Application of the Research to Landscape Practice
Much of the output to date from my research into naturalistic urban vegetation has been published in peer review journals which are rarely available to, let alone read by landscape professionals. In the current environment of the RAE (Research Assessment Exercise) these journals are inevitably UK university academics first priority for publication. Whilst I have, in conjunction with my colleague Dr Nigel Dunnett ran a number of conferences to transfer knowledge of our work to practitioners (New Herbaceous Vegetation for Urban Public Spaces in Britain, 2000; New Herbaceous Vegetation for Urban Public Spaces in Britain, II, 2003; Nature Enhanced, June 2004) the work is probably still seen as fringe by most landscape practitioners, such is the dominance of planting. In order to both challenge these assumptions and to “road test” the research, I have undertaken a number of consultancies to create examples of naturalistic vegetation in practice.
2004- Consultant to Tom Stuart Smith re: Development of extensive North American Prairie at Heveningham Hall.
2003- Consultant to Tony Jellard and Associates re: developing naturalistic herbaceous plant communities at Markham village park in South Wales.
2003-2005 Consultant to Tom Stuart Smith re: Development of extensive dry North American Prairie vegetation at Woodperry, Oxford
1998-2002 Consultant to Roger Griffiths and Associates re: development of prairie-like vegetation at Ove-Arup Corporate Headquarters, Solihull
1999, Co-designer Saracens Cross Park for Glasgow 99 City of Art and Architecture Festival
1998-2003 Consultant to Eden Project, St. Austell, re: design and implementation of Steppe-Prairie exhibits
1997-1999 Consultant to Earth Centre, Doncaster, re: ecological herbaceous vegetation
1997-1999 Consultant to Welsh National Botanic Gardens, re: ecological herbaceous vegetation
As the research has progressed, and my understanding and experience likewise, the outcomes have become both more sophisticated and successful. The most successful projects thus far have been those undertaken in conjunction with organisations with a strong desire to see the projects succeed, and able to deliver the essential establishment and management inputs. To date the highest profile success has been the development of a large area (1500 square metres) of sown prairie at the Eden Project in Cornwall. This was sown in March 2000, and whilst relatively simply in structure and species composition, it has created a substantially self-sustaining and from July to October, very attractive plant community. The Eden project was also successful because the waste china clay substrate was largely without a weed seed bank, although one was quickly introduced by adjacent plantings of weedy container grown stock. Early projects involving prairie vegetation sown on topsoil without effective initial control of weeds failed in the same way that conventional planting of the same species would.
The site prior to sowing, after the planting of widely spaced container grown prairie plants to provide some initial impact for the opening of the project. The carbonised tree trunks provide a dramatic rhythmical element to visually structure the sowing.
Sowing the prairie over the top of the planted vegetation at the Eden Project, Cornwall in March 2000.
The same site in September 2003. The prairie is cut down and weeded in spring through a combination of hand removal and burning over to defoliate-kill heat senstive weeds (such as clover, Trifolium repens, and Epilobium spp.)
This experience combined with ongoing research has led to the development of an establishment protocol to ensure effective suppression of soil weed seed banks. My first large scale application of this was in the Prairie garden at Sheffield Botanic Garden (sown in February 2004). This involved elimination of perennial and other vegetative weeds, followed by blanketing of the site with a 50-60 mm deep layer of sand. Seed is sown onto this sand layer in winter and raked in, then covered with jute erosion matting, and rolled to provide good seed-sand contact. This approach has proved to be extremely successful.
