Premium
A further analysis of the spatio‐temporal variability in aerosols across North America: Incorporation of lower tropospheric (850‐hPa) flow
Author(s) -
Sheridan Scott C.,
Power Helen C.,
Senkbeil Jason C.
Publication year - 2008
Publication title -
international journal of climatology
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.58
H-Index - 166
eISSN - 1097-0088
pISSN - 0899-8418
DOI - 10.1002/joc.1628
Subject(s) - geopotential height , environmental science , climatology , troposphere , aerosol , advection , atmospheric sciences , precipitation , synoptic scale meteorology , spatial variability , principal component analysis , tropopause , geopotential , meteorology , geography , geology , statistics , physics , mathematics , artificial intelligence , computer science , thermodynamics
In our previous research (Power et al. , 2006), we analysed the variability in spectral aerosol optical depth (τ a ), as well as Ångström's wavelength exponent (α), which represents the relative size distribution of aerosols, at 27 sites across North America. The variability was assessed by incorporating the Spatial Synoptic Classification (SSC) surface weather type across each of these sites. In this research, we further our analysis by examining the effect of different 850‐hPa flow patterns within the lower troposphere on variability in τ a and α. To accomplish this, an 850‐hPa flow classification was first developed from principal components analysis (PCA) and cluster analysis on National Centers for Environmental Prediction (NCEP)/The National Center for Atmospheric Research (NCAR) reanalysis geopotential‐height data, normalized by month of year. Nine flow types were identified. Cluster analyses yielded five clusters for both τ a and α. For τ a , spatially cohesive regions were identified, with a western cluster with low aerosol optical depth and minimal synoptic variability, and eastern clusters that showed higher turbidity as well as synoptic variability. In these clusters, higher turbidity was associated with southerly advection. For α, results were less cohesive, with clusters largely representing the different seasonal patterns of the variable, with synoptic variability difficult to interpret. Jackknifing was also utilized to evaluate whether τ a and α could be predicted based on SSC weather type and 850‐hPa flow type. Results varied at the three stations evaluated, with the method yielding potentially useful results for τ a across the two eastern sites, where synoptic variability at the surface and 850‐hPa is greater. At the western site, τ a could not be successfully predicted. Synoptic variability is too weak for α to be predicted effectively at any of the sites examined. Copyright © 2007 Royal Meteorological Society