FALL 2020
Wednesdays at 3:00 pm, Seminar Room SLAB 103 / Virtual SLAB 103
(unless stated otherwise)
Aug 19: NO SEMINAR
Aug 26: NO SEMINAR
Sep 02: Heather Hunter
Department of Ocean Sciences, RSMAS
(one-hour AMP student seminar)
Deep Learning for Object Classification and Detection in
Satellite-Based Synthetic Aperture Radar Images
Recording Available at COMPASS ON DEMAND
Object classification and detection in satellite imagery are important tasks for a wide range of applications, from traffic monitoring over land and sea to monitoring of environmental concerns, such as the spatiotemporal characteristics of sea ice and oil spills. Owing to its ability to learn rich hierarchies of discriminative features, deep learning (DL) is emerging as a powerful tool to accomplish these tasks in both optical and synthetic aperture radar (SAR) imagery. Unfortunately, DL algorithms require vast amounts of annotated training data to generalize well, and collecting a suitable amount of satellite imagery is typically not feasible. Further, applications of DL to SAR face additional challenges because SAR's imaging mechanism depends on the electric permittivity of the imaged scene, its surface roughness, and the nuances of any physical structures present. Consequently, unlike optical images, SAR images of the same scene look substantially different under different weather conditions and imaging angles, which precludes the ability to generate comprehensive training data for a well-generalized model. In this presentation, I will discuss our investigation of several approaches for mitigating the training data limitation for SAR object classification, including transfer learning from multiple domains. Specifically, I will address the question of whether artificially generated SAR data can be used to supplement small amounts of real-world data. I will show how we use these classifiers as the basis for SAR object detection models, as well as how the features learned by each classifier impact those detection models. Finally, I will address the question of whether we can develop a SAR object classifier that learns domain invariant features between artificially generated data and small amounts of real-world data.
Sep 09: NO SEMINAR
Sep 16: STUDENT SEMINARS (rescheduled from Spring 2020)
Wei-Ming Tsai (ATM)
Composite Views of Tropical Convective Evolution Dependence on Aggregation
Using a Gross Moist Stability Framework
The clustering of tropical clouds may strongly modulate radiative forcing, the hydrological cycle, and large-scale circulations. Most studies on convective aggregation are conducted under idealized modeling frameworks, whereas observation-based ones are few and rarely discuss dynamic features. This study aims to explore the dependence of convective evolution on convective aggregation from observations and the potential role of aggregated convection in moist dynamics. 5-year multiple gridded datasets of a 3-hourly resolution, including satellite and reanalysis data, are used to identify convective events in 5°×5° boxes spanning over the tropical Indian Ocean (50°E-90°E, 10°S-10°N). Convective events are determined when the observed box-averaged precipitation maximum exceeds 5 mm/d and is centered in a 4-day time window. The degree of aggregation is quantified by the Simple Convective Aggregation Index (SCAI) based on connected Infrared pixels (BT<240K). Results indicate that given the same box-averaged precipitation intensity, a more aggregated status (smaller SCAI) shows (1) a drier environment with a larger horizontal gradient of moisture and (2) a more bottom-heavy structure of large-scale upwelling, with a larger portion of low clouds and slightly higher cloud top maximum. Gross moist stability (GMS) acts as a useful estimate of atmospheric stability and is highly associated with structures of large-scale circulation and moisture. Based on the above features with the GMS framework, it is proposed that the evolution of aggregated convection is related to importing moist static energy (MSE) into the atmosphere, while scattered convection shows nearly zero or slight export in MSE, suggesting the potential role of aggregation in MSE transport.
Luna Hiron (MPO)
Study of Ageostrophy During Strong, Nonlinear Eddy-Front Interaction
in the Gulf of Mexico
The Loop Current (LC) system has long been assumed to be close to geostrophic balance despite its strong flow and the development of large meanders and strong frontal eddies during unstable phases. The region between the LC meanders and its frontal eddies was shown to have high Rossby numbers indicating nonlinearity; however, the effect of the nonlinear term on the flow has not been studied so far. In this study, the ageostrophy of the LC meanders is assessed using a high-resolution numerical model and geostrophic velocities from altimetry. A formula to compute the radius of curvature of the flow from the velocity field is also presented. The results indicate that during strong meandering, especially during the LC shedding in the presence of frontal eddies, the centrifugal force becomes as important as the Coriolis force and the pressure-gradient force; LC meanders are in gradient-wind balance. The centrifugal force modulates the balance, and thus modifies the flow speed, resulting in a subgeostrophic flow in the LC meander trough around the LCFE and supergeostrophic flow in the LC meander crest. The same pattern is found when correcting the altimetry geostrophic velocities to account for the effect of the centrifugal force, but with smaller magnitudes. The ageostrophic ratio associated with the centrifugal force in the cyclonic and anticyclonic meanders was of ~46% in the model and ~28% in the altimetry dataset. Thus, the ageostrophic velocity is an important component of the LC flow, and, therefore, cannot be neglected when studying the LC system.
