The use of indicators, triage, and clinical decision rules to assess data-limited fish populations

点击翻译按钮获取中文摘要

In 2010, nearly 38.5 million people derived economic benefits from wild-capture fisheries, with 3 billion people obtaining 20% of their animal protein from these fisheries. Despite this global importance, the conservation status is unknown nearly 85% of the harvested taxa. This lack of status assessment is largely due to the lack of data on these populations. In order to assess the status of fish populations, scientists attempt to apply sophisticated statistical methods to data on the volume of catch, results of scientific surveys, size and age distributions from the catch and the surveys, etc. The goals of these assessments are to estimate the exploitation rate for the population as well as the current abundance, often in terms of biomass (B), relative to the abundance at which the stock is most productive in the long term, which is referred to as the biomass that will produce the maximum sustainable yield (Bmsy). Many of the management conservation triggers and population status determinations in fisheries are based on the ratio, B/Bmsy. For the unassessed populations, however, much of the required data does not exist so traditional assessment methods cannot be applied. Data on the volume of catch taken from a population is often the only data available for these data-limited populations. As such, there has been a great deal of effort to develop methods to assess populations solely on this catch data, which has in turn generated quite a bit of controversy. The goal of these methods is to produce a value for either B/BMSY or the exploitation rate relative to the optimal rate, and the quality of these methods has been determined by comparing their output to either known values from simulated data or estimates derived from more complicated assessment models. However, when it comes to these data-limited populations, managers are mostly concerned about whether the population is above or below some specified threshold value (e.g., B/Bmsy= 1.0) or whether the population is increasing or decreasing; the specific value for B/Bmsy is rarely of concern for these data-limited populations. Medical science has been dealing with very similar questions since the early 19th century, first in the form of battlefield triage and more recently in the form of clinical decision rules. In the triage paradigm, a suite of diagnostics (e.g., breathing, bleeding, loss of limbs, etc.) is assessed essentially simultaneously, and a final score, called the risk score, is generated. The risk score assumes that the results of all the diagnostics are available; there is no missing data. The risk score is then used to place patients into different categories of severity. Clinical decisions rules use more of a step-wise hierarchical structure to help doctors determine when additional tests may be necessary and help them balance the overall probability of different types of incorrect decisions (false negatives and false positives) This research program proposes to adapt techniques from medical triage and clinical decision-making to develop multiple-indicator frameworks to determine the status and trend in status for fish populations when trade-offs between false positives and false negatives are asymmetric. The development such frameworks and their ability to assess the false positive and false negative rates associated with the decisions in these frameworks, will help fisheries scientists and managers make more informed decisions about the need for action and the value of increased data acquisition. While the focus of these methods is on data-limited fish populations, the lessons learned will be applicable to data-limited situations beyond the marine realm, particularly in the area of endangered species.