There is mounting evidence that terrestrial arthropods are declining rapidly in many areas of the world. It is unclear whether freshwater invertebrates, which are key providers of ecosystem services, are also declining. We addressed this question by analysing a long-term dataset of macroinvertebrate abundance collected from 2002 to 2019 across 5,009 sampling sites in English rivers. Patterns varied markedly across taxonomic groups. Within trophic groups, we detected increases in the abundance of carnivores by 19% and herbivores by 14.8%, whilst we estimated decomposers have declined by 21.7% in abundance since 2002. We also found heterogeneity in trends across rivers belonging to different typologies based on geological dominance and catchment altitude, with organic lowland rivers having generally higher rates of increase in abundance across taxa and trophic groups, with siliceous lowland rivers having the most declines. Our results reveal a complex picture of change in freshwater macroinvertebrate abundance between taxonomic groups, trophic levels and river typologies. Our analysis helps with identifying priority regions for action on potential environmental stressors where we discover macroinvertebrate abundance declines.

Abundance data for riverine macroinvertebrates in England were extracted from the EA’s ecological monitoring database. The data were filtered to only include three-minute kick-sample data as the primary method for sampling freshwater invertebrates (approximately 99% of samples). Three-minute kick-samples are a standardised, internationally-recognised, semi-quantitative approach to assessing macroinvertebrate ecology and water quality in rivers using invertebrate diversity indicators. Prior to the implementation of the European Union Water Framework Directive in 2000, abundance estimates were based on categories (0–9, 10–99, 100–999 etc.). In 2002, the EA started recording more exact abundance estimates and enacting improved quality control procedures, whereby one in every ten samples was independently re-analysed. Hence, although the original dataset covered sampling years from 1991, our analysis was restricted to the years 2002–2019. Data were further filtered to only include sites sampled for a minimum of three years out of a total of eighteen in both spring (March to May) and autumn (September to November). After filtering the dataset according to these criteria, our final dataset from 2002–2019 included 67,757 individual macroinvertebrate samples from 5,009 sites (out of 10,136 sites in the original dataset). This equates to an average of 3,764 samples a year, covering 2,774 waterbodies distributed across the 10 river basins defined under the European Union Water Framework Directive in England: Anglian, Humber, North West, Northumbria, Severn, Solway Tweed, South East, South West and Thames (Fig. 1). The final dataset provides a wide national distribution of sites representative of the main river conditions, albeit with a bias towards mid to lower perennial reaches (reflective of the purpose of the monitoring programmes instigated for environmental quality monitoring, rather than a river’s intrinsic biodiversity). 

The identification of macroinvertebrates in the original database, including within individual samples, is given at a mixture of taxonomic levels. We pooled the data at two different levels: 1) wider taxonomic groups (non-insect freshwater macroinvertebrates: annelids, molluscs, Turbellaria and crustaceans, and individual insect orders: Ephemeroptera, Trichoptera, Plecoptera, Coleoptera, Diptera, Megaloptera, Hemiptera and Odonata); and 2) taxonomic families, representing observations for which this level of identification was available. We also pooled data considering main trophic groups (carnivores, herbivores and decomposers).