About analysis domains

An analysis domain defines the initial set of data points (samples or single targets) that are available for analysis. Analysis domains can take the form of:

• Selections
• 2D regions
• 3D regions (on scenes)
• Cells
• Region-cell intersections
• The entire echogram

Analysis domains can be defined on single beam echograms, single target echograms and scenes.

Note: Analysis is affected by the settings defined on the Analysis page of the Variable Properties dialog box.

The topics discussed on this page include:

• What is included in an analysis domain?
  • Single beam and single target echograms
  • Multibeam echograms
  • Scenes (multibeam data)
• Analysis settings
  • Analysis exclusions based on lines
  • Analysis exclusions based on regions and samples
    • Special case
    • Integration cues and outcomes - examples
    • Intersections
  • Analysis exclusions based on surfaces
• Excluded data points
• Good data points

What is included in an analysis domain?

Single beam and single target echograms

Data points with time/distance or depth coordinates that are inside the boundary of the analysis domain are included in the analysis domain.

Data points with time/distance or depth coordinates that place them on a boundary of the analysis domain are treated as follows:

If the data point is on the...	Then...
Upper boundary	it is included in the analysis domain
Lower boundary	it is not included in the analysis domain
Left boundary	it is included in the analysis domain
Right boundary	it is not included in the analysis domain

Notes:

• In echograms, the range of the sample represents the sample data point. For more information refer to About stop and start ranges.
• In single target echograms, the range of the single target (the top of the rectangle) represents the single target data point. From more information refer to About single targets: Single target display.
• Selections, 2D regions and cells are all defined in time/depth coordinates.
• Lines in Echoview have a higher precision than Sv sample ranges. As a result the (line) boundaries of an analysis domain may not lie on sample boundaries or may bisect samples. Similarly, the (line) boundaries of an analysis domain may not lie on single target boundaries.
• Heave compensation can change the start depth of a ping, and may affect the membership of pings in an analysis domain and hence affect analysis variable calculations.
• The boundary rules apply to cells in the Cell statistic operators.

Multibeam echograms

Created, detected or virtual surfaces may be used as boundaries to include data points for detections and analysis. Data points inside the boundary are included. Data points outside and on the boundary are excluded.

Scenes (for multibeam data)

Analysis domains on scenes take the form of 3D regions.

Data points are included in an analysis domain if:

• • the latitude, longitude and depth coordinates of the data point are inside the boundary of the analysis domain, and
  • the time of the ping on which the data point is, lies within the valid time of the analysis domain

Data points with latitude, longitude and depth coordinates that place them on the boundaries of the analysis domain are highly unlikely to occur and inclusion is undefined in this circumstance.

Data points can be excluded from analysis domains on scenes with the usage of Exclude above and Exclude below surfaces.

Analysis settings

Data points (samples or single targets) that are included in an analysis domain are then retained or discarded for further analysis based on the settings defined on the Analysis page. The content of this section may benefit from the topics and background covered by About samples and echo integration.

Exclusions based on lines

Data points are excluded from further analysis if they are:

• Above the Exclude above line (if one is defined)
• Below the Exclude below line (if one is defined)

Analysis based on regions and samples

Data points (samples or single targets) are treated as no data if they are:

• Within a Bad data (no data) region and Apply bad data regions is applied.
• Samples with a value of no-data or within a range interval for no-data (on single target echograms)
• Within a ping where the Exclude below line (if one is defined) has a status of “bad” and Set pings with bad line status on the exclude below line to no-data is applied

These samples are assigned a value of no data and they have no sample thickness. In effect, such samples are not counted in the calculations of Thickness_mean and Height_mean. See also the special case below.

Selecting Apply bad data regions also affects:

• Samples within a Bad data (empty water) region. These samples are assigned a value of -999 dB (effectively zero backscatter) and are counted in all calculations.
• Single targets within a Bad data (empty water) region. These single targets are excluded from further analysis.

Special case

Samples within a Bad data (no data) region or with a value of no data in concert with the selection of the Analysis page settings Apply bad data regions and Include the volume of no-data samples.

These samples are assigned a value of no data and have a sample thickness. The sample thickness is included in calculations of the Thickness_mean and Height_mean. The settings can affect both entire pings and parts of pings. The key intent of this special case is the robust assumption (in the absence of any further information) about samples in Bad data (no data) regions. You are assuming, in effect, that these samples have a backscatter value equal to the mean of the good samples in the domain.

