Community Fishers Vertical Cross Section 

Overview

This data product presents the vertical cross section of an ocean property (e.g., temperature, salinity, etc.) along a transect. 

Transects

A transect is a predefined geospatial area, often consisting of spatially consecutive stations, with an associated origin (latitude, longitude) point.,

Note: Transects may overlap and can include casts that are not associated with a specific station but fall within the geospatial extent of the transect.

Data Requirements

In order to produce this product the data must meet the following requirements:

• All cast data must be sampled using the same device,
• A minimum number of casts within a predefined time duration, set on a transect per transect basis,
• All cast data must have been reviewed by a data specialist. 

Formats

The data product is distributed in the following formats:

1. PNG Plot - A colour contour plot of the parameter against distance vs depth axes.
2. CSV Table - A tabular summary of the data.
3. MAT File - A binary, compressed file containing structured arrays with the data.

All formats are based on the same data which includes:

• Metadata,
• Depth bins,
• Distances along the transect,
• Cast data within the geospatial extent of the transect,
• DIVA interpolated values of the ocean property with values deeper than cast measurements masked out.

Note: The amount of metadata varies slightly from format to format. PNG plots have the least metadata, presented in the header and comment. CSV tables have it in the table headers. MAT files have raw metadata as a separate structured array.

Data Collection Methodology

• Sensor Deployment: Data is collected using a profiling device that is lowered into the ocean, recording values of the ocean property as it descends.

Data Processing Workflow

Individual Casts

Each cast (sensor deployment) yields a profile of the ocean property vs depth at a particular location. The measurements from the various sensors attached to and within the instrument are processed to filter, smooth, align, and average data into depth bins. . Recorded measurements are averaged into 1-meter depth bins, each centered at 1-meter intervals (e.g., 0.5–1.5 m centered at 1 m).

Casts can either be assigned to a station or not (unassigned casts that fall within the transect geospatial area). 

Cast Groupings

Profiles from each station are grouped together. Unassigned casts are evaluated and added to an existing group if they fall within a specified distance of any cast in that group (2% of the transect length). After processing the station-based groups, any unassigned casts that remain are grouped independently, using the same proximity-based logic.

Once all groups are formed, the best cast from each group is selected. The "best" cast is defined as the one with the highest number of valid (non-NaN) values, ensuring the most reliable data is used in the final product.

DIVA Interpolation

Values between cast locations are interpolated from profile data using a Data Interpolating Variational Analysis (DIVA) method:

The two-dimensional domain (distance, depth) is discretized by a Finite Element Method (FEM) grid. A spatial density function is applied to the grid generation process in order to cluster nodes more densely near cast locations. The bathymetry line, interpolated using a Piecewise Cubic Hermite Interpolating Polynomial (pchip) from profile depths, is used as a mask such that no values of the ocean property are interpolated for depths deeper than cast measurements. 

Each FEM grid node, defined by a unique (distance, depth) pair, serves as a query point where ocean property values are interpolated as follows: 

1. For each cast location, vertical values are linearly interpolated for the query depths,
2. Each query node is assigned a weight based on its horizontal distance from the surrounding cast locations,
3. For each query node, a weighted average of the vertically interpolated station values is computed and assigned as the interpolated ocean property value.

Error Calculation

For each value interpolated from more than a single cast measurement, a weighted standard deviation and a distance-normalized weighted standard deviation are computed. These values indicate the local uncertainty of the interpolation, based on the degree of agreement between contributing stations and their proximity to the interpolation point. 

The standard deviations are only included in the csv and mat formats of this data product.

Examples

PNG Plot

Text within the plot

The first title string: "TikKuatikkak Bay • TN001 (TRANSECT 1) • Transect Origin: 56.7583° N, -62.5101° W"

• TikKuatikkak Bay: The name of the area where the transect is located.
• TN001 (TRANSECT 1): The transect where the cast profiles were taken.
• Transect Origin: 56.7583° N, -62.5101° W: The origin latitude and longitude of the transect.

The second title string: "Date Range: 2022-07-05T14:45:15Z to 2022-07-05T17:44:22Z (UTC) • Location Range: 56.6713 to 56.756° N, -62.4887 to -62.0957° W • 1 to 122 m"

• Date Range: 2022-07-05T14:45:15Z to 2022-07-05T17:44:22Z (UTC): Timestamp range of casts.
• Location Range: 56.6713 to 56.756° N, -62.4887 to -62.0957° W: The range of latitude and longitude where casts were taken.
• 1 m to 122 m: The minimum to maximum depth of all casts within the plot.

