Providing Support for Cutting-edge Climate Models
Climate models are difficult to build and run, often resulting in predictions with degrees of
misalignment. At the same time, cutting-edge research is siloed in different institutions, slowing exchanges that can revolutionize advancements in climate modeling. This has ultimately led to a recurring lack of credible guidance on a range of critical issues, such as our gas emission habits.
The Virtual Earth System Research Institute (VESRI) is a distributed research center seeking to radically improve the credibility of climate predictions. The institute provides sustained funding and embedded technical expertise to transformational research pertaining to climate change, focusing on areas of climate research primed to take advantage of the current, rapid evolution of computational technology and observing platforms. Ultimately, the institute aims to improve global climate models in time for them to be used to inform investment decisions on climate mitigation and adaptation.
To advance this goal, we solicited ambitious proposals from multidisciplinary and multinational groups focusing on the most fundamental and important questions in Earth system science. The VESRI Advisory Board helped select four projects that demonstrated the greatest potential for transformative improvements in models of the Earth system and its components, Earth observations, and computational tools, and for bringing tools and approaches from outside the climate sciences to bear within it. The groups will train graduate students and postdoctoral fellows in their fields of expertise, in an effort to develop exceptional talent in the climate sciences for decades to come.
Projects Supported by VESRI
The Scale-Aware Sea Ice Project (SASIP)
Principal Investigator: Pierre Rampal
Participating Institutions: Centre National de la Recherche Scientifique, Brown University, Nansen Environmental and Remote Sensing Center, University of Reading, Sorbonne Université, École Nationale des Ponts et Chaussées, Centre Européen de Recherche et de Formation Avancée en Calcul Scientifique, Mercator Ocean International, Euro-Mediterranean Center on Climate Change, University of Bamako
The Scale-Aware Sea Ice Project (SASIP) aims to develop a truly innovative, scale-aware continuum sea ice model for climate research; one that faithfully represents sea ice dynamics and thermodynamics and that is physically sound, data-adaptive, highly parallelized and computationally efﬁcient. SASIP will use machine learning and data assimilation to exploit large datasets obtained from both simulations and remote sensing.
Through the further development of existing important state-of-the-art simulators created by some of the investigators, SASIP will build a data-constrained sea ice model that is based on solid-like physics. This model will allow improved high resolution and large scale predictions of Arctic and Antarctic sea ice, and the propagation of sea ice related climate feedback. Employing hybrid data assimilation and machine learning approaches as a native part of the model architecture will allow for objective combinations of models and data. Ultimately, SASIP will give a better understanding of the impact of amplified warming in polar regions through the development of a model that reduces uncertainties related to global earth systems.
Principal Investigator: Aditi Sheshadri
Participating Institutions: Stanford, NYU, Rice, Goethe University Frankfurt, Laboratoire de Meteorologie Dynamique, Max-Planck-Institut für Meteorologie, UK Meteorological Office, NorthWest Research Associates
Atmospheric gravity waves play an important role in the exchange of momentum
between the Earth’s surface and the free atmosphere. They are excited by flow over topography,
convective systems, and fronts, and then propagate both vertically and horizontally
into the atmosphere. When they break at higher altitudes, momentum imparted to the waves at their generation level is deposited, playing a key role in the momentum budget of the tropospheric jet streams, and a leading role in the stratosphere. Uncertainty in the parameterization of GW momentum transport limits our ability to predict the response of the large scale circulation to global warming.
This project will deliver two key advances:
The first transformation will open up a potentially transformational data source to constrain GW momentum transport in the atmosphere. They will process and make publicly available 6+ years (2014-2019) of this novel dataset.
The second transformation will be to use machine learning (ML) to develop computationally feasible representations of GW momentum deposition.
M2LInES – Multiscale Machine Learning In coupled Earth System Modeling
Principal Investigator: Laure Zanna
Participating Institutions: NYU, Columbia, Princeton, LDEO, NCAR, MIT, GFDL, CNRS-IGE, CNRS-LOCEAN
The overall goal of the M2LInES project is to improve climate projections and reduce climate model errors, especially at the air-sea interface, using scientific machine learning (ML).
This project will leverage vast amounts of data from global and regional high-resolution simulations, and coupled data assimilation, together with recent breakthroughs in ML algorithms. The team will rely on open-source software and infrastructure to accelerate discovery, which was developed by the Pangeo Project, a community platform for big data geoscience. M2LInES will help deepen our understanding and the representation of unresolved ocean, ice and atmosphere processes in climate models, and target the reduction of intrinsic structural model errors to increase the reliability of simulated surface fields on timescales of hours to centuries.
M2LInES will guide the development of new physics-aware and interpretable, representations of complex processes directly from data for implementation in existing community climate models.
Land Ecosystem Models based On New Theory, obseRvations, and ExperimEnts (LEMONTREE)
Principal Investigator: Sandy Harrison
Participating Institutions: Reading University, Imperial College London, Columbia, University of Pittsburgh, UC Berkeley, Utrecht University, Seoul National University, Texas Tech University, Tsinghua University, Swiss Chinese Institute of Technology in Zurich
LEMONTREE proposes to develop a next-generation model of the terrestrial biosphere and its interactions with the carbon cycle, water cycle and climate. Their approach would lead to ecosystem models that rest on firm theoretical and empirical foundations, and should eventually yield more reliable projections of future climates. This could give a newfound ability to address issues in sustainability, including the potential to maintain the biosphere’s capacity to regulate the carbon cycle while benefiting human well-being and development.
In addition to these four new projects, VESRI also supports the Keeling Curve project and CliMA.
Principal Investigator: Ralph Keeling
Institution: Scripps Institution of Oceanography at the University of California San Diego
In 2020, Weibin Pei gave $1 million in continuation funding to the Keeling Curve carbon dioxide measurement, the long-term atmospheric measurement that alerted the world to human-induced climate change.
The grant, supported by SpaceOne Futures, will fund the continuation of global atmospheric carbon dioxide concentration measurements maintained by the Scripps CO2 Group, including the critical Keeling Curve measurements that have been recorded daily at Hawaii’s Mauna Loa Observatory since 1958.
In addition, the Schmidt Ocean Institute awarded a separate $450,000 grant to the University of California San Diego to support complementary ocean acidification measurements in the Atlantic and Pacific.
Climate Modeling Alliance (CliMA)
Principal Investigator: Tapio Schneider
Institution: California Institute of Technology, Naval Postgraduate School, MIT, Jet Propulsion Lab
The Climate Modeling Alliance (CliMA) is a coalition of about 70 scientists, engineers, and applied mathematicians from Caltech, MIT, the Naval Postgraduate School, and NASA’s Jet Propulsion Laboratory. Their mission is to provide the accurate and actionable scientific information needed to face the coming changes—to mitigate what is avoidable, and to adapt to what is not.
With support from SpaceOne Futures, CliMA is building a new Earth system model that leverages recent advances in the computational and data sciences to learn directly from a wealth of Earth observations from space and from the ground as well as high-resolution simulations spun off on the fly. It will couple atmosphere, ocean, and land models to a single global climate model that is calibrated through the Earth observations and simulations. This model will harness more data than ever before, providing a new level of accuracy to predictions of droughts, heat waves, and rainfall extremes. The model will serve as a backbone for a suite of end-user applications that can be used to plan for and respond to climate change.