This page provides a short introduction to population viability analysis. It is extracted from the textbook Applied Population Ecology by H. R. Akçakaya, M. Burgman and L. Ginzburg.
Population viability analysis (PVA) is a process of identifying the threats faced by a species and evaluating the likelihood that it will persist for a given time into the future.
Population viability analysis is often oriented towards the conservation and management of rare and threatened species, with the goal of applying the principles of population ecology to improve their chances of survival. Threatened species management has two broad objectives. The short term objective is to minimize the risk of extinction. The longer term objective is to promote conditions in which species retain their potential for evolutionary change without intensive management. Within this context, PVA may be used to address three aspects of threatened species management:
Components of population viability analysis
There is no single recipe to follow when doing a PVA, because each case is different in so many respects. Below we discuss some of the main components that a PVA might have. Not all PVAs will have all these components, and some will have others that are not discussed here.
Identification of question
Any scientific inquiry starts with a question, and population viability analysis is no exception. The question is likely to change in the course of a PVA. Initially, the question might be very general, such as "Is this species threatened, and if so, why?" The less we know about the species, the more general the questions will be. At this step (Step 1 in the figure) a PVA should concentrate on the identification of factors (including natural factors and human impacts) that are important in dynamics of the specific populations and metapopulations under study, as well as conservation and management options. The methods to be used for this depend on the specific case at hand, and might include statistical analysis of historical data, comparison of populations that are declining with those that are stable, and correlating recent changes in the environment (climatic or habitat changes, introduced species, changing harvest patterns, etc.) with changes in the species.
After the available information about the ecology of the species and its recent history is collated and reviewed, the questions are likely to become more specific. Examples of such questions include
Determining model structure
The identification of the problem and the specific management options determine the model structure to use (Step 2 in the figure). The most appropriate model structure for a population viability analysis depends on the availability of data, the essential features of the ecology of the species or metapopulation, and the kinds of questions that the managers of the population need to answer. For more information, see A Short Introduction to Modeling, the online papers listed below, and a list of common mistakes in modeling.
Estimating model parameters
The next step is to estimate the model parameters with field studies (and sometimes experiments). The kind of parameters that need to be estimated will depend on the model structure, and the type of data already available (see a list of common mistakes in modeling).
For most PVA studies, this is the limiting step, because data are often insufficient. However, if a decision will be made no matter what, it is better if the decision-maker has some input from a PVA, even if the data are not perfect. If a parameter is not known very well, then a range of numbers can be used for that parameter instead of a single number. For example, if the average dispersal distance of California gnatcatchers is about 3 km, but is not known accurately, we can use a range of 2 m to 4 km. These ranges can be used in a sensitivity analysis (see below).
Running the model
Building a model is a method of combining the existing information into predictions about the persistence of species under different assumptions of environmental conditions and under different conservation and management options (Steps 4 and 5 in the figure). When building a model, it is important to keep a list of assumptions made.
The structure of the model and the questions addressed usually determine how the results will be presented. In most cases, the model will include random variation (stochasticity), which means that the results must be presented in probabilistic terms, i.e., in terms of risks, probabilities or likelihoods.
Risk curves (see figure for an example) provide a convenient way of presenting results of a simulation. The example risk curve gives the risk of decline of a population as a function of the amount of decline. The right end of the figure (100% decline) represents extinction. In this example, there is about 10% risk of extinction. According to this figure, the probability that the population will decline by 60% or more in the next 50 years is about 0.32, and the probability that it will decline by 40% or more is about 0.65.
Often, the model must be run many times, with different combinations of the low and high values of each parameter to make sure that all uncertainty in parameter values is accounted for. This provides a way to measure the sensitivity of results to each parameter. Sensitivity analysis (Step 6) is useful for determining which parameters need to be estimated more carefully.
If, for example, the risk of decline is very different with the low value and high value of adult survival rate, then the results are sensitive to this parameter, and we can conclude that future field studies should concentrate on adult survival rate in order to estimate it more accurately. This feedback from modeling to field work is represented by an arrow from Step 6 to Step 3.
When simulations include those with different management options, sensitivity analysis also gives information about the effectiveness of these options. For an example, see "Future directions" in the California Gnatcatcher Study.
Implementation, Monitoring, Evaluation
With the selection of the best course of action under a given set of conditions (Step 7), the function of modeling is completed, but only temporarily. The next step is the implementation of the plan (Step 8). It is important that the field studies continue during and after the implementation to monitor the species (Step 9). The results of monitoring can give valuable information about the response of the species to management, as well as provide more data to refine model parameters and improve the model (Step 10). For example, we might discover, upon evaluation of demographic data after the implementation of the plan, that gnatcatcher vital rates increase faster than predicted in response to removal of cowbirds, but the carrying capacity responds slower than expected to the improvement of the habitat. Such a finding would definitely require modifying the model, refining its parameters, and re-estimating the extinction risks under different management options.
PVA ExamplesBird Modeling Studies at Applied Biomathematics
Applications of RAMAS to specific cases (list of references, including pre-prints and reprints of publications)
Species Conservation and Management: Case Studies (edited volume of 37 case studies)
Akçakaya, H.R. 2002. Estimating the variance of survival rates and fecundities. Animal Conservation 5:333-336.
Brook, B. W., M. A. Burgman, H. R. Akçakaya, J. J. O'Grady, and R. Frankham. 2002. Critiques of PVA ask the wrong questions: throwing the heuristic baby out with the numerical bathwater. Conservation Biology 16:262-263.
Brook, B.W., J. J. O'Grady, A. P. Chapman, M. A. Burgman, H. R. Akçakaya, R. Frankham. 2000. Predictive accuracy of population viability analysis in conservation biology. Nature 404:385-387.
Akçakaya, H.R. 2000. Population viability analyses with demographically and spatially structured models. Ecological Bulletins 48:23-38.
Akçakaya H.R. and P. Sjögren-Gulve. 2000. Population viability analysis in conservation planning: an overview. Ecological Bulletins 48:9-21.
Akçakaya, H.R. 2000. Viability analyses with habitat-based metapopulation models. Population Ecology 42:45-53. (The original publication is available at http://www.springerlink.com)
|PVA software · Prices · Training · What's New · Forum|
|Research · Support · Index · Contact Us · Home|
|©1999-2005 by Applied Biomathematics|
Date modified: 2-20-05