Turning back the demographic hands of time for an endangered species

Steve Beissinger discusses his and Zachariah Peery’s article “Reconstructing the historic demography of an endangered seabird” in the February 2007 Ecology (88(2): 296-305). Read the full article at http://www.esa.org/esablog/blog_docs/Beissinger_article.pdf.

It’s a simple question that I often get asked about an endangered species: “What caused it to decline?” but I find it to be one of the hardest to answer without giving a hand-waiving response. Determining causes of decline for a species based on data-driven conclusions rather than informed opinion is challenging because it first requires figuring out which demographic rate is depressed and then requires evidence linking it to one or more causes. Yet, to provide clear recommendations for recovering a threatened species, is there any more meaningful question to answer than what is causing it to decline?

I like to use an analogy with medicine to understand how ecologists can produce credible evidence to begin to diagnose the cause of population decline. Demographic rates are basic measures of population health for a declining population. Estimating demographic rates is equivalent to taking the temperature and blood pressure of a patient. Deviations from “healthy” birth and survival rates are symptoms that inform us how forces causing population declines are acting.

Unlike medicine, ecologists have no set of established reference points for healthy demographic rates for threatened species that have no thriving populations. Until about 30 years ago, systematic studies of the birth and death rates of wildlife species were rare. That changed after 1972 with the passage of the Endangered Species Act, which required biologists to assess a species’ risk of extinction. It would be incredibly useful if we could turn back the hands of time and estimate demographic rates for species when their populations were presumably healthy.

We tried to do exactly that in our paper ( Ecology 88:296-305) by using age ratios of museum specimens collected 100 years ago to reconstruct rates of reproduction and survival for the Marbled Murrelet ( Brachyramphus marmoratus ), an endangered seabird that flies inland to lay its single egg in depressions on mossy, platform-like limbs 50-120 m above the ground in old growth forests of the Pacific Northwest, Canada and Alaska. Because nest sites are usually impossible to detect from the ground, this species was among the last North American birds to have its nest discovered in 1974. Thought to have declined from deforestation, the murrelet became a cause-celebre for conserving old growth forests, but oil spills, gill-netting, and nest predators were also considered to be limiting factors.

Over the past decade, studies in my lab on Marbled Murrelets were designed to determine why their populations have declined, examining both terrestrial and marine causes. My former Ph.D. students Ben Becker and Zach Peery deserve much of the credit overcoming considerable logistical difficulties involved with studying murrelets. They and their assistants launched a 4 m inflatable zodiac from beaches through the surf at night to catch, band, and attach telemetry units to murrelets in the ocean using spotlights and dipnets, and during the day to conduct population surveys. They tracked the movements of telemetered birds day and night from airplanes and trucks. They sampled prey and measured trees. Why do graduate students get to have most of the fun?

Results from our work provided evidence that murrelets were suffering from reproductive failure. Most telemetered birds did not nest apparently from food limitations, those that did nest had high rates of failure, and we saw few juveniles on the water during our surveys. Our models and comparisons with other species suggested reproductive rates were greatly depressed. Still, they were just models, and their assumptions and demographic estimators were subject to challenge. We lacked the smoking gun of the expected rates of reproduction for a healthy population. Age ratio analysis of museum specimens collected 100 years ago in our study area provided it.

Age-ratio analysis estimates birth and survival rates from ratios of the numbers of individuals in different age classes. It has fallen in and out of vogue in ecology, and recent applications in Ecology rekindled useful debate over the validity of demographic estimates derived from this approach because critical assumptions may be violated. So we knew we had to rigorously examine the potential for bias that could arise from errors in aging individuals, population fluctuations, and trapping or collecting bias, and to analyze our field data for evidence of differential habitat use or movements by age classes. We even dug out the notebooks and correspondences of the collectors to assess their methods and motivations. We wanted to set the bar high.

What strengthens our work is not just the comparison of historic estimates of birth and survival rates with contemporary estimates, but comparisons with predicted rates from life history theory using allometry, which also provides benchmarks for demographic rates under pre-decline conditions. When demographic estimates for murrelets from museum skins and life history theory showed strong concordance, and when most assumptions of age ratio analyses were upheld for murrelets, we developed confidence in the reliability of our conclusions. Evidence from multiple analyses triangulated on the result that birth rates for Marbled Murrelets were 8.5 times greater about 100 years ago than they are today, while adult survival has not changed substantially.

Reconstructing the demography of a threatened species prior to its decline using age-ratio analysis of museum specimens can provide unique insights for recovering threatened species. Although it can only work for species with identifiable ages or stages, museum collections for a broad array of taxa are available spanning time scales that could not be studied with conventional field techniques. The power of this approach is magnified when combined with habitat, genetic, isotope, disease, or other studies of museum specimens, which we have done and are currently doing for murrelets. Our confidence in the estimates derived from age-ratio analysis increases greatly when they are combined with ancillary data to support its assumptions.

With this information in hand, the next step is to diagnose the specific causes of decline. This involves testing mechanisms or limiting factors causing changes in demography (and population trends) by: (1) experimentally manipulating them and measuring demographic responses; (2) modeling how populations should respond to environmental changes that have occurred; (3) comparing the levels of candidate causes for populations with different demographic rates or population trends; (4) examining demography or population trends before and after changes in a limiting factor; or (5) assaulting sets of predicted impacts on behavior or demography of competing candidate factors with multiple forms of field data.

Contributed by Steven R. Beissinger, University of California at Berkeley

For additional stories on this paper, see also:

UC Press Release:

Sidebar on murrelet conservation

TV coverage: