For the detection of marine ship objects in radar images, large-scale networks based on deep learning are difficult to be deployed on existing radar-equipped devices. This paper proposes a lightweight convolutional neural network, LiraNet, which combines the idea of dense connections, residual connections and group convolution, including stem blocks and extractor modules. The designed stem block uses a series of small convolutions to extract the input image features, and the extractor network adopts the designed two-way dense connection module, which further reduces the network operation complexity. Mounting LiraNet on the object detection framework Darknet, this paper proposes Lira-you only look once (Lira-YOLO), a lightweight model for ship detection in radar images, which can easily be deployed on the mobile devices. Lira-YOLO's prediction module uses a two-layer YOLO prediction layer and adds a residual module for better feature delivery. At the same time, in order to fully verify the performance of the model, mini-RD, a lightweight distance Doppler domain radar images dataset, is constructed. Experiments show that the network complexity of Lira-YOLO is low, being only 2.980 Bflops, and the parameter quantity is smaller, which is only 4.3 MB. The mean average precision (mAP) indicators on the mini-RD and SAR ship detection dataset (SSDD) reach 83.21% and 85.46%, respectively, which is comparable to the tiny-YOLOv3. Lira-YOLO has achieved a good detection accuracy with less memory and computational cost.

Munk (1966) showed that the deep (1000-3000 m) vertical temperature profile is consistent with a one-dimensional vertical advectiondiffusion balance, with a constant upwelling and an interior diapycnal diffusivity of O(10 −4 ) m 2 s −1 . However, typical observed diffusivities in the interior are O(10 −5 ) m 2 s −1 . Recent work suggested that the deep stratification is set by Southern Ocean (SO) isopycnal slopes, fixed by SO eddies, that communicate the surface outcrop positions to the deep ocean. It is shown here, using an idealized ocean general circulation model, that SO eddies alone cannot lead to the observed exponential temperature profile, and that interior mixing must contribute. Strong diapycnal mixing concentrated near the ocean boundaries is shown to be balanced locally by upwelling. A one-dimensional Munk-like balance in these boundary mixing areas, although with much larger mixing and upwelling, leads to an exponential deep temperature stratification, which propagates via isopycnal mixing to the ocean interior. The exponential profile is robust to vertical variations in the vertical velocity, and persists despite the observed weak interior diapycnal mixing. Southern Ocean eddies link the surface water mass transformation by air-sea fluxes with the deep stratification, but the eddies do not determine the stratification itself. These results reconcile the observed exponential interior deep temperature stratification, the weak diapycnal diffusivity observed in tracer release experiments, and the role of Southern Ocean dynamics.

The long‐standing paradigm for the large‐scale time‐averaged ocean circulation in the world oceans includes intensified currents at the western boundary, and a much slower interior flow elsewhere. However, poleward deep boundary currents in the eastern limits of the deep South Pacific, Atlantic, and Indian Oceans have been observed repeatedly at 2‐ to 4‐km depth. They carry up to a third of the total deep ocean transports, implying a significant role in climate, yet their dynamics are still not well understood. Here we develop a theoretical understanding for these currents, using a hierarchy of realistic and idealized models, focusing on the South Pacific Ocean. The deep eastern boundary current there is shown to be driven by a traditional interior balance together with a narrow boundary layer scale near the eastern boundary, which exists only when stratification and topography are both included. A simplified semianalytical vorticity model is developed for the deep eastern boundary current.