Education and Appointments
- Cornell University, B.A. in physics, 1988
- California Institute of Technology, M.S. in planetary science, 1990
- California Institute of Technology, Ph.D. in planetary science, minor in physics, 1994
- NASA/Jet Propulsion Laboratory, post-doc and scientist, 1994–2004
- Danish Meteorological Institute, visiting scientist, 2000–2001
- Harvard School of Engineering and Applied Science (Anderson Group), project scientist, 2004–present
My professional interests have taken me from planetary science to internal gravity waves and their generation to radio occultation to climate monitoring and finally to Bayesian inference. I remain interested in all of these topics. Most recently, I have focused on the information content in climate benchmark data types.
Climate benchmark data types have the unique property that they can be used to infer climate change with nearly absolute certainty by comparing to future climate benchmarks. Climate benchmarks are established by empirical determination of observational uncertainty while the observations are being made and entails traceability to international standards of units. Not many data types can be turned into climate benchmarks. Climate benchmarking marks a change in the climate monitoring paradigm from one dependent on climate data records, which is based on the assumption of stability of calibration, to one based in SI traceability.
“What is a climate benchmark?”, you ask. It is a measurement made in space with a chain of calibration back to the standard that defines its unit of measure. And you must fully account for all uncertainties along the way, preferably empirically. I have experience with GPS radio occultation. In a radio occultation, you can measure the Doppler shift of GPS signals as they propagate through the Earth’s atmosphere. The only reason you get Doppler shifts in the signal, other than ordinary vacuum propagation, is that the atmosphere bends the GPS signals downward. The measurement is one of timing, the units of which are inverse seconds, or Hertz, so the measurement must be calibrated against the international definition of the second by a chain of comparisons. This is done using atomic clocks, which are traceable to the international definition of the second with accuracy easily sufficient for climate monitoring. Of course, Doppler shifts are not terribly exciting, but we can easily invert them to get vertical profiles of the atmosphere’s index of refraction. The index of refraction is, for the most part, determined by atmospheric density. You can do a lot with profiles of atmospheric density, including measuring thermal expansion of the atmosphere. Another great example of a climate benchmark is the Earth’s emitted thermal infrared spectrum. My colleague John Dykema focuses on it among other things. Here is a more competent description of climate benchmarking.
But what can be learned from measured climate change in these climate benchmark data types? Most of the research community interested in multi-decadal climate change would like to know the equilibrium sensitivity of climate, or how much surface air temperature would increase with a doubling of carbon dioxide. (They would also like to know how rapidly oceans soak up atmospheric heat.) Some are satisfied with determinations of equilibrium sensitivity derived from paleoclimatic data. I’m not, though, because it is really difficult to prove the accuracy of paleoclimatic records. Thus, I’d rather see more credible determinations made with modern instruments as soon as possible. Theoretically, this can be done with climate benchmark data types in space, particularly with thermal infrared spectra and a shortwave measurement. This is the foundation of the Climate Absolute Radiance and Refractivity Observatory (CLARREO), a multi-spacecraft NASA mission slated for launch in 2017 and 2020.
The methods used to interpret climate benchmark data types are very much the same as those used in climate signal detection and attribution, which is what researchers do to figure out who and what is responsible for climate change. Eventually, methods of data analysis must be directly linked to climate prediction. Bayesian inference, the ultimate statistical implementation of the scientific method, should provide the answer, and optimally so. I expect the frontiers of climate change analysis to be based in Bayesian inference, so that’s what I’m working on in close collaboration with Yi Huang.
Leroy, S.S., and J.G. Anderson, 2010: Optimal detection of regional trends using global data. J. Climate, In Press. [preprint]
Huang, Y., S. Leroy, P.J. Gero, J. Dykema, and J. Anderson, 2010: Separation of longwave climate feedbacks from spectral observations. J. Geophys. Res., In Press.
Leroy, S.S., Y. Huang, and J.G. Anderson, 2009: Radio occultation data: Its utility in NWP and climate fingerprinting. Proceedings of ECMWF Seminar on Diagnosis of Forecasting and Data Assimilation Systems, 7–10 September 2009, Reading, United Kingdom. [reprint]
Ho, S.-P., G. Kirchengast, S. Leroy, et al., 2009: Estimating the uncertainty of using GPS radio occultation data for climate monitoring: Intercomparison of CHAMP refractivity climate records from 2002 to 2006 from different data centers. J. Geophys. Res., 114, doi:10.1029/2009JD011969. [reprint]
Leroy, S.S., J.G. Anderson, and G. Ohring, 2008: Climate signal detection times and constraints on climate benchmark accuracy requirements. J. Climate, 21, 841–846. [reprint]
Leroy, S.S., J.G. Anderson, J.A. Dykema, and R.M. Goody, 2008: Testing climate models using thermal infrared spectra. J. Climate, 21, 1863–1875. [reprint]
Leroy, S.S., J.G. Anderson, and J.A. Dykema, 2006: Testing climate models using GPS radio occultation: A sensitivity analysis. J. Geophys. Res., 111, D17105, doi:10.1029/ 2005JD006145. [reprint]
The Inner Postmodernist
Very often people stop me in the streets, interrupt my haircuts, or sidetrack me from washing pots in the kitchen with a question like, “So...do you believe in global warming?” My knee-jerk reaction is to respond, “It shouldn't be a matter of belief,” but mostly I'm more patient than that. In considering a response, I feel it is a mistake for a scientist to even remotely come off as an environmental advocate when discussing his/her work as it can only lead to loss of credibility of our profession in the long run. Instead, I simply steer my inquisitor toward specifics. Climate change isn't just one question. To the public, it is really three questions:
- Is climate changing?
- Are humans responsible?
- Can we predict climate change?
Well, the answers are “yes,” “highly likely,” and “rather poorly.” My work dwells on the last question. With appreciation for the integrity of my inquisitor, I pose my answers in this framework and find that he/she inevitably responds positively.
The question of belief doesn't really go away, though. I highly recommend the book Scientific Method in Practice by Hugh Gauch on the topic of belief and objectivity in science. Here is one way it arises in my business. Once one appreciates the seriousness of what’s happening out there, there is a strong and understandable feeling that something must be done. How does one determine what must be done? Different people and societies place different values on the preservation of species, sea level rise, the disappearance of the winter Olympics, etc. Many are rightly concerned about the spread of infectious disease, desertification, etc. Then again, some are more interested in human survival beyond age 30 than in the survival of harlequin frogs. Where are your values? What do you believe? The response is not strictly an objective one, but it is still important. Science doesn’t do everything for us.