12 
(3) wave development in the Northeast Atlantic, North Sea, and Baltic Sea (Sea State 
Forecasting Service), and 
(4) the drift and dispersion of substances and floating objects in the North and Baltic Seas 
(drift forecasts). 
Circulation data from the operational model for the North and Baltic Seas are also made 
available for the monitoring tasks under the BSH GOOS sub-programme BSH-P1. 
BSH in its capacity as the German Oceanographic Data Centre (DOD) collects the 
oceanographic data provided by German institutions and agencies and makes them available to 
users requesting such data ( not included are geological7geophysical data and modelling results 
from circulation models of the oceans and atmosphere, which are being archived at Alfred- 
Wegener-Institut in Bremerhaven and Deutsches Klimarechenzentrum (German Climate 
Computer Centre) in Hamburg). 
Present scientific knowledge 
The prediction quality - using as criterion the probability of providing an accurate prediction 
within a particular time interval - does not always satisfy user requirements. An improvement 
of the prediction quality can only be achieved by means of innovative or improved prediction 
tools. However, the extent to which improvements have actually been achieved can be 
assessed on the basis of accurate, specific routine measurement data from the monitoring 
networks (e.g. MARNET). 
The model results of the "operational model”, which is forced by realistic data, provide 
complete time series of the current dynamic status (water levels and currents) as well as 
temperature and salinity throughout the North and Baltic Seas. These data, provided that they 
have been checked against data from the monitoring networks or remote sensing systems, can 
be used as basic data for integral statements (e.g. exchange of water masses, heat content of 
the surface layer, frequency distribution and spectra of individual ocean dynamic parameters). 
They complement measured data and may be substituted, where necessary, for oceanographic 
variables that are difficult or impossible to measure, e.g. currents in navigation channels, 
temperature in very deep layers etc. 
Monitoring and investigation methods, analytical methods 
The tools used in forecasting services and in status assessments employ time-tested methods 
based on scientifically valid data. Short-term meteorological forecasts of the German Weather 
Service are required, as are environmental surveillance data from vessels, satellites, and the 
MARNET monitoring network. Besides, the measured data are used to verify and improve the 
predictions. 
In the following, the data and methods used by the individual services are briefly described. 
Water level prediction and storm surge warning services 
Water levels and storm surges are predicted on the following basis: 
- wind and water level data from the national BSH MARNET monitoring network and from 
foreign networks 
- the results of different prediction tools, based on numerical, statistical models and neural 
network models. 
- a computer-controlled system fed with measured data and with the results of independent 
prediction tools serves as a decisionmaking aid helping the scientist on duty to make