Coastal mangrove forests provide numerous ecosystem services, which can be disrupted by natural disturbances, mainly hurricanes. Canopy height (CH) is a key parameter for estimating carbon storage. Airborne Light Detection and Ranging (LiDAR) is widely viewed as the most accurate method for estimating CH but data are often limited in spatial coverage and are not readily available for rapid impact assessment after hurricane events. Hence, we evaluated the use of systematically acquired space-based Synthetic Aperture Radar (SAR) and optical observations with airborne LiDAR to predict CH across expansive mangrove areas in South Florida that were severely impacted by Category 3 Hurricane Irma in 2017. We used pre- and post-Irma LiDAR-derived canopy height models (CHMs) to train Random Forest regression models that used features of Sentinel-1 SAR time series, Landsat-8 optical, and classified mangrove maps. We evaluated (1) spatial transfer learning to predict regional CH for both time periods and (2) temporal transfer learning coupled with species-specific error correction models to predict post-Irma CH using models trained by pre-Irma data. Model performance of SAR and optical data differed with time period and across height classes. For spatial transfer, SAR data models achieved higher accuracy than optical models for post-Irma, while the opposite was the case for the pre-Irma period. For temporal transfer, SAR models were more accurate for tall trees (>10 m) but optical models were more accurate for short trees. By fusing data of both sensors, spatial and temporal transfer learning achieved the root mean square errors (RMSEs) of 1.9 m and 1.7 m, respectively, for absolute CH. Predicted CH losses were comparable with LiDAR-derived reference values across height and species classes. Spatial and temporal transfer learning techniques applied to readily available spaceborne satellite data can enable conservation managers to assess the impacts of disturbances on regional coastal ecosystems efficiently and within a practical timeframe after a disturbance event.

Planar wall plumes are gravity-driven flows where a fluid of lower (or higher) density than the ambient rises (or lowers) along a vertical or inclined wall. This study investigates planar wall plumes at five different wall slopes, ranging from a vertical wall (θ=90°) to a shallow inclination of θ=3°, using highly resolved direct numerical simulations. The three-dimensional turbulent structure of these supercritical flows is investigated in detail along with the streamwise evolution of the depth-averaged quantities. Simulations were performed in very large domains in order to focus attention on the behavior of the plumes in the near self-similar state, far downstream of the inlet. In the self-similar state, key quantities such as the entrainment rate, the basal drag coefficient, the Richardson number (or equivalently the Froude number), and the shape factors reach constant values, which dependent only on the slope. The present simulations, along with earlier results for subcritical currents at shallower slopes, provide a complete description of this dependency.

Infrasound propagation through various atmospheric conditions and interaction with environmental factors induce highly non-linear and non-stationary effects that make it difficult to extract reliable attributes for classification. We present featureless classification results on the Library of Typical Infrasonic Signals (LOTIS) using several deep learning techniques, including long short-term memory (LSTM), self-normalizing (SNN), and fully convolutional (FCN) networks with statistical analysis to establish significantly superior models. In general, the deep classifiers achieve near-perfect classification accuracies on the four classes of infrasonic events including mountain associated waves, microbaroms, internal atmospheric gravity waves and volcanic eruptions. Our results provide evidence that deep neural network architectures be considered the leading candidate for classifying infrasound waveforms which can directly benefit applications that seek to identify infrasonic events such as severe weather forecasting, military threat detection, natural disaster early warning systems, and nuclear weapons monitoring.

Coastal mangrove forests provide numerous ecosystem services, including carbon uptake and storage, which can be disrupted by natural disturbances, mainly hurricanes. Canopy height (CH) is a key parameter for evaluating the carbon storage of mangroves. Still, it cannot be measured at a regional scale due to mangrove areas' vast extent and inaccessibility. Remote sensing observations assist in better understanding and quantifying the impacts of natural disturbances at the ecosystems and landscape scales. Airborne LiDAR is widely viewed as the most accurate sensor for estimating CH, but it is expensive, limited in spatial coverage, and not readily available for rapid assessment of impacts after natural disturbances. Conversely, space-based Synthetic Aperture Radar (SAR) and optical remote sensing observations are available at higher temporal frequency, but coarser in spatial resolution. In this study, we developed multi-temporal predictions of CH across the expansive mangrove areas in the Everglades National Park and the Ten Thousand Islands regions of South Florida, which were severely impacted by Hurricane Irma, a category-3 hurricane that made landfall on September 10, 2017. We used canopy height models (CHM) derived from Goddard's Light Detection and Ranging (LiDAR), Hyperspectral & Thermal Imager (G-LiHT) data to train machine/deep learning models using Sentinel-1 SAR time series and Landsat-8 optical data as feature variables. Estimated CH derived from LiDAR data and feature variables were acquired across pre- and post-Irma time periods, enabling us to conduct (1) spatial transfer learning to predict CH for the entire extent of mangrove areas across four different species for each period and (2) temporal transfer learning by training a model with pre-Irma data and predicting the post-Irma CH. We achieved an optimal accuracy of 1.9 m and 1.8 m of RMSE for spatial transfer learning on pre-Irma and post-Irma, and 1.7 m for temporal transfer learning using a combination of SAR, optical, and mangrove classification features. Predicted CH losses occurred across 62% of South Florida mangrove areas, with tall mangroves experiencing the most severe CH losses. This study establishes correlations between high precision airborne LiDAR data and more frequently available orbital SAR and optical data for reliable application of calibrated models to estimate regional and multi-temporal CH of mangrove areas.

This study maps canopy height (CH) of mangrove forests using multi-sensor remote-sensing observations with deep learning methods. The multi-sensor data include satellite observations of Synthetic Aperture Radar (SAR) and Landsat-8 optical, and airborne Light Detection and Ranging (LiDAR) data. The study focuses on the south Florida Everglades mangrove forests, which experienced severe damage caused by Hurricane Irma in September 2017. We developed two innovative neural network models to predict CH from SAR and optical data. The models mapped mangrove CH the year following Hurricane Irma with an RMSE of about 3 m. The innovative deep learning methods achieves better accuracy than other traditional machine learning models.