Search

The Northwest Tropical Atlantic (NWTA) is a region of complex surface ocean circulation. The most prominent feature is the North Brazil Current (NBC) and its retroflection at 8°N, which leads to the formation of numerous mesoscale eddies known as NBC rings. The NWTA also receives the outflow of the Amazon River, generating freshwater plumes that can extend up to 100,000 km2. We show that these two processes influence the spatial variability of the region's surface latent heat flux (LHF). On the one hand, the presence of surface freshwater modifies the vertical stratification of the ocean, the mixed layer heat budget, and thus the air‐sea heat exchanges. On the other hand, NBC rings create a highly heterogeneous mesoscale sea surface temperature (SST) field that directly influences the near‐surface atmospheric circulation. These effects are illustrated by observations from the ElUcidating the RolE of Cloud‐Circulation Coupling in ClimAte ‐ Ocean Atmosphere (EUREC4A‐OA) and Atlantic Tradewind Ocean‐Atmosphere Interaction Campaign (ATOMIC) experiments, satellite and reanalysis data. We decompose the LHF budget into several terms controlled by different atmospheric and oceanic processes to identify the mechanisms leading to LHF changes. We find LHF variations of up to 160 W m2, of which 100 W m2 are associated with wind speed changes and 40 W m2 with SST variations. Surface currents or heat release associated with stratification changes remain as second‐order contributions with LHF variations of less than 10 W m2 each. This study highlights the importance of considering these three components to properly characterize LHF variability at different spatial scales, although it is limited by the scarcity of collocated observations. Plain Language Summary: The Northwest Tropical Atlantic (NWTA) is a region with a complex ocean circulation. It is dominated by the North Brazil Current (NBC), which parallel to the South American coast and changes its direction at 8°N. This leads to the formation of closed swirling circulations known as NBC rings. The NWTA also receives the outflow of the Amazon River. These two features affect the heat exchange between the ocean and the atmosphere associated with water evaporation (latent heat flux, LHF) as they modify sea surface temperature, salinity and the near‐surface atmospheric circulation. Here, we use the observations collected from the ElUcidating the RolE of Cloud‐Circulation Coupling in ClimAte ‐ Ocean Atmosphere (EUREC4A‐OA) and Atlantic Tradewind Ocean‐Atmosphere Interaction Campaign (ATOMIC) experiments, satellite data and combined observations with models to identify the key mechanisms leading to such LHF variations. More of 60% of them are associated to surface winds whilst sea surface temperature is behind a 25%. The Amazon outflow accounts for less than 10%. Although this study is limited by the paucity of oceanic, atmospheric and air‐sea interface observations located at the same point in time and space, it highlights the importance of considering these three components to properly describe LHF variability. Key Points: Latent heat flux (LHF) presents strong spatial variations in the northwest tropical Atlantic (NWTA), which has a complex ocean circulationSurface winds and sea surface temperature are the major drivers of LHF changes. The Amazon plume remains as a second‐order contributorIt is necessary to distinguish between spatial scales (mesoscale and below vs. large‐scale) when assessing the ocean's influence on LHF [ABSTRACT FROM AUTHOR]

