| 總結: | Dissolved organic carbon (DOC) in Arctic rivers plays an important role in the sea-land carbon cycle. In this study, in situ measured datasets and Landsat images were used to construct an efficient remote sensing retrieval model for DOC in Arctic rivers. After comparing multiple empirical and machine learning models, the random forest (RF) model with best performance was selected. This study investigated the spatiotemporal distribution of DOC in 15 important Arctic rivers during 2000–2023 and analyzed the relationship between DOC and basin characteristic variables such as climate, hydrology, vegetation, soil, and slope. Europe had the highest riverine DOC, followed by Siberia and North America, with average values of 9.22 ± 1.23, 7.85 ± 1.50, and 6.17 ± 0.77 mg/L, respectively. The annual mean DOC concentration exhibited an increasing trend (+0.04 mg/L yr−1), with an average of 7.96 ± 1.64 mg/L. The DOC concentration changed significantly during the different seasons, and the DOC concentration in spring was significantly higher than that in summer. Spatially, riverine DOC concentration is affected by soil texture and slope. In the ice-free period, the driving factors of riverine DOC are runoff, vegetation, snow cover, and permafrost. This study deepens our understanding of the Arctic carbon cycle through a comprehensive analysis of the spatiotemporal changes and driving factors of Arctic riverine DOC.
|
|---|