This blog is the first in a series of modules that I will present on using simulations to evaluate experimental designs. The series will mostly consist of fairly simple simulations that I have built over the last decade to answer questions about experimental design and analysis for my own research as well as the research of my colleagues. I have found simulation exercises to be a tremendous tool for learning about methods in applied ecology and my hopes are that whoever reads this will enjoy it and perhaps take away some useful ideas and skills that will improve their research as well.
Often in ecological research we start with a question and then propose an experimental design to address the question. The experimental design is usually made up of a sampling design and a statistical analysis; however, it is typically unknown, a priori, how the design will perform for answering the question of interest. Continue reading
Recently I taught a class on estimating fisheries exploitation rates with tagging studies. A common study design to estimate exploitation rates is to release tagged fish into the system and allow anglers to return the tags when they harvest a fish. The simplest estimate of exploitation rate is obtained by dividing the number of tags returned by the number of tagged fish released into the fishery. However, three key sources of uncertainty are mortality of fish due to the tagging process, tag loss, and non-reporting by anglers. Studies can be designed to estimate these quantities by combining both passive and active tags (e.g. Kerns et al. 2013) or performing additional side studies. Often tag loss, tag reporting, and non-reporting rates are adopted from other studies and systems and used to correct exploitation rates for these unknowns. The following method is a parametric bootstrap that incorporates external estimates of tag loss, tag mortality and non-reporting rates into the exploitation rate estimate. Furthermore, it incorporates the uncertainty in these quantities using the bootstrap procedure. Continue reading
Posted in fisheries management, recreational fisheries, Uncategorized
Tagged ecological statistics, exploitation rate, Fish, Fisheries, fisheries management, fisheries science, mark-recapture, National Marine Fisheries Service, Recreational fishing
Abundance of animals in an area (i.e. density) is one of the most basic population parameters desired in applied ecology. Many methods have been developed to estimate abundance within an area such as mark-recapture; however, it is often unclear the exact area within which the sampled population is contained. The issue is that animals move. If you imagine that each animal in the population has a region of activity (i.e. home range) within which it moves, you can further imagine how animals whose region of activity existing on the boundaries of the sampling area may move in and out of the sampling area between sampling events (Figure 1).
Figure 1. An example of a sampling reach in a river. The distributions over each fish represent their home range with the dashed vertical line representing the center of activity.
Posted in fisheries management, population modeling, Uncategorized
Tagged ecological monitoring, ecological statistics, environmental monitoring, Fish, Fisheries, fisheries management, fisheries science, mark-recapture, Population model, simulations, Statistical power
Black basses such as largemouth bass and smallmouth bass support popular recreational fisheries across North America. In addition to length limits and bag limits, bass stocks are often managed with seasonal closures during spawning, particularly in their northern regions. Spawning season closures are viewed as important regulations for bass (Quinn 1993, 2002) because they are nest brooders, with the males guarding the eggs and fry until they become free swimming. There is a wealth of literature demonstrating that a nest guarding male can be very aggressive and therefore more vulnerable to being caught by recreational anglers (e.g., Suski and Philip 2004). When removed from the nest, the eggs are susceptible to predation by other fishes and the stress on the captured male can lead to nest abandonment even when released. It is therefore believed that this disruption can lead to an overall reduction in the population level fecundity for that year and perhaps lead to reduced fish stocks (Suski et al. 2003). Whether the protection of fecundity through spawning season closures is effective at protecting fish stocks depends on the strength and timing of density dependent compensation in juvenile survival. Continue reading
Describing the growth of fish is important for population modeling, particularly when evaluating policy options for fisheries management. The most common way to describe the growth of fish is with a von Bertalanffy growth curve, but estimates of growth parameters can be bias when specimens are collected with size-selective sampling gears. Age-length samples have been corrected for size selective sampling, but these methods typically required multiple years of data, tagging information, or independent estimates of size selectivity of the sampling gears and are, therefore, not feasible for some studies. Continue reading
Recently I coauthored a paper in the journal Fish and Fisheries comparing the relative performance of harvest slots verses minimum-length limits as policies for regulating recreational fisheries. The paper is in response to the paucity of evaluations of harvest-slot regulations and the apparent default use of minimum-length limits to prevent overfishing. The aim is to provide general guidance to managers of a range of recreational fisheries. The novelty of the paper is that we evaluated the performance of harvest slots in terms of multiple fisheries and conservation metrics, across a range of fish life-history strategies, for moderate and high-effort fisheries, and across a range of fisheries objectives.
Full text copy can be downloaded by clicking here. Continue reading
Recently I have been involved with the evaluation of two monitoring program designs where the objective of the program is to detect when a metric has crossed a specific threshold that is used to trigger management. For these programs the question of how 80% statistical power can be determined was asked. Determining statistical power within the context monitoring thresholds has not always been immediately obvious to me so I built a simple simulation and wrote a summary to demonstrate how I formulated the problem.
In the context of monitoring when the abundance of a population has crossed a threshold, statistical power is the probability of the abundance estimate in a given year reaching or exceeding the monitoring threshold when in fact the true population has, in reality, reached or exceeded this value. Thus, we will correctly conclude that the management trigger has been met. A key difference between this way of thinking about statistical power and more common approaches is that we aim to detect when a threshold has been crossed rather than detecting a difference between two samples. Continue reading