What are the trends and drivers of riverine CO2 emissions in the Tibetan Plateau?

Sammanfattning

The release of carbon dioxide (CO2) from inland waters plays an important role in the global carbon cycle. Yet, there are uncertainties related to estimating riverine CO2 evasion (FCO2), especially in data scarce regions such as the Tibetan Plateau. It is unclear how global change in the form of climate warming and anthropogenic disturbance affects emissions. This study reports annual CO2 fluxes as well as their change from 2000 to 2019 for streams and rivers across the Tibetan Plateau by comparing two representative but distinct approaches to calculate each of the commonly requested FCO2 modelling parameters, namely river surface area, gas transfer velocity, riverine partial pressure of CO2, and atmospheric CO2 concentration, yielding a total of 16 combinations. Across the Tibetan Plateau, the annual riverine CO2 emissions increased significantly from 2000 to 2019 for all model scenarios, with average values for the 16 scenarios ranging from 0.26 to 43.2 Tg C yr-1. The most realistic annual CO2 evasion was 37.7 ± 2.84 Tg C yr-1, based on the combination of a geometric equation for river surface area, energy-dissipation based gas transfer velocity, a machine learning model for aquatic partial pressure of CO2, and the local Tibetan Plateau-specific atmospheric CO2 concentration. The presented emissions were twice as high as the previous estimate, suggesting a serious underestimation of CO2 dynamics on the Tibetan Plateau. The differences between the CO2 emission models were caused by the CO2 concentration gradient between atmosphere and water, as the interplay of global/local atmospheric CO2 concentrations and the two models for riverine CO2 concentrations (empirical/machine learning) determined the direction (i.e., whether there was CO2 release or uptake) and strength of the emissions, while effective river surface area and gas transfer velocity played a lesser role. There was a clear dichotomy between the low emissions/slight uptake in the central regions of the Tibetan Plateau and the high CO2 evasion in the southeastern parts. The trend in aquatic CO2 emissions was driven by the river surface area, in turn dependent on the discharge. However, the combination of the large range of riverine CO2 emissions and the lack of direct pCO2 measurements in rivers and streams highlights the associated uncertainties in emission calculations, especially when estimating the riverine CO2 concentrations. With a changing climate, discharge might increase on the Tibetan Plateau, owed to the warming-induced melting of permafrost and glaciers, as well as the projected increase in precipitation. Following the connection between discharge and CO2 release, this might lead to higher CO2 emissions from the Tibetan Plateau. To improve the estimates of riverine CO2 dynamics on the Tibetan Plateau, the partial pressure of CO2 in rivers and streams needs to be modelled more accurately, using a refined model with higher temporal variation, as well as more direct observations.