Evaluation of ERA-40 Reanalysis Data on a Regional 
Scale - Total Cloud Cover in the North Sea Area 
N. H. Schade, BSH, Hamburg, Germany, M. Stengel, R. Hollmann, DWD, Offenbach, Germany 
H. Heinrich, BSH, Hamburg, Germany, G. Rosenhagen, DWD, Hamburg, Germany 
Departmental Research Programme 
National Meteorological Service in Germany 
(DWD) 
German Maritime and Hydrographic Agency 
(BSH) 
German Federal Institute of Hydrology (BfG) 
German Federal Waterways Engineering 
and Research Institute (BAW) 
www.kiiwas.de 
Background 
■ Existing climate models have difficulties in reproducing the spatial variability, e.g. of clouds, in the North Sea area. Especially, no regional coupled atmosphere-ocean 
models (RCAOMs) are available to describe their complex interactions with solar radiation, that influence momentum, freshwater and heat fluxes, and subsequently the 
SST. Therefore, within the KLIWAS research programme, three RCAOMs are actually implemented for in-depth analyses. 
■ Reference data are needed for validation purposes, and can be based on high quality controlled in-situ observations, satellite data and/or reanalyses. Since marine in-situ 
observations are unevenly distributed in space and time, satellite data provide the sole homogeneous observational data source over the North Sea. 
■ In this study, we use long-term cloud property data sets derived by the EUMETSAT Satellite Application Facility on Climate Monitoring (CM SAF, www.cmsaf.eu) to 
investigate differences to ERA-40 reanalysis cloud cover data. ERA-40 is used to drive the hindcast runs of the regional models. 
The Data 
CM SAF AVHRR CDR 
(Satellite product) 
ERA-40 
(Re-analysis) 
Total Cloud Cover (TCC) 
Total Cloud Cover (TCC) 
Temporal resolution: Daily mean 
1982-2009 
Temporal resolution: 6H mean 
1957-2002 
Spatial resolution: 0.25°x 0.25° 
Spatial resolution: 1,125°x 1.121 ° 
Coverage: Global 
Coverage: Global 
CM SAF satellite-based climate data 
records (e.g. Karlsson etal., 2012) 
ECMWF 40 Year Re-analysis data 
archive (e.g. Uppala etal., 2005) 
CM SAF AVHRR CDR vs. ERA-40 | 
CM-SAF AVHRR ERA-40 Difference 
Fig-1: 0 .25° grid points marking the four investigated North Sea boxes (left). Warm water effect 
over the North Sea, January 6th, 2003, resulting in cloud formation: A VHRR composite of two 
visible and one infra red channel (right) 
Source: http://www.keesfloor.nl/artikelen/meteorol/lake-effect_noordzee, 30.04.2012. 
Fig.2: Mean CM SAF AVHRR CDR and ERA-40 cloud covers in % in the North Sea area for the 
period 1982-2002 (left, middle) and the difference (right) 
Fig.3: Left four graphs: Annual Cycle of total cloud cover from CM SAFA VHRR CDR satellite 
data (blue) and ERA-40 (back) for the period 1982-2002. The whiskers show the daily mean 
standard deviations. Right four graphs: Difference of the annual cycle (CM SAF - ERA-40) 
■ Mean Cloud Cover values show similar distribution over the investigated area (Fig.2) but about 
10 % more clouds for the satellite data, except for mountainous areas 
■ The annual cycle (Fig.3) shows more clouds in the winter and less in the summer months for 
both data sets but an obvious underestimation in the reanalysis TCC during the whole year with 
maximum difference in the summer months of about 10-12 % and better agreement in the 
winter months with about 5 %. Standard deviations of both data sets vary between 40 % in the 
winter and 70 % in the summer months. 
CM SAF AVHRR CDR cloud cover data have been validated against globally 
distributed synoptic observations on airports and other satellite products (available 
on demand: Contact.Cmsaf@dwd.de). Here, they are interpolated on the ERA-40 
grid to calculate mean values for the period 1982-2002, and are compared to ERA- 
40 cloud cover data in the four North Sea boxes displayed in Figure 1. 
Frequency Distributions (Fig.4) based on monthly means show a shift to higher cloud amounts 
for the satellite data with maximum values at 70-75 % TCC for ERA-40 compared to the 
satellite data peaking at 75-80 % CFG. Daily mean frequency distributions show a large 
underestimation of overcast situations and an overestimation of all situations between 30 and 
95 % TCC for ERA-40. 
Summary & Conclusion 
Fig.4: Frequency Distribution of CM SAF AVHRR CDR and ERA-40 monthly mean (left four 
graphs) and daily mean (right four graphs) data for the period 1982-2002 
■ For the validation of regional climate models, high quality observations, satellite data and/or reanalyses data, are needed as reference of the actual 
climate's conditions. 
■ Comparisons of daily mean CM SAF AVHRR CDR total cloud covers (TCC) with ERA-40 data show an underestimation of the reanalysis TCC of 
about 10 % in the monthly mean values, less in the winter, more in the summer months. This can be seen in all four North Sea regions and may be 
caused by the parameterisation of humidity processes in the reanalysis and/or the coarse resolution. 
■ Monthly mean frequency distributions show a shift to lower cloud amounts by ERA-40. Due to the pathways and the velocity of low pressure systems 
crossing the North Sea area, maximum TCC values are lower for monthly means than for daily means in both data sets. 
■ The mismatch in the overcast situations in daily mean frequency distributions may result from the modelled positions of the clouds / cloud layers 
inside the ERA-40 grid boxes. 
■ Further investigations will involve comparisons with in-situ measurements and solar global radiation data. The influence of ERA-40 driving regional 
models will also be in focus. 
Contact: 
Dr. Nils H. Schade 
German Maritime and 
Hydrographic Agency (BSH) 
Bernhard-Nocht-Strasse 78 
D-20359 Hamburg, Germany 
Tel.:++49 (0) 40 6690 1264 
Fax: ++49 (0) 40 6690 1954 
Email: nils.schade@bsh.de 
www.bsh.de / www.dwd.de 
Literature: Karlsson, K.G., et al. (2012): The CM SAF cloud and radiation data set from 28 years of global AVHRR data. In preparation. 
Uppala, S.M. and co-authors (2005): The ERA-40 re-analysis. Quart. J.R. Meteorol. Soc., 131, 2961-3012. 
# 1 Federal Ministry 
I of Transport, Building 
I and Urban Development 
4 th WCRP International Conference on Reanalyses 
7-11 Mai 2012, Silver Spring, Maryland, USA 
BUNDESAMT Fdi; 
SfiSCH Iff FAHRT 
UND 
HYDROGRAPHIE