Title: Parameter estimation for a global tide and storm-surge model

Author: Martin Verlaan (Delft University of Technology, Deltares)
Julius Sumihar (Deltares)
Andrea Lalic (Université Nice-Sophia Antipolis)
Yvonne Ruckstuhl (Delft University of Technology)

Tides and storm-surges affect coastal safety, shipping, tidal power generation and many other coastal activities. Hydrodynamic tide and storm-surge models can be used for operational forecasting as well as estimating the impacts of sea-level rise and changes to storm patterns caused by climate change. The performance of these models is affected by friction parameters that are poorly known. In addition, the physical process of Self Attraction and Loading, is still under development for the Delft3D-FM model, that we used for this application. The SAL-term affects the propagation speed of the tidal wave, as does the bathymetry. We decided to compensate for the SAL-effect with regional changes to the bathymetry, at least until the SAL-term becomes available for Delft3D-FM. We estimated these parameters with an iterative least-squares estimator (DUD by Ralston and Jennrich 1978) was applied to assimilate satellite observations of sea-level as well as coastal tide-gauges. Interesting aspects of the application are large spatial variability in the sensitivity to the parameters and the fact that friction needs to remain positive. To constrain the values of the parameters, the available DUD algorithm in OpenDA was extended with the option for hard constraints in a manner similar to Quadratic Programming (Janjic et. al. 2014). The strategy of temporarily compensating for the missing SAL-effect in the model resulted in significant improvements of the accuracy of the sea-level, although the adjustments of the bathymetry are most likely not realistic. Still, as a temporary solution in this stage of the development, the results are considered sufficient. The computational requirements of the DUD method increase linearly with the number of estimated parameters, which restricted the number of parameters that could be estimated. Future work will consider localization in this context to allow for more parameters.

GMAO Head: Steven Pawson Global Modeling and Assimilation Office NASA Goddard Space Flight Center

Curator: Nikki Privé Last Updated: Feb 9 2015