• acidification;
  • acid rain;
  • insects;
  • invertebrates;
  • Llyn Brianne;
  • restoration;
  • recovery;
  • streams


  • 1
    Streams affected by acid deposition should now be recovering biologically, but long-term assessments are scarce. Here, we use the experimental catchments at Llyn Brianne (Wales, UK) to evaluate trends over 25 years (1981–2005) in the chemistry and macroinvertebrates of acid moorland and forest streams restored by liming relative to those responding ‘naturally’ to reduced deposition.
  • 2
    Mean H+ concentrations in acid moorland streams fell by c. 15–16 µquiv L−1 over the study, increasing mean winter pH by 0·8–1·3 units to pH 5·6–6·1. Liming moorland streams in 1987 and 1988 increased mean pH to 5·5–6·4, but differences from naturally recovering streams diminished over 12–18 years.
  • 3
    In limed and acid moorland streams, changes in invertebrate composition were consistent with recovery, and near-identical. Four acid-sensitive species, from a local pool of 29, increased significantly in abundance or occurrence, but effects were too small to increase similarity with circumneutral reference streams.
  • 4
    Mean H+ in acid forest streams declined by 8–15 µequiv L−1, but mean winter pH increased by only 0·4 units and remained too acid for invertebrate recovery (mean pH 4·8–5·2; Al > 0·3–0·6 mg L−1). One forest stream limed in 1987 and 1988 remained at mean pH > 5·9 thereafter, but there was no invertebrate response.
  • 5
    Climate affected recovery pattern. After accounting for time trends, wet winters increased acidity in moorland and forest streams sufficiently to offset 21–41% of the total 25-year decrease in H+ concentration.
  • 6
    Synthesis and applications. These data from one of the world's longest running experiments on acidification confirm that upland British streams are recovering, but ecological effects are marginal and vary with land use. Conifer forest streams at Llyn Brianne remain too acid for sensitive invertebrates, while moorland streams are still at risk from acid events. In this example, liming had few long-term benefits compared with natural recovery, and we suggest that this should be a key, general criterion in evaluating the outcomes of ecological restoration.