Over the past 10 years, Mexican officials and scientists have promoted the project of protecting Mexican forests in order to mitigate climate change, forests acting to absorb carbon dioxide from the atmosphere. This article compares existing policies around mass reforestation and markets for environmental services, and their relationships to a policy in construction – Reduced Emissions through Degradation and Deforestation. Mass reforestation policies collapsed in the face of politicized audits and stories about corruption; markets for environmental services continued with little criticism, stabilized in part by the charisma of Reduced Emissions through Degradation and Deforestation policies. I explain the collapse of mass reforestation policies as being due to failed knowledge performances by officials and scientists; such failures are assessed by more or less skeptical publics who expect specific ways of performing credible public knowledge. Areas of nonknowledge can be tamed as calculable uncertainty, or alternatively transformed into ontological indeterminacy, scandals, and stories of corruption. Areas of nonknowledge are not pathological: they may support, as well as undermine, climate science, the authority of institutions, or the credibility of carbon accounts.