Biomass burning is a widespread practice. During burning

fire combusts organic matter and emitsalarge amount of carbonaceous gases into the atmosphere. Biomass burning not only changes the structure and process of the ecosystem but also affects the carbon cycle of the entire system. To elucidate the impact of wildfire on global carbon cycle

large-scale carbon emissions from biomass burning have been estimated using satellite remote sensing. Many remote-sensing-based models have been developed to estimate biomass burning emissions at different scales. The most widely used model contains four key parameters: burned area

fuel load

burning efficiency

and carbon fraction. The first three parameters can be retrieved from satellite data. This paper discussesmethodologies for the retrieval of these three key input parameters anddescribesthe advantages and disadvantages of each methodology. Methods for the estimation of burned area can be categorized into three types: reflectance-

emission-

and backscatter feature-based methods. Fuel load mapping can be classified as direct and indirect. Indirect fuel load mappingclassifies satellite data to determine the fuel typeand then assigns fuel load value to each pixel depending on the fuel type in the fuel models. This method strongly relies on fuel model and ismostly not suitable for large-scale areas. Direct fuel load mapping estimates fuel load value on the basis of the relationship among fuel load

relative factor of fuel load

and satellite data. Burning efficiency or combustion completeness is usually estimated through direct and indirect retrievalmethods. The direct retrievalmethod is difficult to be usedat alarge scale

whereasthe indirect retrievalmethod maps the burn severity firstand then adjusts the preset fixed burning efficiency on the basis ofburn severity. Finally

suggestions are provided to improve the accuracy of remote sensing in estimating carbon emissions from biomass burning. Many studies have been conducted to retrieve carbon emission-related parameters throughremote sensing. However

the adaptability and uncertainty of theseestimationsfor large-scale areas remain unclear

andthe estimationaccuracy ofglobal carbon emission does not satisfy the demand of research on carbon cycle.