A Method for the Design of Fixed Time-Area Closures to Reduce Salmon Bycatch
Salmon bycatch in the U.S. Bering Sea pollock fishery has reached record levels in recent years, and the NPFMC has recently considered implementing time-area closures that would attempt to reduce salmon bycatch. To assist in this process, Dr. Alan Haynie has written a paper that offers a discussion of important issues for consideration in marine closure design and develops and implements a methodology to identify potential candidate closures.
The starting point for the design of closures in this analysis was to determine whether or not there are any time and area combinations that, if closed, would have reduced bycatch. A fundamental assumption of this methodology is that vessels reallocate effort from closed areas to open areas proportional to other effort. For example, if there were only three areas with one-third of the catch caught in each area, closing one area would lead to half of the catch being caught in each of the two areas that remain open. This is very different from assuming that the pollock effort vanishes with a closure and it means that in order for closures to be effective, there must be low-bycatch fishing areas available at the time of the closure. Of course, depending on which areas are closed, the proportional reallocation assumption may be limiting. We discuss this assumption in greater detail in the paper but believe that it is a good first approximation. Temporally, we consider closures lasting 2-8 weeks and spatially from 1 to 10 ADF&G statistical areas.
The results of this method may be considered "optimal" in the sense that it considers all of the potential area closures that could be created (using data from 2001-2006) and then presents the costs of salmon avoidance, in terms of both the size of the closure (in number of areas) and in the proportion of pollock catch reallocated by the closure. We use ArcGIS to identify neighboring areas and Matlab to systematically explore the bycatch reduction from different closures. "Inferior" closures, where fewer salmon are avoided for the same or greater relocation cost, can be eliminated from consideration, and policy makers are offered a range of closures that represent different policy trade-offs of salmon reduction and avoidance costs. The most effective of the closures here reduced bycatch by approximately 10% per year, on average.
Given the significant size of the most effective closure, nine statistical areas, this is a small reduction, which demonstrates the limitations of static time-area closures in the context of dynamic target and bycatch populations. This work was presented at the Fourth International GIS/ Spatial Analysis Symposium this summer and final results are being prepared for publication.
By Alan Haynie
Climate Change and Changing Fisher Behavior in the Bering Sea Pollock Fishery
One component of the recently initiated Bering Sea Integrated Ecosystem Research Project (BSIERP) is a spatial economic model that will predict changes in fishing activity in the Bering Sea pollock fishery that may result from climate change. Random utility models such as the model employed here have been used in the Bering Sea and elsewhere to model how fishers make decisions about where to fish. Commercial fishers choose different areas to fish based on myriad observable and unobservable characteristics of the area and the fisher. We commonly model location choice as a function of the expected catch (or revenue) in an area, fuel and fish prices, distance to an area, vessel characteristics, and to a more limited degree, institutional and environmental conditions. In the Bering Sea pollock fishery, climate variables affect many aspects of the fishing decision. Key among these impacts is the role that climate has on fish location and abundance and the impact that weather plays in daily participation choices for smaller vessels.
In this project, we are working to expand a robust spatial economic model to include climate data (e.g., ice cover, sea surface temperatures, wind). Including this information in the model will allow us to determine the relative impact of observable contemporaneous environmental conditions on location choices. We will also develop a framework to include predictions of changing pollock abundance in the model, which will allow us to estimate fisher response to scenarios developed by oceanographic and ecosystem modelers involved in the BSIERP project. An overview of the model and data to be utilized in this paper was presented in Gijon, Spain, in May 2008 at the PICES/ICES Conference on the Effects of Climate Change on the World’s Oceans.
