Ocean Institute

ABSTRACT Neural networks have the potential in approximating the solutions of the Navier–Stokes equations. To address the high computational complexity of high‐order derivatives based on automatic differentiation and the high memory requirements of deep neural network training, a novel Deep Collocation Method (DCM) that combines reversible neural network (RevNet) and spectral differentiation is proposed. The memory‐efficient RevNet is employed to approximate the solutions, while Chebyshev differentiation matrices are derived to rapidly and accurately compute the spatiotemporal derivatives of the approximate solutions, making the computational complexity of high‐order derivatives comparable to that of the forward pass, and significantly reducing memory requirements. A unified spectral discretization scheme is applied in both temporal and spatial dimensions, enabling seamless integration of the neural network and the spatiotemporal Chebyshev differentiation matrices. It effectively incorporates global spatiotemporal information into the residual computation of the governing equations at each collocation point, enabling the neural network to actively account for nonlocal effects of the flow field. Validation results demonstrate that DCM performs well various benchmark flow problems and enables efficient training of deep models even on memory‐constrained devices.

We have examined three prominent near-field tsunamis that devastatingly impacted the contiguous Pacific coasts of Mexico. The 1985 tsunami was generated by a major (Mw 8.0) earthquake off the Mexican state of Michoacán on September 19 that caused serious damage and killed more than 5000 people. Tsunami waves from this event were observed at numerous sites along the coast, including Lázaro Cardenas, Zihuatanejo, Acapulco, and Manzanillo. The normal-fault earthquake that occurred on 8 September 2017 in the Gulf of Tehuantepec (Chiapas, Mexico) was an even stronger event (Mw 8.2), resulting in tsunami waves that were measured by a large number of high-resolution tide gauges on the Pacific coasts of California, Mexico and Central America and by three open-ocean DART stations located in the offshore region. The third tsunami was produced by a thrust fault Mw 7.6 earthquake on 19 September 2022 within the coastal zone of Michoacán, Mexico, i.e., on the same date and almost at the same location as the 1985 event. This tsunami was recorded by six coastal tide gauges and by DART 43412. All three tsunamis have been thoroughly examined and numerically simulated. Both the observational data and the modelling results demonstrate that the “strength” of the tsunami waves was mostly determined by the distance from the source rather than by the specific resonant characteristics of the individual recording sites. Numerical modelling of the events closely reproduced the coastal and offshore tsunami records. Our modelling also reveals markedly different anisotropic features for the tsunami energy radiation patterns, whereby the low-frequency energy was mostly concentrated in trapped edge modes propagating along the shelf while higher frequency leaky waves radiated tsunami energy outward from the source with the main beam directed like a “searchlight” normally to the mainland coast. The “trapping coefficients” – a measure of the trapped mode contribution to the tsunami wave energy – were estimated theoretically for all events, with values ranging from 60 to 80.

Bimetal cluster catalysts present boosted catalytic performance in numerous reactions due to their coordinatively unsaturated metal sites and synergy. Nevertheless, constructing ultrafine and stable bimetal cluster catalysts that can be well applied under harsh high-temperature reaction conditions is still a tough challenge. Herein, we propose a spatial and chemical dual-confinement strategy to design hierarchical ZSM-5 zeolite-confined PtZn bimetal catalysts with ultrafine clusters (ca. 1.2 nm) and high stability against sintering, which can be controllably synthesized via a microwave-assisted one-pot hydrothermal approach. The optimized PtZn@ZSM-5 cluster catalyst exhibits remarkably boosted propane dehydrogenation (PDH) performance, with a propane conversion of ca. 60% and a propylene selectivity of >96% (WHSV = 3 h–1, 600 °C). Both experimental and theoretical studies demonstrate that Pt species are well anchored by the hierarchical pores and framework zinc, constructing highly dispersed active Pt–(O–Zn)x species responsible for the boosted catalytic performance. This work affords a facile and general dual-confinement strategy to synthesize ultrafine bimetal cluster catalysts for PDH, which can be well extended to other catalyst systems, such as RhZn@ZSM-5 and IrZn@ZSM-5.

Lindenia tetraphylla, a dragonfly species belonging to the family Gomphidae, is recorded for the first time in Romania. This study documents the discovery and discusses its potential distribution and future perspectives in Romania. Given the changing climatic conditions and habitat modifications, this finding raises questions regarding the species’ range expansion and conservation status in the region.

