The effect of seed mix composition and sowing rate on the establishment of an MG13 grassland on a grazing marsh with high phosphorous availability Barn Elms, Greater London, England

  • Published source details Gilbert J.C., Gowing D.J.G. & Bullock R.J. (2003) Influence of seed mixture and hydrological regime on the establishment of a diverse grassland sward at a site with high phosphorus availability. Restoration Ecology, 11, 424-435


High phosphorus availability in soil is associated with low plant species diversity in grassland swards. This study investigated whether creation of a diverse sward is possible on a soil with very high phosphorous concentrations at a grazing marsh in southeast England. Three experiments (see also Cases 563 and 565) were undertaken to investigate the influence of seed mixture on the establishment of a target grassland community. The experiment summarised here investigated if altering the composition and sowing rate of a commercial seed mix were significant in the establishment of a sward similar to British National Vegetation Classification (NVC) Agrostis stolonifera (creeping bent)–Alopecurus geniculatus (marsh foxtail) grassland (NVC code MG13).

Study site: The effect of sowing rate and seed mixture composition on the establishment of an Agrostis stolonifera (creeping bent)–Alopecurus geniculatus (marsh foxtail) grassland (NVC code MG13, Rodwell 1992) was undertaken on the grazing marsh at the Wildfowl and Wetland Trust's (WWT) Wetland Centre, Barn Elms, London.

Seeding treatments: The experiment aimed to assess the effect of varying the proportions of species in the grass seed mix and sowing rate on the development of the MG13 target community. Three different mixes of the same eight grass species typical of MG13 were used (see Table 1, attached). A. stolonifera was varied in its proportion of the total mix (10, 40 and 70%) with the other seven species (kept in the same relative proportions) making up the rest. A.stolonifera was chosen as it was the most abundant species in the mix aimed at producing MG13 grassland. Seed was sown at rates of 10, 25, and 40 kg/ha (nine different combinations of mixture and sowing rate) in unreplicated plots (approximately 15 × 15 m) with 10, 1 m² permanent quadrats positioned randomly in each.

Monitoring: The presence of all plant species and a visual estimate of percentage cover in each 1 m² quadrat were recorded each June over 3 years. The location and height of the quadrats were recorded to enable an estimation of the water regime. The analytical methods used (a correlation matrix and principle components analysis) allowed identification of the factors significantly influencing the vegetation community developing under each sowing rate/seed mixture combination.

Calculation of similarity coefficients to vegetation communities: The percentage cover estimates for all the species recorded in the quadrats were entered into a vegetation analysis program which compares the vegetation to that of NVC floristic tables.

Hydrological regime: A network of ditches intersects the grazing marsh and influences the water table. A hydrological model was produced to predict the water table depth. This required weekly records of rainfall, evapotranspiration and ditch water level, plus a range of soil parameters (hydraulic conductivity, depth of soil layers, drainable porosity) and the elevation and distance from the ditch of each location. Ditch water level and rainfall were measured by WWT staff, and 18 dip wells were installed to enable the water table to be measured in representative locations. Evapotranspiration data were provided by the nearby weather station at the Royal Botanic Gardens, Kew. These data were used to validate the water-table predictions produced by the model.

MG13 similarity coefficient: In the first 2 years, there was a relationship between the proportion of A.stolonifera and the similarity of the vegetation to that of the target MG13 community, with the mixes containing the greatest proportion of A.stolonifera exhibiting the greatest similarity. However, this was not significant by the third year.

Sowing rate did not significantly affect the similarity of the vegetation to that of the MG13 community. By year 2 it became apparent that hydrological regime was the dominant factor affecting the vegetation similarity. Vegetation experiencing a high aeration stress (i.e. wetter areas) achieved the greatest similarity, whilst that experiencing a high degree of drought stress had the lowest MG13 similarity coefficients.

In the first year water regime and the proportion of A.stolonifera in the seed mix, appeared to be the main factors affecting the distribution of the species. However, other unexplained factors also influenced the results. Perhaps unsurprisingly, A.stolonifera cover was positively related to the proportion sown, whereas the other species were negatively related to the proportion of A.stolonifera sown. The only exception was cock's-foot Dactylis glomerata, which appeared not to be strongly affected. Tufted hair-grass Deschampsia cespitosa and A.stolonifera appeared to be positively related to aeration stress, whereas Yorksire fog Holcus lanatus was positively related to drought stress.

By the third year water regime had become the most important factor influencing the majority of species. A.stolonifera was more prevalent in wetter areas, whilst H.lanatus, D.glomerata and rough meadow-grass Poa trivialis in drier ones. Sowing rate did not appear to affect the distribution of the species. The highest D.cespitosa cover occurred in the plots with the smallest proportion of A.stolonifera sown, suggesting perhaps some competitive interaction.

Hydrological influence: The hydrological regime varied widely across the grazing marsh due to variable topography and restriction of water movement in some areas caused by soil compaction. Measures of aeration ranged over a continuous scale showing that all conditions from no flooding to prolonged flooding occurred. Likewise, measures of drought ranged widely from no drought to very droughty, but few quadrats experienced anything more extreme than moderate drought.

Conclusions: The composition of the grass seed mixes sown on the grazing marsh at Barn Elms, affected the establishment species within the sward in the first year after sowing. After 3 years the composition of the seed mixture ceased to be an important factor with plant species cover and distribution most strongly affected by the water regime. The vegetation was less diverse than predicted from germination tests (see Case 563) and decreased in diversity over the monitoring period. This was perhaps due to the water regime being unsuitable for most species or the extremely high P concentration in the soil allowing the most competitive species to dominate.

The Barn Elms studies highlight the need to be aware of soil and hydrological conditions of a site before choosing a target community and designing a seed mixture. It appears that in attempts to achieve the desired MG13 grassland community, the proportion composition of the seed mix was less critical to vegetation establishment than the water regime and soil nutrient status.

Rodwell J.S.E. (1992) British plant communities. Volume 3. Grasslands and montane communities. Cambridge University Press, Cambridge, UK.

Note: If using or referring to this published study please read and quote the original paper, this can be viewed at:

Output references

What Works in Conservation

What Works in Conservation provides expert assessments of the effectiveness of actions, based on summarised evidence, in synopses. Subjects covered so far include amphibians, birds, terrestrial mammals, forests, peatland and control of freshwater invasive species. More are in progress.

More about What Works in Conservation

Download free PDF or purchase
The Conservation Evidence Journal

The Conservation Evidence Journal

An online, free to publish in, open-access journal publishing results from research and projects that test the effectiveness of conservation actions.

Read the latest volume: Volume 18

Go to the CE Journal

Subscribe to our newsletter

Please add your details if you are interested in receiving updates from the Conservation Evidence team about new papers, synopses and opportunities.

Who uses Conservation Evidence?

Meet some of the evidence champions

Endangered Landscape Programme Red List Champion - Arc Kent Wildlife Trust The Rufford Foundation Save the Frogs - Ghana Bern wood Supporting Conservation Leaders National Biodiversity Network Sustainability Dashboard Frog Life The international journey of Conservation - Oryx British trust for ornithology Cool Farm Alliance UNEP AWFA Butterfly Conservation People trust for endangered species Vincet Wildlife Trust