- Email: [email protected]
Effects of trampling by humans on animals inhabiting coralline algal turf in the rocky intertidal
Journal of Experimental Marine Biology and Ecology, 235 (1999) 45–53
L
Effects of trampling by humans on animals inhabiting coralline algal turf in the rocky intertidal Pam J. Brown a , Richard B. Taylor b , * a
b
Leigh Marine Laboratory, University of Auckland, P.O. Box 349, Warkworth, New Zealand Institute of Marine Sciences, University of North Carolina at Chapel Hill, 3431 Arendell Street, Morehead City, NC 28557, USA
Received 4 September 1996; received in revised form 10 August 1998; accepted 17 August 1998
Abstract This paper investigates the effects of trampling by humans on the fauna associated with articulated coralline algal turf. Patches of intertidal turf in a low-use area of the Cape Rodney to Okakari Point Marine Reserve (in north-eastern New Zealand) were experimentally trampled over 5 days at three levels that fell within those measured in a part of the reserve subject to heavy visitor use. Two days after trampling ended there were | 2 ? 10 5 individual macrofauna ( . 500 mm) per m 2 in control plots, but densities declined with increasing trampling intensity in the treatment plots, and were reduced to 50% of control values at the highest trampling intensity. Densities of five of the eight commonest taxa were negatively correlated with trampling intensity, with polychaetes being particularly susceptible to low levels of trampling. Three months after trampling ended densities of all taxa had returned to control values, with the exception of polychaetes. Reductions in animal densities are tentatively attributed to the loss of turf and associated sand caused by trampling, rather than direct destruction of the organisms. Given the likely importance of these abundant and productive animals in the rocky reef ecosystem, and their vulnerability to low levels of trampling by humans, we conclude that the effective management of marine protected areas may necessitate total exclusion of humans in some cases. 1999 Elsevier Science B.V. All rights reserved. Keywords: Algae; Coralline turf; Epifauna; Human impact; Intertidal; New Zealand; Marine reserve; Trampling
1. Introduction Recent studies have shown that trampling by humans can reduce abundances of rocky *Corresponding author. Tel.: 1 1 252 7266841; fax: 1 1 252 7262426; e-mail: [email protected] 0022-0981 / 99 / $ – see front matter 1999 Elsevier Science B.V. All rights reserved. PII: S0022-0981( 98 )00186-5
46
P. J. Brown, R.B. Taylor / J. Exp. Mar. Biol. Ecol. 235 (1999) 45 – 53
intertidal organisms such as macroalgae, molluscs and barnacles (Beauchamp and Gowing, 1982; Ghazanshahi et al., 1983; Povey and Keough, 1991; Brosnan and Crumrine, 1994; Keough and Quinn, 1998). Most of this research has focused on large conspicuous organisms, but smaller cryptic animals also merit attention due to their great abundance (Hicks, 1986), high productivity (Edgar and Moore, 1986), and importance as food for higher trophic levels (Coull and Wells, 1983; Jones, 1988). In the rocky intertidal highest densities of small animals are typically found on macroalgae (Gibbons and Griffiths, 1986), which provide the epifauna with a range of resources, such as food and a refuge from predation and desiccation (Gibbons, 1991). Abundances of epifauna are therefore likely to be reduced where trampling by humans reduces the biomass of their host plants. Coralline algal turfs are a dominant feature of many subtropical and temperate rocky shores (Johansen, 1981). The turf matrix is typically inhabited by high densities of small mobile and sessile animals (e.g. Chapman, 1955; Dommasnes, 1969; Hicks, 1971). Moderate levels of trampling by humans does not reduce the area occupied by coralline algal turf, but it does reduce the height of the turf (Povey and Keough, 1991), thus altering the habitat available for epifauna. In this study we examine the effects of trampling by humans on the macrofauna inhabiting coralline algal turf within a popular marine reserve in north-eastern New Zealand.
2. Methods This study was conducted in articulated coralline algal turf on the intertidal reef flat at Knot Rock, a low-use part of the Cape Rodney to Okakari Point Marine Reserve, in north-eastern New Zealand (368179S, 1748489E). The reserve receives up to 3000 visitors per day during the summer, and large numbers of these people explore rocky intertidal regions adjacent to the main public entry point [Brown and Creese, in preparation]. During the course of this study few visitors were observed several hundred meters away on the intertidal flats at Knot Rock. The turf there formed apparently homogeneous patches of several square meters in extent and | 5 mm in height. Four 0.09 m 2 quadrats were placed randomly within each of three patches, and the corners marked with small plastic disks nailed into the rock. Within each patch the quadrats were trampled with a total of 0 (control), 10, 50 or 150 footsteps. These trampling levels were designed to fall within the number of footsteps estimated to occur during the 2-month peak visitor season (January–February) at various parts of the Channel Reefs, an area of the marine reserve subject to heavy use by the public. At three sites on the Channel Reefs remote video footage revealed average trampling rates of 45–215 footsteps / m 2 / h, equivalent to 16–77 footsteps / 0.09 m 2 quadrat / 4 h low tide period [Brown and Creese, in preparation]. It is clear therefore, that our experimental trampling intensities were well within the range experienced by some regions of nearby shore, as our maximum total number of footsteps (150) would be reached in only 2 days during the peak visitor season at some parts of the Channel Reefs. Trampling commenced on September 22 1995, with quadrats trampled during low tide at 0, 2, 5, or 30 steps per day for 5 days.
