FALL 2025
Wednesdays at 3:00 pm, Seminar Room SLAB 103 / Virtual SLAB 103
(unless stated otherwise)
Aug 20: Dr. Givo Alsepan
Department of Earth, Ocean and Atmospheric Science, Florida State University, Tallahassee
A Potential Pathway for AMOC Influence on Seasonal Atmospheric Variability
The importance of Gulf Stream (GS) variability for Northern Hemisphere weather and climate has received considerable support in recent years. Recent observational advances show that significant GS variability occurs on multiple timescales, including seasonal. Recently, a high degree of seasonal variability has also been observed in the Atlantic Meridional Overturning Circulation (AMOC), suggesting a potential connection between the seasonal AMOC and mid-latitude atmosphere via the GS. Modeling studies, however, are yet to agree on the exact relationship between AMOC and GS variability, partly due to the lack of eddy-scale resolution. Here, a novel ensemble of eddy-rich model simulations is used to investigate the patterns of co-variability between and fine-scale sea surface temperature (SST) anomalies over the GS on seasonal timescales as a function of time lag. Our results reveal that wintertime variability in GS SST and its meridional gradient are preceded by AMOC variability in the previous summer. These AMOC-induced SST gradient anomalies then drive air temperature gradient anomalies through changes in the differential air-sea sensible heat flux. The findings indicate a potential pathway through which deeper ocean circulation precedes atmospheric variability via the GS. The identification of this pathway may provide a potential source of untapped predictability in weather forecast and climate prediction systems.
Aug 27: NO SEMINAR
Sep 03: NO SEMINAR
Sep 10: NO SEMINAR
Sep 17: Dr. Lisa Beal
Department of Ocean Sciences, Rosenstiel School
More Eddying in Subtropical Western Boundary Currents is
Cooling Adjacent Shelf Waters, Warming Surface Waters Downstream,
and Reducing Oceanic Meridional Heat Transport
Lisa Beal1 and Kathryn Gunn2
1Department of Ocean Sciences, University of Miami, 2University of Southampton, UK
Recording Available at COMPASS ON DEMAND
Observations indicate more ocean eddies with climate change, especially in swift boundary currents like the Gulf Stream. Eddying drives cross-shore fluxes of heat, salt, and nutrients that impact the current and the adjacent shelf-sea environment. Yet, it is unclear whether more eddies will tend to warm or cool the current, suppress or stimulate coastal ecosystems. To approach this question, we study Agulhas Current instabilities off South Africa, quantifying their cross-shore eddy heat and salt fluxes for the first time, using two years of continuous mooring data. We find that eddies converge heat and salt towards the current core over time. On the inshore edge, 10-km frontal instabilities dominate eddy fluxes, pumping cold, nutrient-rich waters up onto the shelf and producing a net offshore heat flux. Farther offshore, 100-km meanders drive net onshore fluxes of heat and salt. Together, these instabilities yield a broader, more stratified Agulhas Current that transports more surface warmth, but less total heat, to higher latitudes. Turbulence closure theory predicts these same tendencies for subtropical boundary currents globally. Combined with greenhouse-driven surface warming, our findings imply that coastal ecosystems will face stronger extremes in future, with cooler more productive bottom waters alongside warmer more deficient surface waters.
Sep 24: Dr. Cassandra Gaston
Department of Atmospheric Sciences, Rosenstiel School
(Candidate for Promotion)
The Impact of Dust and Beyond on Marine Biogeochemical Cycles
Aerosols supply key nutrients, such as phosphorus and iron, to receptor ecosystems that can stimulate primary productivity and, consequently, lead to carbon sequestration in the ocean and terrestrial biosphere. This so-called "aerosol indirect biogeochemical effect" has not been studied as well as the direct radiative properties of aerosols or aerosol-cloud interactions. Mineral dust has been assumed to be the most important contributor of nutrients to receptor ecosystems. However, nutrients in dust can be present in poorly soluble forms leading to several outstanding research questions including, what processes are most important for transforming insoluble nutrients into bioavailable forms? and are non-dust aerosols more important than dust for the delivery of bioavailable nutrients? To answer these questions, we combined traditional, bulk chemical analysis from aerosol filters with state-of-the-art single particle methods to obtain mechanistic information as to what processes are most important for driving aerosol nutrient solubilization. The results from our work have challenged current hypotheses regarding dust aging as well as the importance of dust aerosols in driving biogeochemical cycles. I will present several works exploring chemical processing of dust thought to increase nutrient solubility as well as nutrient delivery from non-dust sources including volcanic ash and smoke from wildfires. I will also discuss the major sampling platform used to conduct this research, the Barbados Atmospheric Chemistry Observatory (BACO), its recent upgrades, and future directions for research at this site.
