Elsevier

Ecological Modelling

Volume 221, Issue 16, 10 August 2010, Pages 1934-1943
Ecological Modelling

On the risk of extinction of a wild plant species through spillover of a biological control agent: Analysis of an ecosystem compartment model

https://doi.org/10.1016/j.ecolmodel.2010.05.003Get rights and content

Abstract

Invasive plant species can be controlled by introducing natural enemies (insect herbivores) from their native range. However, such introduction entails the risk that the introduced herbivores attack indigenous plant species in the area of introduction. Here, we study the effect of spillover of a herbivore from a managed ecosystem compartment (agriculture) to a natural compartment (non-managed) and vice versa. In the natural compartment, an indigenous plant species is attacked by the introduced herbivores, whereas another indigenous plant species, which competes with the first, is not attacked. The combination of competition and herbivory may result in extinction of the attacked wild plant species. Using a modelling approach, we determine model parameters that characterize the risk of extinction for a wild plant species. Risk factors include: (1) a high attack rate of the herbivores on the wild non-target species, (2) niche overlap expressed as strong competition between the attacked non-target species and its competitor(s), and (3) factors favouring large spillover from the managed ecosystem compartment to the natural compartment; these include (3a) a high dispersal ability, and (3b) a moderate attack rate of the introduced herbivore on the target species, enabling large resident populations of the insect herbivore in the managed compartment. The analysis thus indicates that a high attack rate on the target species, which is a selection criterion for biocontrol agents with respect to their effectiveness, also mitigates risks resulting from spillover and non-target effects. While total eradication of an invasive plant species is not possible in the one-compartment-one-plant-one-herbivore system, natural enemy spillover from a natural to a managed compartment can make the invasive weed go extinct.

Introduction

Invasive plant species pose a great problem to global agriculture and ecosystems, threatening valuable indigenous species and productivity in agricultural and natural systems (Callaway and Aschehout, 2000, Pimentel, 2002, Sheppard et al., 2003). Classical biological control, by introducing natural enemies from the native range, is widely regarded as a valuable method for managing invasive species (Ehler, 1998, Thomas and Willis, 1998, Pemberton, 2000). Classical biological control avoids the use of herbicides and can be highly cost-effective (Charudattan, 2001). Chalak-Haghighi et al. (2008) have recently shown that an insect herbivore (Apion onopordi) can increase the net present value obtained from pastures in New Zealand by reducing the growth rate of Californian thistle (Cirsium arvense).

Many authors have discussed the environmental risks of classical biological control (e.g. Thomas and Willis, 1998, Follett and Duan, 1999, Wajnberg et al., 2001). For example, natural enemies may attack non-target species. In order to assess this risk we need to understand the ecological dynamics of the biological control agent in the ecosystems where they are introduced, including their interactions with other species. These interactions include both local population interactions as well as spillover of enemies from one ecosystem compartment to another.

Many of the biological control agents introduced for pest control in agricultural areas can feed on alternative host plants in natural habitats and are likely to spill over from agricultural into natural systems (Henneman and Memmott, 2001, Symondson et al., 2002, Rand et al., 2006, Wirth et al., 2007). This spillover can result in important adverse consequences (Suarez et al., 1998, Cronin and Reeve, 2005, Rand et al., 2006). For instance, the weevil Rhinocyllus conicus, introduced for the biological control of Platte thistle (Cirsium canescens) in the United States, attacked a protected and rare relative, the Pitcher's thistle (Cirsium pitcheri) (Louda et al., 2003, Louda et al., 2005). Adult beetles of the corn rootworm (Diabrotica ssp.), which feed in agricultural land as larvae, spillover into tall-grass prairie causing damage to native plants (McKone et al., 2001). Thus, before introducing a herbivore to a managed system, it is important to consider potential spillover effects to the natural environment, resulting in attack on endangered or protected species in the natural environment.

Because ecological conditions of the managed and natural systems differ, a variety of plant species interactions can prevail in managed and natural systems. Thus, a herbivore may be able to build large populations in one compartment, spill over to another compartment, and affect species interactions and survival in this other compartment. There is a need for analysis of the conditions under which dispersal of a biological control agent from a managed to a natural system produces a spillover effect that is large enough to threaten biodiversity (Rand et al., 2006).

Here, we use a two compartment modelling approach to elucidate risks when releasing a biological control agent to a managed compartment that can spill over to another ecosystem compartment and attack a valuable indigenous species in this other compartment. We focus on the wild plant species’ risk of extinction. There have been some studies on two compartment model systems (Vellend et al., 2003). While compartmentalised ecosystem models appear a very suitable tool to study spillover and its effect on ecosystems, such models have, to the best of our knowledge, to date not been used to study extinction risks resulting from spillover.

We develop here a model consisting of two compartments, that represents key processes such as the interaction between a herbivore and its target and non-target plant species, dispersal of the herbivore between ecosystem compartments, and the competitive relationships between a non-target species and other species in a natural compartment. The objectives are to identify those system characteristics that enhance or mitigate the risk of extinction of the non-target plant species in the natural compartment, and to gain insight in the interrelationships between the different dynamic processes involved. In the next section the model system is described, followed by a mathematical analysis. Next, a numerical analysis is presented and finally, conclusions are drawn.

