A third of England and Wales’s threatened sand dunes have a brighter future thanks to a pioneering National Lottery funded conservation and restoration project, unveiled today.
Sand dunes are listed as the habitat most at risk in Europe. They are a sanctuary for endangered plants and animals like the dune gentian and sand lizard. However, many are being smothered and fixed by a tide of invasive scrub, becoming sterile, grassy hillocks.
Now The Heritage Lottery Fund has given initial support* for a £4.1m grant to an innovative four year partnership project, Dynamic Dunescapes, which will restore some of the most important sand dune landscapes in England and Wales.
Natural Resources Wales, Natural England, National Trust, Plantlife and the Wildlife Trusts and will work with local coastal communities to conserve and restore up to 7000ha in nine areas across England and Wales.
Sand dunes are a naturally dynamic habitat. As the coastal winds blow, new ‘embryo’ dunes are formed at the top of the beach and they slowly grow and shift to create the landscapes we know. The early stages of the project will explore how we can re-establish the natural movement within dunes, to create the conditions that some of our rarest wildlife relies upon.
Kevin Ingram, Natural Resources Wales interim Chief Executive said:
"Sand dunes are some of the richest places for wildlife in Wales but many of the rarest species have declined and in some cases disappeared as the dunes have become more stable.
“This project will bring many of these species back, benefiting not only our environment but also our people and our economy as it builds on the strong links between coastal communities and their dunes"
The project will concentrate on improving the condition of nine identified dune cluster sites at: Lincolnshire Sand Hills; Studland Dunes, Dorset; North Cornwall Coast; Braunton Burrows; North Devon Coast; Swansea/Neath Port Talbot; Carmarthen; Anglesey/Gwynedd; Sefton Coast; and Cumbrian Coast/Solway.
Natural England’s Chairman, Andrew Sells, said:
“We’re really excited about this fantastic project to save our dunes and give more people the opportunity to learn about our fascinating wildlife. Dunes are not only a backdrop to a day at the seaside – they are home to some of our rarest species and are in desperate need of help. That’s why we’re working with our partners from across the conservation movement – spanning two countries – to tackle this problem together.”
The projects will include:
- conservation work to re-establish natural processes;
- a programme of removal of native and non-native invasive species;
- restoration and creation of dune slacks and dune wetlands;
- turf stripping and sand scraping to create bare sand patches;
- on site interpretation and a national promotion programme;
- community education activity, including a schools programme;
- a programme of adult and youth volunteering
Drew Bennellick, head of landscapes and natural heritage at the Heritage Lottery Fund said:
“This is a really exciting project as it is pioneering a new approach to dune management. In recent decades many sand dunes have become smothered by thick vegetation. We now know that this is bad news for some of the rare species that make their homes among our dunes and need a mixture of open sand, pools of water and varied vegetation to thrive. Thanks to this National Lottery funding and the expertise of the partnership organisations, we can begin to find ways of addressing these pressing issues.”
Hugely popular among beachgoers, there are 200m visits a year to sand dunes in England and Wales. Less well known is their role as a sanctuary for endangered plants and animals. Species that make their home in the dune landscape include:
- colourful flowers like purple milk-vetch and dune gentian;
- invertebrates such as the spectacular silver-studded blue butterfly and snail-killing flies;
- mosses and liverworts (with the diminutive petal-wort found only in dune slacks); and
- reptiles and amphibians such as natterjack toads, sand lizards and great crested newts
Dunes have a long place in the cultural history of the United Kingdom. These special places include prehistoric sites around the coast and medieval religious settlements. Sand dunes are popular destinations for seaside recreation, from sand-castle building to surfing. The England Coast Path will open in its entirety by 2020. This will allow more people than ever to access the sand dunes along our shores. This project will ensure those visitors can enjoy our sand dunes, now and in the future.
In the previous sections, we have described two independent approaches for modelling dune groundwater systems, (1) a three-dimensional, numerical groundwater flow model (ZOOMQ3D) with spatially distributed parameters for dune system properties and (2) a one-dimensional single-point / lumped parameter model, to simulate groundwater flow in small sand dunes systems at Braunton Burrows and Ainsdale, respectively.