Sep 23: Greg Koman
Department of Ocean Sciences, RSMAS
(one-hour MPO student seminar)
Circulation and Overturning in the Eastern North Atlantic Subpolar Gyre
Zoom Recording Available at COMPASS ON DEMAND
As part of the multi-national Overturning in the Subpolar North Atlantic Program (OSNAP), this presentation will reveal transport estimates from the first continuous multi-year observations of the East Reykjanes Ridge Current, and new estimates of the North Atlantic Current, from the OSNAP mooring line near 58°N. Together with satellite altimetry and Argo profile and drift data, the mean transport, synoptic variability, water mass properties and circulation pathways of the region are examined. Including recent results from other nearby locations, mass budgets and overturning estimates are created to gain a better understanding of the exchanges in the eastern subpolar gyre. Results find that nearly half of the Atlantic Meridional Overturning Circulation (AMOC) occurs in the eastern subpolar gyre due to progressive water mass modification as warm and salty waters gradually transform into the denser waters that constitute the AMOC's lower limb.
Sep 30: BEST STUDENT SEMINAR AWARD CEREMONY
for Academic Year 2019-2020 (rescheduled from Spring 2020)
Oct 07: Dr. Victoria Treadaway
Department of Atmospheric Sciences, RSMAS
The Role of Deep Convection on Upper Tropospheric Chemical Composition
Recording Available at COMPASS ON DEMAND
Rapid transport by deep convection is an important mechanism for delivering surface emissions of trace species to the upper troposphere (UT) which can impact atmospheric composition far from the boundary layer origin. Over the past decade, different atmospheric field campaigns have observed how convection lofts trace species to the UT. This presentation discusses two campaigns that demonstrated this process. The first campaign, the Deep Convective Clouds and Chemistry Experiment (DC3), was in summer 2012. Observations extended from the surface to 13 km over the central and eastern United States. Formic acid (HFo), a soluble trace gas, measurements are presented for a flight where we observed 700 ppt near convective outflow, ~300 ppt above surrounding UT air. This observation was notable as HFo is traditionally considered scavenged in deep convection but if lofted the lifetime increases from 3 to 20+ days. Possible explanations for this were investigated with the Weather Research and Forecasting Model v. 3.7 coupled with chemistry (WRF-Chem). Based on WRFChem, aqueous chemistry played a minor role in the UT elevated HFo compared to the lofting of a large, unknown surface source. The second campaign, Pacific Oxidants, Sulfur, Ice, Dehydration, and cONvection (POSIDON), was in October 2016 and sampling focused from 14-18 km over the Western Pacific Ocean. POSIDON whole air sample measurements serve as an example of the long-range transport possible after deep convective lofting. Nonmethane hydrocarbons were elevated 60-400% above background mixing ratios. Short-lived chlorinated organic species were elevated above background mixing ratios, dichloromethane (110%), 1,2dichloroethane (190%), and chloroform (75%), a chemical composition indicative of biofuel / biomass burning and industrial emissions. Back-trajectories and chemical fingerprinting show that convectively lofted Asian emissions, including shortlived chlorocarbons, can be transported to the remote tropical UT.
