Contributed PaperUse of Surrogates to Predict the Stressor Responseof Imperiled SpeciesSETH J. WENGERUniversity of Georgia River Basin Center, 110 Riverbend Road, Athens, GA 30602-1510, U.S.A., email [email protected]: Rare or narrowly distributed species may be threatened by stressors to which they have neverbeen exposed or for which data are very limited. In such cases the species response cannot be predictedon the basis of directly measured data, but may be inferred from the response of one or more appropriatesurrogate species. Here, I propose a practical way to use the stressor response of one or more surrogate speciesto develop a working hypothesis or model of the stressor response of the target species. The process has 4 steps:(1) identify one or more candidate s urrogate species, (2) model the relationship between the stressor andthe response variable of interest for the surrogate species, (3) adapt the stressor–response relationship fromthe surrogate species to a model for the target species, possibly using Bayesian methods, and (4) incorporateadditional data as they become available and adjust the response model of the target species appropriately. Iapplied the approach to an endangered fish species, the amber darter (Percina antesella), which is potentiallythreatened by urbanization. I used a Bayesian approach to combine data from a surrogate species (the bronzedarter [Percina palmaris]) with available data for the amber darter to produce a model of expected amberdarter response. Although this approach requires difficult decisions on the part of the manager, especiallyin the selection of surrogate species, its value lies in the fact that all assumptions are clearly stated in theform of hypotheses, which may be scrutinized and tested. It therefore provides a rational basis for institutingmanagement policy even in the face of considerable uncertainty.Keywords: amber darter, endangered species, management policy, Percina antesella, storm-water runoff,substitute species, surrogate speciesUtilizaci´on de Sustitutos para Predecir la Respuesta de Especies en Peligro a Factores EstresantesResumen: Las especies raras o de distribuci´on restringida pueden estar amenazadas por factores estresantesa los que nunca han sido expuestos o para los cuales existen datos muy limitados. En tales casos, la respuestade las especies no puede ser pronosticada sobre la base de datos medidos directamente, pero pueden serinferidos de la respuesta de una o m´as especies sustitutas apropiadas. Aqu´ı propongo una manera pr´acticade utilizar la respuesta a factores estresantes de una o mas especies sustitutas para desarrollar una hip´otesisde trabajo o modelo la respuesta de la especie a factores estresantes. El proceso comprende cuatro pasos: (1)identificar uno o m´as candidatos a especie sustituta, (2) modelar la relaci´on entre el factor estresante y lavariable de respuesta de inter´es para la especie sustituta, (3) adaptar la relaci´on factor estresante-respuestade la especie sustituta a un modelo para la especie objetivo y (4) incorporar datos adicionales a medidaque sean disponibles y ajustar el modelo de la respuesta de la especie objetivo. Apliqu´eelm´etodo a unaespecie de pez en peligro, Percina antesella, que potencialmente est´a amenazada por la urbanizaci´on. Utiliceun enfoque bayesiano para combinar datos de una especie sustituta (Percina palmaris) con los de P. antesellapara producir un modelo de la respuesta esperada. Aunque este enfoque requiere de decisiones dif´ıciles porparte del manejador, especialmente la selecci´on de la especie sustituta, su valor yace en el hecho de que todoslos supuestos est´an claramente establecidos en forma de hip´otesis, que pueden ser examinadas y probadas.Por lo tanto proporciona una base racional para el establecimiento de pol´ıticas de manejo aun en situacionesde considerable incertidumbre.Paper submitted December 7, 2007; revised manuscript accepted April 10, 2008.1564Conservation Biology, Volume 22, No. 6, 1564–1571C2008 Society for Conservation BiologyDOI: 10.1111/j.1523-1739.2008.01013.xWenger 1565Palabras Clave: escorrent´ıa de aguas pluviales, especie en peligro, especie sustituta, Percina antesella,pol´ıticade manejoIntroductionIn the past decade the use of surrogate, substitute, indica-tor, and umbrella species in the practice of conservationbiology has increased (reviewed by Caro & O’Doherty1999; Carignan & Villard 2002) as has the volume of lit-erature warning of the pitfalls of such practices (e.g.,Andelman & Fagan 2000; Hitt & Frissell 2004; Favreauet al. 2006). The majority of this debate concerns theuse of surrogates or umbrella species to identify conser-vation priorities. Nevertheless, surrogate species can alsobe used to predict the response of a rare or poorly studiedtarget species to anthropogenic stressors (Wahlberg et al.1996; Landres et al. 1988; Caro et al. 2005). (I use sur-rogate species in the same sense Caro et al. [2005] usedsubstitute species.) This latter practice has been criti-cized as “incautious” (Caro et al. 2005) or “inappropriate”(Landres et al. 1988) because there are no guarantees thatthe target species will respond in the same manner as thesurrogate.Nevertheless, there are cases in which the response ofa species to a potential stressor cannot be known fromexisting data, and suitable experiments cannot be con-ducted without jeopardizing survival of the species. Thisis especially true for narrowly endemic species, manyof which are imperiled. For example, the amber darter(Percina antesella)—a small freshwater fish—only oc-curs in 2 rivers in the headwaters of the Coosa Riversystem in the southeastern United States. Populationsin both rivers are potentially threatened by increasingurbanization, but because of the limited distribution ofthe species, it has not yet been sufficiently exposed tothe effects of urbanization to determine thresholds of re-sponse. A “natural experiment” in which the species isexposed to increased urbanization could threaten its sur-vival. A potential solution is to examine the relationshipbetween urban cover and a surrogate species that is re-lated to the amber darter or that has similar habits but ismore widely distributed so that it has been exposed tourbanization over part of its range. A working hypothesisof the stressor response of the amber darter can be builton this surrogate stressor response and provide a rationalbasis for