Oikos 116: 75 86, 2007 doi: 10.1111/j.2006.0030-1299.15230.x, Copyright # Oikos 2007, ISSN 0030-1299 Subject Editor: Dag Hessen, Accepted 29 August 2006
The Ulva connection: marine algae subsidize terrestrial predators in coastal Peru Alessandro Catenazzi and Maureen A. Donnelly A. Catenazzi (
[email protected]) and M. A. Donnelly, Dept of Biological Sciences, Florida International Univ., Miami, FL 33199, USA.
How can terrestrial animals survive in a desert with scant primary productivity? The Peruvian coastal desert is hyper-arid, but faces one of the world’s most productive marine ecosystems, the Peru Chile cold current. Given the stark difference in productivity between these adjacent ecosystems, we expected to find strong linkages connecting the terrestrial and marine food web. We investigated how marine resources are incorporated in the diet, and influence the distribution of terrestrial consumers (geckos, scorpions, solifuges and darkling beetles). Stomach contents from geckos, and d13C and d15N values of geckos and other terrestrial consumers suggest that marine green algae of the genus Ulva provide energy and nutrients to the terrestrial food web. Isotopic values suggest that amphipods, which feed on stranded Ulva , make marine resources available to terrestrial predators by moving between the intertidal and supratidal zones. The relative contribution of terrestrial and algal carbon sources varied among terrestrial predators, because scorpions assimilated a lower proportion of energy from Ulva than did geckos and solifuges. These d13C patterns reflected differences in the spatial distribution of consumers. Our study supports the idea that in places where ecosystems with contrasting productivity levels are spatially juxtaposed, it is not possible to understand the structure and dynamics of food webs without taking into account the effects of energy and nutrients flowing from adjacent ecosystems. In contrast to other studied systems, especially those in Baja California, our site in Peru receives very little rainfall and the amount of precipitation is not affected by El Nin˜o events. The near absence of rainfall promotes an extreme dependence of terrestrial consumers on marine resources, and causes permanent indirect food-web effects that are affected by temporal variability in marine productivity, rather than temporal patterns of plant growth.
Food webs in coastal and riparian habitats are linked to aquatic ecosystems by energy, nutrients and material flowing through abiotic and biotic vectors (Bastow et al. 2002, Polis et al. 2004). Ecologists have long recognized the presence of these linkages (Summerhayes and Elton 1923, Lindeman 1942, Koepcke and Koepcke 1952, Odum 1957), yet only recent studies have proposed a theoretical framework (Polis et al. 1997, Loreau and Holt 2004) and generated empirical evidence (Polis and Hurd 1996a, 1996b, Nakano and Murakami 2001) to support the idea that the movement of nutrients and materials have profound influences on the structure and dynamics of ecosystems. Linkages between coastal habitats and adjacent land include a variety of flows, often mediated through rivers and streams. Estuaries and bays are critical ecosystems
where inputs of pollutants and of excessive amounts of nutrients from anthropogenic sources cause important changes in the structure and dynamics of nutrient cycles and food webs (McClelland and Valiela 1998, Vo¨ro¨smartry and Meybeck 2000). However, coastal and adjacent terrestrial systems are also linked through fluxes from sea to land by the transport of sand, salt, sea aerosols, guano, energy and debris from hurricanes, stranding of marine organisms and carcasses through tidal action, and movements of marine organisms from sea to land (Polis et al. 2004). Marine subsidies to near shore communities influence consumers at almost any trophic level (Mizutani and Wada 1988, Polis and Hurd 1996a), including plants (Sa´nchez-Pin˜ero and Polis 2000, Farin˜a et al. 2003), herbivores (Onuf et al. 1977, Lindeboom 1984), scavengers (Koepcke and
75
Koepcke 1952, Heatwole 1971), omnivore consumers (Stapp and Polis 2003) and predators (Polis and Hurd 1995, Rose and Polis 1998, Cree et al. 1999, Barrett et al. 2005). Seaweeds are important contributors to near shore food webs (Duggins et al. 1989, Bustamante et al. 1995, Bustamante and Branch 1996). Seaweed drift, especially of kelp in upwelling systems, can amount to more than 2 000 kg m 1 yr1 (Koop and Field 1980). Stranded algae are converted into abundant populations of intertidal and supratidal algivores and predators that inhabit sandy and rocky beaches (Polis et al. 2004). Places where tremendously productive oceanic waters meet unproductive habitats are among the most remarkable examples of energy and trophic flows among ecosystems (Polis and Hurd 1996a). The Pacific coast of Peru is an extreme example where highly productive marine waters associated with upwelling of the Peru-Chile current face one of the most arid areas in the world, the Atacama desert. The Peru Chile current system has a phytoplankton primary productivity of 3.84 mg C m2 day1 and a bacterioplankton biomass of up to 4.05 g C m2 day 1, making it one the world’s richest marine areas (Arntz et al. 1985, Tarazona and Arntz 2001), whereas recorded rainfall precipitation in southern coastal Peru is less than 2 mm year 1 (Craig and Psuty 1968). Based on Lieth’s (1978) model of primary productivity as a function of annual precipitation, we estimate that desert primary productivity is less than 10 g m2 yr 1 dry mass. Several authors have described the dependence of terrestrial consumers on marine resources in Peru (Murphy 1925, Vogt 1942, Koepcke and Koepcke 1952), but no recent studies have focused on the effects of these flows on the dynamics of terrestrial communities along the coastal desert. The input of marine-derived energy and nutrients can support a disproportionate abundance of consumers and predators, as shown by Polis and Hurd (1995) for island systems in Baja California. In this study, we hypothesized that marine green algae of the genus Ulva would provide an important source of energy and nutrients to, and would influence the spatial distribution of, terrestrial consumers in the Peruvian coastal desert. In contrast to studies in Baja California, our site does not receive annual rainfall above 10 mm year 1 and the amount of precipitation does not vary during El Nin˜o Southern Oscillation events. The near absence of rainfall should promote a stronger dependence of terrestrial consumers on marine-derived resources than what has been reported from other coastal systems. We combined four approaches in our study. We surveyed the distribution of terrestrial consumers with respect to the intertidal zone. We counted the number of consumers in the intertidal mats of stranded Ulva to
76
estimate the amount of invertebrate prey available to desert consumers. We collected stomach contents from geckos to document the occurrence of intertidal algivores in the diet of a top predator. Finally, we used analyses of stable isotope ratios of carbon (d13C) to estimate the proportion of carbon that was ultimately derived from Ulva , and of nitrogen (d15N) to estimate the trophic position of desert consumers.
