Defining Historical Baselines for Conservation: Ecological Changes Since European Settlement on Vancouver Island, Canada
Anne D. Bjorkman, Mark Vellend (2010) Defining Historical Baselines for Conservation: Ecological Changes Since European Settlement on Vancouver Island, Canada. Conservation Biology, Volume 24, Issue 6, pages 1559–1568, December 2010
Full text [here]
Selected excerpts:
Abstract
Conservation and restoration goals are often defined by historical baseline conditions that occurred prior to a particular period of human disturbance, such as European settlement in North America. Nevertheless, if ecosystems were heavily influenced by native peoples prior to European settlement, conservation efforts may require active management rather than simple removal of or reductions in recent forms of disturbance. We used pre-European settlement land survey records (1859–1874) and contemporary vegetation surveys to assess changes over the past 150 years in tree species and habitat composition, forest density, and tree size structure on southern Vancouver Island and Saltspring Island, British Columbia, Canada. Several lines of evidence support the hypothesis that frequent historical burning by native peoples, and subsequent fire suppression, have played dominant roles in shaping this landscape. First, the relative frequency of fire-sensitive species (e.g., cedar [Thuja plicata]) has increased, whereas fire-tolerant species (e.g., Douglas-fir [Pseudotsuga menziesii]) have decreased. Tree density has increased 2-fold, and the proportion of the landscape in forest has greatly increased at the expense of open habitats (plains, savannas), which today contain most of the region’s threatened species. Finally, the frequency distribution of tree size has shifted from unimodal to monotonically decreasing, which suggests removal of an important barrier to tree recruitment. In addition, although most of the open habitats are associated with Garry oak (Quercus garryana) at present, most of the open habitats prior to European settlement were associated with Douglas-fir, which suggests that the current focus on Garry oak as a flagship for the many rare species in savannas may be misguided. Overall, our results indicate that the maintenance and restoration of open habitats will require active management and that historical records can provide critical guidance to such efforts.
Introduction
In many parts of the world, conservation and restoration goals are often based on perceptions of ecosystem states prior to intense disturbance by people of European origin (Foster 2000). Although relict patches of relatively undisturbed habitat are often used in studies of the structure and functioning of past ecosytems, such areas may be exceedingly rare and remain only in a highly biased subset of environmental contexts relative to the past (e.g., Vellend et al. 2008) or they may have been influenced in important ways by human activities that are not obvious through observation of their current state (e.g., Dupouey et al. 2002). An alternative is to look directly at historical records, which can provide critical insights into former landscape conditions and processes, and reasons for subsequent changes (Foster 2000; Whitney & DeCant 2005; Rhemtulla et al. 2009).
When using past ecosystem states to define conservation goals in North America, a key debate concerns the degree to which the pre-European settlement (henceforth presettlement) landscape was shaped by natural versus anthropogenic processes (Vale 2002). The pre-1492 landscape is often thought of as a pristine wilderness, where small groups of indigenous people had little ecological impact (Stankey 1989; G´omez-Pompa & Kaus 1992). This view has strongly influenced conservation strategies (Foster 2000; Hobbs & Cramer 2008). Many of these strategies rest on the assumption that removing the human factor will restore pre-European (“natural”) conditions (Foster 2000). Other studies suggest, however, that many areas of North America were “cultural landscapes” and as such were heavily influenced by native peoples through land cultivation or prescribed fire (Denevan 1992; Vale 2002). This implies that conservation strategies may need to consider not only current human disturbances, but also the possibility of reinstating cultural practices that were historically important in maintaining ecosystems (Higgs 1997; Anderson & Barbour 2003). Natural and anthropogenic processes have always been temporally dynamic, and both factors have influenced historical vegetation composition in North America (Vale 2002). The primary issue is the scale of anthropogenic impacts: were they highly localized around areas of intense land use (e.g., settlements) or were entire landscapes fundamentally transformed (Whitlock & Knox 2002). …
