General context from biodiversity to ecosystem restoration

Restoration ecology: definition, aims and semantics

Restoration ecology is the science that develops and tests a body of theory focused on repairing damaged ecosystems (Palmer et al., 1997) and is therefore closely linked to ecological restoration which, according to the Society for Ecological Restoration, is the process of assisting the recovery of an ecosystem that has been degraded, damaged, or destroyed (Society for Ecological Restoration International Science and Working Policy Group, 2004). The distinction between certain terms may require definition, mainly those concerning differences in restoration objectives. Restoration sensu stricto is the re-establishment of all attributes of the reference ecosystem, including species-richness, composition, structure and function; whereas rehabilitation objectives focus on the re-establishment of some functions or services or partial re-establishment of ecosystem attributes (Figure I.2). Many projects have sensu stricto restoration objectives but in view of their actual results they should perhaps be reclassified as rehabilitation projects (Hobbs, 2007). Both restoration sensu stricto and rehabilitation focus on a historical pre-existing reference ecosystem, while reclamation can be used in the context of industrial or mine degraded lands for the purposes of terrain stabilization, public health and safety, or landscape improvement (e.g. the re-establishment of plant cover in mining sites in order to prevent toxic dust dispersion) (Society for Ecological Restoration International Science and Working Policy Group, 2004). Reclamation can also be used as a synonym for reallocation (Aronson et al., 1993; Muller et al., 1998) with a target ecosystem as an objective, but in contrast to rehabilitation or restoration, the target is an ecosystem which has been chosen for various reasons, such as improving biodiversity or providing ecosystem services (e.g. creation of wetlands in the context of mitigation banks in the United States) (Figure I.2).

Reference ecosystem

All the different terms used to describe restoration sensu lato are linked to the definition of restoration objectives. The vision of how an impaired ecosystem should be after restoration is called the reference (Clewell and Aronson, 2007). The choices of restoration objectives are unavoidably subjective (Choi, 2004), are always a trade-off between different objectives (Bullock et al., 2011) and are determined by historical considerations, ecological values, social acceptance and economic feasibility. Historical considerations can differ geographically: in the New World or in Australia, the historical reference is usually the ecosystem before European settlement (Swetnam et al., 1999) whereas in Europe the historical reference is usually the state before an anthropogenic severe disturbance (e.g. intensive cultivation). Ecological values include species or habitat with conservation values, biodiversity or potential habitat for rare and/or endemic and/or threatened species. A restoration project needs social acceptance to be successful, a counter-example is the restoration of the wetlands on which Chicago was built that would clearly not prove acceptable (Choi, 2007). Economic feasibility is an important limiting factor which will determine the scale and intensity of active restoration taking into account current technical knowledge (Cairns, 2000).

The reference can be defined as an actual area or as a written description (Society for Ecological Restoration International Science and Working Policy Group, 2004). Several authors have criticized a too narrow view of the reference, which could be either unsuited to the current environmental conditions (i.e. in view of global change) or an unattainable goal (Pickett and Parker, 1994; Hobbs and Norton, 1996). On the other hand, the reference ecosystem is viewed as a valuable tool as a basis for setting objectives, identifying restoration needs and assessing restoration success (Aronson et al. (1993); Clewell and Aronson (2007); Giardina et al. (2007) and Chapter 3). More than a simple static state, the reference should reflect the range of variability potentially illustrated by spatial or temporal variation of the natural ecosystem (Society for Ecological Restoration International Science and Working Policy Group, 2004; Hilderbrand et al., 2005). In order to express the natural dynamics, the goal could be a reference trajectory (Aronson et al., 1993), and the restored ecosystem should therefore show the same resilience to common variability under environmental conditions (Figure I.2). The notion of target species is directly linked to the concept of reference ecosystem. These species are the species present in the reference and are usually contrasted with non-target species which are species absent from the reference ecosystem. Reducing the number of non-target species can be an objective but just like the native/non-native distinction, it should be used with caution (Davis et al., 2011). It should be stressed that even target species can become competitive and their over-abundance can represent a threat to successful restoration. For this reason in Chapter 3, the indices developed will distinguish target and non-target abundances rather than species.

