For a formatted version of the methods please see attached pdf file SLS_Metadata.pdf
1.) Data Collection methods:
A single 10 minutes stepped-oblique tow of the net is conducted at each sampling station. Environmental conditions data, including tide, surface turbidity, surface and bottom conductivity, surface water temperature, and Secchi depth are collected at each station at the time of sampling. A General Oceanics (GO) flowmeter is mounted across the net's mouth to estimate the water volume filtered. At the end of each tow, net contents are washed into a cod-end jar attached at the end of the net. Large debris and adult fish (greater or equal 50 mm) are removed if positively identified. If salmonids are caught, lengths and presence of adipose fin are immediately recorded, and the fish are released gently and alive. The remaining contents in the jar are preserved in 10% formalin for identification in the lab. The distinctively labeled sample jars are taken to the laboratory at the California Department of Fish and Wildlife's Bay Delta Region, Stockton. The complete contents are sorted, and any larval fish present are identified and counted. All fish are identified to species or lowest possible taxon. The first 50 fish of each species from each tow are randomly selected and measured (FL) to the nearest millimeter and the rest enumerated. All Delta Smelt are measured regardless of catch size. All Longfin Smelt are measured unless there is an excessive amount ( greater than 200) then only 100 are measured and the rest are plus counted. Only non-zero catch of species caught are recorded.
2.) Link to blank datasheet:
Available upon request (see data contact information)
3.) Instrument and Equipment Specifications, including QAQC methods and frequency:
The IEP-SLS uses a cone shaped net 3.35 meters in length with a mouth area of 0.37 m2. The net itself is composed of 505 microm NitexR and is mounted on a fixed metal tube frame with skis. The mesh was altered prior to the 2014 season to 500 microm NitexR, when new nets were purchased, and the original mesh was no longer available (see 2014 changes below). These new nets were incorporated as old nets became unusable. The net is connected to the frame by a canvas mouth. Nets are checked for tears after each tow. A YSI is used to measure top water temperature (degreeC), surface turbidity (FNU), and electrical conductivity (microS/cm, normalized to 25 degreeC). The probe is rinsed between samples with distilled water. Hach 2100P turbidimeters are used to measure turbidity in nephelometric turbidity units (NTU). Turbidimeters are calibrated annually with commercially available standards designed specifically for the model before each field season. Secchi discs are black and white discs that measure water clarity by measuring the depth of disappearance from the water's surface, or the distance light can penetrate. Secchi discs are mounted onto rigid meter sticks, with a maximum depth of 200 cm. Measurements are taken in the shade, with no sunglasses on, and are done by the same person the entire day for consistency. All General Oceanics flowmeters are checked at the end of the season to determine if readings are correct or not. Prior to 2015, all flow meters were calibrated at UC Davis before the start of the season; after 2015, all flow meters were sent back to GO for refurbishing (2015-2019) or replaced with new units once readings become inaccurate (2019-current).
4.) Analysis Methods and SOPs:
Water volume sampled in each tow is calculated using data collected by a flowmeter mounted to the net. First, the difference in flowmeter counts (distance traveled) is calculated by subtracting the flowmeter reading when the net is put into the water from the flowmeter reading when the net is taken out of the water. This value is then multiplied by a factory calibration factor to convert the reading into a number of meters of flow. This is finally multiplied by the area of the net to estimate the volume of water sampled in each tow (m3):
Vt = A * K * Dt
Where: Vt = volume of water (m3) filtered through the net per tow (t)
A = mouth opening of the net (0.37 m2)
K = calibration factor for the flow meter (0.026873027 since 2015)
Dt = difference in flow meter counts from start to finish of tow
Total number of fish caught per volume water sampled (CPUE) standardized to 1000 m3 is calculated for the SLS using the following equation:
nt = Ft / Vt * 1,000 m3
Where: nt = number of fish per 1,000 m3 per tow (t)
Ft = fish sampled per tow
Vt = volume of water filtered through the net (m3) per tow
5.) Project History:
2009 - Project start. Five biweekly Delta-wide (35 stations) surveys conducted from early January to early March.
