Linking Climate Change and Fish Conservation Efforts Using Spatially Explicit Decision Support Tools

Abstract

Fisheries professionals are increasingly tasked with incorporating climate change projections into their decisions. Here we demonstrate how a structured decision framework, coupled with analytical tools and spatial data sets, can help integrate climate and biological information to evaluate management alternatives. We present examples that link down-scaled climate change scenarios to fish populations for two common types of problems: (1) strategic spatial prioritization of limited conservation resources and (2) deciding whether removing migration barriers would benefit a native fish also threatened with invasion by a nonnative competitor. We used Bayesian networks (BNs) to translate each decision problem into a quantitative tool and implemented these models under historical and future climate projections. The spatial prioritization BN predicted a substantial loss of habitat for the target species by the 2080s and provided a means to map habitats and populations most likely to persist under future climate projections. The barrier BN applied to three streams predicted that barrier removal decisions—previously made assuming a stationary climate—were likely robust under the climate scenario considered. The examples demonstrate the benefit of structuring the decision-making process to clarify management objectives, formalize assumptions, synthesize current understanding about climate effects on fish populations, and identify key uncertainties requiring further investigation.

RESUMEN

los profesionales de las pesquerías están siendo presionados para incorporar proyecciones de cambio climático en sus decisiones. En este trabajo se demuestra cómo un marco de decisiones bien estructurado, acoplado con herramientas analíticas y bases de datos espaciales, puede ayudar a integrar información climática y biológica para evaluar alternativas de manejo. Se presentan ejemplos que relacionan escenarios de cambio climático con poblaciones de peces, con el fin de abordar dos tipos comunes de problemas: (1) priorización espacial estratégica de recursos limitados para la conservación y (2) decidir si la remoción de barreras migratorias beneficiaría a los peces nativos, los cuales también están amenazados por la introducción de competidores foráneos. Se utilizaron redes Bayesianas (RBs) para traducir cada problema de decisión en una herramienta cuantitativa y se implementaron estos modelos bajo proyecciones climáticas históricas y hacia el futuro. La priorización espacial por medio de RB predijo una pérdida sustancial de hábitat de las especies objetivo para el año 2080, y proveyó medios para mapear tanto los hábitats como las poblaciones que más posibilidades tienen de persistir considerando los distintos escenarios climáticos en el futuro. La simulación de barreras mediante RB aplicadas a tres ríos predijo que las decisiones que implicaban una remoción-previamente hechas asumiendo un clima constante-serían, muy posiblemente, robustas bajo el escenario climático considerado. Estos ejemplos demuestran los beneficios de estructurar el proceso de toma de decisiones con la finalidad de clarificar objetivos de manejo, formalizar las suposiciones de los modelos, sintetizar el entendimiento que hasta la fecha se tiene acerca del efecto del clima en las poblaciones de peces e identificar piezas clave de incertidumbre que requieren de investigación ulterior.

ACKNOWLEDGMENTS

K. Rogers, an anonymous reviewer, and the Science Editor provided helpful reviews that improved the manuscript. D. Peterson was supported by the U.S. Forest Service (Agreement 06-IA-11221659-097) and the U.S. Fish and Wildlife Service's Abernathy Fish Technology Center. S. Wenger was supported by grant G09AC00050 from the U.S. Geological Survey and contracts from the U.S. Forest Service Rocky Mountain Research Station. D. Isaak was supported by the U.S. Forest Service Rocky Mountain Research Station. B. Rieman was supported by an “emeritus” association with the U.S. Forest Service Rocky Mountain Research Station. The findings and conclusions in this article are those of the authors and do not necessarily represent the views of the U.S. Fish and Wildlife Service, U.S. Forest Service, or Trout Unlimited.

Notes

^a Climatically driven nodes that are equivalent to the same nodes in the Cutthroat Trout BN (see Figure 4) but have different state or threshold values

^a Node definition and/or states are listed in Table 1.

^b Values in parentheses are mean summer air temperatures (mean air temperature) estimated for the watershed (wtemp; Wenger et al. 2011b). We generated air temperature categories corresponding to those water temperature states by examining the relationship between Brook Trout occurrence and the mean summer air temperature at a point (ptemp; Wenger et al. 2011b). Additional details are found in Appendix C (see http://fisheries.org/appendices).

^c A threshold value of two events per winter delineated hydrologic regimes as either predominantly snowmelt (less than two) or mixed rain-on-snow and snowmelt (more than two). The threshold value was based on ad hoc interpretation of the geographic distribution of modeled winter high flow frequencies across the Pacific Northwest and Intermountain West United States. Similar approaches have been used to approximate transition points between so-called hydrologic regimes (e.g., Mantua et al. 2010).

^d “Hydrologic regime” is defined as the seasonal pattern of runoff and flooding that might influence bed scour and subsequent incubation or emergence success of fall spawning salmonids like Brook Trout. Hydrologic regime has two states: Snowmelt and mixed rain-on-snow and snowmelt. See Peterson et al. (2008) for additional details.

^e “Stream width” is defined as mean wetted width over the stream network during base flow. Stream width has three states: <3 m (small), 3–10 m (medium), and >10 m (large). See Peterson et al. (2008) for additional details.