Accuracies and error statistics for measured waves under various met ocean conditions — Area covering X-band radar wave measurements vs point measurements

Characteristics of the sea state are generally described by statistical parameters. Even though these parameters are known to be limited in describing the actual wave field, they form the basis for many off-shore decisions and safety regulations. Furthermore it is under discussion how accurate these parameters can be measured and how well they agree when comparing data from different wave sensors. Point measurement devices (buoys, pressure sensors, etc.) have been established in the scientific community for decades and their statistical values are well trusted. Nevertheless their application is limited and new wave measurement devices have been developed such as wave monitoring radars (HF-Radar, X-Band Radar, etc.). As these wave sensors use different sampling strategies and analysis algorithms, they allow a different view on ocean waves and make comparability challenging. In this paper an error statistic for statistical sea state information obtained by point measurements (buoy) and areal measurements (nautical X Band radar) will be presented. In contrast to point measurements, which contain limited directional wave information, nautical radar images include a spatial view on the ocean wave field. To evaluate the sensor´s independent expected variations of Hs, firstly buoy raw data are analyzed with respect to different sampling times and analysis concepts. Our analysis reveals that the variance depends as much on the applied sampling interval as it depends on the used analysis method. The obtained variation of about 10% is in agreement with other research [1] [2] [3]. In a next step a comparison of buoy and X-Band radar (WaMoS II) data for different sites is presented. The statistical analysis yields that the variance of 10-15% between the two measuring concepts are in the same order of magnitude as the obtained variance from a buoy using different sampling intervals. A link will be drawn between the observed data variations and the expected variability of the s- a state.