Journal of Geophysical Research-Oceans最新文献

The vertical distribution of subsurface chlorophyll (Chla) is crucial for understanding the marine primary productivity estimation and the ocean carbon cycle. Here we developed a dual-branch neural network (TCB-MHA) that integrates satellites, reanalysis, and BGC-Argo data to accurately estimate Chla profiles in tropical-subtropical oceans. The model shows exceptional performance in estimating subsurface Chla concentration (R² = 0.824–0.859, RMSE = 0.0525–0.0980 mg/m³), capturing the depth (R² = 0.919, MAPE = 7.23%) and intensity (R² = 0.795, MAPE = 12.97%) of the subsurface chlorophyll maximum (SCM). However, SCM thickness prediction remains challenging (R² = 0.513, MAPE = 16.82%), likely due to the lack of phytoplankton photoacclimation parameterization. Independent validation in the South China Sea confirmed the model's strong generalizability. Climatological analysis reveals clear latitudinal SCM patterns across tropical-subtropical oceans: deeper and weaker in subtropical regions (SCM depth >100 m, SCM intensity = 0.32 ± 0.10 mg/m³) versus shallower and stronger in tropical zones (SCM depth <80 m, SCM intensity = 0.48 ± 0.10 mg/m³). Seasonal and vertical SCM variations are also consistently captured. The model further quantifies the impact of mesoscale processes and spring blooms of the Western Mediterranean on Chla distribution. This study highlights the advantages of deep learning in integrating multi-source heterogeneous ocean data, and the resulting monthly 3D Chla products can be applied to improve carbon cycle and productivity studies.

Oceanic fronts significantly affect primary production. While surface fronts are well-studied, subsurface fronts have received relatively little attention. The impacts and underlying mechanisms of subsurface fronts on phytoplankton distribution and nitrogen cycle remain unclear, limiting our understanding of primary production. Based on data from in situ sampling, satellite, and reanalysis, pronounced thermal fronts occurred in the subsurface layers but weakened and disappeared toward the surface and deeper layers of the northern South China Sea. Despite differing formation mechanisms (i.e., dipole eddies and warm offshore water intrusion), both frontal zones exhibited substantially higher chlorophyll a (Chl a) levels than non-frontal zones (on average, Chl a concentrations increased by 77.78% and inventories rose by 88.56%). Positive correlations between frontal intensities and Chl a concentrations, along with enhanced convergence-divergence and vertical processes, suggested that Chl a aggregate relates to physical accumulation. Additionally, evident nitrate (NO₃⁻) loss and isotope enrichment factors (¹⁵$��\epsilon �$ = 3.4‰ and ¹⁸$��\epsilon �$ = 4.5‰) supported that Chl a increase were also associated with NO₃⁻ assimilation. However, 22.7% of the total NO₃⁻ pool in frontal zones was from nitrification, representing an increase of 15.2% compared to non-frontal zones. Elevated regeneration was attributed to enhanced oxygen exchange related to frontal dynamics, as suggested by comparable dissolved oxygen and ammonium levels in both zones, yet with elevated apparent oxygen utilization and NO₃⁻ in frontal zones. This study highlights that subsurface fronts not only facilitate phytoplankton aggregation but also drive active NO₃⁻ regeneration, thereby leading to an overestimation of new production in oligotrophic oceans.

With the rapid loss of Arctic sea ice in recent decades, the pack ice zone (PIZ) is gradually transitioning to the marginal ice zone (MIZ), but its impacts on heat exchanges within the atmosphere-sea ice-ocean system remain unquantified. This study identifies the transition region from PIZ to MIZ using a positive difference in MIZ occurrence frequency between 1979–2010 and 1992–2023, and compares changes in heat exchanges in this region to those over the pan Arctic Ocean. The transition from PIZ to MIZ in summer increases shortwave radiation absorption by the ice-ocean surface (0.7 W m⁻² yr⁻¹, P < 0.05), further increasing upper ocean heat content (0.04 × 10⁸ J m⁻² yr⁻¹, P < 0.05); while that in winter enhances the exchanges of longwave radiation (about 0.8 W m⁻² yr⁻¹, P < 0.05) and surface upward latent and sensible heat fluxes (0.4 and 0.5 W m⁻² yr⁻¹, P < 0.05), partly leading to an increase in 2-m air temperature (about 0.2 K yr⁻¹, P < 0.05), more than twice the average of the pan Arctic Ocean, although the transition region occupies less than 25% of the pan Arctic Ocean. The amplification of heat exchanges is more pronounced in the transition region from PIZ to MIZ than in the MIZ defined by ice concentration between 15% and 80%. Results highlight that the transition process from PIZ to MIZ is more critical for the thermodynamic coupling of the Arctic atmosphere-sea ice-ocean system, compared to changes in the location and extent of MIZ.

