Measurement and modelling of regional or country-level carbon dioxide (CO 2 ) fluxes are becoming critical for verification of the greenhouse gases emission control. One of the commonly adopted approaches is inverse modelling, where CO 2 fluxes (emission: positive flux, sink: negative flux) from the terrestrial ecosystems are estimated by combining atmospheric CO 2 measurements with atmospheric transport models. The inverse models assume anthropogenic emissions are known, and thus the uncertainties in the emissions introduce systematic bias in estimation of the terrestrial (residual) fluxes by inverse modelling. Here we show that the CO 2 sink increase, estimated by the inverse model, over East Asia (China, Japan, Korea and Mongolia), by about 0.26 PgC year . Independent results from methane (CH 4 ) inversion suggested about 41% lower rate of East Asian CH 4 emission increase during 2002-2012. We apply a scaling factor of 0.59, based on CH 4 inversion, to the rate of anthropogenic CO 2 emission increase since the anthropogenic emissions of both CO 2 and CH 4 increase linearly in the emission inventory. We find no systematic increase in land CO 2 uptake over East Asia during 1993-2010 or 2000-2009 when scaled anthropogenic CO 2 emissions are used, and that there is a need of higher emission increase rate for 2010-2012 compared to those calculated by the inventory methods. High bias in anthropogenic CO 2 emissions leads to stronger land sinks in global land-ocean flux partitioning in our inverse model. The corrected anthropogenic CO 2 emissions also produce measurable reductions in the rate of global land CO 2 sink increase post-2002, leading to a better agreement with the terrestrial biospheric model simulations that include CO 2 -fertilization and climate effects.