Methods Study site Fieldwork was conducted in Paracas Bay, Ica Region, Peru (13851?S/76816?W) during the austral summer (January to April) of 2003. The climate is characteristic of the arid coastal desert of Peru and northern Chile, and our site receives less than 2 mm of rain per year. Temperatures are mild and range between an average of 22.98C in February to 16.38C in August (Environmental Resources Management 2002). The coast of Paracas Bay includes sandy beaches, small cliffs, and beaches composed of shells, pebbles, and decaying algal material (Duffy et al. 1981). These beach types are representative of beaches found along the coast of southern and central Peru. We worked where a gentle slope of arid land meets the ocean over a distance of 6 km. The intertidal zone is narrow and subject to a semi-diurnal tidal cycle. Waters enclosed within the bay are calm, because they are protected from marine currents by the northern edge of the Paracas Peninsula. For the purpose of this study, we defined the shore as the mean high tide level. The supralitoral zone is interspersed with series of small sand dunes stabilized by sea purslane (Sesuvium portulacastrum ). No plants are found growing in the adjacent desert. The surface is a mixture of sedimentary and intrusive coarse sand in a matrix of silt and is interspersed with pebbles and boulders of variable sizes and shapes, providing a diversity and abundance of potential retreats for terrestrial consumers. We focused on the nocturnal food web found in the desert that includes algivores that feed on stranded Ulva sp., intertidal predators consuming these algivores, and desert consumers feeding on detritus, terrestrial plants, intertidal consumers, and other terrestrial consumers (Table 1). Abundance of algivores Drift inputs of Ulva in this system are conspicuous along shore, though we did not quantify this input. We observed stranded Ulva every month during 2003, and this alga was often the only macroalga we could find in the intertidal zone. Drift inputs of other seaweeds were
Table 1. Common consumers in the intertidal and nocturnal desert food web at Paracas Bay, Peru. Group Collembolans Silverfish Flies Colydid beetles Darkling beetles Rove beetles Beach hoppers Centipede Mites Solifuge Spiders Scorpion Gecko
Species
Paraxenylla sp. (Hypogastruridae) one unidentified Thysanura species at least three unidentified Diptera species one unidentified Colydiidae species Cordibates fuscus, Psammetichus costatus, Atahualpina peruana (Tenebrionidae) Cafius sp., Carpelimus sp. (Staphylinidae) Orchestoidea spp. (Talitridae) Thindyla litoralis (Schendylidae) one species of Uropodidae and an unidentified Mesostigmata Chinchippus peruvianus (Ammotrechidae) Odo sp. (Zoridae), Sitticus sp. (Salticidae), an unidentified Ctenidae Brachistosternus ehrenbergii (Bothriuridae) Phyllodactylus angustidigitus (Gekkonidae)
very small compared to Ulva throughout the year. Algivores were extracted in April 2003 by washing 30 samples of 100 g of wet Ulva with a solution of water and soap and by filtering the washout on a 20 mm mesh size filter. Algivores were stored in ethanol and later sorted into broad taxonomical categories (ordinal level) using a stereoscope. We estimated biomass (in mg) for each invertebrate group by using the average mass of 20 specimens dried to constant mass. Abundance of terrestrial consumers We surveyed populations of centipedes, solifuges, darkling beetles, scorpions and geckos along the intertidal zone and adjacent desert. We sampled centipedes and solifuges in ninety 1-m2 plots in the intertidal zone. Before adopting this sampling design, we observed that densities of solifuges were higher in the intertidal zone than in the desert, and that centipedes were only found in the intertidal zone. We placed plots at random positions along three transects running along the shoreline. Within each plot, we captured and counted centipedes and solifuges. Means and SE were calculated for the 90 intertidal plots. Scorpions and geckos were counted during the austral summer in four (scorpions) and three (geckos) 150 /150 m plots extending from near shore inland to the desert (distance between plots 12 km). Scorpions and geckos could be located at distances up to 165 m from shore, because the seaward boundary of each plot was traced at approximately 10 15 m from shore. We also sampled an equivalent number of 150 /150 m plots in the desert several km away from shore as control sites without any marine algal input. These sites, similarly to the near-shore plots, lack plants and were interspersed with pebbles and boulders.
