Open AccessA New Statistical–Dynamical Downscaling Procedure Based on EOF Analysis for Regional Time Series Generation
Author(s) -
Yosvany Martinez,
Wei Yu,
Hai Lin
Publication year - 2013
Publication title -
journal of applied meteorology and climatology
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.079
H-Index - 134
eISSN - 1558-8432
pISSN - 1558-8424
DOI - 10.1175/jamc-d-11-065.1
Subject(s) - downscaling , empirical orthogonal functions , meteorology , geopotential height , wind speed , environmental science , terrain , series (stratigraphy) , time series , climatology , geology , mathematics , geography , statistics , precipitation , paleontology , cartography
A new statistical–dynamical downscaling procedure is developed and then applied to high-resolution (regional) time series generation and wind resource assessment. The statistical module of the new procedure uses empirical orthogonal function (EOF) analysis for the generation of large-scale atmospheric component patterns. The dominant atmospheric patterns (associated with the EOF modes explaining most of the statistical variance) are then dynamically downscaled or adjusted to high-resolution terrain and surface roughness by using the Global Environmental Multiscale–Limited Area Model (GEM-LAM). Regional time series are constructed using the model outputs. The new method is applied to the Gaspé region of Québec in Canada. The dataset used is the NCEP–NCAR reanalysis of wind, temperature, humidity, and geopotential height during the period 1958–2004. Regional time series of wind speed and temperature are constructed, and a numerical wind atlas of the Gaspé region is generated. The generated time series and the numerical wind atlas are compared with observations at different masts located in the Gaspé Peninsula and are also compared with a numerical wind atlas for the same region generated in Yu et al. The results suggest that the newly developed procedure can be useful to generate regional time series and reasonably accurate numerical wind atlases using large-scale data with much less computational effort than previous techniques.