Influence of climate change and accidents on perception differs among energy technologies.

Risk perceptions of energy systems, and their evolution under climate change and after accidents, affect public acceptance of generation technologies. Despite this, little is understood about how such factors impact public perception at different timescales and the drivers for perception. We use state-of-the-art natural language processing to measure temporal changes in sentiment toward energy technologies using the full Twitter archive for 2009-2022. We find that perception of natural gas and wind has changed little as discussion of climate change on social media increased. However, climate-linked sentiment toward coal, solar, and hydropower has become more negative, while that for nuclear has improved. We also find that all generation technologies experience a drop in supportive discourse after definable accidents, but this typically rebounds with a half-life of <3 days. Yet, nuclear power is an exception in how it reacts to large-scale events. After Fukushima, sentiment returned to its positive preaccident levels with an 11.3-month relaxation half-life.