Herbivory, hunting, and long-term vegetation change in degraded savanna

B I O L O G I C A L C O N S E RVAT I O N

1 4 1 ( 2 0 0 8 ) 2 1 7 4 –2 1 8 3

available at www.sciencedirect.com

journal homepage: www.elsevier.com/locate/biocon

Review

Herbivory, hunting, and long-term vegetation change in degraded savanna A.S. MacDougall* Department of Integrative Biology, University of Guelph, Guelph, Ontario, Canada N1G 2X8

A R T I C L E I N F O

A B S T R A C T

Article history:

Large ungulate populations are associated with the degradation of many forest plant com-

Received 11 December 2007

munities, but it is unclear if these population sizes are strictly a contemporary phenome-

Received in revised form

non. Human exploitation models predict they are not, with ungulate numbers varying with

17 June 2008

long-term fluctuations in hunting pressure. Alternatively, human disturbance models pre-

Accepted 5 July 2008

dict that abiotic limitations normally restrict herbivores, with contemporary increases

Available online 16 August 2008

reflecting increased productivity associated with agriculture and forestry. Both can explain ungulate abundance, but may have different implications for plant conservation because

Keywords:

they predict different levels of prior evolutionary exposure to herbivory. Here, I review his-

Herbivory

torical records and stand structure studies from degraded oak savanna of western North

Hunting pressure

America to examine whether current ungulate levels are strictly a contemporary phenom-

Oak savanna

enon. Although it was impossible to quantify pre-European herd sizes, all evidence indi-

Oak recruitment failure

cates a strong relationship between hunting pressure and ungulate abundance. Historical

Historical documents

accounts repeatedly describe large herds of deer and elk at first European contact, followed by sharp declines immediately after colonization, and then rapid recovery beginning in the early 1900s as subsistence hunting waned. Stand structure data for oak woodland appear to support this model. Present-day oak woodlands mostly derive from mass recruitment from 1850 to 1910, coinciding with the near elimination of ungulates by hunting. Although these results suggest that large ungulate herds are not strictly a contemporary phenomenon, browsing intensity appears to be unprecedented given limited hunting, predator extirpation, and savanna fragmentation within productive pasture and early successional forest. Hunting pressure thus continues to be important, in that it is now largely absent. Ó 2008 Elsevier Ltd. All rights reserved.

Contents 1. 2.

3.

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.1. Study region . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.2. Literature review and stand descriptions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . First contact descriptions of herbivore abundance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

* Tel.: +1 519 824 4120. E-mail address: [email protected] 0006-3207/$ - see front matter Ó 2008 Elsevier Ltd. All rights reserved. doi:10.1016/j.biocon.2008.07.003

2175 2176 2176 2176 2178

B I O L O G I C A L C O N S E RVAT I O N

4. 5. 6.

1.

1 4 1 ( 2 0 0 8 ) 2 1 7 4 –2 1 8 3

2175

Post-settlement herbivore declines . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Oak stand structure data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Implications for conservation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Acknowledgements. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Introduction

Herbivore effects on vegetation are ubiquitous to most grassdominated ecosystems, and the transformation of these effects has been associated with the loss of structure, function, and diversity in grasslands worldwide (Fleischner, 1994; Cote et al., 2004; Maron et al., 2006; Parker et al., 2006). Transformations include herbivore expansion following hunting reductions or predator loss, herbivore collapse after predator introduction, or the replacement of native grazers with exotic livestock. Two primary challenges for quantifying herbivore effects on grasslands are that detailed descriptions of herbivore dynamics prior to European settlement are lacking or vague, and that they can co-occur with other changes (e.g., plant invasion, habitat loss, fire suppression). From a conservation perspective, both obscure appropriate management practices including restoration targets. In systems undergoing substantial herbivore expansion, as is the case in many regions of North America (e.g., Odocoileus spp.), the degree of impact on grassland plants may depend on whether these increases are strictly a contemporary phenomenon (Mack and Thompson, 1982; Daubenmire, 1985; Martin and Szuter, 1999). Models emphasizing the importance of hunting pressure predict wide fluctuations in ungulate populations in the past, depending on variation in human population sizes (Martin and Szuter, 1999). Models emphasizing the importance of contemporary human land use practices predict that abiotic limitations previously restricted herbivores, with contemporary increases reflecting increased productivity associated with agriculture and forestry (Daubenmire, 1985; Van Vuren, 1987; Lyman and Wolverton, 2002). Although both can explain present-day ungulate abundance, they can have different management implications for plants. Because hunting models imply that ungulate numbers have been high previously, resident plants could be less sensitive to browsing compared to other changes (e.g., invasion, fire suppression). Management actions targeting browsers could have limited effect on plant recovery if these other factors are more limiting. Human disturbance models, in contrast, predict that ungulate populations may never have been larger, implying that plants may be less likely to tolerate intensive browsing and, thus, rapidly decline in its presence (Rooney, 2001; Rooney and Waller, 2003). Because we often lack information on long-term population trends for ungulates, the relative importance of these models for regulating browser populations, and their affect on plants, remains debated. Here, I examine evidence of former herbivore levels in coastal oak savanna of the Pacific Northwest of North America, using historical records and current vegetation structure as an indication of past events. Determining regulatory mechanisms for ungulates in northwestern grasslands has been

2179 2179 2181 2181 2181

especially problematic. As is typical of degraded grasslands, there has been invasion, species loss, and altered function (e.g., nutrient cycling). However, the role of herbivory in these changes is unclear. Some historical evidence suggests that ungulate herds were largely absent in the northwest prior to European settlement (Burroughs, 1961). As a result, Mack and Thompson (1982) hypothesized that extensive plant invasion derived from the arrival of grazers into a plant community lacking evolutionary exposure to intense herbivory. Alternatively, other records reported large but localized ungulate populations in some areas, especially towards the coast (e.g., Menzies, 1923; Ludrin, 1928; Lamb, 1984; Brown, 1989; Gough, 1992; Rollins, 1995). There could be a number of explanations for these discrepancies in the historical literature, ranging from differences in the timing and location of the observations, to different socio-cultural histories (e.g., some observations may have been preceded by disease epidemics, which reduced hunting pressure). I focus this investigation mostly on southeastern Vancouver Island (Fig. 1), British Columbia. Because European colonization occurred there relatively recently (mid-1800s), there are extensive records of

Fig. 1 – Distribution of oak savanna in northwestern North America, with locations of place names referenced in the text. Historical reports on ungulate abundances came mostly from areas within and around the major settlements (Victoria, Cowichan Valley, Comox). Modified from MacDougall et al. (2004).

