Final Report: Evaluating the Impact of Multiple Stressors on Common Loon Population Demographics - An Integrated Laboratory and Field ApproachEPA Grant Number: R829085
Title: Evaluating the Impact of Multiple Stressors on Common Loon Population Demographics - An Integrated Laboratory and Field Approach
Investigators: Meyer, Michael W.
Institution: Wisconsin Department of Natural Resources
EPA Project Officer: Packard, Benjamin H
Project Period: October 1, 2001 through March 31, 2005 (Extended to November 25, 2005)
Project Amount: $490,759
RFA: Wildlife Risk Assessment (2001) RFA Text | Recipients Lists
Research Category: Ecological Indicators/Assessment/Restoration , Ecosystems , Biology/Life Sciences
In this research project, an intensive effort was focused on collecting information on the abundance, distribution, and health of common loons inhabiting an area of Wisconsin where lakes vary in Hg contamination and degree of habitat alteration and human disturbance. A rigorous field sampling scheme was used to produce a random sample of loon breeding pairs, from which the common loon population density and critical population demographic parameters (adult survival, fecundity, and juvenile recruitment) were quantified. Concurrently, the impact of stressors on these parameters was quantified. Several stressors are likely associated with impacts on loon demographic parameters including Hg exposure, habitat quality, and human disturbance.
The goals of this project were to: (1) conduct research to improve predictions of loon population dynamics in regions impacted by multiple stressors, (2) advance techniques for assessing the relative risk of Hg exposure and other stressors on loon populations in the Upper Midwest United States, and (3) predict the population level benefits of reducing or controlling the impact of stressors on loon populations. The specific objectives of the research project were to: (1) estimate the population size of adult common loons in an 8,600 km2 region of northern Wisconsin impacted by Hg deposition, habitat alteration, and human disturbance; (2) quantify loon population demographic parameters within this study area, including adult survival, fecundity, and juvenile recruitment rates; (3) assess the impact of stressors (Hg exposure, habitat alteration, human disturbance) on the measured demographic parameters; and (4) predict the population-level benefits that could be achieved by reducing stressors demonstrated to impact demographic parameters
The results include a Common Loon Projection Matrix Model, which provides output predicting population growth rates as a function of estimates of loon population abundance, demographic parameters, and impacts of stessors on the parameters.
Field surveys were conducted from May 1 to September 30 in 2002, 2003, and 2004 to estimate the loon population density and population demographic parameters (adult survival, fecundity, and juvenile recruitment) within the area of risk. Geographic information system (GIS) tools were used to create an 8,600 km2 sample grid, which was subdivided into 344 25-km2 cells. Cells were stratified into four categories of loon nesting habitat quality based on surface water area (hectares). Cells were randomly selected (22 in 2002, 32 in 2003, 36 in 2004) from each strata and all lakes greater than 4 ha within or intersected by the cells were surveyed to count and map the number and location of loons present. There are a total of 1,582 waterbodies greater than 4 ha within the study area; 4 ha is the minimum size lake used by territorial loons in the risk assessment region. The dual frame sampling method (Haines and Pollock, 1998) was used to estimate population abundance. Demographic parameters (adult survival, juvenile recruitment, fecundity) were estimated via reobservation of color-marked individuals and weekly surveys of lakes occupied by territorial loons, which quantified nest success and chick survival. A two-stage site specific Wisconsin Loon Projection Matrix model was developed that produces an estimate of the annual growth rate of the loon population within our study area in 2002-2004.
The impact of Hg exposure on demographic parameters was assessed in the field and laboratory and integrated into the population model to evaluate the benefits of reducing Hg exposure on the loon population growth rate. Hg risk is established on the basis of exposure (as predicted by laboratory-derived pharmacokinetic model and as indexed by lake pH and loon egg, blood, and feather Hg concentrations in the field) and compared to thresholds of effects determined in controlled studies with common loons. Effect thresholds are established via controlled dosing studies of loon chicks and loon eggs in the laboratory as well as field correlations of Hg exposure, reproduction, and survival.
