Using the School Detection module

Prior to detecting schools

Platform: Fixed

Platform: Determined by GPS

You should ensure that the cruise track has been processed (and optionally smoothed) using appropriate settings. Cruise track data affects distance calculations and allows you to determine whether the survey vessel has doubled back on its track, which could result in repeat observation of the same school(s).

Specify an item from the Distance mode list on the Schools page of the EV File Properties dialog box.

Detect schools settings

In the Data page of the Variable Properties dialog box, Minimum threshold and Maximum threshold settings should be entered. Maximum integration threshold is generally not important to school detection. However, the minimum threshold is of great importance, since sub-threshold data points surrounding the school will be excluded from analysis as the threshold increases, resulting in important changes in estimates of school shape and energetic properties. Numerous approaches have been taken to the question of determining what processing threshold to use for school detection (see Richards et al., 1991; MacLennan and Simmonds, 1992; Barange, 1994; Swartzman et al., 1994; Scalabrin et al., 1996). The corrections for beam width effects used by Echoview (Diner, 1998) generally perform best at processing thresholds between -60 and -65 dB (that is, when dST = (Sv mean - Th) is between 10 and 30 dB)+. The ICES study group on echo trace classification has suggested a threshold for school detection of -60 dB for pelagic schooling fishes (Anon. 1998).

Notes:

Lawson et al. (2001) discuss the usefulness of analyzing schools at two different processing thresholds, in order to use the ratio of measurements of school features at the two thresholds for the purpose of species identification.
+ Where Sv mean is the school density and Th is the processing threshold.

In the Analysis page of the Variable Properties dialog box, the Exclude-above line setting can be used to avoid detected schools being linked to surface noise. The Exclude-below line can similarly be used to ensure no bottom echo is attached to detected schools. The Apply minimum threshold and Apply maximum threshold checkboxes on the Data page of the Variable Properties dialog box must in general be applied prior to school detection. In general a maximum threshold is not required. The minimum threshold should also be applied during echo integration at the same value as during school detection for appropriate use in the corrected schools variables.

In the Schools page of the EV File Properties dialog box, the parameters for school detection are best reached through empirical preliminary analysis. Repeated analysis of some preliminary data files will give an indication of the performance of various settings. It should perhaps be kept in mind that lengths are measured as the sum of individual ping widths (width of ping = difference in position of ping and next ping), and height in multiples of the distance corresponding to the digitizing rate. Echoview does not assume that speed or ping widths are constant over the length of a school when calculating the length of a school.

Some consideration of the biology of the species of interest (for example, what schools smaller than the chosen minima represent to stock biomass, whether an aggregation smaller that 1m in length is in fact a 'school') may also help in these decisions.

Post-processing

Not all 'schools' detected by the module will be true biological schools. Noise, aggregations of organisms other than the species of interest, and so on, may be detected and treated by the module as 'schools'. Some scrutiny may be necessary before or after exporting school variables. In particular regions that have uncorrected lengths smaller or mean energies lower than those appropriate for calculating some corrected schools variables will still be processed and corrected variables exported without filtering. You should filter the exported schools where required.

Volumetric density (individuals/m3), area density (individuals/m2) and school abundance can be calculated in the usual manner with knowledge of species mix in school, target strength of species and an estimate of school volume.

To further understand the calculation of schools variables please read Notes about School Detection.

Warning: It is advised that schools regions shouldn't extend beyond the data. This may occur when you detect schools then remove data files and add new data such that schools regions extend beyond the new data. In this case, the calculation of inconsistent Uncorrected_length and Uncorrected_thickness may have undesired effects on other schools analysis variables. For more information refer to School Detection module algorithms.

Live viewing

Detections can be applied in near real time with live-viewing data. For more information refer to About Live viewing and School detection while live viewing algorithm.