The Bay Area Stream Fish database contains historical and modern information about native and introduced fish in the San Francisco Bay Region. The data was collected by Region IX of the U.S. EPA and the San Francisco Estuary Institute. The database will evolve into the first on-line mapping application created by the San Francisco Estuary Institute. Stream survey station data of numbers and lengths of 37 species of fish will be enhanced by a new USGS National Hydrography High-Resolution Dataset and USGS DRG and DOQ image base maps.
In 1992, the U.S. Environmental Protection Agency awarded a grant to UC Berkeley for an inventory of streams and fishes in the San Francisco Bay basin. These funds were provided by Congress to the San Francisco Estuary Project for early implementation of the the Comprehensive Conservation and Management Plan. This project represented an expansion of an earlier landmark study published by Mr. Robert Leidy in the journal HILGARDIA, Distribution and Ecology of Sream Fishes in the San Francisco Bay Drainage (October 1984).
Between 1992 and 1998, Mr. Leidy, a wetlands scientist with the USEPA, revisted many of the sites he had surveyed in th 1980s and explored new locations that were not previously accessible. These sites profiled mainstems and tributaries in both rural and urban settings.
The data resulting from this extensive research are intended to:
This database contains information about native and introduced fishes collected by Mr. Leidy on Bay Area streams from April 1992 to July 1998. The database was designed and built by the San Francisco Estuary Institute, in collaboration with Region IX of U.S. EPA, in 1999-2000. The database reports on the numbers and lengths of a total of 37 species of fish observed at 263 stations on 79 Bay Area streams.
Selected fields from the original project field datasheets were compiled into a relational database using Microsoft Access. An ArcInfo point coverage was also created by SFEI to represent the sample stations spatially. SFEI is currently developing an internet mapping application to make the Stream Fishes data easily accessible. Additional components, including historical data, are also being developed to extend the utility of the database. U.S. EPA, SFEI, and other local, state, and federal partners will be exploring how these data can be well-used, and updated, to support ongoing watershed management. This database is the most comprehensive to-date for Bay Area stream fishes. However, its limitations should be recognized. Please see the "Data Use" section for more information. The original datasheets recorded additional data about stream habitat, fish parasites, amphibians, and other parameters, as described in the documentation (see "Data Collected but Not Compiled"section below). The data selected for entry into the database were chosen by U.S. EPA based on input from local interest groups. The modular design of the database should permit incorporation of additional data as needed in the future. For further questions about the Fish Survey 1992-1998 data, please contact Mr. Robert Leidy of U.S. EPA at 415-744-1970, or Leidy.Robert@epamail.epa.gov.
A local assemblage of fishes may be defined as those fishes that would be found together at one particular place or locality. Fish assemblage structure (e.g., number of species, distribution or abundance of species at a locality, number of prey species versus piscivorous species, proportional composition of the assemblage by trophic groups or functional groups, body-size patterns for an assemblage, etc.) may vary considerably over time (i.e., morning to night, season to season, year to year) and space (e.g., riffle vs. pool, high-elevation headwater stream reachversus low-elevation reach near the stream mouth). Therefore, caution should be exercised in the data interpretation and any subsequent application of the data to define both broad- and fine-scale ecological patterns. Specific attention should be given to the limitations of the survey objectives and methods of this study when extending the data to other applications. For example, data on the distribution and abundance of fishes in this study was collected during the years1992-1998. However, for many streams sampling was limited to a single sampling event. Any conclusions drawn from the data must therefore be drawn within the context of the natural temporal and spatial variability of stream ecosystems, as well as the objectives and methods of the study design.
Field data were collected by Robert A. Leidy of U.S. EPA and others under his direct supervision. Data were collected between April 1992 and July 1998. A total of 277 survey events took place at 263 stations on 79 creeks. Information collected during each survey event was reported on an individual, two-sided field datasheet.
The primary goal of the fish survey was to document the present distribution and abundance of native stream fishes in the Bay Area. Stream locations known to contain native fishes based on historical records were surveyed to document local species distribution and abundance. Additional locations were surveyed where historical information was incomplete. At each location, an effort was made to survey the full range of habitat types (e.g., riffle, run, pool) within representative stream reaches and geomorphic settings (e.g., high-elevation, high-gradient, bedrock controlled; low-elevation, low gradient, alluvial unconsolidated bottom, etc.). Stations were selected to maximize the diversity of habitat types in different geomorphic settings. Stream reaches were typically sampled above the influence of diel tidal fluctuations, though a few exceptions occurred.
Selection of fish survey techniques depended on site-specific physical characteristics and conditions. The method employed was the one that would survey a site most thoroughly. Fish surveying relied on one or more of the following techniques: electrofisher, minnow or beach seine, dip net, gill net, and snorkel and mask.