Oct 01: SPECIAL ATM & OCE FACULTY PRESENTATION SERIES
Dr. John van Leer
Department of Ocean Sciences, Rosenstiel School (retired)
My Life as an Oceanographer, Engineer & Advocate – So Far
A Life of Adventure, Options, Branchpoints & Agency
Recording Available at COMPASS ON DEMAND
As I plan for the next 20 years of my – all-electric – robotically assisted life, it's time to reflect upon my previous 85 years. I grew up during WW2 before TV, web, or AC, with Victory Gardens and Gold Star Mothers and Wives. By 4, I knew I would be a scientist or engineer – thanks Smithsonian. I attended National Geographic Lectures at 7. By 14, I designed and built a homemade SCUBA (Self-Contained Underwater Breathing Apparatus) and had my first tank filled at WHOI. By 15, I was an apprentice machinist. At 16, I was the youngest Fuller Brush Salesman. In high school I raised tropical fish and built drag racing cars commercially.
I graduated Case BSME '62, worked on Reentry Inertial Guidance System design / testing at MIT Instrumentation Lab, MINUTEMAN-MIRV for 3.5 years. I graduated from MIT–WHOI Joint Program in Oceanography '71, studying Shear & Density Gradients in Thermocline Microstructure under Henry Stommel. For the following 51 years, I was in the RSMAS faculty, designing, building, and using the first robotic oceanographic profilers – Cyclesondes, including solar-powered telemetry. As an observational oceanographer, I participated in multiple West Florida Shelf Experiments, CUEA (Coastal Upwelling Ecosystems Analysis) experiments – two in Oregon and two in Peru – and tropical oceanic observations during the Global Atmospheric Research Program's Atlantic Tropical Experiment (GATE). I participated in six months of under-ice observations in the Marginal Ice Zone Experiment (MIZEX) and the Coordinated Eastern Arctic Experiment (CEAREX) during the Cold War, as well as an over-winter under-ice study of the roll structure in the bottom boundary layer near Copper Harbor in Lake Superior.
As UM faculty, I taught over a hundred graduate students and over a thousand undergraduate students, with first-hand knowledge gained during four and a half years at sea and two decades of lobbying the US Congress for carbon pricing legislation. I served as Vice Chair of the School Council during the Berman years and Chair of the Marine Committee and UNOLS Representative. I also headed the Calanus Replacement Committee during the Rosendahl Deanship, helping raise $0.5M from the Nason Foundation to construct an efficient catamaran ship with the first robotic motion isolation system, described in my first COMPASS Seminar.
Oct 08: Dr. Marybeth Arcodia
Department of Atmospheric Sciences, Rosenstiel School
Coming Full Circle: Harnessing Data Science to Advance Climate Prediction
Through Forecasts of Opportunity
Recording Available at COMPASS ON DEMAND
The climate system is chaotic and noisy, yet within this noise lie windows of predictability that enhance our ability to make accurate weather and climate predictions. This talk highlights innovative machine learning and data science approaches that identify forecasts of opportunity – climate states where prediction skill and confidence are elevated – ultimately advancing prediction beyond traditional limits. I will overview how neural networks trained on large ensemble climate model data, validated with observations, reveal subseasonal (2 week – 3 month) U.S. precipitation predictability. North Atlantic sea surface salinity, a relatively untapped source, holds predictive information for Midwest precipitation forecasts of opportunity. Explainable AI techniques provide insights into model decision-making, building trust while uncovering new sources of predictability and improving our understanding of the climate system.
Building on these insights, my group is developing a machine learning-based real-time subseasonal U.S. precipitation forecasting tool that identifies when, where, and why a forecast is a forecast-of-opportunity. By quantifying forecast confidence and effectively communicating predictions to forecasters and users, this tool integrates scientific understanding with actionable information. Ongoing and future research in the group spans multiple domains: predicting lead times for coral heat stress events, applying machine learning for uncertainty quantification in coastal sea level predictions, predicting the timing of critical temperature threshold crossings in future climates, and utilizing purely AI-based weather models to push the frontiers of climate prediction. Coming full circle to the place where my research journey began, I will conclude with future directions at the Rosenstiel School and the Institute for Data Science and Computing (IDSC), highlighting opportunities for interdisciplinary collaboration at the intersection of data science and climate.
