Marine Biology

, Volume 147, Issue 2, pp 509–515

Smashing tests? Patterns and mechanisms of adult mortality in a declining echinoid population

Authors

    • Department of Zoology, Ecology and Plant ScienceUniversity College
    • Marine Invasions Research LaboratorySmithsonian Environmental Research Center
  • David K. A. Barnes
    • Biological Sciences DivisionThe British Antarctic Survey
  • Anne C. Crook
    • Centre for the Development of Teaching and LearningThe University of Reading
Research Article

DOI: 10.1007/s00227-005-1563-3

Cite this article as:
Verling, E., Barnes, D.K.A. & Crook, A.C. Marine Biology (2005) 147: 509. doi:10.1007/s00227-005-1563-3
  • 49 Views

Abstract

Mass mortality of echinoids is well documented, and has potentially profound effects on benthic communities. However, no study to date has quantitatively investigated how regular, predictable events such as winter storms might lead to large mortality events in pivotal echinoid species. Hydrodynamic disturbances can be major drivers of crucial biological processes in benthic communities. For echinoid populations in particular, wave action in shallow waters generated by high winds (winter storms) can cause displacement, damage and even death to individuals. However, evidence for displacement-mediated mortality is scant in the literature, in part because it is so difficult to demonstrate in exposed environments where echinoids are frequently found. In this study, we examined mortality in a sheltered subtidal population of the European purple sea urchin Paracentrotus lividus over a 3-year period, and examined the role that dislodgement by wave action or predation might play in these mortality patterns. Because our study population has been in decline for the past three decades, we considered it important to evaluate its current status in addition to assessing the contribution that adult mortality makes to that decline. We sampled twice per month, using the density of freshly dead echinoid material to assess the extent of adult mortality. The Irish Meteorological Service provided our estimates of wind speed data. We compared historical and recent data on P. lividus size frequency data to investigate change the population structure (Poor recruitment would be caused by failure to spawn over a prolonged period). Our data suggest ongoing declines in this population, and support the theory that the decline of the P. lividus population of Lough Hyne is a result of persistent recruitment failure linked to repeated cool maximum sea surface temperatures. Although we found peaks of P. lividus mortality were coincident with spikes in wind speed, mortality was low, and seems unlikely to have contributed significantly to the dramatic decline in P. lividus in the past three decades.

Introduction

Fluctuating densities of echinoids act as powerful structuring agents in marine ecosystems, often causing major shifts within benthic communities (e.g. Lawrence 1975). It is not surprising therefore, that in the last 25 years, the phenomenon of echinoid mass mortality, which globally has been both widespread and dramatic, has received considerable attention from the scientific community (see Lessios 1988; Elner and Vadas 1990). In turn, this has led to an increased interest in echinoid mortality as an important community-shaping process (Hughes 1994). Nevertheless, the extent of echinoid mortality outside of that directly related to mass-mortality events has been little examined, though this may have important consequences both at population and community levels.

Population collapse has been reported for echinoid populations in many parts of the world (e.g. Lessios 1988). In most cases this has been attributed to over-exploitation for commercial purposes (Southward and Southward 1975; Andrew et al. 2002), as well as to epizootics (Scheibling and Stephenson 1984). In some circumstances however, massive declines in echinoid population densities have occurred, which cannot be explained by these conventional theories. Paracentrotus lividus population change at Lough Hyne marine nature reserve in southwest Ireland is an example of such a decline. Regular population censi have revealed that for some years (since the mid-1980s) the Lough Hyne P. lividus population has been experiencing a prolonged decline (Barnes et al. 1998, 2001, 2002), though neither over fishing (the lough has been a marine reserve since 1981) nor epizootics can be blamed. This reduction has been attributed to, and is likely caused by recruitment failure (Barnes et al. 2001, 2002). However, no data exist to evaluate the extent to which the removal of adult echinoids (by mechanisms other than fishing) might add to this decline. This is important, considering the remarkable reduction in the numbers of adult echinoids observed during recent censi, and may be essential information in evaluating how close this population is to ‘extinction’. Evidence has already been presented to suggest that the population has become extinct in the south basin of the lough (Barnes et al. 2002).