Surveying with a Smith-Root Type XI backpack electrofisher was the most often employed method, primarily in reaches with depths of less than 1 meter (m) and water velocities of less than 3 second-feet. Electrofishing was conducted in a downstream-to-upstream direction for a minimum distance of 30 m, except in two situations.
First, many survey locations were characterized by isolated pools less than 30 m in length, which were embedded within a matrix of dry stream bed. In these situations single or multiple pools were surveyed. Although an effort was made to survey a minimum of 30 m, survey distance often exceeded this.
Second, distances of greater than 30 m were electrofished when either no, or few, fishes were collected within the first 30 m surveyed. This additional effort was directed at certifying the presence or absence of species. An effort was made to survey all habitats within a reach with equal effort (i.e., survey time and area surveyed), however, habitats immediately adjacent to stream banks often received more intensive surveying because these areas typically provided the most complex habitat for fishes, and therefore required the most attention. Block nets were used at the upstream and downstream ends of the survey location if physical conditions (e.g., high water velocities, poorwater clarity, etc.) warranted their use.
Electrofishing was conducted by either 2 or 3 person crews, with all members of the crew collecting stunned fish by dip net. Fish surveying at depths greater than 1 m was typically accomplished by minnow and/or beach seine, gill nets, or snorkeling with a mask. Surveying effort varied among these methods although most surveys conformed to the minimum 30 m survey distance. A student minnow seine was used to depths of 1 m while depths of between 1 m-2 m were surveyed with a beach seine. Student and beach seines were constructed of 6-mm mesh and both were 1.5 m x 2 m and 2 m x 4m in depth and width, respectively.
Very shallow habitats (typically < 5 cm depth), where electrofishing and larger nets were not effective were surveyed with either small-size mesh dipnets (4-6 mm mesh) or fish were identified in-situ. Depths greater than 2 m were typically surveyed with gill nets (13- and 19-mm mesh nets, 2 m deep and 10 and 20 m in length) or by in-situ observation with a snorkel and mask. Gill nets were set for a period of two hours.
The time of day of net set and retrieval varied. A mid-water trawl was used to survey sites in the Napa and Sonoma Marsh complex.
For each collection site the following information was recorded: survey date and time; streamname(s) and narrative location; survey station (numbered sequentially for all sites surveyed onthe same date); county name; and USGS 7.5 minute quadrangle name. The following data were collected: mm fork length (FL) or total length (TL), unless more than 30 individuals of a species were caught, in which case a representative 30 fish were measured and additional information were summarized with a narrative comment; the total number of species; the percentage native fish species in the collection. Estimates were made of the abundance in the sample of young-of-the-year, salmonids, and sculpins (0-3 scale, where 0 = no individuals, 1 = 1-24 individuals, 2 = 25-50 individuals, and 3 = >50 individuals); percentage of sample consisting of native species byabundance (1-3 scale, where 0 = no native fishes, 1 = 1-33% of the total number of individualscollected were native, 2 = 34-66% of the total number of individuals collected were native; and 3= 67-100% of the total number of individuals collected were native).
A number of additional types of data were collected in the field, but have not been incorporated into the database. These include:
Estimates and measurements of substrate composition, water depth, and wetted-channel, bankfull, and floodplain widths were taken at 9-15 points at equal intervals of 3 measurements per transect along 3-5 equally spaced transects, set up perpendicular to stream flow. Percentage substrate composition was first visually estimated independently, and then confirmed collectively, by two observers centered on 1 square meter quadrat at each survey point. Channel confinement was calculated as average flood plain width (m) divided by average bankfull width (m). An estimate of the number of Oncorhynchusmykiss mykiss/O. m. irideus per 30 m of stream based on captured fish and visual counts of fish were made at each survey site. Amphibians that were captured or observed at each site were alsonoted by species. Unless otherwise noted, all habitat inventory protocols in this study followed those described in Flosi and Reynolds 1994.
All fish were identified using one or more of the following references: Moyle 1976 (as per 1997 key revisions); Lawrence and Burr 1991. All collected fish were released, with the exception of a few voucher specimens, which were preserved in 10% formalin. Each survey site was photo-documented by taking a minimum of 2 photographs with a wide angle lens; one looking up stream and one looking downstream from a location at the centerpoint of the site.
