Using process-based indicator species to evaluate ecological corridors in fragmented landscapes

Highlights

• We provide a framework to define three types of indicator species that can evaluate the functionality of ecological corridors

• The efficiency of the corridor increases with the quality of hedgerows, and decreases with increasing land use intensity in adjacent lands

• The method is easily applicable and provides a robust baseline assessment of species richness, and composition

Abstract

Increasing connectivity among habitat patches is assumed to mitigate the effects of fragmentation on biodiversity, leading to the emergence of green and blue infrastructures in public policies as corridors facilitating movements of species within fragmented landscapes. But still, the scientific knowledge and tools for identifying critical features that make a corridor efficient for biodiversity conservation, and for guiding their establishment and management are still scarce. Here, we define three types of indicator species based on their ecological requirements and dispersal traits in the context of forest metacommunities embedded into agricultural landscapes in N France. We then evaluate whether hedgerows can act as corridors for forest herb species and if this can be predicted from the occurrence of the selected indicator species. We show that among each socio-ecological group of forest herb species, (i) the best disperser (“scout” species) indicates habitat suitability for the other species of the group, hence predicts a colonization credit; (ii) the species with intermediate dispersal traits (“median” species) indicates forest species richness and composition of a given hedgerow. In contrast, the worst dispersers (“focal” species) have a low indicator power, as a probable consequence of extinction events occurring in hedgerows, where ecological conditions are suboptimal with respect of forest species. The quality of the corridor increases with the width, height and age of hedgerows, but decreases with increasing land use intensity in adjacent lands. We conclude that process-based indicator species can be valuable tools to assess the efficiency of ecological corridors in fragmented landscapes.

Introduction

The increasing fragmentation of natural and semi-natural habitats together with the intensification of landscape management are widely acknowledged as major threats to biodiversity and associated ecosystem services (Saunders et al., 1991). Fragmentation can magnify the effect of climate change on biodiversity by impeding dispersal and other movements of species across landscapes (Bertrand et al., 2011). Increasing connectivity among habitat patches is assumed to promote species movement among patches, hence to reduce the deleterious effects of patch isolation on biodiversity (Niemelä, 2001, Bennett et al., 2006). In this context, the concept of green and blue infrastructures (GBIs) has emerged as a mean for facilitating movements of species within fragmented landscapes. GBIs are ecological networks composed of one type of habitat with a spatially coherent structure which facilitates species movements, hence promoting the conservation or the enhancement of biodiversity, ecosystem functions, and services delivered to human populations (Benedict and McMahon, 2006, Opdam et al., 2006). Typically, an ecological network is composed of core areas connected by corridors. These GBIs may have various spatial structures but often consist of linear corridors more or less interconnected and interacting with the landscape matrix into which they are embedded. Surprisingly, whilst GBIs have progressively permeated public policies in many countries around the world, the scientific knowledge and tools for identifying the critical features of GBIs, for evaluating their functioning and efficacy for biodiversity conservation, and for guiding their establishment and management are still scarce.

Designing functional connections is complex since connectivity is a species-specific attribute of landscapes with respect of species' dispersal capacities. It is thus challenging to group species sharing similar habitat requirements and dispersal characteristics and determine indicator species (i.e. a species whose requirements and response to environmental changes encapsulate those of many additional species; Landres et al., 1988), for which the landscape connectivity can be determined and extrapolated to the other species of the group.

Indicator species have been widely used to assess environmental conditions, to detect environmental changes or to indicate the diversity of other taxa (Lawton and Gaston, 2001, Halme et al., 2009). For example, umbrella species are species requiring a large area of habitat whose conservation thus confers a protective umbrella to co-occurring species (Murphy and Wilcox, 1986). This concept was further extended by Lambeck (1997) who defined focal species as a subset of the total pool of species in a given landscape with the most demanding survival requirements for several factors threatened by anthropogenic stressors. Though this “surrogate species” scheme is a popular conservation strategy, empirical validation is rare and hardly provides support to their superiority over randomly selected species (Andelman and Fagan, 2000, Fleishman et al., 2001a, Roberge and Angelstam, 2004). Surprisingly, indicator species have been rarely used to diagnose ecosystem functioning and forecast future changes. A notable exception is Bani et al. (2002) who used focal bird and mammal species to design woodland ecological networks in an Italian lowland area.

