Browsing by Author "Li, X."

Barium/calcium ratios in the skeletons of scleractinian shallow-water corals have been used as proxies for coastal and oceanic processes such as river discharge, oceanic upwelling and surface ocean productivity. However, the variations in Ba/Ca ratios in aragonitic coral skeletons remain difficult to interpret as an environmental proxy. This difficulty is mainly due to the influence of internal (biomineralization) and multiple external (environmental) processes on Ba incorporation into coral skeletons, and these processes are hard to constrain with Ba/Ca alone. Here we present the first annually-resolved records of the Ba isotopic compositions (δ138/134Ba) in shallow-water corals (Porites) collected alive in the field, supplemented by the analysis of Ba/Ca ratios. Seven coral cores were recovered at different oceanic settings in the South China Sea, extending from the northern inner shelf to the central and southern deep basin. The annual δ138/134Ba records of six corals fell within a narrow range from 0.24 ± 0.03‰ to 0.38 ± 0. 03‰ (2SD), with a mean value of 0.33 ± 0.08‰ (2SD, N = 21). One single inner-shelf coral revealed low δ138/134Ba values (0.10 ∼ 0.11 ± 0.03‰), which might reflect the influence of terrestrial water/sediment. In contrast, the coral Ba/Ca ratios showed a wide range of intercolony differences, from 2 to 14 μmol/mol. This variation is too large to be ascribed to the changes in the Ba concentrations of seawater or other environmental parameters. Rayleigh fractionation between corals and seawater during biomineralization was proposed to explain the anomalous variations in the Ba/Ca ratios observed in coral skeletons. However, this result is incompatible with the relatively constant Ba isotopic compositions in coral. Instead, we suggest that the probable precipitation of witherite (BaCO3) within the domains of aragonite under oversaturated calcifying fluid could explain the large variability in the coral Ba/Ca ratios. The coral δ138/134Ba records from diverse oceanic settings were largely unaffected by biomineralization processes and temperature and displayed a relatively constant negative offset from typical surface seawaters. Our results suggest that Ba isotopes in Porites could be a proxy for reconstructing the δ138/134Ba of seawater and hence provide new insights into Ba cycling in the upper oceans in the past. © 2018 Elsevier Ltd

Phenolic compounds are a large class of plant secondary metabolites with various health-promoting effects, and are known for their structural diversity. Therefore, high efficiency characterization of phenolic profiles is of key importance in identifying their potential bioactivity. In the present study, Global Natural Products Social (GNPS) Molecular Networking was applied to trace the phenolic compounds in plants, which allowed the characterization of 9 procyanidins and 11 flavonoid glycosides (di-, tri-, or tetra-saccharides of kaempferol, quercetin, isorhamnetin and myricetin) in litchi pulp extracts. Six compounds were reported for the first time in litchi pulp. In addition, quercetin-3-O-rutinoside-(1 → 2)-O-rhamnoside, the most abundant flavonoid glycoside in litchi pulp, was proved to have considerable α-glucosidase inhibitory activity, illustrating the anti-diabetic potential of phenolic-rich litchi pulp extracts. ? 2019 Elsevier Ltd

Understanding convective–radiative–mixing processes is crucial in making better predictions about tropical climate. The cloud–resolving model and the mixed–layer model, combined with observations, are powerful tools for studying these physical processes interacting with climate. In this article, the authors’ research work of the past 15 years on tropical climate processes is reviewed. The topics reviewed include climate equilibrium study, tropical convective responses to radiative and microphysical processes, the diurnal cycle, cloud clustering and associated cloud–microphysical processes, precipitation efficiency, air–sea exchanges and ocean–mixing processes at diurnal–to–intraseasonal scales, and coupled boundary layer and forced oceanic responses. Representation of these processes in climate models and future perspectives are also discussed. © 2008 by World Scientific Publishing Co. Pte. Ltd. All rights reserved.

The Kuroshio Current (KC), one of the most important western boundary currents in the North Pacific Ocean, strongly affects regional hydroclimate in East Asia and upper ocean thermal structure. Limited by few on-site observations, the responses of the KC to regional and remote climate forcings are still poorly understood. Here we use monthly coral δ18O data to reconstruct a KC transport record with annual to interannual resolution for the interval 1953–2004. The field site is located in southern Taiwan on the western flank of the upstream KC. Increased (reduced) KC transport would generate strong (weak) upwelling, resulting in relatively high (low) local coral δ18O. The upstream KC transport and downstream transport, off Tatsukushi Bay, Japan, covary on interannual and decadal time scales. This suggests common forcings, such as meridional drift of the North Equatorial Current bifurcation, or zonal climatic oscillations in the Pacific. The intensities of KC transport off southeastern and northeastern Taiwan are in phase before 1990 and antiphase after 1990. This difference may be due to a poleward shift of the subtropical western boundary current as a response to global warming. © 2017. American Geophysical Union. All Rights Reserved.

This paper presents a Discriminative Deep Dyna-Q (D3Q) approach to improving the effectiveness and robustness of Deep Dyna-Q (DDQ), a recently proposed framework that extends the Dyna-Q algorithm to integrate planning for task-completion dialogue policy learning. To obviate DDQ's high dependency on the quality of simulated experiences, we incorporate an RNN-based discriminator in D3Q to differentiate simulated experience from real user experience in order to control the quality of training data. Experiments show that D3Q significantly outperforms DDQ by controlling the quality of simulated experience used for planning. The effectiveness and robustness of D3Q is further demonstrated in a domain extension setting, where the agent's capability of adapting to a changing environment is tested.1 © 2018 Association for Computational Linguistics