William L. Thompson

Monitoring vertebrate populations

Preface. Basic Concepts. Sampling Designs and Related Topics. Enumeration Methods. Community Surveys. Detection of a Trend in Population Estimates. Guidelines for Planning Surveys. Fish. Amphibians and Reptiles. Birds. Mammals. Glossary of Terms. Glossary of Notation. Sampling Estimators. Common and Scientific Names of Cited Vertebrates. Subject Index.

Chapter 5 – Detection of a Trend in Population Estimates

This chapter discusses some types of trends that might occur in population size. To be effective in detecting a trend, it is important to have some ideas about what types of trends might occur in a time series of population sizes. It describes the sources of variation that make the detection of a trend difficult because of the resulting stochasticity of the observations and also examines some statistical approaches to detecting a trend. This chapter examines some statistical approaches to detecting a trend. In addition to the most common approach that is based on regression, , it explores randomization of the data with regression and a non-parametric procedure based on ranks that are robust to violations of assumptions that are necessary for regression methods to be valid. The main objective of most biological monitoring surveys is to detect changes or a trend in the population size, survival, or recruitment of certain species. The purpose may be to link the observed decline in habitat with population size, such as reduction in the amount and patch size of sagebrush habitat with a decline in sage grouse population size. Conversely, monitoring may be performed to document that changes in grazing practices result in an increase in cutthroat trout populations.

Chapter 1 – Basic Concepts

This chapter discusses basic concepts and terminology associated with monitoring spatial distribution, abundance, and density of species in a given area during some period of time. The occurrence and spatial arrangement of a species within a defined area at a particular time are called its spatial distribution. The most basic distributional information may be obtained from previous records of trapped, harvested, sighted, or other form of documented occurrence of a given species. Validly assessing spatial distribution, abundance, or density requires either a survey or a census. A survey is a partial count of animals or objects within a defined area during some time interval, whereas a census refers to a complete count within a particular area and time period. These two terms are not synonymous, although they often are used incorrectly as such. Spatial distribution, abundance, and density are parameters, that is, they are fixed but unknown quantities within a defined area and time period. Obviously, the number and spatial distribution of animals will change over time and space; therefore, these parameters are “fixed” only over a short time within a defined space.

Chapter 7 – Fish

This chapter provides an overview of potential methods to survey fish and summarizes some appropriate sampling designs, fish collection methods, and available methods to generate abundance estimates. The goal of any sampling design is to select samples such that appropriate inference can be made to some larger target population, assuming that the sampled and target populations are the same. For example, one possible target population could be defined as all fish of a particular species in the state of Colorado. The species may inhabit different bodies of water throughout the state. It is impossible to estimate numbers of fish in every stream, so a sample of streams must be selected and this list is called the sampling frame. It is probably impossible to sample the entire length of even one stream, and so one must subsample within streams. In statistical terms, it is a multistage sampling problem. The first stage occurs when selecting which streams to survey and the second when selecting which locations to sample within a particular stream. A third stage may occur if one cannot completely count fish from a selected stream unit but must rely on incomplete counts such as from a capture–recapture or removal method.

Chapter 6 – Guidelines for Planning Surveys

This chapter provides guidelines for determining the level of effort required to conduct a survey. The goal is to allocate resources to each of the components of the survey: number of population estimates (surveys) across time; number of plots for each survey; and the amount of effort to allocate to each sampling unit in enumerating the number of animals on each plot. Allocation of effort depends on the amount of variability that one encounters for each of the three components. Because the precision of surveys is affected by enumeration variation, and because the power of the test for trends is affected by the precision of the population estimates used to construct the test, the process is somewhat iterative. The first step in undertaking any type of monitoring program is to clearly define its goals. How many species will be monitored? Will it be a baseline study whose purpose is to gather background information on a little-known species, or will it attempt to monitor a species for important decreases in abundance over time? In a multispecies scenario, funding will likely limit the number of species that can be intensively monitored, and therefore some criteria for prioritizing species will have to be used.