Oct 15: Dr. Jeremy Klavans
Department of Atmospheric Sciences, Rosenstiel School
The Signal-to-Noise Error in Decadal Regional Climate
Recording Available at COMPASS ON DEMAND
To support informed decision-making in a changing world, climate scientists would like to offer detailed predictions of near-term, regional climate changes. It is widely assumed that near-term changes in regional climate are primarily generated by variations internal to the atmosphere-ocean system, implying that there is limited predictability to guide this decision-making. However, using extraordinarily large ensembles of multiple climate models, we show that there is a significant role for externally forced variability in decadal modes of regional climate (e.g. AMV, NAO, and PDO) and their impacts, including changes in the environment in which hurricanes develop as well as the atmospheric circulation patterns responsible for regional drought. This newfound role for external forcing in these decadal modes was previously obscured by models' erroneous penchant for underestimating the amplitude of externally forced multidecadal modes relative to internally generated noise. We propose that accounting for and resolving this signal-to-noise error in climate models will unleash new predictive skill for critical regional climate changes.
Oct 22: NO SEMINAR
Oct 29: Eric Mischell
Department of Atmospheric Sciences, Rosenstiel School
(one-hour ATM student seminar)
Mean and Extreme Precipitation Under Global Warming
Recording Available at COMPASS ON DEMAND
It is expected that, in the global mean, the concentration of water vapor will increase by ~7% per K of global warming and the surface precipitation rate will increase by ~2% per K. From the equation P = Mq (where P is the precipitation rate, q the mixing ratio, and M the convective mass flux), the slower increase of precipitation compared to water vapor implies a decrease of the convective mass flux. Here we revisit this equation and show that a slightly altered form, which introduces the humidity-weighted vertical mass flux rather than the convective mass flux, better describes the circulation response. We then turn from the mean precipitation to the response of the precipitation distribution to global warming. Because no fundamental theory describes this response, we look to observations of rainfall as detected by the Special Sensor Microwave Imager (SSM/I). We show, using four decades of satellite data over the ocean, that heavy and light rain have increased in frequency, while moderate rain has decreased in frequency, leading to a modest rise in the mean rain rate. We furthermore analyze a proxy for convective intensity and find that convective events have shifted from higher to lower intensities, except for the strongest events where the trend is near-neutral. We conclude by discussing the possibility of extending this analysis to rainfall over land by making use of space-based precipitation radars.
Nov 05: Tyler Tatro
Department of Atmospheric Sciences, Rosenstiel School
(one-hour ATM student seminar)
Determining the Dominant Smoke-Cloud Interactions in the Southeast Atlantic
Across Time Scales
Recording Available at COMPASS ON DEMAND
The southeast Atlantic Ocean is home to one of the largest stratocumulus cloud decks on the planet. Seasonal agricultural burning over southern Africa produces biomass burning aerosol (smoke), and the highly shortwave absorbing aerosol can promote or degrade cloud development depending on its location relative to the cloud. In this talk, we utilize a combination of in-situ and satellite observations with meteorological and aerosol reanalysis (ERA5, CAMS) to uncover where, when, and how smoke promotes or decreases cloud reflectivity. On a daily time scale from June to August, smoke presence in the marine boundary layer (MBL) is well correlated with synoptic variations in a weak thermal low over Angola. Low-level easterlies over a weak Angolan thermal low advect aerosol into the marine boundary layer, where they are further carried west, increasing cloud droplet number concentration but contributing to lower cloud fractions. Synoptic variations constrain implications for using shipping tracks as marine cloud brightening proxies. On a decadal time scale (2003-2024), regional circulation trends are more robust from August to October, when increased springtime land heating has strengthened the free-tropospheric winds lofting the aerosol above the stratocumulus deck. Burned area data also reveal a more peaked burning season, and more aerosol has transported over the southeast Atlantic in the past 20 years. Circulation adjustments are tied to the expansion of the subtropical dry zones and continental warming, and southwest movement of the south Atlantic subtropical high. The circulation impacts on aerosol transport help explain a previously-noted positive decadal net aerosol radiative effect impact over the southern subtropical Atlantic, mostly from more aerosol present above the southward extension of the stratocumulus deck.
