图书馆知识库

Accurate assessment of the strength of China's terrestrial ecosystem carbon sink is key to understanding its regional carbon budget. However, large uncertainties in current carbon sink estimations still exist, which hinder the prediction of future climate change trajectories. In this study, we generated a high-resolution (1 km x 1 km) dataset of China's net ecosystem productivity (NEP) in the last decade via a model tree ensemble approach combined with data from 46 flux sites in China and neighboring regions. The upscaling also included detailed information on nitrogen (N) deposition and forest age that have often been neglected in previous studies. The performance of MTE algorithm in simulating NEP at the site level is relatively high for both training (R-2 = 0.81, RMSE = 0.73 gC m(-2) day(-1)) and validation datasets (R-2 = 0.76, RMSE = 0.81 gC m(-2)day(-1)). Our data-driven estimation showed that roughly 70% of the area is a carbon sink, and the largest carbon sinks are found in the southeast and southwest monsoon regions. The total annual NEP in China in the last decade was 1.18 +/- 0.05 Pg C yr(-1), which is similar to the results found by another foundational global-scale study. Yet, the two studies significantly differ in the spatial distribution of carbon sink density. The seasonality of China's NEP is characterized by region-specific kurtosis and skewness in most areas. Furthermore, ecosystem carbon use efficiency (CUE), defined as the annual NEP/GPP ratio, also showed high spatial variation. For example, the Xiaoxing'anling and Changbai Mountains in northeastern China, the eastern edge of the Tibetan Plateau, and bordering areas of the southeast and southwest monsoon regions have a larger CUE than the rest of China. On average, China's terrestrial ecosystem CUE is approximately 0.17. Our data-driven NEP and CUE estimates provide a new tool for assessing China's carbon dioxide flux. Our study also highlights the necessity to incorporate more environmental variables related to vegetation growth and more data derived from flux sites into NEP upscaling to reduce uncertainties in carbon budget estimations.