Loon habitat models were constructed to evaluate the potential impact of habitat alteration (associated with human settlement) on loon population growth rates. First, spatially explicit variables such as shoreline housing densities were incorporated into the habitat model, along with lake morphometrics, water chemistry, and clarity for 377 randomly selected lakes. Shoreline habitat was mapped into categories of nest habitat quality. Logistic regression analysis then evaluated the relationship between these lake features and loon presence/absence and reproductive success. An index of human-related disturbance rate also was developed via passive observation and quantification of significant disturbance events during incubation and chick rearing.
Finally, we quantified the benefits of Hg exposure reduction to the loon population within the risk assessment region by predicting the change in annual growth rate that will occur if fish Hg concentrations are reduced to the No Observable Adverse Effect Level (NOAEL), 0.1 μg Hg/g fish wet weight. Also, we assess the current and predict the future population-level risk posed by habitat alteration as a result of human shoreland housing development with the risk assessment region.
This research project is the first to demonstrate the population level benefits of Hg exposure reduction for a wildlife species using the multistressor approach. Simultaneously, we predict the risk posed by habitat alteration caused by a record rate of housing construction on lakes within the study area between 1990-present. To attain this information, the size of the population at risk is estimated, a site-specific projection matrix population model is developed, the levels of stressors within the risk assessment region are quantified, the impact thresholds for the stressors are determined, and the population benefits of stressor reductions are predicted via changes in the calculated annual growth rate. To our knowledge, this is the first risk assessment project to use this novel approach.
Project collaborators at the U.S. Environmental Protection Agency (EPA) National Health and Environmental Effects Research Laboratory Atlantic Ecology Laboratory, Narragansett, Rhode Island, are applying this approach to the loon population in New England; however, they are using different data sources. Upon completion, results of both projects will be compared, and when appropriate, results will be pooled to strengthen the predictive power of the models.
It is estimated that there are approximately 1,200 adult loons present within our 8,600 km2 risk assessment region (2003 estimate); 80 percent are pair-bonded, and 20 percent are floaters. The two-stage Wisconsin Loon Projection Matrix Model developed for the loon population within the risk assessment region using demographic parameters measured in 2002-2004 predicts an annual growth rate of 1.0133 (± 0.0008). This slow rate of growth also is predicted by the 5-year LoonWatch Wisconsin Loon population estimate 1995-2005, providing an independent estimate of growth (1.2%/year) very similar to that predicted by our population model (1.3%/year).
We have demonstrated that there is no discernable impact of Hg exposure on adult loon survival in New England and Wisconsin, though the approach was unable to detect differences of 1-3 percent in adult survival rates that could be demographically significant. Controlled laboratory dosing studies conducted by research colleagues at the U.S. Geological Survey (USGS) La Crosse in 1999, 2000, and 2003 have allowed for establishment of the Lowest Observable Adverse Effect Level (LOAEL) for Hg to be 0.4 μg MeHgCl/g diet for a critical life-stage of the common loon, the growing chick from hatch to day 105. No toxic effects were measured when chicks received diets containing 0.1 μg MeHgCl/g diet fed, thus establishing the NOAEL for loon chicks. In addition, USGS LaCrosse and Wisconsin Department of Natural Resources (WDNR) are conducting an ongoing loon Hg egg injection study to quantify the level of MeHg in eggs that causes reduction in embryo survival. Preliminary results in 2005-2006 show a significant reduction in egg hatching rate when 1.3 μg MeHgCl/g wet weight egg is injected into loon eggs via the air cell 3 days post-laying. These eggs have background levels of 0.4 μg Hg/g wet weight, thus our preliminary prediction is that loon egg hatching rates will be lower when eggs contain greater than 1.7 μg MeHg/g wet weight. This 3-year experiment is ongoing. The final year of this experiment will occur in 2007 or 2008 pending funding, and final results will be available upon completion. Other controlled dosing studies, including an ongoing egg injection experiment at USGS Patuxent, have found significant reductions in avian embryo survival when egg Hg concentrations exceed 1 μg/g. A pressing research need is determination of comparative toxicity of mechanically injected MeHg versus maternally deposited MeHg in ovo.