The third title string: "Cast Direction: Down Cast • 7 Casts in Plot"

• Cast Direction: Down Cast: The direction of the casts within the plot indicating whether the cast data was collected by a descending or ascending device.
• 7 Casts in Plot: The number of casts included in the plot.

The comment in the footer of the plot: "Comments: Data attribution - Nunatsiavut Government (owner), Ocean Networks Canada (owner, custodian, distributor) Interpolation Method: DIVA with values deeper than casts measurements masked out. See documentation (DP 201) for details."

• Data attribution: The roles of companies and communities regarding the data used for the plot, in alphabetical order.
• Interpolation method: A comment regarding the interpolation method used to produce the plot.

CSV Table

File:

Origin Section

• SOURCE: Citation author.
• HTTP: Citation publication site.
• HOME: Citation publication location.
• FLDATE: Creation date of the file.
• CITATION: DOI citation text as it appears on the Dataset Landing Page. If the file contains data from multiple datasets, additional lines are included in CITATION.
• DATAATTRIBUTION: Companies and communities with their roles regarding the data used for the data product, in alphabetical order.
• SEARCHID: DMAS search ID from Oceans3.0.

Location Section

• LOCNAME: Location Name.
• TRANSNAME: Name of transect.
• TRANSORIGLATITUDE: Origin latitude of the transect.
• TRANSORIGLONGITUDE: Origin longitude of the transect.
• STNNAMES: Names of stations or unassigned casts included in the product.
• STNAVGLAT: Average latitude of each station.
• STNAVGLON: Average longitude of each station.
• DEPTH: Depth range of casts (m).

Device Section

• DEVCAT: Device Category.
• DEPLDATE: Device deployment date.
• DEVNAME: Full device name.
• DEVCODE: Device code.
• DEVID: Device ID.

Data Section

• DATEFROM: First timestamp (using ISO8601 Extended format) contained within the data.
• DATETO: Last timestamp (using ISO8601 Extended format) contained within the data.
• PROCESSINGCOMMENT: Data product comment regarding the calculations of statistical values.
• QAQCCOMMENT: Quality assurance comment with the meanings of QAQC flags.

Measured and Interpolated Data Table

• SENSORID: Sensor IDs.
• CASTTIME: Timestamp when the casts started.
• SENSORNAME: Column name (same as the sensor name for individual casts).
• SENSORUNIT: Units of measurement for the data in the column.
• DISTANCEALONGTRANS_KM: Column distance along transect in km.

Weighted Standard Deviation Table

• COLUMNNAME: Column name.
• UNIT: Units of the data column.
• DISTANCEALONGTRANS_KM: Column distance along transect in km.

Distance Normalized Weighted Standard Deviation

• COLUMNNAME: Column name.
• UNIT: Units of the data column.
• DISTANCEALONGTRANS_KM: Column distance along transect in km.

MAT File

File:

Metadata: Structure array with metadata

• creationDate: Date and time (using ISO8601 format) that the data product was produced. This is a valuable indicator for comparing to other revisions of the same data product.
• deviceDetails: Structure array with the following fields:
  
  • deviceDeploymentDateFrom
  • deviceDeploymentDateTo
  • deviceID: Unique identifier to represent the instrument within the ONC observatory.
  • deviceName: Name given to the instrument.
  • deviceCategory: Unique name given to the category of devices, such as 'CTD'.
  • deviceCategoryCode: Code representing the device category. Used for accessing webservices, as described in Oceans 3.0 API.
  • deviceCode: Unique string for the instrument which is used to generate instrument search data product file names.
• location: Structure array with the following fields:
  
  • locationName: Location name.
  • transectName: Name of the Transect
  • transectCode: Code representing the Transect. Used for accessing webservices, as described in Oceans 3.0 API.
  • depth_metres: Obtained at time of deployment. If NaN, the device is mobile and this position is a variable, the data for which is supplied by a sensor in the data struct.
  • lat_degrees: Range of latitudes of transect.
  • lon_degrees: Range of longitude of transect.
• Attribution: Structure array with information on any contributors, ordered by importance and date. If an organization has more than one role it will be collated. If there are gaps in the date ranges, they are filled in with the default Ocean Networks Canada attribution.
• citation: Char array containing the DOI citation text as it appears on the Dataset Landing Page. The citation text is formatted as follows: <Author(s) in alphabetical order>. <Publication Year>. <Title, consisting of Location Name (from searchTreeNodeName or siteName in ONC database) Deployed <Deployment Date (sitedevicedatefrom in ONC database)>. <Repository>. <Persistent Identifier, which is either a DOI URL or the queryPID (search_dtlid in ONC database)>. Accessed Date <query creation date (search.datecreated in ONC database)>
• totalScalarSamplesReceived: Number of valid (non NaN) samples in cast. 

