We use the Atmospheric Model Intercomparison Project (AMIP-II) experimental protocol to perform six ensemble simulations for four different modes of land surface energy balance complexity. We calculate the land-only mean, minimum and maximum temperature variance for each ensemble and compare these with AMIP-II results. We find no evidence that the surface energy balance complexity leads to systematic differences in the simulated mean, minimum or maximum temperature variance at the global scale, or in the zonal averages. We therefore conclude that the differences simulated by the AMIP-II models cannot be attributed to variations in surface energy balance complexity. Our results indicate that the simulation of mean temperature variance and the variance of extreme temperature are not limited by uncertainties in how to parameterize the surface energy balance, a conclusion that adds confidence to the use of climate models for detection and attribution studies.