The Wisconsin Loon Habitat model was developed to assess the population level impacts of habitat alteration caused by rapidly increasing human settlement within the risk assessment region. Logistic regression analysis shows several variables are strong predictors of lake use by territorial loons and reproductive success. These parameters include lake size and shape, presence of good nest habitat, high water clarity, and lake depth. No indices of human settlement (shoreline building density or boating activity) were statistically significant predictors of impaired lake use by loons in our risk assessment region. Many examples of loon pairs habituating to high levels of human disturbance were evident during the study. A plot of shoreline housing density versus loon pair presence and reproductive success, however, does show no lake use by territorial loons or reproduction success when shoreline building densities exceed 25 buildings/km. This level of building density was found on only 5 percent of lakes surveyed during our study in 2002-2004, but the potential exists for this level of density to be achieved on 67 percent of lakes in our risk assessment region, with potentially large population consequences for loons. Current Wisconsin shoreland zoning (WDNR NR 115) permits a density of up to 33 buildings/km of shoreline (e.g., 1 building/100 ft of lake shoreline). As 67 percent of lakes within our study area have less than 10 percent of shoreland in public ownership, these lakes all are at risk of exceeding 25 buildings/km shoreline if completely subdivided into 100 foot x 300 foot parcels, the minimum size currently allowed by zoning regulations.
We found that approximately 10 percent of the loon population within the Wisconsin risk assessment region exhibits Hg exposure levels that are consistent with toxicity in the laboratory. The toxicity impacts are most likely to affect chick survival and nest success via reduced egg hatchability. The potential benefits of reducing fish Hg concentrations to the NOAEL (0.1 μg MeHg/g whole fish wet weight , 4-8 inch yellow perch) was simulated by increasing chick survival and nest success rates 10 percent and recalculating the annual growth rate. We estimate the annual growth rate of the loon population within our risk assessment region will increase 1.5 percent if fish Hg levels are reduced to less than or equal to 0.01 μg/g. The population-level improvement could be even greater in loon populations receiving higher Hg exposure, such as the population sampled at Kejimkujik National Park, Nova Scotia, in 1999 (Burgess and Hobson, 2005), where loons have 2-5 times higher Hg levels than in Wisconsin. A simulation exercise was conducted, and it is estimated that reducing fish Hg concentrations to the NOAEL could increase the annual growth rate of Kejimkujik population by up to 5 percent; however, density dependent limiting factors and essential habitat requirements will mitigate actualized growth rates. The final calculation of benefit for the Nova Scotia population awaits quantification of site-specific population demographic parameters and Hg exposure profiles.
Despite the fact that we measured no effect of the current level of human settlement on lake use by loons or habitat quality, a large portion of the risk assessment region is at risk to future impacts. This is because much of the lake shoreline in the risk region is in private ownership, subject to subdivision into 100 ft x 300 ft parcels (as allowed by current statewide zoning regulations). Current market pressure for undeveloped shoreland in the risk region is enormous, with vacant land on some lakes selling for $1,500 - $2,000 per linear foot of lake frontage (up to $250,000 for vacant parcels on some lakes). This potential monetary windfall, plus the increase in property appraisals and consequent tax burdens, has led many lakefront property owners to subdivide large properties into smaller parcels. The only constraints on development currently are statewide shoreland zoning (WDNR NR 115) and local zoning ordinances at the county and township level.
