Abstract
Previous climate research concluded that causal influences which have contributed to changes in frost risk in south-eastern Australia include greenhouse gas concentration, El-Niño southern oscillation and other effects. Some of the climatic indices representing these effects have spatiotemporal misalignment and may have a spatially and temporally varying effect on observed data. Other indices are constructed from grid-referenced physical models, which creates a point-to-area problem. To address these issues we use a spatiodynamic model, which comprises a blending of spatially varying and temporally dynamic parameters. For the data that we examine the model proposed performs well in out-of-sample validation compared with a spatiotemporal model.
Original language | English |
---|---|
Pages (from-to) | 755-778 |
Number of pages | 24 |
Journal | Journal of the Royal Statistical Society. Series C: Applied Statistics |
Volume | 64 |
Issue number | 5 |
DOIs | |
Publication status | Published - Nov 2015 |