Attribution of climate change to a nation’s historical emissions is a part of ongoing policy discussions within the UN Framework Convention on Climate Change (e.g., the Brazil proposal) and tends to arise in acrimonious negotiations, including the most recent COP in Durban. Such arguments require a quantifiable link between inventoried human activities and climate change. While the observed increase in the atmospheric abundances of greenhouse gases from pre-industrial era to present day is ascribed with high certainty as the major cause of recent climate change (e.g., IPCC’s Detection and Attribution Chapters), an equivalent attribution of this increase to human activities has not been made. A major uncertainty is the absolute magnitude of the current anthropogenic emissions of the greenhouse gases CH4 and N2O, whose inventories have uncertainties of 25% to 50% at a minimum. A new, top-down approach is taken to better constrain these emissions, using current knowledge of atmospheric chemistry lifetimes and budgets for both present day and preindustrial era. Propagating uncertainties from 23 primary observed or modeled quantities, we derive a present-day atmospheric lifetime for CH4 of 9.1±0.9 y and anthropogenic emissions of 350±65 Tg/y (64% of total emissions), providing the best current constraint on the human impact and providing a method for projecting future abundances with uncertainties. In addition, a soup-to-nuts propagation of errors is presented, tracking uncertainties from human activities to greenhouse gas emissions, to increasing abundances of greenhouse gases, to radiative forcing of climate, and finally to climate change. This example follows the causal chain for greenhouse gases emitted by developed nations since national reporting began in 1990. We find that the global-mean surface temperature increased by +0.11C in 2003 due to the developed nations’ emissions of Kyoto greenhouse gases from 1990 to 2002. The uncertainty range is large considering that the developed countries emissions are well known for this period and climate system modeling uncertainties are constrained by observations.