Trapping and counting fish

The design of a fish-trap is simple, age-old and used worldwide. It is a cylinder blocked at one end, with a funnel-shaped entrance (Fig. 8.6). I used a heavy-duty, plastic gauze. Each trap was 50 cm long with a diameter of 27 cm. One side of the cylinder consisted of a funnel ending in an opening 7 cm across; at the other end of the cylinder was a small flap for removing fish. Inside we tied a large stone, which kept the trap on the bottom and prevented it from moving about too much. The traps were not baited, attached to a rope, and on days when my students and I were fish sampling we threw ten or twenty of them from the shore into the sea below the low-tide mark, and checked them a day later.

Later, the model was improved by making the cylinder longer (80 cm) and adding a second funnel in line with the first one so that the trap effectively had two chambers: an entrance porch and a holding chamber. This twin-funnel type was more effective in holding fish once they were caught (Kruuk et al. 1988). In 1493 'trap-nights', at all times of year, we caught 414l fish and 1122 crabs, a total of 5263. For further analyses, these figures needed various correction factors, to allow for trap type and for the variation in 'trappability' between species (see below). Table 8.1 gives an overall impression of the species composition of the fish fauna in the Shetland study area, but it does distort.

One of the first things I wanted to know was whether the species that were relevant to otters were, indeed, more active at night, as the aquarium

Figure 8.6 Standard plastic two-funnel fish-trap in situ underwater, for sampling fish populations in coastal habitats.

observations suggested. One would expect to catch more overnight in the traps than during daytime and, as shown in Figure 8.7, there was a striking difference, confirmed for several species by a Finnish study (Westin and Aneer 1987).

Table 8.1 Fish caught with funnel traps in otter study area at Lunna Ness, Shetland 1983-1987

Species

% of total (n = 4141)

Mean weight (g)

Eelpout Zoarces viviparus

44

11

Butterfish Pholis gunnellus

8

10

Five-bearded rockling Ciliata

13

27

mustela

Northern rockling Ciliata

1

27

septentrionalis

Shore rockling Gaidropsarus

2

81

mediterraneus

Three-bearded rockling

3

70

Gaidropsarus vulgaris

Sea scorpion Taurulus bubalis

2

16

Bullrout Myxocephalus scorpius

1

89

Sea stickleback Spinachia

5

3

spinachia

Saithe Pollachius virens

13

25

Pollack Pollachius pollachius

3

19

Eel Anguilla anguilla

1

58

Yarrell's blenny Chirolophis

<1

n.d.

ascanii

Two-spotted goby Gobiusculus

2

n.d.

flavescens

Montague's seasnail Liparis

<1

n.d.

montagui

Topknot Zeugopterus punctatus

<1

n.d.

Plaice Pleuronectes platessa

<1

n.d.

Cod Gadus morhua

<1

n.d.

Poor cod Trisopterus minutus

<1

n.d.

Three-spined stickleback

<1

n.d.

Gasterosteus aculeatus

Lumpsucker Cyclopterus

<1

n.d.

lumpus

Conger eel Conger conger

<1

n.d.

Pipe fish Syngnathus acus

<1

n.d.

Dogfish Sycliorhinus canicula

<1

Figure 8.6 Standard plastic two-funnel fish-trap in situ underwater, for sampling fish populations in coastal habitats.

Eelpout Butterfish -Rockling Eel

Sea scorpion Sea stickleback -Crab

Total bottom fish

I Night catch ] Day catch

No. per trap per 12 hours

Figure 8.7 Daytime and night-time fish-trap catches of Eurasian otter prey in Shetland over 108 days show that the main prey species are significantly nocturnal (P < 0.001, Mann-Whitney U test).

Along the sea shore, animal life appeared to be geared to the state of the tide. It affected the landscape, the birds, the places where we could walk, our ability to see otters and the activity of the otters themselves. The question was whether it affected the activity of fishes. Traps were left out, well below the low-tide line and in the areas where otters habitually foraged, for 3 hours around high tide, 3 hours around low tide and 3 hours for periods in between when the water was rising or falling. Time of day had to be allowed for, so that we had tidal observations both at night and during daytime.

Again, the results were significant: in general, and allowing for the time of day and night, bottom-living fishes and crabs were most active at high tide (0.30 fish caught per 3-hour period per trap) and least active at low tide (0.17 fish per 3 hours), with catches at falling and rising tides intermediate. Shore crabs, too, were least active during the ebb tide (Kruuk et al. 1988). Clearly a predator that specialized in catching bottom-living fish or crabs during their inactive period would find most available prey at low tide, and in daytime.