GridData: Structure array with data interpolated from cast data with fields:

• interpolationComment: Data product comment regarding the calculations of statistical values.
• distanceVals: Distance values of FEM grid.
• depthVals: Depth values of FEM grid.
• ColorMappedVals: DIVA interpolated values of the ocean property.
• ErrGrid: Structure array with DIVA interpolation errors as fields:
  
  • weightedStd: Array of weighted standard deviation values.
  • NormalizedWeightedStd: Array of distance normalized weighted standard deviation values.

ProfileData: Structure array with profile data, aligned and depth-binned.

• depth: Structure array with the following fields:
  • sensorID: Unique identifier number of sensor used for profiling.
  • sensorName: Name of sensor.
  • sensorCode: Code of sensor, can be null
  • sensorDescription: Description of sensor.
  • sensorType: Type of sensor as classified in the ONC data model.
  • sensorTypeID: ONC ID given to sensor type.
  • units: Unit of measure for the calculated data.
  • sensorDerivation: String describing the source of the sensor data: derived from calibration formula (dmas-derived), calculated on the device (instrument-derived), calculated by an external process (externally-derived), or direct from the instrument.
  • propertyCode: Property code string for the sensor as used in the Oceans 3.0 API.
  • isMobilePositionSensor: Boolean (flag) to determine if sensor is a mobile sensor. 
  • deviceID: Unique identifier number for the parent device.
  • searchDateNumFrom: Start of the time interval queried with the search in MATLAB datenum format.
  • searchDateNumTo: End of the time interval queried with the search in MATLAB datenum format.
  • dataProductOptions: String specifying any options selected for this data product.
  • qaqcFlagDescription: Quality assurance comment with the meanings of QAQC flags.
  • time: Vector of data timestamps in MATLAB datenum format.
  • dat: Cast data.
  • qaqcFlags: Flags indicating the quality of the data.
  • dataDateNumFrom: Start (first time-stamp) of the cast in MATLAB datenum format.
  • dataDateNumTo: End (last time-stamp) of the cast in MATLAB datenum format.
  • Calibration: Structure array with fields:
    • dateFrom: Start date of calibration.
    • dateTo: End date of calibration.
    • sensorID: Unique identifier number of sensor.
    • name: Name of calibration.
    • formula: Formula used for sensor calibration.
    • qaqcdependencies: A comma separated list of sensor ids. QAQC dependencies are normally only defined if there is no reference either directly or indirectly to a dependent sensor. For example, sensor 100 Salinity is dependent on sensor 101 Depth but the Salinity sensor has no calibration formula. This dependency would be defined in the QAQC Dependencies field for sensor 100 as $101.
• <ocean property name>: Structure array with data corresponding to the ocean property. The fields are the same as the depth struct above.
• longitude: Structure array with data corresponding to the longitude. The fields are the same as the depth struct above.
• latitude: Structure array with data corresponding to the latitude. The fields are the same as the depth struct above.
• startDate: Start date and time of cast.
• endDate: End date and time of cast.
• minDepth: Minimum cast depth.
• maxDepth: Maximum cast depth.
• direction: Cast direction.
• CastSiteDeviceSubset: Struct array with the following fields:
  • siteDeviceSubsetId: Unique identifier number for site device subset.
  • isCastReviewed: Boolean (flag) indicating if cast was manually reviewed by a data specialist.
  • searchTreeNodeName: Name of search tree node.
  • siteDeviceSubsetName: Name of site device subset.
  • avgLatOfStation: Average latitude of station.
  • avgLonOfStation: Average longitude of station.
• station: Station name.
• meanLatOfCast: Mean latitude of cast.
• meanLonOfCast: Mean longitude of cast.
• distanceAlongTransect: Distance of cast along the transect in meters. The distance is calculated using a running sum, starting from the transect origin point, and the mean longitude and latitude of each cast.