Coral Reefs

, Volume 25, Issue 1, pp 77–84

Depth limit for reef building corals in the Au’au Channel, S.E. Hawaii

Authors

    • Department of OceanographyUniversity of Hawaii
Report

DOI: 10.1007/s00338-005-0073-6

Cite this article as:
Grigg, R.W. Coral Reefs (2006) 25: 77. doi:10.1007/s00338-005-0073-6

Abstract

In this paper, the relationship between reef building (accretion) and depth in an optimal inter-island channel environment in Hawaii is analyzed. For accretion, the growth rate of Porites lobata is used as a proxy for the reef community, because it is the most abundant and dominant species of reef building coral in Hawaii. Optimal growth of P. lobata occurs at a depth of 6 m, below which both growth rate and abundance decrease with increasing depth. A lower depth limit for this species is found at about 80–100 m, yet reef accretion ceases at ~50 m depth. Below 50 m, rates of bio-erosion of colony holdfasts equal or exceed the growth of basal attachments, causing colonies to detach from the bottom. Continued bio-erosion further erodes and dislodges colonies leading to their breakdown and ultimately to the formation of coralline rubble and sand. Thus, within this channel environment in Hawaii, a threshold for reef building exists at ~ 50 m depth, where coral accretion is interrupted by bio-erosion. Conceptually viewed, this depth horizon is analogous to a vertical Darwin Point, although quite narrow in space and time. More importantly, it explains the history of reef morphology in the Au’au Channel where a chronological hiatus exists at a depth near 50 m. This hiatus separates shallower modern growth (about 100 years or less) from the deeper reef which is all due to accretion during the early Holocene or Pleistocene epochs.

Keywords

HawaiiDepth limitCoral growthReef accretionDrowningDarwin Point

Introduction

It is well known that the growth of coral reefs is light limited and rapidly attenuates with increasing depth (Buddemeier et al.1974; Baker and Weber 1975; Dustan 1975; Chalker 1983; Barnes and Chalker 1990; Kleypas et al. 1999). This, of course, is a result of the relationship between calcification and symbiotic zooxanthellae which can be several times faster in the light versus the dark (Barnes and Chalker 1990). In fact, corals were originally described as plant–animal associations; lithophytes or zoophytes (Quoy and Gaimard 1825). Clearly, any factor that potentially affects light on the reef, such as turbidity, would also be expected to affect the growth of reef building corals.

Temperature may also limit the growth of corals with increasing depth. Stratification in some tropical latitudes may produce a shallow thermocline that sets a depth limit for corals. For example, at the northwestern end of the Hawaiian Archipelago, a shallow thermocline develops during summer months and limits reef growth to depths of 20m or less (Grigg 1981). Upwelling in some settings may also inhibit reef development. In contrast to these examples, at the southeastern end of the Hawaiian chain, temperature does not appear to exert an effect on the profile of vertical growth of reef corals. This is because the mixed layer there normally exceeds 100 m which is below the depth limit set by light for most corals. The same argument would apply to the carbonate saturation state because a depth gradient for this factor does not exist within the mixed layer (Kleypas et al. 1999). In terms of reef growth, the Au’au Channel off Maui at the southeastern end of the Hawaiian chain exemplifies an environment where optimal conditions for coral growth exist (Grigg 1982).

The question addressed in this paper deals with the relationship between coral colony potential growth and reef accretion (Chave et al. 1972). Is the depth limit for coral reef accretion (reef building), the same as the depth limit for coral colony growth? In other words, do coral reefs accrete as deep as individual reef building corals can grow?

Materials and methods

In this study, Porites lobata was selected as representative of the total reef community as it is the most abundant and dominant species of reef building coral in Hawaii (Grigg 1982). P. lobata is also considered representative of other Hawaiian corals, because its growth approximates the growth in mean solid radius of corals of substantially different growth form, i.e., branching and encrusting species (Maragos 1972) and because it is intermediate in growth rate (Buddemeier et al. 1974).

