At all sites the zone between the neap low tide mark (upper limit of the mussel zone) and the upper limit of the occurrence of the limpet Putellugrunuluris.
Estuarine,
Coasta
and
Shelf Science (1986) 23,283-294
A Field Experiment Influence of Seabird Primary Production
A. L. Bosman”, G. M. Branchb “FitzPatrick and bZoology Africa Received
J. T. Du Toitb,
P. A. R. Hockey”
and
Institute, University of Cape Town, Rondebosch 7700, South Africa Department, University of Cape Town, Rondebosch 7700, South
22 May
Keywords:
Demonstrating the Guano on Intertidal
1985 and in revisedform
birds; guano; nutrients;
8 October
intertidal
198.5
environment;
algae
Nutrient concentrations in intertidal rockpools on the shores of seabird breeding islands off the west coast of South Africa are significantly higher than concentrations in similar pools on mainland shores. This is due to the presence of terrestrial and upper intertidal guano deposits which are washed through the intertidal zone. Growth rates of intertidal macroalgae are significantly faster on the rocky shores of seabird breeding islands than on mainland rocky shores, and the extensive, permanent algal mats that develop on island shores are absent from mainland shores. The hypothesis that guano run-off enhances intertidal algal growth was tested experimentally at a site on the west coast of South Africa where no seabird colonies exist. Seawater, artificially enriched with guano, was trickled down an intertidal rocky slope during a period of two months. Algal settlement and growth on artificial surfaces were significantly enhanced by the guano treatment when compared with controls. The conclusion is that increased intertidal primary production on seabird islands, and the consequent modification of intertidal community structure, are due in part to nutrients which are introduced into the intertidal zone by run-off from the guano cap and by direct deposition of guano in the high-shore region by roosting birds.
Introduction Large numbers of colonial seabirds aggregate on islands off the southwestern Cape coast of South Africa (Cooper et al., 1984). In these protected areas where the birds breed and roost, guano deposition is considerable and guano may accumulate to the extent where it can be removed and exploited commercially as an agricultural fertilizer (Rand, 1963). Under natural conditions, rain, sea spray and waves wash guano from bird colonies into the sea, where it can cause local nutrient enrichment (Golovkin & Garkavaya, 1975; Golovkin, 1976; Golovkin et al., 1976) which may enhance marine phytoplankton production (Golovkin, 1967; Zelickman & Golovkin, 1972; Golovkin & Garkavaya, 1975). Little is known of the effects that these leaching nutrients have on the settlement, growth rates and production of intertidal macroalgae. 283 0271-7714]86/090283+
12 $02.00/O
01986
Academic
Press Inc. (Londoni
Limited
284
A. L. Bosman et al.
Extensive, permanent mats of macroalgaegrow on the rocky shoresof the southwestern Cape seabird islands; in marked contrast, such mats are absent from the shoresof nearby mainland sites where no large collections of seabirds occur. It is possible that on island shoresgeneral elevation of nutrient status occurs when seabird guano enters the intertidal and nearshore waters. As a result of nutrient enrichment, algal growth would be more rapid and the formation of extensive algal mats would be facilitated. To test this hypothesis, comparisons of nutrient concentrations in intertidal rockpools were made at two siteson seabird breeding islandsin Saldanha Bay and at two nearby mainland sites(Figure 1). The rates of algal growth at different shore levels at eachof the siteswere measuredand compared. A field experiment was undertaken to determine the relationship between macroalgal settlement and growth, and the levels of intertidal nutrients on a rocky shore artificially enriched with guano run-off. Methods Study sites
Two study areaswere used. In Saldanha Bay, monthly measurementsof nutrient concentrations and algal growth rates were made at two intertidal sites on islands (Jutten and Marcus Islands, Figure 1) where Cape cormorants Phalacrocorax cupensis (Sparrman), jackasspenguins Spheniscus demersus (L.) and other seabirds breed. Nest sites may total 50 000 for Cape cormorants and 7000 for jackasspenguins on the two islands (Cooper et al., 1982; Shelton et al., 1984). Many thousands of non-breeding birds alsoroost on these islands. Control sites were chosen on the mainland outside the Bay and away from any seabird colonies (Mauritz Bay and Cape Columbine, Figure 1). At all sitesthe shoreswere gently sloping granite and were exposed to strong wave action. In addition, a field experiment was run at Menskoppunt, a rocky promontory on