Development of a Practical Modeling Framework for Estimating the Impact of Wind Technology on Bird Populations

Report

Title: Development of a Practical Modeling Framework for Estimating the Impact of Wind Technology on Bird Populations
Publication Date:
November 01, 1997
Document Number: NREL/SR-440-23088
Pages: 29

Document Access

Website: External Link
Attachment: Access File
(587 KB)

Citation

Morrison, M.; Pollack, K. (1997). Development of a Practical Modeling Framework for Estimating the Impact of Wind Technology on Bird Populations. Report by California State University and North Carolina State University. pp 29.
Abstract: 

The goal of this project is to develop a useful, practical modeling framework for evaluating potential wind power plant impacts that can be generalized to most bird species. We accomplish this by (1) reviewing the major factors that can influence the persistence of a wild population; (2) briefly reviewing various models that can aid in estimating population status and trend, including methods of evaluating model structure and performance; (3) reviewing survivorship and population projections; and (4) developing a framework for using models to evaluate the potential impacts of wind development on birds.

 

We begin with a review of demography. Demography is the study of population statistics, including births, deaths, immigration, and emigration. From demography, we know that conditions leading to extinction are most likely to occur in small populations. Demographic rates vary because individuals do not survive for the same length of time, individuals vary in the number of offspring they bear, individuals often have low birth rates, and so forth. Adult survivorship is usually very high, especially in long-lived species (such as raptors). Therefore, estimating adult survivorship tells one a lot about population status. In addition, in most monogamous species, it is female survivorship that is most important to population persistence. At a minimum, then, quantifying adult survivorship provides a preliminary, basic indication of the status of the population. Modeling genetics is not likely to be as important as modeling demographic and ecological processes in evaluating population persistence. This is based, in part, on the lack of our sufficient understanding of genetics to use it as a basis for management. Thus, practical considerations were the overriding factor in this conclusion. Still, genetics may be a priority in small, isolated populations.

 

Random environmental events such as catastrophic fires, hurricanes, and disease can also have pronounced effects on small populations. Such factors can also have pronounced effects on large populations that are spatially divided into subpopulations. Here, factors such as dispersal will determine the fate of a subpopulation driven to very low numbers, or even to extinction, by a catastrophic event. Thus, the relative importance of environmental stochasticity must be based on an understanding of the spatial distribution of the population under study.

 

Next we review the parameters necessary to develop rigorous population-projection models. Life-history parameters are essential components of population-projection models. The characteristics that we collectively call life-history parameters of animals include quantifiable longevity, lifetime reproductive output, the young produced per breeding attempt, the age of dispersal, survivorship, sex ratio, and the time between breeding attempts. For example, combining various ranges of parameters can yield substantially different rates of population change. Such analyses provide guidance on whether the population can be sustained under varying expressions of life history traits. Once such relationships are understood, researchers have the opportunity to monitor selected life history traits as part of an assessment of the status of a population.

 

A central part of impact assessment--such as in wind power plants--is developing a model that estimates the survival rates required to maintain a constant population. The strategy is to determine the survival rates required to sustain the populations that exhibit the various combinations of the other parameters governing population size. To be useful in a wide range of environmental situations and useable for people with varying expertise, the model should be based on simple mathematics.

 

Leslie matrix and similar stage-structured models can give great insight into the processes of population growth. For example, the sensitivity of the population growth rate, r, to perturbations in vital rates for a Leslie-type model can be solved analytically. Understanding how growth rate changes in response to perturbations at various stages in the life table may help direct management strategies. For example, adult survival tends to be a parameter to which a model is extremely sensitive in long-lived species, whereas fecundity can be more important in short-lived species.

 

To aid in providing general guidelines concerning the potential impacts of wind developments on bird populations, we developed Leslie matrix models and conducted sensitivity analyses to determine the effects of survival of age classes on population growth rates. We gathered data from the literature on passerines, ducks, geese, gulls, and eagles. These analyses provide a first approximation of how populations of these types of birds respond to hypothetical changes in fecundity and survivorship. They can be used to help focus attention on species most likely to be adversely affected by changes in fecundity and survivorship.

 

The simplest models assume that the number of animals in a population goes up or down by a constant ratio, usually designated as lambda, with each unit of time. The annual geometric growth rate of a population is thus represented by lambda, which is also known as the finite rate of population increase. The population is increasing if lambda >1, is constant if lambda = 1, and is decreasing if lambda <1. For example, if lambda = 1.04, then the population was growing at the rate of 4% during the time period sampled.

 

With the models in place, each survival rate parameter was allowed to vary from zero to one while the remaining parameters were held constant. The new value of lambda was calculated at each new value of the changing parameter. Once these data were obtained for each survival parameter, the results were plotted on a graph so as to see the different effects each parameter had on the population growth rate. This can be viewed as a way of expressing the sensitivity of lambda to the different survival parameters.

 

The curves for the passerine show that lambda is much more sensitive to changes in the juvenile survival rate than to changes in the adult survival rate. Also, the juvenile survival rate curve has a very steep slope as the juvenile survival gets very small. The curves for the duck show that lambda is roughly equally sensitive to changes in the juvenile and adult survival rates. For geese, the nonadult age classes survival rates seem to have little impact on the value of lambda. For the adult age class, lambda is extremely sensitive to changes in the adult survival rate. For gulls, except for very small survival rates, the changes in the adult age class gives the largest change in lambda. The other classes all have very similar curves. The situation for the eagle is very similar to the situation for the gull, but even more extreme: there is great sensitivity of lambda to changes in adult survival rate.

 

One of our objectives was to evaluate the use of surrogates, or indices, of survival and population trends. We found a highly significant negative relationship between adult survival and annual fecundity. This analysis indicates that fecundity might be a suitable surrogate for survival in passerines and woodpeckers. This does not imply, however, that fecundity is a suitable indicator of abundance (i.e., increasing fecundity does not necessarily compensate for lower survival). Raptors will leave poor habitat (e.g., low food availability), often moving many kilometers in search of a suitable nesting site. In addition, raptors tend to change territories more often when nesting is unsuccessful. Thus, as a generality, constancy of territory occupancy seems to be an indicator of good habitat quality in raptors. The number of nonbreeding, adult "floaters" in an area is an indicator of the general health of the bird population. This holds if territory availability is constant or increasing. Additionally, an increase in the age of first breeding, as well as an increase in adult aggression, are possible indicators of a population at or above carrying capacity. In long-lived species with delayed age at first breeding, such as in many raptors and some waterbirds, changes in survival rates have a greater effect on the population than changes of similar magnitude in reproductive rates. Thus, the use of reproductive success in long-lived species as a population indicator should likely be supplemented with other indicators, such as territory occupancy and floater individuals. The use of surrogates that we recommend here is not designed to determine the cause of population change. Rather, surrogates are intended to only identify that change has occurred; whether or not such change is caused by wind development will usually require more rigorous research (e.g., field work and experimentation). Surrogates serve primarily as a coarse filter to help narrow the scope of subsequent research.

 

We also present a series of steps that can be used to develop a strategy for evaluating the influence of a project on a bird population. Based on our review it seems that the appropriate hierarchical framework for evaluating population responses to perturbations is: (1) empirical data, (2) surrogates, and (3) model with available data (Leslie matrices). A large set of empirical data is, of course, the optimal situation.

Find Tethys on InstagramFind Tethys on FacebookFind Tethys on Twitter
 
CAPTCHA
This question is for testing whether or not you are a human visitor and to prevent automated spam submissions.