ECOSYSTEM RESTORATION IN A FRAGMENTED LANDSCAPE
Most of the Piedmont north of the Potomac River has been settled for over three centuries. Therefore, it is not surprising that at the beginning of the 21st century the vast majority of the province is directly controlled by humans for urban and agricultural uses and that only about 30% of the land is wooded or in some stage of succession. This long period of settlement and use also means that the natural areas that remain have a history of heavy and repeated disturbance which continues into the present thanks to encroaching urbanization, air pollution, acid rain, nitrogen deposition, altered hydrographs, and ground and surface water pollution.
Larger natural areas in the northern Piedmont generally have been assembled by aggregating abandoned agricultural fields, associated woodlots, and selectively logged forest. The juxtaposition of meadows and forest produces an “edge” at which the forest abruptly stops and gives way to fields dominated by low-growing, short-lived herbaceous species. Prior to the wholesale dominance by human activities, disturbances in the forest such as fire, a gap in the canopy produced by the death of a large tree, or abandonment of an agricultural field also produced “edge” habitat, but the openings, gaps, and fields quickly regrew to thickets, young woodlands and, in less than a century, to forest because the disturbed area was small relative to the natural surroundings. The preponderance of “edge” habitat in the northern Piedmont that is our legacy of historical agriculture and contemporary suburbanization has now set the stage for serious problems with forest preservation and restoration. Two characteristics in particular related to forest fragmentation and “edge” habitat often stall succession in its earliest stages.
The first of these features is the presence of large herds of white-tailed deer. Deer are edge-adapted, crepuscular animals that feed in fields and meadows then retreat for cover into the forest. Deer in numbers approaching or even slightly exceeding the biological carrying capacity of the land will prevent forest regeneration in forest gaps and abandoned fields. Furthermore, deer selectively browse vegetation, decimating native shrubs and wildflowers first and leaving many non-native species undisturbed. Hunting is the only feasible means currently available for controlling deer and hunting may be difficult to implement in small natural areas or those that have a strong anti-hunting constituency. In many cases, however, tree-stand based archery hunting or a combination or archery and limited shotgun hunting may be possible if the hunting is managed carefully.
The second and certainly more insidious and intractable characteristic of Piedmont forests which delays or even stops succession is the overwhelming abundance of non-native plants. Most of the plants that are problematic were imported from eastern Asia where similar climatic conditions virtually “pre-adapt” the species to ambient conditions of the North American Piedmont. Alien plants can produce problems at both the edge of the forest and deep within the woodland.
Those plants that invade the edge of the forest typically are vines that prosper in disturbed conditions and direct sunlight. Multiflora rose (Rosa multiflora, a shrub), Asian bittersweet (Celastrus orbiculatus), porcelainberry (Ampelopsis brevipedunculata), kudzu (Pueraria lobata), and the newest threat, mile-a-minute (Persicaria perfoliata) can quickly overwhelm and destroy a forest edge once they become established. Deeper within the forest, the herbaceous layer and the spring ephemeral flora is endangered by species adapted to low light intensities. Garlic mustard (Alliaria petiolata) is a Eurasian biennial herb that spreads rapidly and limits space for native vegetation through crowding. Japanese honeysuckle (Lonicera japonica) and five-leaved akebia (Akebia quinata) grow profusely in the sun, but are able to invade woodlands and sprawl over the forest floor excluding desirable native species. Microstegium vimineum, an annual grass with an extraordinarily dense growth habit, has the unusual ability to grow as well in full sun as it does with 84% shading, thereby allowing it to invade woodlands with incomplete canopy closure.
Floodplains and riparian areas are especially prone to problems. The constant natural disturbances produced by flooding have been exacerbated by changes in the hydrographs of many urbanized watersheds which lead to unusually severe floods. In addition, riparian corridors tend to be narrow, so that all of the floodplain is “edge” habitat especially prone to invasion. Two herbaceous species in particular have taken advantage of these conditions and have become widespread in riparian corridors in the Northeast: Japanese knotweed (Fallopia japonica) forms huge bamboo-like clumps and lesser celandine (Ranunculus ficaria) carpets the forest floor in early spring. Both species effectively exclude desirable vegetation.
Options for controlling invasive plants
The options for bringing invasive vegetation under control are dictated by conditions present at each site and by the human and financial resources available to the manager. In natural areas that are heavily wooded or support forest ecosystems that are largely intact, “surgical” removal of individual plants with hand tools is effective. Volunteers can be used successfully for this work with a modicum of education and training. However, energetic volunteers cannot simply be “turned loose” in the forest without guidance because of the damage they can cause to desirable vegetation and to themselves. Power tools generally are not needed for alien plant control, although chain saws are often helpful if vines are large and thick. A tool that has proven to be useful is the portable gasoline-powered hedge trimmer (about $300) which can be used effectively on invasive shrubs like privet (Ligustrum spp.), multiflora rose, and the shrub honeysuckles (Lonicera spp.) and has the added advantage of mincing the vegetation into very small pieces that fall to the ground; using loppers or chain saws, intact branches or rosebush canes have to hauled away for disposal. Of course, hand labor is expensive and slow, which limits its usefulness to areas where no other options are available.
