This article presents a study of the content, use of sources and information sharing about climate change analysing over 60,000 tweets collected using a random week sample. We discuss the potential for studying Twitter as a communicative space that is rich in different types of information and presents both new challenges and opportunities. Our analysis combines automatic thematic analysis, semantic network analysis and text classification according to psychological process categories. We also consider the media ecology of tweets and the external web links that users shared. In terms of content, the network of topics uncovered presents a multidimensional discourse that accounts for complex causal links between climate change and its consequences. The media ecology analysis revealed a narrow set of sources with a major role played by traditional media and that emotionally arousing text was more likely to be shared.