Oct 14: Valeria Donets
Department of Atmospheric Sciences, RSMAS
(one-hour ATM student seminar)
Chemical Composition of the Tropical Pacific Troposphere
and Tropical Tropopause Layer
Recording Available at COMPASS ON DEMAND
The tropical Pacific Ocean is an important region of the world's atmosphere that has significant implications for atmospheric chemistry, climate, and aspects of atmospheric dynamics due to its role as the primary gateway for tropospheric air into the stratosphere. Studies have shown that even mild changes in the chemistry and amount of water vapor reaching the stratosphere can significantly affect the climate by absorbing thermal energy radiating from the surface. The stratosphere also contains high concentrations of ozone that at these altitudes filters out ultraviolet radiation. The latter is of particular interest to atmospheric chemists, due to lagging recovery rates of the ozone layer, which were estimated based on phasing out of ozone depleting substances (ODPs), i.e. chlorofluorocarbons (CFCs), hydrochlorofluorocarbons (HCFCs), etc. The observed lag had been linked to natural emissions of halogenated hydrocarbons, particularly prevalent in tropical latitudes, as well as other short-lived (<6 months) halogenated substances with a mixture of natural and anthropogenic sources. Several other chemical species can also provide reservoirs of reactive radicals in the Tropical Tropopause Layer (TTL) and sources for Stratospheric Aerosol formation. For these reasons, understanding chemical composition of the tropical Pacific troposphere and the TTL is of crucial importance, as this allows not only to keep track of current trends but also sets up a framework for prediction of future changes. However, despite its importance, the chemical composition of the TTL, in particular, remains poorly characterized, as very few research platforms can reach these altitudes for in-situ data collection, while remote sensing techniques are limited to a fairly short number of species. This presentation focuses on chemical data that has been collected in the tropical Pacific troposphere and the TTL during a course of seven airborne campaigns conducted between 2006 and 2015, under different seasonal conditions. I will discuss previously published work featuring horizontal transport of anthropogenic species into the tropical latitudes of the western Pacific during boreal winter, followed by analysis of the chemical composition of the TTL and the lower stratosphere over the eastern Pacific. We investigate chemical distributions of a series of trace gases with different atmospheric lifetimes and source origins as a function of air-mass history, based on HYSPLIT trajectory calculations. We also investigate how seasons affect source distributions of various tracers with changing circulation patterns, and how these seasonal changes translate into observed chemical trends.
Oct 21: Andrew Smith
Department of Ocean Sciences, RSMAS
(one-hour MPO student seminar)
Sea Surface Structure Mediation of Turbulent Kinetic Energy and Dissipation
in Near-Surface Boundary Layers
The air-sea interface is a fundamentally important dynamic boundary in regards to the balance and physical exchange of energy between the atmosphere and ocean, accomplished by mechanical and thermodynamic processes that transfer, redistribute or dissipate the energy. Fluxes of heat, momentum, and gases at the interface contribute energy to large-scale phenomena such as convection, hurricanes, wave growth, surface ocean currents, mixing, and climate. These fluxes contribute to the production and vertical redistribution of turbulent kinetic energy (TKE) in near-surface layers, and thus its local dissipation. The TKE budget is explicitly linked to the vertical structure and thermodynamic stability of near-surface boundary layers, which can be modified by the evolution and structure of the sea surface and its collapse through wave breaking. Wave breaking also entrains air into groups of bubbles, which can act as an additional vector for air-sea material and energy transfers and can interact with sub-surface turbulent flow at radius-dependent scales. Although TKE and its dissipation have been experimentally, numerically, and observationally investigated since the 1960s, there remains disagreement as to how production and dissipation terms in the TKE budget balance under different conditions, a lack of turbulence and dissipation observations in hurricane-force conditions, and difficulty in measuring TKE dissipation rate robustly without arbitrary selection of the appropriate contributing turbulent scales. Therefore, we present findings from field data and laboratory experiments to address these challenges; specifically, we compare turbulence and dissipation measured in non-hurricane and explicitly hurricane atmospheric surface layers to parameterize the TKE dissipation rate and close observed imbalance in production and dissipation. Laboratory measurements of TKE and its dissipation rate are made in the presence of wave breaking induced bubble populations, whose distribution we parameterize in terms of the wave breaking intensity. Ultimately, the TKE and its dissipation rate are explicitly linked to evolution of the sea (water) surface structure in both open-ocean wind-sea and swell conditions as well as high-wind breaking-wave laboratory conditions.