P. J. Brown, R.B. Taylor / J. Exp. Mar. Biol. Ecol. 235 (1999) 45 – 53
47
The trampler wore rubber-soled shoes and weighed | 60 kg. Four turf samples were taken from each quadrat on September 29, 1995, 2 days after trampling ended. Sampling was repeated 3 months (92 days) later on December 28 1995, to determine whether the epifaunal populations had recovered from trampling. Samples were collected by cutting through the coralline algae to the rock surface with an open-ended plastic cylinder (internal diameter 5 42 mm, area 5 13.9 cm 2 ). The coralline turf within was collected using a metal scraper, with care taken to ensure that all coralline turf and underlying sediment was removed down to the basal crust or bare rock. This method collects at least 95% of meiofaunal individuals associated with coralline turf (McCrone, 1987), and appeared to be at least as effective for the larger animals sampled in the present study (we noticed no remaining animals on the rock that had been scraped clean, and the epifauna had no opportunity to escape following placement of the sampling cylinder). Samples were preserved in 5% formalin in seawater. They were subsequently rinsed on a 500-mm mesh sieve, and the animals retained were identified to coarse taxonomic levels and counted under a stereo microscope. An additional experiment was run in August 1996 to assess the effects of trampling on physical attributes of the turf matrix that are likely to be important to epifauna. We measured turf height, biomass, and sand content, all of which could potentially affect the amount of space available for epifauna to inhabit within a given area of rock surface. The experimental site, protocol, and sampling were as described above except that trampling was carried out over 3 days instead of 5 (with the same total number of steps) from August 12–14 1996, only three samples were taken from each quadrat, and post-trampling samples were only collected once (2 days after the cessation of trampling). Samples were preserved in 70% ethanol, then transferred to a beaker for removal of soft-bodied animals and detritus by decanting in freshwater. The leftover material comprised turf fragments and sand, along with a few large gastropods and polychaetes that were removed and discarded. The turf and associated sand was washed on a 1-mm mesh sieve that retained the turf, but allowed the sand to pass through, to be trapped on a 100-mm mesh sieve below (following McCrone, 1987). Sand and turf were dried separately to constant weight at 808C. One-way nested analysis of variance (ANOVA) was used to examine the effects of trampling on densities of total animals, densities of the eight commonest taxa (averaged across all trampling levels and both sampling occasions), and physical characteristics of the turf (factor 5 trampling intensity (fixed), samples nested within quadrats). If Cochran’s test detected significant heterogeneity of variances (a 5 0.05) data were log (X 1 1) transformed prior to analysis. Since we predicted that trampling would have a negative impact on epifaunal densities (see Section 1), it was possible to increase the power of the (non-directional) ANOVAs by incorporating information on the ranks of the means. Following Rice and Gaines (1994a, 1994b), directional P values were calculated from the test statistic r s Pc , the product of (1) Spearman’s rank correlation between the observed ranking of the means and the expected ranking given the alternative hypothesis (i.e., densities: control $ 10 $ 50 $ 150, with at least one inequality), and (2) the complement of the P value from the ANOVA. Data for each sample date were analyzed separately.
48
P. J. Brown, R.B. Taylor / J. Exp. Mar. Biol. Ecol. 235 (1999) 45 – 53
3. Results Total numbers of animals in the control plots ranged from 340662 (mean61 S.E.) individuals per 13.9 cm 2 sample at the start of the experiment to 288624 3 months later when the experiment was terminated. These numbers are equivalent to 2.1–2.5 ? 10 5 individuals / m 2 . The natural epifaunal assemblage (in the control plots at the first sampling) was dominated by small gastropods [39.863.2% of total individuals (mean61 S.E.)], polychaetes (26.863.3), bivalves (15.562.4), gammarid amphipods (7.163.4), nematodes (3.061.5), isopods (2.361.3), anemones (1.960.8), and ostracods (1.560.5). Together these eight taxa comprised 97.9% of total individuals. There was a strong negative effect of trampling on total animal densities 2 days after experimental trampling ceased (P 5 0.004), with densities at the highest trampling intensity declining to 50% of control values (Fig. 1). There was no apparent effect of trampling on total animal densities after 3 months (P 5 0.18, Fig. 1). The effects of trampling on the eight most abundant taxa are shown in Fig. 2. The first post-trampling survey (at 2 days) found that trampling caused statistically significant (a , 0.05) reductions in densities of gastropods, polychaetes and ostracods, with densities at the highest trampling intensity being 54, 44 and 37% of control values for the three taxa, respectively. Polychaetes appeared to be particularly vulnerable to trampling, with a substantial decline in density evident at the lowest trampling level. Bivalves (P 5 0.09) and nematodes (P 5 0.10) also tended to be negatively affected by trampling. In contrast, densities of gammarid amphipods, anemones, and isopods were relatively unaffected by trampling (P $ 0.15), although densities of gammarid amphipods and isopods were lowest at the highest trampling intensity. By the second survey (3 months after trampling) densities of animals in the trampled quadrats had returned to near control values for all taxa except polychaetes, which still showed a strong negative effect of trampling (P 5 0.01). This difference in polychaete densities among treatments was not present in samples collected 1 day prior to experimental trampling (data not shown). Two days after trampling ended (in a separate experiment), turf dry weight and sand
Fig. 1. Densities of total macrofauna 2 days and 3 months after cessation of experimental trampling of intertidal coralline turf.
