OOS 24-3 - Anticipating the timing and magnitude of cyanotoxin production by benthic cyanobacteria in rivers

Blooms of toxin-producing benthic (i.e. bottom-dwelling) cyanobacteria in rivers are increasingly detected across the globe, yet our ability to predict the timing of their occurrence and the magnitude of cyanotoxin production is currently limited. The growth of benthic algae in rivers is a dominant process driving temporal variation in whole river ecosystem productivity, and river ecosystem productivity is under similar environmental controls (e.g., flow, light, temperature) as benthic cyanobacterial mats. Therefore, whole river ecosystem productivity, which can be estimated from in-situ high frequency dissolved oxygen (DO) and temperature sensor data, may be synchronous with the timing of benthic cyanobacterial blooms. This potential synchrony presents an opportunity to use ecological forecasting to inform targeted sampling for benthic cyanotoxins that pose a risk to public and ecological health. Here, we address the question, to what degree do overall productivity dynamics in rivers predict the occurrence and dynamics of toxin-producing cyanobacteria? From June to September 2022, we monitored three river reaches with previously reported benthic cyanobacterial blooms and toxin production in the South Fork Eel, Russian, and Salmon rivers of northern California, USA. We estimated daily metabolism using measurements of discharge, temperature, and DO from high-frequency sensors. We collected biweekly samples of surface water for nutrient concentrations and benthic Microcoleus and Anabaena mats for anatoxin-a (ATX) concentrations in the reaches upstream of the sensors. Peak ATX concentrations in either benthic Microcoleus and Anabaena mats lagged behind peaks in river ecosystem productivity. In the South Fork Eel River, declines in productivity from a summer peak of ~4 mg O2 m-2 d-1 began in late July, while ATX concentrations, primarily in Microcoleus mats, peaked in mid-August (range: 64 - 11,219 μg ATX/L of mat material). In the Russian River, productivity was relatively constant at ~5 mg O2 m-2 d-1 until declining in late August, which preceded an increase in ATX concentrations in Anabaena mats (range: 27 - 49 μg ATX/L of mat material). Productivity in the Salmon River did not peak until September, but wildfire interference limited our ability to capture a complete time series of ATX concentrations. These lagged patterns suggest that toxin production may be predictable from patterns in river ecosystem productivity, and we further explore how greater mechanistic insight into environmental drivers and within-mat community dynamics may improve the potential for forecasts.