Browse

You are looking at 1 - 10 of 5,081 items for :

  • Journal of Atmospheric and Oceanic Technology x
  • Refine by Access: All Content x
Clear All
Travis Miles, Wayne Slade, and Scott Glenn

Abstract

Suspended particle size and concentration are critical parameters that are necessary to understand water quality, sediment dynamics, carbon flux, and ecosystem dynamics, among other ocean processes. In this study we detail the integration of a Sequoia Scientific, Inc., Laser In Situ Scattering and Transmissometry (LISST) sensor into a Teledyne Webb Research Slocum autonomous underwater glider. These sensors are capable of measuring particle size, concentration, and beam attenuation by particles in size ranges from 1.00 to 500 μm at a resolution of 1 Hz. The combination of these two technologies provides the unique opportunity to measure particle characteristics persistently at specific locations or to survey regional domains from a single profiling sensor. In this study we present the sensor integration framework, detail quality assurance and control procedures, and provide a case study of storm-driven sediment resuspension and transport. Specifically, Rutgers glider RU28 was deployed with an integrated LISST-Glider for 18 days in September of 2017. During this period, it sampled the nearshore environment off coastal New Jersey, capturing full water column sediment resuspension during a coastal storm event. A novel method for in situ background corrections is demonstrated and used to mitigate long-term biofouling of the sensor windows. In addition, we present a method for removing schlieren-contaminated time periods utilizing coincident conductivity temperature and depth, fluorometer, and optical backscatter data. The combination of LISST sensors and autonomous platforms has the potential to revolutionize our ability to capture suspended particle characteristics throughout the world’s oceans.

Open access
Zaid R. Al-Attabi, George Voulgaris, and Daniel C. Conley

Abstract

An examination of the applicability and accuracy of the empirical wave inversion method in the presence of swell waves is presented. The ability of the method to invert Doppler spectra to wave directional spectra and bulk wave parameters is investigated using one-month data from a 12 MHz WERA High Frequency (HF) radar system and in-situ data from a wave buoy. Three different swell inversion models are evaluated: LPM (Lipa et al. 1981), WFG (Wang et al. 2016) and EMP, an empirical approach introduced in this study. The swell inversions were carried out using two different scenarios: (1) a single beam from a single radar site and two beams from a single radar site, and (2) two beams from two sites (a single beam per site) intersecting each other at the buoy location. The LPM method utilized using two beams from two different sites was found to provide the best estimations of swell parameters (swell height RMS error 0.24m) and showed a good correlation with the partitioned swell in-situ values. For the wind wave inversion, the empirical method presented here is used with an empirical coefficient of 0.3 which seems to be suitable for universal application for all radar operating frequencies. The inverted swell parameters are used to create a swell spectrum which is combined with the inverted wind wave spectrum to create a full directional wave spectrum. The wave inversion method presented in this study although empirical does not require calibration with in situ data and can be applied to any beam forming system and operating frequency.

Restricted access
Ryan Lagerquist, David Turner, Imme Ebert-Uphoff, Jebb Stewart, and Venita Hagerty

Abstract

This paper describes the development of U-net++ models, a type of neural network that performs deep learning, to emulate the shortwave Rapid Radiative-transfer Model (RRTM). The goal is to emulate the RRTM accurately in a small fraction of the computing time, creating a U-net++ that could be used as a parameterization in numerical weather prediction (NWP). Target variables are surface downwelling flux, top-of-atmosphere upwelling flux (FupTOA), net flux, and a profile of radiative-heating rates. We have devised several ways to make the U-net++ models knowledge-guided, recently identified as a key priority in machine learning (ML) applications to the geosciences. We conduct two experiments to find the best U-net++ configurations. In Experiment 1, we train on non-tropical sites and test on tropical sites, to assess extreme spatial generalization. In Experiment 2, we train on sites from all regions and test on different sites from all regions, with the goal of creating the best possible model for use in NWP. The selected model from Experiment 1 shows impressive skill on the tropical testing sites, except four notable deficiencies: large bias and error for heating rate in the upper stratosphere, unreliable FupTOA for profiles with single-layer liquid cloud, large heating-rate bias in the mid-troposphere for profiles with multi-layer liquid cloud, and negative bias at lowzenith angles for all flux components and tropospheric heating rates. The selected model from Experiment 2 corrects all but the first deficiency, and both models run ~104 times faster than the RRTM. Our code is available publicly.

Restricted access
Mark Curtis, Sandy Dance, Valentin Louf, and Alain Protat

Abstract

For mechanically scanning weather radars, precise pointing of the antenna is a key factor in ensuring accurate observation of the atmosphere at far range. Since operational radars typically scan the atmosphere using a series of 360° sweeps at fixed elevation angles, level scanning during azimuthal rotation is also important, but often not actively monitored after installation.