Forest patches are important habitats for the maintenance of species diversity within lowlands (Benton et al., 2003) and the delivering of services to populations, given the increasing urbanization, construction of transport infrastructures, and agriculture intensification. It has been suggested that hedgerows can act as ecological corridors between forest patches, allowing forest species to migrate along them (Corbit et al., 1999, Jamoneau et al., 2011), but also as refuges for some forest species (Peterken and Game, 1984, McCollin et al., 2000). However, the role of hedgerows as corridors for forest species is still controversial, with several studies that failed to find forest specialists in hedgerow habitats (e.g. Fritz and Merriam, 1993, French and Cummins, 2001, Wehling and Diekmann, 2008). Two hypotheses may explain these conflicting results. First, due to their dispersal limitation, forest species may miss from hedgerows that are too young, simply because they did not have sufficient time to reach the corridor. Comparative studies of recent vs. ancient forests (reviewed in Flinn and Vellend, 2005) indeed revealed that the latter contain more forest habitat specialists than the former, a pattern mainly explained by the low dispersal capacities of many forest herb species (Peterken and Game, 1984, Grashof-Bokdam, 1997). Good dispersers (e.g. wind- and bird-dispersed species) are over-represented relative to weak dispersers (e.g. gravity- and ant-dispersed species) in the species composition of recent forests (Flinn and Vellend, 2005). Therefore, higher species richness in older sites is typically thought to be a result of the accumulation of weaker dispersers over time (Verheyen et al., 2003). Such dispersal limitations are widely documented for recent forest patches (Hermy and Verheyen, 2007), but whether they apply to linear woody habitat such as hedgerows remains unexplored. Second, due to their recruitment limitation, forest species may be unable to establish or to persist in the hedgerow due to low habitat quality, since high nutrient levels in the soil, especially phosphates, combined to highlight levels on the floor, stimulate the growth of highly competitive species such as Rubus fruticosus coll. and Urtica dioica L. (Verheyen et al., 2003, Van der Veken et al., 2004). Moreover, since they are embedded in more or less intensively managed farmlands, hedgerows are exposed to the drift of agrochemicals, especially biocides (Kleijn and Snoeijing, 1997).

Plant species can be categorized into socio-ecological groups, when they share similar habitat requirements, a tenet of phytosociology which groups species according to their observed co-occurrence over ecologically homogeneous land portions (van der Maarel, 2004). Species of a same socio-ecological group thus have similar requirements in terms of light, soil pH, soil nutrient content, soil moisture, light, etc. They can also be classified according to their shared life-history traits into plant functional types (Lavorel et al., 1997). Whenever a species from a given socio-ecological group is found in a given habitat, then it predicts that all species of the same group can be found as well (i.e. the habitat is suitable for them). However, due to dispersal limitation, species that are good dispersers have a higher probability to be found than bad dispersers. In other words, the least dispersal-limited species of a given socio-ecological group is expected to colonize a suitable habitat the first, thereby acting as a “scout” species, a species which predicts that other species of the same socio-ecological group and belonging to the regional species pool will be able to establish later on. In contrast, the most dispersal-limited species of the group is expected to establish the last, and thus predicts the co-occurrence of all other species of the same group. If the habitat is not suitable for a given socio-ecological group, then all species of this group should be absent. However, a suitable habitat may become unsuitable over time, under the pressure of surrounding agricultural disturbances (e.g. eutrophication, soil trampling or ploughing, herbicides); in this case, newly dispersed species may become recruitment limited and/or established species may go extinct over time, deterministically or randomly (i.e. independently from their traits; McCune and Vellend, 2015).

The overall objective of this study is to identify a set of indicator species, which can help in determining whether a linear woody habitat can act as an efficient corridor within a forest metacommunity, with respect of forest vascular plant species. More precisely, we aim at testing the following hypotheses:

H1

All forest plant species can live in hedgerows, as long as they can disperse into them. If this is true, then all species of the local forest species pool will be found in hedgerows, and their frequency therein will increase with their frequency within the forest metacommunity, due to greater diaspore pressure (mass effect).

H2

Within a given socio-ecological group, the least dispersal-limited species (hereafter scout species), whenever it occurs in a given hedgerow, indicates habitat suitability with respect to species of this group. If this is true, then a scout species will be the most frequent species of its group and whenever it is missing, all other species of the group will also be absent. The number of scout species found in a hedgerow will correlate with but underestimate total species richness.

H3

Within a given socio-ecological group, a species exhibiting intermediate dispersal capacities (hereafter median species), whenever it occurs in a given hedgerow, is a good surrogate for the number and identity of co-occurring species from the same group. If this holds true, then median species will be good predictors of total species richness and species composition of a hedgerow.

H4

Within a given socio-ecological group, the most dispersal-limited species (hereafter focal species), whenever it occurs in a hedgerow, predicts the presence of all other species of its group. If this holds true, then a focal species will be the least frequent species of its group, and hedgerows hosting focal species will be more species-rich than those missing them.

H5

Habitat quality in hedgerows, as defined by hedgerow's features and adjacent land use intensity, explains the difference between observed and predicted species richness. If this is true, then the observed richness will be lower (greater) than the predicted one whenever the hedgerow is narrow (wide), without (with) a tree layer (cf. species–area relationship; Rosenzweig, 1995), young (old) (cf. species–time relationship; Rosenzweig, 1995), exposed (not exposed) to intensive adjacent land use.