To the Editor

By compiling the emission inventories of China's 30 provinces (excluding Tibet, Hong Kong, Macao and Taiwan) and the nation as a whole in 2010, Guan and his co-workers reported an 18% difference in estimates of China's CO2 emissions1. Although several possible reasons have been suggested2, the researchers were unable to resolve the source of the discrepancy and could not identify which value was the most accurate3. Such discrepancies are apparent not only in energy consumption but also in other economic and environmental datasets, such as gross domestic product (GDP). Throughout the past decades, the data in China's statistical yearbooks never equal the sum of the numbers shown in the provincial statistical yearbooks. For example, just in the first half of 2012 the gap in GDP between the country data announced by the National Statistic Bureau (NSB) and the aggregation of its 31 provinces (excluding Hong Kong, Macao and Taiwan) is about RMB3,000 billion, about 14% of the national total4, whereas the difference in CO2 emissions in 2010 reported by Guan et al.1 is about 18% compared with the national figure. In both cases the sum of the provinces is greater than the national total.

To understand the possible reasons for the reported inconsistencies, we must take into account the differences between the national and local statistical systems. All the indicators are counted both at national and provincial level and it is the job of NSB to validate the provincial data and announce the national data after removing duplicate entries. Since 2000, international organizations such as the World Bank, as well as domestic institutes have admitted that the national-level statistical data should be adapted when we study the whole of China due to this duplicate counting at the local level. Researchers should not drop hints to favour the 'bigger' number of China's carbon dioxide emissions or just focus on describing the global impacts resulting from the discrepancies; we have to show objective caution regarding such uncertainty, especially with respect to CO2 emissions.