Effects of Carp, Cyprinus carpio L., on Communities of Aquatic ...

0 downloads 0 Views 897KB Size Report
Present address: Resource and Planning Branch, P.O. Box 41, East Melbourne, Vic. ... Victoria, Australia to study the effects of carp on invertebrates (Fletcher, ...
Aust. J. Mar. Freshw. Res., 1985, 36, 311-27

Effects of Carp, Cyprinus carpio L., on Communities of Aquatic Vegetation and Turbidity of Waterbodies in the Lower Goulburn River Basin Andrea R. Fletchee, A . K. MorisonAand D. J. HumeB Arthur Rylah Institute for Environmental Research, 123 Brown Street, Heidelberg, Vic. 3084. Present address: Warmwater Ecology Program, 22 High Street, Shepparton, Vic. 3630. Present address: Resource and Planning Branch, P.O. Box 41, East Melbourne, Vic. 3002.

A

Abstract Densities of carp, ranges of turbidity, and details of communities of aquatic vegetation from 1979 to 1982 are given for several waterbodies in the Goulburn River valley including the Broken River, near Shepparton, Victoria. The turbidity values at all sites were high, typical of Australian inland waterbodies. There was no association between high carp densities and high turbidity, and populations of carp did not appear to increase turbidity. Observed turbidity increases at each site appeared to be related to hydrological changes. Fluctuation of water levels was also an important factor determining the extent of aquatic vegetation communities. However, circumstantial evidence is presented that shallow-rooted and soft-leaved aquatic vegetation such as Potamogeton spp. have been reduced by carp.

Introduction McCrimmon (1968) stated that the feeding activity of carp disturbs bottom sediments and probably increases turbidity. Excessively turbid waters absorb more heat and light than clear waters (Ruttner 1952), increase sediment deposition, and have less aesthetic value (Anon. 1972). An increase in turbidity may affect plant growth by reduction of photosynthesis due to a decrease in transmitted light (Ruttner 1952). Carp are also reported to affect macrophytes by uprooting plants during feeding (Cahn 1929; Tryon 1954; McCrimmon 1968) or by direct feeding (Anderson 1950; King and Hunt 1967). One approach used in the United States to determine whether carp have had these effects, has been to set up experiments with carp in enclosures (Thrienen and Helm 1954; Mraz and Cooper 1957; Robe1 1961; King and Hunt 1967; Forester and Lawrence 1978). This experimental design was repeated by Crivelli (1983) in France to compare results with the studies from the United States. Similar experiments were conducted in ponds near Shepparton, Victoria, Australia to study the effects of carp on invertebrates (Fletcher, unpublished data). Another approach has been to monitor changes in parameters such as turbidity and vegetation in the field, associated with different densities of carp and with the removal or addition of carp (Jessen and Kuehn 1960). Only circumstantial evidence relating to the effects of carp on the environment has been documented in Australia and reports have been contradictory. Roberts (1969) observed a decrease in plant growth after the introduction of carp in a lake in Gippsland, Victoria. Malcolm (1971) stated that there were no differences in macrophyte growth in waters with or without carp in Gippsland. There has also been speculation as to whether carp have increased the turbidity of Australian waters. Butcher (1967) and Roberts (1969) believed that the waters that now have carp have become more turbid but they had no measurements to support this. Malcolm (1971) recorded some increases in the turbidity in farm dams after stocking with carp in Gippsland. But Apps (1976) stated that the water quality of lakes containing carp had not decreased. 0067-1940/85/030311 $02.00

Andrea R. Fletcher et al.

In the study reported in this paper, nine lentic sites and one lotic site with various densities of carp in the Goulburn River catchment of northern Victoria, were visited regularly from 1979 to 1982. Communities of aquatic vegetation are described and, for some sites, seasonal changes outlined. Influence of carp, from existing carp populations and in one site from carp that were experimentally introduced, on aquatic vegetation communities is examined. Turbidity was measured quantitatively at all sites over 2 years. This gives a better evaluation of any relationship between turbidity of waters and carp populations than previously reported in Australia.

I

Locality Map

Fig. 1. Map of the Shepparton area, showing the location of sampling sites within the catchment of

the lower Goulburn River, Victoria.

