©2002 CRC Press LLC Effects of Chronic Stress on Wildlife Populations: A Population Modeling Approach and Case Study Diane E. Nacci, Timothy R. Gleason, Ruth Gutjahr-Gobell, Marina Huber, and Wayne R. Munns, Jr. CONTENTS 10.1 Introduction 10.2 A Population Matrix Modeling Approach 10.3 A Stressor of Ecotoxicological Concern 10.4 A Case Study 10.4.1 Toxicological Responses 10.4.2 Matrix Model Projections 10.4.3 Compensatory Mechanisms 10.4.3.1 Life History Shifts: Compensatory Demographic Responses 10.4.3.2 Physiological Response Shifts: Compensatory Toxicological Responses 10.4.4 The Scale of Evolutionary Effects 10.4.5 Risks of Selection and Adaptation 10.5 A Population Modeling Approach and Case Study: Conclusions Acknowledgments References 10.1 INTRODUCTION As a society, we have made commitments to preserve environmental quality, not only for its direct value to humans, but also to support aquatic and other wildlife species. These commitments have become legal mandates in the form of legislation 10 ©2002 CRC Press LLC such as the Clean Water Act. In the most general sense, adverse effects on wildlife species caused by human activities, or anthropogenic stress, result in changes to their densities and distributions. Although such changes can be measured at varying levels of biological organization, populations have been defined as a valued unit for wildlife protection and management. There has been controversy as to whether or not wildlife protection at levels of biological organization higher than the individual truly reflects societal values and whether or not it is effective. 1 However, the impetus toward using the population as the protection unit has occurred for scientific as well as political reasons. Scientif- ically, some ecologists regard populations as sustainable units, valued for important properties beyond those inherent to individuals (i.e., emergent properties). Others regard this move as a practical response to the increased recognition that environ- mental management involves choices and costs. In any case, implicit in this approach is the philosophy that the loss of some individuals does not affect population, and therefore species, persistence, except when population sizes are very low (i.e., threatened or endangered species). While a healthy debate on the value of population protection continues, approaches to quantify effects on populations should be devel- oped and evaluated. This chapter describes a matrix modeling approach to characterize and project risks to wildlife populations subject to chronic stress. Population matrix modeling was used to estimate effects of one class of environmental contaminants, dioxin- like compounds (DLCs), to populations of an ecologically important estuarine fish species, Fundulus heteroclitus, or mummichogs. This approach was applied to a case study site highly contaminated with polychlorinated biphenyls (PCBs), includ- ing DLCs. Model projections suggested high risks to populations of mummichogs subject to intense DLC exposures. However, field observations of mummichog populations indigenous to this site appeared to be inconsistent with these projec- tions. This apparent disparity provided an opportunity to use the population model structure to develop and test hypotheses on how wildlife populations respond to chronic stress. The directed research that followed has resulted in a more holistic assessment that integrates the perspectives of the contributing toxicologists, biol- ogists, and ecologists. 10.2 A POPULATION MATRIX MODELING APPROACH An essential component of the analysis phase of risk assessment is the development of a quantitative relationship between the stressor of concern and an ecological response 2,3 (Figure 10.1). To assess risks to populations, this response should reflect some attribute of population health, such as size, growth rate, or probability of persistence. However, direct measurements of population responses are difficult to acquire and often unavailable. Instead, wildlife toxicologists and risk assessors have used laboratory bioassays to develop quantitative relationships between stressors and individual responses. In these bioassays, adverse effects are often defined as reduced reproductive output or increased mortality for individuals exposed to stres- sors throughout vulnerable portions of their life cycle. 4 However, these measures of FIGURE 10.1 Stressor–response profile illustrating effects on wildlife populations, shown as one component of the analysis phase of the U.S. EPA’s ecological risk assessment (ERA) framework. (Adapted from Reference 2.) ©2002 CRC Press LLC ©2002 CRC Press LLC impaired performance or loss of individuals do not provide sufficient information to define a quantitative relationship between stressors and population health. Life history theory provides the theoretical basis to link performance of indi- viduals and population dynamics that is fundamental for understanding how envi- ronmental stressors can affect population regulation and life history evolution. 