Climate Models: A Spatial Data Story Where Deep Learning Appears

• Activity Number: 261 - Statistical Learning for Environmental Data Science
• Type: Invited
• Date/Time: Wednesday, August 11, 2021 : 1:30 PM to 3:20 PM
• Sponsor: Section on Statistics and the Environment
• Author(s): Douglas Nychka*
• Companies: Colorado School of Mines
• Keywords: spatial process; climate model ; spatiotemporal

Abstract

Simulations of the motion and state of the Earth's atmosphere and ocean yield large and complex data sets that require statistics for their interpretation. Typically climate and weather variables are in the form of space and time fields and it is useful to describe their dependence using methods from spatial statistics. A companion problem is to analyze observational and remotely sensed data to match up with model simulations and assess model biases and uncertainty. Throughout these problems is the need for estimating covariance functions over space and time and accounting for the fact that the covariance may not be stationary. This talk gives some examples of how spatial data methods work their way into interpreting climate model output and climate data and focuses on a new computational technique for fitting covariance functions using maximum likelihood. Estimating local covariance functions is a useful way to represent spatial dependence but is computationally intensive because it requires optimizing a local likelihood over many windows of the spatial field. In this work we show how a neural network (aka deep learning) model can be trained to give accurate maximum likelihood