Methods Study Sites

Study sites were selected near Shepparton, northern Victoria (Fig. 1) and included nine lentic habitats (Lake Cooper, Greens Lake, Loch Garry, Lake Mokoan and five riverside billabongs along the Goulburn River in Darcys State Forest) and 10 stations along the Broken River (which runs into the Goulburn River at Shepparton).

Effect of Carp on Vegetation and Turbidity of Waterbodies

Lake Cooper, a small exposed lake of 1106 ha and 3 m maximum depth, is relatively shallow with brackish water (conductivity up to 7 8 0 0 ~ cm-I). s Greens Lake has an area of 1070 ha, a maximum depth of 7.5 m and fresh water. Lake Mokoan is an impoundment which floods a large area (7760 ha), covering three swamps to a maximum depth of 7.3 m. The lake is an off-river storage of water diverted from the Broken River for flood and irrigation control. The five billabongs studied are on the floodplain of the Goulburn River in Darcys State Forest. We named them Waugh Road, Pogues, Dead Tree, Black and Bruces billabongs. Their form varies from river cutoffs (Bruces) to shallow (1.0 m) swamp areas (Waugh Road). The other billabong studied was Loch Garry (80 ha, 3.5 m maximum depth), an old channel of the Goulburn River, in Loch Garry State Forest. The Broken River is wide (5-30 m), with a substrate of sand, silt and clay and an average gradient of 1:1500 along its sinuous channel. The 10 sampling stations were spread between a weir at Benalla and the junction with the Goulburn River, 90 km downstream. In analyses of turbidity described later, the 10 stations along the Broken River were reduced to five sections, each defined by artificial barriers that would, at most times, restrict movement of carp to within a section. These sections were: 1, 0-24 km, to Goulburn Channel; 2, 24-44 km, to Gowangardie Weir; 3, 45-80 km, to Caseys Weir; 4, 81-89 km, to Benalla Weir; and 5, 90-100 km, Benalla to Lake Mokoan inlet channel. The distance upstream along the river from each site to Benalla was measured from National Topographic maps (1:100 000) using an opisometer.

Densities and Relative Abundance of Carp at Each Site Relative density of carp was estimated from information gathered from surveys, described below, at each site. Additionally, in the Broken River, a value of catch of carp per unit effort (total numbers of carp per hour of electroshocking) was calculated for each of the 10 sites. The carp population was estimated at three sites - Lake Cooper, Loch Garry and Pogues billabong - using a modified Schnabel method (Ricker 1975). At these three sites, the methods of capture included gill nets with mesh sizes 4.0-25.5 cm, rectangular traps (100 by 90 by 50cm) with mesh size 2.2 by 3.5 cm, and a 100-m seine. Gill nets and a beach seine were most successful at Pogues billabong. At Loch Garry, traps were checked every week, with a total of 1025 trap hours over 2 years. At Lake Cooper, a beach seine was the most successful method for capturing carp. These were hauled at least every 6 weeks, usually every 2 weeks during summer months. Population estimates based on numbers of recaptured fish were calculated for selected size classes of fish. These had previously been either marked (fin clipped) or tagged with a plastic anchor tag. Population estimates of other size classes were made by using the same relative proportions of each size class in the total population shown by length-frequency distributions. The total weight (kg) of the population was estimated by combining estimated mean weights of each size class. Biomass density was computed using this estimate of total weight and the area (ha) of each waterbody. The range of biomass density was computed using the 95% confidence limits of the population estimate. These confidence limits for the total number of recaptures had been substituted for R, (number of recaptures in the catch of the day) in the Schnabel equation to give the confidence limits for the estimate of N (number in the population) (Ricker 1975). At all the remaining sites, the fish populations were regularly sampled from February 1980 to March 1982 (either every 6 or 12 weeks) to obtain an estimate of the relative abundance of carp. All gear and their performance are described by Morison et a / . (1985). Methods for sampling the billabongs included use of a backpack-mounted electroshocker, a boat-mounted unpulsed electroshocking unit powered by a 6.5-kVA generator, a 30- and a 100-m beach seine, and gill nets with mesh sizes from 4.0 cm to 25.5 cm. A permanent trap, 3 m by 3 m with two wings 5 m long, was erected in Waugh Road billabong. The Lake Mokoan outlet channel was sampled with a backpack-mounted electroshocker. Lake Mokoan and Greens Lake were sampled with a 100-m seine and gill nets with mesh sizes of 4.0-25.5 cm. In the Broken River, a total of 114 h of electroshocking was undertaken during the period May 1980 to March 1982. It was not possible to calculate population estimates in these sites. Also as different fish capture techniques were used in different waterbodies, values of catch per unit effort could not be compared. However, allowance could be made for these different methods so that a reasonable ranking of all sites was possible, based on relative dominance and apparent density of carp. Possible effects of carp on vegetation was assessed at Waugh Road billabong by increasing the biomass of carp in the billabong. Initially it was thought that no carp were present. The billabong had constant water levels with abundant macrophytes. Carp, collected by seine in Pogues billabong and tagged with numbered plastic anchor tags inserted below the dorsal fin, were released in Waugh Road billabong on 24 June 1981 (Table 1). Additional tagged fish were introduced on 21 September 1981 from Loch Garry, and on 20 October 1981 from Pogues billabong because flooding from the Goulburn River