One method of gauging pointing accuracy of a radar is to use solar interference which occurs as the antenna sweeps past the sun. By comparing the observed position of the sun with its known position, an estimate of pointing error in both elevation and azimuth can be obtained. A basic model for this error assumes that the radar sweep is perfectly level and that biases in elevation are therefore independent of azimuth. We extend this model to allow for the possibility that the plane of rotation may not be level. Consequently, the direction and severity of tilt may be diagnosed in addition to any constant error in elevation and azimuth pointing.

The extended model was applied to a subset of radars from the Australian weather radar network resulting in the discovery of several out of level radars. One radar, Captains Flat near Canberra, showed a severe tilt of 0.81° prompting inspection by a technician. This revealed that mounting studs on the pedestal of the radar tower were badly worn and loose. Correction of this issue resolved the tilt component of the diagnosed elevation error and prevented further mechanical damage to the instrument.

Restricted access
F. Joseph Turk, Ramon Padullés, Estel Cardellach, Chi O. Ao, Kuo-Nung Wang, David D. Morabito, Manuel de la Torre Juarez, Mayra Oyola, Svetla Hristova-Veleva, and J. David Neelin

Abstract

Observationally, a major source of uncertainty in evaluation of climate models arises from the difficulty in obtaining globally distributed, fine scale profiles of temperature, pressure and water vapor, that probe through convective precipitating clouds, from the boundary layer to the upper levels of the free troposphere. In this manuscript, a two-year analysis of data from the Radio Occultations through Heavy Precipitation (ROHP) polarimetric RO demonstration mission onboard the Spanish PAZ spacecraft is presented. ROHP measures the difference in the differential propagation phase delay (Δ𝜙) between two orthogonal polarization receive states that is induced from the presence of non-spherically shaped hydrometeors along the Global Navigation Satellite System (GNSS) propagation path, complementing the standard RO thermodynamic profile. Since Δφ is a net path-accumulated depolarization and does not resolve the precipitation structure along the propagation path, orbital coincidences between ROHP and the Global Precipitation Measurement (GPM) constellation passive MW radiometers are identified to provides three-dimensional precipitation context to the RO thermodynamic profile. Passive MW-derived precipitation profiles are used to simulate the Δφ along the ROHP propagation paths. Comparison between the simulated and observed Δφ are indicative of the ability of ROHP to detect threshold levels of ray path-averaged condensed water content, as well as to suggest possible inferences on the average ice phase hydrometeor non-sphericity. The use of the polarimetric RO vertical structure is demonstrated as a means to condition the lower tropospheric humidity by the top-most height of the associated convective cloud structure.

Restricted access
Ali Tokay, Annakaisa von Lerber, Claire Pettersen, Mark S. Kulie, Dmitri N. Moisseev, and David B. Wolff

Abstract

Performance of the Precipitation Imaging Package (PIP) for estimating the snow water equivalent (SWE) is evaluated through a comparative study with the collocated National Oceanic and Atmospheric Administration National Weather Service snow stake field measurements. The PIP together with a vertically pointing radar, a weighing bucket gauge, and a laser-optical disdrometer was deployed at the NWS Marquette, Michigan office building for a long-term field study supported by the National Aeronautics and Space Administration’s Global Precipitation Measurement mission Ground Validation program. The site was also equipped with a weather station. During the 2017-18 winter, the PIP functioned nearly uninterrupted at frigid temperatures accumulating 2345.8 mm of geometric snow depth over a total of 499 hours. This long record consists of 30 events, and the PIP-retrieved and snow stake field measured SWE differed less than 15% in every event. Two of the major events with the longest duration and the highest accumulation are examined in detail. The particle mass with a given diameter was much lower during a shallow, colder, uniform lake-effect event than in the deep, less cold, and variable synoptic event. This study demonstrated that the PIP is a robust instrument for operational use, and is reliable for deriving the bulk properties of falling snow.

Restricted access
Hans van Haren, Roel Bakker, Yvo Witte, Martin Laan, and Johan van Heerwaarden

