Experimental study designs to improve the evaluation of road mitigation measures for wildlife
Introduction
Roads and traffic have negative impacts on a wide range of animals (reviewed in Trombulak and Frissell, 2000, Spellerberg, 2002, Fahrig and Rytwinski, 2009, Benítez-López et al., 2010, Rytwinski and Fahrig, 2012, van der Ree et al., 2015a). The main focus of road ecology research is to quantify these negative impacts, with the aim of avoiding, minimizing, mitigating, or offsetting negative impacts on individuals, populations, communities, and ecosystems (van der Ree et al., 2011). Options to avoid or mitigate these negative impacts are numerous and have been widely and increasingly implemented around the world (van der Ree et al., 2015b). Examples of mitigation measures include: animal detection systems, wildlife warning signs, changes in road-verge management, measures to reduce traffic volume, speed and/or noise, temporary road closures, wildlife crossing structures, wildlife fences [e.g., barrier fencing (or exclusion fencing) that prevents wildlife from accessing the road, or funnel fencing that primarily funnels animals to wildlife crossing structures but can also prevent wildlife from accessing the road], wildlife reflectors, wildlife repellents, and modified road designs/viaducts/bridges/lighting (Clevenger and Ford, 2010, Huijser and McGowen, 2010). Wildlife crossing structures (e.g., under- or over-passes: amphibian tunnels, badger pipes, ledges in culverts, land bridges, rope bridges, glider poles), combined with fencing to prevent mortality and funnel wildlife towards crossing structures, have gained considerable recent attention by transportation agencies because they enhance landscape connectivity without affecting traffic flow (van der Grift et al., 2013).
There is compelling evidence that many wildlife species regularly and frequently use crossing structures (reviewed in van der Ree et al., 2007), and that well-designed and maintained fencing greatly reduce rates of wildlife mortality and funnels animals towards the crossing structures (reviewed in Glista et al., 2009). Unfortunately, documenting use of a crossing structure (i.e., ‘success’ at the individual level) is so far removed from higher level quantities of interest (i.e., population size and persistence), that such studies provide little information as to whether the structure actually mitigates the effect of the road enough to ensure a viable population (van der Grift et al., 2013). Consequently, the influence of these mitigation structures on population viability is unclear for most road-affected species.
Ultimately, we want to be confident that the predicted impact of a road on a wildlife population will be at least partially mitigated by the proposed road design and that the investment in crossing structures and/or fencing is justified. For example, if the mitigation for an endangered species is ineffective such that the population of the target species declines, the road agency must respond and retrofit the road or modify the mitigation structures. In such cases, it is essential that road agencies have reliable evidence to make informed decisions about which feature of the road or mitigation should be implemented or modified and by how much.
Here we identify seven key questions road planners commonly have about crossing structures and/or fencing that for many species and structure types remain largely unanswered at the level of ultimate concern (e.g., population or community) and at the required level of certainty by existing research (Box 1). These questions must be answered not only so that resources for road mitigation are allocated in the most effective manner, but that they indeed have the predicted (desired) effect.
There are two main reasons why these questions have remained unanswered. First, the existing approach to road mitigation is to simply adopt current best-practice in terms of the type, number, and location of mitigation. While this approach identifies the best known mitigation for installation, it does not explicitly facilitate learning about the effectiveness of mitigation because the mitigation was installed to solve a problem, not generate new information. Second, studies evaluating the effectiveness of mitigation structures have low inferential strength, and, as such, comparatively low predictive power. For example, studies often lack: (1) comparisons between treatment sites (also referred to as ‘impact’ sites in Before-After-Control-Impact (BACI) study designs (Roedenbeck et al., 2007, van der Ree et al., 2015b)) and control sites (i.e., sites that have not been affected by the treatment – these will vary depending on the question and goals of the road mitigation, but may include e.g., road-free areas, areas with narrow or low-traffic volume roads, unmitigated roads, and/or unmanipulated mitigation measures; see section 5 on experimental designs for more detail); (2) data on population sizes or trends prior to mitigation; (3) replication in both space and time; and (4) randomization of treatment and control sites across the pool of potential study sites. Moreover, many study designs confound mitigation variables (e.g., overpass width, density of shrubs at culvert entrance) such that their independent effects cannot be evaluated (reviewed in van der Ree et al., 2007, Glista et al., 2009). For road agencies to make informed and reliable decisions, we need to improve the rigor of studies that evaluate the effectiveness of mitigation measures.
