Extensive wildfires in the Pacific Northwest in the summer of 2017 unleashed a vast plume of smoke that ascended high into the stratosphere, persisted for more than eight months and provided researchers a rare opportunity to evaluate current models of smoke ascent. Their study reveals gaps in the way smoke plume rise and duration is modeled now, they say. Powerful firestorms will occasionally cause pyrocumulonibus clouds (pyroCbs) to erupt violently into the atmosphere. The rapidly rising, super-heated air of the fires below results in a towering, smoke-infused, thunderstorm-like cloud, which - like a chimney - funnels smoke particles directly into the Earth's stratosphere, with lingering global implications. While pyroCb events have been previously observed, they are relatively rare and, outside of model simulations, little is known about their physical and chemical impacts. To assess current model-based assumptions of these events, Pengfei Yu and colleagues compared direct observations of the 2017 Pacific Northwest wildfires from the Stratospheric Aerosol and Gas Experiment III (SAGE III-ISS) and the Cloud-Aerosol Lidar and Infrared Pathfinder Satellite Observation (CALIPSO) satellite platforms with output from model simulations of these fires. Yu et al. report that the solar heating of the black carbon particles warmed the air within the smoke plume, causing it to self-loft, ascending from 12 kilometers to nearly 23 kilometers within two months, increasing its ability to spread latitudinally throughout the stratosphere. However, the observed smoke lifetime in the stratosphere was 40% shorter, the authors say, than what would be calculated using a standard model. This is because standard models do not necessarily consider photochemical loss of organic carbon - a phenomenon apparent in the 2017 plume rise. For future large wildfire events, photochemical reaction rates are important characteristics to measure, say the authors, to improve predictability of the 3-D transport of the smoke. They note their observational data confirms predictions of numerous models of Nuclear Winter, related to how smoke injected into the upper troposphere from urban fires will self-loft high into the stratosphere. However, the persistence of the smoke in the 2017 fire plume also calls into question assumptions of Nuclear Winter models related to how organics in smoke can be ignored due to their rapid loss, which was not observed here.