Oct 28: NO SEMINAR
Nov 04: Rebecca Evans
Department of Atmospheric Sciences, RSMAS
(one-hour MPO student seminar)
Diurnal Oscillations in Tropical Cyclones and Their Influence on Gravity Waves
in Linear and Nonlinear Models
Zoom Recording Available at COMPASS ON DEMAND
The diurnal cycle of solar radiation has recently been shown to influence all manner of features in tropical cyclones (TCs), from structure and intensity to modulating the outflow and production of convective gravity waves. Diurnally varying heating in the cirrus canopy and eyewall has been suggested to cause much of this variability. Satellite observations have confirmed the presence of a pulse in convection that forms in the cirrus canopy around the time of local sunset and propagates outwards overnight. Recent work on this subject has used models to explore the stability regions of waves produced in response to diurnal heating in TCs. Other work on the diurnal cycle in TCs has illustrated that diurnal variability is ubiquitous throughout TCs, with diurnal modulation of the area of the cirrus canopy, storm intensity, amount of rainfall, and strength of surface inflow. The nature of the wave-like responses to the diurnal heating remains unclear, and more generally the field is absent of a holistic quantitative overview of the exact timings and magnitudes of diurnal oscillations in TCs. In the work presented here, we aim to address these questions in modeling frameworks, in order to test the underlying mechanisms causing the diurnal variability. Firstly, we explore the diurnal modulation of waves in TCs in a non-hydrostatic, linear model forced with realistic diurnal heating, with the aim of addressing whether the nighttime pulse in the cirrus canopy is a radiating or balanced wave response. Secondly, we explore diurnal variability in the troposphere of TCs in general using two high-resolution WRF simulations. Diurnal modes of variability are studied using Fourier filtering, and Empirical Orthogonal Function analysis to provide a three-dimensional view of the spatio-temporal evolution of diurnal variability in TCs. Finally, having discussed the influence of the diurnal cycle on a TC in the troposphere, we will extend our focus to the overlying environment of the stratosphere. Specifically, how diurnal modulation of the production of convective GWs, as well as diurnal modulation in the outflow, structure, and intensity may influence propagation of gravity waves from TCs in to the stratosphere.
Nov 11: Dr. Cathleen Jones
NASA Jet Propulsion Laboratory, Pasadena, California
Musings on the Past, Present, and Future of Oil Spill Studies With L-Band SAR
Zoom Recording Available at COMPASS ON DEMAND
It has long been known that oil damps the capillary and gravity-capillary waves on the ocean surface, smoothing the small-scale roughness that causes radar backscatter. However, until the Deepwater Horizon (DWH) spill and the spate of radar remote sensing research that followed, it was thought that L-band synthetic aperture radar (SAR), with its ~24 cm wavelength, was a poor tool for detecting oil slicks, being more sensitive to short wavelength gravity waves. During the DWH accident, data acquired with NASA's airborne L-band SAR, UAVSAR, showed this not to be the case. Since then the instrument has been used in oil spill detection and characterization studies in the Gulf of Mexico and Norway. Results from those studies will be presented and future plans discussed.
Nov 18: James Hlywiak
Department of Atmospheric Sciences, RSMAS
(one-hour MPO student seminar)
Tropical Cyclone Wind Field Decay and
the Evolution of the Boundary Layer Structure During and After Landfall
Zoom Recording Available at COMPASS ON DEMAND
First, the sensitivity of the inland wind decay to realistic inland surface roughness lengths and soil moisture contents is evaluated for strong, idealized tropical cyclones (TCs) making landfall. Results show that the relative sensitivities to roughness and soil moisture differ throughout the decay process, and are dependent on the strength and size of the vortex. Within 12 h of landfall, increased surface roughness and decreasing surface enthalpy fluxes result in rapid weakening. During this time, the sensitivity of the decay to roughness dominates the sensitivity to soil moisture. After TCs decay to tropical storm intensities, weakening slows and the sensitivity of the wind field decay to soil moisture increases. Additionally, increasing surface roughness accelerates the decay of the strongest winds, while increasing soil moisture slows the decay of the larger TC wind field. Lastly, the pre-landfall effects of the upstream surface roughness on the TC intensity and coastal boundary layer structure are discussed. Increasing roughness leads to elevated inflow, which supports a brief acceleration of the near-surface winds, as well as the formation of a deeper internal boundary layer. Results have implications for inland forecasting of TC winds and understanding the potential for damages.
