Including variability in air temperature warming scenarios in a lake simulation model highlights uncertainty in predictions of cyanobacteria

Abstract

AbstractDespite growing evidence that climate change will increase temperature variability and the frequency of temperature extremes, many modeling studies that analyze the effects of warming scenarios on cyanobacteria in lakes examine uniform warming temperature scenarios without including any variability. Here, we used the one-dimensional hydrodynamic General Lake Model coupled to Aquatic EcoDynamics modules (GLM-AED) to simulate 11 years of nitrogen-fixing and non-nitrogen-fixing cyanobacterial biomass in Lake Mendota (Madison, WI, USA). We developed climate scenarios with either uniform (constant) warming or variable warming based on random sampling of daily air temperatures from either a normal or Poisson probability distribution. We found that while the median cyanobacterial biomass among repeated simulations for each of the years was similar regardless of whether or not air temperature variability was included in the climate scenarios, the randomly-sampled air temperature distribution scenarios exhibited much greater variability in the year-to-year cyanobacterial biomass estimates. Our results suggest that including temperature variability in climate scenarios may substantially change our understanding of how climate warming may alter cyanobacterial blooms over yearly to decadal time scales. To more effectively predict the range of possible future cyanobacterial dynamics, both the magnitude and variability of warming must be considered when developing climate scenarios for lake modeling studies.