Chapter 10 – Mammals

This chapter provides an overview of some potential methods to survey mammals. It is organized by methods as opposed to size of the mammal that are being considered. It expects readers to browse through the material to help them decide on an approach for the mammal species in which they are interested. It also provides a dichotomous key to enumeration methods as a general aid to biologists for choosing an appropriate technique. A complete count across an entire area is rarely possible under current technology. However, technologically innovative methods have been evaluated for counting mammals, such as remote sensing techniques for ungulates and ultraviolet photography. Infrared thermal imaging has been applied to small to medium-sized mammals and to large animals. Some mammals, particularly large mammals, lend themselves to complete enumeration on the plot. The most common example of this approach is the aerial counting of large herbivores. Typically, sampling units are quadrats or strips, although irregularly shaped sampling units have been used for moose. Examples of species that were counted by observers from the air include mule deer, white-tailed deer, moose, and pronghorn. Helicopters are typically used for mule deer and moose, whereas fixed-wing aircraft often are used for animals in more open habitats like pronghorn.

Chapter 8 – Amphibians and Reptiles

This chapter provides an overview of some potential methods for surveying populations of amphibians and reptiles. Availability of funds and species of interest will largely dictate which techniques can be employed; index methods may be all that can be realistically applied if many species must be monitored concurrently. However, methods that account for incomplete detectability of individuals must be used in cases that require unbiased estimates at a specified level of precision for detecting a certain percentage change over time. Herpetofauna are particularly sensitive to weather changes, particularly temperature and precipitation. Hence, surveys should be scheduled at appropriate times and conditions to maximize the number of individuals that can be detected. Both highly trained observers and complete detectability of individuals are required for complete counts to be possible. Observers must be knowledgeable in techniques used for both searching and species identification. Further, areas to be searched must be small enough to allow a census. Area size will largely depend upon the density of the species of interest and the searching method employed. Counts of any kind should be conducted when individuals are most visible; for temperate terrestrial salamanders, this would occur immediately after warm rains during spring.

Chapter 4 – Community Surveys

This chapter discusses the community survey approach to assessing both current status and long-term trends in distribution and numbers of species in an area. One approach to assessing status of wildlife species within an area is through collection of information at the community level. Before this can be attempted, however, “community” must be defined in quantifiable terms. A community is defined as all individuals of every species of interest within one or more vertebrate groups (for example, fish, herpetofauna, birds, and/or mammals) occurring in a specified area during a particular time period. The objective of a “community survey” is to then collect data on the distribution and perhaps the abundance or density (or indices of these two) for all target species in an area. This is contrasted by a “single species survey,” in which efforts are concentrated on only one species at a time. A “multispecies survey” refers to a survey of more than one species at a time and may or may not be expanded to include a community. In theory, one would like to collect population data concurrently for all species in either multispecies or community surveys.

Chapter 2 – Sampling Designs and Related Topics

This chapter discusses methods to obtain abundance estimates once the target population has been identified and sampling frame constructed. This protocol includes a sampling plan for selecting plots and estimators that use collected data for calculating parameter estimates. Constructing a sampling frame requires selection of an appropriate shape and size for plots. One also must determine if plots will be the same size or different sizes. Both of these aspects of plot design may affect precision of estimators, which in turn affects the ability to detect a population trend. A number of factors may influence the choice of shape and size of plots, and some of these may suggest entirely different optimal designs. Choice of plot shape is influenced by the ratio of a plots perimeter to its inside area, the method for obtaining plot counts, the detectability of individuals, and underlying arrangement of individuals. All four factors may be related to one another as well as to plot shape. Perimeter to area ratio is a measure of edge effect. A plot with a large perimeter to area ratio is more susceptible to having individuals wrongly included or excluded in counts than the same size plot with a small ratio.

Chapter 3 – Enumeration Methods

This chapter describes how each of the sampling units is assessed to obtain either the actual abundance or an estimate of the abundance on each sampling unit. It discusses complete counts (censuses), indices, and enumeration methods that adjust for incomplete detectability of individuals. Efficient population estimators are needed to minimize the variance of the estimated population size of each sampling unit. Inefficient estimators will result in high sampling variation because there is a lot of noise associated with each sampling unit. That is, the population estimates all have large standard errors, and as a result, the variation among sampling units is large. The enumeration variance contributes to the overall variance of the survey, so the aim is to keep the enumeration variance as small as possible. Methods for assessing the numbers of individuals within a sampling unit can be categorized as complete and partial counts. Complete counts are a complete enumeration (census) of individuals within a sampling unit. Thus, a random sample of quadrats might be drawn, and all the individuals counted on each of the quadrats. Such counts are rarely possible in studies of animal populations.