Transcript The local response to the NAO in a RegCM 30-year run

The local response to the NAO in a
RegCM 30-year run
Roxana Bojariu and Liliana Velea
National Institute of Meteorology
Bucharest, Romania
E-mail:[email protected]
Large scale NAO features
Data
• Simulated data:
•
– air surface temperature, precipitation and sea level pressure
from the control run of RegCM forced by the HadCM 3 run (with
observed SSTs and sea-ice for the interval 1960-1990).
– Resolution 50 km
– 119 grids in longitude and 98 in latitude (PRUDENCE domain)
– Winter (December to February)
Observed data:
– air surface temperature and precipitation from CRU (1960-1990)
– air surface and temperature from 59 stations over the Romanian
territory (1960-2000)
Analysis methodology
Canonical correlation analysis (CCA):
1. The CCA selects a pair of spatial patterns of two variables such
that their time evolution is optimally correlated (Preisendorfer
1988; Zorita et al. 1992; Bretherton, 1992; Kharin 1994; Von
Storch 1995).
2. Before canonical correlation analysis, the original data are
usually projected onto their Empirical Orthogonal Functions
(EOFs), retaining only a limited number of them in order to
minimize the noise.
3. The CCA patterns are normalized such that the coefficients
have standard deviation units, so the patterns represent typical
anomalies in their specific units.
1st CCA CRU Data
Air surface temperature (˚C) and SLP (hPa)
anomalies
Precipitation (mm/day) and SLP (hPa)
anomalies
Hurrell’s NAO index (black) and the time series (green)
associated with the 1st CCA of SLP and T (CRU data)
r=0.70
CCA patterns of air surface temperature
CRU r=0.93 vslp=40% vT=38%
RegCM r=0.95 vslp=50% vT=32%
CCA patterns of precipitation
CRU
r=0.93 vslp=42% vp=26%
RegCM r=0.98 vslp=43% vp=31%
The local response to NAO type variability over Romanian
territory
RegCM: 1st CCA SLP/T
Observations: difference of high and
low NAO composites
Conclusions
• The data simulated by RegCM capture features of the
local response to NAO type circulation.
• In this context, the downscaling of climate change
scenarios becomes more reliable for European area.
Follow up
• The analysis of other sources of variability for the
European regions (e.g. Eastern Atlantic pattern).
• The analysis of other fields (e.g. snow cover)
• The analysis of climate change scenarios.