Colonies of P. lobata were collected at depths of 3, 6, 10, 12, 24, and 50 m off Lahaina, Maui in the Au’au Channel at stations at increasing depth and distance from shore to mid-channel about 5 km offshore (Fig. 1 and Table 2). The deepest and most seaward station was Stonewall, a limestone outcrop that rises from 90 to 32 m (Fig. 5, Grigg et al. 2002). Samples of carbonate substrata and observations of the substratum and bottom morphology down to 120 m throughout the Au’au Channel were conducted with the use of the Pisces V submersible in which over 30 h of bottom time were spent by the author. All dives shallower than 50 m were accomplished with SCUBA.
https://static-content.springer.com/image/art%3A10.1007%2Fs00338-005-0073-6/MediaObjects/338_2005_73_Fig1_HTML.gif
Fig. 1

The Au’au Channel separates the islands of Maui and Lanai and is characterized by oceanic circulation producing maximum water clarity and optimal conditions for coral growth. Stations (numbered 1–6) were located at increasing depths and distances from the shore along a 5-km transect ranging from Maui toward Lanai along a 250° bearing

The growth rate of P. lobata was determined by measuring the width of about ten annual growth bands (Knutson et al. 1972) in skeletal cross sections of between five and seven colonies at each station (depth). Cross sections 6–7 mm in thickness were cut along axes of maximum growth in the corals. These sections were X-rayed producing positive contact prints and a clear record of annual linear extension. The relationship between this measure of annual growth and depth was analyzed by least squares linear and non-linear regression (Fig. 2).
https://static-content.springer.com/image/art%3A10.1007%2Fs00338-005-0073-6/MediaObjects/338_2005_73_Fig2_HTML.gif
Fig. 2

The best-fit regressions with 95% confidence limits (dotted lines) for P. lobata growth data at stations on the Lahaina Transect. The linear plot (gray) includes all data points (n=384); r2=0.64; growth=13.37−0.21×depth. The exponential plot (black line) excludes the data from 3 m (n=345); r2=0.68; growth=16.2e−0.032depth. The open circles are the mean growth data for colonies. Values for growth predicted from both regressions at 50 m are shown in the box

Dominant taxa responsible for bio-erosion were identified and described in terms of their abundance and the degree to which they led to the destruction of holdfasts. Bio-erosion was measured by deduction. At depths below about 50 m, it was found that colonies failed to remain attached to the substratum. It was therefore reasoned that the rate of bio-erosion of colony holdfasts was equal to or exceeded the radial growth rate at this depth. At 50 m, vertical and horizontal (radial) growth at the colony base are nearly equivalent. In other words, the rate of bio-erosion on the undersurface of colonies was assumed to have exceeded the rate of linear accretion at a size and depth when the holdfast was no longer attached to the substratum. While qualitative and approximate, this method proved to be a direct measure of the depth limit of reef accretion. Colonies that do not remain fixed to the substratum do not contribute to the reef building process. In the Au’au Channel, colonies below about 50 m are unattached and gradually break down into sand due to further biological and physical erosion. The sand is ultimately washed out of the channel by strong bottom currents onto the deeper shelves that surround the islands.

Along the lower section of Stonewall, samples of bare limestone substratum were collected at depths between 47 and 60 m for 14C age dating. At these depths, the age of the substratum is important in order to identify the depth at which reef accretion stops. Samples of bare limestone were not collected at shallower depths for age dating because there the reef is clearly modern in origin being dominated by a very high percentage of live coral cover. Samples at greater depths were not age dated due to budgetary limitations. All carbonate samples collected for age dating were cleaned in H2O2, acid etched and X-rayed. They were then sent to the NOSAM Laboratory at Woods Hole Oceanographic Institution for age dating (See Table 1).
Table 1

Samples of carbonate substratum and 14C ages from Station 6 off Lahaina, Maui

Depth (m)

Aragonite/calcite(%)

Calcite age (years before present)

