Climate Research Initiative - photo of a man with a tripod looking up at the Northern Lights

Microsoft Climate Research Initiative

Researchers working together to fight climate change

These collaborative projects with academia are focused around materials engineering for carbon reduction and removal, data fusion to improve carbon accounting, and causal machine learning for understanding and predicting climate risk and related interventions.

Carbon accounting

Tsinghua University: Jia Xing (PI)
Wuhan University: Siwei Li (Co-PI)
Microsoft: Shuxin Zheng, Chang Liu, Yu Shi

Understanding the change in CO₂ emissions from the measurement of CO₂ concentrations such as that done by satellites is very useful in tracking the real-time progress of carbon reduction actions. Current CO₂ observations are relatively limited: numerical model-based methods have very low calculation efficiency. The proposed study aims to develop a novel method that combines atmospheric numerical modeling and machine learning to infer the CO₂ emissions from satellite observations and ground monitor sensor data.

Related papers:
Tsinghua University: Zhu Liu (PI), Piyu Ke (PhD student)
LSCE: Biqing Zhu (Co-PI), Philippe Ciais (Co-PI)
UC Irvine: Steven J. Davis (Co-PI)
Microsoft: Xiaofan Gui, Jiang Bian

Project Summary:

The initiative focuses on improving the computation of the carbon budget, which refers to the permissible amount of carbon dioxide (CO2) emissions required to limit global warming to a targeted level. The existing methods for calculating the carbon budget are limited by a lag of 1 to 2 years, and challenges in forecasting regions with sparse data. This project aims to address these problems by enhancing the accuracy and speed of carbon sink forecasts, particularly those related to oceans.

Two key advancements define this project:
1. Speeding Up Carbon Sink Estimates: Traditionally, ocean carbon sink estimates, calculated by the state-of-the-art Global Ocean Biogeochemistry Model (GOBM), take 1 to 2 years due to data lag. This project has remarkably reduced this time to 1 to 2 months by leveraging a deep learning approach. By utilizing data with only a maximum lag of 1-2 months, this method bypasses the need for the 1-2 year lag data, achieving near real-time forecasting.
2. Forecasting in Regions with Limited Observational Data: Existing models struggle with limited observational data, hampering accurate forecasts in specific regions. To overcome this, the project introduces a semi-supervised deep learning model that enables generalization into broader regions and extended time frames. This is achieved through a specialized teacher-student learning mechanism that augments limited labeled data.
The ultimate objective of this innovative project is to provide governments and administrative agencies with accurate and timely carbon budget computations. The anticipated outcome of this work would allow for more responsive and informed policy adjustments, directly contributing to global efforts in controlling climate change. By achieving near real-time forecasting and overcoming data sparsity challenges, this project represents a significant step forward in the field of carbon sink/budget forecasting.

Related papers:

Carbon reduction and removal

University of California, Berkeley: Jeffrey Long (PI), Katerina Graf (PhD student), Hiroyasu Furukawa (Research Scientist), Xiang Fu (MIT PhD student), Andrew Rosen (Miller Research Fellow)
Microsoft: Jake Smith, Bichlien Nguyen, Kali Frost, Karin Strauss, Tian Xie

Removing CO₂ from the environment is expected to be an integral component of keeping temperature rise below 1.5°C. However, today this is an inefficient and expensive undertaking. This project will tailor design of new metal–organic frameworks (MOFs) that exhibit new structures and mechanisms of CO₂ capture relevant to the low-cost removal of CO₂ from air and other dilute gas streams.

Related paper:
- MOFDiff: Coarse-grained Diffusion for Metal-Organic Framework Design
North Carolina State: Michael D. Dickey (PI), Dhwanil Vaghani (Graduate Research Assistant), Chemical Engineering Seniors (Senior design team)
University of New South Wales: Kourosh Kalantar-Zadeh (Collaborator/advisor)
Microsoft: Kali Frost, Bichlien Nguyen, Karin Strauss, Jake Smith

The CO₂ reduction process can be used to convert captured carbon into a storable form as well as to manufacture sustainable fuels and materials with lower environmental impacts. This project will evaluate liquid metal-based reduction processes, identifying advantages, pinch-points, and opportunities for improvement needed to reach industrial-relevant scales. It will lay the foundation for improving catalysts and address scaling bottlenecks.

