KLIWAS 
Seite 31 
6 Conclusions, Outlook and Summary 
KLIWAS 
Climatology 
of North Sea 
Fronts 
The new KLIWAS climatology of sea surface temperature (SST) and ocean colour 
(OC) fronts in the North Sea provides a reliable reference data set for the assessment 
of changes in frontal position, gradients, and seasonal variability due to climate 
change on the basis of satellite data. It is also a valuable addition to the new KLIWAS 
North Sea Climatology for oceanic and atmospheric in-situ data. The presented 
climatological data sets with annual, seasonal and monthly products clearly 
demonstrates that the GRADHIST algorithm can be used for SST as well as for OC 
data and can also be used with data from different sensors and from different oceanic 
areas. 
The GRADHIST algorithm is based on a combination and modification of the 
gradient algorithm of Canny (1986) and the histogram algorithm of Cayula and 
Cornillon (1992). The main principles of both algorithms have been preserved and 
three significant improvements have been added: (1) the Scharr-Operator (Scharr 
2004) for the computation of the gradient is used instead of the Sobel-Operator; (2) 
the use of an improved non-maximum suppression for optimising the fronts and 
finally (3) an iteration over different window sizes has been introduced when 
applying the histogram algorithm. The validation and testing of the new algorithm 
was carried out using both synthetic data as well as with case studies using real 
AATSR, AVHRR, MERIS and MODIS data (see Table 1). The GRADHIST 
technique provides the possibility to combine fronts derived from different sensors, 
however, the quality of the geo-location has to be taken into account, an aspect which 
is not discussed in the present paper. 
Statistical measures likes probability of occurrence, mean front gradient magnitude as 
well as direction and magnitude of mean front gradient vector, allow for a 
quantitative determination of the frontal positions and gradient strength. It also allows 
for the determination of their natural variability, i.e., annual or seasonal cycles can be 
derived from the climatological data sets. Furthermore, a SST and OC fronts can be 
compared and potential relationships between both types of fronts can be analysed. 
During periods of strong solar radiation in spring or summer for instance, SST 
gradients can be strongly smoothed but a river plume front can still be tracked by the 
difference in TSM or yellow substance concentration between sea water and river 
run-off. 
The compiled climatological data set demonstrates that the GRADHIST allows for an 
automated processing of comprehensive EO data volumes for front detection in the 
North Sea and other shelf sea areas. GRADHIST also enables the establishment of 
front maps as an operational downstream product which will help in monitoring 
potential change due to climate change or human impacts such as the installation of