There were other variations in our fish catches that were more difficult to account for, and probably nothing to do with tide, time or season, and nothing obvious with the weather. Some days all of our traps would catch many fish, of different species, and perhaps the next day with ostensibly the same weather, none would catch anything. One could expect stochastic variation in the numbers, but it was obvious that the observed variation was much greater than that. This is something that every fisherperson knows about, and there must be as many different explanations for these fluctuations in the catch as there are fisherfolk. This variation is just as likely to affect the otters' fishing success as it did our own.

As one would expect, there also were differences in fish numbers along the coast, between bays, shallow sloping coasts, or steep cliffs falling into deep water, between a sandy bottom or large rocks, and many different vegetations of algae. We set fish-traps in various places in the study area that represented the different coastal types, and analysed how the catch in our traps was associated with these differences.

Every species of fish showed its own pattern (Kruuk et al. 1988, 1995). For example, the eelpout was most common along sheltered coasts (i.e. there was a negative correlation with 'exposure'), it avoided Gigartina stellata, a typical seaweed of exposed shores, but it was to be found especially amongst the knotted wrack Ascophyllum nodosum and bladder wrack Fucus vesiculosus, the typical vegetation of the sheltered voes. Shore crabs showed the same distribution as eelpout. However, the other important prey of otters, the five-bearded rockling, was found closely associated with algae of the really exposed coasts in Shetland, Gigartina, and the thong-weed Himanthalia elongata. The other rocklings showed even more avoidance of sheltered areas with Ascophyllum, but they were associated with the indicators of exposure, and with large boulders.

Butterfish and bull-rout were found everywhere, with no particular preference for, or avoidance of, any of the factors measured. However, the sea scorpion was a fish of the wilder shores, with almost exactly the same habitat preference as the five-bearded rockling. The two more mid-water fish, saithe and pollack, showed another pattern again; they were found along steep slopes where the water was deep, but with a preference for sheltered areas rather than exposed ones.

Usually the traps were set for 1-2 metres deep at low tide. The otters, however, fished over a much wider zone, sometimes where the water was more than 10 metres deep, or in more shallow places. They would spend different amounts of energy according to the depth at which they were fishing, and their success rates varied with depth (see Chapter 9). It was important, therefore, to discover what the differences in fish numbers were over a range of depths. We compared the catches in a line of six trap sites, from l.5 to 11 metres deep, in one of the bays, and in summer only (Kruuk etal. 1988).

There was considerable variation in the occurrence of different species of fish with depth (Fig. 8.8). Eelpout and butterfish were more common in shallow water, but not significantly so, whereas five-bearded rockling, sea scorpion and shore crabs were found only in places that were less than 4 metres deep at high tide. The larger fish, such as three-bearded and shore rocklings, occurred more often deep down, with some other larger species. Overall, there were many more fish in the shallow strip along the shore, but the mean weight per fish increased with depth, and consequently the total biomass of fish caught per trap did not change significantly with depth.

One of the important implications of these observations for a predator is that, in order to exploit eelpout as well as rocklings, sea scorpions as well as saithe, access is needed to different types of shore, to exposed areas as well as sheltered voes, to steep coasts as well as gently sloping ones. Greater depths

Figure 8.8 Eurasian otter prey species in Shetland, in fish-traps at various depths in catches over 15 days in June 1987.There were more fish in shallow waters, but they were smaller. (After Kruuk etal. 1988.)

Figure 8.8 Eurasian otter prey species in Shetland, in fish-traps at various depths in catches over 15 days in June 1987.There were more fish in shallow waters, but they were smaller. (After Kruuk etal. 1988.)

should be fished if larger prey are required, for example to provision cubs, and shallow waters, close inshore, are to be exploited for larger numbers of small fish. Ideally, all such areas should be included in an otter's home range.

Seasonal variation in fish numbers, distribution and weights also played a significant role. Otters need to exploit a range of different species of fish, rather than specialize in just one of them. This became evident when we compared the seasonal availability of each species over a 4-year period (Kruuk et al. 1988). The data showed that, overall, most species of bottom-living fish (except for the five-bearded rockling) were least common in late winter and spring—a tough time for otters (see Chapter 12). However, every fish species had its own pattern of seasonality, and there was no significant overall correlation between them.

Every year we found a large increase in the total catch of fish in traps during August (see below). This was due largely to one species, the eelpout, which completely dominated the underwater scene during that month. There were no clear seasonal patterns for the five-bearded rockling, but butterfish was most abundant in summer, sea scorpion and shore crabs in autumn, and saithe in winter. Only the sea stickleback, a fish that is a rather small and bony prey for an otter, was clearly more abundant in spring than at any other time.

