Abstract:Solar geoengineering, also termed as solar radiation management or albedo modification, has been proposed as a potential means to mitigate anthropogenic global warming. Proposed solar geoengineering schemes include deployment of mirrors in space, increase of low-level cloud albedo, and enhancement of land or ocean surface albedo. One of the most widely discussed solar geoengineering methods is to inject sulfate aerosol or its gaseous precursors into the stratosphere to scatter more sunlight to space, and thus cool the Earth. To date, all studies on stratospheric aerosol injection have been climate modeling, engineering assessments, and social science research (including study of economic impacts, ethical/philosophical implications, governance issues, public perceptions). As an attempt to study fundamental processes related to stratospheric aerosol geoengineering, a team of scientists from Harvard University plans to release about 100 to 1000 grams of calcium carbonate particles into the stratosphere at an approximate height of 20 km above the Southwest United States as early as 2019. This experiment, named Stratospheric Controlled Perturbation Experiment (SCoPEx), would be the first field study intended to be relevant to solar geoengineering. Nature selected ten favorite feature stories in 2018: One of them is "First Sun-dimming experiment will test a way to cool Earth", which refers to SCoPEx. Calcium carbonate aerosols, instead of sulfate particles, will be used in this experiment because past research showed negative effect of sulfate aerosols on stratospheric ozone. In my view, the SCoPEx experiment is useful in several aspects. First, the direct observation of the fate of injected particles into the stratosphere would help to deepen our understanding of the microphysics (such as aerosol dispersion and aerosol size distribution and evolution) of injected aerosols and the effect on ambient stratospheric dynamics and chemistry. This observation is important for understanding fundamental sciences relevant to climate response to stratospheric aerosol injection. Second, observational results from SCoPEx can be used to bridge the gap between plume-scale and GCM (General Circulation Model)-scale aerosol processes. These results can be used to test GCM-simulated physical and chemical processes of injected stratospheric aerosols, and help improve model's ability in predicting climate response to stratospheric aerosol injection. Third, the insight gained from SCoPEx will not only be useful for solar geoengineering research, but in general provide new insights into our understanding of aerosol microphysics and its climate implication as well. Many useful insights into the climate response to solar geoengineering have been gained from climate modelling studies. However, model representation of climate physics and chemistry, in particular small-scale processes, rely on real world observations. Field experiments like SCoPEx will provide new opportunities for understanding climate response to stratospheric aerosol injection.

Stratospheric Aerosol Injection as a Deep Reinforcement Learning Problem

Deep reinforcement learning in World-Earth system models to discover sustainable management strategies

Small Scale Stratospheric Aerosol Injection Experiments Benefits Solar Geoengineering Research

Crafting desirable climate trajectories with RL explored socio-environmental simulations

Deep Reinforcement Learning for Greenhouse Climate Control

Integrated Sensing and Communications for Low-Altitude Economy: A Deep Reinforcement Learning Approach

How large is the design space for stratospheric aerosol geoengineering?

Stochastic control of geological carbon storage operations using geophysical monitoring and deep reinforcement learning

A Domain-Knowledge-Aided Deep Reinforcement Learning Approach for Flight Control Design

Deep Reinforcement Learning With Application to Air Confrontation Intelligent Decision-Making of Manned/Unmanned Aerial Vehicle Cooperative System

RAIN: Reinforcement Algorithms for Improving Numerical Weather and Climate Models

Using reinforcement learning to improve drone-based inference of greenhouse gas fluxes

AI for Global Climate Cooperation: Modeling Global Climate Negotiations, Agreements, and Long-Term Cooperation in RICE-N

Kicking the Can Down the Road: Understanding the Effects of Delaying the Deployment of Stratospheric Aerosol Injection

Injection strategy – a driver of atmospheric circulation and ozone response to stratospheric aerosol geoengineering

Tradeoffs When Considering Deep Reinforcement Learning for Contingency Management in Advanced Air Mobility

Deep Reinforcement Learning for Artificial Upwelling Energy Management

Optimization of Pressure Management Strategies for Geological CO2 Sequestration Using Surrogate Model-based Reinforcement Learning

Assessing GFDL‐ESM4.1 Climate Responses to a Stratospheric Aerosol Injection Strategy Intended to Avoid Overshoot 2.0°C Warming

An Intellectual Aerodynamic Design Method for Compressors Based on Deep Reinforcement Learning

Green Deep Reinforcement Learning for Radio Resource Management: Architecture, Algorithm Compression and Challenge