2176

B I O L O G I C A L C O N S E RVAT I O N

1 4 1 ( 2 0 0 8 ) 2 1 7 4 –2 1 8 3

conditions during, and after, first contact, thereby providing an opportunity to examine ungulate population trends in some detail. There are also extensive dendrochronological and palynological studies (Gedalof et al., 2006; Pellatt et al., 2007), which can be used to infer past herbivore influences on current stand structure, and their possible interaction with other factors known to influence savanna dynamics (e.g., fire, grass invasion).

2.

Methods

2.1.

Study region

Oak savanna of southeastern Vancouver Island is part of an oak-grassland complex of western North America that extends south to California (Fig. 1). Habitat degradation (invasion, habitat loss, species extirpation) characterizes its entire range (Barbour and Major, 1977; Dunn and Ewing, 1997; MacDougall et al., 2004; Maret and Wilson, 2005). This oak savanna of Vancouver Island once covered at least 20,000 ha but less than 10% remains (Lea, 2006). Vegetation structure varied from open grassland to partially treed savanna (MacDougall et al., 2004). Now, remnant areas are mostly on south-facing hillsides with shallow soils, which limited agricultural expansion and infilling by woody plants. Oak woodland (canopy over >75%) appears to be more common now than previously (Fig. 2). Native flora includes a diverse pool of forbs and grasses, with relative abundances affected by plant invasion, browsing, soil depth, aspect, and fire frequency (Erickson and Meidinger, 2007; Gonzalez, 2007). Many of these plants have distributional centers within the California Floristic Province (Hickman, 1993). On Vancouver Island, exotic grasses, forbs, and woody plants dominate most remnants that maintain native plant populations. Formerly in this system, there were unknown population sizes of Roosevelt elk (Cervus canadensis roosevelti), black-tailed deer (Odocoileus hemionus columbianus), cougars (Puma concolor vancouverensis) and gray wolves (Canis lupus nubilus). Today, only deer persist within remnant savanna. Current estimates of deer density range from 10 to <40/km2 (Gonzalez, 2007), with densities above 4–15/km2 associated with negative effects on tree recruitment on Vancouver Island (MacTaggartCowan, 1945; Martin and Baltzinger, 2002). These densities are considered high, as is typical of many regions of western North America (McCullough et al., 1997), but it is unclear whether this is a contemporary phenomenon. Deer increases are associated with a number of factors, including increased productivity caused by agriculture and forestry, limited or no hunting, and predator eradication.

2.2.

Literature review and stand descriptions

First-hand historical descriptions of southeastern Vancouver Island were obtained from a range of sources, including provincial archives and land-survey offices, book reprints (e.g., MacFie, 1972; Brown, 1989; MacLachlan, 1998), and recent syntheses that focus on the settlement history of indigenous (the Hul’q’umi’num central coast Salish, hereafter referred to as Hul’q’umi’num) and European people in this region (e.g., Anonymous, 1986; Mackie, 1995). The descriptions spanned

Fig. 2 – Infilling of oak canopy between 1952 and 1992 at the Cowichan Garry Oak Reserve, Duncan British Columbia. The black arrows show the location of a house, built in 1893. The black square indicates the location of a barn, which remains standing but is now obscured by the oak canopy. Most of the trees in these pictures established during the late-1800s; no new trees have established since 1910 (Pellatt et al., 2007).

the period beginning with sailing expeditions of the late 1700s, to the first colonization by Europeans at Victoria in 1842, and ending in the early 1900s (Fig. 3). I focus in particular on historical accounts of the Cowichan Valley (Figs. 1, 2 and 4), where deer numbers are currently high and historical accounts from the 19th century are numerous, including land surveys, botanical descriptions, and ethnographic reports (e.g., Curtis, 1913; Anonymous, 1986; Brown, 1989; Arnett 2007). The Cowichan Valley was colonized by Europeans in 1862; prior to that it was inhabited by indigenous peoples for at least several thousand years. The descriptions were examined for references to herbivore densities, hunting practices including the influence of European colonization on the land use activities of the Hul’q’umi’num, vegetation appearances including species

B I O L O G I C A L C O N S E RVAT I O N

1 4 1 ( 2 0 0 8 ) 2 1 7 4 –2 1 8 3

2177

Fig. 3 – Timeline of historical events affecting the Garry Oak Ecosystem of southeastern Vancouver Island, including qualitative reports of ungulate abundance in the historical literature (shaded areas). High deer abundances were reported beginning in the late 1700s. After the 1850s, reports of deer scarcity were common until the 1890s. Major disease epidemics within the indigenous population preceded the first European sailing expeditions of George Vancouver, and periodically occurred throughout the 1800s.

Fig. 4 – Distribution of oak savanna prior to European colonization and present-day, plus the locations of the Cowichan First Nation’s winter camps in 1858, in the Cowichan Valley, British Columbia. The Cowichan Garry Oak Reserve is the remnant near the northeastern corner of Quamichan Lake. Most remnant areas are on south-facing steep slopes unsuitable for agriculture. Modified from MacDougall et al. (2006).

composition, disturbances including fire, and, for the Cowichan Valley, changes to these factors that occurred following

colonization by Europeans. The land use practices of the Hul’q’umi’num were also determined from the first-hand

2178

B I O L O G I C A L C O N S E RVAT I O N

accounts of the European settlers and, later on, from the ethnographic surveys that began in the late-1800s. Because historical documents can be prone to observer bias (e.g., over-emphasizing deer numbers or land productivity to entice settlement), each document was assessed qualitatively in an attempt to determine the purpose of the document, the circumstances of the observation, and whether different observations at the same locations or times were concordant (Egan and Howell, 2001; Foster and Motzkin, 2003). Evidence for long-term changes to the stand structure and vegetation composition of the savanna was derived from dendrochronological and palynological surveys that cover most of the areas on, or adjacent to, Vancouver Island (Gedalof et al., 2006; Pellatt et al., 2007).

