ANNUALS AND SEDUM-CUTTINGS IN SEED-MIXTURES FOR EXTENSIVE ROOF GARDENS

In March 1999 five seed-mixtures were sown on a shallow substrate bed on a roof near Bernburg. The aim was to study the effect of adding annuals, or cuttings of several sedum-species, respectively, instead of the usual high percentage of grassseeds, in order to optimise the coverage and density, even without a high amount of grasses in the first and second year after sowing. A second factor involved two different fertilisers. Only on fertilized plots in the second year was a covering rate > 60% achieved. The influence of fertilizer in the second year was substantially larger than the effect of the seed-mixtures. Suitable annuals were Gypsophila elegans, Iberis amara, Silene armeria and Xeranthemum annuum, which were able to form aspects in the first year. Gypsophila elegans provided a stable population in the second year. Even without increasing the coverage significantly, it is highly recommended as an addition to seed mixtures because of its flowering aspect. The trial confirms the importance of fertilisation on shallow-substrate roof-gardens. Cuttings of Sedum album, S. sexangulare and S. reflexum can reach high coverage relatively fast, but on larger areas involve higher costs. Even a relatively low addition of 10 g/m2 Sedum to seed mix can result in a significant effect and reduce the proportion of grasses, if fertilised even in the first vegetation period after sowing. INTRODUCTION According to guidelines from the “Forschungsgesellschaft Landschaftsentwicklung Landschaftsbau e.V.“ (FLL, Research association of landscape development and landscape construction) green roofs need to demonstrate a covering of more than 60% in order to be approved (FLL 1995). To achieve this in the shortest time possible, mostly high percentages of grasses, especially Festuca ovina agg. are mixed into the seed compositions. Their rapid growth unfortunately leads to a rather poor flowering aspect because of the low presence of forbs. On a roof garden near Bernburg (SachsenAnhalt;), a central German drought area, a comparative sowing trial was tried to answer the following questions: 1. Can seed mixes with less grasses but additions of annuals or Sedum cuttings respectively optimise the coverage in the 1 and 2 year after sowing? and; 2. What is the effect of fertilisation on these seed mixtures? MATERIALS AND METHODS Research was conducted on the roof of a garden-shelter occupying an area of 42 m2 in a private garden near Bernburg (ca. 480 mm average annual rainfall). The site comprised the following construction profile: basic drain layer (3 cm): shredded styrofoam, filter layer: filter fleece, 90g/m2, and substrate (7 cm): expanded slate (8/16, 80%) + Perlite (15%) + Flora-roof-substrate (5%). The roof had a 10% slope northward. There were twenty plots, each measuring 2 m2 for 10 variations in 2 repetitions. The trail was irrigated only in 1999 every two weeks in dry periods. The basic mix was Roofgarden according to the seed producing company “RiegerH and the trail involved 17 annual species, and cuttings from 5 Sedum species. Treatments were: A1: 2g basic mix –original, A2: 1g basic mix (without annuals and grasses) + 1g grasses, Proc. IC on Urban Horticulture Eds: R. Junge-Berberovic et al. Acta Hort 643, ISHS 2004 302 A3: 1.8g basic mix (without annuals and grasses) + 0.5g annuals (about 40 seeds each species) + 0.2g grasses, A4: 0.35g basic mix (without annuals and grasses) + 0.15g grasses + 100 g Sedumcuttings, A5: 1g basic mix (without annuals and grasses) + 1g grasses + 0.5 g annuals and + 50 g Sedum-cuttings. A list of all species is published in Kircher (2002). Seeding date was in week 13/1999. In the case of fertilisation the treatments were: B1: 1999: without fertiliser, week 17/2000: 4 g N/m2 as Osmocote Plus 5-6M (15:10:12:2), B2: week 18/1999: 10 g N/m2 as Nitrophoska permanent (15:9:15:2), week 17/2000: 10 g N/m2 as Osmocote Plus 5-6M (15:10:12:2). Compared with the recommendations of Fischer & Jauch (2001) and Kolb (2002) a relatively high dosage of fertiliser was chosen, because annuals as R-strategists (Grime 1986) grow naturally on disturbed sites with high availability of nutrients, in particular nitrogen. The degree of coverage (%), number of individuals and height (cm) of each germinated species were estimated or counted on the 28.07.99 and 31.05.00. Obviously dead plants were not evaluated. RESULTS AND DISCUSSION Overall Coverage Mixes with annuals as well as with blends of Sedum-cuttings showed distinctly higher coverage rates than variant A1. On all fertilised plots – except variant A1 – in the second year a coverage rate of > 60% was reached. The influence of fertilising in the second year is more significant than that of species composition. There was a distinct lower coverage of the plots sown with basis-mix without annuals or Sedum-cuttings (A1). Grasses and Forbs The grasses Festuca ovina and F. pallens appeared to dominate. Amongst the forbs Petrorhagia saxifraga was the most successful. Dianthus, Silene and Thymus species, as well as Anthemis tinctoria germinated well, but most of the seedlings were eliminated during the first summer through drought. Some Allium schoenoprasum seedlings became recognisable in the third year after sowing. The other species appeared not to be able to withstand the extreme stress caused by drought. Annuals Only Gypsophila elegans, Iberis amara, Silene armeria and Xeranthemum annuum made reasonable plants in the first vegetation period, but without a significant contribution to increased coverage. Erodium cicutarium, a native R-strategist in poor drymeadows was inconspicuous. The same applied to Arenaria serpyllifolia, which germinated in large numbers, but visually remained in the background because of its low growth. However it did maintain its presence even into the third year on the roof. Though Iberis amara and Xeranthemum annuum germinated well they didn’t reach suitable heights for flowering. Silene armeria was only found in few specimens, but was able to produce attractive flowers on the fertilised plots. Gypsophila elegans surprised with stable populations even in the second year. Despite not seriously contributing to a higher coverage, its flowering in autumn enriched the visual effect. The effect of fertilising to the appearance of the annuals was unexpectedly low. With Gypsophila the low fertilised plots of the second year had better ratings. Sedum-Cuttings Sedum floriferum 'Weihenstephaner Gold' and S. hybridum 'Immergrunchen' did not grow, and S. forsteranum hardly succeeded at all. Several cuttings were blown off by