The Estuary SpaTial LandscapE Evolution Model (ESTEEM) simulates the morphological evolution of estuaries over timescales of the order 10 to 100 years (the mesoscale defined within the iCOASST project; Nicholls et al., 2015). ESTEEM adopts a novel 'hybrid complexity' approach (Thornhill et al., 2015; French et al., 2016) that combines the computational efficiency and high spatial resolution of GIS-based spatial models (e.g. SLAMM; Clough et al., 2010), with the more robust physical basis of morphodynamic models recently developed for tidal inlets and lagoons (Di Silvio et al., 2010). It predicts the evolution of estuary morphology, especially within the intertidal zone, and also incorporates detailed representation of engineered structures. Its computational efficiency allows whole estuary simulations simulations at high spatial resolution (of the order 5 to 20 m grid size) at timescales of decades to centuries. Model scenarios can include not only the effects of changing sea-level, wave climate and sediment supply but also human behaviour as articulated through flood and coastal defence policy.
Modeling approach and underlying data model
Segmentation of the DEM
ESTEEM has been developed at UCL as part of the iCOASST project. It is a hybrid in terms of both its modeling approach and underlying data model . A raster Digital Elevation Model (DEM) is the basis of the model discretization and simulation of erosion and deposition within the various estuary sub-environments allows this DEM to evolve over time. The DEM is usually a composite data product constructed from airborne lidar altimetry for higher terrain and the intertidal zone, stitched to bathymetric surveys (e.g. multibeam soundings). Segmentation of the DEM within specified elevation ranges yields landform 'objects' that define the main subtidal channel, tidal flat and saltmarsh units. The subtidal channel system is represented as a vector of channel centerline coordinates and orthogonal sections, the area and free surface width of which are defined as functions of water level. Intertidal storage volumes are also resolved. A separate vector layer specifies the coordinates of engineered structures (flood embankments, bulkheads etc.) and, where applicable, the areas that they protect.
Simulations start with computation of tidal hydrodynamics based on the 1D form of the shallow water equations, taking account of the lateral (intertidal) storage areas. The computed variation in tidal amplitude along the estuary is used to adjust the initial elevation-based segmentation of landform/habitat types. Following the approach advocated by Di Silvio et al. (2010), suspended sediment dynamics are then computed using a 1D advection-diffusion scheme, and a time-averaged concentration along the estuary computed over an appropriate time interval.
Time-averaged concentration along the estuary
Time-averaged suspended sediment concentration along the estuary a boundary condition for separate simulation of intertidal flat and saltmarsh profiles adjacent to each of the channel sections in the 1D model. Bed evolution is simulated using a simple advection-diffusion scheme that includes the bottom stress due to locally-generated waves computed using a parameterised model drive by wind speed, direction and fetch (Young and Verhagen, 1996). Evolution of the intertidal flats and saltmarsh can then be locally interpolated from 1D to 2D to provide a spatial model output that is comparable to models such as SLAMM but which has a mechanistic underpinning in that key erosional processes are represented
Interpolated from 1D to 2D
Initial application of ESTEEM to iCOASST pilot sites and lessons learnt
ESTEEM is intended to be a generic model code, although the initial version is most applicable to estuaries that have a well-defined central channel flanked by intertidal flats and marshes. The sediment dynamics routines are currently developed to handle muddy estuaries and the estuaries of the Suffolk coast of eastern England (the Deben, Alde/Ore and Blyth) have been the test-bed for the iCOASST development effort. Model development has benefited from the adoption of a participatory approach (Voinov and Bousquet, 2010) in which stakeholder views, especially on key output indicators, interoperability with existing hydrodynamic modelling software used by consultants, and scenarios, have been captured and used to refine the final product.
Development of a complex model from scratch is a big job and we are only now getting to see some of the initial results. Early simulations for the Deben estuary over a 200 year time interval have shown how the infilling of the estuary following realignment of defences is quite strongly dependent on the assumed wind climate and further work is needed to explore the threshold between different end states. We have also shown how ESTEEM can be used to achieve rapid computation of morphological change that can be used in conjunction with a full 2D (or even 3D) hydrodynamic scheme, such as Telemac (which we have implemented for each of estuaries).
Early simulations for the Deben estuary