2.1 Introduction
2.2 Intership calibration using the sea bottom as common target
2.3 Intercalibration on fish layers
2.4 System comparisons - mini surveys
2.5 Concluding remarks
For cooperative surveys it is imperative to have detailed and accurate information on the acoustic instruments and systems used by the participating vessels and to harmonize their use. Only then can one expect to achieve results which are comparable between vessels, fit for a joint total analysis and which can be related to set standards. To this end, all cooperative surveys should include a special programme of instrument calibration and comparisons at different levels. This chapter describes the intercalibration programme for the CECAF 1986 Cooperative Surveys. The participating vessels in these surveys were:
From Spain: CORNIDE DE SAAVEDRA; 60 m; Simrad EK400 38/120 kHz, QDA planned participation from Morocco had to be cancelled due to technical problems with the vessel.
From Senegal: LOUIS SAUGER; 36 m; Biosonics 101, 120 kHz
From Mauritania: N’DIAGO 35 m; Simrad EK400 38 kHz, Agenor
From Norway: DR FRIDTJOF NANSEN; 50 m; Simrad EK400 38/120 kHz, QD, QM
The following elements of calibrations and system comparisons can be identified:
1) Measurements of the electrical properties of the instrument components.For a final biomass estimate by groups of species comes in addition:
2) Measurements of the on-axis sensitivity of the echo sounding- and integrating system.
3) Time-varied gain function of the receiver.
4) Equivalent beam angle of the transducer.
5) Difference in echo integrator output caused by difference in sound frequency used (if any).Points 1 through 3 can be dealt with by each vessel as a preparatory exercise to the cooperative survey. For comparability on-axis calibration should be undertaken by means of a standard target sphere.6) Estimate of the part of the total-integrator output caused by fish as apart from plankton and spurious echoes.
7) Estimates of the species-and size composition of the fish for conversion of integrator output to fish biomass by species or groups of species.
By an intership calibration, i.e. having two or more ships sail over a common target, the same aggregations of fish or an even bottom and afterwards compare their acoustic observations, one can test the systems up to and including point 5.
The total systems can be compared by an inter-system calibration in which two or more vessels undertake a full simultaneous survey of the fish aggregations in a defined area under as far as possible identical conditions of fish behaviour and environment (mini-surveys). In this exercise the vessels should adopt agreed normal working procedures. If only one vessel has been absolutely calibrated inter-ship or inter-system calibration allows the others to be calibrated against it.
In order to obtain a good result of inter-ship calibration, it is necessary to find an area with layered or dispersed fish aggregations preferably of varying density and varying depth where simultaneous runs can be made in an appropriate formation of the vessels.
The success of an inter-ship calibration depends especially on the properties of the fish aggregations worked on. Favourable conditions of fish layers may be difficult to find. The sea bottom may be used as an alternative common target. One must then choose a fairly even bottom where variation in back scattering is not too irregular.
For the full system-intercalibration or simultaneous mini-surveys, two or more ships undertake a full survey of a defined area containing suitable aggregations of fish. Identical or similar survey-tracks should be followed and also the time coverage should be the same, particularly as regards day/night period of work. Fishing for identification and size sampling should form part of the exercise. System-calibration will cover the elements up to 7 listed above, but will in addition include some random survey variability since the ships will not work continuously close up as in inter-ship calibration. It will, however, include comparisons of important routines of both survey execution and data processing and will enable attempts to harmonize these.
The four vessels were available for joint intercalibration exercises during some days at the end of August-beginning of September 1986. Detailed plans were agreed at a meeting in Dakar, 28 August. Since suitable fish aggregations had not been identified within easy reach, it was decided to base the intership calibrations on bottom back scattering. An area on the shelf north of Dakar was chosen and here successive runs were made with each of the four vessels over a distance of about 15 nm. The instrument settings used and the results are set out in Annex 1. These were discussed at various meetings of the participants and the findings are as follows:
Comparison DR FRIDTJOF NANSEN/CORNIDE DE SAAVEDRA
The expected difference from performance and settings would be 9.8 dB, but the data show a difference of about 11.6 dB. Possible causes were thought to be: a) Saturation in the DR FRIDTJOF NANSEN data b) Different beam characteristics c) Inaccurate calibration data. a) was tested by an experiment using full and 1/10 power in bottom integration on the shelf off Panama where the vessel operated in February. No saturation was detected. c) was tested by a new calibration experiment for the CORNIDE DE SAAVEDRA in which the level of SL + VR was found to be 131.4 dB, 1.6 dB higher than previously reported. The main difference between the results from the two systems is thus explained by the differences in performance and settings.
The experiments disclosed a difference in the readings of the nautical logs of the vessels, that of the CORNIDE DE SAAVEDRA giving some 10 per cent too high estimates of the distance. This would not affect the fish abundance estimates, but in order to obtain the best possible correlation in the intercalibration experiment, the two sets of values were plotted over an identical range and new pairs of observations selected. Figure 2.1 shows the regression of these observations. The correlation is good with r = 0.93.
DR FRIDTJOF NANSEN - N’DIAGO
The difference found corresponded well with the estimates based on performance and settings. Figure 2.2 shows the regression with r = 0.96.
LOUIS SAUGER
Although the results from this vessel demonstrated the same general trend as those from the other vessels, a more detailed comparison does not show a good correspondence. This is probably explained by different properties of bottom back scattering with a frequency of 120 KHz as compared with that of 38 KHz used by the other vessels. Further processings of these data were therefore not attempted.
