Addressing and mitigating the effects of climate change requires a collective effort, leveraging our strengths across industry, government, academia and civil society. As we continue to explore the role of technology in advancing the art of the possible, we are launching the Microsoft Climate Research Initiative (MCRI)† This community of multidisciplinary researchers works together to accelerate groundbreaking research and transformative innovation in climate science and technology.
MCRI enables us to leverage Microsoft’s research skills and computational capabilities for deep and continuous collaboration with domain experts. To kick off this initiative, we focus on three critical areas in climate research where computational advances can drive important scientific transformations: overcoming decarbonization constraints, reducing uncertainties in carbon accounting, and assessing climate risks in more detail.
Through these joint research projects, we hope to develop and maintain a highly engaged research ecosystem that encompasses a diversity of perspectives. Researchers will provide transdisciplinary and diverse expertise, particularly in areas beyond traditional computer science, such as environmental sciences, chemistry, and a variety of engineering disciplines. All results of this initiative are expected to be made publicly and freely available to fuel even wider research and progress on these important climate issues.
“As researchers, we are excited to collaborate on projects specifically selected for their potential impact on global climate challenges. With Microsoft’s compute capabilities and the domain expertise of our employees, our complementary strengths can accelerate progress in incredible ways.”
– Karin Strauss, Microsoft
Microsoft researchers will collaborate with collaborators worldwide to explore priority climate-related topics and bring world-class innovative research to influential journals and venues.
Phase one collaborations
Real-time monitoring of CO . control progress2 and air pollutant observations with a physically informed transformer-based neural network
The change in CO. to understand2 emissions when measuring CO2 concentrations as done by satellites is very helpful in tracking the real-time progress of carbon reduction actions. current CO2 observations are relatively limited: methods based on numerical models have very low computational efficiency. The proposed study aims to develop a new method that combines atmospheric numerical modeling and machine learning to determine the CO. to distract2 emissions from satellite observations and data from ground monitor sensors.
AI-based Near-Real-time Global Carbon Budget (ANGCB)
Zhu Liu, Tsinghua University; Biqing Zhu and Philippe Ciais, LSCE; Steven J. Davis, UC Irvine; Wei Cao, and Jiang BianMicrosoft
Climate change mitigation will depend on a carbon emissions trajectory successfully reaching carbon neutrality by 2050. To this end, a global assessment of the carbon budget is essential. The AI-based, near-real-time Global Carbon Budget (ANGCB) project aims to provide the world’s first global carbon budget assessment based on artificial intelligence (AI) and other data science technologies.
Carbon Reduction and Removal
Computational discovery of new metal-organic frameworks for carbon capture
CO . remove2 of the environment is expected to be an integral part of keeping the temperature rise below 1.5°C. Today, however, this is an inefficient and expensive undertaking. This project will apply generative machine learning to the design of novel metal-organic frameworks (MOFs) to optimize for low-cost CO removal2 from air and other dilute gas streams.
An assessment of liquid metal catalyzed CO2 Reduction
the CO2 The reduction process can be used to convert captured carbon into a storable form and to produce sustainable fuels and materials with a lower environmental impact. This project will evaluate liquid metal reduction processes and identify benefits, bottlenecks and improvement opportunities needed to achieve industrially relevant scales. It will lay the foundation for improving catalysts and tackling scaling bottlenecks.
Computational Design and Characterization of Organic Electrolytes for Flow Battery and Carbon Capture Applications
Energy storage is essential to enable 100% CO2-free 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.
Proprietary Prediction of Recyclable Polymers
Despite encouraging progress in recycling, many plastic polymers often become single-use materials. The plastics that make up printed circuit boards (PCBs), which are ubiquitous in every modern device, are among the most difficult plastics to recycle. Vitrimers, a new class of polymers that can be recycled multiple times without significant changes in material properties, represent a promising alternative. This project will utilize advances in machine learning to select vitrimer formulations that can withstand the requirements imposed by their use in PCBs.
Accelerated discovery of green cement materials
The concrete industry is a major contributor to greenhouse gas emissions, most of which can be attributed to cement. The discovery of alternative cements is a promising way to reduce the environmental impact of the industry. This project uses machine learning methods to accelerate the mechanical property optimization of “green” cements that meet application quality requirements while minimizing the environmental footprint.
Causal inference to understand the impact of humanitarian interventions on food security in Africa
The Causal4Africa project will examine the problem of food security in Africa from a new point of view on causal inferences. The project will illustrate the utility of causal discovery and estimation of effects from observational data through intervention analysis. Ambitiously, it will improve the usability of causal ML approaches for climate risk assessment by enabling the interpretation and evaluation of the likelihood and potential impacts of specific interventions.
Improving sub-seasonal forecasting with machine learning
Water and fire managers rely on subseason forecasts two to six weeks in advance to allocate water, manage wildfires, and prepare for droughts and other extreme weather events. However, competent predictions for the subseasonal regime are lacking due to a complex dependence on local weather, global climate variables and the chaotic nature of the weather. To meet this need, this project will use machine learning to adaptively correct the biases in traditional physics-based predictions and adaptively combine the predictions of disparate models.