Event Scheduled for Mar 8, 2013
Event: Cancelled - Environmental Engineering Colloquium - “Climate Extremes: The Science, Impacts, and Policy Relevance”
Location: FLC, Room 212
Time: 12:00 pm
Details of Event:
“Climate Extremes: The Science, Impacts, and Policy Relevance” presented by Dr. Auroop Ganguly, Associate Professor, Civil & Environmental Engineering Department, Northeastern University.
Abstract: Climate extremes may be defined inclusively as severe hydrological or weather events, as well as significant regional changes in hydro-meteorology, which are caused or exacerbated by climate change, and which may in turn cause severe stresses on regional resources, economy and the environment. While regional warming and heat waves, and perhaps heavy precipitation, can be attributed to climate change with a degree of credibility and projected relatively reliably, significant uncertainties continue to exist for regional hydrology, including floods and soil moisture, as well as tropical cyclones or hurricanes and droughts. Arguably the most significant knowledge-gap in climate science relevant for informing stakeholders and policy makers is the inability to produce credible assessments of local- to regional-scale climate extremes. Recent results from the new generation of global climate model runs do not suggest the possibility of significant improvements in the near future, while regional climate models remain a promise. On the other hand, climate-related data, from archived model simulations, and remote or in situ sensors, have already moved into the petabyte scale and are projected to reach 350 PB by 2030. Thus, hypothesis examination, generation of novel insights, and perhaps even hypothesis generation may need to leverage computational methods that are adapted to handle massive and complex data. The methods can succeed in a climate extremes context if they are adapted to handle data that are not just massive but also generated from nonlinear processes, exhibit both proximity-based space-time correlations as well as long-memory and long-range dependence in time and space, and focus on predictability and predictive analysis at smaller scales and for rare or extreme events and change. Statistical approaches like extreme value theory have not been well developed for a majority of climate extremes. Nonlinear dynamical approaches are better at characterizing the climate system rather than generating projections, and even so are not well-developed for predictability assessment in climate. Traditional spatial and spatiotemporal data mining in computer science, while well-suited to certain kinds of geographic data, cannot handle the complex dependence structures, low frequency variability, and nonlinear data generation processes relevant for predicting climate extremes. The barriers are particularly challenging given the so-called deep uncertainties in climate arising from both natural variability in the climate system, e.g., from oceanic oscillators, combined with our lack of understanding of the relevant processes, e.g., convection, and inability to encapsulate them within models. Thus, regional shifts in climate patterns, or the behavior of extreme events over time and space, remain difficult to understand or project, especially at decadal or regional scales. In particular, some of the climate extremes that are of maximum concern are precisely the ones that are not well predicted. On the other hand, uncertainties in climate extremes propagate to consequence analysis. Uncertainty reduction may depend on the ability to leverage maximum information content from climate model simulations, reanalysis outputs, and sensor-based observations. In addition, the inherent variability in climate and impacted systems motivates a characterization of cascading uncertainty. This presentation will discuss the science, impacts and policy relevance of climate extremes and their uncertainties.
Biosketch: Auroop R. Ganguly is an Associate Professor of Civil and Environmental Engineering at Northeastern University, with 15 years of full-time experience spanning academia, a DOE national laboratory, and the private sector, specifically Oracle Corporation and a semi-startup. His research on climate extremes have been published in journals such as Nature Climate Change and Proceedings of the National Academy of Sciences, highlighted in journals such as Nature and Nature Climate Change, and reported by the mainstream media in the United States and across the globe. Over the last 9 years, he has focused on climate extremes and uncertainty, predictability and nonlinear dynamics, as well as global water sustainability. He has industry and research experience in data-intensive sciences, with an interest in extremes and nonlinear space-time processes. He has published an edited book on Knowledge Discovery from Sensor Data and papers in journals such as IEEE Transactions, and won best paper awards in computer science. Ganguly’s research, and his Sustainability and Data Sciences Lab, were highlighted in October 2012 by the National Science Foundation in their national newsletter, Current, in “Faces of NSF Research”. He serves as an Associate Editor for AGU Water Resources Research and ASCE Journal of Computing in Civil Engineering, in the AMS AI committee, and has been invited for review panels of DOE, NSF, and United Nations. He has a Ph.D. in Hydrology from MIT and a B. Tech. (Hons.) in Civil Engineering from IIT, Kharagpur, India.
Target Audience: Open to All
Sponsored By: Environmental Engineering Program
No Pamphlet/Flyer Available