During the night 2nd-3rd August, a ship to ship inter-calibration was made between DR FRIDTJOF NANSEN and LOUIS SAUGER south of Dakar in an area with fish layers and schools of varying density. Also plankton of varying density was present and recorded especially by the 38 KHz system. Three different runs were made of 12, 15 and 11 run respectively and with interchange of lead vessel. Because of low fish- and high plankton values, the first two, four and five observations of each run were rejected. The remaining sets of observations are shown in Table 2.1. With rejection of three unlikely sets of data, the regression obtained is MFN = 1.9 MLS + 66, r = 0.96, see Figure 2.3. Thus the estimates of abundance from DR FRIDTJOF NANSEN must be expected to be about twice those of LOUIS SAUGER.
Table 2.1. Records of observations from DR FRIDTJOF NANSEN
and LOUIS SAUGER during ship to ship calibration 2nd-3rd August
1986.
Units: m2/nm2
|
1st RUN |
2nd RUN |
3rd RUN |
|||
|
LS |
FN |
LS |
FN |
LS |
FN |
|
48 |
70 |
147 |
440 |
279 |
670 |
|
79 |
70 |
263 |
920 |
695 |
1650 |
|
124 |
70 |
450 |
1380 |
2482 |
4550 |
|
71 |
70 |
274 |
180 |
2347 |
9850* |
|
336 |
400 |
226 |
150 |
1282 |
3880 |
|
2639 |
4720 |
70 |
480 |
|
|
|
102 |
290 |
131 |
260 |
|
|
|
87 |
70 |
239 |
580 |
|
|
|
290 |
70* |
415 |
520 |
|
|
|
|
|
784 |
220* |
|
|
MFN = 1.9 MLS + 66
r = 0.96
* Not used in regression.
Simultaneous comparative mini-surveys of defined shelf areas were made both north and south of Dakar. Figure 2.4 shows a fish distribution chart of the area with indications of the species present (data from DR FRIDTJOF NANSEN). The distribution charts of all the vessels show the same main features.
Previous survey results from NW African waters have revealed that night-observations of fish biomass tend to be higher than observations made during the day. Although in general simultaneous, some difference in the day/night coverage of areas of fish occurrence may have taken place between the vessels and this could give rise to some variability.
Mini-Survey No. 1: 28 August-1 September
The shelf north of Dakar between 15°00’ and 15°35’ was covered to 200 m depth by all four vessels in a nearly simultaneous operation. Data from the N’DIAGO have not yet been presented.
CORNIDE DE SAAVEDRA - DR FRIDTJOF NANSEN
The data processed independently gave the following indices of abundance (unit:m reflecting surface/square run).
|
CORNIDE DE SAAVEDRA |
132 |
390 |
m2/nm2 |
|
DR FRIDTJOF NANSEN |
82 |
500 |
m2/nm2 |
A “blind” processing of the DR FRIDTJOF NANSEN data and echo-diagrams by the Spanish team resulted in an abundance index of 71 710 m2/nm2, a value somewhat lower than that obtained originally by DR FRIDTJOF NANSEN.
LOUIS SAUGER - DR FRIDTJOF NANSEN
The transformed observations resulted in the following indices:
|
LOUIS SAUGER |
46 |
931 |
m2/nm2 |
|
DR FRIDTJOF NANSEN |
82 |
500 |
m2/nm2 |
Mini-Survey No. 2: 2-3 September
This took place south of Dakar between N 13°32.5’ and 14°02.5’ from 10 m of depth to W 17°20’ Night time survey was 30 per cent for both vessels. The distance steamed was 178 and 190 nm for LOUIS SAUGER and DR FRIDTJOF NANSEN respectively. The estimates of total fish abundance were:
|
LOUIS SAUGER |
178 |
000 |
m2/nm2 |
|
DR FRIDTJOF NANSEN |
689 |
000 |
m2/nm2 |
The experience from this program of intercalibration and system comparisons, and the results obtained seem to allow the following conclusions to be drawn:
The sea bottom can serve as a suitable common target for intership testing of instruments, at least when a common frequency is used. The application of this method to instruments of different frequencies must be the subject of further studies.
An intership calibration between two of the vessels involving three runs on schools and layers of small pelagic fish was successful in revealing an otherwise unidentified and unexplained difference of nearly 100 per cent which could be caused by a difference in the performance of the two vessels systems and/or a difference in data processing.
Comparison of the simultaneous results of the two vessels using the same instrument systems, CORNIDE DE SAAVEDRA and DR FRIDTJOF NANSEN first of all revealed the importance which must be attached to the process of distinguishing between back scattering from fish and from plankton and other non-fish sources. In tropical waters, plankton often represents a major source of “interference” when using frequencies of the order of 38 kHz. The use of an additional sounder of a higher frequency for identification of fish is therefore an advantage. Dual frequencies were only available to one of the vessels, and an analysis by the Spanish team of the data from DR FRIDTJOF NANSEN gave results largely similar to the original estimate.
The mini-survey comparison between LOUIS SAUGER and DR FRIDTJOF NANSEN for which a difference in output of about 100 per cent had been revealed, showed consistently higher abundance estimates for DR FRIDTJOF NANSEN, the ratios being 1.75 and 3.87. The latter is nearly the double of that established by inter-calibration, but obtained in an area with high density schools and resulting high variability of the estimates.
A final conclusion would be that the results clearly demonstrate the importance of giving programmes of intercalibration and other systems comparisons high priority in cooperative survey work.
