Influence of experimental, environmental, and geographic factors on nutrient-diffusing substrate experiments in running waters

Abstract

Freshwater algal growth is often limited by the availability of nitrogen (N), phosphorus (P), or both nutrients (NP). For over 30 years, investigators have conducted nutrient-diffusing substrate (NDS) experiments to quantify algal nutrient limitation or co-limitation in rivers and streams. Previous meta-analyses of NDS have shown that algae are commonly co-limited by N and P and that water column nutrients are weakly predictive of limitation. These analyses have not, however, comprehensively addressed the experimental, environmental, and geographic covariates affecting nutrient limitation results. We surveyed the literature and extracted data for algal biomass effect sizes and a suite of covariates across a total of 649 experiments. We built meta-regression models to identify important controls on NDS results and to gain insights about algal nutrient limitation patterns over space and time. We also reviewed potential mechanisms for the reported result that NDS N and P treatments can inhibit algal growth. Experimental variables including substrate type, chemical concentration, and experimental length significantly affected P and NP effect sizes, while NDS chemical compound influenced N, P, and NP effect sizes. We also found that environmental variables such as in-stream nutrients and riparian canopy cover significantly affected limitation by N, P, and NP. Temperature, stream discharge, and stream velocity only affected limitation by NP. Land use, ecoregion, and season showed clear trends in nutrient limitation for all treatments that could generally be tied to environmental factors like in-stream nutrients and riparian canopy cover. Most experimental and environmental variables that were statistically significant in the meta-regression models produced very low R2 index values, indicating that the models explained little variation in among-site effect sizes. Spatial factors including stream order, ecoregion, and climate classification had the highest R2 index values, but these models still produced a large amount of unexplained variance. In light of these findings, we provide recommendations for improving NDS experimental design and pursuing future research avenues using NDS. We also highlight the need for future experiments to consider algal stressors that may interact with nutrient limitation experiments.