Anthony D Hawkins

Like other members of the cod family (Gadidae), the haddock is a noisy fish. The male produces a diversity of sounds over the spawning season, with distinctive sounds associated with particular behavioural acts (Hawkins and Amorim, 2000). This behaviour offers the opportunity to detect the presence of spawning fish simply by listening for sounds. A search was therefore carried out in coastal waters by listening for the characteristic sounds of haddock. By this means an aggregation of spawning haddock was located at the upper end of Balsfjord, a sub-arctic fjord in Northern Norway.

Spawning of captive haddock was observed in a 10m diameter annular aquarium tank at the FRS Marine Laboratory Aberdeen. Water depth was 1.5m and the water temperature was maintained at 8° C. The fish were observed from above by means of a low light level TV camera and their behaviour recorded on a time-lapse video tape recorder.

The sounds of the fish were detected with an omni-directional broad-band hydrophone (ITC, 6050C), amplified (Stanford SR560 pre-amplifier, bandwidth 30Hz to 10kHz), sampled at 8 kHz and recorded directly as WAV files on the hard disk of a lap-top computer by means of Avisoft Recorder. Sound analysis was performed with Avisoft SASLab Pro.

In the aquarium, individual female haddock spawned repeatedly over several weeks. Spawning was accomplished through a close spawning embrace, and preceded by elaborate courtship behaviour. Sounds were recorded in the aquarium from male and female haddock, and from juveniles. However, during the spawning period sounds were predominantly made by the male fish.

Haddock sounds have been described as a series of ‘knocks’ (Hawkins & Chapman, 1966), repeated regularly at different rates. It has since become apparent that each knock can be subdivided into two short, low-frequency pulses of sound, spaced closely together (Hawkins & Amorim, 2000).

The individual knocks produced by male fish were regularly repeated at a range of different rates, depending on the behaviour of the fish. Short sequences of repeated knocks were emitted during agonistic encounters. At spawning time, male fish produced much longer sequences, lasting from several seconds to several minutes, the knocks being produced at intervals varying from 500 ms to 30 ms. At the very fastest rates, with intervals of less than 50 ms, the sounds were heard as a continuous humming. Different behavioural acts leading up to the spawning embrace were associated with different repetition rates. This rich diversity of sounds produced by the male haddock appears to be characteristic of this species.

Male haddock showed a distinctive solitary display. Dominant males adopted a characteristic pattern of pigmentation, occupied a favoured area and showed a characteristic pattern of movement, moving in tight circles or a figure of eight. During this behaviour the male uttered an almost continuous train of regularly repeated knocks repeated at intervals of between 140 and 60 ms (Figure 1). During the spawning season males spent much of their time in solitary display (9h out of 24, 75% at night), interrupting the display only when other fish entered their territories.

Figure 1 Repetitive knocks from a male haddock during solitary display

Differences between the sounds of individual male haddock were analysed by measurements on the waveform or through wavelet analysis (see Wood, this workshop). In most instances the knocks were composed of two pulses separated by intervals of 10-20ms. The two pulses often differed in frequency, the first being higher than the second. Within a given call, or from day to day, there was little variation in the waveform for an individual fish. From month to month, however, there was a significant change in the detail of the waveform, though the double pulse structure was usually retained. There were often striking differences between the sounds of individual males (Figure 2).

Figure 2 Waveforms of ‘knocks’ from three individual males (A, B & C)

The characteristic sounds described from haddock in the aquarium provided clear criteria for the location of spawning male haddock in the sea. The short low frequency sounds (below 1kHz), made up of two pulses separated by intervals of 10 — 20 ms, regularly repeated at intervals of 300 — 30 ms, often for more than several seconds, provided unequivocal evidence of the presence of haddock. Moreover, changes in the repetition rate of the sounds were indicative of different stages in the behaviour of the haddock. The sounds were quite different from those described for other gadoid fish (Hawkins & Rasmussen (1978).

Searching took place in Balsfjord, Tromsø, Northern Norway from a small research vessel (the FF Hyas, Norges Fiskerihøgskole, Tromsø, length 12m). Balsfjord is a subarctic fjord (90 km²) with a 30m sill at its entrance, with depths in the inner fjord dropping to 190m. Trawling surveys have shown a preponderance of cod (Gadus morhua), but also significant catches of haddock. There is a small local fishery for cod and haddock in the spring, and reports from fishermen suggested that haddock spawned at the head of the fjord.

Sampling took place when individual fish echoes were detected on or close to the seabed on the echo-sounder. The ship was stopped, anchored by the bow, and the hydrophone hung 2m above a weight on the seabed by a cord attached to a small submerged buoy. Sound recordings were made for a minimum of 15 minutes at each station with the main engine and auxiliary generator of the ship shut down. The position of the ship was determined by GPS.

Four surveys were carried out at Balsfjord (17-19 April 2000; 10-12 May 2000; 9-10 December 2000; 3-7 April 2001). The distinctive sounds of haddock were detected at some time during each of the 4 surveys, though not at all stations and with varying incidence.

In the first survey, 12 stations were examined. Distinctive haddock sounds were recorded at the majority of stations within the main basin, especially close to the head of the fjord. Slowly repeated knocks, made up of double pulses, were the most common (Figure 3). Some were short (a few seconds), others extended over several minutes. Occasionally, sounds with a faster repetition rate were recorded, suggesting that the haddock were engaging in agonistic and courtship activities. At one station in the main part of the fjord, repeated grunts were recorded which lacked the double pulse structure characteristic of haddock. These were tentatively identified as coming from cod.

Sounds were recorded at all times of the day and night. However, in two areas a continuous low frequency rumbling was detected at night, within which individual haddock knocks could be detected.

During the second survey, sounds were recorded at four of the five stations surveyed. All stations at the head of the fjord yielded haddock sounds, and at three of them, the low frequency rumbling sound was audible at night. At one station it proved possible to record for 10 minute periods every hour over a 24 hour period. This revealed a 10 dB increase in ambient noise level at night, which was attributable to the simultaneous production of sound by many haddock.