Sample characteristics

45.7

100/0

Modern

P. lobata skeleton

45.7

100/0

143

Clean coral skeleton

45.7

100/0

50

Clean coral skeleton

48.8

98/2

Modern

P. lobata skeleton

48.8

100/0

Modern

Montipora skeleton

48.8

100/0

Modern

Montipora skeleton

48.8

98/2

Modern

Coral skeleton

48.8

98/2

Modern

Coral skeleton

51.8

97.7/2.3

9,617

P. lobata skeleton

54.8

96.8/3.2

8,978

P. lobata skeleton

54.8

88.2/11.8

8,910

P. lobata skeleton

57.9

94.9/5.1

8,032

P. lobata skeleton

57.9

89.1/10.9

9,025

P. lobata skeleton

The relationship between depth and temperature (Fig. 3) and depth and light (Fig. 4) was also determined. Temperature was measured with a combined temperature and depth sensor (CTD) mounted on the submersible or a remotely operated vehicle (ROV) during dives in the Au’au Channel in November, 2001. The CTD records collected in summer months show a similar pattern of temperature with depth except that surface values are closer to 28°C. In the absence of direct field measurements of underwater irradiance, downwelling irrradiance was computed from the relationship (Kirk 1994). \( E_{{\text{d}}} (Z) = E_{{\text{d}}} (0^{ - } ){\text{e}}^{{ - K_{{\text{d}}} Z}} \)Where Ed(Z) is the downwelling irradiance at depth Z, Ed(0) is the irradiance just below the sea surface, and Kd is the diffuse attenuation coefficient. A diffuse attenuation coefficient (Kd) for PAR (photosynthetically active radiation 400–700 nm) of 0.032 m-1 was used. This value of Kd was the average recorded 7 km offshore in a similar oceanic water type from nearby Oahu, Hawaii (Bienfang et al. 1984)
https://static-content.springer.com/image/art%3A10.1007%2Fs00338-005-0073-6/MediaObjects/338_2005_73_Fig3_HTML.gif
Fig. 3

Temperature versus depth on November 28, 2001 in the middle of the Au’au Channel between Maui and Lanai. Four readings are shown from separate submersible dives (RCV 122–125). Over the depth range of P. lobata, from 0 to 90 m, temperature ranged between 26.3°C (surface) and 24.0°C (90 m), well within the tolerance range for this species

https://static-content.springer.com/image/art%3A10.1007%2Fs00338-005-0073-6/MediaObjects/338_2005_73_Fig4_HTML.gif
Fig. 4

Computation of downwelling photosynthetically active radiation (PAR: 400–700 nm) versus depth for the Au’au Channel using a diffuse attenuation co-efficient for PAR measured at Oahu, Hawaii. Six percent of surface PAR irradiance remains at 90 m

Results

The growth of P. lobata in the Au’au Channel off Lahaina, Maui (Fig. 1) can be considered optimal relative to other locales in the Hawaiian Islands for the following reasons. Water quality is optimal, because the Au’au Channel is flushed by tidal currents that transport pure oceanic water between the islands of Maui and Lanai on a semi-diurnal basis. Run-off from the land is minimal and infrequent and rarely diminishes water quality in the channel habitat. Light transmission is therefore optimal. At 80–100 m where P. lobata drops out, percent surface irradiance is about 5% (Fig. 3), approximating the compensation point of this species.

Being near the southern-most latitude in the archipelago, temperature is also non-limiting (Fig. 4). From the surface to 90 m, the annual temperature range is approximately 28–24°C, well within optimal temperature limits for this species (Kleypas et al. 1999). The location of stations running offshore from Lahaina, Maui can also be considered optimal for coral growth because the entire area is well sheltered from both summer and winter long period swell.

The growth data for P. lobata is based on 34 colonies collected from six stations in the channel (Table 2). Collectively, these colonies represent 384 years of growth data. Except for a slight inhibition in growth at the shallowest station at 3 m, the pattern of change with depth is negative (Fig. 2). If all data are included in the analysis, the regression of the best-fit curve is linear (Fig. 2). Growth at 3 m depth appears to be inhibited relative to the rate at 6 m (Table 2; two-tailed t test P=0.034), and the best-fit regressions indicate that growth at this depth should be 12.75 mm year-1 (linear) and 14.71 mm year-1 (exponential) compared to a measured mean of 11.60 mm year-1. This may be due to a negative effect of factors associated with shallow water, such as high levels of solar ultraviolet radiation, increased turbidity, or episodic sedimentation (Jokiel and Coles 1990). Ignoring the data at 3 m produces an exponential pattern of decreasing growth with depth (Fig. 2). A goodness of fit statistic (r squared) for the linear model is 0.64 compared to 0.68 for the exponential model. The exponential model provides a slightly better fit to the data between 6 and 50 m. A similar pattern of negative exponential growth has been found for other massive reef building corals in both the Western Pacific, P. lobata (Buddemeier et al. 1974), and the Caribbean Sea, Montastraea annularis (Dustan 1975).
Table 2

Mean linear extension rates of P. lobata at six depths in the Au’au Channel, Maui, Hawaii

Station

Depth (m)

Number of colonies

Cumulative annual growth bands (years)

Mean linear extension (mm years-1)

Standard error (± mm years-1)