Succession

The simplest way of defining succession is species change over time, or turnover (Walker and Moral, 2003). Although it is tempting to define it as evolution, as a synonym of development, this should be avoided in order to avoid confusion with the Darwinian sense of evolution. In the present definition of succession, there is also a matter of scale: historical reconstructions of very long-term changes (i.e. paleoecological studies), and temporal variability around a relatively stable state (i.e. the carousel model of Maarel et Sykes (1993)) are two kinds of vegetation changes which are generally not included in the definition of succession (Walker and Moral, 2003). Among the numerous ways of describing succession, a widely used dichotomy is the primary/secondary distinction, in which should both be considered as endpoints of a continuum rather than a clear-cut distinction (White and Jentsch, 2001). This discriminates successions through their characteristics at the beginning. Primary succession occurs on sterile substrate with low nutrient content (e.g. lava or glacial moraine) (Walker and Moral, 2003) whereas secondary succession occurs on substrate were biotic content is not null: a seed bank or soil biota may remain, and soil nutrient content is not a limiting factor (e.g. vegetation recovery after cultivation abandonment or forest regeneration after a hurricane) (Cramer et al., 2007).

One of the first succession models is the relay floristic model of Clements (1916) in which early-successional species establish at the beginning, then perish as the late-successional species establish and persist to form the succession endpoint. Another model is the initial floristic model (Egler, 1954) where all, both early- and late-successional species, are present at the beginning of succession. In the early stages, early-successional species dominate, and late-successional species remain at very low abundance. As time goes by, early-successional species decline and late-successional species tend to dominate. A very influential explanation of succession was the three models of succession developed by Connell et Slatyer (1977): the facilitation, tolerance and inhibition models. The facilitation model is when early-successional species establish and alter the environment in such a way that it becomes more suitable for late-successional species. The tolerance model is when early-successional species establish and alter the environment in such a way that it excludes other species except those that can tolerate the competition of early-successional species, which then, with time, end up dominating. The last model is the inhibition model, when early-successional species establish and alter their environment in such a way that it prevents any other species from establishing; late-successional species will establish only when early-successional species die. As a disturbance is basically at the origin of almost all successions, it is essential to the succession concept.

Disturbance Many authors have given definitions of disturbances, two of which are widely used: the definition by Pickett and White (1985): a relatively discrete event in time that alters the structure of a population, community or ecosystem; and the definition by Grime (1977): a constraint that limits the plant biomass by causing its destruction. The disturbance is usually opposed to stress which is a limitation of the production of biomass (Grime, 1977). The distinction between stress and disturbance is not always straightforward, as the same event can be considered as a stress or a disturbance, depending on the organism of interest. For a given organism, the event can be considered as stress until organism tolerance is exceeded, when the tolerance is exceeded, this leads to its death or a significant loss of biomass and is therefore considered as a disturbance (Sousa, 1984). A disturbance has to be characterized in order to assess its effect on a population, community or ecosystem. Sousa, (1984) defined several attributes of disturbance: the extent, i.e. the size of the disturbed area; the magnitude, including intensity: i.e. the strength of the disturbing force, and severity: i.e. the damage caused by the disturbance; the frequency, i.e. the number of disturbances per period of time and the predictability, i.e. the variance of the mean time between disturbances. Two distinctions are made to characterize disturbances: i) disturbance may be either exogenous: i.e. if the disturbance event originates outside of the system (e.g. avalanche, storm) or the disturbance can be considered as endogenous when the disturbance event originates inside the system (e.g. a senescent tree fall). Sometimes the cut-off line between the two is difficult to determine: for instance a fire can be considered exogenous when its origin is a thunderstorm, but fire intensity can be increased by organic compounds released by the vegetation and therefore the fire can be considered as endogenous for ecosystems such as matorral. ii) natural or anthropogenic disturbance is also a common dichotomy which distinguishes man-made disturbance from other disturbances, even if a natural disturbance can have the same effects as an anthropogenic disturbance (e.g. large herbivore herds and traditional itinerant mixed domestic grazing).