3. First contact descriptions of herbivore abundance The early records by European explorers and colonists indicate that large ungulate populations are not only a contemporary phenomenon. Although these records tend to focus on the availability of resources (fresh water, timber, coal, hay) or agricultural suitability, when game was described, the emphasis was on abundance (Fig. 3). Prairies were not expansive, more pocket-like, but the ‘‘number of deer was almost incredible for the small expanse of country’’ (Mackie, 1995). George Vancouver reported in 1792 that ‘‘deer could be seen. . .in great numbers’’, and that ‘‘nature had here (Whidbey Island) provided a well-stocked park’’. This was repeatedly supported by accounts along southeastern Vancouver Island in areas of, or adjacent to, oak savanna (Fitzgerald, 1848; MacFie, 1972; Anonymous, 1986; Brown, 1989; Mackie, 1995). Settler accounts from the 1860s commonly report herds of deer and elk that were in ‘‘astounding abundance and easily shot’’ (MacFie, 1972; Brown, 1989; Mackie, 1995; Arnett, 2007). Reports from the Cowichan Valley in 1864 indicate ‘‘great herds of elk and many deer’’. They also describe the strong reliance of the Hul’q’umi’num on hunting of both deer and elk in the fall and winter, when ungulate herds converged on the coastal lowlands where most savanna is found. The Hul’q’umi’num were often loath to reveal popular hunting areas to European settlers (MacFie, 1972; Brown, 1989; Mackie, 1995), and, by the late-1800s, bemoaned the scarcity of deer compared to earlier times (Mackie, 1995; Arnett 2007). The historical records suggest several possible reasons for high ungulate abundance, despite (or because of) what was assumed to be an area with a densely concentrated indigenous population (e.g., Harris, 1997). Two of the more prominent explanations were seasonal migrations by the Hul’q’umi’num that left parts of Vancouver Island unoccupied for long periods of the year, and the effects of human disease and conflict beginning in the late 1700s. Land-management practices by the Hul’q’umi’num may have also contributed to large ungulate populations, via burning to promote forage or the regulation of hunting effort based on herd densities (e.g., Schultz, 1980), although neither practice was directly described in the historical literature reviewed here. Ethnographic and historical reports uniformly agree on the highly seasonal variation of local human population densities, which could reduce hunting pressure. In the Cowichan

1 4 1 ( 2 0 0 8 ) 2 1 7 4 –2 1 8 3

Valley, most people were absent from spring to autumn (Curtis, 1913; Brown, 1989; Arnett, 1999). In April–June, many moved to the adjacent Gulf Islands or Victoria to harvest bulbs of the Camassia lily, where it was phenologically more advanced due to the earlier onset of warmer and drier conditions (Curtis, 1913; Arnett, 2007). In June 1864, Robert Brown reports that ‘‘the whole population (of the Cowichan Valley) with the exception of about a dozen old women and children were absent gathering clams. . .for winter use and camas in the district round the Victoria. . .’’ (Brown, 1989). These accounts highlight the seasonal abundance of other types of food in this region, also including fish, whales, sea-lions and seals, and other plants, further reducing the need to hunt. After June, many people moved to the lower Fraser River (Fig. 1), where salmon runs start several months earlier than on the Cowichan River. Some of the best fishing sites on the lower Fraser River were controlled by the Cowichan people (MacLachlan, 1998; Lamb, 2007). Families tended to return to the Cowichan Valley in autumn for the onset of the salmon runs, and stayed there for the winter. The winter settlements were close to the river to facilitate fishing (Wells, 1858; Harris, 2001; Fig. 4). The nearby savanna was used for hunting and plant gathering, but was otherwise unoccupied year-round (Arnett, 1999). Deer were important staples for food and hides (Suttles, 1987), and were sometimes hunted by large parties that used nets or drove deer into deep pits (Anonymous, 1986; Arnett, 2007). It is difficult to determine the exact effect of seasonal migrations of hunters on deer and elk densities, but levels of exploitation must have been lower compared to levels of hunting reported in the post-settlement period (see below). Disease and conflict among indigenous populations drastically reduced their populations, and altered where they lived and how they managed the land (Archer, 1978; MacLachlan, 1998; Marshall, 1999; Keddie, 2003). Epidemics occurred in the 1770s, the mid-1830s, and 1850s–1860s (Harris, 1997; Keddie, 2003) (Fig. 3). Mortality estimates vary, with maximum losses over 90% of the entire population (Harris, 1997). The response to disease included dispersal away from afflicted settlements to small coastal islands or inland villages. Following the 1862 smallpox outbreak, many ‘‘retreated to their main villages, leaving settlements deserted along. . .hundreds of miles of coastline (along southeastern Vancouver Island)’’ (Mackie, 1995). Similar responses occurred to warfare. Although there was always inter-tribal conflict, warfare appears to have intensified as early as the 1790s in response to the increased availability of firearms and competition for trade (Archer, 1978; Marshall, 1999; Keddie, 2003). During conflict, European observers reported that ‘‘little hunting can be expected’’ (MacLachlan, 1998), and it also caused the coalescence of populations from peripheral settlements to the main inland villages (Pellatt et al., 2007). Again, it is difficult to fully extrapolate the effects of disease and conflict among hunters on ungulates prior to the onset of European settlement. However, the implication is that hunting effort was greatly reduced (Martin and Szuter, 1999). It is unclear whether these descriptions from Vancouver Island reflect general trends elsewhere in the northwest. Descriptions of abundance partially reflect the circumstances

