|
|
|
Abstracts
Workgroup: Aquaculture and Hatchery Issues
Improved Forecasts and Dissemination of Oyster Fisheries Closure Lines
Principal Investigator(s):
James E. Mitchell, Louisiana State University, jmitche@lsu.edu
Co-Investigator(s):
E. Conrad Lamon, Louisiana State University
Funding Period: 1999
Louisiana accounts for almost 40 percent of the Nation's oyster production, leading all other states. Production is from state owned water bottoms, including public grounds and almost 9,000 privately held leases.
The Department of Wildlife and Fisheries (DWLF) administers the leasing program and manages the public grounds. The Department of Health and Hospitals (DHH), charged with safeguarding public health, conducts an extensive water quality monitoring program and uses the results to prepare seasonal forecasts for closure areas in which oysters may not be harvested because predicted fecal coliform bacteria in the growing waters exceed safe threshold levels.
Project efforts concentrated on Hydrologic Basin 4, an estuarine unit bounded by levees of the Mississippi River and the Mississippi River-Gulf Outlet Canal, as well as Breton and Chandeleur Sounds. This is a region of major importance to the oyster industry, in which the Caernarvon Diversion Project has created dynamic water quality conditions by diverting fresh water from the Mississippi River.
The goal of this pilot project, which was undertaken in cooperation with DWLF and DHH, is to automate the procedure for (a) identifying leased areas affected by closures and (b) issuing closure notices to leaseholders and other concerned parties. This was accomplished by developing a topological model of oyster lease boundaries from records maintained by DWLF, and overlaying it with predicted closure lines determined from the water quality monitoring records maintained by DHH. An attribute field associated with each lease included the lease holder's address.
The project was accomplished in two phases. Phase I involved spatial analysis of the DWLF lease records, including their conversion from an Intergraph (computer aided design) format based on meets and bounds, to an ARC/INFO GIS system; and quality control measures to eliminate topological conditions associated with the lease boundary descriptions (such as open polygon boundaries) that could propagate through subsequent data processing steps to yield erroneous results.
Phase II concentrated on the statistical analysis and refinement of time series models used to forecast fecal coliform counts at the water quality sampling stations. In particular, these efforts investigated the application of Dynamic Linear Modeling (DLM) methods, which use an adaptive fitting procedure that takes into account linear trends, seasonal cycles, daily cycles, hydrologic variations, and natural (unexplained) variability through a Bayesian learning process. DLMs were developed for 28 locations in the pilot area.
PROJECT PUBLICATIONS:
Conference Paper: Beck, Michael, 2002. Forecasting fecal coliform contamination in Louisiana oyster leases using a dynamic linear model. Presented at the 2002 AWRA Spring Specialty Conference on Coastal Water Resources, New Orleans, LA May 13-15, 2002. LSU-R-02-xx.
Unpublished Paper: Beck, Michael, 2001. Forecasting fecal coliform contamination in Louisiana oyster leases using a dynamic linear model. Special Problem Series No. SP-01-237, Dept. of Experimental Statistics, LSU, Baton Rouge, LA.
![[Maryland Sea Grant]](../mdsg_logo.gif){kind=logo}
![[NOAA]](../noaa_logo.gif){kind=logo}
![[Virginia Sea Grant]](../vasg_logo.gif){kind=logo}