The Labrador Current is a major pathway for export of North Atlantic Deep Water across 53°N. While direct measurements of the integrated transport of the current were made from 1997 to 2014 at the 53°N Moored Observatory and since 2014 as part of the Overturning in the Subpolar North Atlantic Program (OSNAP), the calculations of integrated transport are typically delayed by years due to the time to recover and process moored instrumentation. Here, we examine a method to compute variations in the transport from ocean bottom pressure gradients determined from the GRACE and GRACE-FO satellite missions since 2002, which will allow near-real-time monitoring. After verifying the method using an ocean state estimate that captures the mean and variable transport similar to the observations, we calculate the variability using the satellite observations. While there is a notable degradation at the end of the GRACE mission after 2012, results from 2002 to 2011 are consistent with interannual variations seen in the 53°N Moored Observatory and results from the GRACE Follow-on mission (starting in 2018) are consistent with observations made by the OSNAP array. Monthly differences are of the order of 3–4 Sv compared to overall variability of ±6 Sv (one standard deviation). No statistically significant trend since 2002 was found.

Ocean currents flowing over the seabed produce a turbulent bottom boundary layer (BBL) which extracts energy from the overlying flow and mediates sedimentary, chemical, and biological processes at the seafloor. The role of the BBL is especially important in relatively shallow shelf seas where stratification, mean flow, and tidally-driven oscillating currents are typically strong. Despite this importance, the dynamics and even the height of the BBL formed under stratified, tidal conditions are not well understood. We address this knowledge gap by performing large-eddy simulations of the evolution of a tidally forced and stratified BBL, initialized from rest, over a range of geophysically relevant parameter values. The BBL has two distinct regions: an actively stirred, near-well-mixed bottom mixed layer (BML) and a weakly stirred and strongly stratified capping pycnocline. A simple steady-state empirical model, dependent on stratification, tidal period, and bottom turbulence intensity, provides a useful estimate of BML height. Motivated by appreciable growth in the BML that can occur on timescales of days to weeks, we also develop a time-dependent model for the BML height. This model accounts for the vertical structure of the BBL turbulence and reproduces the BML growth with high fidelity. We also present a simple model for the cycle-averaged eddy diffusivity, which depends on the BML height. Both the steady-state and time-dependent models of BML height—and the accompanying expression for eddy diffusivity—can be used to develop tidal BBL parameterizations and inform grid resolution requirements within ocean models, as well as guiding sampling for future observational studies.

As a crucial component of ecosystems, the distribution of zooplankton is closely related to environmental conditions and ocean currents. Despite its unique role in global circulation, zooplankton data for the Eastern Indian Ocean (EIO) remain scarce. To better characterize zooplankton composition and distribution in EIO, we conducted a survey in the region between 5°S and 14°N and 80°−93°E from March–May 2022. To our knowledge, this is among the first applications of ecological models to reveal the preliminary characteristics of zooplankton community assembly mechanisms in EIO. Through microscopic examination, we identified 427 species of adult zooplankton and 24 zooplankton larvae taxa. The zooplankton abundance in the surveyed ranged from 137 to 1,326 ind./m³. Among them, copepods, particularly small-bodied species, were the dominant components, contributing most significantly to the abundance and species richness of EIO zooplankton. Redundancy Analysis (RDA) and Random Forest (RF) results indicated that Dissolved oxygen (DO), temperature, and chlorophyll a were the primary factors influencing the abundance and diversity of EIO zooplankton. According to cluster analysis, the EIO zooplankton could categorized into three ecological groups: Group A (GA), Group B (GB) and Group C (GC). The results of Neutral Community Model revealed that the community assembly of EIO zooplankton was primarily influenced by stochastic processes. However, certain deterministic factors still influence the community assembly mechanisms of GA and GB. Correlation analyses between environmental factors and characteristic species in each group showed that deterministic assembly in GA is mainly driven by interspecific interactions, whereas in GB it is primarily driven by environmental selection.

Atlantic Water (AW) inflow plays a central role in Arctic Ocean warming and Atlantification, yet its transport into the Amerasian Basin remains poorly constrained. Using four ocean–sea ice reanalyses, including the new SODA4, we assess the structure, long-term trends, and variability of AW transport. While three of the reanalyses are consistent with prior transport estimates upstream of the Lomonosov Ridge, only SODA4 realistically captures the observed transports into the Amerasian Basin across the Lomonosov Ridge. One of the reanalyses, GLORYS12, significantly underestimates the heat content of Atlantic Water upstream of the Lomonosov Ridge as a result of anomalously high heat loss in the Barents Sea and excess cooling in the Barents Sea. We derive the first 40-year time series of heat transport into the Amerasian Basin. The ensemble mean of this time series shows that AW heat transport has increased by 0.36 TW/year (1984–2016) yielding a 5 ZJ increase in Amerasian Basin heat content. Interannual and interdecadal modes strongly modulate AW heat transport across the Lomonosov Ridge over the 40-year period. The leading mode of variability is associated with the phases of the Arctic Dipole anomaly, an atmospheric climate pattern which modulates the strength of the Siberian and Beaufort high-pressure systems. These results suggest that accurately resolving Atlantic Water structure and transport across the Lomonosov Ridge and in the Barents Sea is essential for accurately characterizing ocean-driven Arctic warming as far east as the Amerasian Basin and Beaufort Gyre.