B I O L O G I C A L C O N S E RVAT I O N

of observation. Lewis and Clark, for example, often travelled rapidly and during the summer, when western ungulate herds, especially elk, tend to migrate to upland areas (Burroughs, 1961). This may have contributed to their infrequent encounters with ungulates in some regions. The impressions of settlers on Vancouver Island, in contrast, were based on continuous observation. There seems little doubt that some regions of the northwest supported relatively sparse herds of deer and elk, mostly between the continental divide and the lower reaches of the Columbia and Fraser Rivers (e.g., Burroughs, 1961; Spry, 1995; Lamb, 2007). The coastal regions, however, did not appear to be a system with sporadic ungulate populations. The relative unavailability of widespread and high-quality forage compared to the Great Plains of North America, for example, must be limiting at some level. Hunting pressure, however, appears to provide a more powerful explanation for varying ungulate densities. Accounts of predator abundance in these areas also tentatively support the assertion that herd sizes were formerly large. Cougars and wolves on Vancouver Island feed primarily on black-tailed deer (MacTaggart-Cowan, 1945; Martin and Baltzinger, 2002), and both predators were reported as abundant on Vancouver Island in the mid-1800s. These predators ‘‘terrorized settlers’’, limited or prevented livestock production, and persisted near the settlements until the 1930s despite aggressive eradication efforts that included bounties and the poisoning of deer carcasses with strychnine (MacFie, 1972; Anonymous, 1986; Mackie, 1995; Arnett, 2007). Given the current strong inter-relationship between the abundances of ungulates and predators on Vancouver Island (Hatter and Janz, 1994), it seems likely that predator culls eventually contributed to the increase in deer numbers that began in the early 1900s. It is impossible to directly determine the impact of browser abundance on the pre-colonization vegetation community. One possible exception, however, may be the repeated description of ‘‘fern prairies’’ (Pteridium aquilinum) across southeastern Vancouver Island (Richardson, 1872; Anonymous, 1986; Kane, 1971; MacFie, 1972; Douglas, 1979; Brown, 1989). Fire and human effects could both explain this. P. aquilinum expands aggressively with repeated burning, and its rhizomes were a primary food staple of the indigenous peoples; trenches dug to gather the rhizomes promoted fern recruitment in subsequent years (White, 1999). However, such ‘‘fern parks’’ have also been interpreted as a signature of selective browsing by deer (Rooney, 2001). Today, P. aquilinum is relatively rare in remnant sites, possibly because it was aggressively eradicated by early settlers (MacFie, 1972; Mackie, 1995). However, recent evidence indicates that other unpalatable species, especially non-native annual grasses such as Bromus spp., are dominating some sites because of selective browsing by deer (Gonzalez, 2007).

4.

Post-settlement herbivore declines

Initial reports suggested that ungulates were abundant; subsequent reports indicate widespread eradication beginning soon after colonization (Fig. 3). In the first two decades following European settlement, there was a heavy reliance on subsistence hunting (Mackie, 1995). Livestock was expensive, in short supply (Douglas, 1979), and sometimes killed by cougars

1 4 1 ( 2 0 0 8 ) 2 1 7 4 –2 1 8 3

2179

and wolves. In the Cowichan Valley, the pace of land clearance was slow, especially in non-savanna areas dominated by Pseudotsuga menziesii, meaning that crop production was initially limited. Most colonists hunted, including some who earned their living by selling meat and hides. Hunting methods by the indigenous populations also changed, including the replacement of the bow-and-arrow, spears, and muskets with rifles (Moon, 1978). They also began hunting more intensively, especially for hides that became a common form of barter. Reports indicate that ‘‘tons of deer hides were being traded’’ in the early 1860s. In some cases, there was ‘‘wanton killing’’ with the carcases left once the hides were removed (Mackie, 1995). Exacerbating the hunting pressure was the settlement patterns of the colonists. Unlike the Hul’q’umi’num, Europeans settled throughout the lowland areas, not just by the rivers, and lived there year-round. In the Cowichan Valley, the focus in particular was on settling the open prairie and savanna, which reminded settlers of home (e.g., pastoral England) and was easier to clear (Anonymous, 1986; Arnett, 1999). Permanent occupation of the coastal lowlands presumably meant more intensified hunting year-round, which may have been especially hard on ungulate numbers if reports of their importance as wintering areas are correct (Brown, 1989). Reports of ungulate scarcity began appearing in the Cowichan and Comox Valleys by the 1860s; presumably this occurred sooner around Victoria. The scarcity was attributed to over-hunting and habitat loss (Mackie, 1995). A Game Act was enacted by 1884 in an attempt to limit hunting pressure, although enforcement did not begin until later. The aggressive eradication of predators also began at this time, in part to reduce competition for the dwindling numbers of deer. However, accounts imply this was insufficient initially to prevent the extirpation of elk, and the drastic reduction of deer. A transition in lifestyle patterns occurred towards the 1890s, which appeared to contribute to recovery of deer. Livestock numbers and agricultural production (hay, butter, and cash crops) had increased following two or more decades of forest clearance and the planting of pastures with non-native perennial grasses. Presumably this reduced dependency on hunting for food and hides. This period also saw the emergence of a sports hunting and angling industry in the Cowichan Valley (Anonymous, 1986; Mackie, 1995; Harris, 2001), marking a sharp contrast from the 1860s to 1870s where subsistence, not recreation, was the focus of hunting. Conservation efforts also intensified, in part to protect the local sporting interests. This included prosecution of Hul’q’umi’num hunters and fishers as early as 1894, as tight restrictions were placed on when deer and other game (e.g., grouse, salmon) could be taken (Harris, 2001). By this time, the barter economy that encouraged the harvesting of hides by local indigenous populations was largely replaced by a more conventional industry-based economy, within which many Hul’q’umi’num people became employed (e.g., forestry, fish canneries, agriculture (e.g., hops production for beer)).

5.

