No abstract

Fifteen herbaceous perennial grass and herb species were planted into experimental green roofs in spring 2001. The species differed widely in their origins, heights, flowering times, life spans and growth forms but all were typical of dry and nutrient-stressed habitats. Three individuals of each species of a standardised size were randomly assigned to a planting grid at 20 cm apart in each experimental replicate plot at substrate depth of either 100 or 200 mm. Each treatment was replicated three times. During each growing season, the mean height and spread of each individual was recorded, together with flowering performance and % vegetation cover. In addition the numbers and % covers of all spontaneous colonised species were recorded. Greatest survival, diversity, size and flowering performance of planted species occurred at 200 mm depth. Bare ground and moss cover was greatest at 100 mm, as was diversity of colonising species. Differences between the early years and the final years of the experiment indicate the need for long-term monitoring of green roofs in addition to short-term experiments.

No abstract

Despite the fact that green roofs are based upon living systems, the majority of published research literature contains little specific information on the contribution of plants to the various functions and properties of green roofs. Furthermore, there has been little investigation of the influence of the composition of vegetation on the physical properties of a green roof system. This paper reviews previously published material that throws light on the role of vegetation composition on green roof function, with particular regard to rainwater runoff. Two experiments at the University of Sheffield, UK, are considered in detail: (a) An outdoor lysimeter experiment that investigated the quantity of runoff from trays containing 100 mm of growing medium and combinations of grasses and forbs, together with bare substrate, and (b) a greenhouse experiment using simulated rainfall to estimate the amount of rainfall intercepted by different vegetation types. In both cases the vegetation ranged from simple monocultures of forbs and grasses through to complex mixtures of both. In both cases, the composition of the vegetation was found to significantly affect both the amount of water retained and released from the system.

No abstract