2010 - Temporal extension of sampling temporarily for this season; six biweekly (35 stations) surveys conducted from early January to late March (this addition lasted only this season)
2010 - Implementation of using a Hach Model #2100P Turbidimeter as Standard Operating Procedure to record turbidity in NTU
2010 - Recorded sampling latitude and longitude on datasheets, but this data was not entered into the database.
2011 - Latitude and longitude of tows recorded into database
2011 - Yolk sac and oil globule presence noted in the data
2012 - Sixth survey permanently added
2013 - NA
2014 - Spatial extension of sampling into the Napa River as part of an agreement with the State Water Contractors (stations 340, 342, 343, 344, 345, 346, 347, 348, and 349)
2014 - Database was revised by Tuongvan Nguyen at ITB as part of the Bay Delta Application Hosting to move public facing data onto a secured Tier 3 server. Data is now entered into SLS_Local.accdb (local server), and appended to the Tier 3 server before uploading to the public webpage
2014 - New nets were incorporated (manufactured on 5/10/2013 by Lodi Tent and Awning) with a different Nitex Mesh purchased from Sefar (500-micron, 47% open space, part #06-500/47)
2015 - Factory k value (0.026873027) used in the MeterCorrections table. Flowmeters were not calibrated at UC Davis due to machinery malfunction. The facility is awaiting repairs.
2016 - Continued using factory k value for MeterCorrections. Flowmeters were sent to General Oceanics for refurbishing prior to field season.
2017 - Continued using factory k value for MeterCorrections. Flowmeters were sent to General Oceanics for refurbishing prior to field season.
2018 - Continued using factory k value for MeterCorrections. Flowmeters were sent to General Oceanics for refurbishing prior to field season
2019 - Continued using factory k value for MeterCorrections now and going forward. Flowmeters were sent to General Oceanics for refurbishing prior to field season or replaced with new meters if readings are inaccurate (assessed at the end of a season)
2019 - Spatial reduction of sampling. Ceased sampling stations within the Napa River (stations 340, 342, 343, 344, 345, 346, 347, 348, and 349)
2019 - On 2019-09-10, two tables were removed from the local copy of the database: Zooplankton and Zoo Catch. These tables were appended to the database from the 20-mm database back in 2013. The SLS survey does not survey for zooplankton. More information and a copy of the tables can be found on the local server: U:/NativeFish/SmeltData/Zooplankton/SLS_Erroneous_ZooTables.xlsx
2020 - NA
2021 - Spatially constrained, temporal extension of sampling: two additional surveys were added in December and were limited in geographic range to the south/central Delta to inform risk of entrainment for larval Longfin Smelt.
2022 - The two additional surveys in December are expanded to encompass all stations. Napa River stations (340, 342, 343, 344, 345, 346, 347, 348, and 349) have been added back to the surveys, including the supplemental December surveys.
2022 - End latitude and longitude of tows recorded datasheets starting in survey1, but this data was not entered into the database.
2022 - End latitude and longitude of tows recorded into database starting in December and continuing into the 2023 sampling season
2023 - Changed from YSI model 30 to YSI ProDSS which can now measure turbidity (FNU). Instead of collecting bottom water sample with a Van Dorn the ProDSS probe is deployed to the bottom to measure bottom EC. After this year turbidity (NTU) will no longer be collected using the HACH 2100p turbidimeter
2023 - Ethanol jugs were brought into the field to preserve any identifiable juvenile or adult osmerid.
2023 - 15 Fall Midwater Trawl stations have been added in the San Pablo Bay region to improve spatial balance in accordance with the Monitoring Design Committee
6.) QA/QC
Field methods: After each tow, the field lead verifies that the flowmeter count is within the appropriate range, and if not, a re-tow is required. If there is an obvious reason that the flow meter was out of range, (e.g., weeds caught in the net or on the flowmeter), it is recorded as the reason for a re-tow in the comments section of the datasheet. The net can clog during algal blooms, jellyfish blooms, or heavy debris loads. If material is overfilling the cod end jar, the tow time is reduced to 5 minutes or 2.5 minutes and the appropriate tow schedule is followed. The acceptable flowmeter check range is reduced accordingly. If a 2.5-minute tow is performed and material like algae, jellyfish, or peat is still overfilling the cod end, that tow at that entire station is dropped.