Oak stand structure data

The historical data suggest periods of fluctuating ungulate abundance, with relatively high numbers at first contact, sharp declines in the decades immediately following coloni-

2180

B I O L O G I C A L C O N S E RVAT I O N

zation, and then gradual but substantial increases starting as early as the 1890s. Stand structure data within remnant oak woodlands appear to support this model. As with oak species across North America, Q. garryana is not recruiting in large numbers. Hypotheses for oak recruitment failure are diverse, including increased herbivore abundance, fire suppression, climate change, seed predators, and weakened ectomychorrhizal associations (e.g., Abrams, 2003). It is often implied that a temporal dichotomy exists between an earlier period when oaks recruited with relative ease vs. currently, when they no longer can due to one or a combination of these factors. Stand data from southeastern Vancouver Island suggest that this dichotomy between pre- and post-colonization recruitment is an over-simplification (Gedalof et al., 2006; Pellatt et al., 2007). Instead, mass oak recruitment at many sites began in the mid-1800s in association with the onset of European colonization. Prior to this period, savanna areas were far more open (Grant, 1849; Wells, 1858; Gedalof et al., 2006). After this, oaks once again show limited recruitment. On the Cowichan Garry Oak reserve (Fig. 2), few of the present-day oaks recruited before 1860 or after 1910 (Pellatt et al., 2007). This property was unoccupied by European settlers until the mid- to late-1860s. Since 1910, oak canopy cover has gradually increased (Fig. 2) but, despite high acorn production and large densities of seedlings <10 cm, almost no individuals have reached the sapling stage. These results indicate that the oak woodlands of today appear to be an artefact of land use patterns in the mid- to late-1800s, and that woodlands (i.e., oak canopy cover >75%) were not a substantial component of the pre-Columbian landscape. There has been widespread tree recruitment since the 1900s, but mostly by Douglas-fir (Pseudotsuga menziesii) that has in-filled many remnant sites (Gedalof et al., 2006; Devine et al., 2007). The coincidence between mass oak recruitment and the decrease in ungulate populations following European settlement is consistent with a hypothesis that hunting pressure influences oak abundance. Black-tailed deer can heavily browse oak saplings, and they consume large quantities of acorns in the late summer and fall, including Q. garryana (Sullivan et al., 1985; Garrison, 1992; Devine et al., 2007). Similar effects have been observed in eastern North America, where intense deer browsing is associated with substantial changes to the vegetation community, including decreased recruitment for several oak species and other trees (hemlock, eastern white-cedar) and loss of ground flora diversity (Anderson and Katz, 1993; Horsley et al., 2003; Rossell et al., 2007). In some eastern regions, browsing by white-tailed deer is projected to determine the composition and diversity of forests in the near future (Rossell et al., 2007). The historical evidence does not as easily accommodate other leading hypotheses (fire, invasives, episodic recruitment), although they likely interact with herbivory (Devine et al., 2007). Fire is ubiquitous with oak occurrences (Abrams, 2003), and this system formerly burned with greater frequency (MacDougall et al., 2004; Pellatt et al., 2007). Increased forage quality following fire could stimulate ungulate reproductive success, thereby supporting larger herd sizes than would occur in the absence of fire. However, changes to the fire regime do not match the oak recruitment patterns de-

1 4 1 ( 2 0 0 8 ) 2 1 7 4 –2 1 8 3

scribed above. Charcoal cores in local lakes reveal relatively consistent inputs until after the 1880s, with declines associated with fire suppression (Pellatt et al., 2007). Prior to this, indigenous peoples used fire for a range of purposes that could have included forage improvements for ungulates, although this is not directly reported (Suttles, 1987). Europeans also relied heavily on fire during the initial stages of land clearance (Grant, 1859; MacFie, 1972). Today, the abundance of small oak seedlings in grass swards, unburned for decades, suggests that fire is not critical for seedling establishment, although these swards may inhibit development at later life history stages (Callaway, 1992; Devine et al., 2007). The more critical former role of fire may have been the suppression of infilling by P. menziesii, which eventually leads to the displacement of all savanna plants including oaks (MacDougall, 2005; Pellatt et al., 2007). Grass invasion is also associated with oak recruitment failure in parts of western North America, due to its effects on elevated fire intensity and resource limitation (D’Antonio and Vitousek, 1992; Gordon and Rice, 2000). On southeastern Vancouver Island, exotic perennial grasses dominate most sites, due in part to their aggressive introduction beginning in the 1840s (MacDougall et al., 2004). Requisition requests for seed by settlers and government officials are common in the early historical records, and seeds were repeatedly introduced in massive quantities in attempt to improve forage levels (MacFie, 1972; Arnett, 1999). As with fire declines, however, the onset of grass invasion in the Cowichan Valley generally coincides with mass oak recruitment, rather than the reverse, suggesting it has not been an absolute barrier to oak recruitment. This is tentatively supported by experimental work that shows that native grasses, which presumably were once more abundant, are functionally very similar to exotic perennial grasses (MacDougall and Turkington, 2004). In the absence of burning, both reach high abundances with canopy heights sometimes exceeding 2 m. At present, no oak seedlings on the Cowichan Reserve occur above the grass canopy (MacDougall, unpublished data), suggesting that the grasses may facilitate survival by obscuring them from browsers. Finally, as with all masting species, acorn production by Q. garryana is episodic (Koenig and Knops, 2005). So may be recruitment, when the combination of high acorn availability and optimal climate could lead to mass establishment. If true, decades of failed recruitment may not be problematic for oak persistence as long as there are occasional bursts of establishment, as occurred in the late-1800s. The historical data cannot comment on climate conditions during this period, although winters were clearly colder given the regular descriptions of frozen lakes that now mostly remain open year-round. The regional sub-mediterranean climate of western North America, however, is characterized by high interannual variability in the timing and intensity of rainfall, drought, and temperature averages, so it is conceivable that recruitment windows rarely open. Also, oak establishment is known to be sensitive to climate in this system. Newly emerged seedlings are susceptible to drought in their 2nd and 3rd years (Gonzalez, 2007), while winter precipitation strongly influences oak production in the subsequent growing season (Pellatt et al., 2007). However, these same climate cues also positively affect recruitment by P. menziesii (Thysell and Carey, 2001) and annual productivity

B I O L O G I C A L C O N S E RVAT I O N

by grasses (MacDougall, unpublished data). As well, climate limitations are normally assumed to limit seedling establishment, but as described, establishment is presently occurring in large numbers at some sites. An alternative scenario is that the mass recruitment by oaks in the late 1880s derives from one or more masting events coinciding with the reduction in browser numbers. Masting can facilitate recruitment success by satiating seed predators (Koenig and Knops, 2005) but if these predators are absent this could result in unusually high recruitment.

