Biomass burning is a large important source of greenhouse gases and atmospheric aerosols, and can contribute greatly to the temporal variations of CO2 emissions at regional and global scales. In this study, we compared four globally gridded CO2 emission inventories from biomass burning during the period of 2002-2011, highlighting the similarities and differences in seasonality and interannual variability of the CO2 emissions both at regional and global scales. The four datasets included Global Fire Emissions Database 4s with small fires (GFED4s), Global Fire Assimilation System 1.0 (GFAS1.0), Fire INventory from NCAR 1.0 (FINN1.0), and Global Inventory for Chemistry-Climate studies-GFED4s (G-G). The results showed that in general, the four inventories presented consistent temporal trend but with large differences as well. Globally, CO2 emissions of GFED4s, GFAS1.0, and G-G all peaked in August with the exception in FINN1.0, which recorded another peak in annual March. The interannual trend of all datasets displayed an overall decrease in CO2 emissions during 2002-2011, except for the inconsistent FINN1.0, which showed a tendency to increase during the considered period. Meanwhile, GFED4s and GFAS1.0 noted consistent agreement from 2002 to 2011 at both global (R 2 > 0.8) and continental levels (R 2 > 0.7). FINN1.0 was found to have the poorest temporal correlations with the other three inventories globally (R 2 < 0.6). The lower estimation in savanna CO2 emissions and higher calculation in cropland CO2 emissions by FINN1.0 from 2002 to 2011 was the primary reason for the temporal differences of the four inventories. Besides, the contributions of the three land covers (forest, savanna, and cropland) on CO2 emissions in each region varied greatly within the year (>80%) but showed small variations through the years (<40%).
Keywords: Biomass burning; CO2 emissions; Fires; Remote sensing; Temporal variation.