Fish ID: In the lab fish are identified from each sample under a microscope. First, fish are separated from debris and other organisms during a process referred to as sorting. Then, the entire sample undergoes a quality control (QC) check to ensure that fish were not missed during sorting. Finally, fish undergo a first ID and count by an identifier, followed by a QC from a larval fish ID specialist to confirm all species identifications and counts. This QC process is dependent on the experience of the first identifier doing the first ID. Fish identifiers will begin with all their identifications QC'ed and transition to having fewer and fewer samples QC'ed with experience, until the identifier is considered a larval fish ID specialist. Once the larval fish ID specialist level is acquired, samples are randomly selected to undergo the QC process. Across all samples (QC required or not), all CESA and ESA fishes and any questionable fish IDs must undergo a second ID. All fish are identified to species or the lowest possible taxon. Only the first 50 randomly selected individuals of each species from each tow are measured for lengths to the nearest millimeter, and the rest of the sample is simply enumerated. However, all Delta Smelt (Hypomesus transpacificus) are measured for lengths regardless of catch size. All Longfin Smelt are measured unless there is an excessive amount ( greater than 200) then only 100 are
measured and the rest are plus counted.
Data: All data is entered into a digital Microsoft Access database once the fish id has been finalized. Immediately after entry, data undergoes two rounds of line-by-line checks, wherein all data fields are checked against the original datasheets for fidelity. At the end of the SLS survey field season, once all the fish samples have been processed in the laboratory and data entry is complete, all data is finalized to be as accurate as possible for public use. The first step in this process is to conduct two additional final line-by-lines checks. Next, a project lead will run a series of coded queries to analyze the underlying data distributions to detect potential
outliers in the environmental data. Not all data is changed if it is flagged as an outlier (generally beyond 2 standard deviations of the mean). In most cases, outliers are real data. These queries simply alert the project lead of potential erroneous data, and care is taken to edit only data that truly needs to be edited, e.g., data that was entered incorrectly or caused by equipment failures. All resulting data edits are documented in a separate log file.
7.) Source of the available data tables:
The Catch.csv, FishCodes.csv, Length.csv, MeterCorrections.csv, Station_Lookup.csv, TowInfo.csv, and WaterInfo.csv are available relational tables from the SLS Access database. These tables are exported directly from Access in R and the only manipulations were to include relevant columns, fix Unicode encoding errors, and fix float formatting errors; all underlying data collected in the field and entered into the database remained unmanipulated. The SLS.csv file is the integrated dataset that combines 6 relational tables (the FishCodes.csv table is not included) together. Users should be aware of the units
of the recorded values between the relational and integrated tables, as they may differ (documented in the metadata section of the EDI publication page). All steps are coded in R and the relevant codes are provided with the EDI publication and/or housed on trinhxuann/CDFW-IEP-Surveys GitHub page.
Zero filling:Zero filling is the process of assigning a count value of 0 for instances of no fish catch during a tow. No fish catch can be defined as two levels: 1) across all fish species (a tow that catches no fishes at all), or 2) specific to a singular species (a tow that catches no individuals of a particular species but does for other species). Instances of no fish catch of any fish species in a tow (level 1) are not recorded in the relational Catch.csv table, but the environmental data associated with that tow is recorded in the relational TowInfo.csv table. The joined SLS.csv table flagged these instances in the Length_NA_flag column and filled in the corresponding catch count value (Count) as 0. This zero-filling was not implemented for instances of no fish catch of a particular species in a tow (level 2) in the integrated SLS.csv file; however, code for this step is provided in the SLSIntegrateEDI.R script for users who are interested.
F_(a,l)=T_c (F_(m,l)/T_m )
Where: Fa,l = adjusted frequency of each recorded length
Tc = total catch
Fm,l = measured frequency of each recorded length
Tm = total number of fish measured
The integration process leverages existing code from the LTMRdata package and is provided with the dataset as SLS.R and housed at this GitHub location. Users should familiarize themselves with the code before attempting to use the integrated dataset.
8.) Contractor Information: N/A
9.) External Review Process: N/A
10.) Methods References:
Samuel M Bashevkin. (2020). LTMRdata: Data for the IEP long term monitoring survey review (v1.0.0). Zenodo. https://doi.org/10.5281/zenodo.3934724