We surveyed scorpions (n /2 visits plot 1) by walking through the plots at night with a UVB tube mounted on a portable 20 W lamp. We measured and recorded body length, weight, sex and position of captured scorpions. Scorpions were individually marked with small spots of paint. We surveyed geckos (n /8 visits plots1) by walking through the plots at night and by searching under stranded algae, rocks, and bird cadavers. We recorded gecko snout vent and tail length, weight, sex, and position within the plots. All captured geckos were individually marked by toe clipping. In a separate analysis, we evaluated the seasonal variation in the distribution and relative abundance of geckos in the three plots by surveying populations from February to November 2003. We conducted a repeated measures ANOVA on the number of geckos subdivided in classes of 40-m distance from shore. We considered four two-month periods from February March to September November, and visited each plot four times per period. We sampled darkling beetle (Cordibates fuscus) and solifuge populations along transects extending 100 m inland from shore between 6 11 April 2003. We placed unbaited pitfall traps at 0, 0.1, 1, 10 and 100 m from shore. Pitfall traps consisted of plastic cups 9 cm in diameter and 10 cm deep filled with a mix of water and detergent. Each transect was composed of three lines of pitfall traps (distance between lines 10 m), and we installed one transect in each of the 150 /150 m plots we used for gecko surveys. Means were calculated by pooling pitfall traps from three lines within each plot, and SE based upon variation in mean numbers among the three plots. Stomach content analyses We flushed the stomachs of all geckos captured between JanuaryApril 2003. Each gecko was sampled only once during the study period. We calculated the frequency of occurrence for each species of invertebrate prey, by counting the number of individuals across stomach contents of all geckos. We counted whole prey items (items with no missing body segments) and all identified items (whole items and parts) separately. Identified parts included mouthparts and appendages, and probably overestimated the frequency of hardbodied prey items. We ranked invertebrate prey by frequency of occurrence and compared this ranking to percentage of volume occupied. Volume was estimated for each prey item by using the formula of an ellipsoid: 2 4 1 1 width V p length 3 2 2
77
where length is the greatest distance along the axis of the body, and width the greatest distance perpendicular to the axis length (excluding appendages). Invertebrate prey items were categorized as intertidal or desert items. Intertidal prey were invertebrates that could only be found below the highest tide line; all other invertebrates, including arachnids that were frequent foragers in the intertidal zone, were considered desert items. We used contingency table analysis to determine differential use of intertidal and terrestrial food items between geckos found near shore (B/50 m from shore) and geckos found away from shore (100 150 m away from shore; G test, Sokal and Rohlf 1995). We calculated the G test with data of whole prey items only, and computed the statistics based on (1) number of occurrences of items and (2) volume of items. Stable isotopes Stable isotopes can provide a continuous measure of all energy and nutrients assimilated through trophic interactions by an organism (Fry and Sherr 1984), predict the trophic position of consumers in the food web (Hobson et al. 1994), and have the potential to trace the energy and mass flow through ecological communities (Peterson and Fry 1987, Finlay et al. 2002). There is a large difference in the isotopic signature in marine vs terrestrial producers and consumers (Anderson and Polis 1998), which can be used to evaluate the relative contribution of marine vs. terrestrial sources of energy and nutrients for consumers. Values of d13C and d15N are typically higher for marine organisms and/or for terrestrial plants and consumers of localities with high inputs of marine-derived energy and nutrients than for terrestrial producers and consumers without marine input (Chisholm et al. 1982, Mizutani and Wada 1988, Michener and Schell 1994). We opportunistically collected samples of beach-cast Ulva and invertebrates on several occasions between January and April 2003. Small invertebrates were sacrificed 12 24 h after capture. No samples of invertebrates were obtained from pitfall traps, and no invertebrate was stored in ethanol prior to drying. Scorpions and geckos were captured at night. Scorpions were left alive for at least 24 h to allow their gut contents to be digested and were then sacrificed. We obtained tail samples from geckos to avoid sacrificing animals. All samples were oven dried for 48 h at 608C. All dried samples were ground to a powder and defatted prior to isotope analysis by using a solution of dichloromethane: methanol (9:1) and sonication. We extracted lipids because they are depleted in d13C
78
compared with other tissues, and because the amount of lipids is variable among animal tissues (Peterson and Fry 1987, Kelly 2000). We sonicated samples for 30 min, centrifuged the sample and solvent solution, and eliminated the solvent and lipid portion. We repeated this defatting procedure three times for each sample; defatted samples were oven dried prior to weighing into tin cups for stable isotope analyses. Amphipod samples were treated with dilute HCl to remove carbonate (Connin et al. 1997). Stable C and N isotope analysis were conducted on a Finnigan MAT Delta Plus continuous flow isotope ratio mass spectrometer. We analyzed 3 4 mg of plant and alga samples, and /2 mg of animal samples. Values of d13C and d15N are expressed relative to their respective international standards, Pee Dee belemnite and atmospheric air. We analyzed 16% of our samples in duplicate, and fitted a regression line on duplicated data to determine accuracy of our data. Correlation coefficients were 0.998 for d13C and 0.997 for d15N. We used a two-source mixing model with d13C data to estimate the contribution of Ulva and Sesuvium to the diet of desert consumers. This simple model assumes that Ulva and Sesuvium are the only sources of carbon for all supratidal and desert consumers. This is a fair assumption concerning terrestrial plants, given the aridity of the desert (no plants other than Sesuvium grow in the desert), but may be limiting for algivores, because other types of algae (albeit very rare based on our observations) wash ashore along the beaches of Paracas Bay. Other potentially important sources of carbon in this system that we did not consider in our mixing model were intertidal diatoms, wind-carried spray from the Bay, and wind-carried detritus from the desert. The form of our mixing model was: d13 Cconsumer F (d13 CUlva f Ulva ) [d13 CSesuvium (1f Ulva )] where F is the trophic fractionation between the primary producer and the consumer, and fUlva is the fraction of carbon from stranded Ulva that our model will estimate. We assumed an average trophic fractionation value of 0.39 (Post 2002). We varied the value of F depending on the assumed trophic position of the consumer. In the case of geckos for example, we calculated fUlva for F ranging between 0.78 (/two trophic levels) and 1.56 (/4 trophic levels). We used Phillips and Gregg’s (2001) model to compute standard deviations for our estimates of fUlva . Our approach differs from previous applications of mixing models because we are estimating the relative contribution of two different sources of primary productivity, rather than identifying specific prey species in the diet of terrestrial consumers. We recognize that our approach relies on several assump-
1.0
A Geckos
0.8 0.6 0.4 Proportion of captures
tions: (1) that Sesuvium and Ulva are the two main sources for primary productivity in the system; (2) that fractionation does not vary with trophic levels (i.e. we assume a constant F of 0.39 for each trophic interaction), between terrestrial and marine-derived food matter, or among consumers; and (3) that measured d13C values are representative of the diet of the consumers and are not affected by the tissue sampled. We are confident that our model is representative of energy flow in the system, because of the paucity of sources of primary productivity. Although mean trophic fractionations are known to vary in nature and in laboratory feeding experiments, Post (2002) reported that, based on a literature survey, an average fractionation of 0.39 for d13C is widely applicable. We used d15N values to estimate the trophic position of consumers. We assumed Ulva and Sesuvium as basal resources for consumers. Trophic position was estimated by using the formula (Post 2002):
0.2 0-39 1.0
where l is the trophic position of basal resource (l /1 for Ulva and Sesuvium ) and Dn is the estimated trophic fractionation, assuming a mean fractionation of 3.4 for each trophic level (Post 2002). We used average values (without standard deviations) of d15N for basal resources to simplify calculations. We applied this model to compare the length of food chains supported by the two basal resources, and to strengthen our interpretation of the d13C mixing model.
