Boreal-Arctic wetland methane emissions modulated by warming and vegetation activity.

Wetland methane (CH4) emissions over the Boreal-Arctic region are vulnerable to climate change and linked to climate feedbacks, yet understanding of their long-term dynamics remains uncertain. Here, we upscaled and analysed two decades (2002-2021) of Boreal-Arctic wetland CH4 emissions, representing an unprecedented compilation of eddy covariance and chamber observations. We found a robust increasing trend of CH4 emissions (+8.9%) with strong inter-annual variability. The majority of emission increases occurred in early summer (June and July) and were mainly driven by warming (52.3%) and ecosystem productivity (40.7%). Moreover, a 2 °C temperature anomaly in 2016 led to the highest recorded annual CH4 emissions (22.3 Tg CH4 yr-1) over this region, driven primarily by high emissions over Western Siberian lowlands. However, current-generation models from the Global Carbon Project failed to capture the emission magnitude and trend, and may bias the estimates in future wetland CH4 emission driven by amplified Boreal-Arctic warming and greening.

Vegetation clumping modulates global photosynthesis through adjusting canopy light environment.

The spatial dispersion of photoelements within a vegetation canopy, quantified by the clumping index (CI), directly regulates the within-canopy light environment and photosynthesis rate, but is not commonly implemented in terrestrial biosphere models to estimate the ecosystem carbon cycle. A few global CI products have been developed recently with remote sensing measurements, making it possible to examine the global impacts of CI. This study deployed CI in the radiative transfer scheme of the Community Land Model version 5 (CLM5) and used the revised CLM5 to quantitatively evaluate the extent to which CI can affect canopy absorbed radiation and gross primary production (GPP), and for the first time, considering the uncertainty and seasonal variation of CI with multiple remote sensing products. Compared to the results without considering the CI impact, the revised CLM5 estimated that sunlit canopy absorbed up to 9%-15% and 23%-34% less direct and diffuse radiation, respectively, while shaded canopy absorbed 3%-18% more diffuse radiation across different biome types. The CI impacts on canopy light conditions included changes in canopy light absorption, and sunlit-shaded leaf area fraction related to nitrogen distribution and thus the maximum rate of Rubisco carboxylase activity (Vcmax ), which together decreased photosynthesis in sunlit canopy by 5.9-7.2 PgC year-1 while enhanced photosynthesis by 6.9-8.2 PgC year-1 in shaded canopy. With higher light use efficiency of shaded leaves, shaded canopy increased photosynthesis compensated and exceeded the lost photosynthesis in sunlit canopy, resulting in 1.0 ± 0.12 PgC year-1 net increase in GPP. The uncertainty of GPP due to the different input CI datasets was much larger than that caused by CI seasonal variations, and was up to 50% of the magnitude of GPP interannual variations in the tropical regions. This study highlights the necessity of considering the impacts of CI and its uncertainty in terrestrial biosphere models.

A hierarchical temporal attention-based LSTM encoder-decoder model for individual mobility prediction.

Prediction of individual mobility is crucial in human mobility related applications. Whereas, existing research on individual mobility prediction mainly focuses on next location prediction and short-term dependencies between traveling locations. Long-term location sequence prediction is of great importance for long-time traffic planning and location advertising, and long-term dependencies exist as individual mobility regularity typically occurs daily and weekly. This paper proposes a novel hierarchical temporal attention-based LSTM encoder-decoder model for individual location sequence prediction. The proposed hierarchical attention mechanism captures both long-term and short-term dependencies underlying in individual longitudinal trajectories, and uncovers frequential and periodical mobility patterns in an interpretable manner by incorporating the calendar cycle of individual travel regularities into location prediction. More specifically, the hierarchical attention consists of local temporal attention to identify highly related locations in each day, and global temporal attention to discern important travel regularities over a week. Experiments on individual trajectory datasets with varying degree of traveling uncertainty demonstrate that our method outperforms four baseline methods on three evaluation metrics. In addition, we explore the interpretability of the proposed model in understanding individual daily, and weekly mobility patterns by visualizing the temporal attention weights and frequent traveling patterns associated with locations.