IMOS BA SOOP - Bio Acoustic Data from RV Aurora Australis in the Southern Ocean from the 16th of April 2012 to the 18th of April 2012.

Brief description

This Bio Acoustic data was collected in April 2012 under the IMOS Ship of Opportunity (SOOP) Bio Acoustic (BA) program on RV Aurora Australis (IMOS platform code: VNAA).

Departed: Australia, April 16, 2012
Arrived: Macquarie Island, April 18, 2012

Bio acoustic signals allow understanding how mid-water prey species (known collectively as micronekton) such as small fish, squid, krill and jellyfish are distributed.

Mid-water prey form the core of the ocean food web, transferring energy from primary producers at the ocean surface to top predators such as tunas, billfish, sharks, seals and seabirds.The mass and distribution of micronekton reflects broad-scale patterns in the structure and function of the ocean, as well as the dynamics of marine ecosystems.

Acoustic mapping is done from fishing and scientific vessels that are equipped with scientifically-calibrated 38 kHz digital echo-sounders that record a slice of acoustic backscatter to a depth of 1500 meters.

Lineage

Statement: I. Data collection:

Vessel calibration:
Vessels are calibrated according to the procedures recommended in the ICES CRR 144 document by (Foote 1987). In the case of ES60 systems, the calibration data is pre-processed to eliminate the possibility of bias of up to +/- 0.5 dB due to the systematic triangle wave error that is embedded in the ES60 data (Ryan and Kloser 2004). This triangle wave error can be significant for calibration data, but for field data it averages to zero over long periods and is not considered a significant source of error. Hence processing to eliminate the triangle wave error from field data is not done.

At a minimum vessels are ideally calibrated annually but logistics may dictate different time intervals.

II. Data management:

In-house tools have been developed to assist with data management and help identify and prioritise subsets of data for post-processing. The data management tool borrows from the open-source multi-beam processing software MB-System (http://grass.osgeo.org/wiki/MB-System) approach by generating from each of the acoustic raw files, a corresponding inf file.

The inf file is in text format and contains the temporal and geographic extent of the associated raw file. The inf files are created during a data registration process using the tool ES60_register.jar. User defined metadata can be included during the registration process (e.g voyage name, vessel name). During registration metadata can be automatically extracted from the binary raw files and included in the inf file (e.g. Echo sounder serial number).

The inf files can be visualised as geo-referenced rectangle blocks using our open-source software Dataview.jar. Dataview.jar has the tools to select blocks of inf files by defining time-windows, spatial extents, and keywords or a combination of these.

III. Data processing:

Data processing is carried out via the following steps:
1. Generate a list of on-transit acoustic files to process using Dataview’s visualisation of inf files.
2. Using Myriax’s Echoview software controlled by a Matlab script via COM objects:
a) Create Echoview ev files using an ev template for manageable blocks for raw acoustic data (nominally a new ev file is created for each 6 hours of raw acoustic data). Note the ev template will have been set up to contain data quality filters and to have the appropriate calibration parameters.
b) Processing to identify and eliminate bad data. In order to calculate both a correct mean Sv and area backscatter (NASC) for the echointegration cell, rejected sample values need to be set to either ‘no data’ or -999 dB depending on which criteria led to the value being rejected.
3. Echo integrate and output to csv format. Echo integration is executed on three different data types. Firstly for the original unfiltered data, secondly for the quality controlled filtered data and finally an output to quantify the number of retained (i.e. unfiltered) samples. The quality controlled filtered Sv values are the ones that should be used as the blessed data. The ratio of retained quality controlled data to original data is used to give a metric of data quality. Similarly, comparisons can be made between filtered and unfiltered Sv values as an indicator of data quality. Data quality is likely to be high where there is little or no difference between filtered and unfiltered Sv values. Conversely where there is a large difference, the data quality is likely to be lower.
4. Convert echo integration csv format data to IMOS Netcdf format, merging in all necessary metadata at the same time. The document SOOP-BA NetCDF manual v1.0.doc details the metadata standard associated with the SOOP-BA data.

Processing to identify and eliminate bad data:
We define two types of noise. Background noise is generally at a consistent value for many pings, but as a minimum is constant throughout the duration of one ping. Intermittent noise consists of signal from unwanted sources that is only present for a portion of a ping. Intermittent noise may only exist for a moment (i.e. at a certain range) within one ping, but may persist across multiple pings at a similar range.

For more information on data collection, data management and data processing, please see attached document.

Notes

Credit
Australia’s Integrated Marine Observing System (IMOS) is enabled by the National Collaborative Research Infrastructure Strategy (NCRIS). It is operated by a consortium of institutions as an unincorporated joint venture, with the University of Tasmania as Lead Agent.

Credit
CSIRO Marine and Atmospheric Research (CMAR)