Thirty years of hourly dissolved oxygen in a large shallow river illustrates discrepancy between a primary producer tipping point and river metabolism

Large-scale efforts to reduce cultural eutrophication of freshwater systems have had varied success because internal feedbacks can stabilize the high nutrient, high productivity, and turbid conditions associated with eutrophic systems. We examined these feedbacks using a unique 40-year water quality data set from the middle Loire River, France, where phosphorus and phytoplankton concentrations have decreased by an order of magnitude from 1980–2018. We focused on ecosystem metabolism as an integrative measure to elucidate cause-effect relationships of both bottom-up (e.g., nutrient concentrations) and top-down (e.g., consumer populations) effects on river trophic state.

The dataset combined both long-term (30 years), high-frequency (hourly) measurements of dissolved oxygen (DO) and long-term (40 years), low-frequency (monthly) measures of nutrients, plus several supporting biological surveys of primary producer and consumer densities. Using hourly measurements of DO, we estimated gross primary production (GPP), ecosystem respiration (ER), and net ecosystem production (NEP = GPP – ER), and from the resulting long time series of metabolic fluxes, we tested the hypothesis that GPP and ER responded to changes in water column concentrations of algal pigments (chlorophyll a) and phosphorus. We further tested the hypothesis that change points in the patterns of ecological behavior were contemporaneous with notable changes in river management.

Despite well-established links between phosphorus, chlorophyll-a and primary production, GPP was resilient to the drastic reductions in both P concentrations and phytoplankton. Indeed, GPP has only recently decreased (~25%), despite chlorophyll-a concentrations reaching a new minima 10 years earlier in response to colonization of the invasive Corbicula sp. clam in the year 2000. Declines in ER are only half (~12%) the decline in GPP, shifting the river from an autotrophic state (i.e., positive NEP) to a heterotrophic state (i.e., negative NEP). Moreover, Granger causality analysis suggested that daily primary production and respiration have decoupled over this period. With earlier phytoplankton dominance, daily ER was strongly linked to recent autochthonous GPP, but more recently daily GPP has far less influence on subsequent ER. We interpret this partially as a reduction in carbon and nutrient turnover rates resulting from the community shift from algae to macrophytes, and attendant changes in nutrient sources (now primarily from sediment) and carbon stocks (now principally in the sediment). This study illustrates the benefit of long-term high-frequency data collection for understanding pattern and process in aquatic ecosystems, and illustrates a compelling example of process resilience contrasted with an ecosystem tipping point in the context of global change.