Cyanobacterial survival following their release in water from major headwaters reservoirs was compared in five New South Wales rivers. Under low flow conditions, cyanobacterial presence disappeared rapidly with distance downstream in the Cudgegong and Hunter Rivers, whereas the other three rivers were contaminated for at least 300 km. Cyanobacterial survival is likely to be impacted by the geomorphology of each river, especially the extent of gravel riffle reaches (cells striking rocks can destroy them) and by the different turbulent flow conditions it produces within each. Flow conditions at gauging stations were used to estimate the turbulent strain rate experienced by suspended cyanobacteria. These indicate average turbulent strain rates in the Cudgegong and Hunter Rivers can be above 33 and 83 s-1 while for the Murray, Edward and Macquarie Rivers average strain rate was estimated to be less than 30 s-1. These turbulent strain rate estimates are substantially above published thresholds of approximately 2 s-1 for impacts indicated from laboratory tests. Estimates of strain rate were correlated with changes in cyanobacterial biovolume at stations along the rivers. These measurements indicate a weak but significant negative linear relationship between average strain rate and change in cyanobacterial biomass. River management often involves releasing cold deep water with low cyanobacterial presence from these reservoirs, leading to ecological impacts from cold water pollution downstream. The pollution may be avoided if cyanobacteria die off rapidly downstream of the reservoir, allowing surface water to be released instead. However high concentrations of soluble cyanotoxins may remain even after the cyanobacterial cells have been destroyed. The geomorphology of the river (length of riffle reaches) is an important consideration for river management during cyanobacterial blooms in headwater reservoirs.