skip to primary navigation skip to content

The relationships between vegetation characteristics and the sea defence value of saltmarsh ecosystems

The relationships between vegetation characteristics and the sea defence value of saltmarsh ecosystems


Global and accompanying regional climatic changes are likely to increase flood and erosion risk on the low-lying coasts of many shores in the near future. To manage such risks, a better understanding of the natural sea-defence capacity of intertidal areas is required. While, at the time of the this study, preceding studies (e.g. Möller and Spencer (2003), Möller et al. (2001)) had indicated that such environments can reduce flood risk by significantly attenuating waves over hydraulically rough vegetated surfaces, no study had, as yet, systematically investigated the relative importance of different types of intertidal vegetation and/or seasonal changes to this bed roughness effect. Funding was obtained via an EPSRC/RGS Geographical Research Grant to address this research gap.

Aims/objectives and methodology

The specific research objectives of this project were to:

  1. test and validate an innovative digital photograph technique for the measurement and parameterisation of saltmarsh vegetation 'roughness'/'density', and
  2. provide information on the relationships between wave attenuation and vegetation 'roughness'/'density' on a macro-tidal saltmarsh.

Field measurements of vegetation and wave attenuation over three and nine consecutive spring tides, respectively, were carried out on a mature open coast marsh near Tillingham, Dengie peninsula, Essex, UK, in September 2004 and December 2004. Measurements were carried out at three ca. 10 m long transects, characterised by different vegetation types: Transect 1 contained a dense Spartina cover, Transect 2 a less dense combination of Spartina, Salicornia, and Suaeda, and Transect 3 a uniform Salicornia cover. All transects were located the same distance from the marsh edge and in close vicinity to each other to minimise potential spatial differences in incident wave conditions.

Vegetation measurements: At 18 locations, digital side-on photographs of a 0.6 cm wide and 0.1 m deep strip of vegetation were obtained. In September 2004 only, all vegetation contained in the photographs at the locations outside the transects was then also harvested for determination of wet and dry biomass and species composition / structure.

Wave recording: Surface-mounted pressure transmitters (PTs) were triggered simultaneously at all transects and recorded water levels at 4Hz frequency throughout the tidal inundation period. Data was stored on Campbell 'CR10' dataloggers and processed to provide wave spectra and wave statistics at 30cm, 40cm, 50cm, and, 60cm water depths on the rising tide.


Results from this study are reported in Möller (2006). A significant positive relationship between the vegetation pixel density (black pixel density of the classified binary image) and dry biomass existed in all three vegetation types and in a test case, where tall Spartina vegetation was removed from a fixed photo location in three successive stages (Spartina clip experiment). The different heights of the vegetation canopy partly explained the relative differences in the nature of this relationship.

The highest tides (up to 70 cm water depths above the marsh surface) were recorded in September. The highest significant wave heights of just over 30 cm were recorded in December in a water depth of 50 cm. Wave attenuation (Hs reduction) varied from 0.08% (in 40cm water depth over transect 3 in December) to 33% (in 20cm water depth over transect 1 on the same tide). At any water depth, in September and December, highest average wave attenuation was observed over Transect 1, although attenuation was highly variable. A statistically significant relationship between the relative wave height (significant incident wave height (Hs) / water depth (h)) and wave height attenuation (% Hs reduction) existed on Transects 1 and 2 (but not 3) in September, when Hs/h measurements were consistently < 0.55. In December, however, when Hs/h measurements reached as high as 0.86, no significant relationship was found to exist between Hs/h and Hs attenuation. When the dataset was restricted to records for which Hs/h < 0.55, however, a positive relationship between Hs/h and Hs attenuation did exist in December, although only for Transects 1 and 2 (as in September).


The results from the vegetation and wave measurements through this study at Tillingham contribute significantly to the debate on the causes and consequences of varying wave attenuation over intertidal surfaces. The study was the first such investigation to quantify the degree to which the natural coastal flood and erosion protection potential of such surfaces depends on both the hydrodynamic setting (critically, the relative water depth (Hs/h)) and the surface roughness induced by seasonally and spatially varying vegetation type and density. It did so by deploying an innovative method for measuring vegetation structure and density that has been successfully calibrated against traditional methods for measuring standing biomass. Of the two factors, relative water depth and vegetation type/density, the former appears to have been the dominant control on wave attenuation at Tillingham during the two field campaigns conducted as part of this study.

Publications arising from this project

  • Möller I 2006 Quantifying saltmarsh vegetation and its effect on wave height dissipation: results from a UK East coast saltmarsh. Journal of Estuarine Coastal and Shelf Sciences 69, 337-351.

Conference presentations

  • Möller I 2004 'Quantifying form-process interactions on saltmarshes - a case study from the UK East Coast', Invited Speaker, AGU Chapman Conference on Saltmarsh Geomorphology, Halifax, Canada, October 2004.

Related CCRU publications

  • Wolters M, Bakker JP, Bertness MD, Jefferies L and Möller I 2005 Saltmarsh erosion and restoration in south-east England: squeezing the evidence requires realignment. Journal of Applied Ecology 42, 844-851.
  • Smith GM, Thomson AG, Möller I and Kromkamp JC 2004 Using hyperspectral imaging for the assessment of mudflat surface stability. Journal of Coastal Research 20, 1165-1175.
  • Smith GM, Thomson AG, Möller I and Kromkamp JC 2004 Hyperspectral imaging for mapping sediment characteristics. Estuarine and Coastal Sciences Association Bulletin 45, 14.
  • Möller, I. and Spencer, T. 2003 Wave transformations over mudflat and saltmarsh surfaces on the UK East coast – Implications for marsh evolution. Proc. Int. Conf. on Coastal Sediments '03, Florida, USA.
  • Möller I, Spencer T, French JR, Leggett DJ, Dixon M 2001 The sea-defence value of salt marshes – a review in the light of field evidence from North Norfolk. Journal of the Chartered Institution of Water and Environmental Management 15, 109-116.
  • Möller I, Spencer T, French JR, Leggett DJ, Dixon M 2000. A new perspective on the sea-defence value of salt marshes. Proc. 35th MAFF Conf. of River & Coastal Eng. 5-7 July, Keele, UK, 11.11.1-11.11.4
  • Möller I 2000 Vegetation survey to support wave attenuation monitoring in The Wash. Report to the Environment Agency, UK.
  • Möller I, Spencer T, French JR, Leggett DJ, Dixon M 1999. Wave transformation over salt marshes: A field and numerical modelling study from North Norfolk, England. Estuarine, Coastal and Shelf Science 49, 411-426.
  • Moeller I, Spencer T, and French JR 1996 Wind wave attenuation over saltmarsh surfaces: Preliminary results from Norfolk, England. Journal of Coastal Research 12(4), 1009-1016.

Related CCRU projects


Figure 1: Location map and positioning of wave recording transects, Tillingham Saltmarsh, Dengie Peninsula, UK

Figure 1

Figure 2: View to the Northeast across Tillingham marshes from the seawall and (smaller image) wave recording Transect 1 (photos: I. Möller, September 2004)

Figure 2(a) Figure 2(b)

Figure 3: Digitally classified vegetation image and resulting parameters

Figure 3