1.) Field Data Collection methods:
The San Francisco Bay Study (Bay Study) was established in 1980 and the initial sampling design included 35 stations; 7 stations were added in 1988, 4 stations were added in 1990, and 6 stations were added in 1994. The current protocol is to sample all 52 stations (25 channel and 27 shoal stations) 12 times per year (approximately once per month; see Figure 1 for a map of stations). See sampling frequency figures in the metadata pdf for deviations from sampling effort. Every effort is made to sample within 0.5 nautical mile of the actual station location. Extenuating circumstances such as vessel traffic or inaccessibility due to tides may force a station to be sampled slightly out of geographic range. A midwater trawl (MWT) and otter trawl (OT) are conducted at each station and the crew leader records the data on a paper datasheet. A plankton net was used from 1980 to 1989 to sample larval fish and crustaceans. These methods are described separately at the end of this section.
Station and Tow Data: At each station, the crew leader records the status of waves and cloud cover. Substrate type is recorded based on the majority substrate of the OT sample. The depth is recorded during the first tow at a station and monitored throughout both tows. Sometimes an average depth is used when the depth varies. At the start of each tow when the net is actively fishing, the crew leader records the tow parameters including: net type, tide (direction of the current) , tow direction relative to the current, time of day, start latitude, start longitude, and bearing (using a GPS unit) , as well as the start flow meter reading (which is usually the end meter from the previous tow). As soon as the tow ends, the crew leader records the end latitude and end longitude (from a GPS unit), tow distance (from a GPS unit) , tow duration , and the end reading from the flow meter. Once sampling processing is complete, a catch code is recorded for each tow.
Midwater Trawl: The midwater trawl is generally towed in the direction of the current for 12 minutes, except for during periods of slack tide (when there is no movement either way in the tidal stream). Tow direction during slack tide is with the direction that the current will be turning, as established by the mate. The length of tow cable deployed varies relative to water depth; cable length is verified by the mate. The lead deckhand retrieves the net at approximately 25 feet per minute during the tow to sample the water column obliquely. Depending on the amount of cable deployed, there are scheduled stops when the deckhand pauses retrieval.
Otter Trawl: The otter trawl is generally towed against the current for 5 minutes, except for periods of slack tide in which the tow is generally against the direction that the tide will be turning. The crew lead or mate may reduce tow time to 3 minutes if a station has a history of excessively large samples or snags. The length of tow cable deployed depends on the water depth and cable length is verified by the mate. In general, the ratio of length of tow cable to depth follows the 5:1 scope ratio where the length includes the 100’ bridle.
Water Quality: Water quality measurements are taken at every station, including a salinity, conductivity, and temperature depth profile, and the Secchi depth. The boat position relative to station is verified prior to collecting the water quality measurements. An electronic CTD meter from Seabird Electronics (Seabird) is used to collect a salinity, conductivity, and temperature profile of the water column. The Seabird is hydraulically deployed from a davit on the boat’s port side. The unit is first pre-soaked by running submerged at the surface for one minute; the unit is then lowered continuously collecting 4 scans per second. When the Seabird reaches the bottom, the unit is drawn back up. A Secchi depth reading is taken on the shaded side of the boat when the boat is in neutral, and the reader has removed their sunglasses. The Secchi depth is determined to the nearest five centimeters (cm) .
Re-tows: When a tow cannot be completed or the sample is not acceptable, a re-tow can be conducted. Reasons for a re-tow include a shortened tow due to vessel traffic, an untied cod end, large holes in the net, evidence that the net did not open or fish properly, or other evidence of an incomplete sample. When a re-tow is conducted, the original sample and net must be checked for listed species and their catch is include in the regular ESA reporting. If the net was snagged or contained very large/heavy objects, the net is closely checked for any damage and is repaired or replaced as needed.
