• 20 July 2005 (050720)--Newport, OR
  Discussed database structure and datasets that should be integrated from GLOBEC and BPA datasets. Ancillary datasets that will be useful in defining/interpreting salmon habitats. Developed a timeline for near-term database development. Discussed integration of data synthesis and modeling. Hongsheng Bi led a discussion about the use of Quantile regression for defining habitat characteristics. Rachel Ruppel led a discussion on using SeaWiFS chlorophyll for predicting salmon habitat. We identified some items to be distributed through a web page, and Hal Batchelder agreed to provide a home page and provide initial content. Web URL for the site will be "globec.coas.oregonstate.edu/groups/nep/synthesis/ccs_salmon/salmon_habitat_hp.html". Suzan agreed to provide some diagrams of the database relations. A PDF of the BPA Salmon database structure is HERE.
• 20 September 2005 (050920)--Newport, OR
  Received an update from Suzan Pool on data organization, ACCESS front end, and progress on integrating additional data types into NOAA Fisheries Access database. Rachel Rupple provided an overview of her preliminary findings defining chinook and coho habitat based on chlorophyll concentration. A PDF of her annotated Powerpoint presentation is HERE. Hongsheng Bi provided a summary of various statistical approaches for defining optimal and suboptimal salmon habitat: a) Logistic regression based on presence/absence, and b) zero-inflated Poisson regression.
• 29 November 2005 (051129)--Corvallis, OR
  Received an update from Suzan Pool on data organization, ACCESS front end, and progress on integrating additional data types into NOAA Fisheries Access database. Suzan demonstrated combined queries among the Pearcy, BPA GLOBEC datasets, which are maintained in separate ACCESS tables. This PDF has a description of the data sets and their status as of December 2005. Cheryl Morgan, Joe Fisher and Bill Peterson suggested identified additional quirks in the fish datasets (treatment of mass fish sets, flagging of unusual seining conditions, etc.) that Suzan agreed to work on with Cheryl and Joe to resolve. Hal asked about the availability of %water or %dry weight composition of fish, since for bioenergetic modeling, those data are better than wet-weight data for establishing weight-dependent changes in fish energy density.

  Rachel Rupple provided an overview of her examination of the use of satellite derived chlorophyll to estimate the spatial extent and distribution of juvenile Chinook and coho salmon. She presented an analysis of the assumptions, showed results that indicate that "area" of suitable habitat for chinook and coho is variable among the years studied (1999-2004), but that years of abundant habitat in Washington are also years of abundant habitat in Oregon (no differences between the states). She concluded the presentation with 1) chlorophyll can be used to predict the distribution of salmon, but, by itself, cannot explain the trends and patterns in abundance; 2) the proportion of shelf habitat that is suitable for salmon is greater off Washington than off Oregon, but there was consistency in the temporal trends in suitable habitat on the shelf in the two states, and 3) coho and chinook juveniles have different habitat preferences (levels of suitable chl-a). The link between the area of suitable habitat and growth rate or survival rate has not been done yet. A PDF of her Powerpoint presentation is HERE. Hongsheng Bi, using his logistic regression model, showed a map that suggested that 2005 was a year when juvenile salmon habitat was very limited off Oregon and Washington. A PDF of his presentation can be found HERE.

  Hal summarized the bioenergetic modelling. Software has been purchased. Initial modeling will be done for coho salmon in 2000, the first of the two intensive years of GLOBEC sampling. Prey compositions for juvenile coho have been summarized by cruise (lumping all size fish from all regions within a cruise). Initial wet-weights of coho from the June cruise are ca. 75 g WW. In August, fish were about 400-450 g WW. Energy density information for most prey has been found; a few still need ED values. Eventually, these models will be run in Matlab, rather than in the FishBIO 3.0 software. An issue with running bioenergetic models is that we have a very poor description of the distribution and abundance of actual salmon prey from the GLOBEC field studies. It is hoped that models will help, but modeling will undoubtably rely on surrogates for describing prey fields. This issue needs to be discussed further with Yvette Spitz and Zack Powell (who are funded to do biological modeling of lower trophic levels (LTL)).