Related papers:
- Liquid metals: fundamentals and applications in chemistry (opens in new tab)
- Liquid-Metal-Enabled Mechanical-Energy-Induced Co₂ Conversion (opens in new tab)
University of Michigan: David Kwabi (PI), Bryan Goldsmith (Co-PI), Anne McNeil (Co-PI), Jessica Tami (PhD student), Cameron Gruich (PhD student), Longbang Liu (Undergraduate), Siddhant Singh (PhD student)
Microsoft: Bichlien Nguyen, Jake Smith, Kali Frost, Karin Strauss, Marwin Segler, Yingce Xia, Ziheng Lu

Energy storage is essential to enable 100% zero-carbon electricity generation. This work will use generative machine learning models and quantum mechanical modeling to drive the discovery and optimization of a new class of organic molecules for energy-efficient electrochemical energy storage and carbon capture.
University of Washington: Aniruddh Vashisth (PI), Yiwan Zheng (PhD student)
Microsoft: Bichlien Nguyen, Jake Smith, Kali Frost, Karin Strauss, Ziheng Lu, Shuxin Zheng, Jiang Bian

Despite encouraging progress in recycling, many plastic polymers often end up being one-time-use materials. The plastics that compose printed circuit boards (PCBs), ubiquitous in every modern device, are amongst those most difficult to recycle. Vitrimers, a new class of polymers that can be recycled multiple times without significant changes in material properties, present a promising alternative. This project will leverage advances in machine learning to select vitrimer formulations that withstand the requirements imposed by their use in PCBs.

Related paper:
- Inverse Design of Vitrimeric Polymers by Molecular Dynamics and Generative Modeling
University of Washington: Eleftheria Roumeli (PI)
Microsoft: Kristen Severson, Yuan-Jyue Chen, Bichlien Nguyen, Jake Smith

The concrete industry is a major contributor to greenhouse gas emissions, the majority of which can be attributed to cement, thereby making the discovery of alternative cements a promising avenue for decreasing the environmental impacts of the industry. This project will employ machine learning methods to accelerate mechanical property optimization of “green” cements which meet application quality constraints while minimizing carbon footprint.

Related paper:
- Amortized inference of Gaussian process hyperparameters for improved concrete strength trajectory prediction
Related article:

Environmental resilience

University of California, Irvine: Julie M. Schoenung (PI), Oladele Ogunseitan (Co-PI), Haoyang He (Postdoc)
Microsoft: Bichlien Nguyen, Winston Saunders, Maria Viitaniemi, Kali Frost

Materials play a crucial role in meeting global decarbonization and sustainability targets. Leveraging expertise across life-cycle assessment, material science, computational design, and policy, this project aims to accelerate sustainable development of materials, especially to advance data-driven, high throughput discovery of materials properties and sustainability attributes.
Universitat de Valencia: Gustau Camps-Valls (PI), Gherardo Varando (Co-PI), Jose Maria Tarraga (Scientific Researcher)
University of Reading: Ted Shepherd (PI), Ros Cornforth (Co-PI), Elena Saggioro (PhD student/Researcher), Genevieve Nartey (MSc student)
Microsoft: Trevor Dhu, Emre Kiciman, Lester Mackey, Eleanor Dillon, Ranveer Chandra, Peeyush Kumar, Swati Sharma

The Causal4Africa project will investigate the problem of food security in Africa from a novel causal inference standpoint. The project will illustrate the usefulness of causal discovery and estimation of effects from observational data by intervention analysis. Ambitiously, it will improve the usefulness of causal ML approaches for climate risk assessment by enabling the interpretation and evaluation of the likelihood and potential consequences of specific interventions.

Related papers:
MIT: Judah Cohen (PI), Dara Entekhabi (Co-PI), Sonja Totz (Postdoc)
Microsoft: Lester Mackey

Water and fire managers rely on subseasonal forecasts two to six weeks in advance to allocate water, manage wildfires, and prepare for droughts and other weather extremes. However, skillful forecasts for the subseasonal regime are lacking due to a complex dependence on local weather, global climate variables, and the chaotic nature of weather. To address this need, this project will use machine learning to adaptively correct the biases in traditional physics-based forecasts and adaptively combine the forecasts of disparate models.

Related research:

More sustainability projects