Sample Processing: The sample is emptied into a tub and any large fish are measured and released right away (e.g., bat rays, sharks, sturgeon). The sample is placed on the sorting board and crew members remove any debris and sort the remaining fish and crabs into buckets of water and gelatinous zooplankton (jellies) are placed a pile on the sorting board. Shrimp are sorted into a separate bucket. When sorting is complete, the remaining organisms are processed. Jellies and Northern Anchovies are only processed from the midwater trawl. Crabs and shrimp are only processed from the otter trawl . For fish, up to 50 randomly selected fish of each species are measured to the nearest millimeter for length (fork length or total length for fish without forked tails) and the rest are counted (referred to as “plus count”). Wing width is measured for Bat Rays. Sharks are measured from the tip of the snout to the tip of the tail (i.e., total length). The minimum fish cutoff length is 20 mm , with the following exceptions: Northern Anchovy = 40 mm, Longfin Smelt = 40 mm, Delta Smelt = 30 mm, Pacific Herring = 30 mm, Catfish = 30 mm, Striped Bass = 25 mm . Fish below these cutoff lengths are too small to be sampled consistently by the mesh size used. Some cutoff lengths were established since 1980 and therefore some prior fish length data are in the database below these thresholds. As Longfin Smelt and Delta Smelt are listed species, any lengths less than the minimum cutoff and greater the 20 mm are also recorded. These fish are not entered into the Bay Study database but are reported as ESA take. Crabs are identified to species. For all Cancer crabs, Portunus xantusii, and Eriocheir sinensis up to 30 randomly selected individuals of each species are sexed and the carapace width (CW) is measured to the nearest millimeter with calipers; the remaining crabs are plus counted. CW for Cancer magister is measured inside the 10th anterolateral teeth while CW for all other Cancer species is measured at the widest point of the carapace. Cancer magister with CW >29 mm are sexed; all other crabs having CW >19 mm are sexed. Crabs below this threshold are labelled “juvenile”. Any unidentified fish or crabs are preserved or frozen for later identification in the lab. Shrimp are preserved in 10% formalin and no more than 1 liter is brought back to the lab for processing. Shrimp are subsampled when the sample is more than 1 liter. A graduated pitcher is used to measure the total volume and enough shrimp are randomly removed until 1 liter remains. The total volume caught and volume kept are recorded on the datasheet, in addition to the size bag used (Whirl Pak or cotton bag). The shrimp sample is returned to the lab for processing. Details on shrimp processing will be provided when the data is published. Subsampling is used to minimize processing time when a very large number of fish are collected or when an otter trawl sample is excessively large or full of mud/tubes or other substrate. Subsampling data is used to calculate the total catch (see below) but is not currently published. A detailed description of the subsampling protocol is available upon request.
Plankton Net: The plankton net was towed for 5 minutes on the bottom and then retrieved obliquely at a constant rate. In general, the ratio of length of tow cable to depth follows the 5:1 scope ratio. A General Oceanics flowmeter suspended in the mouth of the net was used to estimate water flow through the net. Plankton samples were preserved in 10% formalin and returned to the laboratory for processing. Depending on the number of larval fish and detritus in the sample, all or a portion of the sample was sorted for larval and juvenile fish. Larval fish were identified to the lowest taxon possible, then 30 randomly selected fish of each species were measured to the nearest 0.1 mm (total length) and the remainder counted. Fish greater than 15 mm total length were measured to the nearest 0.5 mm and any fish >50 mm were not included. Larvae were categorized as yolk-sac larvae, post yolk-sac larvae, or juveniles. Only flatfish were measured from 1987 to 1989 while all other species were enumerated by life stage. Northern anchovy eggs were sorted and counted from 1980 through 1987.
2.) Instrument and Equipment Specifications, including QAQC methods and frequency:
Midwater Trawl: The MWT net has a mouth opening of 12’ by 12’ (3.7m by 3.7m), is approximately 58’ (17.7m) in length and is constructed from 9 stretch mesh panels, with mesh sizes decreasing from 8” at the mouth opening to ½” at the cod end. The panels are made with single knot, bonded nylon thread dyed green. Bay Study uses 2 top and 2 bottom doors to keep the MWT mouth open while under tow. A Willapa marine shrimp buoy is attached to each top door to help with flotation and the bottom doors have extra weight to help them sink. All 4 doors are made from 1/8” thick steel plate.
Otter Trawl: The OT net has a 16’ (4.9m) wide headrope and weighted lead line and is 35’ (10.7m) in length. The net body is constructed from 1” stretch mesh (2.5 cm) and the coded is 3/16” delta nylon knotless mesh. A lazy line is attached to the net once the cod end is tied and consists of a red (~12-16”) buoy on a 100’ piece of nylon dock line. The otter trawl doors are made from oak wood, are approximately 30” x 15”, and weigh 28.5 pounds each.