Results Distribution of geckos and scorpions The average number of geckos in near-shore plots (14.53 geckos) was much higher (DF /2, t /5.92, p/ 0.027) than the average number in desert plots several km away from shore (0.17 geckos). The proportion of geckos captured decreased significantly with increasing distance from shore during the summer (Fig. 1a; DF / 3,84, F /330.33, p B/0.001). Average densities in the first 40 m and from 40 80 m from shore were significantly higher than those beyond 80 m between February and April (Tukey multiple comparison test pB/0.1). On average, 80.6% of the geckos were found between 0 and 40 m from shore; if we break down classes into 10-m distance classes, on average 67.8% of the geckos were found in the first 10 m from shore. The overall number of captures of geckos did not vary with season, but changes in the distribution of geckos across seasons were significant (Table 2;
a
40-79
80-119
120-159
B a
0.8
Scorpions
0.6 ab 0.4 0.2 b
b
0.0 0-39
15
trophic position l(d Nconsumer [d15 Nbasal resource 1 f Ulva d15 Nbasal resource 2(1 f Ulva )])=Dn
a 0.0
40-79 80-119 120-159 Distance from shore (m)
Fig. 1. Distribution of captures of the gecko Phyllodactylus angustidigitus (A) and the scorpion Brachistosternon ehrenbergii (B) in the first 160 m from shore, expressed as a mean proportion of captures (/SE) in 40-m distance classes of all individuals (excluding recaptures) found in 3 (geckos) and 4 (scorpions) plots of 150/150 m in Paracas Bay from February April 2003 (n /8 visits plot1 for geckos, n/2 visits/plot for scorpions). Columns with same letters were not significantly different (Tukey multiple comparisons test, p /0.05).
repeated measures ANOVA, season /distance interaction, DF /27, F /6.45, pB/0.001). A decrease in the average number of captures after the first 40 m was sharpest in February March, AprilMay and September November, and the aggregation near shore was most pronounced in February March (Fig. 2). During the winter months, geckos were more evenly distributed among the four distance classes than they were during the rest of the year. The average number of scorpions was not significantly higher in near-shore vs desert plots (DF /3, t / Table 2. Repeated measures ANOVA on seasonal patterns in gecko distribution, based on 40-m classes of distance. Shown are results from within-subjects analysis for the effect of month periods on distribution. Month periods are February March, April May, June August and September November, 2003.
Season Distance Plot/distance Season/distance Error
DF
MS
F
p
9.00 9.00 18.00 27.00 54.00
1.33 0.55 0.65 1.12 0.44
1.44 148.20 2.32 6.45
0.295 B/0.001 0.077 B/0.001
79
14 Apr-May
Jun-Aug
Darkling beetles Solifuges
Sep-Nov
10 Number of beetless/trap
Number of captures visit
–1
12
8 6 4 2 0
1 2 3 4 1 2 3 4 1 2 3 4 1 2 3 4 Distance from shore (classes of 40m)
Fig. 2. Seasonal distribution of captures of Phyllodactylyus angustidigitus in the first 160 m from shore in three 150 / 150 m plots in Paracas Bay. Columns are mean numbers of captures within a plot in one visit (/SE, n/4 visits plot 1 per time period). All recaptures excluded within each period. Distance classes are (1) 0-39, (2) 40-79, (3) 80-119 and (4) 120-160 m from shore.
2.90, p/0.062). Average number of scorpions per plot ranged from no scorpions in two desert plots to 7 scorpions for a near-shore plot. In near-shore plots, scorpion numbers declined significantly with increasing distance from shore (DF /3, 16, F /10.11, p B/0.001; Fig. 1b). However, the decrease was not as steep as it was for geckos in the period from February to April, and there was no difference in the number of scorpions between 040 m and 4080 m, or among distance classes beyond 80 m from shore. We compared the absolute rate of decrease in abundance as a function of distance from shore between scorpions and geckos by fitting a regression line of abundance on log-transformed distance data. The slopes of the two regression lines differed between geckos and scorpions (DF /1, F /24.94, p /0.001). The absolute rate of decrease was steeper for geckos (slope / /18.61) than it was for scorpions (slope / /1.78). Darkling beetles and solifuges in pitfall traps Darkling beetle populations increased with distance from shore (Fig. 3). Darkling beetles were extremely rare in the intertidal traps (0 and 0.1 m) and in the first desert trap (1 m from shore). In contrast, solifuges were captured in the intertidal zone only, from 0 to 1 m from shore (Fig. 3). Our field observations also support the finding that the distribution of solifuges is restricted to the intertidal zone, because during our nocturnal visits to sample scorpions and geckos in the 150 /150
80
1800
7
1600
6
1400
5
1200 1000
4
800
3
600
2
400
1
200
Number of solifuges trap–1
Feb-Mar
0
0 –3
–2
–1 0 1 Log (distance from shore)
2
3
Fig. 3. Distribution of captures of the darkling beetle Cordibates fuscus and the solifuge Chinchippus peruvianus in three series of transects with pitfall traps at 0, 0.1, 1, 10 and 100 m from shore (n /3 lines of pitfall traps transect 1) in Paracas Bay in April 2003. Number is mean number of beetles found in pitfall traps after a 5-day trapping period (9/ SE).
m plots, we never observed solifuges more than 10 m away from shore. Algivores and consumers in the intertidal zone We captured a total of 107 solifuges and 383 centipedes in 90 plots, for an average density of 1.199/0.14 solifuges m2 (n /90) and of 4.269/0.96 centipedes m2 (n /90). The maximum number of individuals captured within a single plot was 6 for solifuges and 49 for centipedes. In the samples of 100 g of wet algae, the most abundant invertebrate was the springtail Paraxenylla sp. (Fig. 4A). However, invertebrates with a much larger biomass such as beach hoppers and flies were also frequent and averaged 21.6 and 33.3 individuals per 100 g of wet algae. Beach hoppers far outweighed the other groups in terms of biomass (Fig. 4B). The number and type of invertebrates was highly variable among algal samples. Stomach contents of geckos Beach hoppers were the most frequent items and represented over 36% of the total number of food items (n /105 stomach contents). Among the five most frequent prey items, only an unidentified beetle of the family Colydiidae was exclusively terrestrial. After considering stomach contents with whole prey items only (n /85), the ranking of prey items changed only
Log (number of invertebrates)
3.0
A
2.5 2.0
their stomachs (12 of 35 total prey items in 11 stomachs). The volume occupied by intertidal items in the stomachs was 84.1% for all geckos (84.5% for geckos close to shore, 65.6% for geckos away from shore).
