Harmful algal blooms (HABs) in large rivers are an emerging global concern, yet the mechanisms driving bloom initiation, composition and toxin production remain poorly understood compared to lakes. The Ohio River, USA has experienced several extreme blooms, with the largest in 2015 covering over 1100 river km and lasting over two months. Our group is developing a hydrodynamic-based river HABs prediction model that integrates river water quality, plankton ecology and physiology, and channel and watershed hydrology. This presentation highlights how nutrient, light, and temperature relationships were developed to parameterize the model using 40-day laboratory incubations of Ohio River plankton communities. Cyanobacteria dominated ending communities at extreme temperatures and light, with Pseudanabaena sp. dominant at low light and temperature, and Microcystis sp. at high light and temperature. As expected, nutrient concentrations regulated cyanobacteria biomass, but nitrogen (N) and phosphorus (P) ratios also regulated composition, as low N:P favored heterocyte containing Dolichospermum sp., and high N promoted non-heterocyte containing Microcystis sp. and Pseudanabaena sp. Unexpectedly, maximum cyanobacteria and total algal biomass were better predicted by nutrient conditions ~35 days prior to blooms than ambient total N and P at peak biomass, including treatments with high N-fixation rates. This laboratory result parallels the prolonged (> 30 day) low-flow conditions observed in the Ohio River prior to blooms with presumably limited external nutrient input, and introduces the concept of “nutrient history” for bloom forecasting models. These findings add mechanistic insight into large river bloom ecology, suggesting initial succession community composition and competition may interact with chemical and physical environmental controls to influence bloom composition and biomass end points.