Abstract for Weekly Colloquium on Thursday, February 1, 2018 at 4 pm in Harshbarger 206 ~ Refreshments at 3:45 pm
Dr. Roger A. Pielke, Sr., is a Senior Research Scientist with the Cooperative Institute for Research in Environmental Sciences (CIRES) and is a Senior Research Associate in The Department of Atmospheric and Oceanic Sciences (ATOC), at the University of Colorado at Boulder, as well as a Professor Emeritus at Colorado State University
Paradigm Shifts in Atmospheric Sciences
There are several developments in atmospheric sciences that suggest major paradigm shifts in the study, interpretation and application of concepts in atmospheric science. The talk will overview four different such developments:
1. Current emphasis on reducing risk from threats from extreme weather in the coming decade have focused on what is referred to as "outcome vulnerability". This is based on downscaling climate change projections to regional and local areas and giving this information to policymakers and the impacts communities. However, the spectrum of possible risks are missed with this approach as exemplified by Hurricanes Harvey, Irma and Maria. A bottom-up, resource-based approach defined as "contextual vulnerability" will be presented. Examples will be shown.
2. The current view in atmospheric science is that sound waves, with respect to weather, are a curiosity that we only think about when thunder or an echo occurs. However, seminal work by Mel Nicholls of the University of Colorado at Boulder has shown that sound waves, which are dynamically forced (e.g. from tornadoes) and thermal forced (e.g. when diabatic heating occurs), are an important part of weather. Ignoring sound waves, in the interpretation of how weather occurs and filtering from numerical models, will be shown to ignore an important part of the fundamental physics of atmospheric processes.
3. Existing public perception and even with some in our professional community accept the claim that the flap of a butterfly can result in a tornado thousands of kilometers away as a result of the nonlinear chaotic behavior of the atmosphere.. Referred to as the "butterfly effect" the talk will discuss the origin of this term and show why it has been fundamentally misinterpreted.
4. The current approach in numerical weather prediction is to construct a physics core based on the pressure gradient force, advection, gravity and the Coriolis effect. Than cumulus clouds, radiative flux divergence, subgrid scale fluxes, etc, are parameterized using tunable coefficients and functions. However, concepts of artificial intelligence (AI) may permit us to not only replace the parameterizations with representations trained from real observed data, but even the physics core can possibly be eliminated with this AI methodology. This would also remove the issue of numerical error associated with grids intervals and time steps. If correct, we will see soon challenges in terms of skill between the traditional numerical modeling approach to weather prediction and this new AI methodology..