Changing coastlines: data assimilation for morphodynamic prediction and predictability

Lead Research Organisation: University of Reading
Department Name: Mathematics and Statistics

Abstract

In 2005, severe flooding in the aftermath of Hurricane Katrina focussed the world's attention on the importance of accurate knowledge of the topography of the coastal zone in natural disaster management and prediction. The topography of the sea floor, generally known as the bathymetry, evolves over time as sediment is eroded, transported and deposited by water action. The change in bathymetry itself changes the motion of the water, which is also influenced by tides and weather patterns, such as storm surges. An accurate, up-to-date knowledge of coastal bathymetry would allow improved flood forecasting. Improved prediction of future bathymetry, and knowledge of the uncertainty in that prediction, would allow construction of better sea defences, better management of coastal habitats, and better understanding of the effects of changes in land use near the coast. It may also provide better understanding of the effects of climate change (e.g. sea level rise, and increased numbers of extreme storm events) on the longer-term evolution of an estuary. Coastal sediment transport models are becoming increasingly sophisticated. However, observed bathymetric samples typically only provide partial coverage of the domain of such a model. Hence, initialisation of such models using only a set of recent observations is not feasible. The effective and efficient use of limited data, such as these, requires state-of-the-art mathematical, statistical and computational methods, known as data assimilation techniques. Data assimilation combines empirical observations with model predictions to give more accurate and well-calibrated forecasts and enables the uncertainties in the forecasts to be calculated. Whilst data assimilation has been in use in the context of atmospheric and oceanic prediction for some years, its use in the context of coastal sediment modelling is novel. This project will use data assimilation techniques with a coastal sediment transport model to maintain up-to-date near-shore bathymetry, predict future bathymetry, answer statistical questions regarding uncertainty and predictability, gain insight into physical processes taking place during intense storm events and to design an optimal observation strategy for coastal monitoring. Three coastal sites have been identified for numerical experiments. Methodologies will be developed and tested using data from the first site and validated using independent data from the other sites, demonstrating the wider applicability of ideas. The novel use of data assimilation will allow improved estimates of the current bathymetry, and improved predictions of future bathymetry via better initialisation, error estimates for the improved bathymetry, and a means to estimate model parameters from indirect observations. The direct involvement of the Environment Agency in the project will ensure that the resulting benefits are transferred into operational practice.

Publications

10 25 50
publication icon
Bannister R (2008) Modelling of forecast errors in geophysical fluid flows in International Journal for Numerical Methods in Fluids

publication icon
Dance S (2007) Unbiased ensemble square root filters in PAMM

publication icon
Gratton S (2007) Approximate Gauss-Newton Methods for Nonlinear Least Squares Problems in SIAM Journal on Optimization

publication icon
Hunter N (2008) Benchmarking 2D hydraulic models for urban flooding in Proceedings of the Institution of Civil Engineers - Water Management

publication icon
Lawless A (2008) Approximate Gauss-Newton methods for optimal state estimation using reduced-order models in International Journal for Numerical Methods in Fluids

publication icon
Livings D (2008) Unbiased ensemble square root filters in Physica D: Nonlinear Phenomena