Echinoids face many challenges to their survival, including red tides (Cross and Southgate 1980), epizootics (Scheibling and Stephenson 1984; Scheibling and Hennigar 1997), limited food availability (Harrold and Reed 1985) and predation (Duggins 1980; Cowen 1983; Scheibling and Hamm 1991). A major challenge to adult P. lividus survival in Lough Hyne has historically been perceived to be predation by the asteroid Marthasterias glacialis. Nevertheless, a potentially important factor in echinoid survival, namely the hydrodynamic forces generated by weather, has received relatively little scientific attention. The risk of dislodgement from the substratum by current or wave action is significant because free-floating echinoids frequently become overturned and are rolled about, making them vulnerable to physical damage (Denny 1996). Moreover, P. lividus is an appropriate study species, because it is known to be more susceptible to wave-induced or storm-induced disturbance than other shallow water echinoid species, such as Arbacia lixula (Regis 1978). This is possibly because P. lividus does not usually tend to form such dense aggregations (Bulleri et al. 1999).

In the high-energy environments in which P. lividus populations are frequently found, namely exposed intertidal zones, direct evidence of recent mortality (in the form of freshly dead individuals, or test fragments still containing soft tissue), is rapidly transported out to sea or is destroyed, making the relationship between echinoid mortality and weather difficult to quantify. At Lough Hyne, however, P. lividus exists under comparatively calm hydrodynamic conditions. Lough Hyne is a sheltered marine lough (see Kitching 1987 for a full description) connected to the Atlantic by way of a narrow channel known as the ‘rapids’. P. lividus currently occurs in its highest densities at the north shore of the lough, where it is found on shallow subtidal (0–2 m) boulder scree (Barnes et al. 1998). Importantly, the water residence time in this section of the lough is in the region of 26–27 tidal cycles (Bassindale et al. 1957), and the tidal flow within the north basin is too low to be measurable (Bell and Barnes 2002). These unusually sheltered conditions mean that test fragments are more likely to remain within close range of the source population, and provide the unique opportunity to examine seasonal echinoid mortality. Despite these prevailing sheltered conditions, Lough Hyne is still vulnerable to significant wave action during periods of high wind and severe storm activity.

In this study, we (1) further examined the size structure of the P. lividus population to assess its current status, (2) systematically monitored spatial and temporal variation in the density of freshly dead P. lividus over a 3-year period, (3) examined the potential contribution of some of the most important mechanisms that might drive such mortality. We hypothesised that mortality would increase during times of high wind speed resulting from winter storms and other predictable or unpredictable seasonal disturbances.

Materials and methods

Current status of Paracentrotus lividus population at Lough Hyne

The size structure of the P. lividus population at Lough Hyne was examined in 2000 and 2001 to assess the current status of this population. In Lough Hyne, the north shore P. lividus population is found in the immediate subtidal in boulder scree habitat and has been the subject of a long-running census with links back to 1920s (see Barnes et al. 2002). Following the protocol of previous surveys, we collected data to examine variability in size structure, and compared these to data collected in a similar manner in 1977, 1995 and 1998 (see Barnes et al. 1998, 2002). The test diameters of P. lividus were measured in June 2000 and June 2001 by means of a snorkeling survey at each of the three study sites. Quadrats measuring 1 m2 were randomly placed on the substratum at each site, and Vernier callipers were used to record the maximum test diameter (mm) of individuals within each quadrat until a total of 600 individuals were measured (200 at each site). Care was taken to turn over all boulders within quadrats, so that the smallest individuals (<10 mm) could be located.

Assessment of adult Paracentrotus lividus mortality at Lough Hyne

Data were collected (also by means of a snorkelling survey) on the north shore of Lough Hyne Marine Nature Reserve, Ireland (51°29′N 09°18′W) over a 37-month study period, between December 1998 and December 2001. To examine spatial variability in mortality, we selected three approximately equal-sized (100×2-m) sites. Observations were made twice per month during the study period at each of the three study sites. During each sampling session, nine 1-m2 quadrats were placed haphazardly on the substratum and the number of (1) living and (2) freshly dead (entire or partial test fragments with some soft tissue still visible on the test) P. lividus were counted. ‘Proportional mortality’ was calculated using the following equation:
$${\text{no}}{\text{. freshly dead }}P. lividus{\text{ per m}}^{{\text{2}}} {\text{/total no}}{\text{. }}P. lividus{\text{ per m}}^{{\text{2}}} $$

Estimation of wind speed at Lough Hyne

Wind speed (knots) data covering the entire study time-period were obtained from the Irish Meteorological Service. These data were recorded at Valentia Island Weather Station, Kerry Ireland (51°53′N 10°22′W), the closest weather station to the study site.