the west coast of the Capeof Good Hope peninsula (Figure 1) where, generally conditions were comparable except that the substratum was of hard Table Mountain sandstonerather than granite. There were no seabird colonies in the vicinity of this site. At all sitesthe zone between the neap low tide mark (upper limit of the musselzone) and the upper limit of the occurrence of the limpet Putellugrunuluris (L.) was studied. At each site the gently sloping rocky shore between the upper and lower limits specified was divided into a low-, mid- and high-shore region ofapproximately equal area. On the shores in and around Saldanha Bay the limpet P. granularis totally dominated the intertidal zone in terms of number and biomass; the limpets P. grunutinu (L.) and P. oculus Born. occurred in the low-shore zones, and Siphonuriu cupensis Quoy & Gaimard inhabited shallow pools, but none of these specieswas as numerous or as widely distributed as P. granularis. The mussels Choromytilus meridionalis (Krauss), Mytilus gulloprovincialis Lamarck and Aulucomyu uter (Molina) were dominant in the musselzone, below the neap low tide mark, at all sitesexcept Cape Columbine. Extensive mats of Enteromorphu sp. and Porphyru cupensis Kiitz occurred in the high- and mid-shore regions at the island sites, covering up to approximately 50% of the bare rock surface. Algal mats were absent from mainland sites. At Menskoppunt the small whelk Littorinu ufricunu Phillipi was abundant on the exposed high-shore, and the dominant mid-shore grazers were the limpet P. grunuluris and the winkle Oxystele vuriegatu (Anton). The low-shore region was dominated by P. granutinu, P. granularis, and Oxystele tigrinu (Chemnitz), and the infratidal fringe supported a sparse population of Putellu urgenvillei (Krauss). There were few intertidal macroalgae at Menskoppunt: small patches of Porphyru cupensis, Ulvu sp.,
285
Influence of seabirdguano
Cape Columbine-, Mauritz
danha Bay 330 s
___L---
Jutten
Island
Republic South
Atlantic
,
I
zokm
of
Africa
Ocean
,
Menskoppunt d l!‘E Figure 1. Map study sites near Menskoppunt.
SPoint
of the southwestern Cape coastline Saldanha Bay, and the experimental
i showing the positipns of the site on the Cape Peninsula at
Chuetangium erinaceum (Turn.) Papenf., C. ovale (Sur) Papenf. and Gigartina radula (Esp.) J. Ag. were present in the low-shore region. Enteromorpha sp. was present in rockpools and gullies together with the moss-like rhodophyte Caulacanthus ustulatus (Suhr). Papenf. The rock surface of almost the entire study site at Menskoppunt was colonized by Hildenbrandia sp., a red encrusting alga (seeStephenson & Stephenson, 1972 for general shore descriptions). Nutrients
Monthly intertidal seawatersampleswere collected from a high-, a mid- and a low-shore rockpool at each site in the Saldanha Bay area during the period February-May 1983. At each site the high-shore pool was chosen at the upper limit of the occurrence of P. granularis and the low-shore pool at the neap low tide mark. Although the pools were not
286
A. L. Bosman
et al.
of a standard size, they were all flushed regularly at high tide. Each month two replicate water sampleswere taken from the chosen pools at each site, during a spring low tide. These were filtered using Watman glassmicrofibre filters, and were then stored frozen in the dark, in freezers free of organic matter, until the concentrations of NO, -N, NO; -N, and PO:- - P were determined using a Technicon autoanalyser. Two replicateswere taken from each rockpool, and the mean of the two values for each nutrient was used in calculations, except in caseswhere the two values differed by more than 200,,, in which case the sample was discarded. The mean of the nutrient levels from the three rockpools at each site was used as an indication of intertidal nutrient status. Island and mainland mean intertidal nutrient levels were tested for statistical difference using Mann-Whitney U tests with significance levels derived from approximations to the normal distribution (two-tailed) (Sokal& Rohlf, 1981). Algalgrowth
Herbivore exclusion plots were established on the rock at the low-, mid- and high-shore levels at each site by painting borders around 0.1 m2 rectangular areaswith copper-based, marine anti-fouling paint. Intertidal grazers such asPatella spp. and Oxystele spp. do not crossstrips of this paint. Each month four high intensity polystyrene ‘productivity’ strips (McQuaid, 1981) were glued to the rock inside each exclusion plot, using rapid-drying Araldite epoxy glue. The rock surface was scraped to remove settled algae and invertebrates, and then was burned dry with a blowtorch before the 30 x 50 x 2 mm strips were glued down. At the end of eachmonth the strips were lifted, wrapped in aluminium foil to exclude light, and frozen. The strips were scraped clean and the chlorophyll-a present in the algae that had settled on each strip was extracted using the method of Strickland & Parsons (1972). The quantity of chlorophyll-a extracted was estimated using the colorimetric techniques of Jeffrey &Humphrey (1975), and wasused asan index of algal growth after corrections had been made for the exact area of the strip and for the number of days the strip had been in position. The final index of algal growth was expressed in ug chlorophyll-a cm-’ month- ‘. Field experiment