Ways to improve the quality and impact of road ecology research and monitoring have been previously discussed. Roedenbeck et al. (2007) provided a research agenda for road ecology, identifying relevant questions (e.g., Under what circumstances do roads affect population persistence?, and Under what circumstances can road effects be mitigated?), and specifying a hierarchy of study designs for answering these questions. van der Grift et al. (2013) used the principles outlined in Roedenbeck et al. (2007) to propose a methodological framework for increasing the inferential strength of mitigation monitoring schemes. Lesbarrères and Fahrig (2012) proposed the use of such monitoring schemes as a type of experiment, but they did not suggest associated experimental protocols. van der Ree et al. (2015b) summarises these papers into an accessible format for practitioners. Here, we set out the need and standards for using experimental approaches to road mitigation to improve knowledge on the influence of mitigation structures on wildlife populations. We first demonstrate the need for an experimental (manipulative) approach to road mitigation projects. We then outline the road/mitigation project stages and describe how flexibility in experimental design depends on the stage in the road project at which researchers become involved. We provide experimental approaches to answering each of the questions in Box 1, highlighting real case studies when possible, and we conclude with a discussion of potential issues in using experimentation to evaluate the effectiveness of crossing structures and fencing.
Section snippets
Why we need an experimental approach to road mitigation
Most road agencies currently evaluate the effectiveness of mitigation efforts through post-implementation monitoring (van der Ree et al., 2015b). Questions commonly asked with monitoring include: Will a particular species use an underpass to cross the road? and/or Which factors are correlated with the number of crossings? However, the state of road ecology has matured since mitigation structures were first developed, and while monitoring rate of crossing is an important first step, and likely
Road project stages and experimentation
There is a direct relationship between the road/mitigation project stage at which researchers become involved, the ease of experiment implementation, and the knowledge gain at the end of the experiment. If the experiment is designed early in the planning process, a wider range of experiments are possible with some at least having comparatively high inferential strength, broader implications, and greater potential to extrapolate to other locations/species. On the other hand, such experiments
Opportunity for improved experimental design
The flexibility to change components of the experimental design will be influenced by the stage within the road/mitigation project when researchers begin collaborating with road planners (Fig. 3). For a description of general study designs for road mitigation see Appendix A. Here, we define treatment sites as sites where mitigation measures are to be installed or manipulated, and there can be different types of treatments used in an experiment e.g., barrier fencing, land-bridges, culverts,
Experimental designs
In this section, we describe potential experimental approaches for each of the research questions in Box 1 and provide, when possible, case studies highlighting mitigation experiments that have been conducted or are proposed. More specifically, we provide example BACI study designs for two stages in the road/mitigation project where researchers may become involved: (1) at the beginning of a road/mitigation project i.e., project development stage, and (2) after the mitigation has been
Challenges with experimentation
Using experimentation to evaluate road mitigation effectiveness does not come without challenges. First, ethical issues may arise when the experimental design includes mitigation measures that the researcher suspects (but has not yet demonstrated) to be ineffective. A possible solution to this issue might be to build all the structures to the design suspected to be effective, but install reversible modifications in some (randomly selected) that ‘downgrade’ them to designs that are assumed to be
Concluding remarks
If road mitigation experiments became a standard part of any new or existing road project, how will the next 20–30 years look? Quite simply, improvements in road mitigation would increase rapidly. Each new mitigation experiment would build on the insights from past experiments (even those in other regions or on other species) and be incorporated in the design of new road mitigation experiments to ensure more effective mitigation. Currently there is simply not enough information on the
Acknowledgements
This paper began through discussions at a meeting in Montreal, QC in 2009 and continued through a road ecology workshop held at Auberge Val Carroll, QC, Canada. We thank the staff of Val Carroll for hosting a wonderful weekend and L. Francisco Madriñan for comments in earlier discussions leading towards this paper.
References (36)
- et al.
Wildlife–vehicle collision mitigation: is partial fencing the answer? An agent-based model approach
Ecol. Model
(2013) - et al.
