Statistics-Based Compression of Global Wind Fields

• Activity Number: 665 - Spatial Methods for Weather, Climate, and Health
• Type: Contributed
• Date/Time: Thursday, August 3, 2017 : 10:30 AM to 12:20 PM
• Sponsor: Section on Statistics and the Environment
• Author(s): Jaehong Jeong* and Stefano Castruccio and Paola Crippa and Marc G. Genton
• Companies: King Abdullah University of Science and Technology and Newcastle University and Newcastle University and KAUST
• Keywords: Compression ; Nonstationarity ; Spatio-temporal covariance model ; Sphere ; Surface wind speed

Abstract

We propose a statistical model that aims at reproducing the data generating mechanism of an ensemble of runs by providing a stochastic approximation of global annual wind data and compressing all the scientific information in the estimated statistical parameters. We introduce an evolutionary spectrum approach with spatially varying parameters based on large-scale geographical descriptors such as altitude to better account for different regimes across the Earth's orography. We consider a multi-step conditional likelihood approach to estimate the parameters which explicitly accounts for nonstationary features while also balancing memory storage and distributed computations, and we apply the proposed model to more than 18 million spatio-temporal data of yearly global wind speed. The proposed model achieves compression rates which are orders of magnitudes higher than traditional algorithms for yearly-averaged variables, and once the statistical model is fitted, decompressed runs can be almost instantaneously generated to better assess the wind speed uncertainty due to internal variability.