Predation by Marthasterias glacialis

The diet of the sea star M. glacialis at the north shore of Lough Hyne was monitored twice per month over a 24-month period from January 2000 to December 2002. During each sampling period, 30 m2 of substrate was sampled at each of four depth categories: 0–1 m (using 30 1-m2 quadrats), 1–2 m, 2–4 m and 4–6 m (using three 1×10-m transect lines) at each of the three sites 1, 2 and 3. For those individual M. glacialis observed to be feeding, the prey item(s) was observed and recorded.

Results

The status of Paracentrotus lividus population at Lough Hyne

The density of P. lividus on the north shore of Lough Hyne was 4.96±0.96 m−2 in 2000 and 4.35±0.99 m−2 in 2001 (means and standard errors). During the course of the study, there was no indication that the P. lividus population decline at Lough Hyne has slowed in the past number of years. Firstly, no mass spawning events were observed during the study period. Second, the mean maximum test diameter of P. lividus was larger in 2001 compared with 2000, which in turn, was greater than the mean diameters recorded in 1998, 1995 or 1977 (Fig. 1). Third, the proportion of 0+ age and 1+ age P. lividus within the population was negligible during both years (Table 1). Finally, the proportion of adult P. lividus making up the population has increased proportionally since 1977, and these data show further increases in 2000 and 2001, though we know that the actual number of individuals belonging to larger size classes has been decreasing dramatically during this time (Barnes et al. 2002).
Fig. 1

Paracentrotus lividus. Size structure at the north shore of Lough Hyne for 1977, 1995, 1998 (from Barnes et al. 2002), 2000 and 2001 (data collected in this study)

Table 1

Age structure of the Paracentrotus lividus population at the north shore of Lough Hyne in 1995, 1998 (from Barnes et al. 1998), 2000 and 2001. Table shows proportion of the population in six size categories. Age approximations are taken from Crapp and Willis (1975)

Year group

Size group

1995

1998

2000

2001

(mm)

(n=200)

(n=406)

(n=600)

(n=600)

0+

1–10

2.0%

2.2%

0.2%

0%

1+

11–20

3.5%

5.4%

0.6%

0.8%

2+

21–35

14.5%

20.4%

8.7%

4.4%

3+

36–50

39.0%

45.9%

64.1%

50.4%

4+

51–60

35.5%

21.7%

24.4%

39.5%

5–9

61–65

5.5%

2.9%

2.0%

4.8%

Paracentrotus lividus mortality at Lough Hyne

In total, 3,010 observations were made of P. lividus during the 3-year study period. There was no significant difference in the proportion of adult mortality between the three study sites (Kruskal-Wallis, H=0.06, P=0.970), and the pattern of mortality at all three sites over the study period was similar. Therefore, data from the three sites were pooled. The mortality of adult P. lividus detected at the north shore was typically very low year round, reaching a maximum of only 4.78% (January 1999; Fig. 2).
Fig. 2 a

Mean and average wind speed in knots, recorded at Valencia Island Weather Station, Kerry (data courtesy of The Irish Meteorological Service). b Percent mortality (bars represent standard error) of P. lividus at Lough Hyne from December 1998 to December 2001. Shaded areas represent those times when maximum wind speeds exceeded 30 knots (marked with dashed line)

Meteorological data showed that winds from a southerly direction occurred with the highest frequency during the study period. Winds from this direction blow directly onto our study sites at the north shore of Lough Hyne (Fig. 3). We observed wave action in the shallows to be closely linked to these prevailing winds and to wind speed. Moreover, mortality peaks were frequently coincident with maximum wind speeds. In particular, mortality increased during periods where maximum wind speeds exceeded 30 knots. Elevated wind speeds were recorded during all three ‘winter’ periods sampled, and correspondingly, mortality was also highest during these periods (Fig. 2). Other mortality peaks that occurred during our study period however, were not associated with these high wind speeds. It appears from Fig. 2 that mortality increases also occurred in May, June and July of 2000 and 2001, though wind speeds were low and moderate. High mortality in spring 2001 may to be linked to the major storms of winter 2000/2001 as these were the first observations made after the February storm (no data could be collected during March or April, 2001 due to foot-and-mouth disease restrictions at nature reserves in Ireland).