In some instances, a lucky combination of factors will allow managers to employ power equipment in the battle to control alien plants. If the terrain is suitable, and more importantly, the native ecosystem is deemed unsalvageable by less drastic measures, a heavy-duty tractor-driven brush mower is very effective at quickly clearing large areas and keeping non-native species under control. The initial assault will probably require careful reconnoitering to avoid hidden obstacles and uneven topography. If living trees remain, some careful hand cutting and removal of vines may be necessary after the mower has provided access to the site.
Herbicides provide another option for control. The Pennypack Trust coordinated efforts of a consortium of natural areas managers in southeastern Pennsylvania who worked together to study the management of troublesome invasive plant species. The Trust investigated techniques for controlling porcelainberry. We divided an abandoned pasture that had been overrun with porcelainberry into a set of 4m2 plots and tested two mechanical controls (intensive mowing and “grubbing” out the above- and below-ground biomass) and four herbicide treatments (glyphosate in summer, glyphosate in autumn, 2,4-D in summer, and a triclopyr/2,4-D mixture in summer). Results showed that grubbing, glyphosate in autumn, 2,4-D alone, and the 2,4-D/triclopyr mixture were equally and highly effective for controlling porcelainberry. While glyphosate has relatively low toxicity, it has two disadvantages: (1) it kills all vegetation, leaving bare ground subject to erosion, and (2) the use of glyphosate allows poison ivy to resurge in the year following application. For these reasons, the Trust has decided to use broadleaf herbicides such as triclopyr where feasible. Other species investigated as part of this program included multiflora rose, Japanese honeysuckle, the wetland invader purple loosestrife (Lythrum salicaria), shrub honeysuckle, English ivy (Hedera helix), and five-leaved akebia.
Control techniques are valuable for reducing alien vegetation to manageable levels, but they are not permanent solutions to the problem of forest deterioration and arrested succession. The only permanent solution is to produce conditions that favor native vegetation and discourage alien species from becoming established and proliferating. In the northern Piedmont, a permanent solution is to restore a complete native forest canopy to areas currently dominated by invasive vegetation, thereby excluding or reducing the vigor of the sun-loving invasives. Of course, planting unprotected seedlings and saplings of native tree species is a futile task in areas with high deer population densities, so newly planted trees must be protected from predation. Cylindrical plastic tree shelters are one option for protection. In addition to protecting the seedling from rodent damage and herbivore predation, tree shelters create an ideal moist and warm microenvironment that accelerates the growth of the tree.
While tree shelters show great promise for restoring forest where deer are a problem, they are not without disadvantages. The shelters range in cost from $4.25 to $5.25 each if purchased in small quantities. They must be kept free of shading weeds in summer and dead grasses in winter that harbor rodent predators. The wooden stakes that support the shelters usually last for only 2-3 growing seasons and must be replaced or the trees will tip over. The tops of the shelters need to be capped with plastic mesh to exclude cavity-nesting birds, and paper wasps find the shelters irresistible for building nests. Furthermore, once the tree inside the shelter has filled the capacity of the shelter, the plastic must be cut away or the tree will be girdled or will rot from the accumulation of precipitation trapped against the trunk. Nevertheless, comparisons of growth between sheltered and unsheltered trees (growing inside a deer exclosure) were truly dramatic and encouraged the Trust to use the shelters extensively in reforestation.
Shelters cannot be used effectively in shaded conditions such as canopy gaps in existing forest. The light diminution attributable to the shelter compounded by the intermittent shading of adjacent trees reduces the light available to the seedling inside the shelter to levels that will not allow the tree to live. Instead, trees planted in canopy gaps can be protected with wire cages (which are more expensive than shelters and subject to being overwhelmed by vines) or the land manager can choose to plant larger trees whose crowns are above the browsing height of deer and to protect the stem from antler rubbing with the use of tree wrapping tape.
Other interesting options for long-term forest restoration and stewardship are under investigation. For example, the Pennypack Trust has been working with researchers from Rutgers University and the University of Pennsylvania to examine several manipulations designed to accelerate the changes that occur as old fields succeed to native forest without planting seedlings. Three soil treatments (reducing soil fertility, sharply increasing shading, and augmenting the organic content of the soil while at the same time reducing sites for seedling germination) combined with three native woodland seedbank enhancement treatments yielded promising results that suggested ways to manipulate soil fertility to reduce the intensity of management.
Restoring forest is not a project to be undertaken lightly; it is an expensive, on-going, and long-term commitment. A woodland canopy that is capable of producing enough shade to reduce the vigor of invasive vegetation will not close over the forest floor for 15 to 20 years, necessitating at least some invasive plant management for up to two decades. Restoring woodland to a floodplain will require permanent stewardship because of the inherently disturbed nature of the riparian corridor.