Coffee Talks Archive

Return to current Seminars and Coffee Talks Schedule

Global climate models (GCMs) are becoming increasingly important in the simulation of climate changes associated with increases in greenhouse gas concentrations in the atmosphere. However, little, if anything has been said about the effects of increased anthropogenic forcing on severe convective weather. Recent improvements in global climate model resolution and increased data storage capabilities have allowed for climate simulations of resolutions similar to those of the NCAR / NCEP global reanalysis to be archived at six hour intervals. Brooks et al. (2003) demonstrated that probability of occurrence of significant severe convective weather increases with increasing Convective Available Potential Energy (CAPE) and 0 to 6 kilometer deep layer shear. Additionally, Brooks et al.(2003) went on to demonstrate that the NCAR / NCEP global reanalysis data can be used as a surrogate for observational fields of CAPE, deep layer shear, and their combination. The methodology used by Brooks et al. (2003) serves as the basis to examine changes in the severe thunderstorm environment in North America under global warming as simulated by the NCAR CCSM3 global climate model.

A comparison of the current CCSM3 severe thunderstorm environment to that of the global reanalysis data will be presented to illustrate the ability of the CCSM3 to simulate the current severe thunderstorm environment qualitatively. This will be followed by a preliminary examination of the severe weather environment at the end of the 21^st century simulated by the CCSM3 under the A2 SRES emission scenario. The focus will be on comparing CCSM3 and global reanalysis distributions of CAPE, deep layer shear, and their combination. This will be followed by a brief discussion of projected changes in the probability of occurrence of severe thunderstorm environments in North America.

Uncertainty is ubiquitous in climate science, impacts and policy. This talk begins with a discussion of the nature and sources of uncertainty, strategies that can be used to characterize and deal with uncertainty, and cognitive issues that can complicate and bias judgments about uncertainty. Then it will turn to a discussion of uncertainty in the context of: 1) climate forcing; 2) the resulting changes in climate; and 3) the impacts that result from those changes. The final portion of the talk will argue that while there are many things that are unique about climate problems, irreducible uncertainty is not one of them. The talk will conclude with some discussion of implications for climate-related decision-making and public policy.

At the edge, far from the centers of power, are the hundreds of thousands of people blasted from places that, while far from perfect, are home. They are scattered and unorganized, living day-to-day. Where are their voices? They alone know the salient qualities of place that need to be preserved, reimagined, or redefined to create more livable and resilient Gulf Coast futures. A reconstruction process that values local culture, geography, construction techniques, and psychology will yield neighborhoods rather than infrastructure. Empowering a bottoms-up recovery from disaster can restore and enhance the legitimacy of all levels of government.

The recovery of Galveston following America’s most deadly hurricane is both remarkable and especially appropriate at this time. The September 1900 storm devastated the fabric of what was a vibrant and cosmopolitan Gulf Coast city. Like the images from New Orleans, corpses and filth filled the streets with stench and the serious threat of serious disease. Historian’s Patricia Bixel and Elizabeth Turner have described how, with federal support and local leadership, Galvestonian’s turned their city into “a laboratory of sorts, a testing ground for new ideas about government, society, and technology”. Drastic solutions included raising the elevation of a large area of the island, the lifting and restoration of over two thousand historic structures, and the construction of an exceedingly robust seawall that has protected the city from subsequent storms. The recovering city was also a nexus for a variety of progressive changes in the structure of local government, the role of women in shaping public policy, and public health reforms. Even today, the Galveston experience should remind us of the power of citizens to mobilize around love of community and the power to create their own future.

This talk describes the prototype design for community collaboration aimed at sustaining livability and resilience at a time of rapid economic, social, and environmental change.

Common tools used by flood planners to assess a location's potential for extreme precipitation are return-levels maps. The n-year return level is the (precipitation) amount which is exceeded on average once every n- years. The maps supply information about potential extreme precipitation by providing return level estimates for locations in the study region. Traditionally, these maps have not provided uncertainty estimates with the return levels.

There is a current effort by the National Weather Service (NWS) to produce updated return levels maps. Using the Regional Frequency Analysis (RFA) methodology of Hosking and Wallis, the NWS has produced return levels maps and uncertainty estimates for two regions in the US.

We have developed an alternative methodology to produce precipitation return levels maps along with uncertainty estimates. Using this methodology, we have produced a precipitation map for the Front Range region of Colorado. To develop the map, we have relied on the theory of extreme values. Specifically, we have used the generalized pareto distribution (GPD) to model precipitation above a threshold at 56 weather stations throughout the region. By constructing a Bayesian hierarchical model which relates each station's GPD parameters to a spatial model we are able to pool the data from all the stations and obtain parameter and return-level estimates which have more spatial consistency. The parameter estimates also take into account the available covariates we have for the model.

In this talk, I will outline our methodology and discuss the differences between our Bayesian framework and the RFA methodology. I hope that the discussin will bring forth new ideas on ways to improve the model and critique what the model (and map) may be missing.