5,6 In general, the relationship between individual and population traits is not linear, and changes in some traits have a greater impact than others on populations. 7 For example, changes in survival often have a greater impact on populations than changes of a similar magnitude in reproduction. 9 A matrix modeling approach provides a mathematical mechanism for integrating individual performance traits into estimates of population size and dynamics. 6,9–12 Specifically, the rate of increase per individual ( r ), is dependent upon vital rates, i.e., survival probabilities, time between reproductive events, and reproductive out- put. The number of individuals ( n t ) is calculated at regular intervals ( n t + 1 ) using matrix algebra, and the rate of change for the total number of individuals over this interval is termed population growth rate ( = e r ). Population projections reflect population dynamics when conditions remain constant and responses are fixed throughout the period of concern. For example, when population growth rate is projected to be less than 1, the trajectory over time shows decreasing population size, and an increased probability for local extinction. Population matrix models can be constructed to reflect varying degrees of com- plexity and specificity. In simple matrix models, the population is assumed to be closed (i.e., no immigration or emigration) and unbounded by carrying capacity (i.e., density independent). Therefore, changes in population size are affected only by initial population size, distribution, and vital rates. Further, the simplest models are deterministic, i.e., demographic or environmental stochasticity is not considered. In matrix modeling, average rates of vital parameters are determined for the population as a whole or for classes within the population. In age- and stage- structured models, populations are considered as aggregates of linked but discrete classes, permitting the incorporation of class-specific vital rates. Matrix algebra is used to solve one or more difference equations that are used to calculate the number of individuals in each class and, through summation, in the population. This struc- turing also permits the incorporation of class-specific stressor responses. For exam- ple, early life stages are often toxicologically more sensitive stages than adults. As an illustration, a matrix model can be constructed to represent a species with a four-stage life cycle, e.g., typical of many fish species (Figure 10.2). The numbers of individuals in each stage for any period will be affected by the starting number and the rates for processes by which individuals move in and out of stages. For example, the number of individuals in stage 0 will be affected by the fecundity rates for mature stages 1, 2, and 3 (described by f 2 , f 3 , and f 4 , respectively). The number of stage 0 individuals also will be affected by the relative probabilities that an individual will remain in that stage (survival probability, P 1 ) or develop into a stage 1–classified individual (transition probability, G 1 ). How the model is constructed and parameterized defines the relative contribu- tions of each stage or process to population dynamics. Sensitivity or elasticity analysis can be used to evaluate and rank parameters for their influence on population ©2002 CRC Press LLC dynamics, i.e., the magnitude by which population attributes change in response to small changes in parameter values. 13,14 Processes that strongly influence population effects may be demographically important regulators of populations under stress. For example, for species with high reproductive output, changes in the survival of reproductively mature life stages will have a greater impact on population growth rate than changes of similar magnitude on survival of early life stages. 15 Demographic data for unstressed populations are often acquired from published studies of populations in field or laboratory conditions. These initial or reference parameter estimates can be replaced or modified to reflect stress responses (Figure 10.3), e.g., as described by stressor–response relationships from laboratory studies. 