We simulated the potential population level impacts of development by displacing 10 percent and 30 percent of the current breeding loon population in our study area. We reduced the proportion breeding in the fecundity parameter of the Wisconsin Loon Projection Matrix Model and calculated changes in annual growth rates. These reductions result in a 1 percent and 3 percent reduction in annual growth rate. It is probable that a 30 percent reduction in the proportion of adults breeding would lead to a fairly rapid population decline, despite the long lifespan of loons. Further, it is likely the decline would be exacerbated by density dependent factors impacting additional demographic parameters, including decreased adult survival rates, nest success rates, and chick survival rates. Displaced adult loons will fight to the death to acquire breeding territories, as well as cause nest abandonment and intentionally kill the young of resident pairs. If a large proportion of the existing breeding population is displaced because of degradation of nesting and chick-rearing habitat quality, it is likely these other demographic parameters will also decline, potentially resulting in a dramatic population decline. Proactive measures that protect habitat quality can prevent these scenarios from playing out. Wisconsin is currently revising the NR 115 Shoreland Management rules. To conserve common loons within our risk assessment region, an effort should be made at the state, county, and township level to protect common loon nesting habitat from degradation (primarily lakeshore wetland habitat and islands) to promote water clarity by controlling nonpoint sources of nutrient runoff and by promulgation of zoning rules that reduce shoreland housing density to 1 building/200 ft of frontage on all lakes of potential use by breeding loons where there is no shoreland in public ownership. Our surveys show that lakes with greater than 10 percent of shoreland in public ownership still are used by territorial loons despite high density housing development on the privately owned shore. These protected shoreland areas can provide secure nest sites and low disturbance areas for early chick rearing.
We noted a high rate of adult loon mortality caused by ingestion of lead fishing tackle within the study area in 2000-2005. Eleven dead adult loons, in body condition allowing for necropsy, were turned in to our group by the public in 2000-2005. Lead poisoning was the cause of death of 5 of these 11 loons (WDNR Health Team, 2006). In New England, lead fishing tackle has been banned in some states because of the high rate of mortality associated with ingestion of lead tackle by fish-eating birds. Two private, nonprofit organizations, Wisconsin LoonWatch, Ashland, Wisconsin, and the Raptor Education Group, Inc., Antigo, Wisconsin, have initiated a lead fishing tackle exchange program where they provide anglers nonlead alternatives when they turn in their lead fishing jigs and sinkers. This effort should be supported at the state level because of the importance of adult loon survival to the future status of loons in Wisconsin.
Stakeholders and Users of Data and Results
Assessing the ecological risk of Hg exposure to piscivorous wildlife is a priority issue for federal and state resource management agencies and industry alike. Currently, Hg emission reduction rules have been promulgated by EPA, but the rules are being contested in court by several states that contend that emission reductions will not achieve benefits as quickly as needed. Our establishment of a fish NOAEL for loons of 0.1 μg Hg/g whole fish wet weight, as well as documentation of population level benefits to be achieved by reaching this target, will be of interest to these parties.
Wisconsin Department of Natural Resources and other state agencies with a management responsibility to protect lakes from degradation will be interested in our assessment of the potential impact of housing development on loon populations. The extrapolation of future risk via model simulation under current shoreland management rules argues for proactive measures to protect key habitat features as well as promotion of less dense housing concentrations on lakes with no shoreline in public ownership. Habitat protection recommendations made here can safeguard the future status of loons in the region, while accommodating additional human population growth.
Natural resource agencies and conservation groups will be interested in the common loon population and habitat models generated by this project, as well as management recommendations that are presented.
How this Project Furthers Natural Resource Science and Management
This project demonstrates that the approach of assessing multiple stressors at the population level is a viable and important new tool in the science of ecological risk assessment and management. Traditional wildlife contaminant risk models often are compromised by a lack of relevant toxicological data from the laboratory and from the field, resulting in the use of numerous uncertainty factors. In addition, measurement of impacts most often are made at the individual level and only occasionally extrapolated to population level effects. This project delivers a scientifically defensible ecological risk assessment for Hg for wildlife, based on an at‑risk species, a combination of laboratory and field studies that establish species-specific thresholds of risk, and a population model that predicts costs and benefits of increases or decreases in stressor levels. This risk assessment paradigm is amenable to many stressor scenarios and species at risk. We recommend its consideration by other scientists in the risk assessment and management field.
Haines DW, Pollock KH. Environmental and Ecological Statistics 1998;5:245-256.