The purpose of this investigation was to examine the effects of a 70-day Speed, Agility and Quickness (SAQ) training programme on swimming performance for pre-adolescent male swimmers associated with a swimming club in Beijing, China. Twenty male swimmers, consisting of 10 male swimmers aged under 10 years and 10 male swimmers aged under 12 years, participated in the study. Using a one-group, repeated measures design, all swimmers incorporated SAQ training into their regular swim training schedule, participating in five SAQ training sessions per each 10-day period throughout the 70-day intervention period. Each SAQ training session lasted between 20 and 25 minutes, and performance was measured by a 100-metre sprint time, collected at each 10-day interval from baseline (day 1) to completion of the study (day 70). Repeated measures analysis of variance (ANOVA) indicated a statistically significant overall improvement in swim times during the study (F = 155.689, p < 0.001, partial η² = .876) with no difference between baseline (day 1) and follow-up (day 10) (p = 1.000), while statistically significant improvements were shown by day 20 (mean difference [MD] = .401, p = .007) and continued to day 70 with the largest improvements recorded (MD = 2.531, p < 0.001) across all 4 data collection points in both age groups. The results of this study indicate that a structured (scientifically based) 70-day (approximately 2.5-month) SAQ programme significantly enhances 100-metre sprint performance in pre-adolescent swimmers. The results of this investigation suggest that the use of SAQ strength training methods leads to improved muscular coordination (more cohesive and quicker). Therefore, the establishment of strength through the use of modalities such as SAQ at an early age will produce successful athletes. Coaches should consider combining aquatic and non-aquatic strength training with swimming training in order to facilitate performance gains at an accelerated rate, and to produce faster performance gains with earlier initiation of such training than later initiation.

Killer whales (Orcinus orca) are cosmopolitan apex predators that interact with numerous species, including sharks, skates, and rays (subclass Elasmobranchii). Interactions between killer whales and elasmobranchs have garnered attention from the scientific community, but the dynamics of these events are poorly understood, as interactions are often challenging to observe. In this study, we collated imagery of interactions between killer whales and elasmobranchs to create the first image-based repository of interspecific interactions between marine predators. A total of 320 photos and videos from 82 unique interactions were collected from 12 countries. A minimum of 18 elasmobranch species across 16 genera were identified from the imagery. Furthermore, 10 different interaction types were documented, encompassing both predatory and non-predatory behavior. This dataset can provide insight into various research topics, including killer whale predation techniques, elasmobranch behavior, and the identification of individual killer whales that specialize in elasmobranch prey. As a new and openly accessible resource, this dataset provides the foundation to support future ecological and behavioral research on killer whale and elasmobranch interactions.

Extreme rainfall and flooding, driven by a powerful atmospheric river (AR) and a persistent Madden-Julian Oscillation (MJO), hit Southern California in February 2024 during the 2023–2024 El Niño, affecting over 10 million people. ARs are key contributors to extreme rainfall and flooding along the U.S. West Coast. Although the AR-MJO link has been documented, its spatio-temporal variability remains a major forecasting and risk-management challenge. Combining precipitation, stream gauge and demographic data, we quantify the physical drivers and population exposure to this extreme event. Leveraging a Lagrangian MJO precipitation tracking algorithm, we unravel the multiscale interactions responsible for the AR’s development. El Niño favored a large, long-lived MJO that interacted with the North Pacific Jet (NPJ) over more than three weeks. The MJO convective outflow modulated the NPJ by inducing negative potential vorticity advection along the tropopause. The ensuing NPJ extension and acceleration induced explosive cyclogenesis, whose AR-driven moisture transport resulted in extreme rainfall.

Abstract We describe a minimally invasive pilot study to characterize the microbiota of exhaled breath from wild Pacific white-sided dolphins. Samples were collected in a site that is far from any agriculture facilities or large human settlements but home to a high concentration of open-net aquaculture facilities for Atlantic salmon. This case study in pathogen surveillance in wild dolphins reveals a wide diversity of pathogens, including several with the potential to cross species and infect critically endangered killer whales, or humans. Common species are often neglected in funding strategies that prioritize species at imminent risk of extinction, but they may be a top priority for disease surveillance. One important cobenefit of integrating biodiversity monitoring in a One Health framework may be to support monitoring of both abundance and zoonotic potential of common species whose biomass alone makes them key players in ecosystem function.

We present Ocean Genomes, a program dedicated to producing reference genome resources to facilitate improved monitoring approaches and management outcomes for marine vertebrate biodiversity. Ocean Genomes will generate high-quality reference genomes of representatives of all marine vertebrate families and additional high-conservation-value species. Draft-quality genomes may be produced for a more comprehensive sampling of species. We include case studies of Enoplosus armatus, Old Wife and Pempheris klunzingeri, Rough Bullseye.