Functional Data Analysis of the Trends and Predictors of Climate in BC

• Activity Number: 294
• Type: Topic Contributed
• Date/Time: Tuesday, August 3, 2010 : 8:30 AM to 10:20 AM
• Sponsor: SSC
• Author(s): Simon Joseph Bonner*+ and Nathaniel Kenneth Newlands and Budong Qian and Nancy Heckman
• Companies: The University of British Columbia and Agriculture and Agri-Food Canada and Agriculture and Agri-Food Canada and The University of British Columbia
• Keywords: Agriculture ; Bayesian hierarchical modelling ; Climate Change ; Functional data ; Teleconnections ; Spatial modelling

Abstract

Weather data are inherently noisy: temperature and precipitation levels vary greatly day-to-day and this may hide longer-term (monthly or seasonal) trends. Data collection may also be inconsistent and strong correlations exist between the observations from nearby sites. For these reasons, standard multivariate methods are difficult to apply and often identify extreme events as the major sources of variation over space and time. We apply functional data methods and hierarchical Bayesian modelling to analyse historical weather data from more than 1500 sites across British Columbia and to study the effects of space, time and global teleconnections on the yearly trends of temperature and precipitation. Results from this work will be used to model future changes in weather patterns across BC and to understand the possible effects of climate change on the province's agriculture industry.