Andrea R. Fletcher et a/.

during June 1981 and August 1981 might have allowed carp from the initial stocking to escape (Table 1). The total biomass density of carp stocked was 38 kg ha-I, although the final density of stocked carp could have been 16-38 kg ha-I depending on the number of fish that had escaped. To monitor the population, a trap was set up in August 1982 and checked each week for 38 weeks. Because recapture rates were low, the billabong was also electroshocked every 8 weeks.

Turbidity In the period February 1980 to March 1982, water samples were taken to measure turbidity. Samples were taken every 2 weeks at the five billabongs, Loch Garry, Lake Cooper and in the Broken River at Gowangardie Weir (44 km upstream from the mouth); every 6 weeks at Lake Mokoan; and every 12 weeks at Greens Lake and the other nine sites along the Broken River. Samples of water were collected in 500-ml plastic bottles and returned to the laboratory. Bottles were shaken before analysis, which usually occurred within 2 weeks of collection. Turbidity was determined by a Fisher turbidometer (DRT100) and expressed as nephelometric units (NTU). The mean, one standard deviation and range of turbidity measurements for all samples were calculated for each site. The outlet channel from Lake Mokoan was also included as a site. This totalled 11 sites whose waters were ranked from high to low turbidity, using the mean values. The significance of the difference between each of the mean values was determined by a Student-Newman-Keuls (SNK) Table 1. Number, weight and biomass density of carp released into Waugh Road billabong (6.7 ha), 24 June-20 October 1981 Date 1981

24 June 21 September 20 October Total

Number

Mean weight f s.d.(kg)

Total weight (kg)

Biomass density (kg ha-')

81 5 48

1+8+1.0 1.7f0.6 2,121.2

146 8 103

22 1 15

257

38

134

procedure, which is a stepwise method using the range as the test statistic. To determine the relationship between carp density and turbidity, the ranking of sites by turbidity was compared, using Spearman Rank correlation, with the ranking of the same sites by carp density. In the Broken River, a multiple correlation was also performed for the 10 stations using the catch of carp per unit effort, mean turbidity values at the station, and distance from the mouth of the river. For one station (below Gowangardie Weir), a correlation between mean monthly values of relative depth with turbidity was calculated. Depth measurements were available monthly for Lake Cooper and daily for the Broken River, at Caseys Weir, 80 km upstream from the confluence with the Goulburn River, from permanent depth gauges read each month by the Rural Water Commission (Bureau of Meteorology, Melbourne). Graduated poles were placed at every site and read at least every 2 weeks to give fluctuations in levels. A mean monthly depth was calculated for each site. A correlation between turbidity and depth fluctuations was calculated.

Vegetation Aquatic and marginal plants were collected from each site and identified according to Willis (1970, 1972), Aston (1973) and Sainty and Jacobs (1981). Species were grouped according to their occurrence together, with a dominant species designated to each group. Persistence and seasonality of these groups from 1980 to 1982 were determined by observations and photographs every 6 weeks. Distribution of these groups in autumn was further checked at Loch Garry in 1980, and at the five billabongs in 1980 and 1981, as groups were distinguishable in aerial photographs taken in March of these years. At Lake Cooper, anecdotal information from locals was obtained regarding the type of vegetation present before the spread of carp into the lake. The vegetation of each of three billabongs in Darcys State Forest was studied in detail to allow comparisons between the vegetation of small waterbodies containing various densities of carp. The changes in distribution of vegetation groups in Waugh Road billabong were monitored, as part of the field experiment to determine the effects of introduction of carp. Pegs were positioned, after random