6.

Implications for conservation

Based on this review, there is strong evidence that human hunting pressure has influenced herbivore populations in coastal savannas of southeastern Vancouver Island for some time. Ungulate populations appear to have varied widely over the past 150 years, in association with human disease epidemics, shifts in human land use patterns, and the waning importance of subsistence hunting (Martin and Baltzinger, 2002). This result appears to be generalizable to other regions of North America. Similar long-term fluctuations in deer numbers were reported in the Willamette Valley of Oregon, where descriptions of ungulate numbers switched from highly abundant in the 1840s to sparse by the 1890s (Bailey, 1936). In eastern North America, white-tailed deer numbers were reported to have sharply declined following initial European colonization, and then subsequently increased as conservation measures were enforced and hunter numbers slowly decreased (Rooney, 2001). Evidence from this study suggests that such fluctuations in herbivory can have long-lasting effects on the vegetation community. The mass recruitment by oaks as deer numbers fell in the late-1800s echoes to the present, as extant woodlands mostly derive from this period. Today, it appears that native ground plants are being affected to similar degrees (Gonzalez, 2007). Many forbs have suffered large population declines, some of which can be directly attributed to herbivory (Gonzalez, 2007). In some grazed sites, flowering does not occur except in soil pockets on cliff faces that are inaccessible to browsers, and unsheltered plants appear to be annually clipped almost to ground level. Those that persist are mostly perennial bulb species that regenerate by vegetative spread, although the demographic consequences of longterm flowering failure are presumably negative. Browsing effects in these areas is further magnified by the presence of feral sheep and goats. There are also indirect effects. The avoidance of untasty exotic plants by deer, such as annual Bromus grasses, appears to be hastening plant invasion at some sites (Gonzalez, 2007). Given that these exotic grasses are also aggressive competitors with native plants (Dyer and Rice, 1999), there may be feedbacks between browsing avoidance and resource sequestration that hastens their dominance. The negative effects of heavy browsing may also extend to Lepidoptera, which are diverse in this system and rely on many of the food plants that deer also consume. These observations suggest that, despite the previous occurrence of large ungulate herds in this region, herbivore populations have likely never been larger, nor their effects more intense. Several factors likely combine to create this

1 4 1 ( 2 0 0 8 ) 2 1 7 4 –2 1 8 3

2181

historically unprecedented situation: predator loss, hunting reductions, and the isolation of remnant savanna within areas converted from mature forest to pasture or young forests with high browse availability. The extirpation of elk may also have an impact. Elk generally graze grass, while deer browse forbs (Johnson and Cushman, 2007). Remnant sites are typically dominated by grass, which was thought to be mostly an effect of fire suppression (MacDougall, 2005), but may also reflect elk extirpation. In situations such as this, where ungulate numbers are high and impacts on plant are substantial, management solutions are increasingly focusing on some form of ungulate control (e.g., fencing, culling). Culling in particular is increasingly being practiced in parts of North America, and elsewhere, to facilitate recruitment by tree and understory plant species. On southeastern Vancouver Island, it seems conceivable that herd reductions will be required in some places to reduce browsing impacts. Such efforts do not seem inconsistent with the regulation of ungulate densities that has persisted for centuries in this system. However, the results of this historical analysis and results elsewhere do raise questions about the impacts that fluctuating herbivore abundances can have on the savanna plant community. Deer are rarely the sole contributors to vegetation change in overgrazed systems (Russell et al., 2001; Abrams, 2003), and this appears to be true on southeastern Vancouver Island. Burning in some deep soil areas, for example, has lead to significant increases in growth and fitness of forbs by reducing grass cover, despite the presence of deer (MacDougall, 2005). There appears to be a threshold effect for deer density on forbs, with recruitment failure and species loss most evident at the highest densities (Gonzalez, 2007). It is also possible that deer browsing is having undetected positive effects, such as the slowing of tree encroachment (Johnson and Cushman, 2007). Removing herbivores, therefore, could lead to different sets of management problems, especially in systems as this where ungulates have long been present and where other substantial transformations are underway (fire suppression, habitat loss). Until these issues are untangled, it is unclear if the benefits of deer reductions will be offset by other unanticipated effects, as witnessed during the mass recruitment of oaks that followed ungulate eradication in the late-1800s.

Acknowledgements This manuscript benefited from discussions with Emily Gonzalez, Peter Arcese, Tom Nudds, Irvin Banman, Collin Elder, Tim Ennis, Ze’ev Gedalof, Mark Vellend and two anonymous reviewers.

R E F E R E N C E S

Abrams, M.D., 2003. Where has all the white oak gone? Bioscience 53, 927–939. Anderson, R.C., Katz, A.J., 1993. Recovery of browse-sensitive tree species following release from white-tailed deer Odocoileus virginianus Zimmerman browsing pressure. Biol. Conserv. 63, 203–208.

