AbstractA set of tools for the statistical assessment of ocean observing networks is presented and applied for the analysis of different instrumentation scenarios in the German Bight. An optimal linear estimator is used to re-construct ocean state parameters from observations taking into account both the prior distribution of the state and measurement errors. The proposed method enables a re-construction of any scalar parameter or vector field with linear relationship to the state. The performance of the observing network is quantified in terms of the re-construction quality. Apart from the capability of the network to provide estimates of state parameters at the time of the observations, the potential of the measurements for forecasts is investigated as well. Furthermore, a generic method to compare single measurements with continuous observations is presented. Finally, a technique is described to quantify the relative importance of different components of an observational network.
The proposed methods are applied to water level measurements in the German Bight. A numerical model is used to estimate the background statistics. Synthetic measurements provided by tide gauges, satellite altimeters, and HF radar are considered in the analysis. The estimation of the complete water level field in the German Bight is compared for altimeter and tide gauge measurements. It is shown that the orientation of the satellite track with respect to the coastline is of high relevance. The importance of water level measurements taken in deeper water, e.g., at the FINO-1 platform, is demonstrated. It is shown that continuous tide gauge measurements provide more information on the area mean water level in the German Bight than altimeter observations taken by ENVISAT and JASON-1/2. It is furthermore shown how the information provided by a tide gauge propagates with the Kelvin wave. Implications for the design of an assimilation scheme are discussed.