P. J. Brown, R.B. Taylor / J. Exp. Mar. Biol. Ecol. 235 (1999) 45 – 53
49
Fig. 2. Densities of eight commonest epifaunal taxa 2 days and 3 months after cessation of experimental trampling of intertidal coralline turf. Taxa are ranked in decreasing order of abundance.
dry weight showed strong declines with increasing trampling intensity (P 5 0.005 for both), and turf height showed a similar trend (P 5 0.08) (Fig. 3). The magnitude of the declines in these physical variables (values at the highest trampling intensity were 41–53% of controls) were comparable to the magnitude of the declines in abundance of the taxa most affected by trampling (37–54%).
50
P. J. Brown, R.B. Taylor / J. Exp. Mar. Biol. Ecol. 235 (1999) 45 – 53
Fig. 3. Turf height, turf dry weight, and sand dry weight 2 days after cessation of experimental trampling of intertidal coralline turf. Bars represent means 1 1 S.E.
4. Discussion Animals . 500 mm were abundant within the intertidal coralline turf ( | 2 ? 10 5 individuals / m 2 ), but experimental trampling at conservative levels caused immediate declines in densities of total animals and most common taxa (first measured 2 days after trampling ceased). There are several ways in which trampling may reduce densities of turf-dwelling animals. The most obvious and direct effect would be the crushing impact of the footsteps. However, most of the taxa are highly mobile, and should be able to recolonize trampled patches within hours or days (Sherman and Coull, 1980; Billheimer and Coull, 1988), so any short-term reduction in densities should have been rapidly compensated for by immigration, at least on the spatial scale of our experiments. Moreover, if the direct impact of trampling determined subsequent community composition we would have expected the vulnerability of individual taxa to be related to their morphologies, but
P. J. Brown, R.B. Taylor / J. Exp. Mar. Biol. Ecol. 235 (1999) 45 – 53
51
this was not always the case. For example, soft-bodied anemones did not appear to be affected by trampling, but hard-bodied gastropods did. It is more likely that the effects of trampling were indirect, through changes caused to the turf itself. In our study trampling reduced the height of the turf by up to 50%. Similar effects of physical damage have been found for trampled turf on an Australian shore (Povey and Keough, 1991), and for wave-battered turf on a Norwegian shore (Dommasnes, 1968). Our data show that the height reduction caused by trampling was due to loss of plant tissue, not compression. The consequent loss of colonizable habitat is likely to result in lower epifaunal densities, due mainly to the dependence of the animals on their host plants for food and shelter (see reviews in Edgar and Moore, 1986; Hicks, 1986; Gibbons, 1991). Most seaweed epifaunal populations appear to be limited by their periphytal or detrital food supply (Edgar, 1993), which would be depleted in trampled turf due to reduction of plant surface area from which to graze periphyton (Edgar, 1993), and lower quantities of detritus trapped by the thallus matrix (Hicks, 1986). Coralline turf also traps large amounts of sandy sediments (Hicks, 1986). In the present study, the average dry weight of sand in the control plots was nearly three times that of the turf itself, which probably explains why coralline turf harbours taxa such as bivalves, which are more characteristic of sedimentary than phytal habitats. Densities of such taxa might be expected to decline following the reductions in sand caused by trampling (Fig. 3), but paradoxically the bivalves were in fact less affected by trampling than many other taxa (Fig. 2). In trampled turf there will also be a reduction in shelter afforded from solar radiation, predation (Coull and Wells, 1983), wave action (Dommasnes, 1968), and desiccation (Gibbons, 1991). It is also possible that fewer passively drifting animals would be captured in a turf mat that had its height reduced by trampling (Dean and Connell, 1987). Further manipulative experiments would be required to determine the relative importance of these various mechanisms, which may vary seasonally. If solar radiation and desiccation are important agents of epifaunal decline in trampled turf then the animals would be most vulnerable to trampling during the summer, which is also when visitor numbers are highest. Since biomasses of many other macroalgal species are also reduced by trampling, it is highly likely that animals inhabiting their fronds will also be detrimentally affected by trampling. In our study we did not consider meiofauna (animals passing through a 500-mm mesh, but trapped on a 63-mm mesh), but they are likely to be more abundant than macrofauna by at least an order of magnitude in the turf (Gibbons and Griffiths, 1986), and are also likely to be negatively affected by reduction of plant biomass due to trampling, for the reasons outlined above. Three months after trampling ended it was found that densities of total animals, and all taxa except polychaetes, had returned to near control values. Our experimental trampling was conducted in 0.09 m 2 quadrats surrounded by patches of relatively natural coralline turf. This turf harboured dense populations of animals, which represented a large pool of potential recolonizers. However, in high-use areas where the entire reef is trampled by large numbers of people this pool may be severely reduced, increasing recovery time. Time taken for the epifaunal assemblage to fully recover probably depends ultimately on the recovery rate of the turf itself. Coralline algae grow slowly compared to other seaweeds (Littler and Littler, 1980) and may not fully recover before
52
P. J. Brown, R.B. Taylor / J. Exp. Mar. Biol. Ecol. 235 (1999) 45 – 53
the next visitor season, especially if low levels of trampling by visitors retard its development in the interim. Small mobile invertebrates have high production:biomass ratios (Edgar and Moore, 1986) and are extremely abundant on intertidal macroalgae (Gibbons and Griffiths, 1986). Common taxa such as amphipods and isopods are major prey items for juvenile and adult fish of many species (Choat and Kingett, 1982; Coull and Wells, 1983; Jones, 1988), as well as some decapods (Dean and Connell, 1987), so it is likely that the removal of a substantial proportion of epifauna from the intertidal by trampling will have consequences for higher trophic levels. A function of marine protected areas is to safeguard resident flora and fauna from the activities of humans. Within such reserves the collection of animals and plants is typically prohibited, but there is normally no restriction on public access, suggesting that managers fail to recognise more subtle impacts of humans such as trampling. Given the potentially serious implications of damaging an abundant and productive assemblage of animals such as the turf-dwelling epifauna, along with the larger organisms also potentially affected, the effects of trampling by humans clearly deserve consideration in coastal management plans. This is particularly critical for areas of coastline that are intended to be maintained in ‘‘natural’’ condition for the purposes of conservation or scientific study. To meet these goals it may be necessary to completely restrict humans from some areas, since even relatively low levels of trampling can severely damage many rocky intertidal organisms (Povey and Keough, 1991; Brosnan and Crumrine, 1994 and this study).
