FALL 2022
Wednesdays at 3:00 pm, Seminar Room SLAB 103 / Virtual SLAB 103
(unless stated otherwise)
Aug 24: NO SEMINAR
Aug 31: NO SEMINAR (COMPASS Student Committee Meeting)
Sep 08 (Thursday, 10:00 am): Dr. Duo Chan
Woods Hole Oceanographic Institution, Woods Hole, MA
Combining Physical and Data-Driven Methods
to Improve Historical Estimates of Earth Surface Temperatures
In addition to using data-driven methods to mine insights from data, data quality is also a crucial aspect of data science. When applying this philosophy to investigate climate science, an obstacle the community faces is the lack of long and accurate climate records, even for the earth's surface temperatures since the late 19th century. In this talk, I will introduce recent progress in understanding historical temperature variability after combining physical knowledge and data-driven methods to remove data bias and improve observational estimates. Specifically, I will answer two questions. First, how much has the earth's surface warmed since the 1880s, and how far away are we from the 1.5°C warming target? Second, how do biases in the pattern of sea-surface temperatures prevent atmospheric models from accurately simulating recorded North Atlantic hurricane variability? Our improved temperature estimates outline a simpler and smoother warming pattern throughout the 20th century. They also open up new opportunities for further constraining climate sensitivity, attributing climate variations, understanding the dynamics of climate patterns, and benchmarking models, which will help to make more accurate future projections and design efficient mitigation and adaptation strategies. I also hope this talk could provide perspectives on performing proper model-data comparisons. When model and data disagree, although models could be problematic, data are also noisy and sometimes biased reflections of reality. Being skeptical about data quality always appears to be a good practice.
Sep 14: Dr. Larissa Patricio Valerio
College of Science and Engineering, James Cook University, Townsville, Queensland, Australia, and
Commonwealth Scientific and Industrial Research Organisation, Brisbane, Queensland, Australia
Exploring Meteorological Satellite Observations for Diurnal
Water Quality Assessments in the Great Barrier Reef
Zoom Recording Available at COMPASS ON DEMAND
A model-based ocean colour inversion algorithm was developed using artificial neural networks to retrieve diurnal Total Suspended Solids (TSS) concentrations in the coastal Great Barrier Reef (GBR) from geostationary Himawari-8 AHI observations. The machine learning algorithm developed in this work allows the direct inversion of AHI top-of-atmosphere reflectance data to derive a wide range of TSS concentrations (0.01 to 100 mg L−1) at a temporal resolution of up to 10 min. The retrieval accuracy of the inversion algorithm was determined by comparing the satellite estimated TSS with concurrent in-situ TSS measurements in the coastal GBR. Himawari-8 AHI signal-to-noise ratios and retrieval limitations of the inversion algorithm were quantified. The 10 min TSS product provided the required temporal resolution for resolving short-lived coastal processes, such as the dispersion of a flood event, revealing an order of magnitude increase in TSS concentrations within one day. In addition, the Himawari-8 TSS product allowed the identification and tracking of short-lived sub-mesoscale resuspension eddies within the reef matrix. For the first time, coastal TSS features were reliably quantified for the entire GBR, at a temporal resolution only possible with biogeochemical and hydrodynamic models.
Sep 21 (10:00 am): Dr. Po-Lun Ma
Global Atmospheric ModelingTeam, Pacific Northwest National Laboratory, Richland, WA
Improving the Predictability of Aerosol-Cloud Interactions for
Climate Relevant Issues Using AI/ML
The role of aerosol-cloud interactions (ACI) in the climate system is a major source of uncertainty in projections of Earth's future climate and in interpreting how the climate has evolved in the past. Over the past decade, efforts have been made to improve understanding and to address deficiencies in Earth system models (ESMs), including increasing model resolution, improving the representation of aerosol and cloud properties and processes, and incorporating observational constraints. Recent studies have highlighted the importance and usefulness of artificial intelligence (AI) and machine learning (ML) in climate and weather science. We show that artificial neural network (ANN) can be used to replace or augment parameterizations, providing better or faster simulations of droplet nucleation, warm rain initiation, and aerosol optics. We find that the key to success is to incorporate domain knowledge in the design of the ANN emulators. Furthermore, explainable AI (XAI), feature / signature detection, and causal inference techniques have been found useful for identifying emergent properties of the real and simulated climate system. They also provide a path toward knowledge discovery, objectively hunting for anomalies in high-dimensional heterogenous datasets to understand what the machine has learned and to reveal missing mechanisms in climate and weather models. We will also discuss research challenges and potential future opportunities in deploying AI methods to improve the predictability of ACI for climate relevant issues.