It is most important however to recognise that high levels of emergence of prairie species sown in sand mulches requires the sand to be kept moist during the late March early April seedling emergence period. If the sand is allowed to dry out during this time, most species will have greatly reduced emergence or even suffer complete failure. In the absence of adequate rainfall during this period weekly or even twice weekly irrigation is required. My research shows that most herbaceous plants, including common native species respond in a similar manner. These "blanketing mulches" are fundamental to applying vegetation establishment by seed sowing technology to practice and the latest research in this area is discussed under "Current Research; Blanketing substrates for managing weed seed banks". (LINK)
As my experience of undertaking this work in practice has accumulated, it has become possible to develop more sophisticated schemes. In the Sheffield Botanic Garden project, three different plant communities were used to deal with the various sun and shade regimes across the site. In a current project involving over 5,000 square metres of sown vegetation at Heveningham Hall, five plant communities have been used to deal with conditions ranging from dry to moist, full sun to dense shade (woodland core; woodland edge; moist prairie; dry prairie; and prairie grass dominated edge). The scheme has been designed so that adjacent mixes always contain some common species, thus avoiding sharp transitions between mixes; there is change but it is not visually dominating.
Up to 30 species are used per mix, although normally the number of species per area is 15-20. By using this number of species the sustainability of the scheme in the longer term is improved, as as species that prove to be poorly fitted decline, their territory is seamlessly occupied by better fitted neighbours. Individuals or even species cohorts may fail but the community continues on. The other innovation in these plant communities has been to develop spreadsheet formulae that allow the seed mix designer to identify target densities of each species per square metre, which are then automatically converted into grams of seed for each species that need to be added to the seed mix. The key and often unavailable information required for these formulae is typical field emergence for that species. Some species convert a very high percentage of their sown seed into seedlings, for example 70%, in other species it may be only 1%. Most species lie between 10 and 30%. I maintain an active screening programme to establish these values for a wide range of species sown under standardised sand mulching protocols.
These ideas are discussed in greater detail on page 154 of Dunnett and Hitchmough (2004). Where typical percentage emergence is known, this provides a high level of control over the outcome of the sowing, and the concrete link between what the designer envisages and what happens in reality on the site. This allows you to design communities as a series of layers and exercise control over composition in each layer.
Dunnett, N. and Hitchmough, J.D. (2004) The Dynamic Landscape, Design, Ecology and Management of Naturalistic Urban Planting. Taylor and Francis, London.
Seed sowing mixes and distribution plan for the prairie garden at Sheffield Botanic Garden. Mix A contains lower more drought tolerant species to compensate for the presence of a tree (not shown on plan) close to the north side of the planting; Mix B is the core prairie mix, and Mix C is a woodland edge community of species that will persist under shady conditions.
Penstemon digitalis and Castilleja coccinea, two early summer flowering species in the prairie in June 2005.
The Botanic Garden Prairie in September 2005.
Application of the sand mulch technique to a very large scale (3000+ square metres), the "white area" is where a prairie seed mix has been sown using a sawdust carrier. The photograph was taken immediately prior to "raking in".
Use of different sowing mixes to cope with a light gradient from full sun to dense shade under trees. A woodland core mix of highly shade tolerant species has been sown under the trees. This is separated from the prairie mix in areas receiving full sun by a transition mix of woodland edge species.
After sowing the area is covered in open weave jute erosion matting to stabilise the surface and provide a better germination and establishment environment.
Where initial weed control was effective and few rhizomes of weedy species persist, the sand technique results in very low maintenance in the first year. In year 2 and subsequent years, the management requirements differ according to the nature of the community and species used. In the case of prairie vegetation, the vegetation is inspected in winter and any evergreen weeds either removed or killed in situ with a glyphosate herbicide. The dead stems of the prairie plants are left in situ till late winter or early spring to provide a food source/habitat for wildlife. The stems are cut down with a brushcutter and either removed from the site or burnt in situ using a propane burner between late March-early April. Burning is not essential every year unless the prairie is very weedy, but it does greatly improve prairie plant regeneration from self sown seed as well as reduce snail population build up and competition from colonising species such as cleavers (Galium aparine) and nursery Epilobium.
A detailed analysis of the effects of various management techniques on sown prairie development can be found in Hitchmough and De la Fleur (in press).
Hitchmough, J.D., and De La Fleur, M., (in press) Establishing North American Prairie vegetation in urban parks in northern England: Effect of management practice and initial soil type on long term community development. Landscape and Urban Planning.