Experimental Techniques for Evaluating the Success of Restoration Projects

The ecological background of a restoration project is complex and difficult to betermine without experimentation. A useful context for experiments is the well-studied process of natural succession, because the factors that drive or inhibit succession are also at work during reclamation (a form of primary succession) and restoration (which often resembles secondary succession). Using experimental studies on urban wasteland reclamation, we have tested for factors that stimulate or inhibit succession during early phases of woodland development in the Northeastern United states. The emphasis has been on mutualisms (seed dispersal, pollination, and mycorrhizae) and microsite limitations in the recruitment, growth, and reproduction of woody plants. Using plantings of seeds, seedlings, and clusters of reproductively mature plants on abandoned landfills, we have observed that (1) soil microsite deficiencies lead to very poor germination (<0.1$\%$) and seedling survival (<0.01$\%$) of most native species; (2) seed dispersal by birds is a significant and reliable source of woody plant recruitment; however (3) proximity effects are strong, with most (up to 95$\%$) of seed rain falling in the vicinity of planted clusters that are closest to putative seed sources; and (4) remnant natural woodlands are critical components of the recruitment process. To emphasize the last point, in one case, we found that the destruction of approximately 50$\%$ of nearby natural woodland vegetation led to a commensurate decline in seed rain. In another case, we found that the species richness of recruits was strictly limited by the species composition of nearby source plant communities, with no evidence of community enrichment by long distance dispersal over 5 years. We conclude from these results that the size and proximity of remnant natural populations are critical considerations when planning reclamation and restoration programs that rely on natural successional processes.