The 1859 notes revealed a landscape mosaic of prairie, plains, open woods, and forest. Of the 155 historical points with habitat descriptors, more than two-thirds were described as prairie, plains or open woods, with the remainder as forest, swamp, or bottom land (Table 1). … [T]he majority of 2007 points (79%) were classified as forest and only 3% as prairies or plains (Table 1). … Even if all undescribed areas were forests, however, the proportion of the landscape in forest still roughly doubled over the past 150 years, whereas the proportion of prairies or plains declined precipitously. …
Finally, although frequent fires do not necessarily imply an anthropogenic cause, our results do indicate that the fire regime was influenced by native peoples. The observed patterns are characteristic of landscapes prone to more frequent fires than expected by lightning strikes. Experiments suggest that the unimodal tree size distribution observed on Saltspring Island occurs at a fire interval of <5 years (Fule & Covington 1994; Peterson & Reich 2001). In contrast, a study in the Douglas-fir forests of Vancouver Island estimated a fire cycle of 5700 years, on the basis of the frequency of lightning strikes between 1950 and 1992 (Pew & Larsen 2001). A more localized study of charcoal in lake sediments (Cowichan Valley), which would not have detected low-intensity grassland fires, estimated a fire-return interval of 27–41 years, although it was difficult to detect distinct fire events relative to the background level of charcoal (McCoy 2006; Pellatt et al. 2007). These results suggest fires were more frequent than expected given the natural fire-return interval and were therefore likely related to anthropogenic activity (Turner 1999; MacDougall et al. 2004). …
Conclusions
Our results have important implications for the development of a perspective on historical human impacts in the New World that is more balanced than the simple dichotomy of landscapes as either humanized or pristine (Vale 2002). Ambiguities concerning the intensity, spatial extent, and timing of anthropogenic impacts were identified by Vale (2002) as being in great need of clarification in order to reconcile opposing viewpoints. Most of our study region appears to represent an intermediate level of intensity of historical landscape modification—less intense than within the confines of a densely populated village, but of sufficient intensity to have modified vegetation structure relative to what one would expect in the absence of humans.
In terms of space, the presence of both forested and open habitats historically suggests considerable spatial variability in the magnitude of human impacts, with prescribed fire likely to have maintained at least half of the landscape as open habitat (Table 1). In terms of time, the duration of impacts has likely varied across the landscape, with sites on relatively deep soils filling in rapidly with forest and sites on shallower soils remaining as open habitats to the present day (Vellend et al. 2008). Our results support the viewpoint that for the conservation implications of historical human impacts to be fully appreciated, conservation professionals need to abandon the simple yes–no dichotomy and embrace the reality of continuous variation in the intensity, spatial extent, and duration of such impacts.
Restoration efforts are often prone to uncertainty about target conditions (Higgs 1997; Hobbs & Cramer 2008), especially in areas with no appropriate reference sites to help define historical conditions. Land managers often follow a do-nothing approach and allow land to return to its “natural” state (Hobbs & Cramer 2008). Nevertheless, our study indicates that the open nature of the endangered savannas on Vancouver Island was likely maintained by fires purposefully set by native peoples.
Thus, restoration of these habitats to their pre-European state cannot be accomplished simply by removing human influences. Achieving the goal of maintaining open savannas would almost certainly need to involve active removal of encroaching trees and shrubs, either through burning or alternative strategies (e.g., mowing, tree removal) (MacDougall et al. 2004; Gedalof et al. 2006). Ideally such measures should be implemented in controlled experiments (i.e., through adaptive management; Walters 1986). Furthermore, the tree species currently most closely associated with savannas (Garry oak) does not appear—in a historical context—to be an appropriate flagship for savannas in general or their rich diversity of associated species. This highlights the necessity of understanding
and preserving ecosystem processes, rather than patterns only, and emphasizes the need for active management to achieve conservation and restoration goals in many ecosystems.
