Mitchell Lecturer 2014

Doug Nychka will be the 2014 Mitchell lecturer.  

Doug is the Director of the Institute for Mathematics Applied to Geosciences (IMAGe) and also a Senior Scientist in Geophysical Statistics Project (GSP). He describes his main task as one to enrich the scientific and educational activity at NCAR through mathematical methods and models, and to use the large scientific projects at NCAR to engage the mathematical science communities in new applications and to motivate new mathematics and statistics.  

He will be visiting the Statistics group from 5th to 10th May and will give two lectures, details as follows:

Tuesday 6th May at 3pm in 203 - Uncertain weather, uncertain climate.

What will the weather be tomorrow? How cold was it 500 years ago?

The first question has a clear relevance to our daily lives and the second is necessary to understand variation in our Earth's climate.  Answers to both of these questions rely on statistical methods that combine observations with geophysical models to understand our physical environment.  Annual temperatures many centuries in the past can be estimated without the benefit of having direct measurements from thermometers. Should we be skeptical of scientific attempts to do this? This lecture will present statistical methods for scientific problems where observational information is limited and characterizing the uncertainty in the results is important. These methods, known as Bayesian hierarchical models, have become a mainstay of data analysis for complex problems and besides being used in the geosciences have wide application in other areas of science.

Thursday 8th May at 3pm in 203 - Multi-resolution spatial methods for large data sets.

Spatial data is ubiquitous and a basic problem is to reconstruct surfaces from irregular observations or measurements and to quantify the uncertainty in the estimates. Standard statistical methods break when applied to large data sets and so alternative approaches are needed that balance changes to the statistical models for increases in computational efficiency. A useful method expands the field in a set of compact basis functions and places a Gaussian Markov random field latent model on the basis coefficients. The impact is that evaluating the model likelihood and computing spatial predictions is feasible even for tens of thousands of spatial observations on a single computational core (e.g. a laptop). Moreover, by varying the support of the basis functions and the correlations among basis coefficients it is possible to entertain multi-resolution and nonstationary spatial models that exploit the rich structure often found in large data sets.