Seasonality in a tidal reach: Existence, impact and a possible approach for design flood level estimation.

Heterogeneity caused by seasonality could lead to the estimation error of the design flood level (DFL). This research intended to examine the existence of seasonality in the extreme water levels in a tidal reach and to quantify its impact on the DFL estimation. The mixed-distribution, a commonly used method for design value estimation with heterogeneous samples, was tested. A case study was carried out in the Pearl River Delta, South China. Results showed that a significant seasonality existed in the extreme water levels that were generated from the flood-tide interactions in the delta. If the DFL was estimated with a single distribution, the DFL with a return period smaller than 1.1 years would be underestimated and more information of the non-flood season would be lost. The mixed-distribution was superior in its consideration of seasonality, however, when the return period was over 10 years or smaller than 5 years, the DFL estimation results of this approach were only shifted by <1% from that of a single distribution.