Acknowledgements We thank R. Cole, R. Creese and the two referees for their comments on the manuscript. We thank the New Zealand Department of Conservation for funding the project (Grant 1947), and for permitting us to conduct experiments within the Cape Rodney to Okakari Point Marine Reserve.
References Beauchamp, K.A., Gowing, M.M., 1982. A quantitative assessment of human trampling effects on a rocky intertidal community. Mar. Environ. Res. 7, 279–293. Billheimer, L.E., Coull, B.C., 1988. Bioturbation and recolonization of meiobenthos in juvenile spot (Pisces) feeding pits. Estuarine Coastal Shelf Sci. 27, 335–340. Brosnan, D.M., Crumrine, L.L., 1994. Effects of human trampling on marine rocky shore communities. J. Exp. Mar. Biol. Ecol. 177, 79–97. Chapman, G., 1955. Aspects of the fauna and flora of the Azores. VI. The density of animal life in the coralline alga zone. Ann. Mag. Nat. Hist. 8 (Ser. 12), 801–805. Choat, J.H., Kingett, P.D., 1982. The influence of fish predation on the abundance cycles of an algal turf invertebrate fauna. Oecologia 54, 88–95. Coull, B.C., Wells, J.B.J., 1983. Refuges from fish predation: experiments with phytal meiofauna from the New Zealand rocky intertidal. Ecology 64, 1599–1609.
P. J. Brown, R.B. Taylor / J. Exp. Mar. Biol. Ecol. 235 (1999) 45 – 53
53
Dean, R.L., Connell, J.H., 1987. Marine invertebrates in an algal succession. III. Mechanisms linking habitat complexity with diversity. J. Exp. Mar. Biol. Ecol. 109, 249–273. Dommasnes, A., 1968. Variations in the meiofauna of Corallina officinalis L. with wave exposure. Sarsia 34, 117–124. Dommasnes, A., 1969. On the fauna of Corallina officinalis L. in western Norway. Sarsia 38, 71–86. Edgar, G.J., 1993. Measurement of the carrying capacity of benthic habitats using a metabolic-rate based index. Oecologia 95, 115–121. Edgar, G.J., Moore, P.G., 1986. Macro-algae as habitats for motile macrofauna. Monograp. Biol. 4, 255–277. Ghazanshahi, J., Huchel, T.D., Devinny, J.S., 1983. Alteration of southern California rocky shore ecosystems by public recreational use. J. Environ. Man. 16, 379–394. Gibbons, M.J., 1991. Rocky shore meiofauna: a brief overview. Trans. Royal Soc. South Africa 47, 595–603. Gibbons, M.J., Griffiths, C.L., 1986. A comparison of macrofaunal and meiofaunal distribution and standing stock across a rocky shore, with an estimate of their productivities. Mar. Biol. 93, 181–188. Hicks, G.R.F., 1971. Checklist and ecological notes on the fauna associated with some littoral corallinacean algae. Bull. Natural Sci. 2, 47–58. Hicks, G.R.F., 1986. Meiofauna associated with rocky shore algae. In: Moore, P.G., Seed, R. (Eds.), The Ecology of Rocky Coasts. Hodder and Stoughton, London, pp. 36–56. Johansen, H.W., 1981. Coralline Algae, A First Synthesis. CRC, Boca Raton, FL. Jones, G.P., 1988. Ecology of rocky reef fish of north-eastern New Zealand: a review. NZ J. Mar. Freshwater Res. 22, 445–462. Keough, M.J., Quinn, G.P., 1998. Effects of periodic disturbances from trampling on rocky intertidal algal beds. Ecol. Appl. 8, 141–161. Littler, M.M., Littler, D.S., 1980. The evolution of thallus form and survival strategies in benthic marine macroalgae: field and laboratory tests of a functional form model. Am. Natural. 116, 25–44. McCrone, A., 1987. Spatial and temporal variation of meiofauna associated with Corallina officinalis Linnaeus. Unpublished M.Sc. thesis, University of Auckland, Auckland, New Zealand. Povey, A., Keough, M.J., 1991. Effects of trampling on plant and animal populations on rocky shores. Oikos 61, 355–368. Rice, W.R., Gaines, S.D., 1994a. Extending nondirectional heterogeneity tests to evaluate simply ordered alternative hypotheses. Proc. Nat. Acad. Sci. USA 91, 225–226. Rice, W.R., Gaines, S.D., 1994b. The ordered-heterogeneity family of tests. Biometrics 50, 746–752. Sherman, K.M., Coull, B.C., 1980. The response of meiofauna to sediment disturbance. J. Exp. Mar. Biol. Ecol. 46, 59–71.