16 Specifically, population models parameterized with stressor responses can be used to project how population attributes like population growth rate would be affected by a constant or chronic level of stress. By integrating stressor–response FIGURE 10.2 Demographic matrix models translate conceptual models of life history char- acteristics into mathematical models ( M ) that integrate stage-specific rates for survival ( P ), development (stage transition, G ), and reproduction ( f ) into projections of population size at some time ( n t , t + 1 ). f f f G G G PP P P P P PP M n n t G G G fff t t t t t t t t t ©2002 CRC Press LLC relationships into a demographic framework, effects on individuals are translated into effects on populations. Therefore, the accuracy of modeling projections is dependent upon the accuracy and completeness of both demographic and toxicolog- ical relationships for the populations of concern. 10.3 A STRESSOR OF ECOTOXICOLOGICAL CONCERN The U.S. Environmental Protection Agency (U.S. EPA) has recognized national concerns about the effects on aquatic and other wildlife species of dioxin and other contaminants that act through similar toxicological mechanisms. 17 These contami- nants, classified as dioxin-like compounds, include polychlorinated dibenzo- p -dioxin (PCDD), dibenzofuran (PCDF), and certain PCB congeners. 18 Like other persistent bioaccumulative and toxic contaminants, DLCs occur in detectable concentrations in many wildlife populations. 2 Concerns about their effects on fish have been reinforced by results of eco-epidemiological studies demonstrating that DLCs have contributed to the decline of lake trout in the Great Lakes. 19,20 A population matrix modeling approach provided an opportunity to predict risk of DLCs for populations of other fish species. Although PCB congeners act through several mechanisms, and vary widely in toxic potencies, the mechanism of action for the most potent congeners that resemble dioxins (i.e., DLCs) has been the subject of much study. 18,21 These congeners are non- and mono- ortho substituted congeners whose toxic potency can be evaluated relative to 2,3,7,8-tetrachlorodibenzo- p -dioxin (TCDD or dioxin) using a toxic- equivalency approach. 18 The toxic effects of DLCs are mediated, in large part, through the aryl hydrocarbon receptor (AhR). 22,24 The DLCs are extremely toxic to FIGURE 10.3 Ecotoxicological/demographic matrix models integrate mathematically stage- and process-specific stressor–response relationships into projections of population-level effects (i.e., population growth rate) relative to reference or unstressed level. n n M tt ©2002 CRC Press LLC the early life stages of fish. 25–27 Specifically, pericardial and yolk sac edema (“blue sac” disease) and subcutaneous hemorrhaging are characteristic pathologies in devel- oping fish exposed to DLCs. 25 Poor growth, “wasting syndrome,” and direct or indirect reductions in reproductive output have also been produced by DLC expo- sures to adult vertebrates, including fish species. 21,27,28 The AhR signal transduction pathway is activated through binding with xeno- biotic ligands that include DLCs, and results in the transcriptional regulation of several proteins. 22,24 Proteins induced by the AhR pathway include a major xenobi- otic-metabolizing enzyme, cytochrome P-4501A1 (CYP1A1). The CYP1A1, induced when ligands bind the AhR, is a specific catalyst for ethoxyresorufin o - deethylase (EROD). Thus, elevated EROD activity has been used as a specific indicator of vertebrate exposure and response to AhR ligands, including DLCs. 22,29,30 10.4 A CASE STUDY New Bedford Harbor (NBH), Massachusetts was selected as a case study site to evaluate the utility of a matrix modeling approach for the projection of population- level effects associated with DLC exposures (Figure 10.4). Although typical in terms of nutrient overenrichment, habitat loss, and other characteristics of anthropogenic disturbance 31 of many urban estuaries of the northeast coast of the United States, NBH sediment and biota contain extraordinarily high concentrations of PCBs. 31–33 These findings suggest that PCBs, especially DLCs, are toxicologically important stressors in NBH. According to historical records, PCBs were discharged into the northern or upper harbor as industrial wastes from the 1940s to the 1970s, producing contamination of sufficient magnitude to warrant listing on the U.S. EPA National Priorities List as a Superfund site. 31 Sediment PCBs in the Superfund site have been measured at levels as high as 2100 g/g dry weight in NBH (total PCBs). 32 This value is four orders of magnitude greater than the sediment guideline value for total PCBs that has been correlated with probable adverse biological effects (180 ng/g dry weight). 