The "ghosts of landscapes past" have been found to haunt the biota of both terrestrial and aquatic systems. Land-use legacies on vegetation communities are especially prevalent and well documented. In central Massachusetts (USA), for example, the legacy of past agricultural land-use from the 18th century is still reflected in the vegetation of today's forests. This landscape has undergone a complete transformation during the past three centuries: from a primarily forested region that was largely cleared for timber and agriculture by the mid-1800s following European settlement, to a now mostly forested region once again as industrialization during the latter half of the 19th century led to farm abandonment and migration to urban centers, thus allowing for natural reforestation over the past 100-150 years in spite of a steadily increasing human population (Foster et al., 1998). Despite the return to forest, the structure and composition of these forests have been dramatically—and perhaps permanently—altered by this relatively brief but intense period of deforestation and agricultural land use, such that current landscapes are much more homogeneous across the region than during Colonial times. Forests are now dominated by trees such as paper birch (Betula papyrifera), red maple (Acerrubrum) and white pine (Pinus strobum) that were relatively uncommon in Colonial forests, but which are essentially "pioneer species" that were capable of broad-scale dispersal and rapid establishment following severe disturbance wrought by clearing, cutting, and cultivation of the region (Foster et al., 1998). In contrast, species such as eastern hemlock (Tsuga canadensis) and beech (Fagusgrandifolia), which are long-lived and poor dispersers, have yet to recover their former abundance or extent.
Looking past the trees, the forest contains other land-use legacies from America's Colonial period. The current impoverishment of forest herbs is also thought to be a legacy of 18th-century forest clearing. The understory flora of woodlands that had been cleared and plowed, but which were allowed to reforest naturally following agricultural abandonment, still bear the signature of past cultivation: they contain more weedy species and fewer ericaceous shrubs (e.g., wintergreen,
Gaultheria procumbens, huckleberry, Gaylussacia baccata; wild raisin, Viburnum cassinoides) than woodlands that escaped the plow (Foster et al., 2003). Many forest herbs have low seed production, lack persistent seed banks, and are dispersal-limited (i.e., their seeds are ant-dispersed or lack morphological structures for longdistance dispersal) because they are adapted to relatively stable forest ecosystems that are characterized by fine-scale patch disturbances (Bellemare et al., 2002). These species have not exhibited rapid recovery and recolonization following their extirpation from areas that had been cleared and cultivated. In a modeling study of land-use change over a 300-year period, Matlack (2005) showed that seed dispersal ability was critical to the regional survival of forest herb species. Further, the model predicted that the legacy effects of agricultural land-use would likely persist for at least another century. Given the spatial and temporal scale of human land-use, slow-migrating species and those lacking gap-crossing abilities are most at risk of regional extinction, raising the possibility of an extinction debt for herbaceous species in these forests. Additionally, past land-use may affect nitrogen cycling and the spatial heterogeneity of soil resources (Fraterrigo et al., 2005), which could produce effects that persist for many decades, resulting in a fundamental shift in the composition and diversity of these forested ecosystems. If true, it may not be possible to recover historical vegetation even if dispersal limitation is eventually overcome in time (see also Dupouey et al., 2002).
Past land use also influences the biological diversity of aquatic systems. Aquatic systems have a strong dependence on the surrounding landscape, and land-use practices throughout the watershed may affect a wide range of conditions, such as hydrology, organic inputs, temperature, and water chemistry, and are thus capable of contributing to strong legacy effects (Allan et al., 1997). For example, patterns of fish and invertebrate diversity within streams draining two watersheds in the southern Appalachians were best explained not by current land use, but by the intensity of agricultural land use some 40 years earlier (Harding et al., 1998). Although some streams currently flow through watersheds that are mostly forested, their complement of fish and invertebrate species more closely resembled those found in agricultural streams. Significantly, these "anomalous" forested streams were in watersheds that had formerly experienced a high degree (~40%) of deforestation and agriculture in the 1950s. Reforestation over the past half-century has thus resulted in little effective recovery of these stream communities. As in terrestrial systems, the recovery of aquatic biota from high-impact disturbances such as deforestation or agriculture—even though seemingly removed from the stream or lake in question—can still take decades to achieve.
It comes as no surprise that landscape transformation has such profound effects on biological communities; rather, the surprise is that these effects are so persistent even after human activities have ceased and vegetation has been allowed to recover (however illusory that recovery may be). Even if transformation was not complete or particularly extensive, human land use can still have other more subtle effects on landscape structure, which may have no less a dramatic effect on patterns of diversity. For example, a seemingly trivial loss of habitat at a critical point can effectively disrupt the habitat connectivity of the entire landscape, which may have consequences for biodiversity that far exceed the actual amount of habitat lost (e.g., nonlinear or critical threshold responses; With and Crist, 1995). A disruption of landscape connectivity can reduce dispersal or colonization success and enhance species extinction risk (With and King, 1999a,b), even when local conditions are not directly affected by land-use activities. As with landscape transformation, a disruption of connectivity has the potential to produce strong legacy effects in diversity patterns. For example, high plant species diversity within the small remaining patches of semi-natural grassland in Sweden is a relic of a formerly connected open farming landscape that existed nearly a century ago (Lindborg and Eriksson, 2004). These grasslands have declined more than 90% during the past 80 years, such that historical grasslands had much higher connectivity than present-day remnants. Subsequently, these grasslands have maintained a higher diversity of plants than might otherwise be expected based on the current amount and distribution of habitat. Similarly, historical habitat connectivity still exerts an influence on the distribution of carabid beetles within hedgerow networks of France (Burel, 1992; Petit and Burel, 1998). Hedgerows have been declining since the 1950s as a result of a shift from traditional to modern farming practices, resulting in increased isolation of beetle populations. It may take many decades, however, before beetle populations disappear from isolated hedgerows. Beetle distributions thus exhibit a 'memory' of past landscape structure, with the result that current carabid beetle assemblages better reflect the historical landscape structure of a half-century ago than the present-day hedgerow network. It is worth noting, then, that land-use legacies are not always negative (e.g., depauperate herb or fish communities in reforested landscapes following agricultural abandonment), but may actually appear to be positive (a retention of native species, such as carabids or grassland plants, in spite of past land clearing), at least in the short term.
Was this article helpful?