L
Effects of trampling by humans on animals inhabiting coralline algal turf in the rocky intertidal Pam J. Brown a , Richard B. Taylor b , * a
b
Leigh Marine Laboratory, University of Auckland, P.O. Box 349, Warkworth, New Zealand Institute of Marine Sciences, University of North Carolina at Chapel Hill, 3431 Arendell Street, Morehead City, NC 28557, USA
Received 4 September 1996; received in revised form 10 August 1998; accepted 17 August 1998
Abstract This paper investigates the effects of trampling by humans on the fauna associated with articulated coralline algal turf. Patches of intertidal turf in a low-use area of the Cape Rodney to Okakari Point Marine Reserve (in north-eastern New Zealand) were experimentally trampled over 5 days at three levels that fell within those measured in a part of the reserve subject to heavy visitor use. Two days after trampling ended there were | 2 ? 10 5 individual macrofauna ( . 500 mm) per m 2 in control plots, but densities declined with increasing trampling intensity in the treatment plots, and were reduced to 50% of control values at the highest trampling intensity. Densities of five of the eight commonest taxa were negatively correlated with trampling intensity, with polychaetes being particularly susceptible to low levels of trampling. Three months after trampling ended densities of all taxa had returned to control values, with the exception of polychaetes. Reductions in animal densities are tentatively attributed to the loss of turf and associated sand caused by trampling, rather than direct destruction of the organisms. Given the likely importance of these abundant and productive animals in the rocky reef ecosystem, and their vulnerability to low levels of trampling by humans, we conclude that the effective management of marine protected areas may necessitate total exclusion of humans in some cases. 1999 Elsevier Science B.V. All rights reserved. Keywords: Algae; Coralline turf; Epifauna; Human impact; Intertidal; New Zealand; Marine reserve; Trampling
1. Introduction Recent studies have shown that trampling by humans can reduce abundances of rocky *Corresponding author. Tel.: 1 1 252 7266841; fax: 1 1 252 7262426; e-mail: [email protected] 0022-0981 / 99 / $ – see front matter 1999 Elsevier Science B.V. All rights reserved. PII: S0022-0981( 98 )00186-5
46
P. J. Brown, R.B. Taylor / J. Exp. Mar. Biol. Ecol. 235 (1999) 45 – 53
intertidal organisms such as macroalgae, molluscs and barnacles (Beauchamp and Gowing, 1982; Ghazanshahi et al., 1983; Povey and Keough, 1991; Brosnan and Crumrine, 1994; Keough and Quinn, 1998). Most of this research has focused on large conspicuous organisms, but smaller cryptic animals also merit attention due to their great abundance (Hicks, 1986), high productivity (Edgar and Moore, 1986), and importance as food for higher trophic levels (Coull and Wells, 1983; Jones, 1988). In the rocky intertidal highest densities of small animals are typically found on macroalgae (Gibbons and Griffiths, 1986), which provide the epifauna with a range of resources, such as food and a refuge from predation and desiccation (Gibbons, 1991). Abundances of epifauna are therefore likely to be reduced where trampling by humans reduces the biomass of their host plants. Coralline algal turfs are a dominant feature of many subtropical and temperate rocky shores (Johansen, 1981). The turf matrix is typically inhabited by high densities of small mobile and sessile animals (e.g. Chapman, 1955; Dommasnes, 1969; Hicks, 1971). Moderate levels of trampling by humans does not reduce the area occupied by coralline algal turf, but it does reduce the height of the turf (Povey and Keough, 1991), thus altering the habitat available for epifauna. In this study we examine the effects of trampling by humans on the macrofauna inhabiting coralline algal turf within a popular marine reserve in north-eastern New Zealand.
2. Methods This study was conducted in articulated coralline algal turf on the intertidal reef flat at Knot Rock, a low-use part of the Cape Rodney to Okakari Point Marine Reserve, in north-eastern New Zealand (368179S, 1748489E). The reserve receives up to 3000 visitors per day during the summer, and large numbers of these people explore rocky intertidal regions adjacent to the main public entry point [Brown and Creese, in preparation]. During the course of this study few visitors were observed several hundred meters away on the intertidal flats at Knot Rock. The turf there formed apparently homogeneous patches of several square meters in extent and | 5 mm in height. Four 0.09 m 2 quadrats were placed randomly within each of three patches, and the corners marked with small plastic disks nailed into the rock. Within each patch the quadrats were trampled with a total of 0 (control), 10, 50 or 150 footsteps. These trampling levels were designed to fall within the number of footsteps estimated to occur during the 2-month peak visitor season (January–February) at various parts of the Channel Reefs, an area of the marine reserve subject to heavy use by the public. At three sites on the Channel Reefs remote video footage revealed average trampling rates of 45–215 footsteps / m 2 / h, equivalent to 16–77 footsteps / 0.09 m 2 quadrat / 4 h low tide period [Brown and Creese, in preparation]. It is clear therefore, that our experimental trampling intensities were well within the range experienced by some regions of nearby shore, as our maximum total number of footsteps (150) would be reached in only 2 days during the peak visitor season at some parts of the Channel Reefs. Trampling commenced on September 22 1995, with quadrats trampled during low tide at 0, 2, 5, or 30 steps per day for 5 days.