34 Consistent with historical records, PCB concentrations in sediments at the Superfund site have been at toxic levels for decades 35 (Figure 10.5). Although the entire harbor is contaminated, 31 there is a steep gradient of PCB concentrations in sediment 36 and biota 37,38 from the northern to southern (Hurricane Barrier) boundaries of the NBH. Despite high levels of contamination, a few fish species, including mummichogs, exist in great abundance in NBH. 39 Although PCB discharge ceased in 1976, 40 biota sampled from NBH more than 20 years later continue to accumulate PCBs. 33 For example, the mean concentration for total PCBs in livers of mummichogs collected in 1996 from the upper harbor Superfund site was 324 g/g dry weight. 38 In com- parison, the mean concentration of total PCBs in livers of mummichogs from West Island (WI), a reference site outside NBH, was 2.4 g/g dry weight. 38 Mummichogs are a nonmigratory fish with no dispersive life stages. 41 Although mummichogs reside in an essentially continuous band along the East Coast of the United States, studies have shown that there is limited gene flow between populations. 42 These findings suggest that mummichogs are subject to the environmental attributes of a limited geographical location throughout their FIGURE 10.4 Case study site, New Bedford, Massachusetts. The northern estuary has been designated a Superfund site by the U.S. EPA because of a high sediment levels of PCBs. A local reference site (West Island, Fairhaven, Massachusetts) is located about 15 km away. km New Bedford, MA Acushnet River, MA (New Bedford Harbor) 15 km Reference Site: West Island Fairhaven, MA Boat Launch Coggeshall St South Northern Estuary 0 500 1000 meters Superfund Site Hot Spot Coggeshall St North ©2002 CRC Press LLC ©2002 CRC Press LLC life cycle. In addition, local populations of mummichogs exist in dense schools 41,43 of highly fecund and genetically variable individuals. 44,45 For these reasons, mummichogs have been used extensively as a model for evolutionary studies on environmental adaptation. 45,46 These attributes, and their amenability to laboratory conditions, make mummichogs an ideal species for evaluating the chronic effects of environmental stressors. 10.4.1 T OXICOLOGICAL R ESPONSES The occurrence of mummichog populations in varied estuarine environments has promoted their reputation as a hardy species. 47 However, they can be quite sensitive to DLCs. Early life stage toxicity tests for dioxin suggest that the sensitivities of seven freshwater fish species ranged over two orders of magnitude. 48 In comparison to these species, the mummichog has an intermediate sensitivity to dioxin. 49 Salomon 50 showed that adult mummichogs from reference populations demon- strated reductions in survival when exposed to dietary dioxin under laboratory conditions. Similarly, Black et al. 51 (Figure 10.6A) reported that injections of a mixture of dioxin-like PCB congeners mimicking the mixture and concentration found in NBH mummichogs produced mortalities and reduced egg production in female fish from reference populations. 51 Gutjahr-Gobell et al. 52 found that dietary exposure of DLCs to adult reference mummichogs also reduced their growth, feed- ing, and survival (Figure 10.6B). Together, these results indicate that exposure to DLCs at concentrations similar to those measured in NBH mummichogs produce toxic effects in reference mummichogs. Similarly, a recent literature review con- cluded that tissue concentrations as high as those measured in NBH mummichogs increased embryonic and larval mortality and altered neurotransmitter concentra- tions, hormone metabolism, and gonadal development in many fish species. 28 FIGURE 10.5 Sediment contamination at the Superfund site at NBH, as inferred by meas- urements of PCBs in sediment cores. Dashed line indicates sediment concentrations of PCBs associated with probable ecological effects (180 ng/g). 34 (Courtesy of J. Latimer, 2000.) ©2002 CRC Press LLC Consistent with the direct toxic effects of DLCs as assessed using short-term laboratory exposures to reference mummichogs, Black et al. 37 demonstrated increased mortality among 3-year-old NBH mummichogs held in the laboratory during a portion of the summer spawning season. However, the laboratory results for DLC-exposed mummichogs from reference populations and NBH mummi- chogs were not identical: NBH mummichogs did not exhibit reduced fecundity. 37 In addition, laboratory-held NBH mummichogs produced larvae with unique developmental abnormalities, 37 unlike DLC-exposed mummichogs from refer- ence populations. 51 FIGURE 10.6 Regressions of mortalities of female F. heteroclitus exposed to a mixture of non- ortho and mono- ortho PCBs as TEQs of PCBs in liver tissue. Fish were exposed by injection (A) or diet (B). (A, adapted from Black, D.E. et al., Environ. Toxicol. Chem ., 17, 1396, 1998. With permission.) (B, adapted from Gutjahr-Gobell, R.E. et al., Environ. Toxicol. Chem ., 18, 699, 1999. With permission.) A B – [...]