I

3

5

39

11.60

0.61

II

6

5

36

13.49

0.64

III

10

6

67

11.57

0.47

IV

12

5

51

10.81

0.48

V

24

6

75

8.02

0.24

VI

50

7

116

3.02

0.05

Total

 

34

384

  
The exponential growth model predicts that P. lobata should be able to grow at depths in excess of 80 m in the Au’au Channel. This is in good agreement with observation since colonies have been found down to 90–100 m. It also accords with the relationship between PAR and depth (Fig. 4) which indicates that 6% of surface irradiance may penetrate to 90 m. Both the exponential growth relationship and the PAR reduction show the same coefficient for the exponential term (0.032), in other words both curves have the same shape. In spite of this, reef accretion or reef building is observed to cease about 50 m depth. While colonies can and do grow much deeper than 50 m, they are not permanently attached to the substratum. Below 50 m, colonies eventually break off the bottom due to bio-erosion undermining their holdfasts (Fig. 5).
https://static-content.springer.com/image/art%3A10.1007%2Fs00338-005-0073-6/MediaObjects/338_2005_73_Fig5_HTML.gif
Fig. 5

X-radiographs of a thin section of two colonies of P. lobata collected between 50 and 55 m at Station 6. Note the platy growth form (wider than tall) and bio-erosion along the undersurface of the colony caused by clionid sponges and bivalves. Annual growth bands are also visible in the figures

During the submersible dive series, no colonies of P. lobata larger than ~15 by 30 cm were observed fixed to the substratum at depths deeper than 50 m, e.g., all were unattached. In all cases, colonies of P. lobata at depths below 50 m were platy in growth form, being broader than tall (Dustan 1975) (Fig. 5). At 50 m, lateral or radial growth was approximately equal to vertical growth.

As mentioned above, the rate of bio-erosion at 50 m was measured indirectly. At 50 m in the Au’au Channel, the vertical growth rate of P. lobata was 3.02 (±0.05 SE) mm year-1(Table 2). Because permanent reef structures were not forming below this depth, the negative inverse of this rate was taken as the average rate of bio-erosion of colony holdfasts. In other words, losses due to bio-erosion at 50 m must equal or exceed gains due to the growth of P. lobata. It is reassuring that the bio-erosion rate calculated in this manner falls well within the range of estimates (0.5–4.0 mm year-1) that have been calculated elsewhere in the world, e.g., Bikini, Guam, Aldabra, Bermuda, Western Australia and the Great Barrier Reef (Davies 1983, Table I).

In summary, as colonies slowly accrete vertically at depths greater than 50 m, bio-erosion actively undermines their basal attachment sites (Fig. 5). The most active bio-eroders are clionid sponges which alone account for about 80% of the loss of calcium carbonate. The large boring bivalves, Lithophaga sp., Arca ventricosa and Rocellaria hawaiiensis also contribute significantly to weakening holdfasts of P. lobata.

Clionid sponges are endolithic borers that excavate a multitude of tiny chambers with even smaller galleries branching off the main chambers. Literally thousands of holes produced by these borers were found on the undersurface of all colonies of P.lobata collected below 50 m (Fig. 5). The thickness of clionid penetration was rarely more than 2–3 cm along the undersurface. Nevertheless, the sponges slowly etch inward eventually pinching off the holdfast. Apparently, the undersurfaces of platy colonies of P. lobata are ideal substrata for clionid sponge larvae to settle since they are not covered by living coral tissue (Wilkinson 1983). Boring clionid sponges are considered to be the most destructive bio-eroding organisms found on coral reefs globally (Wilkinson 1983; Wood 1999).

The boring bivalves Lithophaga and Arca are also commonly found within the holdfasts of P. lobata. They excavate bare oval areas up to 3 by 6 cm. The mantle glands of Lithophaga secrete acid that dissolves and weakens the surrounding limestone structure making holdfasts more prone to fracture. There appears to be no significant secondary infilling due to cementation or lithification of sediment which in this environment is swept away by vigorous bottom current. Like the clionid sponges, all of the species of boring bivalves found in P. lobata are filter feeders and are well adapted to high current environments.

