The data from the tape recorder cassettes were processed by an analogue to digital converter and a programme written by the third author gives a print out of the fish per 1 000 m3 and the fish per hectare in any specified water layer.
Two runs were made as nearly as possible over the same transect in the early evening of 5 September 1980. The second run was completed within 30 minutes of the first run. The results are given in Table 5.
Table 5
The number of fish per hectare in two replicate echo-surveys on 5 September 1980
| Depth layer (m) | Run 1 | Run 2 | Percent difference from mean |
| 4–8 | 118 | 140 | 8.5 |
| 8–16 | 396 | 449 | 6.3 |
| 16–23 | 1 472 | 1 484 | 0.4 |
The variability between the two runs is very small indeed.
The numbers of fish per 1 000 m3 estimated for the appropriate water layers on three dates are given in Table 6. A diagram showing the relation of the trawl and acoustic density estimates is shown in Figure 5. If the two methods gave the same results, the points would lie on the 45° diagonal line. Those points below and to the right of the line are the estimates in which either the trawl underestimated or the acoustic survey overestimated, the density, or both, and similarly those above and to the left of the line are estimates in which either the trawl overestimated or the acoustic method underestimated the densities, or both.
Table 6
The density of fish estimated per 1 000 m3 from the acoustic surveys, for all fish and those with echoes < -49 dB
| Series I (5–6 September 1980) | |||||
| Depth layers (m) | 4–7 | 9–12 | 13–16 | 17–20 | 20–23 |
| Fish density | 10.28 | 2.04 | 13.47 | 24.66 | 22.59 |
| < -49 dB | 2.92 | 0.91 | 11.29 | 9.84 | 0.0 |
| Series II (7–8 September 1980) | |||||
| Depth layers (m) | 6–9 | 6–10 | 16–20 | 17–20 | |
| Fish density | 7.75 | 8.91 | 14.94 | 15.90 | |
| < -49 dB | 5.11 | 7.10 | 5.36 | 3.80 | |
| Series III (8–9 September 1980) | |||||
| Depth layers (m) | 6–9 | 6–10 | 16–20 | 17–20 | |
| Fish density | 14.45 | 13.32 | 18.91 | 20.83 | |
| < -49 dB | 11.35 | 10.38 | 7.82 | 8.01 | |
The three hauls above the line are ones relating to estimates from 4 m to 8.5 m below the surface and hauls made with the small mesh Y.F.T. (Numbers 5, 9 and 14) which contained a great many smelt. It is very probable that the boat towing the echo-sounder frightened these fish from its path, and so the acoustic survey underestimated these fish. The surface hauls with Commercial Vendace Trawl, even though it had a small mesh cover, would allow these smelt to escape (Hauls 7 and 11) through the meshes and the density was therefore underestimated by the trawl. The underestimates with the small mesh Young Fish Trawl are probably due to net avoidance.
The echo-sounder picked up echoes in the range -56 dB to -49 dB, which probably come from fish of 3.5 cm to 8 cm. The Y.F.T. certainly retained fish above this size, but even so the densities of these larger fish were below those obtained by the acoustic method.
The alternative hypothesis, that the acoustic survey overestimated the fish density, is difficult to explain.
The acoustic method appears to give a more reasonable population estimate than the trawl catches except in the surface layers (probably down to 10 m). It is not possible from these experiments to comment on the absolute accuracy of the acoustic method in estimating the actual number of fish present, but the high degree of precision reported in section 7.2, especially at a depth below 16 m, from which multiple echoes were not registered, is very suggestive of good accuracy.
