxa per sample, barely higher than family richness (20 taxa per sample) and very small compared to studies with more intensive sampling eforts. Independently of the efect of anthropogenic impacts, I question the adequacy of the current biomonitoring design and sample sampling efort to study river macroinvertebrate biodiversity.

Implications of new findings . Linear trend estimates in taxonomic richness (independently of the time period) were dependent on taxonomic resolution, higher at ‘species’ than family level. Neither the abundance nor the anomaly in taxonomic richness showed signs of recovery during the period 1990-2020. Current sampling eforts for rapid bioindicators, such as those developed for the European Water Framework Directive, are inadequate to address the needs of the EU 2030 Biodiversity Strategy. Macroinvertebrates would be right to demand more from us.