Prior to Echoview 5.4 it has always been possible to make this (commonly-desired) assumption by scaling the reported Thickness_mean of the domain (and hence the NASC) using the various analysis domain analysis variables (no_data_samples etc.).

Integration cues and outcomes - examples

	Use Apply bad data regions	Use Apply bad data regions Use Include volume of no data samples
	Hatching cue, Outcomes, Assumptions	Hatching cue, Outcomes, Assumptions
Bad data (no data)	Purple border, grey hatch. Does not use no-data sample intervals in Thickness mean. This data effectively reduces Thickness mean and reduces reported volume for NASC.	Purple border, grey hatch. Use no-data sample intervals in Thickness mean. This data effectively is converted to the mean of the analysis domain. This gives you greater flexibility in calculating your NASC based results.
Bad data (empty water)	Green hatch. Uses empty water samples in Sv mean count and Thickness calculations. This data is deemed to have no acoustic return.	Green hatch. Uses empty water samples in Sv mean count and Thickness calculations. This data is deemed to have no acoustic return.
Analysis	Red hatch. Good samples used in analysis calculations.	Red hatch. Good samples used in analysis calculations.

Intersections

Data points in intersections of regions or other designated areas may change in nature, for calculation purposes, according to an order of precedence:

1. Good
2. Bad data (empty water)
3. Bad data (no data)
4. Above Exclude above line
5. Below Exclude below line

With the Good category being dominated by any of the following categories, and the Below Exclude below line category dominating any of the preceding categories.

The Processed data operator can often make clear how samples are affected by Analysis page settings.

The screenshot to the right shows region types, integration cues and intersections. Regions include: Bad data (empty water) region Analysis region with integration hatching Bad data (no data) region

Exclusions based on surfaces

Data points that are included in an analysis domain are then retained or discarded for further analysis based on the settings defined on the Analysis page.

Data points are excluded from further analysis if they are:

• Above the Exclude above surface (if one is defined)
• Below the Exclude below surface (if one is defined)

Excluded data points

Excluded data points are ignored for these primary calculations:

• Thickness_mean and Height_mean
• Sv_mean
• Beam_volume_sum
• Wedge_volume_sampled
• TS mean, TS minimum, TS maximum and TS standard deviation
• Target depth mean
• Target length mean
• Number of single targets
• 2D School Detection module analysis variables - Uncorrected_area and 3D_school_volume

In turn, these quantities variously affect the calculation of:

• Area_backscatter_strength
• ABC
• NASC
• Density_number
• Density_weight
• 2D School Detection module analysis variables - Corrected_area and Image_compactness

Notes:

• Thickness_mean, Sv_mean, ABC and NASC are key quantities in echo integration. Excluded samples do not contribute to these quantities. If you have bad data, we recommend that you fully understand the implications of any quantities you assign to a given volume of water, such as the fish density in a given alongtrack interval. See About samples and echo integration for further discussion, or contact support@echoview.com.
• Excluded samples do not alter the definition of the analysis domain itself. The analysis domain's boundaries remain unchanged. These exclusions apply to the analysis of samples within the analysis domain.
• Echoview 5.4 introduced general analysis variables Bad_data_no_data_samples and Bad_data_empty_water_samples to represent samples in the region types Bad data (no data) and Bad data (empty water). In Echoview 5.2 samples within bad data regions with a value of no data are labeled as bad_data_samples. In versions 5.1 and older, these samples are labeled as no_data_samples.

Good data points

The data points that are included in an analysis domain and not subsequently excluded are known as good data points.

When you perform an on-screen integration or an on-screen analysis of an analysis domain, all good data points will be overlayed with analysis-hatching (providing that the Show Analysis option on the Echogram page of the EV File Properties dialog is selected).

Tip

You can view the good data points for an entire echogram using the Processed Data operator. This operator assigns a value of no data to excluded data points.

You can view information about a data point as follows:

• To view time information about a data point, move the mouse point over the data point and check the status bar.

• To view depth and range information about a data point display the Details dialog box (F9).

See also

About analysis
About analysis variables
Samples
About integration
About samples and integration
About no data
Thickness mean and Height mean
ABC and NASC
About biomass estimation
About single targets
About no data
Sv_mean
Analysis page
Bad data regions
Special data values