Research focuses primarily on the interdisciplinary application of engineering feedback analysis, dynamics, and control tools to problems in climate; principally solar climate engineering (aka geoengineering) and climate dynamics/variability. Additional interests include control of fluid dynamics, vibration and noise, and telescope control.
Climate Engineering refers to large-scale intentional intervention in the climate system as a possible additional tool to help manage some impacts of climate change; an example would be adding aerosols to the stratosphere to reflect some sunlight. This doesn’t reduce the need to cut greenhouse gas emissions, nonetheless deploying some amount of Climate Engineering might reduce climate damages, and more research is needed to evaluate it.
Main research questions:
- How can one “design” stratospheric-aerosol climate engineering, using available degrees of freedom (e.g., latitudes, times of year to inject material) to achieve desired objectives?
- What would the climate impacts be of different deployment choices?
- How uncertain are our predictions; what are the dominant uncertainties?
I am also interested in collaborations in social-science and governance of these technologies.
A few potential research opportunities are listed below; come talk to me if any of these sound interesting or you’d like more than the terse description below!
- We have recently conducted a 20-member ensemble of geoengineering simulations to evaluate robust regional responses and variability. A number of variables have been analyzed, some that haven’t been looked at in detail include sea ice, permafrost,…
- Using an emulator (reduced-order dynamic model), predict the response for different scenarios, including for example predicting the rate-of-change of temperature and precipitation to understand stressors on ecosystems, or predicting the impact of a 1-, 2-, or 3-year interruption in deployment.
- One of the next steps in research is to explore what happens if one only injected aerosols in a single season (and then use that information to combine injection at different seasons and latitudes to explore outcomes). This requires first conducting new climate model simulations to explore how the season of injection affects stratospheric and tropospheric climate
- If you optimize for one set of metrics, what happens to other metrics?
- How can we design multivariable feedback algorithms to simultaneously manage multiple metrics?
- Most simulations have injected SO2, which oxidizes over ~1 month to form sulfate aerosols. Injecting sulfate aerosols directly should give better control of the resulting aerosol size distribution. How much does this “help” (e.g., reduce the amount of sulfate needed, or improve the ability to compensate for climate change)
- Current simulations use feedback to manage surface temperature goals. Design and demonstrate feedback control of aerosol optical depth instead.
- How robust are the feedback algorithms used to date in Climate Engineering? Do they still work when the aerosol properties are changed? One option is to change the properties by injecting sulfate directly, another would be to change the parameters of the aerosol model.
- What would happen if there was a volcanic eruption during a stratospheric aerosol deployment?
- What would the first years of a stratospheric aerosol deployment look like? What is the signal-to-noise ratio for detecting aerosol optical density; what is the smallest initial deployment that would provide useful information, and what observations would be required? (This isn’t quite my expertise.)
- What other experiments would be useful to understand stratospheric aerosol geoengineering? (This definitely isn’t my expertise!)
- Another next step in research is to better assess the extent and impact of uncertainty, e.g., through perturbed physics simulations or multiple models. This is also a bigger project that would require computational support, but a plausible first example would be changing the parameters of the aerosol microphysical growth model and seeing how much of an impact that makes on surface climate.
(The ideas below aren’t current, though I’ll leave them here to spark ideas.)
- Geoengineering control design… papers such as this one (single input) and this one (multivariable) use a simple proportional-integral control; how should one design a strategy using more information about either the dynamics (requires estimating those from model output) or the state?
- Using System identification to assess regional climate response to Marine Cloud Brightening (geoengineering by “brightening” clouds in particular regions and seasons). Some preliminary simulations and analysis have been conducted, see paper here, but analysis so far has only scratching the surface. Nothing like this has yet been done in climate science!
- “Big data” and climate science: can we use ARGO float data to understand ocean eddies? The ARGO float trajectories provide massive (though sparse in space and time) data on currents, but most researchers have only looked at temperature/salinity profiles – can we build some spatiotemporal estimate of ocean eddies?
- Application of Engineering System identification tools to understand dynamics underlying ENSO or AMOC (requires funding for computational resources); a useful starting point is here.
- Efficacy – how do we compare different forcing agents in the climate system, such as comparing CO2 versus methane? One metric is how much of each cause the same change in global mean temperature; this can be efficiently computed with a feedback loop. The concept has been written down here, but there’s a lot of different forcing to consider, and different metrics that could be evaluated.
- If we sprayed sea water on top of existing sea ice in late fall and early winter to generate a thicker layer of ice, would this extend the life of the ice through the following summer?
- Holistic Assessment of SO2 injections into the stratosphere: can we combine aerosol injection at different latitudes to improve climate outcomes or meet specific goals? (A start towards “well-designed” geoengineering!) Joint with PNNL and NCAR.
- Geoengineering on a Regional Scale: With significant impacts projected from global warming and melting ice, the Arctic is a critical region for evaluating possible future global cooling techniques, such as injecting aerosols into the stratosphere to boost “albedo” and reflect some of the sun’s energy. Combining social science, engineering, and communication, this team will engage Arctic communities in a participatory discussion about these emerging technologies, identify public concerns, and evaluate regionally specific geoengineering strategies that address them. (Atkinson Center for a Sustainable Future, Academic Venture Fund, 2015, with Bruce Lewenstein and Holly Buck)
- How Do You Construct a Strategic Approach to Climate Change by Coupling Geoengineering to Mitigation and Adaptation? (Atkinson Center Impact through Innovation Fund, with Environmental Defense Fund)