Perhaps, the most unique result of this study was the discovery of a sharp hiatus in the age of samples of bare limestone substratum at a depth of about 50 m. Above 50 m, all samples carbon dated turned out to be modern, younger than about 100 years (Table 1). In sharp contrast, samples collected below 50 m, were all between 8,000 and 9,600 years old. Although these samples are few in number, they clearly demonstrate that reef accretion below 50 m during most of the Holocene has not occurred. In essence, the modern reef stops at 50 m even though the growth of individual corals may extend considerably deeper. Colonies of Leptoseris spp. which grow as deep as 110 m in the Au’au Channel are also unattached.

Another strong indication of the absence of Holocene reef building in the Au’au Channel below 50 m is the retention of Pleistocene morphology there (undercut notches and wave cut terraces). Holocene accretion had it occurred would have otherwise masked these features by overgrowth (Fig. 6a, b). These results indicate that all of the karst topography below about 50 m in the Au’au Channel is of early Holocene or Pleistocene age (Grigg et al. 2002).
https://static-content.springer.com/image/art%3A10.1007%2Fs00338-005-0073-6/MediaObjects/338_2005_73_Fig6_HTML.gif
Fig. 6

a and b Underwater photographs from the Au’au Channel showing sea caves, near 50 m depth, eroded by paleo-sea level stands. These are examples of Pleistocene morphology that has not been altered or covered by Holocene growth. Right hand scale is depth in meters

Discussion

A decrease in the rate of growth of reef building corals as a function of increasing depth is not a new finding. Similar patterns have been described for massive corals in both the Pacific and Caribbean (Buddemeier et al. 1974; Dustan, 1975). What is new, is the limiting effect of bio-erosion which effectively decreases by about 50% the critical depth to which coral reefs can build permanent structures (accrete) at least in this Hawaiian inter-island channel environment. Reef building corals can and do grow at depths to over 100 m in this environment, yet reefs do not accrete deeper than 50 m under present day conditions. Limestone structures deeper than this represent fossil reef growth deposited during the early Holocene or the Pleistocene when the reefs were in shallower water (Grigg et al. 2002). The extent of structures below 50 m covers almost 100 sq km of bottom area extending across the entire Au’au Channel.

The depth limit to which reef building corals develop permanent reefs has long been considered a function of decreasing light (Wells 1957; Dustan 1975; Buddemeier et al 1974; Kleypas et al. 1999). This study shows reduced growth rates of P. lobata with increasing water depth which mirrors the reduced availability of PAR. Both the measured rate of growth (Fig. 2) and the calculated light availability (Fig. 4) suggest that there is still potential for coral growth down to about 100 m in the Au’au Channel. However, the results also demonstrate that bio-erosion, by many endolithic boring organisms, intervenes by setting a shallower depth limit of 50 m for reef building than would otherwise occur by light availability alone. At this threshold for reef building, net carbonate production approaches zero. This threshold is produced because growth rates are exceeded by rates of bio-erosion. In other environmental settings where bio-erosion is of less significance, the depth limit of reef building may be more a direct function of light, or even temperature in areas of upwelling or shallow stratification.

Conceptually viewed, the depth horizon at 50 m in the Au’au Channel is analogous to a vertical Darwin Point, although quite narrow in space and time. A broader zoogeographic Darwin Point exists at the northwestern end of the Hawaiian Archipelago, where islands drown because net carbonate production of the reefs there approaches zero due to reduced coral growth, bio-erosion, mechanical losses and subsidence (Grigg 1982). At the southeastern end of the Hawaiian chain, reefs drown if they fall below 50 m due to rising sea level. This threshold for drowning is set by reduced coral growth due to decreasing light, and the inability of coral colonies to permanently attach to the substratum due to bio-erosion. In less optimal environments for coral growth in Hawaii, the depth limit for reef building may even be shallower (Grigg and Epp 1989).

Acknowledgements

This research was supported by a grant from the National Sea Grant Program to the author. I would also like to thank Eric Grossman for helping to collect and analyze the samples of carbonate. Robin Lee provided vessel time at no cost to the project and provided diver support. Submersible time (five dives) in the Pisces V was awarded to the project by the Hawaii Undersea Research Laboratory. Terry Kirby and Chuck Hathoway were the pilots of the Pisces V. Eric Hochberg greatly assisted with the modeling of down-welling irradiance and the preparation of Fig. 4, Nancy Hulbirt prepared Figs. 1 and 5 of the manuscript, and Richard Dunne assisted with the preparation of Fig. 2. Eric Hochberg and Sam Kahng reviewed the manuscript and provided helpful suggestions for improvement.

Copyright information

© Springer-Verlag 2005