P. J. Brown, R.B. Taylor / J. Exp. Mar. Biol. Ecol. 235 (1999) 45 – 53
47
The trampler wore rubber-soled shoes and weighed | 60 kg. Four turf samples were taken from each quadrat on September 29, 1995, 2 days after trampling ended. Sampling was repeated 3 months (92 days) later on December 28 1995, to determine whether the epifaunal populations had recovered from trampling. Samples were collected by cutting through the coralline algae to the rock surface with an open-ended plastic cylinder (internal diameter 5 42 mm, area 5 13.9 cm 2 ). The coralline turf within was collected using a metal scraper, with care taken to ensure that all coralline turf and underlying sediment was removed down to the basal crust or bare rock. This method collects at least 95% of meiofaunal individuals associated with coralline turf (McCrone, 1987), and appeared to be at least as effective for the larger animals sampled in the present study (we noticed no remaining animals on the rock that had been scraped clean, and the epifauna had no opportunity to escape following placement of the sampling cylinder). Samples were preserved in 5% formalin in seawater. They were subsequently rinsed on a 500-mm mesh sieve, and the animals retained were identified to coarse taxonomic levels and counted under a stereo microscope. An additional experiment was run in August 1996 to assess the effects of trampling on physical attributes of the turf matrix that are likely to be important to epifauna. We measured turf height, biomass, and sand content, all of which could potentially affect the amount of space available for epifauna to inhabit within a given area of rock surface. The experimental site, protocol, and sampling were as described above except that trampling was carried out over 3 days instead of 5 (with the same total number of steps) from August 12–14 1996, only three samples were taken from each quadrat, and post-trampling samples were only collected once (2 days after the cessation of trampling). Samples were preserved in 70% ethanol, then transferred to a beaker for removal of soft-bodied animals and detritus by decanting in freshwater. The leftover material comprised turf fragments and sand, along with a few large gastropods and polychaetes that were removed and discarded. The turf and associated sand was washed on a 1-mm mesh sieve that retained the turf, but allowed the sand to pass through, to be trapped on a 100-mm mesh sieve below (following McCrone, 1987). Sand and turf were dried separately to constant weight at 808C. One-way nested analysis of variance (ANOVA) was used to examine the effects of trampling on densities of total animals, densities of the eight commonest taxa (averaged across all trampling levels and both sampling occasions), and physical characteristics of the turf (factor 5 trampling intensity (fixed), samples nested within quadrats). If Cochran’s test detected significant heterogeneity of variances (a 5 0.05) data were log (X 1 1) transformed prior to analysis. Since we predicted that trampling would have a negative impact on epifaunal densities (see Section 1), it was possible to increase the power of the (non-directional) ANOVAs by incorporating information on the ranks of the means. Following Rice and Gaines (1994a, 1994b), directional P values were calculated from the test statistic r s Pc , the product of (1) Spearman’s rank correlation between the observed ranking of the means and the expected ranking given the alternative hypothesis (i.e., densities: control $ 10 $ 50 $ 150, with at least one inequality), and (2) the complement of the P value from the ANOVA. Data for each sample date were analyzed separately.
48
P. J. Brown, R.B. Taylor / J. Exp. Mar. Biol. Ecol. 235 (1999) 45 – 53
3. Results Total numbers of animals in the control plots ranged from 340662 (mean61 S.E.) individuals per 13.9 cm 2 sample at the start of the experiment to 288624 3 months later when the experiment was terminated. These numbers are equivalent to 2.1–2.5 ? 10 5 individuals / m 2 . The natural epifaunal assemblage (in the control plots at the first sampling) was dominated by small gastropods [39.863.2% of total individuals (mean61 S.E.)], polychaetes (26.863.3), bivalves (15.562.4), gammarid amphipods (7.163.4), nematodes (3.061.5), isopods (2.361.3), anemones (1.960.8), and ostracods (1.560.5). Together these eight taxa comprised 97.9% of total individuals. There was a strong negative effect of trampling on total animal densities 2 days after experimental trampling ceased (P 5 0.004), with densities at the highest trampling intensity declining to 50% of control values (Fig. 1). There was no apparent effect of trampling on total animal densities after 3 months (P 5 0.18, Fig. 1). The effects of trampling on the eight most abundant taxa are shown in Fig. 2. The first post-trampling survey (at 2 days) found that trampling caused statistically significant (a , 0.05) reductions in densities of gastropods, polychaetes and ostracods, with densities at the highest trampling intensity being 54, 44 and 37% of control values for the three taxa, respectively. Polychaetes appeared to be particularly vulnerable to trampling, with a substantial decline in density evident at the lowest trampling level. Bivalves (P 5 0.09) and nematodes (P 5 0.10) also tended to be negatively affected by trampling. In contrast, densities of gammarid amphipods, anemones, and isopods were relatively unaffected by trampling (P $ 0.15), although densities of gammarid amphipods and isopods were lowest at the highest trampling intensity. By the second survey (3 months after trampling) densities of animals in the trampled quadrats had returned to near control values for all taxa except polychaetes, which still showed a strong negative effect of trampling (P 5 0.01). This difference in polychaete densities among treatments was not present in samples collected 1 day prior to experimental trampling (data not shown). Two days after trampling ended (in a separate experiment), turf dry weight and sand
Fig. 1. Densities of total macrofauna 2 days and 3 months after cessation of experimental trampling of intertidal coralline turf.