Sep 28: NO SEMINAR
Oct 05: Dr. John van Leer
Department of Ocean Sciences, Rosenstiel School (retired)
A Research Vessel Consistent With the IPCC's Net Zero Carbon Target
Recording Available at COMPASS ON DEMAND
The IPCC is clear on the need to bring CO2 and other greenhouse gas emissions to net zero, or below zero. Why should seagoing oceanography be exempt from the economic and moral consequences of Climate Change? As major contributors to global climate science, oceanographers should lead by example and join the great transition, with low carbon research vessels. Serious post oil thinking has emerged, with the passage of the Inflation Reduction Act. Russia Oil sanctions and global shortages have led to upward pressure on fuel prices in the foreseeable future.
Sun 21 was the first vessel to cross the Atlantic using solar photovoltaic power for propulsion paired with battery storage.
Oct 12: Dr. Ori Adam
Fredy and Nadine Herrmann Institute of Earth Sciences, Hebrew University of Jerusalem, Israel
Single and Double ITCZs and Everything in Between
Recording Available at COMPASS ON DEMAND
The tropical rain belt varies between unimodal (single ITCZ) and bimodal (double ITCZ) meridional precipitation distributions, both regionally and on seasonal to geological timescales. In this talk I will cover some of the theory and implications of modal variations of the tropical rain belt, and argue that "tropical modality" is a fundamental characteristic of tropical climate. Specifically, in the present climate, the prevailing tropical circulation regime is the Hadley circulation, which under ideal conditions leads to a unimodal precipitation distribution (single ITCZ) that follows the Sun in its seasonal migrations. However, land-ocean contrast, cloud effects, and atmosphere-ocean coupling introduce asymmetries that cause the underlying characteristics of the tropical rain belt to significantly deviate from the idealized Hadley paradigm. For example, in past climates and currently in the western Pacific, ITCZs reside on either side of the equator year-round, yielding seasonal variations that significantly differ from Hadley-like circulations. Similarly, modern climate models tend to overestimate precipitation south of the equator in the Pacific, leading to an overly bimodal precipitation distribution – a problem commonly known as the "double-ITCZ bias". Key aspects of these modal variations are captured by simple dynamic and energetic constraints, which I will cover.
Oct 19: Amie Dobracki
Department of Atmospheric Sciences, Rosenstiel School
(one-hour ATM student seminar)
Characterization of Aged Biomass Burning Aerosol Over the
Southeast Atlantic Ocean in the Free Troposphere and Marine Boundary Layer
Recording Available at COMPASS ON DEMAND
Aerosol over the remote southeast Atlantic (SEA) is some of the most sunlight-absorbing aerosol on the planet. When overlying the large subtropical stratocumulus cloud deck, the absorbing aerosol warms the region by 20-30 Wm–2, although model estimates can vary widely. When located within the boundary layer, the aerosol both diminishes the cloud fraction while brightening the remaining clouds through microphysical effects. We seek to explain the chemical, optical, and physical properties of the highly absorbing biomass burning aerosol sampled over the SEA, and to characterize the prevailing wind structure that transports the aerosol to the sampling locations. Here we examine data from an aircraft campaign which focused more on the free tropospheric aerosol, and information on boundary layer aerosol gathered at Ascension Island (8ºS, 14.5ºW, on the mid-Atlantic ridge). Results from the September 2016 aircraft campaign indicate the single-scattering albedo (SSA) of aged free-tropospheric biomass burning can be robustly estimated from the ratio of organic aerosol to black carbon aerosol (OA:BC). Measurements taken near-coast within a significant plume support an inference of OA mass loss through photolysis. The simple aerosol SSA depiction can be easily implemented into large-scale models, with the caveat that most models currently overestimate the OA:BC ratio, leading to too much scattering of sunlight. We further investigate the aerosol chemical and physical properties alongside the optical properties of biomass burning aerosol between June and September 2017 in the marine boundary layer at Ascension Island. Despite the lack of strong free-tropospheric winds, the June and July boundary layer can nevertheless be very smoky at times. The significant smoke loading in austral winter boundary layer is unexpected. The smoke is either advected directly west at low altitudes off of the continent when the south Atlantic high is well-developed and near the African coast, or from the north as part of convectively-induced circulations over western Africa. The boundary layer smoke also has low single scattering albedos that correlate well to the fraction of black carbon. The aerosol composition appears to be mostly informed by burning conditions and fuel types, and less by aging time scale and transport pathways. Further work will more completely integrate the findings from both campaigns.
