Evaluation of carbon sink in the Taklimakan Desert based on correction of abnormal negative CO2 flux of IRGASON

Studies showing that deserts can sequester CO2 through non-photosynthetic processes have contributed to locating missing carbon sinks. However, the contradiction between the desert CO2 flux obtained by different observation methods leads to uncertainty in evaluating desert carbon sequestration. This has caused scepticism regarding desert carbon sequestration after years of research. Through a comparative experiment in the non-vegetated shifting sand of the Taklimakan Desert (TD), it was found that if the abnormal negative CO2 flux observed by IRGASON during the day was not corrected, the carbon sequestration of the TD would be overestimated. The CO2 flux observed by EC155 is highly consistent with that of LI-COR8100A and can reflect the real CO2 exchange in the desert. The CO2 flux observed by EC155 was used to correct the results of IRGASON. Results show that the expansion/contraction of soil air containing CO2 caused by the change in the daily average soil temperature difference (T0-10cm) drives CO2 exchange in shifting sand. This results in diurnal vari-ation of CO2 release caused by shifting sand during the day and CO2 absorption at night, anda unimodal distribution of CO2 exchange caused by shifting sand throughout the year. From April to September (T0-10cm > 2 ?), the shifting sand releases CO2 as a carbon source. In the other months (T0-10cm < 2 ?), the shifting sand absorbs CO2 as a carbon sink. The stronger absorption shows that the shifting sand in the TD provides carbon sequestration, with a CO(2)uptake rate of ~148.85 x 10(4 )tons a-1. This suggests that deserts play an active role in locating the missing carbon sinks and miti-gating climate change, and that the status of deserts in the global carbon cycle cannot be ignored.