ANNEX II State of acoustic instruments and conversion of echo-intensity to fish abundance.

While two echo-sounder/integration systems were used (38 kHz and 120 kHz) the outputs from the 120 kHz system provided the data for computing the indices of echo-abundance of fish. Nearly all fish is found in waters of less than 100 m depth and this high frequency sounder was found to be less influenced and biased by plankton scatterers than the 38 kHz system. The following settings were used:

Basic range: 0-100 m, Transmitter 1/1. Bandwidth and pulse length: narrow, 0.6 m/sec TVG and gain 20 logR-0 dB, Recorder gain 3, Integrator treshold 8, Integrator gain 20 (x10), Integrator channels 4-25 and 25 m-bottom or 4-50 m and 50 m-bottom.

The acoustic systems of the "Dr.Fridtjof Nansen" were intercalibrated with those of R/V "Johan Hjort" in March 1978. Calibrations of the source level and the voltage response of the sounders have been made regularly since then and have demonstrated that the systems have remained stable. The intercalibration resulted in a conversion factor C of 10 tons/n.m² with reference to a cod type fish size of 40 cm in length. The conversion factor should, however, be adjusted according to the size of the fish because the scattering cross section per unit weight decreases approximately linearly with length. Ideally each acoustic observation of fish density should be related to the size composition of the fish within the area of observation but this is not practicable. For assessment of the fish size the length sampling of the catches were analysed as follows:

For small pelagic fish the samples were grouped by families, areas and seasons. The mean lengths were:

	No of samples	Mean length cm:				Tennasserim
		Arakan		Delta
		Aut.	Spring	Aut.	Spring	Aut.	Spring
Carangidae	42	13.1	11.9	19.2	22.1	14.7	21.9
Clupeidae	42	12.1	12.8	13.2	13.7	11.4	18.1
Engraulidae	44	6.8	8.6	8.5	8.4	6.9	8.9
Scombridae	25	10.3	23.1	15.6	20.7	16.8	16.1
Weighted mean		10.4	11.3	11.1	11.6	7.6	10.1

For each area and season the mean lengths were then weighted with the corresponding proportion of the taxonomic group in the total catch. This procedure is not strictly correct, but will give a more representative mean length than an unweighted mean. Since the variations in length are relatively small and the fish is only one of several factors which contribute to a variance in the conversion factor it was decided to use an overall mean of 10 cm as representing the size of small pelagic fish.

For the demersal and semi-demersal fish a similar procedure was used, but the observations from each season were pooled.

	No of samples	Mean lengths cm:		Tenasserim
		Arakan	Delta
Scianidae	50	22.9	31.3	18.8
Pomadasyidae	60	25.2	37.4	29.4
Leiognatiadae	56	9.5	6.2	12.6
Lactaridae	12	16.7	15.5
Nemipteridae	53	16.1	16.8	19.0
Harpadontidae	2		17.5
Synodontidae	8	23.3	17.9	16.8
Lutjanidae	25	61.3	43.0	53.5
Mullidae	8	15.2	13.9	19.9
Ariidae	2		62.1
Muraenesocidae	8		133.4
Polynemidae	18		73.5	93.0
Weighted mead			39.2	22.0

The mean lengths were weighted as for small pelagic fish according to the groups contribution to the overall total catch of demersal fish. This procedure will result in an overrepresentation of large-sized fish as has probably been the case in the Delta area. It was therefore decided to use an overall mean of 20 cm as representing the size of "other fish". The respective conversion factors C will accordingly take the values of 2.5 for small pelagic fish and 5.0 for "other fish"