The majority of Earth's land surface has been modified by human land use. The effects of changing land use on carbon budgets are relatively well known, but the effects on the energy budget are less well-explored. Vegetation affects the energy budget mainly through its control of albedo and transpiration. Usually, these effects are studied in terms of deforestation, where forested lands are replaced by crop or urban cover. However, long term forest development has led to forest regrowth and species composition shifts in much of the temperate northern hemisphere. In this dissertation, I use remote sensing data, sap flux data, and model output to investigate the effects of long-term, post-land-use-change forests development on the energy balance of the Upper Midwest. I show first that forest regrowth and composition shift occurred in more than half of the historically deforested land in the Upper Midwest, and that composition shift affected albedo and surface temperature differently than deforestation. Investigating vegetation control over transpiration with composition shift, I found that neither species nor stem size explained transpiration differences between historical and modern forests; instead, within-species differences in hydraulic conductance were most important. Finally, exploring the ecological and biophysical mechanisms of composition shifts in two ecosystem models, I find disagreement in leaf phenology and vegetation albedo relations that point to a need for more explicit treatment of composition shift in predictive models. Combined, I show that composition shifts have novel effects on energy balance that will be important to quantify in the future as forest regrowth and composition shift proceed.