P. J. Brown, R.B. Taylor / J. Exp. Mar. Biol. Ecol. 235 (1999) 45 – 53
49
Fig. 2. Densities of eight commonest epifaunal taxa 2 days and 3 months after cessation of experimental trampling of intertidal coralline turf. Taxa are ranked in decreasing order of abundance.
dry weight showed strong declines with increasing trampling intensity (P 5 0.005 for both), and turf height showed a similar trend (P 5 0.08) (Fig. 3). The magnitude of the declines in these physical variables (values at the highest trampling intensity were 41–53% of controls) were comparable to the magnitude of the declines in abundance of the taxa most affected by trampling (37–54%).
50
P. J. Brown, R.B. Taylor / J. Exp. Mar. Biol. Ecol. 235 (1999) 45 – 53
Fig. 3. Turf height, turf dry weight, and sand dry weight 2 days after cessation of experimental trampling of intertidal coralline turf. Bars represent means 1 1 S.E.
4. Discussion Animals . 500 mm were abundant within the intertidal coralline turf ( | 2 ? 10 5 individuals / m 2 ), but experimental trampling at conservative levels caused immediate declines in densities of total animals and most common taxa (first measured 2 days after trampling ceased). There are several ways in which trampling may reduce densities of turf-dwelling animals. The most obvious and direct effect would be the crushing impact of the footsteps. However, most of the taxa are highly mobile, and should be able to recolonize trampled patches within hours or days (Sherman and Coull, 1980; Billheimer and Coull, 1988), so any short-term reduction in densities should have been rapidly compensated for by immigration, at least on the spatial scale of our experiments. Moreover, if the direct impact of trampling determined subsequent community composition we would have expected the vulnerability of individual taxa to be related to their morphologies, but
P. J. Brown, R.B. Taylor / J. Exp. Mar. Biol. Ecol. 235 (1999) 45 – 53
51
this was not always the case. For example, soft-bodied anemones did not appear to be affected by trampling, but hard-bodied gastropods did. It is more likely that the effects of trampling were indirect, through changes caused to the turf itself. In our study trampling reduced the height of the turf by up to 50%. Similar effects of physical damage have been found for trampled turf on an Australian shore (Povey and Keough, 1991), and for wave-battered turf on a Norwegian shore (Dommasnes, 1968). Our data show that the height reduction caused by trampling was due to loss of plant tissue, not compression. The consequent loss of colonizable habitat is likely to result in lower epifaunal densities, due mainly to the dependence of the animals on their host plants for food and shelter (see reviews in Edgar and Moore, 1986; Hicks, 1986; Gibbons, 1991). Most seaweed epifaunal populations appear to be limited by their periphytal or detrital food supply (Edgar, 1993), which would be depleted in trampled turf due to reduction of plant surface area from which to graze periphyton (Edgar, 1993), and lower quantities of detritus trapped by the thallus matrix (Hicks, 1986). Coralline turf also traps large amounts of sandy sediments (Hicks, 1986). In the present study, the average dry weight of sand in the control plots was nearly three times that of the turf itself, which probably explains why coralline turf harbours taxa such as bivalves, which are more characteristic of sedimentary than phytal habitats. Densities of such taxa might be expected to decline following the reductions in sand caused by trampling (Fig. 3), but paradoxically the bivalves were in fact less affected by trampling than many other taxa (Fig. 2). In trampled turf there will also be a reduction in shelter afforded from solar radiation, predation (Coull and Wells, 1983), wave action (Dommasnes, 1968), and desiccation (Gibbons, 1991). It is also possible that fewer passively drifting animals would be captured in a turf mat that had its height reduced by trampling (Dean and Connell, 1987). Further manipulative experiments would be required to determine the relative importance of these various mechanisms, which may vary seasonally. If solar radiation and desiccation are important agents of epifaunal decline in trampled turf then the animals would be most vulnerable to trampling during the summer, which is also when visitor numbers are highest. Since biomasses of many other macroalgal species are also reduced by trampling, it is highly likely that animals inhabiting their fronds will also be detrimentally affected by trampling. In our study we did not consider meiofauna (animals passing through a 500-mm mesh, but trapped on a 63-mm mesh), but they are likely to be more abundant than macrofauna by at least an order of magnitude in the turf (Gibbons and Griffiths, 1986), and are also likely to be negatively affected by reduction of plant biomass due to trampling, for the reasons outlined above. Three months after trampling ended it was found that densities of total animals, and all taxa except polychaetes, had returned to near control values. Our experimental trampling was conducted in 0.09 m 2 quadrats surrounded by patches of relatively natural coralline turf. This turf harboured dense populations of animals, which represented a large pool of potential recolonizers. However, in high-use areas where the entire reef is trampled by large numbers of people this pool may be severely reduced, increasing recovery time. Time taken for the epifaunal assemblage to fully recover probably depends ultimately on the recovery rate of the turf itself. Coralline algae grow slowly compared to other seaweeds (Littler and Littler, 1980) and may not fully recover before
52
P. J. Brown, R.B. Taylor / J. Exp. Mar. Biol. Ecol. 235 (1999) 45 – 53
the next visitor season, especially if low levels of trampling by visitors retard its development in the interim. Small mobile invertebrates have high production:biomass ratios (Edgar and Moore, 1986) and are extremely abundant on intertidal macroalgae (Gibbons and Griffiths, 1986). Common taxa such as amphipods and isopods are major prey items for juvenile and adult fish of many species (Choat and Kingett, 1982; Coull and Wells, 1983; Jones, 1988), as well as some decapods (Dean and Connell, 1987), so it is likely that the removal of a substantial proportion of epifauna from the intertidal by trampling will have consequences for higher trophic levels. A function of marine protected areas is to safeguard resident flora and fauna from the activities of humans. Within such reserves the collection of animals and plants is typically prohibited, but there is normally no restriction on public access, suggesting that managers fail to recognise more subtle impacts of humans such as trampling. Given the potentially serious implications of damaging an abundant and productive assemblage of animals such as the turf-dwelling epifauna, along with the larger organisms also potentially affected, the effects of trampling by humans clearly deserve consideration in coastal management plans. This is particularly critical for areas of coastline that are intended to be maintained in ‘‘natural’’ condition for the purposes of conservation or scientific study. To meet these goals it may be necessary to completely restrict humans from some areas, since even relatively low levels of trampling can severely damage many rocky intertidal organisms (Povey and Keough, 1991; Brosnan and Crumrine, 1994 and this study).