Predation by M. glacialis did not appear to play a significant role in P. lividus mortality. The echinoid did not feature prominently in the diet of M. glacialis during the 2-year study period. In total, 5,289 observations of M. glacialis were made, of which 598 were observed to be feeding. The diet included a wide variety of taxonomic groups (see Fig. 4), and there was a significant difference in the contribution of different taxonomic groups to the M. glacialis diet in both years (χ=17.08, df=7, P<0.025). However, only nine individual sea stars were found feeding on P. lividus during the study period; the echinoid formed only 4.9% of the total M. glacialis diet in 2000 and 4.3% in 2001 (Fig. 4).
Fig. 3 a

Map of Lough Hyne showing prevailing wind direction. b Wind direction frequency during the 2-year study period (empty bars show prevailing wind direction)

Fig. 4

Relative abundance of P. lividus in the diet of Marthasterias glacialis for 2000 and 2001

Discussion

Patterns in biotic communities are frequently generated by natural physical disturbances such as weather (Dayton 1971; Sousa 1979). In particular, winter storms that increase wave action are known to (1) strongly affect diversity and composition within marine communities, (2) create free space within communities, and (3) interrupt successional processes (Sousa 1979). Echinoids inhabiting wave-swept shores in particular are vulnerable to displacement by drag and lift. The extent to which hydrodynamic forces can play the role of ‘predator’ (sensu Denny 1987) amongst echinoid populations has not, to date, been investigated. Data presented in this study have shown that mortality of adult P. lividus in Lough Hyne increases in coincidence with, and may be driven by, winter storms. We do not, however, consider that such weather-mediated mortality has played a major role in the dramatic decline of this echinoid population in recent decades.

Historically, P. lividus has thrived in the Lough Hyne environment, and was for a long time considered to be the most dominant echinoderm in the lough’s shallow sublittoral (Kitching and Ebling 1961; Ebling et al. 1966). The species plays a central role in shaping algal assemblages within the lough and elsewhere, preventing space monopolisation by dominant competitors, such as Enteromorpha clathara (Kitching and Ebling 1961). However, historical census data compiled in recent years have shown that there has been an effective extinction of P. lividus in many parts of the lough where it was once abundant (Barnes et al. 2002). Data presented here signal a continuing decrease in numbers and point towards negligible recruitment in both 2000 and 2001. In 1995 and 1998, the north shore P. lividus population was dominated by individuals belonging to larger size classes (>21 mm, 2+ age group and older), with fewer than 8% of observed individuals belonging to smaller size classes (1–20 mm, 0+ and 1+ age groups; Barnes et al. 1998). Similarly, extremely low numbers of 0+ age (six individuals in 2000 and none in 2001) and 1+ age individuals were recorded during 2000 and 2001, and the proportion of the population made up by these age classes in 2001 was lower than that found in any previous survey year. Moreover, the mean test diameter of P. lividus was greater in 2000 and 2001 than in previous years, revealing the greater abundance of larger size classes and providing further evidence of the underlying decreased number of new recruits.

For the majority of sampling periods over the course of this study, adult P. lividus mortality appeared to be negligible. Although the north shore is more susceptible to wave action than any other shore at Lough Hyne (Little 1991), mortality through dislodgement still appears to be rare. It appears from these data that mortality through dislodgement may occur only during severe storm disturbance, and evidence of this is seen in peaks of mortality (of varying intensity) occurring in November/December of all years, which coincided with elevated wind speeds. Moreover, during those months where peaks in mortality occurred, the pattern was consistent across at least two and usually all three of our study sites, which suggested that the causal factor acted across a relatively large spatial scale, and was not caused by a localised biotic factor such as predation. Although disturbances within communities are typically heterogeneous in space (Sousa 1979), our data did not show this, probably due to the relatively small spatial scale used in this study.