2182

B I O L O G I C A L C O N S E RVAT I O N

Anonymous, 1986. Memories Never Lost: Stories of the Pioneer Women of the Cowichan Valley and a Brief History of the Valley 1850–1920. Frieson and Sons, Altona, Manitoba. Archer, C.I., 1978. Spanish exploration and settlement of the northwest coast in the 18th century. In: Langlois, W.J. (Ed.), Captain Cook and the Spanish Explorers on the Coast. Province of British Columbia, Victoria, BC, pp. 33–53. Arnett, C., 1999. The Terror of the Coast: Land Alienation and Colonial War on Vancouver Island and the Gulf Islands, 1849– 1963. Talon Books, Vancouver, BC. Arnett, C., 2007. Two Houses Half-Buried in Sand: Oral Traditions of the Hul’q’umi’num’ Coast Salish of Kuper Island and Vancouver Island. Talon Books, Vancouver, BC. Bailey, V., 1936. The Mammals and Life Zones of Oregon. North American Fauna, No. 55, United States Department of Agriculture, Bureau of Biological Survey, Washington D.C. Barbour, M.G., Major, J., 1977. Terrestrial Vegetation of California. John Wiley and Sons, New York. Brown, R., 1989. Journal of the Vancouver Island exploring expedition 1864. In: Hayman, J. (Ed.), Robert Brown and the Vancouver Island Exploring Expedition. University of British Columbia Press, Vancouver, BC, pp. 37–155. Burroughs, R.D., 1961. The Natural History of the Lewis and Clark Expedition. Michigan State University press, East Lansing. Callaway, R.M., 1992. Effect of shrubs on recruitment of Quercus douglasii and Quercus lobata in California. Ecology 73, 2118–2128. Cote, S.D., Rooney, T.P., Trembley, J.-P., Dussault, C., Waller, D.M., 2004. Ecological impacts of deer overabundance. Annu. Rev. Ecol. Evol. Syst. 35, 113–147. Curtis, E.S., 1913. The North American Indian. The Salishan Tribes of the Coast, vol. 9. Plimpton Press, Norwood, MA. D’Antonio, C.M., Vitousek, P.M., 1992. Biological invasions by exotic grasses, the grass/fire cycle, and global change. Annu. Rev. Ecol. Syst. 23, 63–87. Daubenmire, R., 1985. The western limits of the American bison. Ecology 66, 622–624. Devine, W.D., Harrington, C.A., Leonard, L.P., 2007. Post-planting treatments increase growth of Oregon white oak (Quercus garryana Doug. Ex Hook) seedlings. Restor. Ecol. 15, 212–222. Douglas, J., 1979. Selected letters. In: Bowsfield, H. (Ed.), Fort Victoria Letters 1846–1851. Hudson’s Bay Record Society, Winnipeg, Manitoba. Dunn, P.V., Ewing, K., 1997. Ecology and Conservation of South Puget Sound Prairie Landscape. The Nature Conservancy, Seattle. Dyer, A.R., Rice, K.J., 1999. Effects of competition on resource availability and growth of a California bunchgrass. Ecology 80, 2697–2710. Egan, D., Howell, E.A., 2001. The Historical Ecology Handbook: A Restorationist’s Guide to Reference Ecosystems. Island Press, Washington, DC. Erickson, W.R., Meidinger, D. 2007. Garry Oak (Quercus garryana) Plant Communities in British Columbia: A Guide to Identification. British Columbia Ministry of Forests and Range Technical Report 040. Victoria, BC. Fitzgerald, J.E., 1848. Vancouver’s Island: The New Colony. University of British Columbia Archives, Vancouver. Fleischner, T.L., 1994. Ecological costs of livestock grazing in western North America. Conserv. Biol. 8, 629–644. Foster, D.R., Motzkin, G., 2003. Interpreting and conserving the openland habitats of coastal New England: insights from landscape history. Forest Ecol. Manag. 185, 127–150. Garrison, B.A., 1992. California Oaks and Deer. Integrated Hardwood Range Management Program. University of California, Berkeley. Gedalof, Z., Smith, D.J., Pellatt, M.G., 2006. From prairie to forest: three centuries of environmental change at Rocky Point, Vancouver Island, BC. Northwest Sci. 80, 34–46.

1 4 1 ( 2 0 0 8 ) 2 1 7 4 –2 1 8 3

Gonzalez, E.K. 2007. The Effects of Herbivory, Competition, and Disturbance on Island Meadows. Ph.D. Thesis, Department of Forestry, University of British Columbia. Gordon, D.R., Rice, K.J., 2000. Competitive suppression of Quercus douglasii (Fagaceae) seedling emergence and growth. Am. J. Bot. 87, 986–994. Gough, B.M., 1992. The journal of Alexander Henry the Younger, 1799–1814. Champlain Society, Toronto. Grant, W.C., 1849. Report on Vancouver’s Island by Captain W.C. Grant to Sir James Douglas. British Columbia Archives, Victoria, BC. Grant, W.C., 1859. Description of Vancouver Island. J. R. Geogr. Soc. 27, 268–320. Harris, C., 1997. The Resettlement of British Columbia: Essays on Colonialism and Geographical Change. University of British Columbia Press, Vancouver, BC. Harris, D., 2001. Fish, Law, and Colonialism: The Legal Capture of Salmon in British Columbia. University of Toronto Press, Toronto. Hatter, I.W., Janz, D.W., 1994. Apparent demographic changes in black-tailed deer associated with wolf control on northern Vancouver Island. Can. J. Zoolog. 72, 878–884. Hickman, J.C., 1993. The Jepson manual: higher plants of California. University of California, Berkeley. Horsley, S.B., Stout, S.L., deCalesta, O.S., 2003. White-tailed deer impact on the vegetation dynamics of a northern hardwood forest. Ecol. Appl. 13, 98–118. Johnson, B.E., Cushman, J.H., 2007. Influence of a large herbivore reintroduction on plant invasions and community composition in a California grassland. Conserv. Biol. 21, 515– 526. Kane, P., 1971. Wanderings of an Artist Among the Indians of North America – 1858. M.G. Hurtig, Edmonton, Alberta, Canada. Keddie, G., 2003. Songhees Pictorial: A History of the Songhees People as Seen by Outsiders 1790–1912. Royal British Columbia Museum, Victoria, BC. Koenig, W.D., Knops, J.M.H., 2005. The mystery of masting trees. Am. Sci. 93, 340–347. Lamb, W.K., 1984. A Voyage of Discovery in the North Pacific Ocean and Round the World, 1791–1795 (George Vancouver). Hakluyt Society, London. Lamb, W.K., 2007. The Letters and Journals of Simon Fraser, 1806– 1808. Dundurn Press, Toronto, Ontario. Lea, T., 2006. Historical Garry oak ecosystems of Vancouver Island, British Columbia: pre-European contact to present. Davidsonia 17, 34–50. Ludrin, N., 1928. The pioneer women of Vancouver Island 1843– 1866. The Women’s Canadian Club of Victoria, Victoria, BC. Lyman, R.L., Wolverton, S., 2002. The late prehistoric-early historic game sink in the northwestern United States. Conserv. Biol. 16, 73–85. MacDougall, A.S., 2005. Response of diversity and invasibility to burning in a northern oak savanna. Ecology 86, 3354– 3363. MacDougall, A.S., Turkington, R., 2004. Relative importance of suppression-based and tolerance-based competition in an invaded oak savanna. J. Ecol. 92, 422–434. MacDougall, A.S., Beckwith, B., Maslovat, C., 2004. Defining conservation strategies with historical perspectives: a case study from a degraded oak ecosystem. Conserv. Biol. 18, 455– 465. MacDougall, A.S., Boucher, J., Turkington, R., Bradfield, G.E., 2006. Plant invasion along an environmental stress gradient. J. Veg. Sci. 17, 47–56. MacFie, M., 1972. Vancouver Island and British Colombia: their History, Resources, and Prospects – 1865. Coles Publishing, Toronto.