Acknowledgements We thank R. Cole, R. Creese and the two referees for their comments on the manuscript. We thank the New Zealand Department of Conservation for funding the project (Grant 1947), and for permitting us to conduct experiments within the Cape Rodney to Okakari Point Marine Reserve.
References Beauchamp, K.A., Gowing, M.M., 1982. A quantitative assessment of human trampling effects on a rocky intertidal community. Mar. Environ. Res. 7, 279–293. Billheimer, L.E., Coull, B.C., 1988. Bioturbation and recolonization of meiobenthos in juvenile spot (Pisces) feeding pits. Estuarine Coastal Shelf Sci. 27, 335–340. Brosnan, D.M., Crumrine, L.L., 1994. Effects of human trampling on marine rocky shore communities. J. Exp. Mar. Biol. Ecol. 177, 79–97. Chapman, G., 1955. Aspects of the fauna and flora of the Azores. VI. The density of animal life in the coralline alga zone. Ann. Mag. Nat. Hist. 8 (Ser. 12), 801–805. Choat, J.H., Kingett, P.D., 1982. The influence of fish predation on the abundance cycles of an algal turf invertebrate fauna. Oecologia 54, 88–95. Coull, B.C., Wells, J.B.J., 1983. Refuges from fish predation: experiments with phytal meiofauna from the New Zealand rocky intertidal. Ecology 64, 1599–1609.
P. J. Brown, R.B. Taylor / J. Exp. Mar. Biol. Ecol. 235 (1999) 45 – 53
53
Dean, R.L., Connell, J.H., 1987. Marine invertebrates in an algal succession. III. Mechanisms linking habitat complexity with diversity. J. Exp. Mar. Biol. Ecol. 109, 249–273. Dommasnes, A., 1968. Variations in the meiofauna of Corallina officinalis L. with wave exposure. Sarsia 34, 117–124. Dommasnes, A., 1969. On the fauna of Corallina officinalis L. in western Norway. Sarsia 38, 71–86. Edgar, G.J., 1993. Measurement of the carrying capacity of benthic habitats using a metabolic-rate based index. Oecologia 95, 115–121. Edgar, G.J., Moore, P.G., 1986. Macro-algae as habitats for motile macrofauna. Monograp. Biol. 4, 255–277. Ghazanshahi, J., Huchel, T.D., Devinny, J.S., 1983. Alteration of southern California rocky shore ecosystems by public recreational use. J. Environ. Man. 16, 379–394. Gibbons, M.J., 1991. Rocky shore meiofauna: a brief overview. Trans. Royal Soc. South Africa 47, 595–603. Gibbons, M.J., Griffiths, C.L., 1986. A comparison of macrofaunal and meiofaunal distribution and standing stock across a rocky shore, with an estimate of their productivities. Mar. Biol. 93, 181–188. Hicks, G.R.F., 1971. Checklist and ecological notes on the fauna associated with some littoral corallinacean algae. Bull. Natural Sci. 2, 47–58. Hicks, G.R.F., 1986. Meiofauna associated with rocky shore algae. In: Moore, P.G., Seed, R. (Eds.), The Ecology of Rocky Coasts. Hodder and Stoughton, London, pp. 36–56. Johansen, H.W., 1981. Coralline Algae, A First Synthesis. CRC, Boca Raton, FL. Jones, G.P., 1988. Ecology of rocky reef fish of north-eastern New Zealand: a review. NZ J. Mar. Freshwater Res. 22, 445–462. Keough, M.J., Quinn, G.P., 1998. Effects of periodic disturbances from trampling on rocky intertidal algal beds. Ecol. Appl. 8, 141–161. Littler, M.M., Littler, D.S., 1980. The evolution of thallus form and survival strategies in benthic marine macroalgae: field and laboratory tests of a functional form model. Am. Natural. 116, 25–44. McCrone, A., 1987. Spatial and temporal variation of meiofauna associated with Corallina officinalis Linnaeus. Unpublished M.Sc. thesis, University of Auckland, Auckland, New Zealand. Povey, A., Keough, M.J., 1991. Effects of trampling on plant and animal populations on rocky shores. Oikos 61, 355–368. Rice, W.R., Gaines, S.D., 1994a. Extending nondirectional heterogeneity tests to evaluate simply ordered alternative hypotheses. Proc. Nat. Acad. Sci. USA 91, 225–226. Rice, W.R., Gaines, S.D., 1994b. The ordered-heterogeneity family of tests. Biometrics 50, 746–752. Sherman, K.M., Coull, B.C., 1980. The response of meiofauna to sediment disturbance. J. Exp. Mar. Biol. Ecol. 46, 59–71.