Nov 25: NO SEMINAR (Thanksgiving Recess)
Dec 02: Anne Barkley
Department of Atmospheric Sciences, RSMAS
(one-hour ATM student seminar)
Understanding Nutrient Transport to the Tropical Atlantic Ocean and
Amazon Basin From Long-Range Transported African Aerosols
The Amazon and equatorial North Atlantic Ocean are nutrient-depleted ecosystems that partially rely on the deposition of atmospheric aerosols for nutrients, which can stimulate primary productivity and sequester carbon dioxide from the atmosphere into the biosphere. In this work, we examine aerosols collected at a coastal field site in South America to answer three main questions: 1) What are the aerosol sources of phosphorus and soluble phosphorus to the Tropical Atlantic Ocean and Amazon? 2) What types of long-range transported particles contain nutrients? and 3) Where does African dust originate from within North Africa? We find that African biomass burning from both southern Africa and northern Africa provide approximately the same amount of phosphorus to the Amazon as dust. Additionally, we find that freshwater diatoms, likely transported from African paleolakes, are internally mixed with dust and provide iron, which is important for stimulating primary productivity in the equatorial North Atlantic. Finally, we show that North African dust originates from largely sub-Saharan Africa, and potentially the region that includes the Bodélé Depression, which has long been hypothesized to be a critical source of dust for fertilizing the Amazon.
Dec 09: Dr. Jennifer MacKinnon
Scripps Institution of Oceanography, University of California San Diego
Tales From a Multi-Scale Ocean
Zoom Recording Available at COMPASS ON DEMAND
The ocean displays immense and compelling variability at an enormous range of scales, from global and basin-wide overturning circulation, down to coastal and regional features with scales of tens of km, down to sharp fronts of order meters, down to turbulent dissipation with scales of centimeters. Though phenomena at these vastly different scales have often been treated quite separately, using different subsets of the fundamental equations of motion, it is becoming increasingly clear that this separation is not always appropriate. Specifically, recent observations theory and modeling point to nonlinear coupling between what initially seem to be very different types of motions, with very different time or space scales, which sometimes can play an order one role in the evolution of phenomena at all scales. In many if not most cases, novel observational methods have often led the way, giving us new and sometimes surprising glimpses into the oceans' inner workings. Here I'll give a flavor of these results from several recent experiments our group has been involved with.
Dec 10 (Thursday): Samantha Ballard
Department of Ocean Sciences, RSMAS
(one-hour AMP student seminar)
Investigation of Satellite-Observed, In-Situ-Measured, and Model-Simulated
Land-Air-Sea Boundary Layer Physics
Air-sea flux parameters are essential for quantifying momentum exchange in the marine atmospheric boundary layer (MABL). Traditional techniques to quantify these parameters in the open ocean assume stationary and homogeneous conditions. However, these conditions break down in coastal areas, where non-stationary conditions, horizontal gradients, topography, breaking waves, etc. are often encountered. Also, traditional in situ instrumentation alone cannot characterize the variety of spatial scales necessary for accurate modeling. Furthermore, the models themselves are often too coarse to resolve small-scale coastal features. To combat these problems, a field experiment named Coastal Land Air-Sea Interaction (CLASI), conducted in Monterey Bay, CA, aimed to combine in situ and satellite data from various instrumentation (on ships, land-based towers, satellites, etc.) to improve forecasting within the Navy’s Coupled Ocean-Atmosphere Mesoscale Prediction System (COAMPS) model and understanding of land-air-sea boundary layer dynamics. A new synthetic aperture radar (SAR) based technique, using wavelet analysis, is proposed to derive state-of-the-art maps of air-sea parameters, such as wave speed, wave age, surface roughness length, drag coefficient, and wind stress at as low as 5-meter resolution. It is validated against traditional buoy and coastal tower anemometer measurements with promising results. This new technique is valuable for improving forecasting models and can be utilized outside of Monterey Bay, in locations where in situ measurements are not available, and in various complex conditions (i.e. hurricanes, etc.).
Dec 14 (Monday, 10:00 am): Dr. Gunnar Spreen
Institute of Environmental Physics, University of Bremen, Germany
MOSAiC – The International Arctic Drift Expedition:
Sea Ice and Remote Sensing Measurements
Zoom Recording Available at COMPASS ON DEMAND
The research icebreaker Polarstern drifted for one year from October 2019 to September 2020 with the sea ice from the Russian Arctic towards Fram Strait between Greenland and Svalbard. The inter-disciplinary measurement program covered atmosphere, ocean, sea ice, ecosystem, and biogeochemistry domains. Here the focus will be on the comprehensive sea ice, snow, and remote sensing program. Observations ranged from micro-physical properties of the snow and ice to the large scale distributions of ice thickness and topography measured by helicopter. A summary of sea ice measurements will be presented together with personal experiences from the first and last leg of the expedition. A focus will be on observations used for satellite remote sensing evaluation and algorithm development. In particular, microwave emission and scattering of sea ice was measured at different frequencies and polarizations.