Effect of Carp on Vegetation and Turbidity of Waterbodies

selection, along the shoreline of the billabong to form transects. Four transects were measured across Waugh Road billabong, in June 1980, January 1981, June 1981 and February 1982; two transects across Black billabong, in May-June 1980 and January-February 1981; and three transects across Dead Tree billabong, in May-June 1980 and January-February 1981. The vegetation along these marked transects was recorded, and again after 6 months. Cover abundance values according to Braun-Blanquet (1928) were recorded for each species within 0.25-m2 quadrats at 10-m intervals along each transect.

Results

Densities and Relative Abundance of Carp at Each Site Population and biomass density estimates varied greatly (Table 2). For stations along the Broken River, the catch of carp per unit effort (Table 3) was greatest at the station below Gowangardie Weir, 44 km upstream, then gradually decreased upstream. Table 2. Approximate densities and biomass of carp in Lake Cooper (1106 ha), Loch Garry (80 ha) and Pogues billabong, Darcys State Forest (18 ha) from population estimates (1980-1982) Waterbody

Population estimate

Biomass density (kg ha-))

Biomass density range (kg ha-')

Lake Cooper Loch Garry Pogues billabong

Table 3. Catch of carp per unit effort and turbidity values at 10 stations along the Broken River from its confluence with the Goulburn River up to Benalla Weir Stations are identified in terms of their distance (km) from the mouth. Total effort was 114 h electroshocking from May 1980 to March 1982 Distance from mouth (km)

Carp catch per unit effort

Turbidity (NTU)

Distance from mouth (krn)

Carp catch per unit effort

Turbidity (NTU)

In Waugh Road billabong, only four carp were captured alive after carp were introduced to the billabong. These did not have anchor tags and may have come in with floods. Sixteen of the carp with anchor tags were found dead in February 1982, immediately after a low percentage saturation of dissolved oxygen had been recorded. The experiment ended with an uncertain final density of carp. In Dead Tree billabong (2 ha), 11 carp were caught (with a total weight of 19 kg) before it dried up in February 1981. When this billabong was evaporating, the biomass density of carp might have increased above an estimated 9.5 kg ha-'. After the floods of June and August 1981, no carp were caught or observed in this billabong. The ranking of sites from highest to lowest, according to the relative abundance of carp was: 1, Pogues billabong; 2, Loch Garry; 3, Lake Cooper; 4-7, Broken River at different distances from the mouth - 4, section 2, 5, section 1, 6, section 3, 7, section 4; 8, Greens Lake; 9, Lake Mokoan; 10, Black billabong; 11, Bruces billabong; 12, Waugh Road billabong, post-floods (experimental introduction); 13, Dead Tree billabong, pre-flood; 14, Broken River, section 5; 15, Waugh Road billabong, pre-flood; 16, Lake Mokoan outlet channel; 17, Dead Tree billabong, post-flood.

Andrea R. Fletcher et al.

Turbidity The mean turbidity of most sites was above 50 NTU, with a large range (Fig. 2). The site with the highest turbidity was Dead Tree billabong, pre-flood, when carp were present (15-20 kg ha-'), and the site with the lowest turbidity was Waugh Road billabong, pre-flood, when the biomass density of carp was very low (less than 1 kg ha-'). The SNK test showed these differences to be significant, identifying nine, possibly ten groups (Fig. 2). These differences in turbidities (Fig. 2) were not significantly correlated with carp densities (rs=O-08, n = 17, P > 0.05). A multiple correlation for data on turbidity, distance from the river mouth and catch of carp per unit effort, for the 10 sites along the Broken River (Table 3), was significant for all comparisons. Catch of carp per unit effort declined with distance upstream (r=0.65, d.f. = 9, P < 0e05), as did turbidity ( r = 0.71, d.f. = 9, P < 0.05). These two variables showed a highly significant positive correlation (rs= + 0.88, d.f. = 9, P < 0.001). Table 4. Values of correlation coefficient r (and degrees of freedom) for correlations between depth and turbidity from monthly mean values at all sites (March 1980-March 1982) Separate correlations before and after the floods in 1981 are given where applicable. n.s., not significant, *P