This study investigates the long-term salinity dynamics of the partially mixed Ems Estuary, focusing on the influence of brine inflow on salt intrusion. Salinity time series from eight monitoring sites (2000–2023) were analyzed to assess variations due to river discharge and brine inflow from two injection sites. Tidally-averaged salinity was evaluated by tracking the positions of the 2 and 13 g$��·�$kg⁻¹ isohalines. Brine effects were identified by separating the data set into four subsets: two no-brine, one low-brine, and one high-brine period. For each, we computed temporal averages, absolute and relative salinity changes, and horizontal salinity gradients. Salinity showed strong seasonal variation driven by river discharge, both with and without additional salt input. The effect of brine inflow was superimposed on river discharge-related variability and depended on prevailing and previous discharge conditions. Although brine volumes were, on average, five orders of magnitude smaller than flood tidal volumes, they increased salt intrusion during low discharge conditions. Compared to no-brine scenarios, salt intrusion shifted up to 10 km further upstream during the high-brine period. Estuarine transport processes caused the salinity to increase 15–35 km upstream (exceeding the tidal excursion) of the brine inflow sites. The longitudinal location of the brine release affected salt dispersion, as demonstrated by a movement of the maximum salinity gradient. These findings emphasize the sensitivity of estuarine systems to small but persistent salt sources, with implications for estuarine ecosystems and management globally under increasing salinity stress from freshwater scarcity, land-use pressure, and climate change.

Large river-influenced coastal seas serve as critical interfaces for particulate organic matter (POM) dynamics within the global carbon cycle, characterized by intricate hydrodynamics and diverse sources. However, the response patterns of these POM processes to typhoon events remain poorly understood. This study conducted two field cruises in Laizhou Bay (LZB), a semiclosed bay with multiple river inputs, to compare the impacts of riverine pulses on POM dynamics during posttyphoon Lekima and a nontyphoon period. Typhoon-induced heavy rainfall significantly altered POM composition and fluxes. Posttyphoon, particulate organic carbon (POC) and particulate nitrogen fluxes increased to 4.9 × 10¹⁰ mmol d⁻¹ and 4.8 × 10⁹ mmol d⁻¹, respectively, compared to 1.5 × 10¹⁰ mmol d⁻¹ and 2.3 × 10⁹ mmol d⁻¹ during the nontyphoon period. Monte Carlo (MC) simulations quantitatively revealed riverine organic carbon (OC_riverine) dominated POC posttyphoon (44.8%), followed by marine OC (OC_marine) (35.0%), while deltaic organic carbon (OC_deltaic) was least (20.3%) in LZB. Under nontyphoon conditions, OC_deltaic dominated (46.0%) with OC_riverine being minimal (25.0%). Both δ¹³C and C/N ratio analyses showed that posttyphoon POM sources included freshwater algae, soil organic carbon, C3 plants, and marine POC, while nontyphoon POM primarily derived from freshwater algae and marine POC. Molecular indicators suggested that typhoon-driven terrestrial inputs of labile organic matter enhanced POM degradation via the priming effect. Consequently, POM degradation was more extensive posttyphoon, transforming the coastal region from a carbon sink to a carbon source. These disturbances likely reduce carbon burial in marine sediments during extreme weather events.

Large-scale offshore wind farms are expected to influence surface waves by modifying local wind forcing through wake effects. We use regional coupled ocean-atmosphere-wave model simulations to investigate a realistic large-scale offshore wind development scenario in the northeastern U.S. during boreal summer. Near-surface wind speeds are reduced by 10% over lease areas and within downstream wake regions, leading to decreases in significant wave height (3%) and wave-supported momentum flux (30%). This further leads to reductions in surface roughness length (16%) and near-surface ocean turbulent kinetic energy (20%). Spectral analysis shows a clear reduction in wind-sea energy, indicating suppressed local wind-wave growth near the wind farms. Weaker winds favor the development of longer-period waves, increasing dominant wave phase speed by 3% and suggesting a transition to an older sea state. Modern bulk flux algorithms often parameterize surface roughness using inverse wave age and/or wave slope. This raises the question of whether wake-driven reductions in inverse wave age and wave height impact air-sea momentum exchange. To assess this, we compare fully coupled simulations with an atmosphere-only run excluding wave coupling. Results show that about one-third of the reduction in roughness length can be attributed to sea state changes, while two-thirds result from lower friction velocity due to lower wind speeds. However, the impact of sea state on the drag coefficient and momentum flux is negligible ($��\sim �$1%), suggesting that wake-induced wind speed reductions are the primary driver, with sea state changes playing a secondary role.