Aim
Evaluating the effectiveness of marine reserves in achieving conservation goals is challenged by the decadal scales over which biological systems respond following protection. Given that trophic interactions underpin community responses following protection and that complex ecological interactions make responses difficult to identify, quantifying changes in species traits may provide detail missed by traditional diversity measures, including information relevant to ecosystem functioning. We determine whether this is the case by comparing community metrics based on functional traits to taxonomic diversity measures associated with ‘no take’ marine reserves and partially protected, fished areas along eight degrees of latitude.

Location
Eighteen ‘no take’ marine reserves and 14 partially protected, fished areas along the east coast of Australia.

Method
We use two independent datasets from shallow and deep coastal rocky reefs to analyse trait‐based metrics and taxonomic diversity from sites inside reserves to sites in partially protected, fished areas.

Results
Taxonomic diversity (species diversity and richness) and trait‐based multimetrics (functional richness and dispersion) showed no difference with level of protection. Total fish abundance responded positively to protection, but only on shallow reefs. Comparing values of individual functional traits implied a return of larger bodied species of fish in protected areas and an increase in trophic level. The latter was significant on deeper reefs and was strongly correlated with age of protected area. Thus, recovery responses were largely associated with community mean functional trait values, highlighting the value of trait‐based approaches for detecting change, when no differences in traditional taxonomic diversity measures were apparent.

Main conclusions
We empirically demonstrate that functional traits can elucidate early conservation outcomes, when traditional multimetric diversity indices do not distinguish protected and fished communities. Ecologically relevant but sensitive metrics are fundamental to allow information to be incorporated into adaptive management strategies, which often occur on political rather than biological time‐scales.