Section snippets

Description of the model system

For our analysis we model our system as two compartments: (1) a managed compartment where a herbivore (zm, numbers m−2) is introduced to control a pest weed (w, shoots m−2), and (2) a natural compartment where the same herbivore species (here denoted as zn, numbers m−2) can attack a valuable wild plant species (species x, shoots m−2) (Fig. 1). The two herbivore populations are linked by dispersal, enabling the introduced species to spill over from one compartment to the other if the densities in

Mathematical analysis

Before conducting numerical analyses, the 5-dimensional system was analysed mathematically to obtain all of its equilibria and determine their stability. To facilitate mathematical analysis, the model was first non-dimensionalized, i.e. units were removed from the state variables as well as from the parameters by substitution of variables (Appendix A). This non-dimensionalization reduces the number of parameters from 13 (Table 1) to nine (system (A1)). Moreover, the combinations of original

Numerical analysis

For the numerical analysis, two equilibria are of particular interest: (1) equilibrium xiv in which all species coexist and (2) equilibrium x in which species x is extinct due to the combination of herbivore attack by zn and competition with y. A dynamic trajectory that starts from disturbance from the positive equilibrium xiv ending in equilibrium x is of special interest because it represents extinction of x, allowing us to investigate which parameter values would lead plant species x to

Comprehensive sensitivity analysis

A comprehensive local sensitivity analysis demonstrates the effect of all parameter values on the equilibrium densities of all state variables (Fig. 9). First, the relationship between percentage change from the default parameter values and the equilibrium density of plant species x is presented (Fig. 9A and B). The density of species x at equilibrium increases as x becomes a stronger competitor and herbivory on x decreases. Thus, on the one hand x is a strictly increasing function of

Discussion

This paper puts forward a theoretical model framework for analysing which factors contribute to extinction risk of a wild non-target plant species due to spillover of a herbivore introduced for biological control in agriculture. Extinction is enhanced by: (1) a large resident population of the herbivore in the agriculture compartment, which is the case at intermediate values of the attack rate on the target weed; (2) a high attack rate of the herbivore on the non-target wild species; (3) a high

Acknowledgments

We thank Dr. A.J.V. Rotteveel and T. Heijerman for helpful discussions on the parameter values. We acknowledge E. Hendrix and D. Pannell for their useful comments. We thank two anonymous reviewers for insightful comments.

References (28)

  • F.J.J.A. Bianchi et al.

    Model evaluation of the function of prey in non-crop habitats for biological control by ladybeetles in agricultural landscapes

    Ecological Modelling

    (2004)
  • L.E. Ehler

    Invasion biology and biological control

    Biological Control

    (1998)
  • R.D. Holt

    Predation, apparent competition, and the structure of prey communities

    Theoretical Population Biology

    (1977)
  • M.B. Thomas et al.

    Biocontrol—risky but necessary?

    Trends in Ecology and Evolution

    (1998)
  • M. Begon et al.

    Ecology: From Individuals to Ecosystems

    (2006)
  • F.J.J.A. Bianchi et al.

    Sustainable pest regulation in agricultural landscapes: a review on landscape composition, biodiversity and natural pest control

    Proceedings of the Royal Society, Series B

    (2006)
  • R.M. Callaway et al.

    Invasive plants versus their new and old neighbors: a mechanism for exotic invasion

    Science

    (2000)
  • M. Chalak et al.

    On the economics of controlling an invasive plant: a stochastic analysis of a biological control agent

    International Journal of Environmental Technology and Management

    (2009)
  • M. Chalak-Haghighi et al.

    Management strategies for an invasive weed: a dynamic programming approach for Californian thistle in New Zealand

    New Zealand Journal of Agricultural Research

    (2008)
  • R. Charudattan

    Biological control of weeds by means of plant pathogens: significance for integrated weed management in modern agro-ecology

    BioControl

    (2001)
  • J.T. Cronin et al.

    Host–parasitoid spatial ecology: a plea for a landscape-level synthesis

    Proceeding of the Royal Society B

    (2005)
  • L. Edelstein-Keshet

    Mathematical Models in Biology

    (1988)
  • P. Follett et al.

    Nontarget Effects of Biological Control

    (1999)
  • M.L. Henneman et al.

    Infiltration of a Hawaiian community by introduced biological control agent

    Science

    (2001)
  • Cited by (11)

    • Eco-evolutionary dynamics of agricultural networks: Implications for sustainable management

      2013, Advances in Ecological Research
      Citation Excerpt :

      This increase of contact zones may, however, have evolutionary consequences for populations inhabiting semi-natural areas, by changing their interactions within their species assemblage (Knight et al., 2005). A recent review also highlighted the frequency of the opposite direction of spillover, from crops to semi-natural areas for a wide range of functional groups (Blitzer et al., 2012), which will depend on species traits, such as dispersal ability and attack rate on the pest prey (Chalak et al., 2010), and on similarity between cultivated and semi-natural areas (Opatovsky et al., 2010; Prevedello and Vieira, 2010). It may have dramatic consequences for ecological networks when crops are the dominant land cover and when they harbour larger populations of organisms than semi-natural habitats.

    • Impacts of insect herbivores on rare plant populations

      2011, Biological Conservation
      Citation Excerpt :

      Tests for density dependence are rare because they require measurement of herbivore impact in multiple populations or at multiple time points (with or without density manipulation by the researcher) – something that is logistically challenging for any system but particularly difficult for rare plants. The possibility that exotic herbivores introduced for biocontrol might expand their host ranges to impact rare native plants is frequently discussed (e.g., Louda et al., 2003; Sheppard et al., 2005; Chalak et al., 2010). However, we could locate just two well-studied examples: attack by the weevil Rhinocyllus conicus on the threatened thistles Cirsium canescens and C. pitcheri (Louda et al., 2005; Rose et al., 2005), and attack by the moth Cactoblastis cactorum on the endangered cactus O. corallicola (Stiling et al., 2004).

    • Invasive Alien Species in Iraq

      2021, Invasive Alien Species: Observations and Issues from Around the World
    View all citing articles on Scopus
    1

    Current address: Water Economics and Institutions Group, Royal Haskoning, P.O. Box 151, 6500 AD Nijmegen, The Netherlands.

    View full text