This study explores the role of Twitter bots (automated users) in online discourse on climate change. We examined 6.5 million tweets posted during the days leading up to and the month following President Donald Trump’s June 1, 2017 announcement of the United States’ withdrawal from the 2015 Paris Climate Agreement. Of a 10% sample of users, we used the machine learning algorithm “Botometer” to identify likely mechanized “bots.” Botometer identified 17,509 suspected bot accounts, representing about 9% of users and 17% of all tweets. Query limits on Botometer and the capacity of STM modeling reduced our final sample size to 167,259 tweets. We then used the ‘stm’ package in the R statistical programming language to implement structural topic models to cluster tweet content into topics. We identified 34 topics. Topics broadly fell into categories related to news of the withdrawal and the responses from various media and government personnel, posts about climate change research, discussions of the denial of climate change, and finally activist topics with campaign goals. Within these topics, bots were often common, representing as much as 38% of tweets. Additionally, among the most common 15 topics, we found evidence that after adjusting for topic size, timing of tweets, and whether a user was suspended, tweets produced by bots were more likely relative to non-bot users to fall into four topics. These topics included tweets sharing news links on the announcement, links related to climate research, links sharing denialist research, and links about White House aids and their views of fossil fuels. These findings suggest a substantial impact of mechanized bots in amplifying denialist messages about climate change, including support for Trump’s withdrawal from the Paris Agreement