Working Group for Bottom Trawl Survey Improvements Reducing Error in Area Swept Estimates
In January 2006, the Groundfish Assessment Program (GAP) formed the Working Group for Bottom Trawl Survey Improvements (WGBTSI) to assess the GAP survey methodology in the context of the best available science. The primary goal of the WGBTSI is to critically review GAP bottom trawl surveys and to make recommendations for reducing systematic errors in survey procedures and data analyses. From January to March 2007, the group focused on reviewing the components of the area swept calculation used for estimating catch per unit effort (CPUE), specifically the methods for calculating the average width and distance traversed by the trawl net.
Currently the GAP uses two different methods for calculating distance fished. The eastern Bering Sea shelf trawl survey estimates the beginning and end points of the towpath and calculates the straight-line distance between the two. All other GAP surveys estimate distance using a moving average of the actual path of the trawl tow. GAP scientist Michael Martin modeled the effects of curved bottom trawl tows and noisy global positioning system (GPS) data on distance fished calculations. Model simulations showed that systematic error could be introduced with curved or noisy towpath data when using a moving average or a simple exponential smoother to calculate distance fished. A cubic spline method, however, produced very little systematic error and was least affected by noisy or curved GPS data.
The GAP calculates average net widths using a simple mean of the acoustic net spread observations obtained from the trawl net mensuration equipment. Similar to GPS data, acoustic net spread data can be noisy, and unrealistic measurements must be filtered prior to averaging. Filtering is currently done using a standard range of netspread values that assumes outliers are symmetrical around the mean. GAP scientist Stan Kotwicki analyzed net spread data from the 2006 eastern Bering Sea bottom trawl survey and found that outlier data are not consistently symmetrical and that simply applying arbitrary bounds on net spread data can cause a systematic bias in net width averages. To minimize potential bias caused by selectively gating raw net spread data, Kotwicki is developing a more robust method of outlier rejection using sequential rejection. The method iteratively removes outliers and fits a cubic spline smoother (the same smoother used in the distanced fished calculations) to the remaining data in an attempt to eliminate bias in the resulting calculation of mean net spread.
A third area of interest within the WGBTSI was changes in trawl performance and catching efficiency during the retrieval period of Bering Sea survey trawl. When winches are engaged at haulback, the speed over ground of the trawl increases and there is a lag period before the trawl net is completely off bottom. The time lag duration can vary with the depth of tow and the skipper operating the vessel. During this period, the trawl performance is different and the net is catching at less than 100% efficiency. Increased speed also increases the distance traversed by the trawl relative to distance covered by the vessel. The WGBTSI is currently discussing ways estimate and model biases associated with the retrieval period.
By Bob Lauth
Passive Acoustics to Study Spawning Essential Fish Habitat
Fishery scientists and managers are mandated by the Magnuson-Stevens Fisheries and Conservation and Management Act (MSFCMA) to identify and describe, both in a spatial and temporal sense, those waters and substrate necessary for spawning. In the context of the new paradigm for an Ecosystem Approach to Management, it is surprising that such basic life history information is lacking for many important commercial fish species in Alaska, and that managers are tending to focus more on the “big picture” while neglecting basic areas of research.
For marine fishes in Alaska waters, the relatively new and rapidly emerging field of passive acoustics has tremendous potential for studying spawning essential fish habitat (EFH). What is passive acoustics? In the September 2006 issue of Fisheries, Rountree et al. define passive acoustics as “the act of listening to the sounds made by fishes and using that information as an aid in locating fish so that their habitat requirements and behaviors can be studied.” In the North Pacific, fishery managers and funding agencies are unfamiliar with this technology. They will need to be educated and convinced that Alaska fishes make sounds and that there are useful management applications using passive acoustics. A first step will be to determine which Alaska groundfish species actually make sounds, and then associate those species-specific sounds to behaviors such as courtship, mating, feeding, or aggression.
From 7 to 9 February, Bob Lauth attended a workshop on Underwater Passive Acoustic Monitoring for Remote Regions. The workshop was sponsored by the Alliance for Coastal Technologies (ACT) and was conducted at the Hawaii Institute of Marine Biology. The workshop provided an opportunity to meet other research scientists in the field and learn more about the passive acoustic tools being used. By collaborating with scientists and technology developers, Lauth was able to build a low cost (<$1,000) underwater passive acoustic device for listening to fish. He is currently working with the RACE Fisheries Behavioral Ecology Program in Newport, Oregon, and the Alaska SeaLife Center in Seward, Alaska, to record fish sounds both in situ and in the laboratory.
By Bob Lauth
Skate Nurseries Research in Eastern Bering Sea Produces New Ph.D.
On 26 January 2007, Gerald Hoff successfully defended his Ph.D. entitled “Reproductive Biology of the Alaska Skate Bathyraja parmifera, with Regard to Nursery Sites, Embryo Development and Predation.” Hoff’s research focused on characterization of an Alaska skate nursery site located in the southeastern Bering Sea, where he identified the extent of the nursery and its habitat, and included seasonal sampling to determine the timing of reproduction and egg deposition; timing of embryo hatching; and mortality sources to young skates.
The findings showed that the nursery site was small in area (<2 km2) and contained egg cases at very high densities throughout the year. Reproduction was seasonal with peak deposition events in the summer months (June-August) and low levels of egg deposition throughout the year. Embryo development time, from deposition to hatching, was protracted and estimated to be more than 3 years. Mortality sources to young skates included snails and large predatory fishes. Snails drilled holes in newly deposited egg cases and devoured the large yolk, and predatory fishes including Pacific cod (Gadus macrocephalus) and Pacific halibut (Hippoglossus stenolepis) consumed newly hatched juvenile skates in the nursery area.
To date, seven nursery sites for three different Alaskan skate species have been located in the eastern Bering Sea. All nurseries were quite small in area, were composed of a single species, and possessed high densities of egg cases. Sites were near canyon heads along the shelf-slope interface between 150 and 380 m. Upwelling along the upper slope and shelf break is highly productive, and nearly constant water temperatures year-round provide an optimal environment for successful embryo development and hatching. Mature adults utilized the nurseries primarily for reproduction. Immature, juvenile and newly hatched skates were relatively rare within nursery sites.
This research has broad implications for identification of essential fish habitat for these slow growing, late maturing, and low fecund fish species. The findings from this study will be incorporated into conservation strategies for elasmobranch species in the North Pacific as management plans are developed. The North Pacific Research Board and the AFSC supported this research.
