Several lines of evidence from a eutrophic lake show how polymixis enables phosphorus (P) released from anoxic, iron (Fe)-poor sediments to lower nitrogen-to-phosphorus (N : P) ratios and stimulate cyanobacterial blooms. Detailed sediment analyses revealed extensive formation of Fe sulfides, which suppressed porewater Fe levels and prevented sequestration of P in Fe minerals. Experimental additions of Fe significantly decreased the flux of dissolved P from warm, anoxic sediments, increasing N : P ratios in porewater and overlying water. The net midsummer effect of polymixis and P release from Fe-poor sediments quickly doubled the total P in the euphotic zone during a period of very low external P loading. This internal "nutrient pump" decreased N : P in surface waters and led to a cyanobacterial bloom comprised primarily of diazotrophic Anabaena and Aphanizomenon spp. along with nonheterocystous and potentially toxic Microcystis icthyoblabe and Woronichinia naegelianum. Concentrations of the cyanotoxin, microcystin, in this lake were typically elevated during, or shortly after, episodes of internal P loading. Our study demonstrates an important mechanism underlying the increasing cyanobacterial dominance of weakly stratified eutrophic north temperate lakes, and warns of further increases under a warming climate.

The effects of reducing nutrient inputs to lakes and reservoirs are often delayed by hysteresis resulting from internal phosphorus (P) loading from sediments. Consequently, controlling harmful algal blooms (HABs) in many eutrophic ecosystems requires additional management to improve water quality. We manipulated iron (Fe) concentrations in a hypereutrophic lake to determine if Fe amendment would suppress HABs by inhibiting P release from sediments. Our experiment consisted of 15 in situ mesocosms, 12 of which each received a different dose of Fe (ranging from 2 to 225 g/m ); the remaining three were unmanipulated to serve as controls. Iron amendment decreased P accumulation in porewaters and the flux of P from sediments, which significantly lowered P concentrations in the water column. Iron exerted significant dose-dependent negative effects on the biomass of phytoplankton and periphyton, and reduced the dominance of cyanobacteria. Even at the lowest doses, Fe appeared to reduce the toxicity of cyanobacterial blooms, as measured by concentrations of hepatotoxic microcystins. Overall, our findings highlight the potential for Fe treatment as an effective strategy for minimizing HABs in eutrophic lakes and reservoirs. More broadly, our study reinforces the importance of Fe in regulating the trophic state of freshwaters, and the sensitivity of certain ecosystems to changes in Fe supply. Finally, we hypothesize that decreases in natural Fe supplies to lakes associated with anthropogenic activities may worsen outbreaks of toxic cyanobacteria.