1.5 1.0
Stable isotope analysis
0.5 0.0 Sp
r in
1.2 Log (biomass of invertebrates)
Number
gta
B
ils
Mi
tes Flies upae pers rvae dults etes rvae eetle a p a a p ho an l tle a och tle l ing b an ter each ipter bee Olig bee ar kl p i D B D ove D ve R Ro
Biomass
1.0 0.8 0.6 0.4 0.2 0.0
Be
h ac
ho
ers ails ites dults Flies upae eetle rvae etes rvae M a pp r ingt b e la cha n la p e o tl a an ing etl Sp ter ar kl bee Olig ipter be p i e D ove D D v o R R
Invertebrate categories
Fig. 4. Mean number (A) and biomass (B) of invertebrates (/SE) found in samples of 100 g of wet Ulva along beaches of Paracas Bay. Samples (n /30) were collected on different days (1, 3, 18 and 27 days) from initial stranding date (8 May 2003) from previously weighed mats of Ulva enclosed in bags with 5-mm mesh size.
slightly, though colydid beetles occurred almost as frequently as beach hoppers (Fig. 5). When prey volume was examined, intertidal invertebrates far outweighed terrestrial invertebrates: beach hoppers, solifuges and centipedes had much larger volumes than terrestrial colydid beetles (reduced to less than 2% of the volume of all ingested prey items). The diet of geckos varied significantly with distance from shore. Geckos found within 50 m from shore had higher proportion (contingency table, DF /1, G /8.44, p/0.004) and higher volume (contingency table, DF /1, G/44.86, p/0.001) of intertidal prey items than geckos found 100 150 m away from shore. Intertidal prey items were found in 58.8% of all stomach contents. They represented 60.3% of all prey items (176 of 292 total prey items in 86 stomachs) in geckos found close to shore. Geckos found away from shore had 34.3% of intertidal prey items in
The d13C of beach hoppers, centipedes, solifuges and geckos were more similar to those of Ulva than those of terrestrial plants (Fig. 6). The d15N values suggested that the darkling beetle Psammetichus costatus , centipedes, geckos and scorpions were occupying high trophic levels in the intertidal and desert nocturnal food web. Cordibates fuscus darkling beetles and silverfish had isotopic signatures that were intermediate between Ulva and terrestrial plants for d13C and between algivores and top predators for d15N. Large variation in isotopic signatures for silverfish samples suggested that these insects could feed on a variety of sources including detritus, plants and animals. We assumed Ulva to be the main source of food for beach hoppers; however there was a mean difference in d13C of 1.17 between beach hoppers and Ulva . The mixing model based on d13C suggested that beach hoppers, solifuges, centipedes and geckos derived most of their C from Ulva (Table 3). Modifying the trophic fractionation factor (F) from 0.78 to 1.56 for predators decreased the estimated amount of C derived from Ulva , but fUlva did not decrease below 60%. The estimated portion of C that predators receive from Ulva was highest for solifuges, but was nearly similar for geckos and centipedes. A second group of consumers had fUlva below 50%, suggesting that these consumers derived most of their energy from the terrestrial sea purslane. This group included scorpion, darkling beetles and silverfish. High variability in d13C and d15N for scorpions and Cordibates darkling beetles suggested that these consumers may be alternating between terrestrial and marine-derived source of food, and that they may be feeding at several trophic levels. Estimated trophic positions ranged between 1.91 for beach hoppers and 3.75 for Psammetichus darkling beetles (Table 4). Among terrestrial predators, geckos and centipedes had similar trophic positions of /3, whereas scorpions and solifuges had values ranging between 2.5 and 2.6. According to this model, consumers in the food chain mainly supported by Ulva (beach hopper, solifuge, centipede and gecko) occupied three trophic levels, whereas consumers in the food chain mainly supported by the sea purslane (silverfish, Cordibates, silverfish and Psammetichus ) occupied four trophic levels.
81
Beach hopper Colydid beetle Other Solifuge Prey species
Centipede Spider Rove beetle larvae Silverfish Flies Pseudoscorpion Sand Darkling beetle
Occurrence Volume 0
10
20
30 40 Proportion (%)
50
60
Fig. 5. Diet of Phyllodactylus angustidigitus in the desert facing Paracas Bay, based on stomach contents of geckos captured in the first 160 m from shore (excluding recaptures) from February April 2003. Gray bars show percentage of frequency of prey items across all pooled stomach contents (n /337 prey items). Black bars show percentage of volume based on calculations for the volume of an ellipsoid. Frequencies of occurrence and volumes were calculated for stomachs containing whole prey items only (n /85 stomachs).
Discussion The distribution of terrestrial predators supports the idea that marine input is subsidizing their populations.
The concentration of prey in the intertidal zone strongly affects the distribution of geckos in the desert facing Paracas Bay, by driving most of the individuals close to shore. A similar distribution occurs with
22 Psammetichus
20 Scorpion
δ15(N‰)
18 16
Centipede Gecko
Cordibates
Solifuge
Silverfish
14 12
Beach hopper Sea purslane Green algae
10 –24
–22
–20
–18
–16 –14 δ13C (‰)
–12
–10
–8
Fig. 6. Average (mean9/SE, n /5 replicates) d13C and d15N values for terrestrial plants (sea purslane Sesuvium portulacastrum ), marine green algae (Ulva sp.), and terrestrial consumers in the beach and desert facing Paracas Bay from January April 2003. Individual plant and algal samples are photosynthetic parts from several pooled plants or thalli. Each insect and beach hopper sample is 5 10 pooled individuals from the same site on the same date. Each individual predator sample (centipedes, solifuges, scorpions) represents a single whole individual. Each gecko sample represents a single individual from analysis of tail tissue.