We propose an isoneutral eddy transport diagnostic framework and apply it in the Southern Ocean to analyze eddy potential vorticity (PV) transport dynamics. We leverage a thickness‐weighted spatiotemporal scale separation to define the generalized eddy flux. A local sampling method is employed and validated to estimate the transport tensor. We introduce Leonard's decomposition in Large Eddy Simulations (LES) to split the process into the Reynolds and non‐Reynolds parts. We also use the stationary‐transient decomposition. The non‐Reynolds term and the stationary effect contribute significantly to the transport properties. The Reynolds and transient parts have distinctions, that show oceanic mesoscale dynamics are not synergistic in time and spatial scales. We investigate the spatial scale dependence of each part and find that the transient and stationary components of the Reynolds (non‐Reynolds) term have consistent (dissimilar) scale dependence. The scale dependence of the generalized eddy flux and the transport eigenvalues is controlled by the non‐Reynolds term. We further investigate the eigenvalues in terms of the potential enstrophy and anisotropy. The potential enstrophy in the Southern Ocean holistically has a forward cascade on the researched scales. The Reynolds and transient effects efficiently promote the forward enstrophy cascade. The non‐Reynolds and stationary effects stimulate drastically anisotropic eddy transport. The PV gradient barrier mechanism dominates the drastic anisotropic case. When the anisotropy is weak, a special anisotropic case with a locally balanced potential enstrophy budget overwhelms the case of real isotropy. We also introduce several LES concepts, hoping to enlighten any potential ocean modeling and parameterization. Plain Language Summary: In this study, a 2D spatial filter is applied to define the sub‐grid scale signals or approximately mesoscale eddying signals as the deviation from the filtered large‐scale field. Then, a Reynolds temporal average is used to divide the eddying effect into transient and stationary parts. Since we use a spatial filter to separate the scales, the form of the eddy flux needs to be generalized and a non‐Reynolds part emerges. This part is invisible under the traditional Reynolds average (e.g., zonal average, time average, ensemble average). We introduce Leonard's decomposition from large eddy simulations to decompose the generalized eddy flux into non‐Reynolds and Reynolds parts and to categorize the interactions of eddies with various sizes. We utilize a flux‐gradient relationship to link the generalized eddy flux with the large‐scale tracer gradient through a locally sampled anisotropic transport tensor. Both decompositions and the transport tensor help provide insights into the multiscale nature of mesoscale eddy PV transport dynamics and prompt a theoretical understanding of designing potential LES‐related parameterizations for ocean models. Key Points: A spatiotemporal thickness‐weighted isoneutral eddy transport diagnostic frame‐work is proposed and applied in the Southern OceanThe stationary‐transient and Leonard's decomposition are leveraged to analyze the multiscale effects on eddy potential vorticity transportThe Reynolds and transient effects promote the forward enstrophy cascade, and the non‐Reynolds and stationary effects stimulate anisotropy [ABSTRACT FROM AUTHOR]

To address the growing demand for accurate high-resolution ocean wind forcing from the ocean modeling community, we develop a new forcing product, ERA*, by means of a geolocated scatterometer-based correction applied to the European Centre for Medium-range Weather Forecasts (ECMWF) reanalysis or ERA-interim (hereafter referred to as ERAi). This method successfully corrects for local wind vector biases present in the ERAi output globally. Several configurations of the ERA* are tested using complementary scatterometer data [advanced scatterometer (ASCAT)-A/B and oceansat-2 scatterometer (OSCAT)] accumulated over different temporal windows, verified against independent scatterometer data [HY-2A scatterometer (HSCAT)], and evaluated through spectral analysis to assess the geophysical consistency of the new stress equivalent wind fields (U10S). Due to the high quality of the scatterometer U10S, ERA* contains some of the physical processes missing or misrepresented in ERAi. Although the method is highly dependent on sampling, it shows potential, notably in the tropics. Short temporal windows are preferred, to avoid oversmoothing of the U10S fields. Thus, corrections based on increased scatterometer sampling (use of multiple scatterometers) are required to capture the detailed forcing errors. When verified against HSCAT, the ERA* configurations based on multiple scatterometers reduce the vector root-mean-square difference about 10% with respect to that of ERAi. ERA* also shows a significant increase in small-scale true wind variability, observed in the U10S spectral slopes. In particular, the ERA* spectral slopes consistently lay between those of HSCAT and ERAi, but closer to HSCAT, suggesting that ERA* effectively adds spatial scales of about 50 km, substantially smaller than those resolved by global numerical weather prediction (NWP) output over the open ocean (about 150 km). [ABSTRACT FROM AUTHOR]