Though wind driven disturbance may be one of the factors leading to P. lividus mortality, this study does not rule out the possibility that other factors may also play a role in adult P. lividus mortality. Although this population is subtidal, it is possible that desiccation occurs during extremely low tides in summer time, particularly when the diurnal migratory pattern of P. lividus at Lough Hyne is considered—P. lividus migrate to the tops of boulders during the day (Barnes and Crook 2001a), rather than at night, as is the case in Mediterranean populations (Dance 1987). Desiccation of this kind could explain mortality peaks in summer 2000 and 2001, which appear not to be associated with peaks in wind speed data. It is unlikely that summer mortality peaks were driven by reproductive stress caused by the energetic demands of spawning, as in recent years SSTs have rarely reached levels high enough (>17°C; Fenaux 1968) to allow spawning events amongst the P. lividus population, and no mass spawning events or P. lividus larvae were observed either during the course of this study or indeed in 1997 or 1998 (Greenwood 2000).

Predation has also been cited as a potential agent of mortality amongst P. lividus populations. Certainly, predation by the predatory asteroid, M. glacialis was common when the population was much more substantial (Ebling et al. 1966; Kitching 1987), though the extent of the influence of the sea star was not quantitatively demonstrated. Ebling et al. (1966) suggested that the diurnal migratory rhythm of P. lividus at Lough Hyne (which is opposite to the rhythm observed in P. lividus populations elsewhere) has been developed as a result of nocturnal feeding raids into shallower water by the sea star, though these data do not support this suggestion. Importantly, over the past five decades, M. glacialis has expanded from occupying chiefly the south basin to also inhabiting the north basin in large densities, such that the highest densities of the M. glacialis and P. lividus are currently coincident (Verling et al. 2003). The diversity of the M. glacialis diet at Lough Hyne (and probably elsewhere) has been demonstrated in recent years (Verling et al. 2003), and this study highlights the low relative contribution of P. lividus to that diet. Despite the abundance of the asteroid, it does not appear that it feeds preferentially on P. lividus, and thus it is unlikely that predation by M. glacialis can effect changes in P. lividus density.

The location of the study population means that mortality due to storm disturbance would in all probability be lower than it is for other Irish west coast populations, where P. lividus is common in exposed settings, often occupying boreholes (which is thought to offer echinoids a level of protection from displacement) inside tidepools on the intertidal zone. However, during a 2-year sampling period at a large exposed (and boring) P. lividus population in Co. Clare on Ireland’s west coast, no freshly dead test material was found (Verling, unpublished data). Although mortality by dislodgement undoubtedly causes significant mortality amongst these populations, any dislodged echinoids and/or test material likely gets quickly carried out to sea. The sheltered conditions at Lough Hyne might offer P. lividus considerable protection from dislodgement. Nevertheless, it is also possible that the absence of boring behaviour amongst the Lough Hyne population might actually make individuals more vulnerable to displacement during violent winter storms than those intertidal boring populations on the nearby open coast. This is an example of how key behavioural components can have important impacts upon other aspects of echinoid ecology (Verling et al. 2002, 2005).

Natural physical disturbances have pivotal effects on ecosystems, and in particular, when species affected by such a disturbance are ecologically important, the disturbance has the potential to have cascading effects that can affect entire communities (Johnson and Keough 2003). The changing density, and thus the shifting role of P. lividus is brought into sharp focus by the coincident change in the density and distribution of the predatory sea star M. glacialis. Barnes and Crook (2001b) hypothesised that changes in the Lough Hyne population could be due to (1) changes in cohort recruitment or (2) large-scale mortality of adults. Our study supports the theory of Barnes et al. (2001) that the massive historic collapse in this echinoid population has been caused by repeated and prolonged recruitment failure linked to SST. We suggest our data provide reasonable evidence for wave- or storm-induced echinoid mortality and that the levels of mortality could have been much higher when P. lividus existed in densities of hundreds per m2 (Renouf 1931). However, removal of mature individuals in this way only seems likely to account for small-scale population changes.

Acknowledgements

The authors would like to thank the following people for invaluable help with data collection: I.C. Davidson, C.J. Murphy, M.G. O’Mahoney, K.A. Rawlinson and D.I. Watson. We are also grateful to D. O’Donnell for providing research permits to study at Lough Hyne, to J. Bohane for his consistent support of the research effort at Lough Hyne and to the Irish Meteorological Service for providing us with data for the study. We are grateful to two referees whose comments greatly enhanced this manuscript. This work was funded partly by The Department of Zoology, Ecology and Plant Science, University College Cork and by a research grant to E.V. from Enterprise Ireland.

Copyright information

© Springer-Verlag 2005