Concurrently with the collection of data in the Saldanha Bay area, a field experiment was run at Menskoppunt to establish whether a direct relationship exists between nutrient enrichment and algal growth. A section of rocky shore from the upper limit of macroalgal growth to the mussel zone was enriched using a solution of dried, sifted seabird guano (from Malgas Island in Saldanha Bay) and seawater. Two kg of dried, sifted guano were dissolved in 1350 1 of seawater, thus significantly increasing the nutrient status of the seawaterthat wasto enter the intertidal zone, such that even after dilution by waves in the intertidal zone, excessnutrients would be available to intertidal algae.After the first week, weekly samplesof the guano solution were taken for nutrient analysis asdescribed above. The guano solution was stored in a 1350 1 capacity asbestoscement tank which was installed above the high-water mark. From this reservoir, the guano solution passed through a constant pressure tank, which regulated the flow, and down a short length of plastic irrigation piping. At the high-shore mark the piping wasjoined, using a T-piece, to the two arms of the drip bar which measured approximately 4 m in length (Figure 2). Nozzles punched in the drip bar provided a flow of 2 1h-’ of guano solution. Similar apparatus was used to drip unenriched seawater down an adjacent section of shore approximately 5 m to one side of the guano-drip area. The tanks were filled every week. A
287
influence of seabirdguano
SEAWATER \
DRY CONTROL
HIGH
[ wstrip
\
MID
jH/
,
I
j/
,
exclusion plot
nnn loi
UUU Eli
__
.
____ LOW
______ -...__
WATER
Figure 2. Plan diagram of the apparatus layout and sampling stations low-shore) in the study area at Menskoppunt (not drawn to scale).
(at high-,
mid-
and
third section of adjacent shore, approximately 5 m to the opposite side of the guano-drip area (Figure 2), was a ‘dry-control’ which had neither seawater nor guano solution running over it from the high-shore. Algal growth rates were measured using productivity strips glued to the rock inside herbivore exclusion plots as described previously. In the guano-drip and dry-control areas there were exclusion plots at the low-, mid- and high-shore levels, but the topography of the shore prevented the establishment of a high-shore plot in the seawaterdrip area. In this instance algal growth was measured at two levels only (Figure 2). Four strips were glued down inside each plot and after a month in position two were removed from each set of four and were replaced with new strips. After a further month all remaining strips were lifted. The development of microflora on the rocky substratum in the three experimental areas was compared qualitatively by chipping small flakes from the rock surface and examining theseby meansof a scanning electron microscope. Flakes were collected from areasof rock that had been scrubbed and burned for useasexclusion plots, and they were collected two
288
A. L. Bosman
et al.
TABLE 1. Mean intertidal nutrient concentrations (ug at l- ‘) and standard deviations (where samples number more than two) at study sites in the Saldanha Bay area. A high-, a mid- and a low-shore rockpool were sampled each month NO,
-N
NO,-N
PO,-P
Mean
SD.
Mean
S.D.
Jutten Feb Mar Apr May
0.5 1.0 2.0 1.9
0.2 0.6 2.1 0.2
21.2 11.5 16.2
Marcus Feb Mar Apr May
1.0 0.9 2.1
0.4 0.4 0.5
2.9 8.2 10.5 24.0
Mauritz Feb Mar Apr May
0.5 0.5 0.8 1.1
0.2 0.1 0.3 0.2
6.8 13.6 6.1 8.3
4.7
Columbine Feb Mar Apr May
0.5 0.3 0.5 1.1
0.3 0.2 0.2
7.3 6.9 2.0 7.6
9.6
8.3
Mean
1.0 3.5 2.4 2.5
2.0 3.1
S.D.
1.1
17.0 1.5
7.7 2.6
5.7 3.3
0.6
0.8 1.5 2.0 1.1
0.6 0.4 0.6 0.2
1.5 1.5 2.1 0.9
1.1 0.2 0.1
months after the plots were established. The rock flakes were fixed in 100/b formalin (in seawater) and then dehydrated with increasing concentrations of ethyl alcohol followed by critical-point drying from lOOoh ethyl alcohol.
Results The mean nutrient concentrations recorded in intertidal rockpools (Table 1) were significantly higher at island sites in Saldanha Bay than at mainland sites in the vicinity (Mann-Whitney U tests 2-tailed, NO, -N: n, = 8, n2 = 7, U= 45.00, p < O-05; NO, -N: n1=7, n2=8, U=45*00, p