Nov 12: Aidan Mahoney
Department of Atmospheric Sciences, Rosenstiel School
(one-hour ATM student seminar)
Investigating the Influence of the Large-Scale Environment on Tropical Cyclones
Recording Available at COMPASS ON DEMAND
Tropical cyclones (TCs) are rare but destructive weather events, responsible for approximately 50% of the total cost of all natural disasters from 1980-2024. Changes in the large-scale environment in response to global warming are likely to result in an increase in the most destructive storms. However, global climate model (GCM) projections of TC activity exhibit substantial differences between models. For example, potential intensity (PI), a key indicator for TC activity, exhibits considerable variability across GCMs even when using identical sea surface temperatures (SSTs), due to differences in tropopause temperatures. To overcome these limitations and develop a better understanding of how the large-scale environment influences TC activity, we present a new Matrix-of-Environments (MoE) based on reanalysis data and numerical modeling that relates climatological variations in the large-scale environment to TC activity. This method has several advantages over other approaches: i) the MoE uses reanalysis data to accurately capture the full spectrum of environmental conditions in TC development regions; ii) the Point-Downscaling (PDS) method accurately initializes each environment within the Weather Research and Forecasting (WRF) model, allowing for simulations that resolve the complex physical processes before and after TC genesis; iii) we can perform a physically-based decomposition of the influence of different environmental factors on observed TC activity. This presentation will demonstrate the application of the MoE for understanding observed changes in TC activity during the historical period and outline future plans for understanding the response of TCs to anthropogenic climate change.
Nov 19: William Downs
Department of Atmospheric Sciences, Rosenstiel School
(one-hour ATM student seminar)
Identification and Prediction of Tropical Easterly Waves Using Machine Learning
Recording Available at COMPASS ON DEMAND
Tropical easterly waves (TEWs) directly impact people through wind, rain, and hurricane formation in the Pacific and Atlantic Oceans. Current TEW tracking algorithms struggle to accurately identify TEWs over the Caribbean Sea. We train a convolutional neural network (CNN) using ERA5 reanalysis data and human-analyzed labels to identify TEWs in the eastern North Pacific and North Atlantic. This methodology successfully captures TEW activity in the Caribbean. We analyze which input atmospheric variables most contribute to the CNN's performance, and examine differences between TEWs identified in the Caribbean and TEWs identified in the open Atlantic. We also present results from a similarly trained CNN that identifies the locations of the intertropical convergence zone and monsoon trough, and show that the monsoon trough has extended further westward from Africa during peak summer months over the past several decades. We use the TEW CNN to generate an archive of TEW tracks from 1981-2023, and we use this dataset to train a new CNN to generate ensemble forecasts of vorticity strength and convective organization in TEWs at lead times of 1-5 days. We compare the performance of this prediction CNN for TEWs in 2023 to the performance of the European Centre for Medium-Range Weather Forecasts' deterministic model and that of a simple climatology model. We assess which input variables contribute most to the CNN's predictions of TEWs, and we investigate discrepancies between which variables are most important for predicting TEWs of different strengths and TEWs in different geographic regions. Physically intuitive mechanisms seen in this investigation can help us better understand how TEWs evolve along an intensity / organization spectrum ranging from weak, dry waves to full-fledged tropical cyclones.
Nov 26: NO SEMINAR (Thanksgiving Recess)
Dec 03 (4:00 pm): Dr. Dennis Hansell
Department of Ocean Sciences, Rosenstiel School
Biogeochemical Dynamics of Organic Carbon Dissolved in the Global Ocean
Marine dissolved organic carbon (DOC) has important roles in the cycling of carbon in the global ocean; it is illuminating the details of those roles that has been the goal of my laboratory for several years. The pool has long been recognized as one of Earth's major carbon reservoirs and as a nutritive substrate supporting the metabolism of numerous heterotrophic microbes. Today, I focus on understanding the dynamics of carbon processed through oceanic DOC at the broadest spatial scales. I explain controls on its spatial distribution in both the surface and deep global ocean. The net production and accumulation of DOC relative to the distribution and timing of marine production are considered. I address the mechanisms and locations for introducing DOC to the deep ocean (as a component of "carbon export production"), both by overturning circulation and by solubilizing sinking biogenic particles. I conclude with recent findings on the abiotic removal of "aged" DOC in the deep Pacific Ocean.
SPRING 2026 PREVIEW
Jan 14: Dr. Lauren Zamora
NASA Goddard Space Flight Center, Greenbelt, MD / University of Maryland College Park
Guest of Paquita Zuidema, Department of Atmospheric Sciences
Feb 18: Dr. Andrew Dessler
Texas Center for Extreme Weather, Texas A&M University, College Station
Guest of the Department of Atmospheric Sciences
