link to AFSC home page

link to AFSC home page link to NMFS home page link to NOAA home page

Resource Ecology & Fisheries Management (REFM) Division

AFSC Quarterly
Research Reports
July-Aug-Sept 2008
Contents
Feature
ABL Reports
FMA Reports
HEPR Reports
NMML Reports
RACE Reports
REFM Reports
Items
Quarterly Index
Quarterly Home

Economics & Social Sciences Research Program

Predicting Fishing with Vessel Monitoring System (VMS) Data

The National Marine Fisheries Service has expanded requirements that vessels fishing in the Pacific cod, Atka mackerel, pollock, and other fisheries own and operate a vessel monitoring system (VMS). The system sends each vessel’s location to NMFS every 20-30 minutes while the transmitter is operating. The VMS consists of two parts. A transmitter/receiver, installed on the vessel, queries GPS satellites and downloads vessel position, as well as estimates the heading and speed. The transmitter then sends these data to NMFS via the ARGOS (Advanced Research and Global Observation Satellite) system of polar orbiting satellites.

Though the VMS tells NMFS the location of each participating vessel, it does not directly determine whether the vessel is fishing or not. However, when a vessel is fishing, its course and speed are generally different from when the vessel is simply transiting an area. These differences produce a "signature" that indicates fishing is taking place. The nature of a given vessel’s signature depends on many factors, including the gear type used (trawl, hook-and-line, or pot), the type of vessel deploying the gear, and the length of time the vessel spends fishing. In addition to VMS, many vessels carry a NMFS-certified observer during 30%-100% of their days at sea. Thus, NMFS can determine directly and independently whether or not fishing is taking place and can thus corroborate whether a given signature indeed demonstrates that fishing is taking place.

The primary purpose of this research is to determine the extent to which the signatures can be used to accurately predict whether fishing is occurring or not. In previous work by Dr. Pat Sullivan for the NMFS Alaska Region, a number of techniques were explored to predict fishing for a select number of vessels. This current project builds upon that exploratory work and develops an operational algorithm. To the extent that a given signature can accurately predict whether fishing is taking place, NMFS will use the signatures to develop computer algorithms that will automatically predict whether a given vessel is or was engaged in fishing operations. The predictive power of the developed algorithms can be expressed as a percentage of predicted fishing events that correspond to actual fishing events. Functions of lagged speed and bearing have been developed which predict spatial effort with relatively low error. Preliminary results from this work were presented at the Fourth International GIS/Spatial Analysis Symposium this summer. Final results are being prepared for publication.

By Alan Haynie and Patrick J. Sullivan


Measuring the Value of a Statistical Life in the BSAI Crab Fisheries

The value of a statistical life (VSL) is revealed by one’s choices regarding monetary returns and fatality risk. Estimates of the VSL have been extensively used to inform public policy and quantify preferences for environmental quality, health, and safety. To date little attention has been paid to investigate the tradeoffs associated with the returns from natural resource extraction activities and the risks incurred, and the effect that changes in safety regulations and the utilization of property rights management has had on the VSL. In this research we model fishing captains' discrete choices to fish on a given day or not, conditional on the observed risk present, in the Bering Sea and Aleutian Islands crab fisheries. We instrument for risk using fatality data from crab and noncrab fisheries, as well as information on wave height, wind speed, air temperature, and sea temperature (to capture ice risk propensity). We examine how the VSL and fatality risk have changed since the inception of both the U.S. Coast Guard (USCG) preseason boarding program and the BSAI crab rationalization program.

Notably, our estimation framework controls for the inherent sample selection bias present in many VSL estimates and when compared to those estimates which do not control for sample selection, illustrate the substantially upward biases which may arise. In addition, these estimates are robust to heterogeneous preferences which can bias homogeneous estimates of the VSL. By expanding our utility theoretic model to reflect captains' preferences for others' well-being, we may be able to recover the value of an altruistic life via the unique data generating process present within these fisheries. Preliminary empirical estimates provide a measure of the captain's value for crew lives and decomposing our estimated VSL (which includes the value of an altruistic life); we were also able to recover the captain's implicit value of his own life.

These estimates may be used to benefit contemporary fisheries policy. For instance, recently The New York Times reported that the fatality rates within the Pacific Northwest Dungeness crab fisheries possessed fatality risk rates that were 60 times greater than the average American worker. This article pointed out that the dockside safety program does not check every vessel participating in this fishery, whereas in Alaska the USCG preseason boarding program does. Presumably the cost of the program may be one of the reasons why this dockside boarding program is not perfectly executed. However, given our range of estimates for the VSL and our estimated reductions in the fatality risk resulting from the complete coverage of the USCG preseason boarding program, our model predicts that the annual benefits derived from this policy within the BSAI crab fishery are quite substantial.

By Kurt Schnier, William Horrace, and Ron Felthoven
 

<<< previous

next >>>


            Home | FOIA | Privacy | USA.gov | Accessibility      doc logo