82
70
Table 3. Estimated percent of assimilated carbon from the marine green alga Ulva sp. (9/ SD, upper and lower 95% CI in parenthesis) for terrestrial consumers, by using a two-source mixing model and assumed values of trophic fractionation. Taxon
Mean trophic fractionation (F) /0.39
Beach hopper 101.19/5.9% (87.2 100%) Solifuge Does not feed on algae or plants Centipede Does not feed on algae or plants Gecko Does not feed on algae or plants Scorpion Does not feed on algae or plants Cordibates 43.19/2.4% (37.6 48.5%) fuscus Psammetichus 40.19/2.6% (34.4 45.9%) costatus Silverfish 38.79/17.0% (0 86.2%)
/0.78
/1.17
/1.56
Algivore only 97.29/5.8% (83.6 100%)
Algivore only 93.19/5.6% (80.1 100%)
Algivore only 88.99/5.5% (75.9 100%)
79.69/5.1% (68.0 91.3%) 75.59/4.9% (63.8 87.2%) 71.49/4.8% (59.9 82.8%) 72.39/9.1% (49.3 96.0%) 68.59/9.0% (43.5 93.6%) 64.49/9.0% (39.5 89.3%) 44.19/7.3% (23.9 64.4%) 40.09/7.3% (19.8 60.2%) 35.99/7.3% (15.8 56.0%) 38.99/2.3% (33.8 44.1%) 34.89/2.2% (29.9 39.7%) 30.79/2.2% (25.9 35.5%) 36.09/2.5% (30.4 41.6%) 31.99/2.5% (26.4 37.4%) 27.89/33.2% (66.8 77.6%) 34.69/17.0% (0 81.9%)
scorpions, but they are less concentrated at the shoreline than geckos (Fig. 1). The slope for the relationship between abundance and distance from shore is steeper for geckos than for scorpions, suggesting that geckos are more strongly subsidized by marine resources than scorpions. Results from the mixing model also support the inference that scorpions are less dependent on marine resources for their diet than geckos. Both lizards and scorpions have been reported as intertidal predators in previous studies: high densities were calculated for populations feeding upon intertidal crustaceans, such as the scorpion Vaejovis littoralis in Baja California (Due and Polis 1985), or the skink Leiolopisma suteri in New Zealand (Towns 1975). These studies however did not specifically address the variation in relative densities of predators as a function of distance from shore, and the potential effects of the availability of intertidal prey to predators on terrestrial consumers at lower trophic levels. A variety of ecological factors could affect gecko activity patterns. Geckos might reduce their movements between shore and their terrestrial burrows during the
30.59/17.0% (0 77.7%)
26.49/17.0% (0 73.6%)
winter because of lower temperatures. However, the number of captures overall did not decrease with season. Alternatively, the frequency of algae stranding ashore could decrease during the winter, because of greater tidal action removing algae from the intertidal zone. High tides and rough sea typically occur during June August along the Peruvian coast, and these tides can remove stranded material that accumulated over several months. In 2003, the first of a series of high tides occurred the last week of May. Following that event, we observed that the amount of stranded Ulva that covered the beach was greatly reduced. A third explanation for the variation in gecko distribution throughout the year could be variations in the population size of algivores, and especially of beach hoppers. The mixing model indicates that solifuges derive almost all of their energy from Ulva . Abundance surveys indicate that solifuges are restricted to the intertidal zone where Ulva washes ashore. The relative density of solifuges in this zone is extraordinarily elevated, given that these arachnids could face strong predation pressures from both scorpions and geckos, as
Table 4. Estimated trophic positions of consumers in the desert food web, by assuming percent of assimilated carbon from the marine green alga Ulva sp. for different d13C trophic fractionations (0.39 1.56). Trophic positions were not estimated above 2 for algivores (beach hopper) and below 2 for predators. d13C trophic fractionation
Taxon
Beach hopper Solifuge Centipede Gecko Scorpion Cordibates fuscus Psammetichus costatus Silverfish
/0.39
/0.78
/1.17
/1.56
SD
1.91 2.50 3.75 1.92
2.59 3.05 2.93 2.59 2.48 3.74 1.91
2.57 3.03 2.91 2.57 2.46 3.72 1.89
2.55 3.01 2.90 2.56 2.44 3.70 1.87
0.03 0.19 0.11 0.13 0.35 0.14 0.06 0.41
83
well as from conspecifics. Solifuges in the intertidal zone also have to cope with wave and tidal submersion of their hiding places and the risk of being swept into the ocean. An endemic species of solifuge from Namibia is known to also live and forage in the intertidal (Griffin 1998); most solifuge species however inhabit arid habitats and are xerophilic (Punzo 1998). Stomach content analyses of geckos coincide with results from stable isotope analyses, suggesting that observed trophic interactions between geckos and their prey are indicative of energy and nutrient flows within the desert food web. Data on prey frequency of occurrence within stomachs overestimate the relative importance of small terrestrial colydid beetles. When prey volume is taken into account, the three most common items in the diet of geckos are intertidal organisms (amphipods, solifuges and centipedes), and colydid beetles are relegated to a negligible fraction. The mixing model estimating the proportion of carbon assimilated from Ulva calculates high percentage values for beach hoppers, solifuges and centipedes (Table 3, Fig. 6). Our study estimated the length of food chains supported by sea purslane and beach-cast Ulva . In the Ulva -supported chain, consumers fell into a narrower range of trophic levels than in the sea purslanesupported chain. A possible explanation for a longer chain length among organisms depending on the sea purslane is the extent of omnivory in the diet of silverfish and darkling beetles. Desert food webs are known for having complex trophic interactions and high frequency of omnivory (Polis 1991), which could explain both an overall large range in trophic positions among consumers, and a large variation in trophic position among individuals of a given consumer. Our ‘‘top predator’’, the darkling beetle Psammetichus costatus , was only found in dunes covered with sea purslane at the study site. We have observed this beetle feeding on detritus, guano, decomposing animal tissues and invertebrates on guano islands, where Psammetichus are extremely abundant (Catenazzi, pers. obs.). We did not observe this beetle feeding in Paracas Bay; however the elevated d15N suggests a similar scavenging and predaceous diet. Predators depending on Ulva were organisms restricted to a strict arthropodivorous diet, making a switch to lower or higher trophic levels more difficult. There could be several problems in tracing single potential food sources, such as the green alga Ulva , in the desert food web. Variation in the physical and chemical properties of Paracas Bay could induce changes in the isotopic signature of living Ulva . Such variation could occur for natural reasons, such as shifts in the prevailing marine currents entering Paracas Bay and high sediment discharge from the Pisco River (approximately 25 km north of the study site), as well as
