15Despite growing evidence that climate change will increase temperature variability and 16 the frequency of temperature extremes, many modeling studies that analyze the effects of 17 warming scenarios on cyanobacteria in lakes examine uniform warming temperature scenarios 18 without including any variability. Here, we used the one-dimensional hydrodynamic General 19 Lake Model coupled to Aquatic EcoDynamics modules (GLM-AED) to simulate 11 years of 20 nitrogen-fixing and non-nitrogen-fixing cyanobacterial biomass in Lake Mendota (Madison, WI, 21 USA). We developed climate scenarios with either uniform (constant) warming or variable 22 warming based on random sampling of daily air temperatures from either a normal or Poisson 23 33 Keywords: climate change; blooms; phytoplankton; GLM; Mendota; probability distributions 34 35 2008; Visser et al., 2016). Cyanobacteria have higher thermal optima than eukaryotic 70 phytoplankton (up to 35°C for most taxa; Reynolds 2006), so higher temperature variability 71could increase the incidence of very warm days on which the optimum temperature for 72 cyanobacteria occurs in the lake (Konopka and Brock, 1978;Visser et al., 2016; Walls et al., 73 2018;Walter Helbling et al., 2015). However, like most lake modeling studies, most work to 74 date studying the effects of warming on phytoplankton has focused on changes to mean 75 temperatures, not temperature variability (e.g., De Stasio et al., 1996; Mooij et al., 2005; Striebel 76 et al., 2016;Thomas and Litchman, 2015), despite evidence that variability in water temperatures 77 can alter phytoplankton (Edlund et al., 2017;Havens et al., 2016;Sahoo et al., 2016; Saros et al., 78 2016; Merchant, 2017, 2018). 79Here, we used a lake ecosystem simulation model to answer the question, How does 80 increased mean temperature and daily temperature variability affect the biomass of lake 81 cyanobacteria? We calibrated a one-dimensional hydrodynamic water quality model to simulate 82 two phytoplankton functional groups that represent the physiological and ecological traits of 83 nitrogen-fixing (N-fixing) and non-nitrogen-fixing (non-N-fixing) cyanobacteria, and quantified 84 the effects of warming and more variable air temperatures on phytoplankton community 85 dynamics. We developed scenarios of variable air temperature forcing by randomly sampling 86 from different distributions of air temperature increases on each day of the simulation. We 87 expected that including air temperature variability would result in an increased incidence of 88 cyanobacterial blooms, because especially warm days might trigger cyanobacterial blooms. We 89 also expected that the results would vary substantially among the multiple years in a simulation 90 due to background variability in other meteorological drivers (e.g., precipitation), which were not 91 modified in the air temperature scenarios. 92Potential effects of global climate change on small north-temperate lakes: Physics, fish, and 595