Developing the groundwater flow of the Braunton Burrows dune system has shown that a higher degree of model complexity/heterogeneity has been required than was initially anticipated based on the conceptual model. Whilst the system itself is of relatively modest size (8 km2) and has been subject to a variety of surveys and studies (e.g. Mcfarlene 1955; Kidson & Carr 1960; Kidson et al. 1989; Burden 1998; May 2001; Stratford et al. 2013; Allen et al. 2014), quite a high degree of uncertainty was discovered during model development relating to the definition of the model boundary conditions (e.g., drainage conditions in the northeast of the model domain) and the transmissivity distribution within the model domain.
Simplifying assumptions had to be made during the early stages of model development, but these were gradually adjusted using sensitivity analysis. The importance of such an analysis in defining the correct model boundary conditions has been demonstrated by Hunt et al., (1998).
Targeted field investigations were an important part of the model development process, specifically the geophysical survey, which provided an improved characterization of the bedrock topography and led to a more detailed delineation of transmissivity zones within the model domain. Constructing models with increasingly more complex spatial distributions of transmissivity within the model domain, led to considerable improvements in the fit between predicted and observed heads.
Calibration of transmissivity (T) (= one parameter) rather than hydraulic conductivity (K) and depths to aquifer base (z) (= two parameters) was chosen here to reduce the degrees of freedom or “free parameters” in model calibration, as proposed by Refsgaard (1997), and to reduce the problem of equifinality or non-uniqueness in numerical modelling (Beven 2006), which says that there are many different conceptualizations of a numerical groundwater flow model that fit the observed data equally well. This problem has been demonstrated by the Monte Carlo analysis (Fig. 7), where many different combinations of T resulted in the same or similar NSE values (despite the reduction in free parameters). This non-uniqueness is also the reason for why numerical groundwater models cannot be truly validated (Konikow and Bredehoeft, 1992) and why their use must be constraint by their inherent uncertainty (Wondzell et al., 2009).
In the Braunton Burrows model, model uncertainty needs to be reduced and the problem of equifinality must be further investigated. While the topography of the aquifer base (i.e. z) is relatively well-constrained by the data from the geophysical survey, there remains large uncertainty relating to the distribution of hydraulic conductivity (K) within the dune system. The inability of the model to reproduce the observed water level response in some parts of the model may indicate that the distribution of hydraulic conductivities across the model domain is not as uniform as Burden (1998) and Allen et al. (2014) suggest. Recent drilling in the central part of the dune system has revealed the presence of coarse sands and gravels at depths within the dunes (D. White, 2015, personal communication), hence challenging the general assumption that sand-blown dune systems consist of uniform, well-sorted granular sediments throughout. Furthermore, the lack of water level data in the north of the dune system must be addressed and an improved understanding of how rainfall events in 2004 influence recharge is required.
Groundwater modelling is often considered to be a linear process, consisting of a series of successive steps as, for example, outlined in Hobma et al. (1995). The present study illustrates that it should be a more iterative process and that modelling should be carried out in conjunction with field studies, i.e. not as a postscript, but as ongoing interaction. This interaction is required throughout the investigation process. The importance of feedback between numerical modelling, conceptual modelling and data collection is increasingly recognised in the modelling of dune systems (e.g. Spring (2005)) as well other groundwater systems (Voss 2011).
The simple, conceptual modelling undertaken for the Ainsdale Dune system demonstrated that such a model could be calibrated with reasonable generic hydraulic parameters (K = 11 m d−1; ne = 0.28). There are a number of observation boreholes on the Ainsdale site and the model transferability was tested by applying it to two more boreholes. However, adjustments were required as at one location the dune slack generation was more pronounced, and the other was covered by pine trees. For the latter, the recharge module of the model needed adjustment to take into account interception. The changes required by the model illustrate the variability within a small site. The subtle variation of land use and surface response required changes to the model structure thus reinforcing the need for interaction and subsequent flexibility within the modelling methodology.
In first runs of models, a poor fit is usually found between the water table levels observed in the field and the model-generated data. This often forces the modeller to return to the conceptual model or to question the data used to parameterise or run the model (e.g. the representativeness of the available dune property data or the scale at which rainfall or evapotranspiration data are employed). This interactivity is key to heuristic learning, i.e. learning by someone doing something themselves, and helps guide the user through developing their understanding of the system and identifying important controls. Voss (2011) points outs that rather than use automatic calibration the interaction between the user and the model should be “learning something from a model-based analysis”.