 
Description In 2005, severe flooding in the aftermath of Hurricane Katrina focussed the world's attention on the importance of accurate knowledge of the topography of the coastal zone in natural disaster management and prediction. The topography of the sea floor, generally known as the bathymetry, evolves over time as sediment is eroded, transported and deposited by water action. The change in bathymetry itself changes the motion of the water, which is also influenced by tides and weather patterns, such as storm surges. Coastal sediment transport models are becoming increasingly sophisticated. However, observed bathymetric samples typically only provide partial coverage of the domain of such a model. Hence, initialisation of such models using only a set of recent observations is not feasible. The effective and efficient use of limited data, such as these, requires state-of-the-art mathematical, statistical and computational methods, known as data assimilation techniques; the application of these techniques to coastal morphodynamic modelling is novel and presents an exciting opportunity for improvements.
In this project we combined a simple decoupled hydrodynamic and sediment transport model with a data assimilation scheme to investigate the ability of such methods to improve the accuracy of the predicted bathymetry for a case study of Morecambe Bay. UK. The observation data used for assimilation purposes comprised waterlines derived from satellite Synthetic Aperture Radar (SAR) imagery and swath bathymetry data collected by a ship-borne survey. A LiDAR survey of the entire bay was used for validation purposes. The comparison of the predictive ability of the model alone with the model-forecast-assimilation system demonstrated that using data assimilation significantly improves the forecast skill. An investigation of the assimilation of the swath bathymetry as well as the waterlines demonstrated that the overall improvement is initially large, but decreases over time as the bathymetry evolves away from that observed by the survey. The result of combining the calibration runs into a pseudo-ensemble provided a higher skill score than for a single optimized model run.
In these experiments the model parameters were set by manual calibration; this is laborious and is found to produce different parameter values depending on the type and coverage of the validation dataset. The second part of the project considered the problem of model parameter estimation in more detail. By employing the technique of state augmentation, it was possible to use data assimilation to estimate uncertain model parameters concurrently with the model state. This approach removes inefficiencies associated with manual calibration and enables more effective use of observational data. We developed a novel hybrid sequential 3D-Var data assimilation algorithm for joint state-parameter estimation and demonstrated its efficacy using idealized models The results of this study are extremely positive and suggest that there is great potential for the use of data assimilation-based state-parameter estimation in coastal morphodynamic modelling.
Exploitation Route Since the end of the project, the application of data assimilation to coastal morphodynamic modelling has already been taken up by researchers at the engineering consultancy/research institute Deltares.

Our new technique for concurrent state and parameter estimation has so far only been investigated in idealized systems and requires further development and testing in more realistic systems. This work is beginning in the new application area of river flood modelling.
Sectors Aerospace, Defence and Marine,Environment,Leisure Activities, including Sports, Recreation and Tourism,Government, Democracy and Justice

 
Description Since the end of the project, the application of data assimilation to coastal morphodynamic modelling has already been taken up by researchers at the engineering consultancy/research institute Deltares, with the development of their OpenDA package. Our new technique for concurrent state and parameter estimation has so far only been investigated in idealized systems and requires further development and testing in more realistic systems. This work is beginning in the new application area of river flood modelling.
First Year Of Impact 2011
Sector Aerospace, Defence and Marine,Environment
 
Description EPSRC Senior Fellowship in Digital Technology for Living with Environmental Change
Amount £1,706,722 (GBP)
Funding ID EP/P002331/1 
Organisation Engineering and Physical Sciences Research Council (EPSRC) 
Sector Public
Country United Kingdom
Start 09/2016 
End 08/2021
 
Description Satellite Applications Catapult (CORSAIR) 
Organisation Satellite Applications Catapult
Country United Kingdom 
Sector Charity/Non Profit 
PI Contribution We have been carrying out research on automatic flood delineation using synthetic aperture radar data provided by the Satellite Applications Catapult. Report to the catapult (March 2018)
Collaborator Contribution Synthetic Aperture Radar (SAR) images were provided under the CORSAIR programme.
Impact Mason, D. C., Dance, S. L., Vetra-Carvalho, S. and Cloke, H. L. (2018) Robust algorithm for detecting floodwater in urban areas using Synthetic Aperture Radar images. Journal of Applied Remote Sensing. doi: 10.1117/1.JRS.12.045011
Start Year 2016
 
Description Ongoing "Magnificent Maths" Days 
Form Of Engagement Activity Participation in an open day or visit at my research institution
Part Of Official Scheme? Yes
Geographic Reach Regional
Primary Audience Schools
Results and Impact Building on the event funded by this grant, the University of Reading continues to hold an annual "Magnificent Maths" day or 2-day summer school for Year 12 school students, featuring workshops, talks and career panels.

Teachers comment that a Maths related school trip is a rare opportunity. Many of the same schools bring their students year after year, showing the value that they place on this activity for enthusing their students about choosing to continue Maths to A-level as well as further on at University.
Year(s) Of Engagement Activity 2010,2011,2012,2013,2014,2015
URL http://www.prospectus.rdg.ac.uk/archive/teachersandadvisors/advisors/taSTEMEvents.aspx