... of wildlife populations and communities, Environ Toxicol Chem., 19, 1703, 2000 2 U.S Environmental Protection Agency, Framework for Ecological Risk Assessment, Risk Assessment Forum, U.S EPA/630/R-92/001, Washington, D.C., 1992 3 U.S Environmental Protection Agency, Guidelines for Application of a Framework for Ecological Risk Assessment, Risk Assessment Forum, U.S EPA/630/R-92/001, Washington, D.C.,... Environmental Protection Agency, Interim Report on the Assessment of 2,3,7,8Terachlorodibenzo-p-Dioxin Risk to Aquatic Life and Associated Wildlife, U.S EPA/600/R-93/005, Office of Research and Development, Washington, D.C., 1993 18 Safe, S., Polychlorinated biphenyls (PCBs), dibenzo-p-dioxins (PCDDs), dibenzofurans (PCDFs), and related compounds: environmental and mechanistic considerations which support the... of Rhode Island, Kingston, 1994 51 Black, D.E et al., Effects of a mixture of non-ortho and mono-ortho-polychlorinated biphenyls on reproduction in Fundulus heteroclitus (Linnaeus), Environ Toxicol Chem., 17, 1396, 1998 52 Gutjahr-Gobell, R.E et al., Feeding the mummichog (Fundulus heteroclitus) a diet spiked with non-ortho and mono-ortho-substituted polychlorinated biphenyls: accumulation and effects,... indirect effects and multiple distinct outcomes in ecological risk assessment, Environ Toxicol Chem., 17, 1640, 1998 ©2002 CRC Press LLC 65 Abraham, B.J., Species profiles: life histories and environmental requirements of coastal fishes and invertebrates (mid-Atlantic) — mummichog and striped killifish, U.S Fish and Wildlife Service Biological Report 82 (11.40), U.S Army Corps of Engineers, TR EL-8 2-4 , 1985 66... in evolutionary and population ecology, Trends Ecol Evol., 14, 467, 1999 15 Calow, P., Sibly, R.M., and Forbes, V., Risk assessment on the basis of simplified life-history scenarios, Environ Toxicol Chem., 16, 1983, 1997 16 Gleason, T.R., Nacci, D.E., and Munns, W.R., Jr., Projecting population-level responses of purple sea urchins to lead contamination for an estuarine ecological risk assessment, J Aquat... M.K and Peterson, R.E., Potencies of polychlorinated dibenzo-p-dioxin, dibenzofuran, and biphenyl congeners, relative to 2,3,7,8-tetrachlorodibenzo-pdioxin, for producing early life stage mortality in rainbow trout (Oncorhynchus mykiss), Aquat Toxicol., 21, 219, 1991 26 Peterson, R.E., Theobald, H.M., and Kimmel, G.L., Developmental and reproductive toxicity of dioxins and related compounds: cross-species... Biochemical and Cellular Perspectives, Malins, D.C and Ostrander, G.K., Eds., CRC Press, Boca Raton, FL, 1994 23 Hahn, M.E., Ah receptors and the mechanism of dioxin toxicity: insights from homology and phylogeny, in Interconnections between Human and Ecosystem Health, Di Giulio, R and E Monosson, E., Eds., Chapman & Hall, London, 1995 24 Hahn, M.E., Mechanisms of innate and acquired resistance to dioxin-like... Munns, W.R., Jr et al., Evaluation of the effects of dioxin and PCBs on Fundulus heteroclitus populations using a modeling approach, Environ Toxicol Chem., 16, 107 4, 1997 54 Gleason, T.R., personal communication, 2000 55 Prince, R and Cooper, K.R., Comparisons of the effects of 2,3,7,8-tetrachlorodibenzo-p-dioxin on chemically-impacted and non-impacted subpopulations of Fundulus heteroclitus: I TCDD... toxicity of 2,3,7,8-tetrachlorodibenzo-p-dioxin to seven fresh water fish species during early life-stage development, Environ Toxicol Chem., 17, 472, 1998 49 Toomey, B.H et al., TCDD induces apoptotic cell, death and cytochrome P4501A expression in developing Fundulus heteroclitus embryos, Aquat Toxicol., 53, 127, 2001 50 Salomon, K.S., Dietary Uptake of 2,3,7,8-Tetrachlorodibenzo-p-Dioxin and Its Effects... contaminant-tolerant and contaminant-sensitive fish populations, presented at the Environmental Toxicology and Chemistry Annual Meeting, Nashville, TN, November 2000, Abstr PWA124 77 Roark, S.A., Guttman, S.I., and Nacci, D., Allozyme analysis of the relationship among contaminant-tolerant and contaminant-sensitive populations of Fundulus heteroclitus, presented at the Environmental Toxicology and Chemistry . (Figure 10. 10A and B). Similarly, mum- michog embryos from NBH are profoundly less sensitive to the lethal effects asso- ciated with DLC exposures than reference fish 38 (Figure 10. 10A and B) congeners are non- and mono- ortho substituted congeners whose toxic potency can be evaluated relative to 2,3,7,8-tetrachlorodibenzo- p -dioxin (TCDD or dioxin) using a toxic- equivalency. Responses 10. 4.4 The Scale of Evolutionary Effects 10. 4.5 Risks of Selection and Adaptation 10. 5 A Population Modeling Approach and Case Study: Conclusions Acknowledgments References 10. 1 INTRODUCTION