Oct 26: Dr. Dalia Kirschbaum
Earth Sciences Division, NASA Goddard Space Flight Center, Greenbelt, MD
Exploring Earth: Your Planet, Our Mission
Zoom Recording Available at COMPASS ON DEMAND
NASA's satellites and telescopes are seeking to uncover the mysteries of the universe, to understand where we all came from, and to find life in our solar system. From the vantage point of space, satellites are also looking down at our own planet to observe how our Earth is constantly moving and changing. The abundance of Earth observations has sky-rocketed, with over 6 million individual data points flowing in every 6 hours and growing by the day. These data from active and passive systems in low earth to geostationary orbits are being ingested, processed, calibrated and archived to provide vital insight of our Earth as an interconnected system. This seminar will provide an introduction into how remotely sensed data and models are being used to gain better understanding of how water moves around our planet in the atmosphere, on land, and in our oceans. Through the integration and synthesis of this information, NASA and partner products and capabilities support a broad range of stakeholders needs that benefit society.
Nov 02: NO SEMINAR
Nov 09: NO SEMINAR
Nov 16: NO SEMINAR
Nov 23: NO SEMINAR (Thanksgiving Recess)
Nov 30: Dr. Erik van Sebille
Department of Physics & Freudenthal Institute, Utrecht University, Netherlands
Whose Plastic Is That?
Combining Ocean Physics With Bayesian Inference
to Attribute Marine Plastic Sources and Sinks
Recording Available at COMPASS ON DEMAND
The world's ocean currents have the potential to transport material like plastic over vast scales, connecting leakage on one continent to impacts on another. Yet, it had recently become clear that most plastics found at any particular location are relatively local, often originating from within the same country. Effective policies to reduce the impact of plastic pollution require knowledge of whose plastic ends up where. In this seminar, I will present some recent work on using a Bayesian framework to analyze the sources of plastics found on beaches around the world. The input to this analysis comes from Lagrangian ocean analysis simulations with the OceanParcels.org tool, which I will also showcase. I will particularly highlight results from the Indian Ocean, the Galapagos, the South Atlantic and the North Sea.
Erik van Sebille is professor of oceanography and public engagement at Utrecht University. He investigates how ocean currents move 'stuff' around. He is co-author of the textbook 'Ocean Currents – Physical Drivers in a Changing World' with Professor Robert Marsh. Until 2022, he led the European Research Council Starting Grant project 'Tracking Of Plastics in Our Seas'. In parallel to his ongoing work in physical oceanography, he has recently started a new research team on how scientists can be effective and inclusive in their communication and engagement with society, specifically on the climate crisis.
Dec 07: Haozhe He
Department of Atmospheric Sciences, Rosenstiel School
(one-hour ATM student seminar)
Constraining the Inter-Model Spread in Radiative Forcing and Feedbacks
Recording Available at COMPASS ON DEMAND
Climate sensitivity is the change in global-mean surface temperature required to restore radiative equilibrium in response to a doubling of CO2 concentration and is the most widely used metric to quantify the susceptibility of the climate to an externally forced change. When evaluating the effect of CO2 changes on the earth's climate, it is widely assumed that variances in climate sensitivity arise from differences in radiative feedbacks. Here we attempt to observationally constrain the feedbacks in an effort to reduce their inter-model uncertainties. The observed interannual variation provides a useful constraint on the long-term cloud feedback, while the combined lapse-rate plus water vapor feedback cannot be constrained with observations of interannual variability, uncovering a large discrepancy among different observations and reanalyses products on the interannual feedback value. In addition, this study suggests the role of radiative forcing in determining climate sensitivity has been greatly underrated. Our results demonstrate the underestimate of the radiative forcing is the cause for the underestimated climate sensitivity by the most widely used method for estimating climate sensitivity. As the most important component of the radiative forcing, instantaneous radiative forcing (IRF) measures the change in net radiative flux that results only from the change in forcing agents. Compared to the widely held assumption that the IRF from a doubling of a given CO2 concentration is constant, our results show the IRF2×CO2 is not constant, but also depends on the climatological base-state, increasing by ~25% for every doubling of CO2, and has increased by ~10% since the pre-industrial era primarily due to stratospheric cooling, implying a proportionate increase in climate sensitivity. This base-state dependence also explains about half of the inter-model spread in IRF2×CO2, a problem that has persisted among climate models for nearly three decades. Meanwhile, a comprehensive evaluation of simulated radiative forcing from CMIP3 to CMIP6 models reveals little progress in radiation schemes in the recent two decades. The stronger IRF in the most recent generation of climate models resulted from finer stratospheric resolution causes the higher radiative forcing, partly contributing an alarming increase in climate sensitivity compared to previous generations, suggesting the importance of a well-resolved stratosphere.
