Controls on productivity of submerged aquatic vegetation in 2 spring-fed rivers

Submerged aquatic vegetation (SAV) is an important structural component of many aquatic ecosystems that also affects a broad array of biogeochemical and ecological processes. In the spring-fed rivers of Florida, where SAV is the foundation of healthy community structure and function, environmental controls on SAV growth remain poorly understood. This lack of knowledge complicates the evaluation and prioritization of management and restoration activities. Our goal was to explore the relative influences of environmental drivers on annual patterns of SAV productivity in 2 spring-fed rivers (Silver River and Alexander Springs Creek) with contrasting nitrate concentrations (1.4 vs 0.05 mg N/L, respectively) but otherwise similar physical and chemical attributes. We used monthly measurements from 16 fixed plots in each river to observe spatial and temporal heterogeneity in SAV growth. Mean growth rates between the rivers were statistically indistinguishable, suggesting that SAV growth is not controlled primarily by water column N concentrations. Despite representing most of the autotroph biomass, measured SAV growth rates were only 25% of contemporaneous estimates of open-channel net primary production (i.e., gross primary production minus autotrophic respiration). This result indicates that the epiphytic assemblages that grow primarily on SAV dominate overall ecosystem metabolic activity. Pairwise associations between measured SAV growth and a suite of physical (light, redox potential, temperature, sediment texture) and chemical (porewater and water column solutes) variables revealed that no single factor can accurately predict SAV growth. Instead, 2 complementary multivariate approaches (regression trees, general linear models) yielded useful models to predict SAV growth. While we constructed models independently for each river, their best-fit model formulations were remarkably similar-light (+), sediment redox potential (-), and porewater orthophosphate (-) predicted over 50% of the variation in SAV growth. System managers could use these model results to aid in screening restoration sites for SAV suitability. Further, these results underscore the need for a broad systems approach to springs management.