Tow Cable: Tow cables are made from threaded 3/16” stainless steel wire aircraft cable inside of ¼” AmSteel Blue (threaded Amsteel cable). Bay Study recently switched tow cable type due to safety concerns. From January 1980 through May 2022, the tow cables were made from #316 stainless steel aircraft cable, 3/16” diameter, 7x19 strand (steel cable). To replace the steel cable, a new cable was made “in house” at the CDFW Stockton lab by threading the original steel cable inside of AmSteel Blue. This allows for the weight of the steel cable with the safety benefits of AmSteel Blue (e.g., less risk of cable snapping, getting snagged, or burrs causing injury to cable handlers). The threaded Amsteel cable was first used during survey in June 2022. Each tow cable has a 100’ bridle and at least 250’ of backing cable. The bridles are attached to the main tow cable with shackles and a large bullet-nose Miller swivel (size BB-414, 0.75 ton). The main tow cable is at minimum 250’ in length and is marked at 25’ increments. Small Miller swivels (size A-414, 0.45 ton) and shackles are attached to each trawl door on the tow cable end. The carabineers at the end of each bridle are attached to the shackles.
Flowmeter: A General Oceanics Flowmeter (model 2030R) is used to measure the volume of water sampled by the net and is deployed for the entire tow duration. For the MWT and OT, it is connected to a short cable that is attached to the end of a davit on the starboard side of the boat. For the plankton net, the flowmeter is suspended in the mouth of the net. Propeller revolutions are checked in between each survey.
Seabird CTD Meter: An electronic CTD meter from Seabird Electronics is used to collect water quality data at every station. Currently, Bay Study uses an SBE 19plusV2 model to measure conductivity, temperature, and depth. The Seabird is mounted in a stainless-steel cage to protect the unit. At the beginning of each sampling day, the Seabird is fastened to stainless steel deployment cable using a shackle, and the plastic tubing and metal switch stabilizer are removed. At the end of the day, the Seabird is rinsed with freshwater, and the plastic tubing and switch stabilizer are attached. The Seabird is calibrated by the manufacturer periodically, but not more frequently than annually.
Secchi Disk: The disk has a diameter of 20 cm, and the dorsal surface is divided into quadrants of alternating black and white color. The Secchi disk is commercially made and designed specifically for evaluating water clarity. The ventral surface has an attached lead weight (13cm diameter anode) to maintain vertical tension on the rope. The Secchi rope is marked with colored twine at 10 cm increments, from 10 to 250, with a different color twine marking the 50 cm increments (ex. 50, 100, 150 and 200 cm).
Crew Leader’s GPS: A Furuno GP-32 is mounted at the biologist’s station and is used to record start and end GPS coordinates, compass bearing and distance travelled for each tow.
Plankton Net: The plankton net was made of 505 micrometer mesh with a mouth area of 0.38 m2, and was mounted on a steel sled with the net bottom 12 cm above the skids.
3.) Analysis Methods and SOPs:
The columns TowVolume and TowArea from the “BoatTow.csv” table are calculated as follows:
Midwater Trawl: Water volume sampled (TowVolume) is calculated for the MWT and is calculated using data collected by a flowmeter towed from a davit alongside the boat. First, the difference in start and end flowmeter counts (number of propeller revolutions) is calculated. This value is then multiplied by a factory calibration factor to convert the reading into meters of water flow through the net. This is finally multiplied by the area of the net to estimate the volume of water sampled in each tow (m3):
Vt = M * K * Ft ,
where: Vt = volume of water (m3) filtered through the net per tow (t); M = mouth opening of the net (10.7 m2); K = calibration factor for the flow meter (0.02687); Ft = difference in flow meter counts from start to end of tow.
Otter Trawl: Area sampled (TowArea) is calculated for the OT and is calculated using data collected by the crew leader’s GPS unit. First, the difference in distance traveled is calculated from start coordinates when the net is set and the end coordinates when the tow duration runs out. This value is collected in nautical miles and converted to meters. This is multiplied by the area of the net opening to estimate the area sampled in each tow (m2):
At = S * Dt ,
where: At = area (m2) covered by the net per tow (t); S = distance between doors when fully spread (3.42 m); Dt = distance (m) between the start and end coordinates of the tow.
Plankton Net: Water volume sampled (TowVolume) is calculated for the plankton net and is calculated using data collected by a flowmeter towed from a davit alongside the boat. First, the difference in start and end flowmeter counts (number of propeller revolutions) is calculated. This value is then multiplied by a factory calibration factor to convert the reading into meters of flow (m). This is finally multiplied by the area of the net to estimate the volume of water sampled in each tow (m3):
Vt = M * K * Ft ,
where: Vt = volume of water (m3) filtered through the net per tow (t); M = mouth opening of the net (0.383 m2); K = calibration factor for the flow meter (0.02687); Ft = difference in flow meter counts from start to finish of tow.
