Sea level variability and coastal trapped waves around southern Africa

Includes bibliographical references.The propagation characteristics of the coastal trapped waves (CTWs) around the coast of southern Africa were investigated by analyzing the observed daily mean sea level data from 16 coastal tide gauges, as well as outputs of sea level anomalies from the Hybrid Coordinate Ocean Model (HYCOM) at a grid point closest to each tide gauge station under consideration. The observed records showed sea level variability dominated by the short time variability with a period shorter than one month. This short time variability varies from season to season with the largest CTW amplitude during austral winter. The short time variability propagates anticlockwise as coastal trapped wave around the coast of southern Africa with a propagation speed ranging from 3 to 6.5 m/s, and from 1 to 7.5 m/s, along the west and south coasts, respectively. These propagation speeds are forced by synoptic atmospheric disturbances mainly in term of wind variability. Coastal trapped waves were observed propagating equatorward in the east coast of southern Africa in the opposite direction of Agulhas current on a few occasions. It can be a result of a good resonance between a strong and persistence of weather system and the coastal trapped wave. It is believed that more precise response and good answers for some discrepancies that were found can be achieved when a longer time records from Inhambane is included in future similar study. The outputs from HYCOM showed very similar propagation characteristics to the observed data. Along the south coast, the behaviour of the CTW is well reproduced. Unfortunately the model does not reproduce very well the variability along the west coast. While it seems to underestimate the west coast response, at same time it seems to overestimate it along the south coast of southern Africa. Although the model demonstrated some CTWs travelling northwards along the east coast, such disturbances were infrequent and difficult to find in the observed data