Description: Current climate models systematically underestimate the strength of oceanic fronts associated with strong western boundary currents, such as the Kuroshio and Gulf Stream Extensions, and have difficulty simulating their positions at the mid-latitude ocean’s western boundaries1. Even with an enhanced grid resolution to resolve ocean mesoscale eddies—energetic circulations with horizontal scales of about a hundred kilometres that strongly interact with the fronts and currents—the bias problem can still persist2; to improve climate models we need a better understanding of the dynamics governing these oceanic frontal regimes. Yet prevailing theories about the western boundary fronts are based on ocean internal dynamics without taking into consideration the intense air–sea feedbacks in these oceanic frontal regions. Here, by focusing on the Kuroshio Extension Jet east of Japan as the direct continuation of the Kuroshio, we show that feedback between ocean mesoscale eddies and the atmosphere (OME-A) is fundamental to the dynamics and control of these energetic currents. Suppressing OME-A feedback in eddy-resolving coupled climate model simulations results in a 20–40 per cent weakening in the Kuroshio Extension Jet. This is because OME-A feedback dominates eddy potential energy destruction, which dissipates more than 70 per cent of the eddy potential energy extracted from the Kuroshio Extension Jet. The absence of OME-A feedback inevitably leads to a reduction in eddy potential energy production in order to balance the energy budget, which results in a weakened mean current. The finding has important implications for improving climate models’ representation of major oceanic fronts, which are essential components in the simulation and prediction of extratropical storms and other extreme events3, 4, 5, 6, as well as in the projection of the effect on these events of climate change.

Description: Sensors, Vol. 18, Pages 712: Saliency Detection and Deep Learning-Based Wildfire Identification in UAV Imagery Sensors doi: 10.3390/s18030712 Authors: Yi Zhao Jiale Ma Xiaohui Li Jie Zhang An unmanned aerial vehicle (UAV) equipped with global positioning systems (GPS) can provide direct georeferenced imagery, mapping an area with high resolution. So far, the major difficulty in wildfire image classification is the lack of unified identification marks, the fire features of color, shape, texture (smoke, flame, or both) and background can vary significantly from one scene to another. Deep learning (e.g., DCNN for Deep Convolutional Neural Network) is very effective in high-level feature learning, however, a substantial amount of training images dataset is obligatory in optimizing its weights value and coefficients. In this work, we proposed a new saliency detection algorithm for fast location and segmentation of core fire area in aerial images. As the proposed method can effectively avoid feature loss caused by direct resizing; it is used in data augmentation and formation of a standard fire image dataset ‘UAV_Fire’. A 15-layered self-learning DCNN architecture named ‘Fire_Net’ is then presented as a self-learning fire feature exactor and classifier. We evaluated different architectures and several key parameters (drop out ratio, batch size, etc.) of the DCNN model regarding its validation accuracy. The proposed architecture outperformed previous methods by achieving an overall accuracy of 98%. Furthermore, ‘Fire_Net’ guarantied an average processing speed of 41.5 ms per image for real-time wildfire inspection. To demonstrate its practical utility, Fire_Net is tested on 40 sampled images in wildfire news reports and all of them have been accurately identified.