4.) QA/QC Methods
Field QC: Bay Study implemented a field quality control (QC) program in March 1996 to check the accuracy and completion of field procedures, including emptying the net, removing organisms from the debris, and identifying and enumerating species. For each survey, 10% of the trawls were randomly selected for QC checks (3 OTs and 3 MWTs), with one QC for each region sampled. Samples identified for QC, were rechecked for missed organisms. All crabs and fish were re-counted by species. Lengths were not re-measure unless a length was in question.
Data QC:
Data Entry and Outliers: Immediately after each survey, station and tow data is entered into an Access database and first checked through “line-by-line”. Then, two proofs of the entered data are conducted by Environmental Scientists to evaluate queries that flag entry and calculation errors, data out of range, outliers and the agreement between net, tide, direction, and bearing, and the datasheets are also reviewed for legibility, completeness, and reasonable species ranges and lengths.
End of Year Proofing: The station and tow data and fish, crab, and jelly catch and length data are double entered at the end of year. The two entries are compared to flag entry errors and the working database is corrected. The final data are checked for measurement errors and range outliers using a full year of survey data in the evaluation. A specific set of queries are used to highlight questionable data, and these are checked against field data sheets. Details on shrimp QC protocols will be provided when the data is published.
5.) Historic Protocol Changes and Exceptions
Long-term monitoring surveys are subject to variety of influences that lead to inadvertent divergences from the establish sampling procedures defined in the methods sections above. Additionally, protocols have changed throughout the survey history. In effort to provide full transparency and facilitate the most appropriate uses of the survey data, known sampling changes and exceptions are defined in footnotes that are included in the PDF version of the metadata provided with the data.
6.) Source of Available Datasets
These tables are exported from Access and manipulated as follows.
“StationConstants.csv”
The table is a combination of station info from the Access database and an excel file with theoretical coordinates for the station location and comments about station location changes.
“SubstrateCodes.csv”
This table is directly from the Access database with no modification.
“SpeciesCodes.csv”
This table is from the fish Access database and some columns were renamed for clarity and irrelevant columns removed.
“BoatStation.csv”
This table is a combination of two tables from the Access database with station information joined with station surface, average, and bottom water quality readings. Irrelevant columns were removed.
“BoatTow.csv”
This table is from the Access fish database and manipulated to convert all coordinates to decimal degrees and calculate the water volume and area sampled, for each gear type as described above in Section 4. Irrelevant columns were removed.
“FishCatch.csv”
This table in the Access database originally stores subsampling information and plus count for each species per tow. Catch from net 3 is originally stored separately from nets 1 and 2, but combined in the published table. The total catch for each species collected from nets 1 and 2 is calculated using the subsample (if applicable) and adding the plus count to the number of fish measured stored in the “FishLength.csv” using the following equation:
Tt,s =((Mt,s + Pt,s))/((Cs + Ss) ),
where Tt,s = the total number of fish of a species caught per tow; Mt,s = the total number of fish of a species measured per tow; Pt,s = the plus count for the species; Cs = the total size of the sample caught; Ss = the proportional subsample processed for that species.
“FishLength.csv”
In the Access database, this table stored the length frequency of fish measured per species per tow. An additional column, AdjustedFreq, is calculated using the following equation:
At,s,l = (Ft,s,l/Mt,s)* Tt,s ,
where At,s,l = the adjusted length frequency of a species in the tow; Ft,s,l = the frequency of that species and length; Mt,s = the total of that species measured in the tow; Tt,s = the total number of fish of a species caught per tow. Additionally, there is an added column for Width because from 1980-1991 Bat Ray total length was measured and the measurement is in the Length column and after 1992 the protocol changed to measure Bat Ray wing width, which is stored in the Width column.
Users interested in using these base tables should be aware of the units of these recorded values. Users should also be aware that zero catches of each species per tow are not recorded in the base “FishCatch.csv” table, but the data associated with that tow is recorded in the base “BoatTow.csv” table.
7.) Sampling Frequency
Throughout survey history, there are multiple gaps in sample frequency. Refer to the metadata PDF for visuals of sampling frequency by net type and reasons for missed surveys and partial survey.