The BASF project has a research component to analyze historical conditions of salmonid populations in the San Francisco Bay Region. The following methods have been used:
Certain problems will need to be addressed to begin to use and interpret the historical data alongside the modern data. The common thread that unites the past and present views is the connection between species and stream, but,
The Access database for the Robert A. Leidy Fish Survey 1992-1998 was designed at SFEI by Samir Arora, with the assistance of Jessica Sisco, Robin Grossinger, Zoltan Der, Julie Ekstrom, and Jung Yoon. Robert Leidy and Josh Collins supervised the interpretation of the datasheets and reviewed data entry and presentation format. The database was modeled off the field datasheets. The design consists of three major groups of linked tables. The first group stores information about species nomenclature, such as scientific name, common name, family name, native designation, and phylogenetic order. The second group of linked tables stores information about the survey stations and other location information, such as stream, watershed, county, and respective USGS 7.5 minute quadrangle. The final group of linked tables stores the survey event data, namely, collection date, collection time, list of observed species and associated measured fork or total lengths, and any survey collect or assigned species ratings. As a group, this latter set of tables serves as the link between the other two sets, the species nomenclature group and station location group. Data were transcribed from copies of the original datasheets during January to May 1999. The datasheet copies were sorted and organized by SFEI and assigned a unique ID. Each datasheet represents one survey event.
The master list of fish species and families was provided by Robert Leidy. This list represents possible species within Bay Area streams, and not just those surveyed in the 1992 to 1998 fish survey. Species information about each survey event was taken from the field datasheets. All surveyed fish have been made to conform to the above master species list. Instances of no observed species at survey event have been flagged in the database. Cases of no individuals or no measured lengths are coded as -9 in the database. All field collector-assigned species abundance ratings were taken directly off the field datasheets. The master list of stations and associated information was compiled and organized by SFEI based upon the field datasheets. Each station has been assigned a unique ID.
Station location narrative descriptions were transcribed off the field datasheets with minimal editing. Stream names were taken primarily from USGS 7.5 minute quadrangle maps. For sample stations in tidal sloughs, the corresponding Baylands Segment (Goals Project 1999, p.95) is identified.
The point, stream and base image GIS layers supporting the relational database were designed by Zoltan Der at SFEI. GIS production at SFEI was peformed by Christina Wong, Ellen Carney and Zoltan Der. The USGS team working on the NHD and DOQ development was Tom Sturm, Vicki Lukas and William Kaiser.
Survey stations represented 20m-50m reaches of streams and were marked and annotated on 7.5 minute USGS quadrangle maps with pencil crosshatches and transfered to an ArcInfo point coverage using head-ups digitizing over registered Quadsheet images. Cumulative error (using the root-mean-squares method) is 27m. This error is within the range of station ground location lengths. It should be noted that error resulting from the relationship between the points marked on the USGS quadsheet and their actual field position was not assessed by SFEI. It should also be noted that the most detailed record of the survey location is the narrative decription recorded in the "Location" field in the database. This field describes the specific location and extent of the survey station in relation to local landmarks. The GIS coverage represents the general location of the survey stations in a digital form, but does not replace the specific narrative location description.
When the BASF project began, there were no San Francisco Bay Region stream spatial datasets with sufficient resolution to show all the streams in the survey. In August of 1999, SFEI began working with USGS to develop a high-resolution National Hydrography Dataset for the San Francisco Bay Region. 128 DLG-3 hydrography files were edgematched by SFEI and transferred to USGS for final processing into the NHD High-Resolution standard. SFEI will use stream names and Reach Codes from the NHD for the initial versions of the BASF internet mapping application. When the NHD for the region is completed, SFEI will act as the regional data steward, updating the BASF stream layer and providing transactional updates back to the USGS.
SFEI also worked with USGS on obtaining and processing DOQs and DRGs for use in the GIS and web production steps. These images (stored as georeferenced MrSID images) will be used as base layers in the mapping application.
Current internet mapping solutions were studied with the goal of immediate off-the-shelf useability, short development time and web standards support. Esri's ArcIMS product was chosen as the best available tool to deliver the BASF data.
An on-line survey was posted on the SFEI website in October of 1999, 150 resource managers, environmental scientists, and creek activists were emailed a request to complete the survey; a total of 45 surveys were completed. Survey results were used to design the application. Overall design goals were to provide a stable, fast, and easy to use internet mapping application to project funders and other regional agencies and to members of the public with current ArcIMS supported browsers. The application will be imbedded within a standard web site design to allow for maximum access to related information.
SFEI is currently working with St. George Consulting to build the BASF internet mapping application. St. George will assist with data migration to ArcSDE and Oracle, and application and associated help system development.
SFEI will serve the non-spatial related tables, line streams and point sampling stations data from Oracle with ArcSDE running on a SUN Enterprise 250 server, the DOQ MrSID image data from a LINUX server, to an ArcIMS map server running on Netscape Enterprise Server on the SUN.
Project funders will begin testing the site in August and September of 2000. SFEI expects that the site will be ready for the general public by the end of 2000.
Geographic Information System Manager
San Francisco Estuary Institute
180 Richmond Field Station
1325 South 46th Street
Richmond, CA 94804
510-231-5791
510-231-9414 (fax)
zoltan@sfei.org
Assistant Environmental Scientist
San Francisco Estuary Institute
180 Richmond Field Station
1325 South 46th Street Richmond, CA 94804
510-231-5742
510-231-9414 (fax)
robin@sfei.org