Low (web) Quality - BALTEX

171
Analysis of surface air temperatures over Ireland from re-analysis data and
observational data
Priscilla A. Mooney 1 , Frank J. Mulligan 2 and Rowan Fealy 1
1 Irish Climate Analysis and Research Units, National University of Ireland Maynooth, Kildare, Ireland.
priscilla.a.mooney@nuim.ie
2 Department of Experimental Physics, National University of Ireland Maynooth, Kildare, Ireland.
1. Abstract
Since the observational data used in both ERA-40 and
NCEP/NCAR Reanalysis data are largely identical, it is
frequently assumed that either dataset is equally valid in the
assessment of climate models. However, comparisons of
these two dataset shows that subtle differences do exist and
it is important to compare them to observational data in the
region of interest to determine which reanalysis dataset
should be used.
In this study, surface air temperatures at 2 m altitude
predicted by ERA-40 and NCEP/NCAR reanalysis (NNRP-
1) have been compared with observations at eleven synoptic
stations in Ireland over the period 1958-2000. Both
reanalysis datasets show good agreement with the observed
data and with each other. Slopes of the least-squares line to
scatter plots of reanalysis data to observational data show
small differences between the two reanalyses, with NNRP-1
slopes and ERA-40 slopes ranging between (0.75-0.97)
±0.01 and (0.79-1.0) ±0.01, respectively. Summary
statistics and the monthly mean temperatures over the 1979-
2000 period showed that both reanalyses predicted
significantly warmer winters than the observations which
caused the slopes of the best fit lines to be consistently less
than unity.
2. Introduction
Over the past decade reanalysis data has found widespread
application in many areas of research ranging from studies
of climatic trends (Ciccarelli et al, 2008) and climate
modeling (Fealy and Sweeney, 2007) to estimation of
renewable energy resources (Henfridsson et al, 2007). It is
advantageous to use reanalysis data in certain research areas
where observational data is sparse or when knowledge of the
state of the atmosphere on a uniform grid is required.
Some of the most well known reanalysis data sets are
NCEP/NCAR (NNRP-1), NCEP/DOE (NNRP-2), ERA-15,
ERA-40 and ERA interim. NNRP-1 and ERA-40 are two of
the most widely used reanalysis archives. Although they use
similar observational data, previous studies have shown
subtle differences between them (Escoffier and Provost,
1998; Li et al., 2004; Gleiser et al., 2005; Ruti et al., 2008).
Simmons et al., 2004 compared surface air temperature
anomalies from CRUTEM2v with ERA-40 and NNRP-1.
They found that ERA-40 showed better agreement with
CRUTEM2v than NNRP-1 over the time periods 1958-2001
and 1979-2001. In addition they reported closer agreement
between ERA-40 and CRUTEM2v over the period 1979-
2001 compared with the period 1958-1979. This was
attributed to the greater observational coverage after 1979.
3. Surface Temperature Datasets
NNRP-1 reanalysis data was produced by the National
Centers for Environmental Predication (NCEP) in
collaboration with the National Center for Atmospheric
Research (NCAR). The assimilation system used in the
NCEP/NCAR reanalysis is described in detail in Kalnay et
al., 1996. The NNRP-1 data used in this study was obtained
from the NOAA/OAR/ESRL PSD, Boulder, Colorado,
USA web site at http://www.cdc.noaa.gov/.
ERA-40 data was produced by the European Center for
Medium range Weather Forecasting (Uppala et al, 2005)
in collaboration with several other institutions with an
interest in climate analysis and weather forecasting. The
ERA-40 data used in this study was obtained from the
ECMWF data server (http://data.ecmwf.int/data/).
The observed station data used in this study for the period
1958-2000 was obtained from Met Éireann, the Irish
national meteorological service, for 11 synoptic weather
stations. The synoptic stations, which are geographically
dispersed around Ireland (figure 1), represent mixture of
both coastal and inland locations. This study did not
perform any homogeneity analysis of the data. However,
the data is from the synoptic network which is manned by
experienced meteorological officers and the data is
considered to be of high quality.
Figure 1. Map of Ireland showing the location of the
eleven meteorological stations and the marine buoys M1
and M4. The map is overlaid with grids for NNRP (red ---
line) and ERA-40 (blue… lines).
4. Collocation Method
Spatially, the parameter values in the reanalysis data sets
represent the value of that parameter in a ‘grid box’
centred on the geographic coordinates given in the dataset.
This presents a difficulty when comparing the station data
with the reanalysis data, since the station data represents a
single site within a grid box. Additional complications are
introduced by the fact that the grid sizes in ERA-40 and
NNRP-1 are different, the grid centres in the two
reanalysis data sets are not coincident and the location of a
station in the grid box can vary from the centre of the grid
box to its edge. This is illustrated in figure 1 where the
grids used by ERA-40 and NNRP-1 are plotted together
with the location of the stations and marine buoys.