Broad-scale evaluations of biodiversity change need reference comparisons

Assessments of broad-scale biodiversity change commonly aggregate communities that differ in their initial impacts and their total degree of change (differing "impact trajectories"). Biodiversity responses may depend on the impact trajectory of a given community; thus, not controlling for these differences could produce a misleading mixture of trends even if all communities degraded or recovered. To evaluate this dependency, we related biodiversity trends for 4062 river invertebrate time series (1971 to 2024) from 23 European countries to impact trajectories for each community, quantified through comparison to least-impacted references. Trends were typically stronger and more consistent in initially more impacted communities that also exhibited more extreme total change. However, responses were weaker when initial impacts were lower and highly inconsistent with lesser total change. Our results show that biodiversity trends should be compared among communities with similar impacts, determined relative to appropriate references. Improving trend comparability is essential to accurately determine how biodiversity generally responds to anthropogenic impacts and conservation actions.