A meta-analysis of the effects of nutrient enrichment on litter decomposition in streams

The trophic state of many streams is likely to deteriorate in the future due to the continuing increase in human-induced nutrient availability. Therefore, it is of fundamental importance to understand how nutrient enrichment affects plant litter decomposition, a key ecosystem-level process in forest streams. Here, we present a meta-analysis of 99 studies published between 1970 and 2012 that reported the effects of nutrient enrichment on litter decomposition in running waters. When considering the entire database, which consisted of 840 case studies, nutrient enrichment stimulated litter decomposition rate by approximately 50%. The stimulation was higher when the background nutrient concentrations were low and the magnitude of the nutrient enrichment was high, suggesting that oligotrophic streams are most vulnerable to nutrient enrichment. The magnitude of the nutrient-enrichment effect on litter decomposition was higher in the laboratory than in the field experiments, suggesting that laboratory experiments overestimate the effect and their results should be interpreted with caution. Among field experiments, effects of nutrient enrichment were smaller in the correlative than in the manipulative experiments since in the former the effects of nutrient enrichment on litter decomposition were likely confounded by other environmental factors, e.g. pollutants other than nutrients commonly found in streams impacted by human activity. However, primary studies addressing the effect of multiple stressors on litter decomposition are still few and thus it was not possible to consider the interaction between factors in this review. In field manipulative experiments, the effect of nutrient enrichment on litter decomposition depended on the scale at which the nutrients were added: stream reach > streamside channel > litter bag. This may have resulted from a more uniform and continuous exposure of microbes and detritivores to nutrient enrichment at the stream-reach scale. By contrast, nutrient enrichment at the litter-bag scale, often by using diffusing substrates, does not provide uniform controllable nutrient release at either temporal or spatial scales, suggesting that this approach should be abandoned. In field manipulative experiments, the addition of both nitrogen (N) and phosphorus (P) resulted in stronger stimulation of litter decomposition than the addition of N or P alone, suggesting that there might be nutrient co-limitation of decomposition in streams. The magnitude of the nutrient-enrichment effect on litter decomposition was higher for wood than for leaves, and for low-quality than for high-quality leaves. The effect of nutrient enrichment on litter decomposition may also depend on climate. The tendency for larger effect size in colder regions suggests that patterns of biogeography of invertebrate decomposers may be modulating the effect of nutrient enrichment on litter decomposition. Although studies in temperate environments were overrepresented in our database, our meta-analysis suggests that the effect of nutrient enrichment might be strongest in cold oligotrophic streams that depend on low-quality plant litter inputs.