Search for a command to run...
A large-scale field experiment on 11 closed old-style landfill sites aims to identify the constraints to tree survival and growth, and the opportunity for restoration to community forestry. This paper analyses survival and growth during the first 3 years. A preliminary site investigation showed that the main environmental constraints to tree growth were soil depth, site exposure, soil compaction, waterlogging and low soil oxygen. None of the sites were suitable for forestry according to current guidelines. Thirty-nine plots (mostly 40 m 40 m) at the 11 landfill sites were each planted with 21 woody species, amounting to 8 ha of experimental plots within 14 ha of buffer zone planting. All planting stock was of the same origin, planted at the same time in a randomised-block design. Despite herbicide applications, weed competition had the largest inhibitory effect on establishment in the first year. Patterns of survival and growth were contradictory; often species with poor survival rates grew well and vice versa. The most successful species were late successional species (compared to pioneers), native species (compared to exotics), shrubs (compared to trees) and broadleaves (compared to conifers). Eight of the 21 species could be recommended to be good general choices for landfill sites. It is concluded that all sites were suitable for community forestry, although two sites would require additional soil amendment. Species recommendations to suit a particular range of environmental determinants at other landfill sites is complicated by the mound-like shape of landfills and a high degree of within-site and within-plot variability; this was a more important determinant of mortality and growth than was the particular species planted. Differences between plots were reflected in mortality data, whereas growth data separated differences between species. Monitoring of the plots and further experimental work at the sites will continue for at least 15 years. # 2004 Elsevier B.V. All rights reserved.