Abstract

The redistribution of matter in the deep-sea depends on water-flow currents and turbulent exchange, for which breaking internal waves are an important source. As internal waves and turbulence are essentially three-dimensional ‘3D’, their dynamical development should ideally be studied in a volume of seawater. However, this is seldom done in the ocean where 1D-observations along a single vertical line are already difficult. We present the design, construction and successful deployment of a half-cubic-hectometer (480,000 m3) 3D-T mooring array holding 2925 high-resolution temperature sensors to study weakly density-stratified waters of the 2500-m deep Western Mediterranean. The stand-alone array samples temperature at a rate of 0.5 Hz, with precision <0.5 mK, noise level <0.1 mK and expected endurance of 3 years. The independent sensors are synchronized inductively every 4 h to a single standard clock. The array consists of 45 vertical lines 125 m long, at 9.5 m horizontally from their nearest neighbor. Each line is held under tension of 1.3 kN by a buoyancy element that is released chemically one week after deployment. All fold-up lines are attached to a grid of cables that is tensioned in a 70 m diameter ring of steel tubes. The array is build-up in harbor-waters, with air filling the steel tubes for floatation. The flat-form array is towed to the mooring site under favorable sea-state conditions. By opening valves in the steel tubes, the array is sunk and its free-fall is controlled by a custom-made drag-parachute reducing the average sinking speed to 1.3 m s-1 and providing smooth horizontal landing on the flat seafloor.

Restricted access
Zepei Wu, Shuo Liu, Delong Zhao, Ling Yang, Zixin Xu, Zhipeng Yang, Dantong Liu, Tao Liu, Yan Ding, Wei Zhou, Hui He, Mengyu Huang, Ruijie Li, and Deping Ding

Abstract

Cloud particles have different shapes in the atmosphere. Research on cloud particle shapes plays an important role in analyzing the growth of ice crystals and the cloud microphysics. To achieve an accurate and efficient classification algorithm on ice crystal images, this study uses image-based morphological processing and principal component analysis, to extract features of images and apply intelligent classification algorithms for the Cloud Particle Imager (CPI). Currently, there are mainly two types of ice-crystal classification methods: one is the mode parameterization scheme, and the other is the artificial intelligence model. Combined with data feature extraction, the dataset was tested on ten types of classifiers, and the highest average accuracy was 99.07%. The fastest processing speed of the real-time data processing test was 2,000 images/s. In actual application, the algorithm should consider the processing speed, because the images are in the order of millions. Therefore, a support vector machine (SVM) classifier was used in this study. The SVM-based optimization algorithm can classify ice crystals into nine classes with an average accuracy of 95%, blurred frame accuracy of 100%, with a processing speed of 2,000 images/s. This method has a relatively high accuracy and faster classification processing speed than the classic neural network model. The new method could be also applied in physical parameter analysis of cloud microphysics.

Restricted access
Seth F. Zippel, J. Thomas Farrar, Christopher J. Zappa, Una Miller, Louis St. Laurent, Takashi Ijichi, Robert A. Weller, Leah McRaven, Sven Nylund, and Deborah Le Bel

Abstract

Upper-ocean turbulence is central to the exchanges of heat, momentum, and gasses across the air/sea interface, and therefore plays a large role in weather and climate. Current understanding of upper-ocean mixing is lacking, often leading models to misrepresent mixed-layer depths and sea surface temperature. In part, progress has been limited due to the difficulty of measuring turbulence from fixed moorings which can simultaneously measure surface fluxes and upper-ocean stratification over long time periods. Here we introduce a direct wavenumber method for measuring Turbulent Kinetic Energy (TKE) dissipation rates, ϵ, from long-enduring moorings using pulse-coherent ADCPs. We discuss optimal programming of the ADCPs, a robust mechanical design for use on a mooring to maximize data return, and data processing techniques including phase-ambiguity unwrapping, spectral analysis, and a correction for instrument response. The method was used in the Salinity Processes Upper-ocean Regional Study (SPURS) to collect two year-long data sets. We find the mooring-derived TKE dissipation rates compare favorably to estimates made nearby from a microstructure shear probe mounted to a glider during its two separate two-week missions for O (10−8) ≤ ϵO (10−5) m2 s−3. Periods of disagreement between turbulence estimates from the two platforms coincide with differences in vertical temperature profiles, which may indicate that barrier layers can substantially modulate upper-ocean turbulence over horizontal scales of 1-10 km. We also find that dissipation estimates from two different moorings at 12.5 m, and at 7 m are in agreement with the surface buoyancy flux during periods of strong nighttime convection, consistent with classic boundary layer theory.

Open access
Joaquin Cuomo and V. Chandrasekar

Abstract

Nowcasting based on weather radar uses the current and past observations to make estimations of future radar echoes. There are many types of operationally deployed nowcasting systems, but none of them are currently based on deep learning, despite it being an active area of research in the last few years. This paper explores deep learning models as alternatives to current methods by proposing different architectures and comparing them against some operational nowcasting systems. The methods proposed here, harnessing residual convolutional encoder-decoder architectures, reach a level of performance expected of current systems and in certain scenarios can even outperform them. Finally, some of the potential drawbacks of using deep learning are analyzed. No decay in the performance on a different geographical area from where the models were trained was found. No edge or checkerboard artifact, common in convolutional operations, was found that affects the nowcasting metrics.

Restricted access