84
from industrial effluents from anchovy processing plants and sewage effluents from towns north of Paracas. The extent of such variation is at present unknown. Stranded blades of Ulva could also change their isotopic values as they are sun dried, fragmented by wind, and decomposed by microorganisms. Because we only observed two sources of primary productivity in our study system, we consider our mixing model a reasonable first approximation of the relative contribution of Ulva to the diet of terrestrial consumers. The concentration of scorpions and geckos near shore could exert predatory pressure on terrestrial consumers at lower trophic levels, and could explain the distribution of Cordibates fuscus darkling beetles in the first 100 m from shore (Fig. 3). Polis et al. (2004) reported that, on Gulf of California islands, density of darkling beetles increased two times in cages excluding vertebrate predators compared with densities in control cages. However, we think that darkling beetles overall may also benefit from the input of energy and nutrients from the sea. We frequently observed these beetles feeding on parts of seabird and crab carcasses partially consumed by foxes and gulls. These resources, especially partially consumed crabs that gulls leave in the desert, are usually found more than 10 m from shore. The intertidal zone does not offer many food sources to these beetles, because they do not consume Ulva and do not prey upon algivores. The distribution of these beetles could therefore be a consequence of the availability of marine-derived resources that these organisms can use, rather than an effect of predation from terrestrial predators. Our study adds evidence to a growing body of literature documenting the contributions of marine resources to coastal food webs worldwide (Polis et al. 2004). Our study shows the strength of bottom-up factors for predators occupying mid to high trophic levels in a donor-controlled (sensu Polis et al. 1997) desert food web, and casts light on the factors that determine an organism’s responses to the availability of marine resources. In contrast to other studied systems, especially those in Baja California, our site receives very little rainfall and the amount of precipitation is not affected by El Nin˜o Southern Oscillation events. Vegetation cover is extremely scarce and likely has limited influence on the dynamics of the desert food web. These conditions promote an extreme dependence of terrestrial consumers on marine resources, and cause permanent food-web effects that are affected by temporal variability in marine productivity, rather than patterns of plant growth. Acknowledgements We thank the Paracas National Reserve and the National Institute of Natural Resources in Lima for releasing research and collecting permits. We appreciate the assistance of J. Carrillo and J. C. Jorda´n in the field. R.
Villanueva assisted with laboratory analyses of stomach contents. We thank L.-F. Bersier, D. Hessen, P. Peterson, M. Pilgrim, M. Power and J. Roth for comments on an earlier version of the manuscript. AC was funded in part by grants from the Organization for Tropical Studies, the PADI Foundation, the American Museum of Natural History, Sigma Xi, the American Society of Ichthyologists and Herpetologists and the FIU Graduate Student Association. This is publication number 114 of the Tropical Biology Program at Florida International Univ.
References Anderson, W. B. and Polis, G. A. 1998. Marine subsidies of island communities in the Gulf of California: evidence from stable carbon and nitrogen isotopes. Oikos 81: 75 80. Arntz, W., Landa, A. and Tarazona, J. (eds). 1985. El Nin˜o. Su impacto en la Fauna Marina. IMARPE and GTZ. Barrett, K., Anderson, W. B., Wait, A. D. et al. 2005. Marine subsidies alter the diet and abundance of insular and coastal lizard populations. Oikos 109: 145 153. Bastow, J. L., Sabo, J. L., Finlay, J. C. et al. 2002. A basal aquatic-terrestrial trophic link in rivers: algal subsidies via shore-dwelling grasshoppers. Oecologia 131: 261 268. Bustamante, R. H. and Branch, G. M. 1996. The dependence of intertidal consumers on kelp-derived organic matter on the west coast of South Africa. J. Exp. Mar. Biol. Ecol. 196: 1 28. Bustamante, R. H., Branch, G. M. and Eekhout, S. 1995. Maintenance of an exceptional grazer biomass in South Africa: subsidy by subtidal kelps. Ecology 76: 2314 2329. Chisholm, B. S., Nelson, D. E. and Schwarez, H. P. 1982. Stable carbon isotope ratios as a measure of marine versus terrestrial protein in ancient diets. Science 216: 1131 1132. Connin, S. L., Virginia, R. A. and Chamberlain, C. P. 1997. Carbon isotopes reveal soil organic matter dynamics following arid land shrub expansion. Oecologia 110: 374 386. Craig, A. K. and Psuty, N. P. 1968. The Paracas papers. Studies in marine desert ecology. Occas. Publ., Dept of Geography, Florida Atlantic Univ. 1: 1 196. Cree, A., Lyon, G. L., Cartland-Shaw, L. et al. 1999. Stable carbon isotope ratios as indicators of marine versus terrestrial inputs to the diets of wild and captive tuatara (Sphenodon punctatus ). N. Z. J. Zool. 26: 243 253. Due, A. D. and Polis, G. A. 1985. The biology of Vaejovis littoralis , an intertidal scorpion from Baja California, Mexico. J. Zool. 207: 563 580. Duffy, D. C., Atkins, N. and Schneider, D. C. 1981. Do shorebirds compete on their wintering grounds? Auk 98: 215 229. Duggins, D. O., Simenstad, C. A. and Estes, J. A. 1989. Magnification of secondary production by kelp detritus in coastal marine ecosystems. Science 245: 170 173. Environmental Resources Management. 2002. Estudio de impacto ambiental del proyecto Planta de fraccionamiento de LGN e instalaciones de carga y alternativa can˜eria