• 12 April 2006 (051129)--Corvallis, OR
  Received an update from Suzan Pool on data organization in the ACCESS database on the NOAA site.

  Hongsheng Bi reported on his analysis of spatial analysis of juvenile coho salmon in the BPA multiyear dataset, with a focus on data from June cruises. He described some of the relationships he has found with various predictor variables (SST, chlorophyll, water mass type, local winds) and also reported on size-dependent spatial distribution of juvenile coho salmon. Wind data for the analysis came from nearby NDBC buoys.

• 12 September 2006 (051129)--Corvallis, OR
  Agenda included Pool on data status, Bi on spatial analysis and prediction, Batchelder on Lagrangian tracking and Teel on salmon genetics.

  Suzan Pool updated the group on progress on getting the early 1980s (Pearcy Seine) data into the ACCESS database. Suzan agreed to provide an updated table on compiled data to Batchelder for posting on the project web page. That table is HERE. Issue of intercalibration of secchi disks arose, since multiple diameter disks have been used from various research programs. Bob Emmett agreed to attempt an intercalibration.

  Suzan Pool also showed analysis results from agglomerative hierarchical cluster analysis and indicator species analysis that were prepared for submission to the special American Fisheries Society "NEP Salmon synthesis" publication. Six regional clusters are identified: WGOA--Western Gulf of Alaska; EGOA--Eastern Gulf of Alaska; NWVI--Vancouver Island; WA--Washington; OR-Oregon; CA-California. Ric Brodeur will be summarizing the AFS salmon synthesis effort at the Pan-Regional GLOBEC synthesis meeting in November.

  Hongsheng Bi reported on his progress in relating salmon distribution to various indices of zooplankton type or species composition, including total copepod abundance, cold water species, warm water species, etc., using generalized linear-mixed regression models. He showed maps of predicted suitable habitat areas off Washington and northern Oregon from 1998-2004, that were developed using water depth, PCA of copepods and chlorophyll-a in the model. Comparison of mean abundance of fish within suitable areas and survival rate were better for chinook than for coho. Analysis will be continued and results written into manuscript form for submission.

  Batchelder presented a progress report (PDF of PPT file available here ) on his coupling bioenergetic codes with outputs of models of ocean physics models (ROMS). He documented the problems encountered, and solutions found in obtaining accurate 3D trajectories of particles from the NEP domain ROMS model. Bioenergetic coupling has not yet been achieved. Some issues still remain, particularly with dealing with grounding of particles in downwelling, onshore flows. Work continues. A first need that is being addressed in the modeling is that there has been no evaluation of the validity/robustness of the NEP model fields--do they resemble nature?

  Daily snapshots of the NEP domain are available for years 1958-2004, and are being evaluated in several ways by Batchelder. A talk will be given at PICES in October on this.

  Dave Teel described progress on genetic tools and datasets from the salmon collected in the BPA and GLOBEC programs. A chinook baseline has been developed (using freshwater collected fish). Statistical packages are used to conduct mixture proportion analysis or to estimate a site of origin of an individual fish. For coho, a baseline is available, all GLOBEC and BPA Plume fish have been genotyped, and genetic sex has been determined. For Chinook, the GAPS (Genetic Analysis of Pacific Salmon) baseline is used, which encompasses from the Central Valley of California to Southeast Alaska, and which is a collaborative effort of many individuals. Priority for analysis will be 1) genotyping of GLOBEC fish, 2) genotyping of BPA Sept. cruise fish (esp. 2000 and 2002 GLOBEC years), and 3) May BPA cruise fish.

  We agreed to meet next in April 2007, but will also probably meet as a group in Seattle during the GLOBEC NEP SI meeting in January 2007.

This page was last updated on September 29, 2006.

Maintained by:
Hal Batchelder
COAS
Oregon State University
Corvallis, OR 97331-5503
phone: 541-737-4500; FAX 541-737-2064