Tree seed dispersal among forest fragments: I. Conifer plantations as seed traps

Abstract

Aim In fragmented landscapes, the survival of plant populations may depend on sufficient rates of migration between fragments with which to counteract local extinctions. Empirical information on the plant migration process is limited because of the difficulties of directly measuring the rare medium to long‐distance dispersal events involved. The potential for hardwood tree species' seedlings growing in conifer plantations to offer `a posteriori' measures of intermediate to long‐distance dispersal events was evaluated. Location Fragmented eastern deciduous forest in a rural, agricultural landscape, Southern Ontario. Methods To indicate the degree to which plantations `sample' incoming seed, seedling populations were enumerated in plantations and their species composition and abundances compared with those of mature trees, the presumed seed sources, in the surrounding landscape (to 150 m distance). To further quantify the degree to which hardwood seedling populations in plantations represent seed input, seeds of twelve hardwood species were sown in pine plantations and their emergence and early seedling survival monitored for 2 years. A subset of four of these species was planted in hardwood woodlots and their survival compared with survival in plantations to indicate how well seedlings in plantations reflect colonization in deciduous forest fragments. Results Hardwood seedling populations in plantations comprised the majority of tree species present in the surrounding landscape. Species' seedling abundances were significantly related to tree abundances around plantations. These findings indicate that most hardwood species can establish in plantations and do so in approximate proportion to the amount of seed supplied. All species produced first year seedlings in plantations, but rates of seedling survival were variable, ranging from under 5 to over 40% per species. Large‐seeded species had significantly greater seedling survival than smaller‐seeded species, suggesting that seed size—survivorship relationships could be used to adjust seedling frequencies to measure seed rain. Mean light levels photosynthetically active radiation (PAR) were significantly higher in plantations than in hardwood woodlots, yet rates of early seedling establishment were comparable between plantations and woodlots for four experimentally sown species. Main conclusions Hardwood seedlings in plantations appear to provide reasonable, although conservative, measures of dispersal to forest fragments, albeit with certain qualifications including adjustments for species' establishment abilities.