submarina en Playa Loberı´o, Pisco, Ica, Peru´ Environmental Resources Management. Farin˜a, J. M., Salazar, S., Wallem, K. P. et al. 2003. Nutrient exchanges between marine and terrestrial ecosystems: the case of the Galapagos sea lion Zalophus wollebaecki . J. Animal Ecol. 72: 873 887. Finlay, J. C., Khandwala, S. and Power, M. E. 2002. Spatial scales of carbon flow in a river food web. Ecology 83: 1845 1859. Fry, B. and Sherr, E. B. 1984. d3C measurements as indicators of carbon flow in marine and freshwater ecosystems. Contrib. Mar. Sci. 27: 196 229. Griffin, R. E. 1998. Species richness and biogeography of non-acarine arachnids in Namibia. Biodiv. Conserv. 7: 467 481. Heatwole, H. 1971. Marine-dependent terrestrial biotic communities on some cays in the Coral Sea. Ecology 52: 363 366. Hobson, K. A., Piatt, J. F. and Pitocchelli, J. 1994. Using stable isotopes to determine seabird trophic relationships. J. Animal Ecol. 63: 786 798. Kelly, J. F. 2000. Stable isotopes of carbon and nitrogen in the study of avian and mammalian trophic ecology. Can. J. Zool. 78: 1 27. Koepcke, H.-W. and Koepcke, M. 1952. Sobre el proceso de transformacio´n de la materia orga´nica en las playas arenosas marinas del Peru´. Publ. Mus. Hist. Nat. ‘‘Javier Prado’’ Ser. A (Zool.) 8: 1 24. Koop, K. and Field, J. G. 1980. The influence of food availability on the population dynamics of a supralitoral isopod, Ligia dilatata (Brandt). J. Exp. Mar. Biol. Ecol. 48: 61 72. Lieth, H. 1978. Primary productivity in ecosystems: comparative analysis of global patterns. In: Lieth, H. (ed.), Patterns of primary productivity of the biosphere. Dowden, Hutchinson and Ross, pp. 300 321. Lindeboom, H. J. 1984. The nitrogen pathway in a penguin rookery. Ecology 65: 269 277. Lindeman, R. L. 1942. The trophic-dynamic aspect of ecology. Ecology 23: 399 418. Loreau, M. and Holt, R. D. 2004. Spatial flows and the regulation of ecosystems. Am. Nat. 163: 606 615. McClelland, J. W. and Valiela, I. 1998. Changes in food web structure under the influence of increased anthropogenic nitrogen inputs to estuaries. Mar. Ecol. Prog. Ser. 168: 259 271. Michener, R. H. and Schell, D. M. 1994. Stable isotope ratios as tracers in marine aquatic food webs. In: Lajtha, K. and Michener, R. H. (eds), Stable isotopes in ecology and environmental science. Blackwell, pp. 138 157. Mizutani, H. and Wada, E. 1988. Nitrogen and carbon isotope ratios in seabird rookeries and their ecological implications. Ecology 69: 340 349. Murphy, R. C. 1925. Bird islands of Peru. Putnam’s Sons. Nakano, S. and Murakami, M. 2001. Reciprocal subsidies: dynamic interdependence between terrestrial and aquatic food webs. Proc. Natl Acad. Sci. 98: 166 170. Odum, E. P. 1957. The ecosystem approach in the teaching of ecology illustrated with sample class data. Ecology 38: 531 535.
85
Onuf, C. P., Teal, M. and Valiela, I. 1977. Interactions of nutrients, plant growth and herbivory in a mangrove ecosystem. Ecology 58: 514 526. Peterson, B. J. and Fry, B. 1987. Stable isotopes in ecosystem studies. Annu. Rev. Ecol. Syst. 18: 293 320. Phillips, D. L. and Gregg, J. W. 2001. Uncertainty in source partioning using stable isotopes. Oecologia 127: 171 179. Polis, G. A. 1991. Complex trophic interactions in deserts: an empirical critique of food-web theory. Am. Nat. 138: 123 155. Polis, G. A. and Hurd, S. D. 1995. Extraordinarily high spider densities on islands: flow of energy from the marine to terrestrial food webs and the absence of predation. Proc. Natl Acad. Sci. 92: 4382 4386. Polis, G. A. and Hurd, S. D. 1996a. Allochthonous input across habitats, subsidized consumers, and apparent trophic cascades: examples from the ocean-land interface. In: Polis, G. A. and Winemiller, K. (eds), Food webs. Integration of patterns and dynamics. Chapman & Hall, pp. 275 285. Polis, G. A. and Hurd, S. D. 1996b. Linking marine and terrestrial food webs: allochthonous input from the ocean supports high secondary productivity on small islands and coastal land communities. Am. Nat. 147: 396 423. Polis, G. A., Anderson, W. B. and Holt, R. D. 1997. Towards an integration of landscape and food web ecology: the dynamics of spatially subsidized food webs. Annu. Rev. Ecol. Syst. 29: 289 316. Polis, G. A., Sa´nchez-Pin˜ero, F., Stapp, P. et al. 2004. Trophic flows from water to land: marine input affects food webs of islands and coastal ecosystems worldwide. In: Polis, G. A., Power, M. E. and Huxel, G.
86
R. (eds), Food webs at the landscape level. Chicago Univ. Press, pp. 200 216. Post, D. M. 2002. Using stable isotopes to estimate trophic position: models, methods, and assumptions. Ecology 83: 703 718. Punzo, F. 1998. The biology of camel-spiders (Arachnida, Solifugae). Kluwer Academic Publishers. Rose, M. D. and Polis, G. A. 1998. The distribution and abundance of coyotes: the effects of allochthonous food subsidies from the sea. Ecology 79: 998 1007. Sa´nchez-Pin˜ero, F. and Polis, G. A. 2000. Donor controlled dynamics on islands: direct and indirect effects of seabirds on tenebrionids. Ecology 81: 3117 3132. Sokal, R. R. and Rohlf, F. J. 1995. Biometry: the principles and practice of statistics in biological research. Freeman. Stapp, P. and Polis, G. A. 2003. Marine resources subsidize insular rodent populations in the Gulf of California, Mexico. Oecologia 134: 496 504. Summerhayes, V. S. and Elton, C. S. 1923. Contributions to the ecology of Spitsbergen and Bear Island. J. Ecol. 11: 214 286. Tarazona, J. and Arntz, W. 2001. The Peruvian coastal upwelling system. In: Seeliger, U and Kjerfve, B. (eds), Coastal marine ecosystems of Latin America. SpringerVerlag. Towns, D. R. 1975. Ecology of the black shore skink, Leiolopisma suteri (Lacertilia: Scincidae), in boulder beach habitats. N. Z. J. Zool. 2: 389 408. Vogt, W. Q. 1942. Aves guaneras. Bol. Compan˜ia Administradora del Guano 18: 1 132. Vo¨ro¨smartry, C. J. and Meybeck, M. 2000. Riverine transport and its alteration by human activities. IGBP Newsletter 39: 24 29.