The impact of roads and other infrastructure on mammal and bird populations: a meta-analysis
Biol. Conserv.
(2010) - et al.
A review of mitigation measures for reducing wildlife mortality on roadways
Landsc. Urban Plan.
(2009) - et al.
Measures to reduce population fragmentation by roads: what has worked and how do we know?
Trends Ecol. Evol.
(2012) - et al.
Do species life history traits explain population responses to roads? A meta-analysis
Biol. Conserv.
(2012) - et al.
Movement re-established but not restored: inferring the effectiveness of road crossing mitigation for a gliding mammal by monitoring use
Biol. Conserv.
(2013) - et al.
Wildlife crossing structures, fencing, and other highway design considerations
- et al.
Effect of cover on small mammal movements through wildlife underpasses along US highway 93 north, Montana, USA
- et al.
Culverts alone do not reduce road mortality in anurans
Ecoscience
(2014) - et al.
Effects of roads on animal abundance: an empirical review and synthesis
Ecol. Soc.
(2009)
Coordinated distributed experiments: an emerging tool for testing global hypotheses in ecology and environ mental science
Front. Ecol. Environ.
Big problems, small science
Conserv. Ecol.
Reducing wildlife-vehicle collisions
Habitat continuity and social organisation of the mountain pygmy-possum restored by tunnel
J. Wildl. Manag.
Evidence that a highway reduces apparent survival rates of squirrel gliders
Ecol. Soc.
Southern Highways Program 2010–2014
Incorporating biodiversity issues into road design: the road agency perspective
The Rauischholzhausen agenda for road ecology
Ecol. Soc.
Cited by (61)
Overestimation of mitigation leads to underestimation of residual impacts
2024, Environmental Impact Assessment ReviewWhere do wildlife cross the road? Experimental evaluation reveals fauna preferences for multiple types of crossing structures
2023, Global Ecology and ConservationThe lost road: Do transportation networks imperil wildlife population persistence?
2021, Perspectives in Ecology and ConservationCitation Excerpt :There is great lack of evidence on the effectiveness of mitigation actions in reducing negative impacts on population persistence, as only 3% of the studies that focused this topic used empirical population-oriented approaches (see examples below). Dedicated studies are essential to confirm the effectiveness of road mitigation to ensure (or not) the viability of populations inhabiting road vicinity (Lesbarrères and Fahrig, 2012; Rytwinski et al., 2015). Comparing the impacts before and after road construction with controls, both for no road and for road with no mitigation measure (i.e., Before-After-Control-Impact, BACI designs) provide less biased findings and allow stronger inferences than simpler observational designs (Lesbarrères and Fahrig, 2012; Christie et al., 2020).
Road-effect mitigation promotes connectivity and reduces mortality at the population-level
2021, Biological ConservationCitation Excerpt :Results from our robust 6-year BACI study design and rigorous field methodology strongly support our prediction that the exclusion fencing reduces reptile and amphibian mortality, but that effects are taxon-specific. Further, we show that crossing structures in Presqu'ile PP effectively promote landscape connectivity for local herpetofauna at the population level quantifying usages rates of captured animals as well as usage relative to estimated population abundance, addressing a crucial gap in our understanding of road-effect mitigation dynamics at the population-level (Rytwinski et al., 2015). Our paired BACI approach allowed us to identify and control for yearly variation, which has confounded previous studies (Lesbarrères and Fahrig, 2012; Rytwinski et al., 2015).
Where wildlife and traffic collide: Roadkill rates change through time in a wildlife-tourism hotspot
2021, Global Ecology and ConservationPairing long-term population monitoring and wildlife crossing structure interaction data to evaluate road mitigation effectiveness
2021, Biological ConservationCitation Excerpt :Though there is widespread agreement on the need for road mitigation measures, installing new wildlife crossings and fencing infrastructure is a costly, long-term, multi-step process, which requires the identification of mitigation goals and a data-driven design for site selection on the landscape (Polak et al., 2019). Another key element of road mitigation is sufficient monitoring to assess effectiveness (Rytwinski et al., 2015; van der Grift et al., 2013). Ultimately, road mitigation measures aim to improve population viability.