B I O L O G I C A L C O N S E RVAT I O N

Mack, R.N., Thompson, J.N., 1982. Evolution in steppe with few large hooved mammals. Am. Nat. 119, 757–773. Mackie, R.S., 1995. The Wilderness Profound: Victorian Life on the Gulf of Georgia. Sono Nis Press, Victoria, BC, Canada. MacLachlan, M., 1998. The Fort Langley Journals 1828–1830. University of British Columbia Press, Vancouver, BC. MacTaggart-Cowan, I., 1945. The ecological relationships of the food of the Columbian black-tailed deer (Odocoileus hemionus columbianus (Richardson), in the coast forest region of southern Vancouver Island, British Columbia. Ecol. Monogr. 15, 109–139. Maret, M.P., Wilson, M.V., 2005. Fire and litter effects on seedling establishment in western Oregon upland prairies. Restor. Ecol. 13, 562–568. Maron, J.L., Estes, J.A., Croll, D.A., Danner, E.M., Elmendorf, S.C., Buckalew, S., 2006. An introduced predator transforms Aleutian Island plant communities by disrupting spatial subsidies. Ecol. Monogr. 76, 3–24. Marshall, D.P., 1999. Those Who Fell from the Sky. Cultural and Educational Center, Cowichan Tribes, Duncan, BC. Martin, J.-L., Baltzinger, C., 2002. Interaction among deer browsing, hunting, and tree regeneration. Can. J. Forest Res. 32, 1254–1264. Martin, P.S., Szuter, C.R., 1999. War zones and game sinks in Lewis and Clark’s West. Conserv. Biol. 13, 36–45. McCullough, D.R., Jennings, K.W., Gates, N.B., Elliott, B.G., DiDonato, J.E., 1997. Overabundant deer populations in California. Wildlife Soc. B 25, 478–483. Menzies, A., 1923. Menzies’ Journal of Vancouver’s Voyage, April to October 1792. University of British Columbia Archives, Vancouver. Moon, B.J., 1978. Vanished companions: the changing relationship of the West Coast people to the animal world. In: Langlois, W.J. (Ed.), Captain Cook and the Spanish Explorers on the Coast. Province of British Columbia, Victoria, BC, pp. 71–77. Parker, J.D., Burkepile, D.E., Hay, M.E., 2006. Opposing effects of native and exotic herbivores on plant invasions. Science 311, 1459–1461. Pellatt, M.G., Gedalof, Z., McCoy, M., Bodtker, K., Cannon, A., Smith, S., Beckwith, B., Mathewes, R., Smith, D., 2007. Fire

1 4 1 ( 2 0 0 8 ) 2 1 7 4 –2 1 8 3

2183

history and ecology of Garry Oak and associated ecosystems in British Columbia. Interdepartmental Recovery Fund Project 733. Parks Canada, Vancouver. Richardson, J., 1872. Report on the coal fields of the east coast of Vancouver Island. Reports of Explorations and Surveys 1871– 1872. Geological Survey of Canada, Montreal. Rollins, P., 1995. The Discovery of the Oregon Trail: Robert Stuart’s Narratives of his Overland Trip Eastward from Astoria 1812– 1813. University of Nebraska Press, Lincoln. Rooney, T.P., 2001. Deer impacts on forest ecosystems: a North American perspective. Forestry 74, 201–208. Rooney, T.P., Waller, D.M., 2003. Direct and indirect effects of deer in forest ecosystems. Forest Ecol. Manag. 181, 165–176. Rossell, C.R., Patch, S., Salmons, S., 2007. Effects of deer browsing on native and non-native vegetation in a mixed oak-beech forest on the Atlantic Coastal Plain. Northeast Nat. 14, 61–72. Russell, F.L., Zippin, D.B., Fowler, N.L., 2001. Effects of white-tailed deer (Odocoileus virginianus) on plants, plant populations, and communities: a review. Am. Midl. Nat. 146, 1–26. Schultz, J.W., 1980. Blackfeet and the Buffalo. University of Oklahoma Press, Norman, OK. Spry, I.M., 1995. The Palliser Expedition. Fifth House, Calgary, Alberta. Sullivan, T.P., Nordstrom, L.O., Sullivan, D.S., 1985. Use of predator odors as repellents to reduce feeding damage by herbivores: 2. Black-tailed deer (Odocoileus hemionus columbianus). J. Chem. Ecol. 11, 921–936. Suttles, W., 1987. Coast Salish Essays. University of Washington Press, Seattle. Thysell, D.R., Carey, A.B., 2001. Manipulation of density of Pseudotsuga menziesii canopies: preliminary effects on understory vegetation. Can. J. Forest Res. 31, 1513–1525. Van Vuren, D., 1987. Bison west of the Rocky Mountains: an alternative explanation. Northwest Sci. 61, 65–69. Wells, O., 1858. Field Notes of the Cowichan Surveys. SurveyorsGeneral Branch, Victoria, BC, Canada. White, R., 1999. Indian land use and environmental change, Island County Washington: a case study. In: Boyd, R. (Ed.), Indians, Fire, and the Land in the Pacific Northwest. Oregon State University Press, Corvallis, pp. 36–49.