This seasonality stood in marked contrast to observations of prey for Eurasian otters along the Portuguese coast. Pedro Beja (1995a and b) established that the otters took mostly corkwing wrasse Symphodus melops, several shannies, blennies and gobies, the shore rockling Gaidropsarus mediterraneus and the conger eel Conger conger. Using fish-traps of similar design as those in Shetland, as well as hand-netting and angling, he found that all of these species were present in larger numbers and were also heaviest during winter and early spring, but least available during late summer and autumn.

In Shetland there were also considerable differences in fish populations between subsequent years, although the seasonal pattern generally remained the same. We put out numbers of traps during August, from 1983 until 1988, and there were large, significant differences in the size of the eelpout 'glut' during that month. This variation was highly important to

the otters, especially to the breeding females (see Chapter 11).

To complicate matters, the seasonality of the various fish species was different along various sections of coast, especially for the otters' two main prey species: five-bearded rockling and eelpout. Thus, to exploit a particular prey species efficiently, an otter would have to fish different parts of the coast at different times of year. A two-way analysis of variance showed that this complicated interaction between seasonal and area effects was highly significant for eelpout, five-bearded rockling, three-bearded rock-ling, saithe and pollack (Kruuk et al. 1988). The picture of seasonal food availability was enhanced by fluctuations in the weights of fish, with median weights of eelpout being about twice as high in August at 16 g, compared with 9 g for January. Rocklings showed a steady 25-30 g throughout the year, but in June butterfish more than doubled their weight compared with that in January (12 versus 5 g).

One problem with the interpretation of catch size from fish-traps is that the species differ in the likelihood that an individual will get caught, in their 'catchability'. This means that differences between trap catches (for example, in seasonality, depth, or the effects of tide and time of day) are largely relevant as they stand, uncorrected. However, to interpret trap catches in terms of actual prey density and biomass, corrections need to be applied. To this end, I compared some of the catches with what we found in the area immediately around the traps. This was done by intensive searching, hand-netting, snorkelling and scuba-diving by several people over an area of 20 m2, where we assumed that we caught every fish present (Kruuk etal. 1988).

We discovered that, in an area where we would catch 10 eelpout per trap per night, there would be 6.7 eelpout per 10 m2; 10 butterfish per trap per night corresponded with 3l.l butterfish per 10 m2; and 10 rocklings per trap per night with only 0.6 rocklings per 10 m2. These differences were due to the variation in behaviour of the species: rocklings are more catchable. The comparison enabled us to calculate correction factors, to translate numbers caught in traps into actual fish densities. Unfortunately, we did not get sufficient data on other species. From the numbers of fish caught, their weights and correction factors, I made an approximate estimate of monthly changes in fish biomass in the otters' habitat (Fig. 8.9). In general, fish biomass was higher in summer (over 10 g/m2) than in winter or early spring (about 5 g/m2), with a peak in August of more than 60 g/m2, a veritable glut, due mostly to the annual invasion of eelpout.

We did one small experiment to see what the effects would be when predators, such as Eurasian otters, removed a large number of fish from a small patch. The question was: would new fishes repopulate? This was relevant when studying otters that were repeatedly fishing in the same small patches along the coast, a common strategy (see Chapter 9). During some very low spring tides, we removed and counted all the fish we could find, from under rocks and seaweeds in one area of 20 m2, at the water's edge. There were 16 fish (butterfish, eelpout and rockling), and we caught and removed 11. The next day we did the same, and again found 16 fish, catching and removing 14 of them. For the count on the third and final day we found 20 fish, a slight increase due to a larger number of butterfish (Kruuk et al. 1988). It suggested that fish caught by a predator from a small 'patch' are replaced within 24 hours. Similar results were obtained for butterfish on the Scottish west coast (Koop and Gibson 199l). It is likely that such a patch has a given number of suitable sites for fishes sheltering under rocks, and vacancies are filled up rapidly.

Figure 8.9 The biomass of available otter prey (eelpout, rocklings and butterfish) in Shetland, by month, as calculated from trap catches. There is an annual invasion of prey (mostly eelpout) in summer, with a glut in august.

The results from Shetland on Eurasian otter prey are useful not only to explain behaviour and ecology of the predators on site, but also to suggest similar approaches elsewhere. Based on the use of conventional-type fish-traps, they are subject to problems with interpretation (being dependent on species-specific fish behaviour), but there are ways round this, and these studies enable comparisons of prey availability between areas, seasons, times of day and other variables. Chapter 9 shows that the trapping results explain several aspects of otter behaviour, such as time of day and tide when they forage, and how prey behaviour enables otters to forage repeatedly in the same patches of habitat. They also explain Eurasian otters' seasonality of reproduction (see Chapter 11), mortality (see Chapter 12), and are relevant for understanding sprainting behaviour (see Chapter 6).

Was this article helpful?

0 0

Post a comment